Shivkumar V.

Iyer

Modeling and
Python Simulation
of Magnetics for
Power Electronics
Applications
Modeling and Python Simulation of Magnetics
for Power Electronics Applications
Shivkumar V. Iyer

Modeling and Python

Simulation of Magnetics
for Power Electronics
Applications
Shivkumar V. Iyer
Hamburg, Germany

ISBN 978-3-030-96767-3 ISBN 978-3-030-96768-0 (eBook)

https://doi.org/10.1007/978-3-030-96768-0

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland
AG 2022
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface

Every electrical engineer deals with magnetics, whether as mere inductors, or as

transformers, or maybe even with motors and generators. It is hard to imagine any
electrical engineering application without magnetic components. Despite magnetics
being ubiquitous in electrical engineering, a large proportion of electrical engineers
do not have an in-depth understanding of the working of magnetic components. Fur-
thermore, despite the increased availability of documentation related to magnetics
provided by manufacturers, it is still difficult to find documentation that enables an
engineer to gain in-depth understanding of magnetics from these application notes.
One of the greatest challenges to learning about electromagnetism is that it
is a phenomenon that is difficult to perceive. Learning magnetism only through
equations and mathematical analysis will not lead to an understanding of electro-
magnetism. In this book, a visual approach to understanding electromagnetism is
chosen by returning to the fundamental laws of physics and imagining the magnetic
field produced in the magnetic components used in electrical engineering. This book
uses only the basic laws of physics that have been known for close to 200 years to
interpret the operation of magnetics.
This book extensively uses simulation as a learning tool. Moreover, by using the
free and open source circuit simulator Python Power Electronics, the simulations
in the book are universally accessible to a reader in any part of the world. The
book uses the fundamental laws of physics to build simulation models for magnetic
components. Therefore, through the process of simulations, the book builds a bridge
between the world of electrical engineering and electromagnetism. Simulations
allow the reader the opportunity to alter simulation models and examine the effects
without any fear of physical damage that might occur in hardware.
This book will not be considered “heavy” reading. This is intentional, as the
target audience of these books are not just students of power engineering but
also practising electrical engineers. This book uses mainly a process of discussion
and questioning, rather than mere mathematical analysis to describe the concept
of electromagnetism. The book examines in every case different scenarios and

v
vi Preface

variations that help the reader to apply fundamental laws in different conditions
and interpret the results. By going back and forth between simulations and theory,
the book will provide a hands-on learning experience for both students of electrical
engineering and practising engineers.

Hamburg, Germany Shivkumar V. Iyer

Acknowledgements

I would like to thank Janamejaya C for being a loyal follower of the project and his
constant encouragement. I would like to thank Krishnan Srinivasarengan, Awhan
Pattnaik and Hareesh Chitikena for their support and encouragement. I would like
to thank Ramdas Murdeshwar for his continuous support as I developed my online
courses and wrote this book. I would like to thank Prof. Bin Wu for being a
supportive mentor as I embarked on becoming an author of technical books. I would
like to thank all the supporters and followers of the Python Power Electronics
project who have seen this project grow from a mere side project into a full-
fledged educational platform. And my greatest thanks are reserved for Rakhat for
her constant support, encouragement and patience as I spent countless hours on this
book project.

vii
Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Why Another Book on Magnetism? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 The Importance of Magnetism in Electrical Engineering. . . . . . . . . . . . 3
1.3 Back to the Basic Laws of Physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Learning Through Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.5 Outline of the Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2 Presenting Basic Magnetic Circuits with Inductors . . . . . . . . . . . . . . . . . . . . . . 23
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2 Revisiting the Basic Inductor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.3 Inductor Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.4 Magnetic Circuits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.5 B–H Relationship . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.6 Simulating an Inductor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.7 Simulations of an Inductor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.8 Simulating an Inductor with the Magnetic Circuit . . . . . . . . . . . . . . . . . . . 45
2.9 Practical Cores with Air Gaps and Multiple Limbs . . . . . . . . . . . . . . . . . . 50
2.10 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3 Simulating Magnetically Linked Circuits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.2 The Concept of Magnetic Coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.3 The Nature of Magnetic Coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.4 Mutual Inductance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.5 Approach to Simulating Coupled Coils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.6 Dot Polarity of Coupled Coils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
3.7 Simulating Magnetically Coupled Coils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.8 Ćuk Converter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.9 Scalable Mathematical Models with Magnetic Coupling . . . . . . . . . . . . 99
3.10 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

ix
x Contents

4 Modelling and Simulating Transformers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
4.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
4.3 Transformer Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
4.4 Simulating the Transformer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.5 Understanding Transformer Winding Currents . . . . . . . . . . . . . . . . . . . . . . 129
4.6 Transformer Turns Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
4.7 Multi-winding Transformers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
4.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
5 Applications of Transformers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
5.2 Three-Phase Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
5.3 Simulation Model of a Three-Phase Transformer . . . . . . . . . . . . . . . . . . . . 168
5.4 Simulation Results of Three-Phase Transformers . . . . . . . . . . . . . . . . . . . . 175
5.5 High Frequency Transformers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
5.6 Flyback Converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
5.7 Simulation of a Flyback Converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
5.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
6.1 Usefulness of the Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
6.2 A Summary of the Course Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
6.3 The Difference in Approach. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
6.4 Scope for Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
Chapter 1
Introduction

1.1 Why Another Book on Magnetism?

Any reader who glances at the title of this book will immediately ask this question—
“Is there any need to publish another book on simulating magnetics?” Either this
book will contain a collection of simulations already found in many other books or
pulled from the documentation of simulation software that offer ready-made models
for magnetic components. This section will address this question. To answer this
question, I will describe my own experiences as a power engineer that motivated
me to write this book. I strongly believe that many other power engineers have had
similar experiences and this book will be of significant use to the power community.
As undergraduates, all electrical engineers learn about magnetic components
such as inductors, coupled inductors and transformers in basic courses introducing
us to electrical engineering [1, 2]. In these courses, we learn how to analyse
electrical circuits and systems using basic network laws. We also learn how to
represent an electrical circuit or system by an equivalent network, which can then be
analysed using these basic network laws. As we progress to more advanced courses,
we apply our basic knowledge to learn about the construction and working of
electrical machines such as transformers, motors and generators [3, 4]. Therefore, as
undergraduates, we have significant exposure to magnetics in electrical engineering
from a theoretical perspective. Additionally, we also learn the practical aspects of
machines through laboratory experiments.
Despite all these courses, most electrical engineers struggle with a deeper
understanding of magnetic components especially with transformers, motors and
generators. This is partly due to the inherent difficulty in perceiving a phenomenon
like magnetism when the lines of force (or flux lines) of the magnetic field are a bit
tricky to visualize. Nowadays, there are software that allow us to plot the flux lines
of magnetic fields [5]. These software are used extensively when designing electrical
machines such as motors and generators, as accurate mapping of the magnetic field
helps to improve the performance of these machines. These software use fairly

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 1

S. V. Iyer, Modeling and Python Simulation of Magnetics for Power Electronics
Applications, https://doi.org/10.1007/978-3-030-96768-0_1
2 1 Introduction

specialized techniques to arrive at these solutions. Moreover, these software are

usually proprietary and not available in all universities and even if they are available,
they may not be accessible by undergraduates for course work.
Without diving into the details of magnetism, most simulation software through
which one can simulate an electrical circuit offer ready-made simulation models
of most magnetic components such as inductors, transformers, motors and gener-
ators [6–9]. Advanced simulation software also offer simulation models of fairly
advanced machines. Using these software, a student can simulate a circuit that
contains an electrical machine without too much effort. All one has to do is to
drag-and-drop the component from the library of the simulation software into a
circuit schematic. Though simulating circuits with machines is now far simpler,
these simulations do not help to understand the inner working of the machines in
the simulation.
At this point, I would like to recount my experiences and why I needed to relearn
the basics of how magnetism and electrical engineering are interconnected. After
my undergraduate studies, I spent many years as a graduate student when I did
research in power quality improvement and microgrids. I performed simulations
using proprietary simulation software using in-built model of transformers and at
times induction motors. All I had to do was enter the parameters of the transformer
in a dialog box that contained fields for specifying the power rating, voltage
ratings of the windings and also optionally the leakage inductance and magnetizing
inductance. Other than that, I just had to connect the transformer to the rest of the
circuit. I must have performed close to a hundred simulation studies during my
time as a graduate student without ever needing to know about what the simulation
model contained as all I needed was the transformation of voltage and current that
was provided by the transformer component.
Later when I was working in the R&D department of a company, I was
assigned a project where I had to design a 5 kW dc–ac converter that would
charge and discharge a 12 V battery from a 120 V ac grid. This dc–ac converter
would feature a 20 kHz H-bridge converter that was interfaced to the grid through
a high frequency transformer. To begin with, a 5–10 kVA, 12 V:120 V, 20 kHz
high frequency transformer could not be directly purchased. I needed to contact
a few manufacturers to design and fabricate the transformer. One manufacturer was
willing to fabricate the transformer and after a week, submitted the first design for
my approval. And this is when I realized I had absolutely no idea how I could
verify the design submitted by the manufacturer. If I approved the design without
rigorously verifying it, I would be held responsible if the converter did not work and
another transformer had to be fabricated.
Thankfully, I could search the internet for articles on transformer design [10–
15]. I read many articles written by practising engineers, professors and electronics
enthusiasts, and also design notes released by a few companies that manufactured
transformer cores. After several days of reading, I finally managed to assemble
the formulas needed to design the transformer. Though I managed to verify the
manufacturer’s design, I had barely managed to understand why those formulas
were being used and why they were what they were to begin with. This is when
1.2 The Importance of Magnetism in Electrical Engineering 3

I realized that if I ever needed a very special design of a transformer in the future, I
would be in big trouble. One should not have to follow a set of formulas like a set
of commandments to arrive at a design. Since transformers and other magnetics are
being used heavily by power engineers, a deeper knowledge of their working was
fairly important.
In the recent years, after having become an online teacher and author, I have
been gradually revisiting the list of topics that I felt needed more in-depth literature.
In this book, I have written about how one can simulate and analyse magnetic
components like inductors and coupled inductors, and an electrical machine such
as a transformer. The book does not describe the design process of magnetic
components as that is quite often closely tied to the application. Rather, the book
focuses on understanding how magnetics plays a role in electrical engineering using
only the most basic laws of physics that are learned in high school.
How does this book stand apart from other books in this topic? This book
has been written by a power engineer and for power engineers. There are books
on magnetics that will present detailed solutions to map the magnetic field. This
book will not cover these techniques, as for static electrical machines, such a level
of detail of the magnetic field is quite often unnecessary. This book will assume
uniform magnetic fields and linear magnetic properties that are quite often used in
the design of magnetics. This book will describe to the reader how we can build a
bridge between electricity and magnetism, by connecting electrical quantities with
magnetic quantities and formulating expressions between them.
As electrical engineers, what we are familiar with is the voltage that is applied
to energize a system, and the voltage measurable at designated terminals. However,
this book will help a reader to fill in the gaps—how does the magnetic field that is
produced lead to these induced voltages. With code samples in Python and a detailed
description of how simulation models can be gradually developed, this book will
give the reader the tools to learn about the working of magnetic components by
simulation and example. The most important aspect of this book is how it describes
the working of a transformer with several windings, using just a few fundamental
laws of physics, and describes to the reader how these laws can be applied to
understand the operation of magnetic machines.

1.2 The Importance of Magnetism in Electrical Engineering

In the previous section, I described the reason why I wrote this book. In this section,
let us talk about why magnetism is important in electrical engineering. Before we
jump into the importance of magnetism, let us broadly talk about how we perceive
the appliances that we use every day. In these modern times, everyone is surrounded
by a vast number of appliances and systems, and most of us would not realize
the nature of these appliances and where magnetism might play a role in them.
Magnetism might be critical to the functioning of these appliances and in some
cases, any magnetic field produced by them might be a nuisance as well.
4 1 Introduction

For a vast majority of us who do not work in the field, exposure to heavy electrical
equipment is limited, let us therefore begin with consumer appliances. Almost
everyone nowadays has a smart phone and a few other gadgets such as tablets
and laptops. One would not expect a magnetic field to be produced by electronic
appliances such as these, as health and safety guidelines are fairly stringent in how
strong a magnetic field an appliance can produce [16, 17]. Moreover, we expect
these appliances to be adhering to Electromagnetic Compatibility (EMC) standards
that require these appliances to not produce a magnetic field strong enough to disrupt
other appliances [18]. Despite all these requirements, none of these appliances
would function without magnetic components.
Let us consider as an example, our smart phones. Even though we would like
to limit the magnetic field produced by our smart phones to the bare minimum,
the power needed to run the mobile circuits is derived from dc–dc converters that
convert the voltage available from the phone battery to a stable power supply for
the mobile circuits. The fact that our mobiles have been getting thinner and lighter
despite the fact that they have been getting more computation power is the result of
higher density batteries and efficient dc–dc converters. These dc–dc converters will
have at least an inductor if not also a high frequency transformer. These inductors
are magnetic components, and their fundamental basis of operation is the magnetic
field produced in the core of the inductor.
The fact that these mobile phones are EMC approved is due to the shielding
that is provided in all these appliances to ensure that magnetic fields produced are
limited and localized. This is just one example of a circuit in the mobile phone that
uses magnetism to function. This points to the fact that even in appliances where we
do not expect a magnetic field to play a role, due to the power conditioning that is
necessary to supply the main circuits, almost every appliance produces a magnetic
field. Moreover, this magnetic field cannot be said to be a side effect or simply
unintentional, rather, it is critical to its functioning. One could say, however, that
the magnetic field that is produced should remain local to the circuit where it is
necessary and not leak into the surroundings.
Let us now progress from a few sample electronic appliances to the power
adaptors or chargers that they use [19, 20]. A mobile phone can be nowadays be
charged by any Universal Serial Bus (USB) cable that is connected either to a
computer’s USB port or to a dedicated charger plugged into a 240 V, 50 Hz (or
120 V, 60 Hz) domestic power outlet. Let us consider the case of the charger plugged
into the domestic power outlet. Such a charger usually has an ac–dc converter that
converts the 240 V, 50 Hz (or 120 V, 60 Hz) ac voltage to a 5 V dc voltage with which
the battery of the mobile can be charged. There are several ways to implement such
an ac–dc converter. One possible implementation is to use a flyback converter that
comprises of a power device in combination with a high frequency transformer. This
will be described in Chap. 5 along with a simulation. In the flyback converter, the
high frequency transformer plays a critical role in its operation [19, 20]. Therefore,
this is an example of magnetism playing a fundamental role in the operation of the
appliance—the mobile phone charger.
1.2 The Importance of Magnetism in Electrical Engineering 5

We have considered a mobile phone and the charger of the mobile as an example
to describe how ubiquitous magnetic components are in our daily lives. Let us
examine also a few other cases of appliances in our homes and offices. Almost
every home will have several appliances that use electric motors. As an example,
a refrigerator has a compressor that compresses freon gas and pumps it around
the refrigerator compartment to keep the compartment cool. This compressor has
a motor that performs the task of compression and pumping. Depending on the
capacity of the refrigerator, the motor can be of different types—single-phase
induction motor or three-phase induction motor. Every motor uses a magnetic field
to convert electrical energy into mechanical energy. Besides refrigerators, motors
are used in a wide variety of other appliances such as fans, mixers/grinders, washing
machines, dishwashers etc. The motors used in these appliances will differ as the
requirements of the application are different. However, in all cases, the presence of
the motor implies the presence of a machine based on magnetism.
Another application of electric motors has been in the steady adoption of electric
vehicles (EVs). A decade ago, an EV was the possession of a few enthusiasts.
However, in current times, due to the decreasing cost of EVs and the incentives
provided for the sale and purchase of EVs, their penetration in the automobile
market has steadily increased. Many countries have gone so far as to pledge to
eradicate fossil fuel-based cars in a few decades. At the heart of an EV is an electric
motor powered by a battery. Popular batteries for EVs are lithium-ion batteries that
provide cars a range of several hundreds of kilometres on a full charge. The motor
used by the EV once again depends on the type of vehicle. A simple e-bike or e-
scooter could use a simple single-phase induction motor, while a full-size Sports
Utility Vehicle (SUV) could use a three-phase induction motor or a three-phase
synchronous motor. In an EV, magnetism plays a role not only in the motor but also
in the converter that supplies this motor and controls the energy needed by the EV
to accelerate and decelerate.
Now that we have described a few applications that a non-engineer might
be exposed to, let us describe a few industrial applications that are specific to
engineering. In industries, close to 70% of the load is considered to be some
form of motor load. As an example, motors are used for applications such as
crushers, rollers, mixers, centrifuges etc. in industries such as chemical factories,
iron and steel factories, pharmaceutical factories, cement factories etc. These motors
can range from a few kilowatts (kWs) to several megawatts (MWs). The type of
motors used also varies greatly depending on the application. A vast majority of
motors used are induction motors with squirrel cage motors being preferred due to
their robustness and low maintenance. Wound-rotor induction motors, synchronous
motors and switched reluctance motors are several other types of motors used. In
an industrial setting, one cannot imagine any industry functioning without motors.
This points to how critical magnetic machines are for industry.
Now that we have started our discussion about industrial applications, let us talk
about the power system that makes it all possible [21]. More than a hundred years
back when industries began to be electrified, every factory had its own dedicated
generator. From those rudimentary beginnings, we have now reached the modern
6 1 Introduction

power system that is a huge network of transmission lines connecting generating

stations scattered over huge landmasses and powering industries and homes in cities
over entire continents. One of the most critical components that makes it possible
to have such a diverse power system is the transformer, which is described in
detail in Chap. 4 [13]. The transformer enables us to raise or lower the voltage
by interfacing isolated electrical systems through a magnetic field. This makes it
possible to generate power at 690 V using a thermal generator and transmitting it to
cities hundreds of kilometres away at hundreds of kilovolts (kVs).
An electrical engineer may find himself or herself working in different domains.
He or she could be designing power supplies for consumer appliances, or could
be working in motor controls, or could be working in a generating station.
Wherever an electrical engineer ends up, he or she is never far away from magnetic
components or magnetic machines. Therefore, to understand the basics of how
magnetic components and magnetic machines work is essential for any electrical
engineer. The detail to which an engineer wishes to learn will vary according
to his or her position and the job requirements. However, a basic knowledge of
magnetism will help in every case. The purpose of this book is to help an engineer
translate fundamental laws of physics into the world of electrical engineering and
learn through the process of simulation the role that magnetism plays in electrical
engineering. To make these initial steps easier, this book will only consider static
magnetic components such as inductors and coupled inductors, and static electrical
machines such as transformers.

1.3 Back to the Basic Laws of Physics

The objective of the previous section was to motivate an electrical engineer to learn
how magnetism plays a role in electrical engineering either through the magnetic
components used in circuits or in the machines that are in use. As already stated in
the first section of this chapter, this topic has already been dealt with in dozens if
not hundreds of books and other forms of literature. This section will describe the
approach used in this book and how that differs from the approach taken in prior art.
This book will describe the working of magnetics through simulations, by
formulating simulation models of these magnetics using the basic laws of physics.
It is important to emphasize that all engineering analysis in general is a process
of expressing physical laws using mathematical equations such that these may be
solved and analysed. Therefore, the basic laws of physics are the building blocks
of all engineering as we know it. Let us clarify what we mean by the title of this
section “Back to the Basic Laws of Physics”. During our studies, most engineers
learn new concepts only through mathematical equations. Unfortunately, most of us
normal human beings do not learn through equations but rather either visually, or
through discourse, or through a process of questions and answers. Therefore, using
a purely mathematical approach to learning about magnetism will lead to gaps in
understanding how things “work under the hood”.
1.3 Back to the Basic Laws of Physics 7

In this book, all discussions and analysis of simulation results will always refer
to the basic laws of physics rather than engineering equations that have been
derived from these. In most cases, engineering equations that are fairly concise
as a result of being simplified can be derived from basic physical laws in just a
few steps. Therefore, a repeated reference to the basic laws of physics will not
increase the burden in creating simulation models. Rather, the application of these
basic physical laws and the discourse surrounding their application will lead to a
deeper understanding of engineering processes. Besides mathematical expressions
and a detailed discussion on them, this book will also use simulations as a tool to
accompany learning. The next section will describe the usefulness of simulations.
If one wants to learn about static magnetic components such as inductors or
coupled inductors, and static electrical machines such as transformers, a fairly robust
understanding can be built using Kirchhoff’s Voltage and Current Laws, Ampere’s
Law, Faraday’s Law and Lenz’s Law. If we were to progress to moving (or rotating)
machines such as motors and generators, a few additional physical laws are needed
to understand how mechanical force is produced (motors) or results in electrical
energy being produced (generators). However, in this book, we will cover only the
modelling and simulation of inductors, coupled inductors and transformers. In this
section, we will describe these laws that will be used, such that they can be used
directly in the later chapters of this book.
Kirchhoff’s Voltage and Current Laws (KVL and KCL) also knows as the
Kirchhoff’s network laws were described by Gustav Kirchhoff in 1845 [1, 2, 22].
These laws became fundamental laws in circuit analysis and it is hard to imagine
electrical engineering without these laws as they are used by every electrical
engineer to solve circuits. These laws are applicable to circuits with lumped
elements. A lumped element is when an element such as a resistor, inductor or
capacitor is represented as an element between terminals across which a definite
law can be applied representing the element as a parameter. As an example, the
following law is applicable for resistors:

v = iR (1.1)

where v is the voltage across the resistor terminals, i is the current through the
resistor and R is the resistance of the resistor.
We have represented a component by a single parameter, namely the resistance
R. However, a resistor can be manufactured in several different ways and using
many different materials. Depending on the construction and the nature of materials
used, the actual relationship between the voltage v across the terminals and the
current i flowing through it may not be representable by a single parameter R. As
an example, there will be a maximum current that can be allowed to flow through
the resistor. If a current larger than this maximum value flows through the resistor,
it will be damaged. If however, we allow a current to flow through the resistor that
is fairly close to the maximum limit, the above relationship may not hold true, as
a certain degree of damage might already have begun. If one wishes to represent
other physical phenomena that occur within the resistor, the above law will not be
8 1 Introduction

Fig. 1.1 Kirchhoff’s Current i1 i2

Law applied at a node in a
circuit
i3

applicable. Subsequently, Kirchhoff’s laws that use lumped parameters such as the
resistance R above will not also be applicable if a circuit does not have lumped
elements, but rather, we must express detailed physical phenomena within each
element of the circuit.
For most engineering analysis, assuming lumped elements in a circuit is quite
reasonable. With this assumption, Kirchhoff’s Current Law (KCL) can be stated
as—“the algebraic sum of currents in a network of conductors meeting at a point
is zero” [1, 2, 22]. Figure 1.1 shows how KCL can be applied in a circuit. In a
circuit, the meeting point of more than two conductors in a circuit is called a node.
Figure 1.1 shows three conductors meeting at a node. As per KCL, the following
expression can be written for the currents i1 , i2 and i3 flowing in the three branches
meeting at the node:

i1 − i2 + i3 = 0 (1.2)

Let us arrive at the above result step by step. KCL can be applied at every node
in the circuit. In order to apply KCL at a particular node of the circuit, we must
assume directions of currents in all the branches that are meeting at the node. While
performing circuit analysis, we need to assume directions of currents in every branch
of the circuit. If the current in reality is flowing in a direction reverse to what has
been assumed, it only means that current will have a negative value. Once we assume
directions of all currents in the branches meeting at a node, we consider the current
entering at a node to be a negative value and the current leaving a node to be a
positive value. KCL states that the sum total of the current entering a node must
be equal to the sum total of the current leaving a node. In other words, all the
charge that enters a node must leave it, or else there will be charge accumulating
at a node. In Fig. 1.1, we could have assumed all the currents entering the node or
even all the currents leaving the node. Any random assumption of current directions
is acceptable as long as we consider the appropriate sign while writing the KCL
equation.
As KCL can be applied at every node in a circuit, we can write N equations for
a circuit with N nodes. Typically, one node in a circuit will be a reference node
and, therefore, only N − 1 equations will be independent, and the Nth equation can
be expressed as a linear combination of the remaining equations [1]. If the number
of branches in the circuit is B, there will be B current variables. Therefore, using
KCL, we can write N − 1 equations in B variables for a circuit with N nodes and
B branches. These equations can be solved simultaneously to calculate the value of
the currents in the circuit. Additionally, if the elements in a branch are known, one
1.3 Back to the Basic Laws of Physics 9

R1 i2
Fig. 1.2 Kirchhoff’s Voltage
Law applied in a closed loop
in a circuit
i4
+
v1
− R2
i1 i3

+ −
v2

can express the current in a branch as an expression with respect to the voltages of
the nodes at the terminals of the branch. To learn about the various techniques by
which KCL can be used to analyse a circuit, the reader is encouraged to read the
references [1, 2, 22]. KCL is a very powerful technique in circuit analysis and is
used extensively by electrical engineers.
While KCL can be applied at every node in a circuit, Kirchhoff’s Voltage Law
(KVL) can be applied in every closed loop of a circuit. The statement of KVL is as
follows—“the directed sum of the potential differences (voltages) around any closed
loop is zero” [1, 2, 22]. Figure 1.2 shows KVL being applied in a closed loop in a
circuit. To begin with, let us define a closed loop in a circuit. As shown in Fig. 1.2, a
closed loop in a circuit is a closed path that we can trace if we were to start at any one
node in a circuit, progress through at least one branch in the circuit and return back
to the same node where we began. Quite obviously, for a circuit with several nodes
and branches, one could trace several such closed loops if we began at different
nodes, passed through different branches connected to the node and eventually find
our way back to the same node we started at. In each such closed loop of the circuit,
we can apply KVL.
Now let us examine how we can apply KVL in a loop with the loop shown in
Fig. 1.2 as an example. We will assume a loop in a particular direction as shown in
Fig. 1.2 by the arced arrow inside the loop. A loop can have voltage sources such as
v1 and v2 as well as passive elements such as the resistors R1 and R2 . The currents
flowing in the branches have been marked as i1 , i2 , i3 and i4 and are either known
values or merely unknown variables with assumed directions of currents as in the
case of KCL before. In order to apply KVL, we need to traverse the loop in some
direction. Let us choose a direction as shown in Fig. 1.2 by the arrow inside the loop.
When progressing along this loop, any voltage we encounter will be considered as
a positive value equal to the value of the voltage if we progress from the negative
terminal to the positive terminal and will be considered as a negative value equal
to the value of the voltage if we progress from the positive terminal to the negative
terminal. Therefore, from Fig. 1.2, voltage source v1 will be +v1 , while voltage
source v2 will be −v2 . For passive elements, we will need the relationship between
the voltage across the terminals and the current through the element.
We can begin with applying KVL for resistors shown in Fig. 1.2 and then
discuss the possibility of other elements. When traversing the closed loop in a given
10 1 Introduction

direction, the voltage drop across a passive element will be positive if we traverse
the element in a direction opposite to the direction of the current in the branch and
will be negative if we traverse the element in the same direction as the direction of
current in the branch. With this logic, the voltage drop across R1 is −i2 R1 , while
the voltage drop across R2 is +i4 R2 . As per KVL, once we have determined the
voltage drops with their polarities, we can assign the sum of these voltage sources
and voltage drops to be zero. This results in the following equation:

v1 − i2 R1 + i4 R2 − v2 = 0 (1.3)

Just like KCL was an application of the conservation of charge at a node, KVL
is an application of the conservation of voltages in a loop. All that KVL states is
that the sum total of voltage produced in a loop is equal to the sum total of the
voltage drops across all elements in the loop. This only makes sense, as the voltage
produced by the sources in a loop has to be dropped across other elements in the
loop. When I was an undergraduate, I was confused about one aspect of KVL—if
one can imagine numerous closed loops in a circuit, will KVL hold true for each
and every closed loop imaginable? Quite surprisingly, the answer is yes. For every
closed loop in a circuit, no matter how roundabout a path we choose until we return
to the origin node, KVL will hold true.
This might seem almost magical and difficult to believe, but a simple way to
interpret it is as follows. Instead of the voltages across each element, let us compute
the voltage across each branch in a closed loop such that the voltage across the
nodes of a branch is equal to the sum of voltage sources minus the sum of voltage
drops across passive elements in that branch. Every branch will have a non-zero
voltage across it unless it has been short-circuited. If we consider another branch
that is incident at one of the nodes of this branch, and we add the voltages across the
two branches together, we will still result in a non-zero voltage as long as the two
free nodes of the branches are not short-circuited. However, if we choose any closed
loop in a circuit and add branches in the loop one after the other until we close the
loop, the net voltage will be zero. This is because the net voltage will be the voltage
of the starting node with respect to itself, as the definition of a closed loop is that it
ends with the starting node.
Figure 1.2 shows only resistors, but KVL can also be applied if there are other
elements with lumped parameters. Let us replace the resistor R1 with an inductor
L1 , and the resistor R2 by a capacitor C1 . Since we are considering elements with
lumped parameters, we can write the following expressions for the voltage across
the inductor and capacitor [1, 2]:

di2
vL = L1 (1.4)
dt

1
vC = i4 dt (1.5)
C1
1.3 Back to the Basic Laws of Physics 11

The KVL expression for the closed loop can then be expressed as:

di2 1
v1 − L1 + i4 dt − v2 = 0 (1.6)
dt C1

KVL is a very powerful circuit analysis technique that is used as widely as

KCL. As a matter of fact, KCL and KVL are used quite often together to analyse
complex circuits. In this book, KCL and KVL will be used many times to express
the voltage applied across a magnetic element with respect to the voltage drop
across resistances, inductances and the emf induced in the magnetic element. These
equations will be found in every chapter of the book—Chap. 2 on inductors,
Chap. 3 on coupled inductors, Chap. 4 on transformers and Chap. 5 on transformer
applications. Now that we have covered the basic network laws that will be used in
this book, we can describe the physical laws related to magnetism that will be used
in this book.
As we start discussing the basics of electromagnetics, it is important to emphasize
that electromagnetics and magnetism in general are very vast domains with a
multitude of different laws. However, this book is for a power engineer who wishes
to learn about the inner workings of magnetic components. The objective of this
book is only to develop a deeper understanding of magnetic components through
simulations. The laws used in this book will be only the basic laws that result in
functional simulation models. This book will not cover electromagnetism to a great
depth, nor will it attempt to understand the finer details of magnetic fields.
The bridge between electricity and magnetism was probably first demonstrated in
1821 by Hans Christian Oersted, who showed that a conductor carrying an electric
current produces a magnetic field [23, 24]. This was demonstrated by an experiment
where he observed that the compass of a magnetic needle turned and pointed in
a direction perpendicular to the current carrying wire. He also observed that the
lines of flux of the magnetic field encircled the wire and that these lines of flux
were in a plane perpendicular to the wire. Furthermore, the strength of the magnetic
field decreased with distance from the wire and also reversed if the direction of
current through the wire reversed. After Oersted, several other mathematicians and
physicists worked on this phenomenon, and several laws resulted—Ampere’s force
law, Ampere–Maxwell law and many others [23]. Without diving very deep into
what happened two hundred years back, let us concisely list out what we can use as
electrical engineers.
One of the simpler visual implementations of many of the laws that resulted
from the magnetic field produced by a current carrying conductor was the Right
Hand Thumb Rule, which was a combination of Oersted’s observations and work
done by Ampere [1, 22, 24]. The Right Hand Thumb Rule states that if we were to
hold a current carrying conductor by our right hand with the thumb outstretched
and pointing in the direction of the current, the direction in which our fingers
will encircle the conductor will be the direction of the flux lines of the magnetic
field. Figure 1.3 shows the flux lines of the magnetic field for a current i flowing
in a conductor in the direction shown. If the reader were to imagine holding this
12 1 Introduction

Fig. 1.3 Right Hand Thumb i

Rule

conductor with his or her right hand, with the thumb outstretched and pointing
upwards in the direction of current, the direction of the magnetic field will be in
the direction shown, which is the same as the direction in which the fingers encircle
the conductor. The reader must note that the magnetic lines of flux will be in the
plane perpendicular to the conductor as shown by the dashed rectangle. These lines
of flux will extend indefinitely radially outwards and will be continuous throughout
the length of the conductor.
This Right Hand Thumb Rule is applicable to any current carrying conductor,
even when it has been twisted or wound as a coil as will be discussed soon. This
book extensively uses this rule to determine the direction of the magnetic lines of
flux. Besides being able to determine the direction of the magnetic lines of flux,
Ampere’s Law is extremely useful in expressing the strength of the magnetic field
with respect to the current. The following is the expression of Ampere’s Law as
borrowed from physics [22, 24]:

H dl = iencl (1.7)
c

where H is the strength of the magnetic field. iencl in the case of a simple current
carrying inductor in Fig. 1.3 is merely the current i flowing through the inductor.
Soon, we will examine when iencl is different from the current i. The integral
notation needs a detailed
 expression.
The notation c implies the integral over a closed curve c. In the above
expression, we are integrating over one particular closed loop of the magnetic field
comprising of one flux line. This closed loop has a magnetic field strength H . As we
move radially away from the conductor, the radii of the closed loops of the magnetic
flux lines will increase. Therefore, as we move radially away from the conductor,
we are integrating over loops of increasing larger lengths. However, every integral
of a loop of a magnetic flux line will be equal to the current that is enclosed by
the loop—namely i flowing through the conductor. Therefore, as we move radially
away from the conductor, with increasing length of the loops, the magnetic field
strength H will decrease. Therefore, the above expression relates the magnetic field
strength with the distance from the conductor.
A single straight conductor as shown in Fig. 1.3 is not of any particular interest
to an electrical engineer. This is due to the fact that the magnetic field produced
1.3 Back to the Basic Laws of Physics 13

Fig. 1.4 Magnetic field due

to the current through a coil
i i
φ
φ

φ
φ

by a single straight conductor carrying a current in the surrounding air will not
be significant to put it to any real use. An exception would of course be if we
were to examine the magnetic field strength in the surrounding of a high voltage
transmission line carrying thousands of amperes of current. In that case, we would
need to determine the strength of the magnetic field around the line so as to define a
safe distance on either side of the transmission line for the sake of safety to humans.
Things begin to get interesting when we take a conductor such as the one in Fig. 1.3
and wind it as a coil.
Figure 1.4 shows the magnetic field produced by a similar conductor as in
Fig. 1.3, but now wound in the shape of a coil. All the laws are applicable in this
case as well. We can apply Right Hand Thumb Rule to determine the direction of
the flux lines of the magnetic field. Since the conductor is wound as a coil with
several turns, we can apply the rule to each turn. As a result, there are flux lines of
the magnetic field passing through the coil. Though we can apply the Right Hand
Thumb Rule to determine the direction of the field, there is another convenient rule
that is easier to apply in the case of such coils and is an offshoot of the Right Hand
Thumb Rule. The flux lines of the magnetic field will enter the coil face where the
current is flowing in the clockwise direction and will leave the coil face where the
current is flowing in the counter-clockwise direction. The reader is encouraged to
verify this by applying the Right Hand Thumb Rule.
The flux lines of the magnetic field that are passing through the coil will complete
their paths through the exterior of the coil as shown in Fig. 1.4. Similar to the case
of the single conductor, the flux lines will extend in all directions radially away from
the coil. As we move farther away from the coil, the strength of the magnetic field
will decrease. We can apply Ampere’s Law in this case as well, though with a very
interesting twist. The integral over a closed loop of the magnetic flux line of strength
H will be equal to the enclosed current. However, in the case of a coil with turns
as shown in Fig. 1.4, the enclosed current is no longer merely the current i flowing
through the conductor. If the coil has N turns, the enclosed current is Ni leading to
this expression [22, 24]:

H dl = iencl = Ni (1.8)
c
14 1 Introduction

I struggled for many years to understand this and for that reason, I would like
to devote a detailed explanation for the above expression. If one applies the Right
Hand Thumb Rule to the coil in Fig. 1.4, it is fairly easy to figure out the direction of
the magnetic field. However, it is important to note that we can apply the Right Hand
Thumb Rule to every turn of the coil. Therefore, each and every turn of the coil is
producing a magnetic field. As a result, there is a cumulative magnetic field due to
the current flowing in all the turns of the coil. An accurate map of the magnetic field
that can be determined from specialized software is very complex and will include
all possible non-linearities and non-uniformities and will not be as neat and clean as
shown in Fig. 1.4. However, if we consider only the simplest effects, it is fairly easy
to see that the magnetic flux lines due to the current through each turn assist each
other especially in the flux lines that pass through the coil. There are other places
where there may be a conflict between the flux lines produced by different turns. By
and large, by winding a conductor as a coil with N turns, it appears as if the strength
of the magnetic field that passes through the coil will be N times stronger than the
current through a single straight conductor. The reader is encouraged to verify this
visually.
By winding a conductor as a coil with N turns, we seem to have magically
enhanced the magnetic field passing through the coil. But the effect of the magnetic
field produced by the consecutive turns of the coil assisting each other is very clear.
If we use Ampere’s Law, we can interpret it as—a single flux line of the magnetic
field will enclose N turns of the coil each carrying a current i. Quite obviously,
the strength of the magnetic field H will be much stronger in the case of a coil as
opposed to a single straight current carrying conductor. This leads to the concept—
ampere turns. When we wind a coil with N turns, we define the ampere turns of
the coil as N i [22, 24]. These ampere turns are now responsible for producing the
magnetic field and not just the current i as was the case for the straight current
carrying conductor.
Since this concept is fairly important and will be used throughout the book,
let us state precisely how a reader can interpret these laws. In this book, we will
mainly be dealing with coils either in the context of inductors, coupled inductors
or transformers. Any coil carrying a current will produce a magnetic field. As per
Right Hand Thumb Rule, the flux lines of the magnetic field will pass through
the coil. This follows from Oersted’s findings in 1821 that the flux lines of the
magnetic field will encircle the current carrying conductor [23]. One can determine
which face of the coil will be the South pole and which will be the North pole by
checking whether the current flows through the coil in that face in the clockwise or
the counter-clockwise direction, respectively. Subsequently, we can determine the
direction of the flux lines of the magnetic field as they enter the coil at the South
pole and leave at the North pole—this is the case with any magnet.
The flux lines of the magnetic field for a current carrying conductor wound as a
coil are shown in Fig. 1.4. This is merely mapping the flux lines that pass through
the coil with closed loops in the surrounding air. For each flux line, we can apply
Ampere’s Law. The integral of the magnetic field strength H over the closed loop of
a flux line will be equal to the enclosed current, which will be equal to the ampere
1.3 Back to the Basic Laws of Physics 15

turns of the coil, i.e. the product of the current in the conductor and the number
of turns of the coil. This ampere turns is therefore the driving force that produces
the magnetic field and is an extremely important term used throughout the book. In
Chap. 2, we will call the ampere turns as Magnetomotive Force (MMF) [25].
Now that we can describe the magnetic field produced by a current flowing
through a conductor, which may either be straight or wound as a coil, we can now
progress to describing what will be the effect of a magnetic field on a circuit. It
should be noted that the magnetic field in this case can be produced by the circuit
itself such as a current carrying coil as described above in some part of the circuit
or can be an external magnetic field that has no direct relation to the circuit. As
stated before, the effects of a magnetic field on a circuit may be very complex
and governed by several phenomena that have been studied over the past century
by many physicists. However, as electrical engineers trying to build a simulation
model of a magnetic component, we can limit ourselves to just a few phenomena
and ignore the rest.
The phenomenon of electromagnetic induction was first discovered and reported
by Michael Faraday in 1831 [23]. Faraday’s Law of electromagnetic induction states
that “the electromotive force around a closed path is equal to the negative of the time
rate of change of the magnetic flux enclosed by the path” [1, 2, 22, 24]. Let us break
this statement up and figure out how we are going to use it. The closed path in the
law is any closed loop in a circuit [1, 2]. If we consider a single straight conductor
of a finite length, which is open at both ends, this would not qualify as a closed
loop, as quite obviously, it is not closed but open at both ends. Suppose we connect
a Voltmeter across the two ends of the conductor, we have a closed loop as shown
in Fig. 1.5. This is the simplest possible conception of a closed loop for the above
law. However, any circuit with at least one closed loop through which current can
flow will also fit the above definition.
Before we talk about the “negative of the time rate of change”, in simplest terms,
if the magnetic field associated with (enclosed by) such a closed path were to change
in any way, an electromotive force (emf) will be induced in that closed path. If were
to imagine a magnetic field with flux lines passing through the closed path—and
this could be the simple connection of a Voltmeter across a conductor—and this
magnetic field were to change in any manner, an emf will be induced in the path.
This emf induced will cause the Voltmeter to indicate a voltage measurement. A
magnetic field may be widespread as we have seen in our discussion above of the
magnetic field produced by a current carrying coil with the lines of force spreading

Fig. 1.5 Electromagnetic

induction
V
16 1 Introduction

far and wide even though the strength of the field will decrease as we move away
from the coil. This is why the law states the magnetic flux enclosed by the path is
what matters, i.e. the flux lines of the magnetic field passing through the enclosed
path.
An emf will be induced in the closed path only if the magnetic flux enclosed by
the path changes. If there is no change, there will be no induced emf. Furthermore,
the law states that the emf induced is equal to the negative rate of change of the
magnetic flux enclosed by the path. For a moment, let us drop the term “negative”.
The magnitude of the induced emf will be equal to the rate of change of magnetic
flux passing through the closed path. Therefore, the more rapidly the magnetic flux
passing through the closed path changes, the greater will be induced emf. For a
single conductor with a Voltmeter connected across it as shown in Fig. 1.5, the
magnitude of the induced emf can be expressed mathematically using Faraday’s
Law as:
dφ
e= (1.9)
dt
where e is the magnitude of the induced emf and φ is the flux in Weber passing
through the closed loop.
To delve further into the “enclosed” term in Faraday’s Law, let us look at our
setup of a single conductor with a Voltmeter across it as shown in Fig. 1.5. The
magnetic flux passing through the closed loop is the magnetic flux enclosed by the
closed loop. No difference between them and thus the expression above. Things
again get interesting if we consider the magnetic flux passing through a coil. In our
discussion before, we had considered a current carrying coil and had described the
magnetic field produced by it. Now let us imaging the reverse as shown in Fig. 1.6.
A coil with no current is in a magnetic field and we are connecting a Voltmeter
across the two free terminals of the coil. We thus have a closed path. Suppose, for
some reason, the magnetic field to which the coil is exposed as shown in Fig. 1.6 is
identical to the magnetic field shown in Fig. 1.4. It is to be noted that we are now
assuming no current flowing through the coil. The magnetic field is produced by
some external agent.

Fig. 1.6 Induced emf in a V

coil placed in a magnetic field

φ φ

φ
φ
1.3 Back to the Basic Laws of Physics 17

If the flux lines of the magnetic field passing through the coil were to change, an
emf will be induced. This is due to the fact that the coil is connected to the Voltmeter
externally and is a closed path. Even if we were to neglect the flux passing through
the part of the circuit with the Voltmeter connected across the coil terminals, the
flux passing through the coil is still considered to be the flux enclosed by the closed
path. By Faraday’s Law, the magnitude of the induced emf will be equal to the rate
of change (negative to be explained soon) of the flux enclosed by the closed path.
In this case, if we assume that the only flux associated with the closed path is the
flux passing through the coil, the emf induced will be equal to the magnetic flux
enclosed by the coil. And here, because we have a coil, there is difference between
the flux passing through the coil and the flux enclosed by the coil.
For a coil with several turns, one could imagine each turn of the coil impacted
by the changing magnetic flux passing through the coil. The emf induced across the
terminals of the coil is the sum total of the emf induced in each turn of the coil.
Therefore, the enclosed magnetic flux is not merely the flux φ but is now multiplied
by the number of turns N of the coil. Once again, by winding a current carrying
conductor as a coil with N turns, we have enhanced the emf induced in the closed
path by a factor of N. We can express the magnitude of the induced emf for the
circuit of Fig. 1.6 as follows:

dφ d dψ
e=N = (Nφ) = (1.10)
dt dt dt
In the above equation, the term Nφ is called the flux linkage ψ. The flux linkage ψ
of a coil is the product of the flux passing through the coil and the number of turns
of the coil and is an indication of the impact of enclosed flux on a coil with more
than one turn.
It is fairly easy to see the analogy between ampere turns and flux linkage. In
both of them, the number of turns N of the coil appears to take into account the
fact that a coil wound with more than one turn will behave differently from a single
straight conductor. Therefore, whenever we have a coil with more than one turn, if
we are trying to calculate the strength of the magnetic field from the current flowing
through the coil, we must calculate the ampere turns Ni and not just consider the
current i. Similarly, if we are trying to compute the emf induced in a coil, we must
consider the flux linkage ψ = Nφ and not just the flux φ passing through the coil.
Now that we have expressed the magnitude of the induced emf, let us talk
about the “negative rate of change” that we have so far ignored. The complete
mathematical expression of Faraday’s Law is as follows:

dψ
e=− (1.11)
dt
To understand this negative sign, one needs to ask the question—what will be the
polarity of the induced emf? Rather than using the negative sign in Faraday’s Law,
if we wish to determine the polarity of the induced emf, we can use another law—
18 1 Introduction

Lenz’s Law. Lenz’ Law was formulated by Emil Lenz in 1834 [23] and states that
the emf induced due to the change in magnetic flux enclosing a closed path will be
such that it will oppose the cause that produces it, i.e. the change in magnetic flux
[22, 24]. It is very similar to Newton’s Third Law of Motion, which states that the
reaction to an action will be opposite to the action.
This book will use Lenz’s Law in combination with Faraday’s Law to determine
the magnitude and polarity of the induced emf. Though Lenz’s Law is fairly easy to
use, we must take into account a fine detail in the law. The induced emf will oppose
the cause that produces it. The cause that produces the induced emf is the changing
magnetic flux. Therefore, the induced emf will oppose the changing magnetic flux.
The induced emf will not oppose the magnetic flux. This is a distinction one must
remember throughout the book and will be repeated over and over again. As an
example, a constant magnetic flux will not induce an emf. Therefore, even though
there exists a flux, there will be no induced emf. It is a common misconception to
assume that the induced emf will always oppose the magnetic flux.
How would one apply Lenz’s Law to determine the polarity of the induced
emf? The induced emf will be of such a polarity that will either produce or
attempt to produce a current flowing in the closed loop, which will oppose the
changing magnetic flux. An emf will always be induced and will have a magnitude
as expressed above by Faraday’s Law when the magnetic flux enclosed changes.
However, the closed path can consist of any elements. In our case of Fig. 1.6, the
only other component in the closed path besides the coil is a Voltmeter. Therefore,
the induced emf will produce a negligible current as the Voltmeter is an extremely
high resistance. This negligible current would quite obviously not be able to produce
much of a magnetic field to oppose the changing magnetic flux. However, by Lenz’s
Law, even if the opposing magnetic flux produced by the induced emf is negligible,
it will still offer this opposition. Therefore, one can always determine the polarity
of the induced emf by asking the question—what polarity should the emf have so
that the resulting current produces a magnetic field, which will oppose the change
in main magnetic field?
With this rather long and detailed section, we have described the basic laws of
physics that will be used to understand magnetism in electrical engineering. We will
be using these laws over and over again in all the chapters of this book. Before we
end this section, let us pause to note that all these laws were formulated close to
200 years back! These centuries-old laws are all that we as electrical engineers need
to develop a basic understanding of magnetics in electrical engineering. Though
one can use modern techniques and results to arrive at a much more complex and
accurate understanding of the magnetic fields in electrical engineering, for most
power engineering applications, such a degree of detail is not necessary. As the
reader progresses through this book, he or she will be amazed at how we can produce
reasonably functional simulation models with laws that are close to two centuries
old.
1.4 Learning Through Simulation 19

1.4 Learning Through Simulation

In the previous section, we described in detail the basic laws of physics that will be
used in this book to understand the operation of magnetic components. We will be
using these basic laws to also produce simulation models that can be used in basic
applications in power engineering. In this section, I will talk about why simulation
has been used as a tool to learn engineering [6–8] and how the simulations in
this book will make the theory presented much more interesting and easier to
understand.
Nowadays, computers are available to a vast majority of students. There was a
time when a significant proportion of students would still need to visit a computer
lab in order to access computational facilities. Nowadays, personal laptops feature
sufficiently powerful processors and internal memory so as to provide in our hands
a computer good enough to perform simulations. Simulations were once performed
mainly by researchers and graduate students who were exploring new technologies
and needed to investigate the feasibility of their innovations before a practical
implementation. Simulations can now be performed by even undergraduate students
as a part of their course work.
Simulations are a very useful learning tool besides also being useful for the
purpose of practical engineering and research. A simulation is the next closest thing
to building a hardware prototype. Building hardware and laboratory experiments are
undoubtedly the best way to learn and also to verify changes in design and control.
However, for many students of electrical engineering, a ready access to laboratory
facilities is not available. Moreover, the costs associated with building hardware
prototypes on a regular basis make it impossible to use hardware implementation as
a continuous learning tool. Simulations on the other hand provide a safe platform
where an inexperienced engineer can try out new designs, control strategies and
custom models without causing any physical damage to a circuit.
To perform simulations, we normally use simulation software for a particular
application. For example, as power engineers, we need to perform simulations of
electrical circuits, and therefore, we use circuit simulators [9]. A circuit simulator
allows us to define a circuit in some convenient manner, which will then be
processed by the simulator to define a set of ordinary differential equations that form
the mathematical model of the circuit. These are solved by the circuit simulator by
a numerical integration technique to provide us with a time-varying behaviour of
the circuit. The convenience offered by such a simulator is that we do not need to
write equations for every circuit or solve the equations. All we need to do is to
represent the circuit in some form of network representation that can be understood
by the simulator. Some advanced circuit simulators allow a user to draw a circuit
schematic almost similar to a regular electrical circuit, thereby making the process
of circuit representation extremely convenient and visually appealing. Some other
simulators might need a nodal representation.
There are numerous circuit simulators that a power engineer can use for simu-
lating electrical circuits. Some are proprietary and fairly expensive, while some are
20 1 Introduction

free and universally accessible. Many major universities maintain university-wide

licenses of scientific software such that any student registered in the university can
use these software. A student, therefore, might have access to multiple simulation
software. On the other hand, there are a significant proportion of students all over
the world who are enrolled in universities that cannot afford to maintain licenses for
some of the more expensive scientific software. This has unfortunately resulted in
the rise of software piracy. The adoption of free and open source software ensures
universal access to software for any interested user in all parts of the world and
therefore is one solution to providing software services in a sustainable manner.
Any student who is reading this book can use any circuit simulator that he or
she may have access to. The circuit simulator being used must allow the user to
create a circuit schematic and create a user-defined component that can contain a
custom model of a magnetic component. The user-defined component needs to be
interfaced to the rest of the circuit. It needs to receive measured quantities such as
voltages and currents as inputs from the main circuit. These inputs will be fed to a
mathematical model. The output of the mathematical model will result in changes
to the main circuit. Such a closed-loop mechanism must be achievable by the circuit
simulator the reader chooses as the simulation tool to accompany this book. A vast
number of proprietary and free circuit simulators will be able to perform this task.
This book will use the free and open source circuit simulator Python Power
Electronics. Python Power Electronics has been written completely using Python,
which is a free and open source high level programming language. Python Power
Electronics has been written especially for power engineers who wish to simulate
power electronic circuits while implementing control digitally. The homepage of
Python Power Electronics can be found at:
https://www.pythonpowerelectronics.com/.
The circuit simulator can be downloaded from the link:
https://www.pythonpowerelectronics.com/?page=softwaredownloads.
The circuit simulator can be used through a web browser by launching a server
on the computer. To learn how to install the circuit simulator and on how to use it
to simulate power electronic circuits with digital control, the reader can view any of
the video series in the following link:
https://www.pythonpowerelectronics.com/?page=videos.
The reader need not use Python Power Electronics to perform the simulations in
this course. Most of the code samples and simulations can be achieved using any
other circuit simulator. The purpose behind using a free and open source circuit
simulator in this book is to promote the use of free and open source software
in electrical engineering. Free and open source software not only increases the
accessibility of software but also builds a community of users, which enables the
sharing of information and knowledge. The reader is encouraged to follow the
project on any number of different social media where regular updates are made.
For example, a weekly video lecture series can be found on YouTube on the
video link above, where different topics in electrical engineering are explored using
simulations.
1.5 Outline of the Book 21

All the simulations in this book have been hosted on GitHub to be freely
accessible to everyone. The link for the repository that contains the simulations is:
https://github.com/opensourceelectrical/simulating-magnetics-for-power-
electronics.
The repository contains folders for every chapter. Within each folder are the
simulations for that chapter. Each simulation has been arranged in a separate folder
and there can also be nested folders in the case of separate cases being simulated
separately. Every simulation will have a README file. The reader must first read
the README file, which describes what the simulation is about, what are the circuit
schematic files used and what are the control files used. In the later chapters, as we
describe each simulation, the exact files will be referenced and described in detail.
The reader is strongly encouraged to perform the simulations while reading this
book. While describing the simulations, potential modifications to the simulations
will also be suggested, and the reader is strongly encouraged to try these out as well.
As already stated before, simulations are a safe platform to try out circuits without
the fear of circuits blowing up. This gives several opportunities for the learner to
look “under the hood” and also poke around with simulation models.

1.5 Outline of the Book

In this final section of the introduction, we will describe the contents of the book
and what the reader can expect from each chapter. The book progresses gradually
from the basics of magnetic circuits and how these can be used to simulate inductors
to finally simulate three-phase transformers and a flyback converter. Each chapter
will use the learnings from the previous chapter to add another layer of detail to be
able to simulate more complex systems. The reader is strongly encouraged to read
the book in the order of the chapters, even if he or she has a specific application that
he or she would like to simulate.
Chapter 2 begins with building a simulation model for an inductor [10, 26].
Though it may appear that building a simulation model for an inductor is a
silly and useless exercise, the purpose of doing so is to introduce the concept
of magnetic circuits and how the magnetic laws can be combined with electrical
laws in a simulation model. Since the inductor is the simplest magnetic component
imaginable, this will ease us into the process of building a mathematical model using
the laws of physics. In Chap. 2, we build an analogy between electrical circuits and
magnetic circuits and describe how knowing the construction of an inductor, we
can compute the flux in every part of the core of the inductor. Several simulations
are presented with different core constructions. The simulations in Chap. 2 can also
be used by a power electronics engineer to verify designs of inductors as, quite
often, power electronics engineers need to wind custom inductors especially while
designing power supplies.
Chapter 3 extends the mathematical model of an inductor described in Chap. 2 to
model more than one inductor wound on the same core and, therefore, magnetically
22 1 Introduction

coupled to each other [11, 12]. Using the basic laws of physics, we examine several
separate scenarios to fully understand the phenomenon of magnetic coupling and
also express coupled inductors mathematically. We will introduce the concept of
mutual inductance, which will provide us with a manner of expressing the effect
of the current flowing in one coil on the flux linking the other coupled coil. The
simulation model of an inductor will then be extended to model coupled inductors.
Using simulation results, we will verify our theoretical discussions. Since the
simulation of magnetically coupled inductors can consist of many inductors wound
on the same core, we will convert our simulations model into a flexible and scalable
model using matrix equations.
Chapter 4 describes how a transformer is merely a special case of a set of
magnetically coupled windings wound on the same core, with the objective being
to maximize the transfer of energy from one winding to the others [13, 14]. Since
transformers are essentially magnetically coupled coils, the mathematical model of
coupled inductors in Chap. 2 will continue to be used. The difference between a
transformer and a random set of magnetically coupled coils is that a transformer
is an electrical machine that has specifications similar to any other machine like
a motor or generator. For example, a transformer will have a rated maximum
power and rated maximum voltages that can be applied to each winding. The
chapter will describe how using these specifications, an equivalent circuit can be
formulated and the parameters of this equivalent circuit can be estimated. Using
this equivalent circuit, we will extract the values of inductances that are necessary
to use the simulation model of the coupled inductors. We will examine several
simulations to understand the working of the transformer. Simulations will examine
the magnetizing current and the other components in the no-load current of the
transformer. Simulations will also examine the effect of turns ratio in transforming
voltages and currents by simulating step-up and step-down transformers.
Chapter 5 describes some of the most common uses of a transformer in
electrical engineering while using the simulation models presented in Chap. 4. For
a power systems engineer, three-phase transformers are ubiquitous in every system
being analysed [27–30, 30, 31]. Therefore, Chap. 5 describes how the transformer
simulation models of Chap. 4 can be extended to simulate three-phase star–star
and delta–star transformers. To make the contents relevant for a power electronics
engineer, the chapter also presents the simulation model of a flyback converter. The
chapter presents the theory behind high frequency transformers and the specific
application of high frequency transformers in the specific case of a flyback converter
[15, 19, 20]. The purpose of describing the operation of the flyback converter is to
describe how even in the context of a non-linear power converter such as a flyback
converter, the operation of the transformer can still be interpreted using the basic
laws of physics.
Chapter 2
Presenting Basic Magnetic Circuits
with Inductors

2.1 Introduction

The inductor is probably the most basic magnetic component in electrical engineer-
ing [1, 2, 22, 24]. An inductor can appear unintentionally as a parasitic element or
can be included intentionally for various reasons. In electrical engineering besides
power electronics, an inductor is usually considered to be a parasitic element. As
an example, it is normal to consider a distribution feeder as a combination of
resistance–inductance (R–L) or resistance–inductance–capacitance (R–L–C) [32].
Such a model is usually to capture the voltage drop across the feeder. When
introduced intentionally, an inductor is usually a component in a filter that either
blocks the flow of current in a particular path or bypasses current into a path [33, 34].
In power electronics applications, it is difficult to imagine a power converter
without at least one inductor [10, 26]. Therefore, despite being an intermediate
chapter to the later chapter on transformers, this chapter on inductors is of significant
importance to a power electronics engineer. In a power electronic converter, an
inductor usually plays a dual role—as a filter and as an energy storage element
[35]. The fundamental basis of a static power electronic converter is that power is
conditioned by non-linear solid state devices. As a result, the voltage of the input
supply is converted into a switched voltage by one or more solid state devices. In
order to condition this switched voltage, we need inductors and capacitors besides
potentially many other components, to store surplus energy during a particular time
interval and transfer the energy from one branch of the circuit to another. Since
power electronic converters differ widely in their mode of operation, the manner in
which inductors and capacitors condition the output voltage will also vary widely.
In a vast number of power electronics applications, the inductors used are
supplied by manufacturers as off-the-shelf components just like resistors, capacitors
or power devices. The two basic specifications of an inductor are the inductance
and the current rating. Manufacturers are continuously expanding their catalogue of
inductors to supply commonly used inductors. However, there are instances when an

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 23

S. V. Iyer, Modeling and Python Simulation of Magnetics for Power Electronics
Applications, https://doi.org/10.1007/978-3-030-96768-0_2
24 2 Presenting Basic Magnetic Circuits with Inductors

inductor cannot be found available directly as a mass-manufactured component. In

such a case, as power electronic engineers, we can have a custom inductor fabricated
by a manufacturer for a particular inductance and current rating. There are also a few
cases, where due to the uncertain nature of the application, we are forced to wind
our own inductors. Many power electronics engineers who work in power supply
design and fabrication would at some point of time have wound their own inductors
[10, 26].
There are several resources available for winding inductors besides this also
being taught in power electronics courses [10, 26]. However, in a vast number
of cases, the resources we use are quite confusing resulting in improperly wound
inductors. Specifically, the resources that I used never explained the process from
fundamentals but rather was just a set of equations. It was for this reason that
the very first chapter in this book starts with the simulation of inductors. Though
the design of inductors will not be covered in this chapter, the simulation models
presented can be used as a very convenient tool to verify if a particular design of
an inductor is close to what we expected. Furthermore, with inductors being basic
components that are easy to analyse in a circuit, developing a simulation model for
an inductor will ease the process for the later chapters where we begin to model
magnetic coupling. The reader is encouraged to simulate every example presented
in this chapter and also modify those simulations for his or her purposes.

2.2 Revisiting the Basic Inductor

In this section, we will use the basic laws of physics to express the most basic
equations of an inductor [1, 2, 10, 22, 24–26]. Modelling and simulating an inductor
might seem like a trivial exercise, but this basic understanding of the inductor will
demonstrate how it is possible to convert basic physical laws into a simulation
model. This section will show, beginning with the inductor, how all that is needed
to develop a basic simulation model for any magnetic component are the basic
laws of magnetism and induction—Ampere’s Law, Faraday’s Law and Lenz’s Law
[10, 22, 24, 26].
The inductor as we know is a coil optionally wound on a core. Most inductors
have a core either of iron or ferrite depending on the application [10, 26]. However,
it is possible to have an inductor that has no core—also called air-cored inductors. In
any case, if a wire is wound as a coil, and if it carries a current, a magnetic field will
pass through the coil. This phenomenon is the result of Ampere’s Law that states
that a current carrying conductor will produce a magnetic field. The direction of the
magnetic field can be determined by applying the Right Hand Thumb Rule. If you
grab the wire with your right hand with your outstretched thumb pointing in the
direction of the flow of current, the direction in which your fingers encircle the wire
is the direction of the magnetic field.
In the case of a current passing through a coil, using the Right Hand Thumb Rule
produces the following rule for determining the direction of the magnetic field. The
2.2 Revisiting the Basic Inductor 25

Fig. 2.1 A coil wound on a

I I
core

Fig. 2.2 A coil wound in air

I I
φ
φ

φ
φ

face of the coil carrying current in the counter-clockwise direction is the North pole
as the magnetic lines of force leave the coil from this side. The face of the coil
carrying current in the clockwise direction is the South pole as the magnetic lines
of force enter the coil from this side. This rule is just a derivative of the Right Hand
Thumb Rule applied specially to the case of a wire wound as a coil. Figures 2.1
and 2.2 show the magnetic lines of force produced by currents flowing through coils
wound on a core and without a core.
For an inductor, two physical laws always hold true. Faraday’s Law states that
when a conductor is placed in changing magnetic field, an emf (electromotive force)
is induced across it [22, 24]. The law further states that the induced emf generated
is equal to the rate of change of flux linked ψ with the conductor:

dψ
e= (2.1)
dt
In the case of an inductor, the flux linked with (flowing through) the inductor takes
into account the number of turns of the coil and is the product of the magnetic flux
φ and the number of turns N of the coil:

ψ = Nφ (2.2)
26 2 Presenting Basic Magnetic Circuits with Inductors

With the unit of flux being Weber, the unit of flux linkage is merely Weber-turn.
Therefore, the induced emf can expressed as:

dφ
e=N (2.3)
dt
From an engineering perspective, an expression involving the magnetic flux
linking an inductor is a fairly abstract equation. To introduce a more practical
approach to the above expression, let us introduce inductance of a coil as a physical
quantity. The inductance L of a coil is the defined as the flux linkages produced per
unit current flowing through it [1, 2, 10, 26]. A coil will have an inductance of 1
Henry, if a current of 1 Ampere flowing through it will produce a flux linkage of 1
Weber-turn:

ψ = Li (2.4)

Therefore, the induced emf can be expressed with respect to the inductance of the
coil as follows:
di
e=L (2.5)
dt
The second law that can be used to model the inductor is Lenz’s Law. Lenz’s Law
states that the induced emf is such so as oppose the cause that produces it [22, 24].
From the above discussion on induced emf being proportional to the rate of change
of current, the cause of the induced emf is the change in current. Therefore, this
usually means that the induced emf will oppose the change in the current through
the inductor as this change in current is what produces the induced emf. Lenz’s Law
helps to determine the polarity of the induced emf, while Faraday’s Law determines
the magnitude of the induced emf. It is important to note that the induced emf will
oppose the changing current, which is the cause that produces it. The induced emf
does not oppose the current itself as a resistor does. Further sections and chapters
will contain detailed description of cases that will illustrate this concept.
Mathematically, if expressing all quantities as phasors, one could combine the
two laws to result in the following expression for the induced emf:

di
e = −L (2.6)
dt
However, the above equation is applicable for a phasor representation, which is not
well suited for a simulation model. Therefore, we will not use the negative sign in
the equation but rather include it in the circuit as we build our simulation model.
In this section, we have described how Faraday’s Law and Lenz’s Law determine
the emf induced in an inductor when the current flowing through the inductor
changes. The laws link the change in magnetic flux linking the coil with the
emf induced in the coil. However, using our definition of inductance, we can
2.3 Inductor Model 27

include inductance into the expression of induced emf. In the next section, we will
demonstrate how this expression of induced emf can be expanded to an inductor in
a circuit.

2.3 Inductor Model

In the previous section, we examined how Faraday’s Law can be used to express the
emf induced in the inductor with respect to the rate of change of flux linkages, and
therefore with respect to the change of current through the inductor. However, the
expressions in the previous section were for the inductor in isolation. In this section,
we will examine how the expressions from the previous section can be used when
the inductor is a part of a larger circuit.
While simulating a circuit, the inductor is modelled by substituting the following
expression into the network equations for the circuit:

di
L (2.7)
dt
The network equations can be written for any circuit based on Kirchhoff’s Voltage
Law (KVL) and Kirchhoff’s Current Law (KCL). By substituting the above
term, the resultant differential equations that represent the circuit can be solved
numerically in a simulation [9].
However, this considers the inductor as a constant and does not consider any of
the magnetic details of the inductor. The assumption used to arrive at the above
expression in the previous section was

φ∝i (2.8)

In reality, this is not the case. The relationship between φ and i depends on the core
material, the structure of the core and most importantly is not even a simple linear
function [10, 22, 24, 26]. In most cases, it is simpler to neglect the non-linear nature
of the relationship between φ and i unless we wish to simulate in detail the effect of
the core magnetic material.
It is possible to include details of the magnetic properties of elements such as
an inductor in a simulation [5]. The level of detail that we wish to include will
determine the complexity of the model. Very complex and detailed models might
need techniques such as finite element methods, which are out of the scope of this
book. However, for those who would like an inductor model that is more than a mere
constant L, it is possible to do so with basic concepts of magnetics from physics.
In this section, we will introduce the basic concepts of magnetism in order to create
an inductor model. However, these concepts can be extended later to any magnetic
component such as coupled inductors and transformers that will be covered in later
chapters [10, 26].
28 2 Presenting Basic Magnetic Circuits with Inductors

Before we get into the details of magnetic characteristics of the inductor core, let
us review the basics of magnetics. In the previous section, we had considered two
inductors in Figs. 2.1 and 2.2. The flux paths shown in the two figures was the result
of applying the Right Hand Thumb Rule. Let us first consider the air-cored inductor
in Fig. 2.2. Let us suppose that this coil has N turns and is carrying a current I . From
Ampere’s Law, any current carrying conductor will produce a magnetic field. In the
case of a coil of N turns, the current flowing through each turn will augment the
magnetic field produced by the others. The very first step to calculating the magnetic
field is to define a term called the magneto motive force (MMF) of the coil as [25]:

MMF = NI (2.9)

In simple terms, the MMF is basically how much of an effect the current flowing
has in terms of the ability of the coil to produce a magnetic field. Since it is a coil,
the current’s effect increases by a factor equal to the number of turns. For a single
straight wire, the MMF would be just the current I .
Now that we have defined the MMF as the driving force that produces the
magnetic field, the next question is what else do we need to know to calculate the
flux? Since this coil of Fig. 2.2 is wound without a core, the lines of flux pass through
the coil and complete their closed paths on all sides through the air. The wider flux
paths will have decreasing field strengths until the field strength or the magnitude
of flux becomes negligible. Therefore, when calculating flux, we need to define the
path of the flux. In the case of an air-cored inductor of Fig. 2.2, the path is difficult
to assign unless we assume the shortest possible path around the coil through which
the flux can complete its path. To make the understanding of flux paths easier, let us
consider the inductor of Fig. 2.1 where the coil is wound around an iron core.
Before we begin calculating the magnitude of the magnetic flux, what is
the difference between the two inductors? As already stated, the magnetic flux
completes a path just like current does. This flux path can be through a single
medium or through multiple mediums. Every medium offers a certain level of
opposition to the flux just like conductors offer opposition in the form of resistance
to the flow of current through them. The opposition offered by a medium to the flow
of magnetic flux through it is called reluctance [1, 2, 22, 24]. The reluctance offered
by the medium is dependent on the length of the path, cross-sectional area of the
path and the permeability of the medium:
l
ρ= (2.10)
μA
The above formula is very similar to the formula for calculating the resistance of a
current-carrying conductor.
In the expression (2.10) for reluctance, the intrinsic property of the medium that
offers opposition to the flow of flux through it is the permeability μ, which has a
unit of Henry per metre (H/m) [22, 24]. This permeability is further expressed as:

μ = μ0 μr (2.11)
2.3 Inductor Model 29

where μ0 is the permeability of free space and is equal to 4π × 10−7 H/m. Free
space (or air) is considered the base in terms of measuring permeability of a medium
towards the flow of flux through it. μr is the relative permeability of a medium with
respect to that of free space and is merely a constant. From (2.10), the larger is the
permeability of the medium, the lower will be the reluctance of the particular path
to the flow of flux.
Magnetism is a subject that has been studied intensively for many decades
and close to a century. The magnetic properties of many different materials have
been studied under various conditions and the interested reader can find an almost
inexhaustible pile of information on this topic [22–24]. However, we need to restrict
our discussion in this book to specific applications with respect to power electronics.
In power electronics, the most popular core materials used are iron and ferrite. The
actual fabrication of the core might need the iron to be laminated or the ferrite to be
a composite or a polymer. In some rare applications, there could be no core as well.
In the case of no core or air core, the permeability of the flux path is μ0 itself. In
the case of iron, the relative permeability is in the range of 1000–5000 depending
on the presence of other metals in cases of alloys. In the case of ferrite, the relative
permeability is usually in the range of 300–500 depending on the presence of other
metals in cases of alloys.
With this background on the permeability of iron and ferrite cores, it is fairly
evident that the reluctance of the core is far less than that of the surrounding air.
Therefore, a vast majority of the magnetic lines of flux will flow through the core
and only a negligible amount of flux will leak into the surrounding air [10, 26]. If
we neglect the leakage of flux into the air, we can now determine the path through
which the magnetic flux will flow in the core and thereby calculate the reluctance
of the path. In the case of Fig. 2.1, we can define the average length of the flux
path (shown as a dotted line) as l and the cross-sectional area (imagine a depth to
the core) of A. We can use (2.10) to calculate the reluctance of the flux path. The
magnitude of the flux can be expressed as [22, 24]:

MMF NI μA
φ= = (2.12)
ρ l

From the above expression, the flux is directly proportional to the current flowing
through the coil, the permeability of the core and the cross-sectional area of the
core. The flux on the other hand is inversely proportional to the mean length of the
flux path through the core. Subsequently, using the definition of the inductance, the
inductance can be calculated as:

Nφ N 2 μA
L= = (2.13)
I l
30 2 Presenting Basic Magnetic Circuits with Inductors

Cores can be more complicated than just a single closed path as shown in Fig. 2.1
or there might be a core with multiple windings. Therefore, the question arises as to
how we will calculate the flux using an expression such as (2.12). The good news
is that the path of magnetic flux can be dealt with as a magnetic circuit in a manner
very similar to how an electric circuit is used to compute the current flowing through
different branches of the circuit. This will be described in the next section.

2.4 Magnetic Circuits

In the previous section, we had used the basic physical laws of magnetism to write
an expression for the magnetic flux given a particular construction of an inductor.
The construction of the inductor is with respect to the construction of the core (cross-
sectional area and length), the permeability of the core and the number of turns of
the coil wound on the core [10, 26]. Figure 2.1 was a sample inductor with a core
that we had chosen for our derivation. In this section, we will extend our discussion
to calculate the flux in case the core is a bit more complex than the one in Fig. 2.1.
A core such as the one in Fig. 2.1 will have a magnetic flux φ expressed by (2.12).
In order to increase the magnetic flux φ for that particular core, we could increase the
number of turns of the coil. In order to decrease φ, we could decrease the number of
turns. Another technique that is used to adjust the flux produced by a given number
of turns is to introduce an air gap as shown in Fig. 2.3 [10, 26]. For the purpose
of illustration, the air gap has been shown to be large. However, when introducing
an air gap to decrease the flux in the core, this air gap is usually much less than a
millimetre. In order to introduce this air gap, two half cores are chosen and joined

Fig. 2.3 Iron core with two lg

air gaps
2

lg
2
2.4 Magnetic Circuits 31

together while separating them by some kind of insulating material [10, 26]. The
gap introduced by the insulating material is usually equivalent to an air gap that has
a permeability equal to that of μ0 .
By introducing this air gap, we are introducing a segment that has a much higher
reluctance than that of the iron core. In Fig. 2.3, we have shown the two air gaps to
l
be of length 2g . If the length of the complete path shown as the dotted line passing
through the centre of the core were to be denoted by l, this would imply the length
of the flux path through the iron core was l − lg . It should be noted that at the air
gap, the flux would experience a fringing effect and would not be as uniform as it
was while passing through the iron core [5]. However, this detail is very difficult
to express and therefore, we will neglect this fringing effect and assume the flux to
be as uniform as it was while passing through the iron core. The reluctances of the
paths through the iron core and the air gap can be expressed as:

l − lg
ρc = (2.14)
μ0 μr A
lg
ρg = (2.15)
μ0 A

The flux flowing due to the MMF generated by the current I flowing through N
turns of the coil will face the sum of the reluctances due to the iron core and the air
gap [1, 2]:

NI
φ= (2.16)
ρc + ρg

Therefore, by introducing an extremely small air gap in the core, the flux through
the core can be adjusted. At times, this form of adjustment results in bringing the
inductance of this component closer to the desired value as opposed to adjusting the
number of turns. The inductance can be calculated as:

Nφ N2
L= = (2.17)
I ρc + ρg

From the above case, it is quite clear that magnetic flux behaves in a manner quite
similar to electric current. The reluctances encountered by the flux path add up when
mediums are in series in exactly the same way resistances add up to oppose the flow
of current when different conducting materials are connected in series. In the case
of parallel limbs in the iron core, the magnetic flux has parallel paths to flow and
again just like electric current, a greater proportion of the flux will flow through the
path of least reluctance. As an example, consider the same coil wound on the iron
32 2 Presenting Basic Magnetic Circuits with Inductors

lg
2

I
a c
lg
b 2

lg
2

Fig. 2.4 Iron core with multiple legs and air gaps

core of Fig. 2.4. If the width of the core was w, we can calculate the reluctance of
each leg as follows:

2a + b + 3w − lg
ρ1 =
μ0 μr A
lg
ρg1 =
μ0 A
b+w
ρ2 = (2.18)
μ0 μr A
lg
2c + b + 3w −
ρ3 = 2
μ0 μr A
lg
ρg2 =
2μ0 A

In the calculation of the reluctance of every limb of the iron core, we have used
the mean length of the flux path. The main flux produced by the coil (flowing in the
left limb) can be expressed as:

NI
φ= (2.19)
ρ2 (ρ3 +ρg2 )
ρ1 + ρg1 + ρ2 +ρ3 +ρg2

In the above expression, the parallel paths available to the main flux result in an
equivalent reluctance that is the parallel combination of the reluctances ρ2 and ρ3 +
2.5 B–H Relationship 33

ρg2 . This parallel combination of reluctances can be computed in exactly the same
manner as for a parallel combination of resistances.
This analogy with electric circuits enables us to compute the magnetic flux
for fairly complex core constructions. Extending our analogy, we can also use
superposition theorem to calculate the flux due to multiple windings on different
legs of the iron core. These relations will be extremely useful when we model
electrical machines such as transformers, which have multiple windings wound on
the same core. Before we use the above expressions to model inductors, we need
to understand that the above expressions contain approximations. In reality, the
magnetic properties of the core material are not constants but can change due to
various reasons. In the next section, we will describe the B–H relationship, which is
one of the most widely used expressions for the magnetic properties of a core.

2.5 B–H Relationship

In the past section, we established a method to calculate the flux in a magnetic

circuit. In order to perform this calculation, we need to know the nature of the
magnetic material, the dimensions of the core, the length and the cross-sectional
area. The only drawback to such a calculation is that it contains the dimensions of
the magnetic material, which makes it a little inconvenient to use as the computation
is coupled to a particular core construction. In this section, we will introduce two
new physical quantities so as to express the flux calculation in a form that is
independent of the core.
The magnetic field strength H for a magnetic circuit is defined from Ampere’s
Law as [22, 24]:

NI
H = (2.20)
l
The strength of the magnetic field is therefore the magneto motive force (MMF) per
unit length of the magnetic path. The next term defined is the flux density B, which
is the flux per unit cross-sectional area of the magnetic path:

φ
B= (2.21)
A
With these two definitions, we can rewrite the magnetic flux equation as:

φ NI
=μ (2.22)
A l
From our definitions of magnetic field strength and flux density:

B = μH (2.23)
34 2 Presenting Basic Magnetic Circuits with Inductors

Fig. 2.5 Relationship B

between B and H with
constant μ
μ

In Eq. (2.23), we have eliminated the physical dimensions of the magnetic core.
Instead we have expressed the flux density (flux per unit area) with respect to the
magnetic field strength (MMF per unit length) and the permeability of the core. This
equation therefore establishes the magnetic behaviour of any magnetic material as
μ is a property of the material. This equation is a very powerful relation that sets
the base for most of magnetic design in electrical engineering. Though this relation
is extremely useful by itself, it assumes that the permeability μ is a constant. Every
constant related to a material is subject to change with environmental conditions
such as temperature and many others. However, given all environmental conditions
to remain fixed, we might be tempted to use (2.23) as the absolute truth with respect
to determining the flux density B in terms of the magnetic field strength H or vice
versa. In simple terms, we expect the relationship between B and H to be as shown
in Fig. 2.5.
Such a linear relationship between B and H with the slope of the line being the
permeability μ would make a very convenient design tool. Such a tool can be used
to calculate how much would be the flux density B in a core made of the material
if the magnetic field strength H were to be of a given value. We could quickly
determine what would be the effect of a given number of turns N of a coil wound
on the core carrying a current I if we assume a mean length of the flux path in the
core. Unfortunately, magnetism is a phenomenon that is much more complex than a
mere straight line as depicted in Fig. 2.5. To capture detailed aspects of magnetism
is not something that will be attempted in this book as our objective is to model
magnetic components for power electronic applications. However, a few details are
worth considering as they do affect us power engineers.
The first complication to Fig. 2.5 is the effect of saturation [22, 24]. As the
magnitude of H increases, the magnitude of B will increase proportionally.
However, for values of H greater than a particular value, the increase in B will no
longer be proportional and will not obey (2.23). This effect is saturation. To visualize
this, if the core is thought of as a container and flux as a fluid, saturation is when the
container is filled to the capacity with this fluid called flux and cannot further take
large quantities of it. Saturation is depicted as shown in Fig. 2.6. The two large dots
in the curve indicate the limits of the linear part of the B–H relationship. Either
these points or two different points within the linear part of the curve are called
the knee points similar to the knee of a human leg. The values of B at these knee
2.5 B–H Relationship 35

Fig. 2.6 Relationship B

between B and H showing
saturation

points are taken as the maximum value of flux density that a core can be subjected
to without causing the core to saturate.
A simulation of the effect of saturation is quite complex and will not be included
in this book [5]. However, at this stage of the discussion, let us discuss two aspects
of saturation—the effect and the relationship. If a very large current were to flow
through a coil wound on a core such that the magnetic field strength H produced
by it was larger than the value of H corresponding to the knee point, the core of the
inductor is said to be saturated. The flux density B will no longer be proportional
to the magnetic field strength H . If a current carrying conductor (wire) is forced to
carry a current larger than the value it is safely allowed to carry, it will experience
greater losses that could potentially damage the conductor. The resultant heating
effect could cause the surrounding insulation to get damaged and if the current was
very high, the wire might simply burn. In a similar manner, if the flux density B in a
core was to be much higher than the knee point value, such a core might experience
large losses that could cause the magnetic properties of the core to deteriorate. From
basic physics, excessive heat can damage a magnet and deteriorate the magnetic
properties of an electromagnet.
The next question is how do we study the behaviour of the core in this region of
saturation if we can no longer count on the permeability μ of the core material to be
a constant? Magnetic materials available as cores for most engineering applications
are usually accompanied by datasheets where the manufacturers provide B–H
curves under a number of different operating conditions [10, 26]. Therefore, to
study the behaviour of the core when saturated, we must rely on experimental data
provided by the manufacturer of the core material. In order to use these non-linear
characteristics in a simulation, we could rely on techniques such as curve fitting
where a non-linear expression is derived so that it matches a given curve. Therefore,
in such cases, we would need curve-fitting techniques to estimate a function f such
that:

B = f (H ) (2.24)

The other aspect of the B–H relationship that deviates from the ideal linear
characteristics of Fig. 2.5 is the fundamental nature of magnetic materials. When
a core made of a magnetic material such as iron is placed in a magnetic field, it
36 2 Presenting Basic Magnetic Circuits with Inductors

Fig. 2.7 Practical B

relationship between B and
H without saturation

Br
Hc
−Hc H
−Br

will become magnetized. If the magnetic field were to be removed, the core will
continue to retain some of its magnetism even if the magnetic field produced by it is
now weaker than as compared to when it was in the external magnetic field. In order
to completely demagnetize the core, a reverse magnetic field will need to be applied.
This is true for a coil wound on a core as well. This retention of magnetization by the
core causes the B–H relationship to become non-linear and follow a path as shown
by arrows in Fig. 2.7 [22, 24].
In Fig. 2.5, if the current I flowing through the coil increases, this will cause
the magnetic field strength H to increase. Subsequently, we expect the flux density
B to increase proportionally. However, in reality, the increase in B will not be
linear but rather will be a curve. This is due to the fact that as the magnetic field
strength increases and the core gets magnetized, the core is sluggish to respond to
the increase in H with a proportional increase in B. One can think of this behaviour
as a form of friction. The result is the B–H curve, which is as shown in Fig. 2.7. If
the current flowing through the coil is an alternating current (ac), when it reaches
its peak, it will decrease back to zero following a sine waveform with respect to
time. As the current decreases to zero, H will decrease until it reaches zero. One
would expect B to decrease proportionally to zero along the straight line of Fig. 2.5.
However, this is where the retentivity of the core as a magnetic material plays a role.
It “‘fights” to retain some of the old B as H decreases.
Once current starts decreasing from its peak causing H to decrease as well, the
operating point follows another curve. This is due to the property of the core to retain
the magnetic flux and therefore, the decrease in B is slower than the decrease in H .
When current becomes zero causing H to be zero, the flux density B has a positive
non-zero value. This value Br is called the retentive flux density. To force B to zero,
the current has to become negative resulting in a negative H . The value of H = Hc
for which B = 0 is called the coercive magnetic field strength. Symmetric values of
−Br and −Hc can be found during the negative half cycle of the current. As with
the case of saturation, these curves are usually provided by the manufacturer of the
core for various operating conditions such as temperature.
In this section, we have described the relationship between the magnetic field
strength H and the flux density B in the core. As already stated, this relationship
2.6 Simulating an Inductor 37

between B and H allows us to define the behaviour of a magnetic core independent

of the core dimensions as we found in the previous section. Depending on our
requirements, we can model the core to the detail we require. We could neglect
saturation and retentivity and use a linear relationship as shown in Fig. 2.5. Or
we could include saturation while neglecting retentivity as shown in Fig. 2.6. Or
we could include both saturation and retentivity—this curve is not shown but the
phenomenon of retentivity extends also into the saturation zone. In the next section,
we will talk about how we can put together all these different expressions described
over the past few sections to form a simulation model.

2.6 Simulating an Inductor

In the past few sections, we have examined the different physical laws and phenom-
ena that determine the operation of the inductor. We examined how Faraday’s Law
and Lenz’s Law can be used to determine the relationship between flux and applied
voltage. We examined how a magnetic circuit can be solved to determine the flux
due to the current flowing in a coil. We also examined the B–H curve of a core and
how saturation impacts this characteristic. In this section, let us put all this together
to create a simulation model for an inductor.
When creating a simulation model for an electrical component (or even a
machine), we need to determine the inputs and outputs [6–8]. An inductor is a
two-terminal component as shown in Fig. 2.8 and is typically connected in series
with other components. In the case of such a two terminal component, there are
two electrical quantities that can be measured with respect to the terminals—the
current flowing into one terminal and out of the other and the voltage across the
terminals. For our simulation model of the inductor to be accurate in the electrical
sense, the current flowing through the terminals and the voltage across the terminals
must conform to the physical laws established for the inductor. If we were to express
the voltage and current as functions with respect to each other:

di
v = L = f (i) (2.25)
dt

1
i= vdt = g(v) (2.26)
L

Fig. 2.8 The inductor as a i

two-terminal component

v
38 2 Presenting Basic Magnetic Circuits with Inductors

In the above equations, f and g would be the detailed magnetic models that we
have discussed in the previous sections. However, the above equations show that
whichever functional representation we choose, one variable will be independent
and the other dependent. In the first equation, the current through the terminals
is the independent variable and the voltage across the terminals is the dependent
variable. In the second equation, the voltage across the terminals is the independent
variable and the current through the terminals is the dependent variable. From an
electrical sense, what does it mean if the voltage is independent or the current is
dependent? To answer this question, we need to ask—is the component voltage
driven or current driven? If we turn on a circuit with an inductor, are we applying
a voltage across the inductor or are we passing a current through it? In most cases,
the voltage is the independent input that is applied to a circuit while the current that
flows is dependent on the components. Therefore, we can begin by labelling the
voltage across the inductor terminals as the independent variable while the current
needs to be calculated using our mathematical model.
Our simulation model must accept the voltage across the terminals as the input,
use this voltage to calculate the current drawn from the mathematical model
based on the physical laws and finally ensure that this current passes through the
terminals. This is achieved using Fig. 2.9 where a Voltmeter measures the voltage
V across the terminals and feeds it to the mathematical model g(V ). The output
of the mathematical model is the current i. To realize this current i, a controlled
voltage source is connected in series with a variable resistor between the terminals.
The output of the controlled voltage source and the value of the variable resistor
are adjusted in relation to the measured voltage across the terminals, such that
the current flowing through the terminals is equal to the value calculated by the
mathematical model.
One could argue that to produce a variable current equal to the value produced
by the mathematical model, all we need is either a controllable voltage source
or a variable resistor. When modelling an inductor and for that matter, any other
electrical component using detailed models, even though we need only voltage to

Fig. 2.9 Simulation model of

the inductor

g(V )
2.7 Simulations of an Inductor 39

be variable, it is still advisable to choose the resistor as a variable resistor so that

the model can be applied to systems of different ratings. This is so as to be able
to perform impedance matching between our model and the system to which we
are connecting the model. This will be better explained in the later section when
we simulate our custom model. Therefore, even if the value of the variable resistor
is calculated and set initially, choosing a value calculated from system ratings will
result in a more stable simulation model.
In this section, we have described how a mathematical model of an inductor can
be implemented in a simulation to produce a customizable component. We have not
yet derived the mathematical model of the inductor, as we must decide the extent
to which we would like to model the magnetic circuit of an inductor. The objective
of this section was to define the interface of the simulation model of the inductor in
terms of inputs and outputs before we formulate the mathematical model. In the next
few sections, we will present a few cases where increasing details of the magnetic
circuit of an inductor are simulated.

2.7 Simulations of an Inductor

In the previous section, we described the interface for the model that we can use for
simulation of an inductor. In this section, let us begin with basic simulations where
the equations pertaining to the magnetic model can be included. The simulations
will gradually increase in complexity as the detail to which the magnetic model is
simulated is increased. All simulations will be performed using the free and open
source circuit simulator Python Power Electronics with all code written in Python.
Figure 2.10 shows the sample circuit that we can use to test our inductor models.
In this circuit, we have connected a voltage source to a resistor R1 and inductor L1
connected in series. The resistor rs is the parasitic resistance of the voltage source.
We are measuring the current supplied by the voltage source and the voltage across
the series combination R1 –L1 . We could choose either an ac voltage source of 50
or 60 Hz or we could choose a dc voltage source. For simplicity, let us choose a
dc voltage source as it will be easier to interpret the simulation results with a dc

Fig. 2.10 Test circuit for i

inductor simulation

rs
L1
V
Vin
R1
40 2 Presenting Basic Magnetic Circuits with Inductors

Fig. 2.11 Simulation result with in-built inductor model

voltage. We can choose some sample values for the simulation—a dc voltage of
100 V, rs = 0.01 , R1 = 10  and L1 = 0.3H. These are just random values and
the reader is welcome to change these. It should be noted that in this test circuit,
we are simulating the inductor L1 . To begin with, let us use the in-built model of
the inductor, which contains no details of the magnetic circuit. This simulation can
be found in the folder inbuilt_inductor within chapter2_inductors in
the following link in the simulation repository:
https://github.com/opensourceelectrical/simulating-magnetics-for-power-
electronics.
Figure 2.11 shows the plot of the current supplied by the voltage source. The
L1
time constant of the circuit is τ = R 1
= 0.03 s. Therefore, the settling time of
the circuit is expected to be 4τ = 0.12 s. The settling time of a system is the
time taken for the system to settle to within 2% of its steady state following a
disturbance. In this particular case, the initial application of the dc voltage at time
t = 0 is the disturbance. From the result of Fig. 2.11, the settling time of the circuit is
R1 = 10 A. This of course
approximately 0.12 s. The steady state will be the current 100
is a fairly trivial simulation. However, our only purpose of showing this result is to
compare it with the results we obtain from the simulation of a custom inductor.
Let us repeat the above simulation while replacing the in-built inductor in
Fig. 2.10 with the custom model of Fig. 2.9. This simulation can be found in the
folder basic_Ldibydt_equation within the folder magnetic_model in
chapter2_inductors inside the repository:
https://github.com/opensourceelectrical/simulating-magnetics-for-power-
electronics.
2.7 Simulations of an Inductor 41

We will need a controllable voltage source and a variable resistor that can be attained
by using the ControlledVoltageSource element and the VariableResistor element,
respectively. Moreover, we also need a control logic for the function g(v) shown in
Fig. 2.9. Repeating the expression of g(v) from the previous section:

1
i= vdt = g(v) (2.27)
L

The above expression needs to be realized in a program and used to regulate the
voltage produced by the ControlledVoltageSource element such that a current equal
to i flows in Fig. 2.9.
Before describing how the program corresponding to g(v) can be written, there
needs to be a change in the expression for g(v) due to an issue specific to the
numerical implementation of the integration in g(v). In g(v), we are integrating the
voltage v applied across the inductor. The ideal case is that v will be a sinusoid of
the fundamental frequency with some arbitrary phase angle. Integrating a sinusoid
will produce a cosinusoid along with an integration offset:

cos(ωt + φ)
Vm sin(ωt + φ)dt = − +C (2.28)
ω

To compute the value of the integration offset, we would use the initial conditions.
However, in a physical system such as an electrical circuit, such an integration offset
will imply a dc component in the current. In practical cases, this dc component
that arises at the time the circuit is excited will gradually decay to zero due to
the resistance in the circuit. However, in the expression for g(v) above, we have
considered a pure inductor without any resistance. Therefore, any integration offset
that is produced will remain and not decay. However, this cannot be allowed to occur
particularly if the inductor is used in ac circuits.
To ensure that any integration offset that appears eventually decays to zero, we
need to add a parasitic resistance to the inductance. Such a parasitic resistance could
represent the winding resistance of the inductor as well as any magnetic losses in
the inductor iron core. Therefore, the modified expression for g(v) is

1
i= (v − ir)dt = g(v) (2.29)
L

It is important to note that the resistance r in the expression for g(v) above is not the
same as the series resistance in Fig. 2.9. The series resistance in Fig. 2.9 is a much
larger resistance that is needed to ensure that the difference between the output of
the controlled voltage source and the voltage measured across the inductor terminals
results in the current i as computed.
42 2 Presenting Basic Magnetic Circuits with Inductors

With this change in the expression for inductor current, we can now begin writing
the control program. It is always best to begin with a skeletal code.
d t = 1 . 0 e−6
L = 0.3
r = 0.01
i f t _ c l o c k >= t 1 :
# I n d u c t o r model
t 1 += d t

In the above block of code, we have defined three variables dt, L and r at the
head of the file. dt is the time interval at which this code needs to be executed. In
Python Power Electronics, control functions can be executed at designated time
instants by defining time events in the configuration of the control file. In the
above code, t1 is the time event variable, which is compared with t_clock.
t_clock is the present simulation time instant and is provided and updated by
the circuit simulator. The inductor model will be executed when the time instant
of simulation is greater than or equal to the time event. At the end of the inductor
model computation, the time event variable is incremented with the time interval
dt, which ensures that the inductor model is computed at the interval of dt.
Since we have not begun simulating power electronic circuits, a 1 µs integration
time step is sufficient for a stable and accurate simulation. In the case of the custom
inductor model, the control file is in reality a model computation file that should be
executed at preferably the same time step as the simulation integration time step. In
contrast, control files that run control algorithms and change the state of components
run at a much larger time step as compared to the simulation integration time step.
In later chapters, the time step dt of the control file and the integration time step
will be reduced to smaller values when power electronic circuits will be combined
with magnetic components.
The code in a control file such as the block above will be inserted into a function
by the circuit simulator at run time. Therefore, all the rules for Python variables
apply. As an example, every variable must be defined either at the head of the file
such as dt, L and r or must be defined as special variables in the configuration of
the control file such as the time event variable t1. The only exception is t_clock,
which is an internal variable of the circuit simulator and it is advisable to not alter
this. It is advisable to define constants and parameters at the head of the file such
as the case with dt, L and r. All other variables that can change as the simulation
proceeds should be defined in the configuration of the control file. For now, we have
started with a definition of the time event variable t1.
With this basic background on the setup of a control file in Python Power
Electronics, we can now include the integral equation g(v) in the placeholder
“Inductor model”. The expression g(v) contains the voltage measured by the
Voltmeter across the component terminals. In the configuration of the control file,
2.7 Simulations of an Inductor 43

we can define an input variable vmeas and connect it to the measurement of the
voltmeter. Numerically, the integral equation will be realized as a summation:


n
1
i[n] = (v[k] − i[k − 1]r)dt (2.30)
L
k=0

In the above equation, consecutive samples are separated by a time interval equal
to dt. In order to calculate the current at a sample n, we must perform the summation
from k = 0 to k = n if we assume the circuit to be at rest at t = 0 and all
variables to be 0. However, the current i also appears on the right hand side within
the summation. In such a case, the sample of current within the summation will
be the previous sample k − 1. The summation can be simplified to the following
recursive form:
1
i[n] = i[n − 1] + (v[n] − i[n − 1]r)dt (2.31)
L
We have already defined the voltage v in the program as vmeas. We need to
define the current i as a variable. From a programming perspective, we need any
Python variable such as ind_current that can be initialized to a value of 0 at
time t = 0 and will be updated with every new value of vmeas. As for the current in
the right hand side, it will hold the previous value of the computation at the previous
time instant. In order to achieve this, we can define a static variable ind_current
and initialize it to 0. The Python code corresponding to g(v) will be
i n d _ c u r r e n t += ( 1 / L ) ∗ ( vmeas − i n d _ c u r r e n t ∗ r ) ∗ d t

We must now use this value of current to compute a voltage that will be generated
by the ControlledVoltageSource. The voltage generated by the ControlledVoltage-
Source must take into account the resistance of the VariableResistor in series with
it. Let us now create two output variables in the configuration of the control file.
The first will be connected to the ControlledVoltageSource and will be called
vsrc and the second will be connected to the VariableResistor and will be called
series_res. The initial value of vsrc can be 0 since at time t = 0 we would
like the system to start from rest. The initial value of series_res on the other
hand can be set to a value such as 100  as the resistance in series with the voltage
source cannot be zero but must be a value that can limit the current when the main
voltage is switched on.
With the inductor current already calculated, the voltage to be generated by the
ControlledVoltageSource can be computed by a simple application of Kirchhoff’s
Voltage Law in the closed loop with the Voltmeter, ControlledVoltageSource and
VariableResistor:
v s r c = vmeas − i n d _ c u r r e n t ∗ s e r i e s _ r e s

With this, we have now configured our control file with an input being the Voltmeter
and outputs being the ControlledVoltageSource and VariableResistor. We have
44 2 Presenting Basic Magnetic Circuits with Inductors

created a time event variable to regulate the intervals at which our model equations
are computed and static variables have been defined, which will update as the
simulation progresses.
Before we conclude on this basic simulation model of an inductor, we need to
elaborate on the numerical integration technique used in the simulation. The induc-
tor model above uses the Backward Euler integration method. In this simulation, we
arrived at the results we expected and for a simulation as simple as this, Backward
Euler provides sufficient accuracy and stability. However, in later chapters, we will
simulate more complex magnetic circuits where accuracy and stability could play
a major role in the final results obtained. Therefore, before we proceed to the next
section, let us replace the Backward Euler integration method with the Runge–Kutta
Fourth Order integration method. For a detailed description of the Runge–Kutta
Fourth Order method, the reader is advised to read the following reference [36]. We
could replace the single statement with multiple statements as follows:
k1 = ( 1 / L ) ∗ ( vmeas − ind_current∗r )
k2 = ( 1 / L ) ∗ ( vmeas − ( i n d _ c u r r e n t + d t ∗ k1 / 2 . 0 ) ∗ r )
k3 = ( 1 / L ) ∗ ( vmeas − ( i n d _ c u r r e n t + d t ∗ k2 / 2 . 0 ) ∗ r )
k4 = ( 1 / L ) ∗ ( vmeas − ( i n d _ c u r r e n t + d t ∗ k3 ) ∗ r )
k = ( k1 + k2 ∗2 + k3 ∗2 + k4 ) ∗ d t / 6 . 0
i n d _ c u r r e n t += k

As can be seen, the Runge–Kutta Fourth Order method calculates four slopes
based on the derivative of current and the final increment is the weighted average of
the four slopes. The complete control code can be listed as follows:
d t = 1 . 0 e−6
L = 0.3
r = 0.01
i f t _ c l o c k >= t 1 :
# I n d u c t o r model
k1 = ( 1 / L ) ∗ ( vmeas − i n d _ c u r r e n t ∗ r )
k2 = ( 1 / L ) ∗ ( vmeas − ( i n d _ c u r r e n t + d t ∗ k1 / 2 . 0 ) ∗ r )
k3 = ( 1 / L ) ∗ ( vmeas − ( i n d _ c u r r e n t + d t ∗ k2 / 2 . 0 ) ∗ r )
k4 = ( 1 / L ) ∗ ( vmeas − ( i n d _ c u r r e n t + d t ∗ k3 ) ∗ r )
k = ( k1 + k2 ∗2 + k3 ∗2 + k4 ) ∗ d t / 6 . 0
i n d _ c u r r e n t += k
v s r c = vmeas − i n d _ c u r r e n t ∗ s e r i e s _ r e s
t 1 += d t

In this section, we have now introduced a control function that can execute
a custom inductor model. The model currently does not contain any details of
the magnetic circuit and is merely a numerical integration of the regular inductor
equation. However, with this simple model, we now have a controller that can
compute the current that is required to flow through the inductor given that a
particular voltage is applied across it. In the next section, we will begin with
introducing details of the magnetic circuit of the inductor coil.
2.8 Simulating an Inductor with the Magnetic Circuit 45

2.8 Simulating an Inductor with the Magnetic Circuit

In the previous section, we had created a custom simulation model of an inductor

using the control interface provided by Python Power Electronics. However, we
had solved the basic L dt
dt equation of the inductor. In this section, we will use the
control interface setup in the previous section and extend the mathematical model
to include the magnetic circuit of the inductor. We will begin with basic simulations
that compute the flux in the iron core and later take into account the shape of the
core and air gaps.
To include the magnetic circuit of the inductor, we need to break up our
computation into several steps. We will still continue to use the model of Fig. 2.9
with the change being that the function g(v) will be replaced by many other
functions. To begin with, we will use Faraday’s Law to express the induced emf
with respect to the applied voltage. The induced emf in turn can be expressed using
Kirchhoff’s Voltage Law:

e = v − ir (2.32)

where r is the parasitic resistance of the inductor as described in the previous section
and is a combination of the winding resistance and also represents the core losses.
The induced emf can then be expressed as the rate of change of flux linkages of
the inductor coil using Faraday’s Law:

dψ dφ
e= =N = v − ir (2.33)
dt dt
We have simplified the above expression further since flux linkage ψ of a coil is
merely the product of flux φ associated with the coil and the number of turns N of
the coil. The physical significance of the above expression is that since the resistance
drop ir is usually negligible, the applied voltage defines the flux in the core that
passes through the coil.
In the past section dealing with the basics of magnetics, we have already
discussed how the flux in the core can be expressed in terms of the MMF of the
coil and the reluctance ρ of the core flux path:

MMF
φ= (2.34)
ρ

The above expression assumes a linear magnetic relationship such as shown in

Fig. 2.5. We can calculate the reluctances of the different segments of the core and
express the core flux as:

MMF Ni
φ= = (2.35)
ρeq ρeq
46 2 Presenting Basic Magnetic Circuits with Inductors

where ρeq is the equivalent reluctance of the core in case it contains an air gap
(Fig. 2.3) or has more than one limb (Fig. 2.4).
From the above expression, calculating the inductor current i results in

φρeq
i= (2.36)
N
In the above expression, we have included details of the inductor core. With this
model, it is now possible to simulate a multi-limb core inductor with an air gap in
one or more limbs as long as the structural details (length of limbs, air gap, cross-
sectional area) are known and the number of turns of the inductor coil is known.
Since, in the case of inductors wound for special purposes, these are usually known,
the above model can be used simulate it directly.
Even at this stage, one can argue that the above model is still fairly simple since
the equivalent inductance of a coil wound on a multi-limb core can be calculated
anyway. However, we are now one step ahead of our basic L di dt inductor model.
For now, let us use this simulation model with an example of the coil wound on
a simple single limb core and then on a multi-limb core with air gaps. Let us
consider a core similar to Fig. 2.3. At first, let us neglect the air gap and consider
the core to be a continuous iron medium. This simulation can be found in the
folder rectangular_core_inductor within the folder magnetic_model
in chapter2_inductors inside the repository:
https://github.com/opensourceelectrical/simulating-magnetics-for-power-
electronics.
Let us assume the core to have the following dimensions—cross-sectional area
Ac = 9 cm2 and mean length of flux path l = 15 cm. Let us assume that the iron
core has a relative permeability of μr = 1000. Let us assume that the coil wound
on the core has 200 turns. This can be translated to the following Python code:
import math
d t = 1 . 0 e−6
r = 0.1
n o _ o f _ t u r n s = 200
c s _ a r e a = 9 . 0 e−4
l e n g t h _ i r o n = 1 5 . 0 e−2
mu_0 = 4∗ math . p i ∗ 1 . 0 e−7
mu_r = 1 0 0 0 . 0

Since in this core, there is single limb without an air gap, the inductance can be
calculated by the formula described before:

N 2 μAc 2002 × 1000 × 4π × 10−7 × 9 × 10−4

L= = = 0.3016H (2.37)
l 15 × 10−2
2.8 Simulating an Inductor with the Magnetic Circuit 47

This design of an inductor should result in a value of inductance very close to

what was used in the previous simulation (0.3 Henry). Our simulation results
should, therefore, not be very different. This would make the subsequent simulations
interesting when we consider cores with multiple limbs and air gaps.
The model computations will however be in multiple steps instead of a single
one. To begin with, we will compute the flux linkages and subsequently the core
flux:
k1 = ( vmeas − i n d _ c u r r e n t ∗ r )
k2 = ( vmeas − ( i n d _ c u r r e n t + d t ∗ k1 / 2 . 0 ) ∗ r )
k3 = ( vmeas − ( i n d _ c u r r e n t + d t ∗ k2 / 2 . 0 ) ∗ r )
k4 = ( vmeas − ( i n d _ c u r r e n t + d t ∗ k3 ) ∗ r )
k = ( k1 + k2 ∗2 + k3 ∗2 + k4 ) ∗ d t / 6 . 0
f l u x _ l i n k a g e += k
flux = flux_linkage / no_of_turns

Here, flux_linkage will need to be defined as a Static variable in the control

configuration to ensure that it can be accumulated between simulation iterations as
required by the process of integration. We are merely integrating the induced emf,
which by Faraday’s Law is the rate of the change of flux linkages of the coil. The
core flux is merely the flux linkages divided by the number of turns of the coil.
We can now use the core dimensions and the magnetic properties defined at the
head of the file:
R _ i r o n = l e n g t h _ i r o n / ( mu ∗ c s _ a r e a )

For now, in the absence of an air gap, the reluctance of the iron core is fairly simple
to calculate. The inductor current can then be updated as:
i n d _ c u r r e n t = f l u x ∗ R_iron / no_of_turns

The entire code can be summarized as:

import math
d t = 1 . 0 e−6
r = 0.01
n o _ o f _ t u r n s = 200
c s _ a r e a = 9 . 0 e−4
l e n g t h _ i r o n = 1 5 . 0 e−2
mu_0 = 4∗ math . p i ∗ 1 . 0 e−7
mu_r = 1 0 0 0 . 0
mu = mu_0 ∗ mu_r

i f t _ c l o c k >= t 1 :
# Runge−−K u t t a F o u r t h Order method
k1 = ( vmeas − i n d _ c u r r e n t ∗ r )
k2 = ( vmeas − ( i n d _ c u r r e n t + d t ∗ k1 / 2 . 0 ) ∗ r )
k3 = ( vmeas − ( i n d _ c u r r e n t + d t ∗ k2 / 2 . 0 ) ∗ r )
k4 = ( vmeas − ( i n d _ c u r r e n t + d t ∗ k3 ) ∗ r )
k = ( k1 + k2 ∗2 + k3 ∗2 + k4 ) ∗ d t / 6 . 0
f l u x _ l i n k a g e += k
flux = flux_linkage / no_of_turns
R _ i r o n = l e n g t h _ i r o n / ( mu ∗ c s _ a r e a )
i n d _ c u r r e n t = f l u x ∗ R_iron / no_of_turns
48 2 Presenting Basic Magnetic Circuits with Inductors

v s r c = vmeas − i n d _ c u r r e n t ∗ s e r i e s _ r e s

indmodel_flux = flux
i n d m o d e l _ e m f = vmeas − i n d _ c u r r e n t ∗ r

t 1 += d t

Simulating with the above model will result in an inductor current that is very
close to the waveform in Fig. 2.11. However, to understand what we have achieved
by introducing a few intermediate computations, we can examine a few other quan-
tities that we had no access to before. For example, we can define VariableStorage
elements in the control configuration by the names of indmodel_flux and
indmodel_emf. Let us assign the flux to indmodel_flux and plot this
variable as shown in Fig. 2.12. Figure 2.12 shows how the core flux follows the
same wave shape as the current since the magnetic flux is directly proportional to
the current flowing through the coil. Figure 2.12 shows the flux being 0 at time t = 0
and attaining a steady state of approximately 15 milli Weber. We could also in the
Python control code compute the flux density B. This will be left as an exercise
to the reader. Plotting the flux density can be useful to know what would be the
maximum flux density that is attained to ensure that this is below the saturation
limit.
Another advantage of computing these in a simulation is that it helps to
understand the theory behind electrical components such as inductors, and in later
chapters, machines such as transformers. Figures 2.11 and 2.12 show how the
current through the inductor and the flux in the core increases gradually until they

Fig. 2.12 Flux in the inductor core

2.8 Simulating an Inductor with the Magnetic Circuit 49

Fig. 2.13 Induced emf produced

reach constant values. Most of us who simulate circuits with inductors rarely stop to
examine what happens at the fundamental level explained using the laws of physics.
To begin with, we are applying a voltage of 100 V to a resistor–inductor combination
at time t = 0. This is a step input to the circuit. However, the current starts at 0. The
reader is encouraged to verify that if the inductor is removed, the current will jump
to the steady state value of 10 Ampere.
The fundamental property of the inductor is that it opposes any abrupt change to
the current flowing through it. This opposition can be inferred from Faraday’s Law,
which states that an emf is induced with a magnitude equal to the rate of change of
flux linkages of the inductor coil and from Lenz’s Law, which states that this induced
emf will oppose the cause that produces it. In the Python code, we have plotted the
induced emf by defining a VariableStorage element indmodel_emf. This induced
emf is shown in Fig. 2.13. As can be seen, the induced emf starts at an initial value
of 100 V. This induced emf of a magnitude equal to the applied voltage and with a
polarity such that it will oppose the flow of current ensures that the current does not
jump at the instant of applying the voltage.
As the current gradually increases, the induced emf gradually decreases until
at steady state, the induced emf is zero. This is fairly intuitive as the highest rate
of change of current is when the current has still not started to flow, while as it
increases, the rate of change will decrease. Eventually, the induced emf is zero, and
the current is limited only by the external resistor of 10  shown in Fig. 2.10. Since
all we have is a dc voltage source connected to a resistor, the result is a dc current,
which does not change. The inductor plays no role in the circuit except for perhaps
its negligible winding resistance and core losses.
50 2 Presenting Basic Magnetic Circuits with Inductors

The simulation results are also helpful in dispelling some misconceptions with
respect to magnetic components. One misconception is that an inductor will offer an
impedance to the flow of current equal to the inductive reactance 2πf L. However, it
is to be noted that the inductive reactance is frequency dependent and in a dc system,
the inductive reactance is 0 in the steady state. It is important to note that an emf
is induced only when the flux linkages of the inductor coil changes, which in turn
implies that the current flowing through it must change. An emf will not be induced
due to the flow of current through it.
Another misconception is that applying a dc voltage to an inductor will cause it to
saturate. In Fig. 2.12, the core flux attains a constant value in steady state. However,
whether this flux in the core will cause saturation can be determined by computing
the flux density in different parts of the core. As shown in Fig. 2.6, every magnetic
material has a value of flux density that defines the saturation level of a core formed
of that magnetic material. Unless the flux density in the core exceeds the value
of flux density at the boundary of the linear portion of the B–H relationship, the
inductor cannot be said to be saturated. Application of a dc voltage alone will not
force the inductor to be saturated. The current that flows through the inductor is also
determined by the external impedance in the circuit. The flux in the core will be
determined by the current flowing through the coil and the reluctance of the core.
In this section, we have gradually included the mathematical model of the
inductor based on the basic laws of physics. Most importantly, we translated these
laws to Python code and used it to simulate a basic inductor as a coil wound on an
iron core. Such a simulation model would help an engineer verify the design of an
inductor as it can also be used to plot quantities such as the flux or flux density in
the core. In the next section, we will examine how the model can be augmented to
account for air gaps and multi-limb cores.

2.9 Practical Cores with Air Gaps and Multiple Limbs

In the previous section, we began programming the basic laws of physics that can
be used to model a magnetic component like an inductor. We had started with the
simplest possible core—a rectangular core with a uniform cross-sectional area and
no air gap. In many cases, inductor cores will have air gaps or multiple limbs. In
this section, we will describe how to include air gaps and multiple limbs into the
simulation model.
Let us simulate an inductor wound as a coil on a rectangular iron core with
an air gap as shown in Fig. 2.3. This simulation can be found in the folder
rectangular_core_with_air_gap within the folder magnetic_model
in chapter2_inductors inside the repository:
https://github.com/opensourceelectrical/simulating-magnetics-for-power-
electronics.
2.9 Practical Cores with Air Gaps and Multiple Limbs 51

l
The core of Fig. 2.3 has two air gaps of length 2g . However, since it is a single
magnetic flux path, we could assume a single air gap of length lg . The code can be
modified to have the following parameters at the head of the file:
l e n g t h _ c o r e = 1 5 . 0 e−2
l e n g t h _ a i r g a p = 0 . 2 e−3
length_iron = length_core − length_airgap

We have assumed an air gap length of 0.2 mm. Therefore, the length of the iron
segment of the core can be calculated from the total length of the flux path and the
air gap.
The calculation of the flux linkages and the resultant flux in the segment of the
core in which the coil is wound remains unchanged. The calculation of the MMF and
the resultant current through the coil will now be with respect to the total reluctance
of the iron segment and the air gap:
R _ i r o n = l e n g t h _ i r o n / ( mu ∗ c s _ a r e a )
R _ a i r g a p = l e n g t h _ a i r g a p / ( mu_0 ∗ c s _ a r e a )
i n d _ c u r r e n t = f l u x ∗ ( R_iron + R_airgap ) / no_of_turns

It should be noted that though the air gap is only 0.2 mm, the permeability of the air
gap is 1000 times lower than the permeability of iron.
This simulation will yield very different results. Figure 2.14 shows the inductor
current. The rate of rise of the current can be seen to be very different. The reader is
advised to calculate the inductance from the calculations presented in the previous
L
section and verify that the settling time is equal to 4 R . Since a dc voltage is applied

Fig. 2.14 Inductor current

52 2 Presenting Basic Magnetic Circuits with Inductors

Fig. 2.15 Flux in the core with air gap

to the R–L combination, the steady state current value is not affected and is still 10
Ampere.
Figure 2.15 shows the flux in the core. The maximum value of flux at steady
state has decreased. This is due to the fact that the steady state value of current is
determined by the 10  resistance. However, the total reluctance of the iron core
is much greater due to the air gap. Therefore, even though the MMF generated is
the same in steady state, the resultant flux is smaller due to the larger equivalent
reluctance of the magnetic circuit. The reader can further expand the results by
plotting the MMF across the air gap and the MMF across the iron segment of the
core to compare how total MMF produced by the coil is used in the magnetic circuit.
The reader is also encouraged to plot the emf and verify that it falls to zero at the
same rate as the flux rises to its steady state value.
Before we proceed to the next simulation, it is worth discussing the simulation
results. To calculate the flux linkages and the flux, we have used Faraday’s Law.
Even though Faraday’s Law equates the induced emf to the rate of change of flux
linkages, for most practical inductor coils with a negligible winding resistance, the
applied voltage is approximately equal to the induced emf. Therefore, at time t = 0,
we are integrating the applied voltage to compute the flux. As the flux increases
from zero with the integral operation in the computation, the current is calculated
according to the MMF needed to overcome the reluctance of the core and the air
gap. As the length of the air gap increases, the MMF and, therefore, the current will
also increase. This accounts for the rapid rise in the current as the MMF needed to
produce the flux is much greater in the presence of the air gap. However, it should
be noted that the current also flows in the external circuit that has a 10  resistor
producing a larger voltage drop across the external resistor, due to which the applied
2.9 Practical Cores with Air Gaps and Multiple Limbs 53

voltage across the inductor will also fall sharply. This lower applied voltage in turn
results in lower flux linkages and core flux that eventually results in a lower steady
state value of core flux.
Now that we have simulated an iron core with an air gap, let us consider a more
complex core such as that of Fig. 2.4. This simulation can be found in the folder
multi_limb_core_with_air_gap within the folder magnetic_model in
chapter2_inductors inside the repository:
https://github.com/opensourceelectrical/simulating-magnetics-for-power-
electronics.
The core has multiple legs and more than one air gap. In the past section, where the
inductor of Fig. 2.4 was presented, the reluctance of the different legs of the core
was also derived from the geometry of the core. We can translate these to Python
code at the head of the control file:
l e n g t h _ a = 5 . 0 e−2
l e n g t h _ b = 4 . 5 e−2
l e n g t h _ c = 4 . 0 e−2
l e n g t h _ w = 3 . 0 e−2
l e n g t h _ l g _ 2 = 0 . 1 e−3

These definitions are similar to the dimensions marked in Fig. 2.4.

The flux linkages of the core will still be computed in exactly the same manner
as until now, as the rate of change of flux linkages will continue to be governed
by Faraday’s Law. Subsequent to the computation of flux linkages, the flux will
also be calculated by dividing the flux linkages by the number of turns of the coil.
However, this flux that has been calculated is the flux passing through the coil. From
Fig. 2.4, this flux is in the left limb of the core on which the coil has been wound. To
determine the MMF that would produce this flux, we can solve the magnetic circuit
as shown in Fig. 2.16.
The flux φ can be expressed as:

MMF MMF
φ= = (2.38)
ρeq ρ1 + ρρ22+ρ
ρ3
3

Fig. 2.16 Equivalent φ

magnetic circuit of
multi-limb core inductor
φ2 φ3
ρ1
ρ2 ρ3
MMF
54 2 Presenting Basic Magnetic Circuits with Inductors

The reluctances of the different segments of the core have already been described in
the past section and can be translated to Python code as:
R1 = ( 2 ∗ l e n g t h _ a + l e n g t h _ b + 2 . 0 ∗ l e n g t h _ w − 2∗ l e n g t h _ l g _ 2 ) / ( mu ∗ c s _ a r e a )
Rg1 = 2∗ l e n g t h _ l g _ 2 / ( mu_0 ∗ c s _ a r e a )
R2 = ( l e n g t h _ b + l e n g t h _ w ) / ( mu ∗ c s _ a r e a )
R3 = ( 2 ∗ l e n g t h _ c + l e n g t h _ b + 2 . 0 ∗ l e n g t h _ w − 2∗ l e n g t h _ l g _ 2 ) / ( mu ∗ c s _ a r e a )
Rg2 = l e n g t h _ l g _ 2 / ( mu_0 ∗ c s _ a r e a )
R_eq = R1 + Rg1 + ( R2 ∗ ( R3 + Rg2 ) / ( R2 + R3 + Rg2 ) )

A potential advantage of calculating the reluctances of each limb of the core

is that we can now compute the flux (and flux density) flowing in each limb.
The computation of the inductor current will now take into account the equivalent
reluctance R_eq:
i n d _ c u r r e n t = f l u x ∗ R_eq / n o _ o f _ t u r n s
indmodel_flux = flux
i n d m o d e l _ f l u x 2 = f l u x ∗ ( R2 ∗ ( R3 + Rg2 ) / ( R2 + R3 + Rg2 ) ) / R2
i n d m o d e l _ f l u x 3 = f l u x ∗ ( R2 ∗ ( R3 + Rg2 ) / ( R2 + R3 + Rg2 ) ) / R3

We can create two more VariableStorage elements indmodel_flux2 and

indmodel_flux3 and assign to them the values of flux in the central limb
and the right limb, respectively.
Figures 2.17 and 2.18 show the simulation results with the inductor current and
the fluxes in the three limbs. The current continues to have a much faster settling
time than the first result of Fig. 2.11. This is due to the smaller inductance owing to
the air gaps in the left and right limbs of the core. Figure 2.18 shows the fluxes in
the three limbs, with the flux φ in the left most limb being the highest as the coil

Fig. 2.17 Inductor current

2.10 Conclusions 55

Fig. 2.18 Fluxes in different core limbs

is wound on that limb. The flux φ2 in the central limb is greater than the flux φ3 in
the right limb as the right limb has an air gap leading to it being the path of greater
reluctance. The reader is also encouraged to verify that φ = φ2 + φ3 as the flux φ
in the left most limb divides into φ2 and φ3 .
With these simulations, we have simulated inductors wound on cores with
multiple limbs with air gaps. The purpose of these simulations was to demonstrate
how a simulation model can be formulated based on the laws of physics. Such a
simulation can be used to verify the design of an inductor that might be wound for
any given application, as the simulation takes into account the number of turns of
the inductor coil and the construction of the inductor core. The simulation enables
us to plot variables such as the flux in different segments of the core and thereby
determine if there is a possibility of the core to be saturated.

2.10 Conclusions

In this chapter, we have introduced the concept of magnetic circuit modelling by

simulating inductors. Besides a theoretical introduction to magnetic circuits, this
chapter provided simulations of circuits with inductors. This chapter might seem
trivial and useless as in most cases, one cannot imagine a need to simulate the
magnetic circuit of an inductor. Therefore, in concluding this chapter, let us discuss
what has been gained from this chapter as well as what are its shortcomings.
Magnetism as a subject has been studied extensively and this chapter has merely
used some of the most basic laws and expressions related to magnetic fields in
56 2 Presenting Basic Magnetic Circuits with Inductors

electric circuits. We used Ampere’s Law to state that the inductor coil will produce
a magnetic field and used the Right Hand Thumb Rule to determine the direction of
the magnetic lines of flux. For a few types of inductor cores, we illustrated the path
of the magnetic flux. It is important to note that determining the magnetic flux is
fairly complex and requires a very different approach. In this chapter, we assumed
the magnetic field to be uniform and neglected all other effects—non-uniformity,
fringing etc. The purpose of describing the magnetic field produced in an inductor
core was to introduce the concept of the magnetic circuit.
With a few examples, we had shown how a magnetic circuit can be solved in a
very similar manner to an electric circuit. We had introduced the concept of magneto
motive force (MMF), reluctance of the flux path, magnetic field strength and flux
density. We had shown how for the case of a simple coil wound on a core, we could
use the MMF generated along with the reluctances of the core segments to determine
the flux flowing in all the segments of the core. With flux being analogous to current,
we found that the reluctances of the core segments can be treated in exactly the same
way as resistances in an electric circuit. Despite the fact that we have neglected all
the non-linearity in the magnetic field, to be able to determine the flux in every part
of the core given a particular current flowing through a coil leads to a better picture
of the operation of an inductor.
Following a discussion on magnetic circuits, we had described how the magnetic
circuit of an inductor can be simulated. Eventually, we need an inductor model
that can be inserted into a circuit. The chapter described how the inductor can be
modelled as a combination of a Voltmeter, a controlled voltage source and a variable
resistor. We described how the control loop takes as an input the voltage across the
inductor terminals and solves a mathematical model that produces a current that will
flow through the inductor. This current is produced by modifying the output of the
controlled voltage source with respect to the voltage measured across the terminals.
Several simulations were presented where we gradually increased the complexity
of the mathematical model of the inductor, until the final simulation presented the
solution of the magnetic circuit whereby the flux and subsequently the flux density
in every part of the core can be computed.
With this chapter, we have described how the basic laws of physics can be
used to produce a simulation model for the inductor. It is important to emphasize
that this simulation model does not model the non-linear aspects of magnetism.
However, for most power electronics applications, a basic insight into the magnetic
circuit of an inductor will be sufficient to determine the operation of an inductor in
certain applications. As an example, in power supply applications, engineers find
themselves winding inductors by hand. The simulation models in this chapter can
be used to roughly estimate how such hand-wound inductors would behave in power
converters. Besides being able to determine the current flowing through the inductor,
other details such as the flux density in the core can be determined as well.
The main purpose of this chapter, however, was to serve as a gradual introduction
to the more complex magnetic circuits that will be found in machines such as in
transformers that will be covered in the later chapters. By beginning the discussion
on magnetic fields, we have laid the foundation for the concept of energy transfer
2.10 Conclusions 57

through the magnetic field. In the next chapter, we will begin the discussion on
magnetic coupling between multiple coils wound on a core. The mathematical
model of coupled inductors will be an extension of the single inductor presented in
this chapter using the same basic laws of physics presented in this chapter. The next
chapter on coupled inductors is also an intermediate chapter before we progress to
simulating transformers. By meandering through this concept of magnetic circuits,
we wish to provide the reader an insight into the importance of magnetism in
electrical engineering, and how it can be understood and modelled in a fairly simple
manner using only the most basic laws of physics.
Chapter 3
Simulating Magnetically Linked Circuits

3.1 Introduction

In the previous chapter, we described how we can simulate a basic inductor while
including some details of the construction with respect to the core structure and
the number of turns of the inductor coil. With inductors being one of the most
fundamental components in any power electronic converter, the ability to simulate
an inductor in detail will be beneficial to power engineers who might have to wind
their own inductors for specific applications. In this chapter, we will extend our
knowledge of magnetic circuits and the flux produced in the core due to current
flowing in the inductor coil to examine the behaviour of magnetically coupled coils.
A vast number of machines in electrical engineering use the concept of magnetic
coupling to achieve transfer of power [1–4]. The next chapter will be devoted to
transformers, which will describe how a transformer can achieve transfer of power
from one isolated electrical system to another [10, 26]. Besides transformers, motors
and generators also use the concept of magnetic coupling to achieve conversion of
electrical energy into mechanical energy and vice versa [3, 4]. Therefore, before
describing the operation of transformers, this chapter is devoted to describing the
concept of magnetic coupling. Even though magnetic coupling is a mere extension
of the basic laws of physics already described in the previous chapter, there are
several nuances to how magnetically coupled circuits can behave. In this chapter,
using a simple setup of two coils wound on a core, the nature of magnetic coupling
is described using both theory and simulations [11].
All the basic laws of physics (Ampere’s Law, Faraday’s Law and Lenz’s Law)
are applicable in the case of more than one coil wound on a core [22, 24]. We will
describe how the fundamental nature of magnetic coupling is the fact that emfs
will be induced in all the coils wound on a core through which a magnetic flux is
flowing [11]. Therefore, by energizing a single coil wound on a core, energy can
be transferred to other coils wound on the core. Furthermore, by using Faraday’s
Law and Lenz’s Law, we can decipher the nature of the emfs induced in all the

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 59

S. V. Iyer, Modeling and Python Simulation of Magnetics for Power Electronics
Applications, https://doi.org/10.1007/978-3-030-96768-0_3
60 3 Simulating Magnetically Linked Circuits

coils and subsequently, the current that will flow through them. We can continue to
translate these systems to magnetic circuits and solve them to determine the system
operation. A number of examples are presented to illustrate these concepts in detail.
In order to be able to express magnetically coupled circuits in a concise manner,
we define the mutual inductance between two coupled coils [1, 2, 10, 26]. We
describe how the mutual inductance between coils along with the self-inductances
of the coils results in concise expressions of flux linkage and subsequently of
induced emfs. To eliminate the need to determine the polarity of induced emfs based
on the physical winding sense of coils, we describe the dot polarity convention
[1, 2, 10, 26]. Using the dot polarity convention, we describe how magnetically
coupled systems can be represented using simple schematics with the only addition
being dots on the terminals of inductors to indicate the nature of coupling between
inductors. Using self and mutual inductances along with dot polarities, we describe
how detailed simulation models can be developed for several cases of coupled
inductors.
Several simulations will be described where the above concepts are translated
to practical Python code. Simulation results are analysed so as to compare the
simulated behaviour with what we expect from theoretical results. To conclude this
chapter, we will present a scalable simulation model using matrix equations with
which we can simulate any number of coupled inductors. In the next chapter, this
scalable simulation model will be used to simulate transformers. The next chapter
based on transformers will examine the practical application of transfer of energy
through magnetic coupling. This chapter will serve as an intermediary chapter where
magnetic coupling is formulated mathematically extending on our knowledge of
magnetic circuits presented in the previous chapter on inductors. As in the previous
chapter, theory will be supported by simulations and code samples that the reader
can implement and tinker with.

3.2 The Concept of Magnetic Coupling

In the previous chapter on simulating inductors, we have examined how a coil

wound on an iron core produces a flux that flows through the entire core [22, 24].
We defined the reluctance of a magnetic medium, which we showed was dependent
on the permeability of the magnetic material, the length of the flux path through the
medium and the cross-sectional area of the flux path. We also examined how flux
can be calculated using magnetic circuits just like the current in an electric circuit,
with the MMF being the driving force and the reluctance of the different parts of the
magnetic core perceived as series/parallel combinations.
In the previous chapter, when we had described Faraday’s and Lenz’s Laws to
determine the magnitude and polarity of the emf induced in an inductor, we had only
considered the current through the inductor, which was responsible for producing
3.2 The Concept of Magnetic Coupling 61

Fig. 3.1 Two coils wound on

the same core
i

the flux. In this chapter, and this section, we will get started with the concept of
magnetically coupled coils [11]. As an example, consider the same rectangular core
as in the previous chapter as shown in Fig. 3.1. However, let us assume that another
coil is wound on the other leg. The coil wound on the left limb is connected to an
external circuit, which is not shown in the figure for simplicity, but this external
circuit results in a current i flowing through the coil. The second coil wound on the
right limb is left open-circuited without any connection either to the coil on the left
limb or to any other external circuit.
As before, the current flowing through the first coil will produce a magnetic flux
that flows through the core. The direction of the flux shown by arrows within the
core can be determined by using the Right Hand Thumb Rule with respect to the
current flowing through the coil. The face of the coil through which current flows
in the counter-clockwise direction forms the North pole of the coil and the other
face of the coil forms the South pole. This flux flows not only through the coil
wound on the left limb that produces it but also with the second new coil wound on
the right limb. The two coils that have no electrical connection are now said to be
magnetically linked. Faraday’s and Lenz’s Laws apply to both coils. If the current
through the first coil is changing, the resultant flux flowing through the core will
also be changing. Therefore, this changing flux will induce an emf in the first coil
as well as the second coil, with the magnitudes of the emfs determined by Faraday’s
Law, i.e. directly proportional to the rate of change of flux linking the coils.
Before we progress to developing the mathematical model for the entire system,
let us express the magnetic coupling mathematically [10, 11, 26]. If the flux in the
core is denoted by φ, the coil of the left limb has N1 turns and the coil on the right
limb has N2 turns, the flux linkages of the two coils can be expressed as:

ψ1 = N1 φ (3.1)
ψ2 = N2 φ (3.2)
62 3 Simulating Magnetically Linked Circuits

The magnitude of the emfs induced in the coils is determined by Faraday’s Law:

dψ1 dφ
e1 = = N1 (3.3)
dt dt
dψ2 dφ
e2 = = N2 (3.4)
dt dt
As can be observed from the above equations, both the induced emfs e1 and e2
are proportional to the rate of change of core flux dφ dt . Therefore, the magnitude of
the induced emfs only differ by the number of turns N1 and N2 of the two coils. This
fact shows the nature of coupling between the two coils. Each coil has an induced
emf with a magnitude that follows the rate of change of core flux. So far we have
examined the magnitude of the induced emfs, but what about the polarity? As in the
case of inductors in the previous chapter, it is better to determine the polarity of the
induced emfs in the circuit rather than express it in the equations. The polarity of
the induced emfs can be determined by using Lenz’s Law, which we have already
used in the previous chapter while simulating inductors. The induced emf in a coil
will be in such a sense that it opposes the cause that produces it. The cause of the
induced emf is the changing core flux passing through the coil. Therefore, both e1
and e2 will oppose the change in flux φ.
With the case of the simple inductor in the previous chapter, we never had to
delve deep into the polarity of the induced emf. The induced emf was merely one
of the intermediate computations while calculating the current flowing through the
inductor. However, in the case of coupled coils, we have many interesting cases
as will be examined soon. Therefore, the polarity of the induced emf is now fairly
important, particularly in the second coil wound on the right limb. To determine
the polarity of the induced emf, we need to ask the question—how can the induced
emf oppose the change in core flux that produces it? Answer—the induced emfs
have to produce core fluxes that oppose the change in the core flux. Here again, it is
important to emphasize that the induced emf will oppose the change in the core flux
and not the core flux itself.
To determine the induced emfs in the two coils, we note that the coil wound on
the left limb is energized by an external voltage v, while the coil wound on the
right limb is left open-circuited. Therefore, for the coil wound on the left limb, the
following expression can be written as:

v − ir1 = e1 (3.5)

where r1 is the winding resistance of the coil wound on the left limb. In the case of
the coil wound on the right limb, since it is open-circuited, there will be no current
flowing, nor is there an external voltage source connected to it. However, if the
voltage across the terminals of the coil wound on the right limb is measured, it will
be the same as the induced emf e2 as there is no resistive drop with the current being
zero.
3.2 The Concept of Magnetic Coupling 63

From (3.5), it appears as though the induced emf e1 opposes the current i flowing
through the coil wound on the left limb. This is true as Kirchhoff’s Voltage Law is
always applicable in any closed loop, and the fundamental basis of the law is that
the sum of voltage sources and voltage drops in a closed loop will be zero. Since the
resistive drop ir1 is usually negligible for most practical coils, the main opposition
to the applied voltage is the induced emf. It is the induced emf that limits the flow
of current through the coil when a voltage is applied across it. Even despite the
fact the induced emf seems to be the only opposition to the flow of current, it is
important to emphasize that the induced emf will not always be there to oppose
the flow of current. The induced emf will only be present when the magnetic flux
associated with the coil changes with respect to time. As an example, if the applied
voltage were to be a dc voltage, in steady state, the induced emf will be zero once
the current rises and settles to the steady state value. We have examined this in the
previous chapter using simulations.
Equation (3.5) can still be used to answer the question—how will the induced emf
oppose the cause that produces it (the changing core flux)? If we were to remove
the applied voltage v by short-circuiting it, we now can examine the effect of the
induced emf e1 alone. Figure 3.2 shows the effect of the induced emf alone. Since
the supply voltage has been replaced by a short-circuit, the direction of current has
reversed. This current is denoted by i  , as it is not the same as the actual coil current
in Fig. 3.1. Due to the reversal of the current, the core flux has also reversed direction
with respect to the core flux in Fig. 3.1. This can be verified by applying the Right
Hand Thumb Rule. Therefore, the induced emf attempts to oppose the changing
core flux by producing a changing flux that flows through the core in the reverse
direction.
A deeper investigation will show the true nature of the induced emf [10, 26].
According to Faraday’s Law, the induced emf is directly proportional to the rate of
change of core flux dφ dφ
dt . If this dt were positive, it would mean that the core flux
is increasing in the direction shown in Fig. 3.1. This would imply that the induced
emf e1 would also have a rate of change that is positive, i.e. de 1
dt > 0. From Fig. 3.2,

such an induced emf will produce a current i , which will also have a positive rate of

Fig. 3.2 Effect of induced

emf on the core flux
i

v=0 e1
64 3 Simulating Magnetically Linked Circuits


change, i.e. di 
dt > 0. An increasing current i will produce an increasing flux in the
downward direction. Therefore, the increasing core flux is opposed by an increasing
induced emf that tries to produce an increasing core flux in the reverse direction.
The reader is encouraged to try out the above steps in the case when the core flux in
Fig. 3.1 were to be decreasing.
In the above case of examining the effect of the induced emf, we ended up
combining Faraday’s Law and Kirchhoff’s Voltage Law to prove that Lenz’s Law is
applicable—that the induced emf will be such that it opposes the cause that produces
it. We began with the argument that if an external voltage source is connected to the
coil, by Kirchhoff’s Voltage Law, the induced emf will oppose the applied voltage
as the only voltages in the closed loop are the applied voltage, the resistive drop
and the induced emf. If we wish to determine the nature of induced emf in the coil
wound on the right limb, we can no longer use Kirchhoff’s Voltage Law, as the coil
on the right limb is left open-circuited. However, even if the coil is open-circuited,
the induced emf in the coil will abide by Faraday’s Law and Lenz’s Law.
We have already expressed the equation for the magnitude of the induced emf as
directly proportional to the rate of change of core flux. We need only to determine
the polarity of the induced emf. Figure 3.3 shows the coil wound on the right limb.
To determine how the induced emf can oppose the cause that produces it, we need
make an assumption. This is due to the fact that when the coil is open-circuited,
no current can flow through the coil and therefore, nothing can be done to oppose
the change in core flux, which is the cause that produces it. Let us assume that the
coil is short-circuited as shown by dashed lines. If we assume the coil to be closed,
a current can now flow through the coil and this current can produce a flux that
opposes the change in core flux. From Fig. 3.1, the core flux passes through the coil
in the downward direction. Therefore, in order to oppose the change in the core flux,
the current flowing through the coil must be such that it produces a flux that passes
through the coil in the upward direction as shown in Fig. 3.3. This is possible only
if the induced emf e2 has a polarity as shown in Fig. 3.3, which will result in the
current i2 .

Fig. 3.3 Induced emf in the

coupled coil
i2

e2
3.3 The Nature of Magnetic Coupling 65

The induced emf e2 can be quantified with respect to the rate of change of core
flux dφ
dt as done for the coil wound on the left limb of the core. Let us consider what
happens when the core flux is decreasing, i.e. dφ dt < 0. In this case, the induced
emf e2 will have a magnitude that is negative. With the polarity of e2 as shown
in Fig. 3.3, an induced emf of negative magnitude would result in a current that
is negative. A negative coil current would try to produce a flux in the core that is
negative, which would imply that for the direction shown in Fig. 3.3, the flux would
be in the downward direction. This is the same direction as the core flux. This might
at first glance seem like a mistake, as we expect the induced emf e2 to produce a flux
that will oppose the core flux. We would like to reiterate that the induced emf will
not oppose the core flux, but instead it will oppose the change in core flux. The core
flux is decreasing and therefore, the induced emf will attempt to produce a flux that
will cause it to increase. Therefore, the induced emf attempts to produce a flux in
the same direction as the core flux, thereby opposing the decrease in the core flux.
The reader is encouraged to examine the case when the core flux is increasing, i.e.
dφ
dt > 0.
In this section, we examined in detail how the basic laws of physics can be
used to show how two coils whose magnetic fields interact with each other end
up being magnetically coupled. Due to one of the coils being energized by a voltage
source, an emf is induced in both coils. The nature of magnetic coupling might seem
confusing to begin with. In the next section, we will examine several different cases
with the same example of two coils wound on a rectangular core. Through these
cases, the concept of magnetic coupling will be clearer to the reader after which we
can proceed to developing the simulation model.

3.3 The Nature of Magnetic Coupling

In the previous section, we had examined how two coils wound on the same core
become magnetically coupled because the magnetic field produced by one of them
interacts with the other. We had used one specific case to examine how emf is
induced in the two coils. As we progress to transformers in the next chapter, the
nature of magnetic coupling can become a bit tricky to fully understand. Therefore,
it would be better to examine through a number of different examples how coupled
coils behave. We will consider the same case of the two coils wound on the
rectangular core in Fig. 3.1 as in the previous section, but we will look at several
different scenarios.
For each of the cases in this section, we will use the basic laws of physics to
determine the nature of coupling, specifically, Faraday’s Law and Lenz’s Law. The
reader should be familiar with these laws and most importantly the nuances related
to these laws [22, 24]. Faraday’s Law states that a changing magnetic field associated
with a coil (or conductor) results in an emf induced in the coil, and the induced emf
is equal to the rate of change of flux linkages of the coil. It is important to emphasize
66 3 Simulating Magnetically Linked Circuits

i1 i2

e1 e2 R
v

Fig. 3.4 Coupled coils

that the magnetic field must change for the emf to be induced. Lenz’s Law states
that the induced emf will be such so as to oppose the cause that produces it. It is
again important to emphasize that the cause of the induced emf is the changing
magnetic field and not the magnetic field itself. Armed with these two laws, let us
now examine several interesting cases with the two coils wound on a rectangular
core.
Let us expand on Fig. 3.1 by adding some components as shown in Fig. 3.4. The
coil on the left limb is connected to an external voltage source v. There could also
be a series impedance R-L-C circuit with an impedance of Z1 along with the voltage
source, but for simplicity, we have neglected the equivalent impedance for now. The
coil on the right limb is connected to a resistor that can be thought of as a dump
resistor. The reader can repeat the logic used in the previous section to arrive at
similar conclusions. Emfs e1 and e2 are induced in the left and right limb coils,
respectively. The currents i1 and i2 will be flowing through the two coils. The flux
in the core will be flowing in the direction shown by arrows.
At first glance, by energizing the coil on the left limb, we are causing energy to
be produced in the coil on the right limb, which is dissipated in the dump resistor R.
Through magnetic coupling between the coils, energy is transferred from one coil
to the other through the medium of the magnetic field. We can write Kirchhoff’s
Voltage Law for the coil on the left limb:

v − i1 r1 − e1 = 0 (3.6)

If we assume the circuit to be completely de-energized initially (at t = 0), the

current i1 at t = 0 is 0 and therefore, the induced emf e1 = v. Another approach
to this relation is to use the fact that the coil has an inductance that will not allow
the current through to it to change rapidly. Therefore, the current i1 through the coil
will change gradually from 0. As a result, the induced emf e1 will be initially equal
to the applied voltage v and will gradually differ as the current flows through the
coil.
3.3 The Nature of Magnetic Coupling 67

As the current i1 flows through the coil in the left limb, a flux will build up in the
core flowing in the direction shown. How can we determine this flux? We could use
the magnetic circuit that we used in the previous chapter while simulating inductors.
The net MMF will be N1 ii − N2 i2 . If we know the dimensions of the core and the
relative permeability of the core material, we can calculate the reluctance of the
core and subsequently the flux. However, instead of using this method, we will use
another method that will be more convenient as we deal with multiple coupled coils.
According to Faraday’s Law, the induced emf is equal to the rate of change of flux
linkages:

dψ1 dφ
e1 = = N1 (3.7)
dt dt
Therefore, the flux in the core can be expressed as:
 
1 1
φ= e1 dt = (v − i1 r1 )dt (3.8)
N1 N1

In the above equation, since the resistor r1 is a negligible winding resistance of

the coil, the quantity that primarily influences the core flux is the applied voltage v.
At this point, we must remember that we have neglected any equivalent impedance
in series with the applied voltage. If there was such an impedance (denoted by say
Z1 ), the applied voltage v in the above equation would be expanded to v − i1 Z1 .
In the simulations of the inductor in the previous chapter, we had considered a
resistance connected in series with the inductance being simulated. As the equivalent
impedance Z1 becomes larger, the influence of the applied voltage v will decrease.
For now, let us assume the applied voltage v to be an arbitrary voltage—not
necessarily a pure dc or a pure ac sine waveform of a particular frequency. If the flux
φ due to the above integral equation is not constant but is changing, by Faraday’s
Law, an emf will be induced in the coil in the right limb due to the flux flowing
through it. This induced emf can be expressed as:

dφ
e2 = N2 (3.9)
dt
This induced emf will result in a current flowing through the coil in the right limb
expressed by Kirchhoff’s Voltage Law as:

e2 − i2 r2 − i2 R = 0 (3.10)

As we have already discussed in the previous section, according to Lenz’s Law,

the induced emf e2 will be of such a nature that it will oppose the cause that produces
it. In the previous section, however, we had considered the coil in the right limb to
be open-circuited due to which, even though an emf was induced in it, no current
flowed through it. In the circuit of Fig. 3.4, the coil in the right limb feeds a dump
resistor and therefore, a current is flowing through it. The reader is encouraged to use
68 3 Simulating Magnetically Linked Circuits

the same logic as in the previous section to verify that the polarity of the induced
emf e2 is as shown in Fig. 3.4. However, now an important question needs to be
asked—if the induced emf e2 is of such a nature that it opposes the change in core
flux and results in a current i2 flowing, how will the core flux be affected?
The MMF produced by the coils in the left limb and right limb are N1 i1 and
N2 i2 , respectively. Given the direction of currents flowing in the two coils, they
would be trying to produce fluxes flowing in the opposite directions. The reader is
encouraged to verify this fact by applying the Right Hand Thumb Rule to the two
coils with their currents in the direction shown. It is very important to emphasize
that we cannot conclude using Lenz’s Law that the fluxes produced by the currents
in the two coils will oppose each other. According to Lenz’s Law, the induced emfs
and subsequently the currents will oppose the changing core flux and not the flux
itself. The fact that the fluxes produced by the two coil currents oppose each other
is the conclusion of using the magnetic circuit once the directions of currents in the
coils have been determined.
Now that we have established that the currents in the two coils will tend to
produce opposing fluxes in the core, it is clear that the current i2 flowing in the
coil wound on the right limb will tend to demagnetize the core by opposing the
flux produced by current i1 flowing in the left limb. For this magnetic circuit, the
net MMF will be N1 i1 − N2 i2 . It is therefore quite natural to expect the flux in the
core to decrease due to this decreased MMF. However, we had expressed the core
flux using (3.8). If the voltage v remains the same, the core flux will also remain
approximately the same (neglecting the drop i1 r1 ). Here arises a confusion—if the
core flux will not change, what effect will the current i2 have?
We need to remember that when coils are wound together in such a manner that
their magnetic fields interact with each other, energy is transferred from one coil to
the other [10, 11, 26]. Energy is not produced in the coil wound on the right limb
just because it is affected by the magnetic field produced by the coil wound on the
left limb. This can only happen if we account for the energy transferred from the
coil on the left limb to the coil on the right limb. If the flux in the core is not going
to change, the net MMF must remain at the value of N1 i1 . The MMF in the coil on
the right limb N2 i2 must additionally flow through the coil on the left limb to ensure
that the net MMF remains the same. The net MMF can be expressed as:

N2
N1 i1 + N2 i2 − N2 i2 = N1 i1 + N1 i2 − N2 i2 (3.11)
N1

As a result, the current flowing through the coil wound on the left limb has
increased to:
N2
i1 = i1 + i2 (3.12)
N1

This increase in the current in the coil on the left limb is understandable as it supplies
the energy consumed by the dump resistor R in the coil on the right limb. The
3.3 The Nature of Magnetic Coupling 69

i1 i2

e1 e2 R
v

Fig. 3.5 Changing the sense of winding of one of the coils

second term in the above equation is called the transformed current [10, 26]. Since
the current in the coil wound on the left limb has increased, by (3.8), the core flux
will decrease. How much the core flux will decrease will depend on the current i1
and if there is an equivalent impedance Z1 in series with the applied voltage v.
Let us examine a few variations to the above coupled coils to dig deeper into
the nature of magnetic coupling. In Fig. 3.4, we used Lenz’s Law to determine the
polarity of induced emf e2 in the coil wound on the right limb. What if we change
the sense in which the coil on the right limb is wound as shown in Fig. 3.5? As
before, applying Lenz’s Law, we can state that the induced emf e2 must be such
that it opposes the change in core flux. Let us consider the case of the core flux to
be increasing, i.e. dφdt > 0. In that case, the induced emf e2 will be increasing in
magnitude. As a result, the current i2 that will flow through the coil on the right
limb will also be increasing in magnitude. The flux produced by this current will be
an increasing flux. By Lenz’s Law, this flux produced by i2 must oppose the change
in the core flux. In order for this opposition to occur, the flux produced by i2 must be
in a direction opposite to that of the main core flux. For this to happen, the polarity
of the current i2 and therefore the induced emf e2 will be as shown in Fig. 3.5.
These polarities are exactly the opposite of the polarities in Fig. 3.4. Therefore, by
changing the sense of winding of a coil, the polarity of the voltage supplied to the
dump resistor R can be reversed.
Let us consider another case where we use the same coupled coils in Fig. 3.4,
but instead of connecting a dump resistor R to the coil on the right limb, we short-
circuit the coil instead. This is shown in Fig. 3.6. The equations related to the coil
on the left limb will remain the same. However, the equations related to the coil on
the right limb will change as follows. Applying Kirchhoff’s Voltage Law on the coil
on the right limb gives us

e2 − i2 r2 = 0 (3.13)
70 3 Simulating Magnetically Linked Circuits

i1 i2

e1 e2 R=0
v

Fig. 3.6 Short-circuiting one of the coils

With the winding resistance of most practical coils being negligible, the above
expression can be approximated as:

e2 ≈ 0 (3.14)

The equations that link the two coils will also remain the same as the nature of
their coupling has not changed. Therefore, the induced emfs of the coils can still be
expressed by Faraday’s Law as (3.7) and (3.9). The nature of the coupling of the
coils is exhibited by the fact that both induced emfs are now proportional to the rate
of change of core flux. If one of the emfs were to be zero, the other emf would also
be zero. This would imply

e1 ≈ 0 (3.15)

Therefore, by short-circuiting the coil on the right limb, we have effectively

short-circuited the coil on the left limb as well. The current i1 will be limited only
by the equivalent impedance Z1 in series with the applied voltage. If this equivalent
series impedance Z1 were to be negligible, the current i1 would be dangerously high
in magnitude and would usually burn the wire of the coil wound on the left limb.
If, however, there were a substantial equivalent series impedance Z1 , the current i1
would be limited. A very interesting question to be asked in this case would be—
what would the core flux be? The core flux would still be determined using Faraday’s
Law and can be expressed by (3.8). If the flux is the integral of the induced emf that
is negligible, the core flux would therefore be negligible as well.
Let us examine another case. In the coupled coils of Fig. 3.4, let us remove
the iron core leaving the coils wound in air as shown in Fig. 3.7. By removing
the core, we have removed the magnetic medium that was providing a relatively
low permeability path for the flux to flow and link the two coils. A core made
of laminated iron would have a relative permeability (μr ) of approximately 1000
compared to the relative permeability of air being 1. With the permeability of the
medium decreasing by a factor of 1000, the flux produced will drastically decrease.
3.3 The Nature of Magnetic Coupling 71

i1 i2

e2 R
v e1

Fig. 3.7 Removing the core

i1 i2

e1 e2 R
v

Fig. 3.8 Applying a dc voltage

For a given external voltage applied to the coil on the left, the induced emf will
much lower due to the fact that the rate of change of flux will also be much lower
at a diminished flux. Moreover, depending on the physical separation between the
two coils, the flux linking the right coil will be lower and in most cases, almost
negligible. As a result, a negligible emf e2 will be induced in the coil on the right
and a very large current will flow in the coil on the left due to e1 being smaller in
magnitude. Unless the two coils are almost merged together, the two coils can be
said to have no practical coupling.
Let us examine one last case before we end this section. In the case of Fig. 3.4,
let us connect a dc voltage such as a dc battery to the coil wound on the left limb as
shown in Fig. 3.8. As before, we can use (3.8) to determine the flux produced. Since
we are integrating v − i1 r1, for a practical coil with a negligible winding resistance
r1 , we are integrating the applied voltage. If v is a dc voltage, it is quite obvious that
the flux will very soon increase to values that will saturate the core. The induced
emf e1 starts with a value equal to the applied voltage v as the coil will not allow an
abrupt rise in the current through it. As the core begins to saturate, the induced emf
will decrease as it is proportional to the rate of the change of flux, thereby causing
72 3 Simulating Magnetically Linked Circuits

the current i1 to increase. Eventually, the current drawn will increase until it will
burn the coil.
There is, however, an interesting case, if we consider the applied voltage v to be
in series with an equivalent series winding Z1 . In such a case, the effective applied
voltage to the coil changes from v to v − i1 Z1 . The core flux will increase due
to the applied dc voltage and the current i1 flowing through the coil will increase
as well in order to produce the MMF N1 i1 that will sustain the core flux. As the
current i1 increases, the voltage v − i1 Z1 applied to the coil decreases. If Z1 is
significantly large, the applied voltage will reach a negligible value, and the flux
will stop increasing and remain constant.
If this core flux does not cause the core to saturate, this will not cause harm to
the coil or the core. If, on the other hand, Z1 is not very large, the core flux will
increase to values that cause the core to saturate. The flux may settle at some value
in the saturation zone of the B–H curve and the current drawn will be proportionate
in order to sustain this flux. Even if such a current does not result in the coil being
damaged, such a state of operation where the core is saturated and currents are high
is not recommended as the excessive losses can deteriorate the magnetic properties
of the core permanently. Therefore, applying a steady dc voltage on a coil is not
recommended and if one wishes to do so, a series impedance must be connected
to ensure that the core does not saturate. In either case, whether the core saturates
or not, the flux will attain a constant value at steady state, and the induced emfs e1
and e2 will become zero. Therefore, in the case of a dc voltage applied to a coil, no
energy will be transferred to a coil that is magnetically coupled to it.
In this section, we examined several possible cases with the two magnetically
coupled coils. The purpose of this section was to understand in detail the nature of
magnetic coupling using the basic laws of physics. The purpose behind including
this section was for the reader to get comfortable with the usage of physical laws
to understand how coupled coils behave. In the later sections and also the later
chapters, this fundamental nature of coupling will form the basis of transformers.
In the next section, we will derive a generalized set of equations for magnetically
coupled coils.

3.4 Mutual Inductance

In the previous section, we had used the basic laws of physics to understand how
magnetically coupled coils behave. We were able to understand how the coils would
behave under different conditions from mathematical equations based on these
basic laws of physics. However, in order to develop a simulation model of coupled
inductors, we need a rigorous set of equations that will capture all these physical
phenomena and that can be applied to any possible condition. In this section, we will
begin to express the magnetic coupling between inductors mathematically, which in
turn can be used to develop a simulation model in the later sections.
3.4 Mutual Inductance 73

In the previous chapter, for an inductor, we can express the flux linkage ψ with
respect to the current i using the property of the inductance as [1, 2]:

ψ = Li (3.16)

This is just how the property of inductance is defined. An inductance of 1 Henry

is the inductance that will produce a flux linkage of 1 Wb-turn when a current of
1 Ampere is flowing through the inductor. Since we are relating the current to the
flux linkage, we are taking everything into account—number of turns, permeability
of the core, mean length of flux path, cross-sectional area of the core. This makes
inductance a very convenient property as it enables us to represent many variables
with just one. Because once we have wound an inductor on a core, you do not change
anything physically, only the current through the coil changes.
In the case of coupled coils, the above equation will change as for a particular
coil, the flux linkages are impacted not only by the current through that coil but
also by the current through the other coil that is magnetically coupled to this coil.
For now, we will consider the example of two coils wound on a rectangular core as
shown in Fig. 3.9. The flux linkage of the coil wound on the left limb (let us call it
coil 1) is expressed as:

ψ1 = L1 i1 ± f (i2 ) (3.17)

The first term is exactly the same as before—inductance of coil 1 and current
i1 through coil 1 affecting the flux linkages of the coil. However, because this coil
1 is magnetically linked to the coil wound on the right limb (let us call it coil 2),
any current i2 flowing through coil 2 will also impact the flux linkages of coil 1.
In the previous section, we had considered specific cases with coil 1 energized by
an external voltage and coil 2 not being independently energized. In this section,
we are assuming the two coils to have their own independent external circuits that
result in currents of i1 and i2 to flow through them. We have included the effect of
the current i2 in coil 2 as ±f (i2 )—some arbitrary function f of the current in coil
2. Moreover, we do not know if the magnetic field produced by the current in coil
2 will assist or oppose the magnetic field produced by the current in coil 1. Though

Fig. 3.9 Magnetically

coupled coils
i1 i2

v1 v2
74 3 Simulating Magnetically Linked Circuits

in Fig. 3.9, we can determine if the magnetic fields produced by the currents will
assist or oppose each other from the sense of their winding, let us ignore the actual
winding details for now.
Before we proceed to detailed mathematical modelling, we need to clarify about
the circuits connected to the two coils. In the previous section, coil 1 was the primary
coil that would result in an induced emf in coil 2, thereby resulting in transfer of
energy from coil 1 to coil 2. This is a very specific case. However, to make the
discussion in this section more general, we would like to bring in the condition
where any circuit could be connected to coil 1 and coil 2. Both of them could be
energized by voltage sources and could have other interfacing components as well.
What we are interested in is expressing through equations the impact of the current
in one coil on the other coupled coil, due to the fact that they are magnetically
coupled.
In such a case, where some arbitrary currents i1 and i2 are flowing through the
coils, coil 1 and coil 2 are both trying to produce fluxes in the core. The direction
of the fluxes can be determined for the currents flowing in coil 1 and coil 2 using
the Right Hand Thumb Rule. In Fig. 3.9, for the manner in which the two coils are
wound, if the currents are flowing in the two coils as shown, the result is that the
fluxes produced by the two coils end up flowing in the same direction through the
core. One could imagine another case, where the current flowing through one of the
coils is reversed. What would happen in that case? The fluxes produced by the coils
would be in the opposite directions. The stronger coil wins and it ends up pushing
a flux through the other. Exactly the same way in an electrical circuit with multiple
voltage sources where one voltage source would force a current through the other.
In this simple case of a rectangular core as shown in Fig. 3.9, the stronger coil would
be the one with the larger MMF—the product of the current and the number of turns.
It is very important to stress that in a more complex core with multiple limbs, the
stronger coil is not just determined by the MMF of the coil. Eventually, we must
solve the magnetic circuit as a whole taking into account the MMF of the coils and
the reluctance of every limb of the core.
To illustrate how we can determine the flux in the core by solving the magnetic
circuit, let us use Fig. 3.10. We could determine the flux in every limb of the core by
considering only one coil at a time and calculating the equivalent reluctance. This

Fig. 3.10 Magnetic circuit

solution for two coils wound
on a three-limb core i1 i2
3.4 Mutual Inductance 75

will result in two fluxes as shown by solid arrows (for the coil wound on the left
limb) and dashed arrows (for the coil wound on the right limb). The flux in the core
due to both coils will be the sum of the two fluxes by superposition theorem. This
will result in the fluxes expressed as functions of the current flowing in the coils
as follows. In the following equations, ρ1 , ρ2 and ρ3 are the reluctances of the left
limb, right limb and central limb, respectively.
The flux flowing in the left limb is expressed as:
ρ1 ρ3
N1 i1 N2 i2 ρ1 +ρ3
φ1 = ρ2 ρ3 + (3.18)
ρ1 + ρ2 +ρ3 ρ2 + ρρ11+ρ
ρ3
3
ρ1

The flux flowing the in right limb is expressed as:

ρ2 ρ3
N1 i1 ρ2 +ρ3 N2 i2
φ2 = + (3.19)
ρ1 + ρρ22+ρ
ρ3
3
ρ2 ρ2 + ρρ11+ρ
ρ3
3

Further simplifications are possible with the above two equations though expressing
them in the above form makes it very simple to understand how the fluxes can be
calculated as the effect of both coils.
Subsequently, the flux linkages ψ1 and ψ2 can be determined as follows:
ρ1 ρ3
N12 i1 N1 N2 i2 ρ1 +ρ3
ψ1 = ρ2 ρ3 + (3.20)
ρ1 + ρ2 +ρ3 ρ2 + ρρ11+ρ
ρ3
3
ρ1

ρ2 ρ3
N1 N2 i1 ρ2 +ρ3 N22 i2
ψ2 = + (3.21)
ρ1 + ρρ22+ρ
ρ3
3
ρ2 ρ2 + ρρ11+ρ
ρ3
3

It is fairly easy to observe that (3.20) is of the same structure as (3.17). We have
expressed the flux linkage ψ1 as a sum of a function of the current i1 in coil 1 and a
function of the current i2 in coil 1.
The reader is encouraged to determine that the first term in (3.20) and the second
term in (3.21) are merely the inductances L1 and L2 of the coils if we ignore the
effect of the other coil. This calculation can be found in the previous chapter. We will
now rename these inductances L1 and L2 as the self-inductances of the coil [1, 2].
The self-inductance of a coil expresses the flux linkages of the coil due to a current
flowing through the coil itself ignoring any other coils that may be magnetically
coupled to it. We can now introduce another term—mutual inductance. The mutual
inductance expresses the flux linkage of a coil due to the current flowing in another
coil magnetically coupled to it [1, 2]. This mutual inductance is a quantity that
expresses the relationship between two coils and we normally use the letter M to
denote it.
76 3 Simulating Magnetically Linked Circuits

The mutual inductance M12 will define the flux linkage of coil 1 due to the current
flowing in coil 2:

N1 N2 ρ3
M12 = ρ1 ρ3 (3.22)
ρ2 + ρ1 +ρ3 ρ1 + ρ3

The mutual inductance M21 will define the flux linkage of coil 2 due to the current
flowing in coil 1:

N1 N2 ρ3
M21 = ρ2 ρ3 (3.23)
ρ1 + ρ2 +ρ3 ρ2 + ρ3

It can be observed from the above equations that the mutual inductances will be
equal when ρ1 = ρ2 . It is usually assumed that when two coils are magnetically
coupled, their effect on each other will be identical or that the mutual inductance
will be same. However, that is not the case—in general, when two entities affect
each other, the effect that one has on the other need not be the same.
The flux linkage equations for the two coils can be written as [1, 2]:

ψ1 = L1 i1 + M12 i2 (3.24)

ψ2 = L2 i2 + M21 i1 (3.25)

The term with the mutual inductance has a positive sign due to the sense of winding
of the coils and the directions of currents chosen. If the sense of winding of one of
the coils was changed or the direction of current was reversed, the term with the
mutual inductance would have a negative sign. The reader is encouraged to verify
this.
Theoretically, we could imagine many strange shapes of cores and could imagine
coils wound in various different limbs. In such cases, the mutual inductance Mxy
and Myx between two coupled coils x and y would not be the same. However, the
practical reason for winding coils on the same core such that they are magnetically
linked is usually to either achieve transfer of energy from one coil to the other
(transformers) or to augment/decrease the effective inductance. In such cases,
practical cores have fairly regular shapes, due to which two magnetically coupled
coils x and y will have the same mutual inductance Mxy = Myx . We can substitute
ρ1 = ρ2 in the expressions for the mutual inductances M12 and M21 above.
This would imply that the left limb and the right limb of the core have identical
dimensions—mean length, cross-sectional area and relative permeability—which is
quite reasonable for a practical core.
3.4 Mutual Inductance 77

Besides the convenience of M12 = M21 , assuming a symmetrical core results in

another very interesting relation. The self-inductances L1 and L2 can be written as:

N12
L1 = (3.26)
ρ1 + ρρ11+ρ
ρ3
3

N22
L2 = (3.27)
ρ1 + ρρ11+ρ
ρ3
3

In the above equations, ρ1 = ρ2 has been assumed to reduce the number of

variables.
The product of the self-inductances in the above expressions results in the
following expression:

N12 N22
L1 L2 =  2 (3.28)
ρ1 + ρρ11+ρ
ρ3
3

If we take the product of the mutual-inductances M12 and M21 , we arrive at the
following expression:
 2
N12 N22 ρ3
M12 M21 =  2 (3.29)
ρ1 ρ3 ρ1 + ρ3
ρ1 + ρ1 +ρ3

Comparing the above expressions, we arrive at the following relation [1, 2]:
 2
ρ3
M12 M21 = L1 L2 (3.30)
ρ1 + ρ3

From the above expression, we have now found a way to relate the mutual
inductances between two coils with respect to the self-inductances of the coils. Since
M12 = M21 , we can express the above equation as:
ρ3
M12 = M21 = L1 L2 = k L1 L2 (3.31)
ρ1 + ρ3

In the above equation, we have expressed the mutual inductance between two coils
with respect to the self-inductances of the coils in addition to a factor k. This k
is called the coupling factor between the coils. k is the ratio of the flux produced
by a coil that links with the other coupled coil. In the case of Fig. 3.10, due to the
presence of the central limb, only a fraction of the flux produced by a coil links with
the other. In the case of the simple rectangular core of Fig. 3.9, the coupling factor
k = 1 because all the flux produced by a coil links with the other coil.
78 3 Simulating Magnetically Linked Circuits

In the above discussion of coupling factor k, we have assumed that all the flux
produced by a coil will flow through the iron core. Usually this is not the case and
there will always be a certain amount of flux that leaks into the air and does not
flow through the core. Since the iron core has a permeability that is approximately
a 1000 times greater than the surrounding air, the flux that leaks into the air will be
negligible. It is normal to further multiply the coupling factor k above with another
coupling factor to take into account the flux that leaks into the air. This coupling
factor can be denoted by kair and can be between 0.995 to 0.98, which would result
in a leakage between 0.5% and 2%, respectively.
In this section, we have quantified the effect of the current flowing through a
coil on the flux linkages of another magnetically coupled coil through the property
of mutual inductance. Furthermore, we derived the mutual inductances between
coupled coils with respect to the self-inductances of the coils and the ratio of
the flux produced by one coil that links the other coil. This flux linkage equation
is a generalized equation that can be written for any number of coils that are
magnetically coupled. In the next section as well as the next chapters, we will use
this equation as the basis of the simulation model for coupled coils.

3.5 Approach to Simulating Coupled Coils

In this section, we will use the concept of mutual inductance and self-inductance
to expand on the method used to simulate a single inductor, which was described
in the previous chapter. We had modelled inductors as variable voltage sources in
series with variable resistors of such values that the current drawn by the inductor is
equal to that computed from the mathematical model. While modelling magnetically
coupled coils, each coil will be represented as a variable voltage source in series
with a variable resistance and their instantaneous values will be such so as to draw
currents as computed from the mathematical model. In this section, we will describe
how we will use this mathematical model in a simulation.
Figure 3.11 shows how we can expand the voltage in series with a resistance
model used for simulating a single inductor to the coupled coils of Fig. 3.9. The
two coils will have independent electrical circuits that are connected to their
separate terminals. The voltage across each coil terminal is measured and fed to
the mathematical model that describes the magnetic coupling between the two
coils. The output of the mathematical model will be the value of the voltages to
be produced by the controllable voltage sources and optionally the value of the
resistances in series with the voltage sources. As can be seen from Fig. 3.11, there
is a single mathematical model for both coils and this mathematical model will be
based on the equations derived in the previous section.
3.5 Approach to Simulating Coupled Coils 79

Fig. 3.11 Simulation model

of coupled coils

V V

g(V )

Let the voltages measured across the terminals of the left and the right coils be
v1 and v2 , respectively. We can independently apply Kirchhoff’s Voltage Law to the
two coils as follows:

e1 = v1 − i1 r1 (3.32)

e2 = v2 − i2 r2 (3.33)

Here e1 and e2 are emfs induced in the two coils, while r1 and r2 are the winding
resistances of the two coils. The directions of currents i1 and i2 have been assumed
to be the same as in Fig. 3.9—entering the upper terminals of each coil.
The emfs induced in the two coils can be expressed by Faraday’s Law as:

dψ1
e1 = (3.34)
dt
dψ2
e2 = (3.35)
dt
Subsequently, the flux linkages can be expressed as integrals of the induced emfs:

ψ1 = e1 dt (3.36)


ψ2 = e2 dt (3.37)

The flux linkages of a coil can be expressed as the cumulative effect of the current
through the coil and the current through the magnetically linked coil using self and
mutual inductances as described in the previous section:

ψ1 = L1 i1 ± M12 i2 (3.38)

ψ2 = L2 i2 ± M21 i1 (3.39)
80 3 Simulating Magnetically Linked Circuits

In the above equations, we have maintained all the generalizations possible. The
± sign with the mutual inductance term takes into account the fact that the coils
can be wound in such a sense that the fluxes produced by the coils could be in the
same direction or in the opposite directions. Moreover, the mutual inductances M12
and M21 between the coils could be different depending on core construction even
though they are the same in the case of the rectangular core of Fig. 3.9.
In Eqs. (3.38) and (3.39), the flux linkages ψ1 and ψ2 are computed from the
applied voltages and the currents already flowing through the coils. This concept has
been described in the previous chapter where the simulation model for the inductor
was developed. Therefore, (3.38) and (3.39) are simultaneous equations with the
variables being i1 and i2 and upon solving them, we would obtain the updated
values of the currents i1 and i2 flowing through the two coils. Solving simultaneous
equations is fairly simple and for this specific case of two coupled coils, we could
solve them by mere observation as:

ψ1 ∓ M12 i2
i1 = (3.40)
L1
M21
ψ2 ∓ L1 ψ1
i2 = M12 M21
(3.41)
L2 ∓ L1

From the above equations, it is fairly obvious that a few simplifications will make
the solution much cleaner. To begin with, there could be any number of coils that
are magnetically coupled in which case such a solution will be very inconvenient. In
a later section, we will generalize the solution by expressing the equations in matrix
form and solving them using matrix manipulations. The second messy aspect of the
above equations is the fact that we have included the ± term to take into account
the sense of winding of the two coils. Though it is essential to include the sense
of winding of the coils in the model, including it in the above equations results in
equations that can be different for every set of coupled coils. If we can write the
flux linkage equations with only positive signs for every mutual inductance term,
we would result in a set of equations that would be identical for any magnetically
coupled system. In the next section, we will introduce the concept of the dot polarity,
which introduces a visual method of depicting the sense of winding of coils while
leaving the equations to be the identical in every case.

3.6 Dot Polarity of Coupled Coils

In the previous section, we had developed a simulation model for the coupled
inductors seen in Fig. 3.9. The equations for the flux linkages of each coil take
into account the mutual inductances between the coils with the sense of winding
of the coils appearing in the equations as positive terms in case the fluxes flow in
3.6 Dot Polarity of Coupled Coils 81

Fig. 3.12 Dot polarity to i1 i2

indicate the sense of winding
of coupled coils with
assisting fluxes
L1
v1 L2 v2

the same direction and as negative terms in case the fluxes flow in the opposite
direction. This is a bit inconvenient as there is a disconnect between the circuit and
the equations. When we represent coupled coils, we would typically not show the
coils wound on a core as this will make circuit diagrams complicated. We would
like to show the coupled coils as mere inductors and denote their sense of winding
in some manner visually on a circuit. In this section, we will introduce the concept
of the dot polarity, which is a convenient manner of marking the relative sense of
windings of two coupled coils [1, 2].
Figure 3.12 shows how the coupled coils of Fig. 3.9 can be represented as mere
inductors without any details of the core or how the coils are wound on the core. As
can be noticed, there is a “dot” on each inductor (coil)—on the left terminal for the
left inductor and on the top terminal for the right inductor [1, 2]. In more complex
circuits with several coupled coils wound on separate cores, we could use a number
of different shapes such as triangles, rectangles, ellipses etc. to indicate that coils
are coupled together [1, 2]. Moreover, several inductors can be coupled since there
could be several coils wound on the same core, in which case, they will all have a
similar shape (such as a dot) on one of their terminals.
By using dot polarities, we have eliminated the need to draw the core and the
coil wound on the core as we did in the previous sections. This makes circuit
diagrams much simpler. The dot polarity indicates that the two inductors shown are
magnetically coupled. Conversely, the presence of shapes such as a dot on inductors
anywhere in a circuit indicates that they are magnetically coupled. In Fig. 3.12, the
inductors have been deliberately drawn such that one is horizontal while the other
is vertical. In complex circuits, inductors can be far apart in the circuit diagram, but
a dot polarity indicates that they are still magnetically coupled. Now that we have
described how dot polarities are used, let us examine what they imply for the nature
of magnetic coupling between them and how the sense of winding can be inferred.
In simple terms, when a dot polarity exists on two inductors as shown in Fig. 3.12,
it implies that currents flowing through the inductors while entering at the terminals
marked by dots will produce fluxes in the core flowing in the same direction [1, 2].
Moreover, the flux produced in one coil due to the current flowing in the other will
be in the same direction as the flux produced due to the current flowing in the coil
itself as long as both currents are entering the terminals marked by dots. With this
82 3 Simulating Magnetically Linked Circuits

Fig. 3.13 Coils wound on a

core that produce opposing
fluxes in the core i1 i2

definition, the reader can verify that the inductors with the dot polarities of Fig. 3.12
precisely depict the coupled coils wound on a rectangular core shown in Fig. 3.9.
Once we have established this basic definition, a number of other effects can also
be deduced in the manner similar to that performed in the previous section. If an
increasing current flows through one of the inductors while entering at a terminal
marked by a dot, the emfs induced in all the inductors will be such that the positive
polarity of the emfs will appear at all the terminals marked by dots. Conversely, if
a decreasing current flows through one of the inductors while entering at a terminal
marked by a dot, the emfs induced in all the inductors will be such that the negative
polarity of the emfs will appear at all the terminals marked by dots. One can also
examine cases when increasing or decreasing currents flow through the inductors
such that they leave a terminal marked by a dot. The reader is encouraged to verify
these and use Fig. 3.9 with the coils on the core so that Faraday’s Law, Lenz’s Law
and Right Hand Thumb Rule can be applied in each case.
To complete the explanation on how dot polarities can be used to denote the
sense of winding of inductors that are magnetically coupled, let us consider the case
of Fig. 3.13 in which the coil on the right limb is wound in a sense that is opposite
to that of Fig. 3.9. By using the Right Hand Thumb Rule, for the current i2 flowing
through the right limb coil as shown, the flux produced will be in a direction that is
opposite to the flux produced by the current flowing in the left limb coil. Such a set of
coils can also be conveniently represented using mere inductors with dot polarities.
We need to only ask the question—which terminal of the inductors (coils) would
the currents have to enter while flowing through the inductors, that the fluxes in the
core due to the currents would be in the same direction? From Fig. 3.13, the current
would have to enter the right limb coil at the lower terminal in order to produce
a flux in the same direction as the current entering the left limb coil at the upper
terminal. Therefore, as shown in Fig. 3.14, the dot polarities are placed such that a
dot appears in the left terminal of the left inductor and at the lower terminal of the
right inductor. It is important to understand that these dots are merely a convention,
and we can place them on the other terminals of either inductor as well, i.e. on the
right terminal of the left inductor and at the upper terminal of the right inductor.
By using dot polarities to depict the sense of winding of magnetically coupled
coils, we can now remove the sense of coupling from the flux linkage equations and
3.6 Dot Polarity of Coupled Coils 83

i1 i2

L1
v1 L2 v2

Fig. 3.14 Dot polarity to indicate the sense of winding of coupled coils with opposing fluxes

i1 i2 i1

L1 L1
v1 L2 v2 v1 L2 v2

i2

Fig. 3.15 Comparison of dot polarity representation of coupled coils with assisting and opposing
fluxes

instead visually depict how coupled coils are connected in their respective circuits.
We must, however, make an assumption while using dot polarities in coupled coils.
We assume that the current through the coil enters the terminal marked by a dot. We
also assume that the voltage measured across a coil is such that the terminal with the
dot has the positive polarity. We depict these assumptions for both cases as shown
in Fig. 3.15. For the set of inductors on the right, the direction of current i2 and
measured voltage v2 have been reversed. Since the currents i1 and i2 are entering
the terminals marked by dots, the flux linkage equations will be the same in both
cases and can be expressed as:

ψ1 = L1 i1 + M12 i2 (3.42)
ψ2 = L2 i2 + M21 i1 (3.43)

In this section, we have described a convenient manner to depict magnetically

coupled coils as inductors while visually depicting how they produce fluxes in
the core. This method can also be used in developing simulation models so as
to simulate any number of coils that are magnetically coupled. Furthermore, by
choosing convenient directions for current flowing through a coil and the voltage
measured across the coil terminals, we can write the flux linkage equations without
having to think about the nature of coupling between the coils. In the next section as
well as in the next chapters, these assumptions help in being able to model several
coils that are magnetically coupled in the simplest possible manner.
84 3 Simulating Magnetically Linked Circuits

3.7 Simulating Magnetically Coupled Coils

In the past few sections, we introduced several concepts that we can now use to
simulate magnetically coupled coils. We had examined how the coupling between
coils can be quantified by calculating the mutual inductance between the coupled
coils. We had also described how dot polarities can be used to visually depict
the effect that a current flowing in a coil has on the flux flowing through another
magnetically coupled coil. In this section, we will use all these concepts to simulate
magnetically coupled coils wound on an iron core. The basis for this simulation will
be very similar to the simulations of inductors performed in the previous chapter.
Figure 3.16 shows the topology of the circuit that will be chosen for
simulation. This simulation can be found in the folder basic_concept within
chapter3_magnetic_coupling in the following link in the simulation
repository:
https://github.com/opensourceelectrical/simulating-magnetics-for-power-
electronics.
We have used the dot polarity convention to depict the nature of magnetic coupling
between the coils. Each coil is connected to its own circuit that may comprise
of voltage sources vs1 and vs2 in series with feeder impedances Zs1 and Zs2 .
Furthermore, the circuits may also be feeding loads such as ZL1 and ZL2 . The
voltage sources vs1 and vs2 can be dc voltages sources, ac voltage sources or even
power converter output voltages. At this point, we will begin our simulations with
the simple case of the voltage sources being dc voltages. Let us also make the
assumptions that the feeder impedances Zs1 and Zs2 are mere feeder resistances
Rs1 and Rs2 and the loads ZL1 and ZL2 are mere load resistors RL1 and RL2 . Let us
choose Rs1 = Rs2 = 0.1 and RL1 = RL2 = 10. Let us begin with the voltage
source vs1 = 24 Volts and vs2 = 0, implying only the circuit on the left is energized.
We must now define the coils and the coupling between them. We could define
the coils in terms of the number of turns in which case we would also need
the dimensions of the iron core in order to define the magnetic circuit. It is
always possible to start every simulation of any magnetic component such as an
inductor, coupled inductors or even machines such as transformers, motors and

i1 i2
Zs1 Zs2

L1 vs2
vs1 v1 v2
L2
ZL1 ZL2

Fig. 3.16 Circuit topology with coupled coils

3.7 Simulating Magnetically Coupled Coils 85

generators, with every detail of their construction. However, unless the objective of
the simulation is to determine the detailed state of the component or machine such
as the flux in different parts of the core, using the construction of the component or
machine in a simulation is inconvenient. Therefore, for simplicity, while simulating
coupled inductors and in later chapters transformers, we will represent them as an
equivalent circuit with inductances and resistances.
In the case of the coupled coils shown in Fig. 3.16, we only need their self-
inductances L1 and L2 , the mutual inductance M12 = M21 = M between them
and their winding resistances r1 and r2 . As already discussed in the previous
section pertaining to mutual inductance, for a simple rectangular core, the mutual
inductance of one coil with respect to the other will be the same as that of the
other. In most practical cases, core shapes are taken to be as simple as possible and
therefore, if two coils need to be magnetically coupled, there is no need to choose a
core of a very complex shape. Such an assumption also simplifies the √ computation
of the mutual inductance with respect to the self-inductances as M = k L1 L2 with
k being the coupling factor. In the case of the simple rectangular core, k accounts
for the leakage flux and therefore can be chosen to be 0.99 as this implies 1% of the
flux generated by a coil does not link the other coil but leaks instead.
With the above definitions, we can get started with the control code for the mutual
inductor magnetic model with the following parameter definitions:
import math
d t = 1 . 0 e−6
L1 = 0 . 3
L2 = 0 . 3
r1 = 0.1
r2 = 0.1
coupling_factor = 0.99
M = c o u p l i n g _ f a c t o r ∗ math . s q r t ( L1 ∗ L2 )
res_output1 = 100.0
res_output2 = 100.0
i f t _ c l o c k >= t 1 :
# Model
t 1 += d t

As with the case of simulation of inductors, since we are going to solve a

mathematical model, which in turn involves numerical integration, we need to run
the code at a time interval. In the above code, t1 is the time event variable that needs
to be defined in the configuration of the control file. In the above code, the other two
variables res_output1 and res_output2 are the variable resistors in series
with the controllable voltage sources as shown in Fig. 3.11. Since the values of these
variables need to be linked to the resistance of the VariableResistor elements in the
circuit, we need to define these variables as Output Variables in the configuration of
the control file. Since this is only a basic simulation proving the concept of mutual
inductances, we can choose an integration time step of 1 µs and we can also choose
the interval at which the above code runs to be the same by setting dt to be 1 µs.
86 3 Simulating Magnetically Linked Circuits

We can now start adding code to the model, which will be solved iteratively. We
need to compute the induced emfs of the two coupled coils and subsequently the
flux linkages of the coils:
i f t _ c l o c k >= t 1 :
e1 = v1 − i n d _ c u r r 1 ∗ r 1
e2 = v2 − i n d _ c u r r 2 ∗ r 2
p s i 1 += e1 ∗ d t
p s i 2 += e2 ∗ d t
...
t 1 += d t

In the above statements, we are using the voltages v1 and v2 measured across
the inductor terminals as shown in Figs. 3.11 and 3.16. These variables v1 and v2
need to be defined as Input Variables in the configuration file and their input source
needs to be defined as the voltages measured by the Voltmeters shown in Fig. 3.11.
The currents ind_curr1 and ind_curr2 need to be defined as static variables
in the control configuration as they will be computed by the model. psi1 and psi2
are flux linkages computed by integrating the induced emfs and therefore, will also
need to be defined as static variables in the control configuration. The induced emfs
e1 and e2 are temporary variables in this model and they can remain as regular local
Python variables that will be created and destroyed in every iteration. However, if
we wish to plot these variables or perform any other computations, it is advisable
to define e1 and e2 also as static variables in the control configuration. To ensure
that the Python compiler does not throw any unexpected errors due to referencing
a variable that does not exist, every variable must be defined in every iteration by
either defining them at the head of the file in case of parameters or using the special
variables in the control configuration.
Now that the flux linkages psi1 and psi2 have been computed, we can
calculate the currents using (3.40) and (3.41):
i f t _ c l o c k >= t 1 :
...
i n d _ c u r r 2 = ( p s i 2 − (M ∗ p s i 1 / L1 ) ) / ( L2 − (M ∗ M / L1 ) )
i n d _ c u r r 1 = ( p s i 1 − (M ∗ i n d _ c u r r 2 ) ) / L1
...
t 1 += d t

We are merely solving a simultaneous set of equations with the variables being
ind_curr1 and ind_curr2. Later in the section, we will examine a more
elegant solution using matrix equations.
Now that we have computed the currents that must flow through the two coils, we
need only to adjust the output voltages of the controllable voltages in Fig. 3.11. We
can define two Output Variables vout1 and vout2 in the control configuration and
specify the target of the variables to be the controllable voltage sources in Fig. 3.11.
3.7 Simulating Magnetically Coupled Coils 87

Fig. 3.17 Coil terminal voltages in the case of coupled coils producing assisting fluxes

We can compute the output voltages using Kirchhoff’s Voltage Law applied either
to Fig. 3.11 or the sub-circuit of Fig. 3.16 as follows:
i f t _ c l o c k >= t 1 :
...
v o u t 1 = v1 − i n d _ c u r r 1 ∗ r e s _ o u t p u t 1
v o u t 2 = v2 − i n d _ c u r r 2 ∗ r e s _ o u t p u t 2
...
t 1 += d t

Figures 3.17, 3.18, 3.19, and 3.20 show the simulation results. Figure 3.17 shows
the terminal voltages v1 and v2 across the coils and can be seen to decay from a value
of approximately 12V to zero at steady state. To understand this behaviour of v1 and
v2 , let us examine the induced emfs in Fig. 3.19, which shows that the induced emfs
are also decaying from a value of around 12 V to zero at steady state. From Figs. 3.17
and 3.19, it can be observed that there is a very short-lived transient when the circuit
is energized. Figure 3.20 shows this transient in the induced emfs. The induced emf
e1 at time t = 0 is equal to the dc source of 24 V. This is understandable, since
the current through the inductor cannot abruptly rise but has to increase gradually.
However, e1 drops drastically soon after and the induced emf e2 in coil 2 increases
equally drastically to attain a value close to e1 . This behaviour may seem very
confusing but can be understood if we follow the basic laws as duone before.
Due to the mutual coupling between the two coils, their induced emf will be
proportional to the number of turns. In this case, we have considered two coils with
the same self-inductance, which implies that they have the same number of turns
assuming the parts of the core on which they are wound have the same dimensions.
88 3 Simulating Magnetically Linked Circuits

Fig. 3.18 Coil currents in the case of coupled coils producing assisting fluxes

Fig. 3.19 Coil-induced emfs in the case of coupled coils producing assisting fluxes

Therefore, once e1 is produced in coil 1, a proportional e2 develops in coil 2 as well.

Once e2 increases in coil 2, this results in a terminal voltage v2 developed across coil
2 as shown in Fig. 3.17. Therefore, the current i2 flowing through coil 2 will increase
depending on the value of the load resistor RL2 . This can be seen from Fig. 3.18. The
3.7 Simulating Magnetically Coupled Coils 89

Fig. 3.20 Coil-induced emfs during circuit energization

reader is encouraged to revise the previous section, which described the different
possibilities with coupled coils in which we saw how the current flowing in one coil
is transferred to the coupled coil. Therefore, the current i2 flowing in coil 2 will be
translated to coil 1.
With a current i1 flowing through coil 1, the terminal voltage v1 across coil 1
will fall due to the load resistor RL1 . This results in the fall of the induced emf e1 as
shown in Fig. 3.20. After this initial transient, the two induced emfs follow the same
path. As the current i1 through coil 1 increases and reaches a steady state, the flux
in the core will also reach a steady state and therefore, e1 and e2 decay to zero. With
e2 decaying to zero, the current i2 flowing through coil 2 will also decay to zero.
It can be seen from Fig. 3.17 that the voltage v1 across coil 1 is not zero in steady
state but had a small non-zero value. This is merely due to the drop across the coil
winding resistance with a steady state current i1 flowing through coil 1.
The source used to energize coil 1 in the above simulation has been a dc
source. However, the reader is welcome to repeat the simulation with an ac supply.
Moreover, the reader can also change the parameters of the coupled inductors such
as the self-inductances and the coupling factor. Since the later chapters will be
dedicated to the operation of transformers energized by ac supplies, these simulation
results will not be presented in this chapter. However, we will repeat the simulation
for the case when the dot polarities of the coupled coils are as shown in right
sub-figure of Fig. 3.15. The circuit being simulated can be redrawn as shown in
Fig. 3.21. It is important to note from Fig. 3.21 that the current i2 in coil 2 has been
reversed in direction such that it enters the terminal marked by the dot. Moreover,
the voltage v2 across coil 2 terminals has also been reversed such that the positive
90 3 Simulating Magnetically Linked Circuits

i1
Zs1 Zs2

L1 vs2
vs1 v1 v2
L2
ZL1 ZL2

i2

Fig. 3.21 Circuit topology with coupled coils while reversing the sense of winding

Fig. 3.22 Coil terminal voltages in the case of coupled coils producing opposing fluxes

polarity is at the terminal marked by the dot. These conventions ensure that the flux
linkage equations always have positive signs and therefore the code representing the
mathematical model does not change. To adapt to these dot conventions, the polarity
of the Voltmeter and the controlled voltage source on the right in Fig. 3.11 will be
reversed.
Figures 3.22, 3.23, 3.24, and 3.25 show the simulation results for the circuit of
Fig. 3.21. Figures 3.22 and 3.23 show the coil terminal voltages and the induced
emf in the coils. Strangely, these are exactly the same as the previous simulation.
However, it is important to remember that because coil 2 does not have a voltage
source, the terminal voltage will be the same as the induced emf. From the definition
of dot polarity, an increasing current entering one of the coils at a terminal marked
by a dot will produce an induced emf in the other coil such that the polarity of the
3.7 Simulating Magnetically Coupled Coils 91

Fig. 3.23 Coil-induced emfs in the case of coupled coils producing opposing fluxes

Fig. 3.24 Coil currents in the case of coupled coils producing opposing fluxes

emf will be positive at the terminal marked by the corresponding dot. Therefore,
since current i1 is initially increasing and entering coil 1 at a terminal marked by
a dot, the induced emfs and the coil voltages will be the same as in the previous
simulation.
92 3 Simulating Magnetically Linked Circuits

Fig. 3.25 Supply voltage and voltage across load RL2

The difference between the simulations will be known when we plot the coil
currents as shown in Fig. 3.24. The current i2 in coil 2 has the opposite polarity as
in the previous simulation. This is due to the fact that the induced emf e2 in coil 2
has a polarity such that it is positive at the terminal marked by the dot. As a result,
the current flowing through the load resistor RL2 will be in the reverse direction. To
get a better understanding of the voltages produced by the coils, Fig. 3.25 shows the
supply voltage of 24 V and the voltage appearing across the load resistor RL2 . The
supply voltage is a constant 24 V as expected, while the voltage across RL2 rises
to approximately −12 V and then decays to zero once the current i1 through coil 1
reaches a constant value at steady state.
In this section, we have used the theory described in the past few sections to
develop simulation models for coupled coils in a very basic circuit. The purpose
of this section was to describe how using Python code, we can simulate coupled
conductors while also including the sense of winding of the coils using dot
polarities. The simulation results support our theoretical discussions and the reader
is encouraged to extend these simulations with variations in the circuit topology. In
the next section, we will discuss a more practical case of when coupled inductors
need to be simulated. The next chapter based on transformers will describe in detail
a practical application of magnetic coupling.
3.8 Ćuk Converter 93

3.8 Ćuk Converter

In the previous section, we simulated magnetically coupled coils that were con-
nected to two independent isolated electric circuits. When we wish to transfer
electrical energy from one circuit to another isolated circuit, we would use a
transformer, which is a special case of magnetically coupled coils designed in such
a manner so as to achieve maximum power transfer with the minimum possible
leakage and losses. In undergraduate courses, as electrical engineers, we often
analyse random circuits that contain coupled coils that are usually marked using
dot polarities. In such circuits, the coupled inductors may not be in isolated sub-
circuits. In this section, we will examine such a case of where coupled inductors
could be used in a power electronic converter [12, 37, 38].
Besides using magnetic coupling to transfer energy from one isolated circuit to
another, another reason for magnetic coupling between inductors is that they are
wound on the same core. In an inductor, the core results in the inductor being
bulky and occupying space in a converter. If a converter is being used as a power
supply, we would like to decrease the size as much as possible to make the power
supply portable and light. In power converters, inductors and capacitors are integral
components as they are energy storage elements that allow us to condition power
while producing a particular output for a given input. If a power converter has
multiple inductors, if each inductor needs a separate core, this would result in the
size of the converter increasing and potentially becoming bulky. On the other hand,
if more than one inductor could be wound on the same core, the size of the converter
could be decreased. Once coils are wound on the same core, there will be a mutual
inductance between the coils. In the circuit equations, the mutual inductance will
appear in addition to the self-inductance of the coils.
Besides the circuit equations needing to be modified, winding more than one
inductor on the same core is practical only under some circumstances. This is due
to the fact that flux in the core is now dependent on the current flowing through
both coils. As a result, the induced emfs in the coils will also be affected by the rate
of change of current flowing through both coils. If a circuit is so designed, that the
current flowing through two or more inductors will have the same pattern—increase
during the same intervals and decrease during the same intervals, these inductors
could be wound on the same core. However, if the nature of current flowing through
inductors is very different and not at all correlated, winding them on the same core
could result in the operation of the circuit being altered and eventually cause the
circuit to malfunction. This concept can be best described using the example of the
Ćuk converter [37, 38].
Figure 3.26 shows the topology of a Ćuk converter named after Slobodan Ćuk
who first presented its design as an improvement over the basic buck-boost converter
[37]. The Ćuk converter can achieve an output voltage vo that is either higher or
lower than the input Vin (buck-boost) though the output is inverted just like with a
conventional buck-boost converter. The main energy storage device in the converter
is the capacitor C1 . This is in contrast to the conventional buck-boost converter that
94 3 Simulating Magnetically Linked Circuits

iL1 L1 C1 iL2
L2
+ −
vC1
+ −
Vin S D Co vo RL
−
+

Fig. 3.26 Ćuk converter

iL1 L1 C1 iL2
L2
+ −
vC1
+ −
Vin S Co vo RL
−
+

Fig. 3.27 Ćuk converter operation when switch S is conducting

iL1 L1 C1 iL2
L2

+ −
vC1
+ −
Vin D Co vo RL
−
+

Fig. 3.28 Ćuk converter operation when diode D is conducting

uses the inductor as the energy storage device, which results in either high ripple
currents or a bulky converter if a large inductor is chosen to decrease the current
ripple. In the Ćuk converter, the inductors L1 and L2 experience much lower ripple
current and additionally their currents iL1 and iL2 are continuous making closed
loop control design simpler.
The working of the Ćuk converter can be depicted by Figs. 3.27 and 3.28. When
the switch S is turned ON, the supply Vin energizes the inductor L1 and the current
iL1 increases. Initially, when all currents and voltages are zero with the circuit being
at rest, the very first action will be the inductor L1 energizing. When the switch S
is turned OFF, the current iL1 will freewheel through the diode D and will therefore
3.8 Ćuk Converter 95

charge the capacitor C1 due to which the voltage vC1 will increase with the polarity
shown. Since the energy in the inductor L1 charges the capacitor C1 , the current iL1
will decrease. When the switch S is turned ON in the next cycle, the current iL1
will increase as before with energy being stored in the inductor. However, now with
the capacitor C1 being charged, this capacitor will charge inductor L2 and current
iL2 will increase. As a result, energy has been transferred from the supply Vin to C1
through the inductor L1 and now the capacitor C1 transfers energy to L2 . Along with
energizing the inductor L2 , when the switch is conducting, the capacitor C1 also
transfers energy to the output capacitor Co and the output voltage vo will increase.
When the switch is turned OFF, inductor L1 transfers energy to the capacitor C1 .
With the inductor L2 also energized, the current iL2 will freewheel and continue to
supply the output capacitor. Since both inductors are freewheeling through the diode
D, the currents iL1 and iL2 will decrease.
From the working of the Ćuk converter, a few inferences can be made. When the
switch is turned ON, the currents iL1 and iL2 increase simultaneously—inductor L1
energized by the supply and inductor L2 energized by the capacitor C1 . When the
switch is turned OFF, the currents iL1 and iL2 decrease simultaneously—inductor
L1 transfers energy to capacitor C1 and inductor L2 transfers energy to the output
capacitor Co . Therefore, the Ćuk converter uses two inductors L1 and L2 whose
currents increase and decrease simultaneously during the same intervals. Moreover,
the function of these inductors is merely to transfer energy to and from the capacitor
C1 . Ripples in the currents iL1 and iL2 will be present due to the transfer of energy.
However, to decrease electromagnetic interference, it is advisable to decrease the
ripple as much as possible. The simplest way to decrease current ripple is to increase
the size of the inductors. However, this would result in a more bulky converter,
which is undesirable. Therefore, a convenient manner of decreasing the ripple while
also limiting the size of the inductors is to wind them on the same core so as to
introduce a mutual inductance between them.
In Fig. 3.26, the inductors L1 and L2 have been marked by dot polarities on their
terminals to indicate that they are magnetically coupled. It is important to note that
polarities are such that currents iL1 and iL2 are entering the terminals marked by
the dot polarities, which implies that the fluxes produced in the core due to the
currents will be in the same direction. The flux linkage equation for L1 and L2 can
be expressed as:

ψ1 = L1 i1 + Mi2 (3.44)

ψ2 = L2 i2 + Mi1 (3.45)

In the above equations, there is now a mutual inductance M term. This term results
in the flux linkages of both inductors being larger than they would have been if the
inductors had been isolated and wound on separate cores. Subsequently, the rate
of change of flux linkages will also be larger resulting in large induced emfs in
the inductors. Since the induced emf in the inductor is the primary opponent of the
96 3 Simulating Magnetically Linked Circuits

change in the current flowing through the inductor, this will result in reduced current
ripple.
The simulation of the Ćuk converter can be found in the folder cuk_converter
within chapter3_magnetic_coupling in the following link in the
simulation repository:
https://github.com/opensourceelectrical/simulating-magnetics-for-power-
electronics
To simulate the Ćuk converter of Fig. 3.26, we need one additional component of
control to generate gate signals to the switch S. Since we need only to demonstrate
the basic operation of the Ćuk converter, we choose a simple constant frequency
operation of the converter in which the switch S is turned ON for a duration of the
switching time period [35]. Therefore, if T is the time period of a switching cycle
and ton is the time interval for which the switch S is turned ON, we define the duty
ratio d as:
ton
d= (3.46)
T
Therefore, d lies between 0 and 1. By varying the duty ratio and therefore varying
the time interval for which the switch S is turned ON, the output voltage vo can be
varied. To regulate the output voltage vo to a desired reference, we vary the duty
ratio d.
In this simulation, we will not implement output voltage regulation as closed
loop control is a fairly vast topic. Instead, we will assume a value of d between
0 and 1 (say 0.3). To convert this value of duty ratio to a gate signal, we use
the Pulse Width Modulation (PWM) technique where we compare the duty ratio
with a fixed frequency carrier waveform. For a dc–dc converter, the fixed frequency
carrier waveform is usually a sawtooth waveform [35]. The PWM technique can be
implemented using the following code block:
d t = 1 . 0 e−8
f_sw = 2 0 0 0 0 . 0
T_sw = 1 / f_sw
c a r r _ s l o p e = 1 / T_sw
i f t _ c l o c k >= t 1 :
c a r r _ w a v e += c a r r _ s l o p e ∗ d t
i f carr_wave > 1 . 0 :
carr_wave = 0.0
duty_ratio = 0.3
i f d u t y _ r a t i o > carr_wave :
s1gate = 1.0
else :
s1gate = 0.0
pwm_carr = c a r r _ w a v e
pwm_dutyratio = d u t y _ r a t i o
pwm_gate = s 1 g a t e
t 1 += d t
3.8 Ćuk Converter 97

In the above code block, we have chosen the switching frequency f_sw to be
20 kHz or 20,000 Hz. This is typical for a dc–dc converter such as the Ćuk converter.
As with other control codes, this has a Time Event variable t1, which is updated
at a constant interval at which we would like the control code to be executed. This
is usually the case when we implement control in hardware using platforms such as
microcontrollers. The control code is executed at a regular interval by configuring
a timer. Since our switching frequency f_sw is 20,000 Hz, we choose the time
interval of the control to be 10 nanoseconds. This is to ensure that we are able to
generate accurate switching gate signals within each switching time period.
The carrier waveform is denoted by the variable carr_wave. This carrier
waveform is compared with the fixed duty ratio of 0.3 (the reader is encouraged
to try out other values). When the duty ratio is greater than the carrier waveform,
the switch S is turned ON by providing a gate signal of 1 while when the duty ratio
is lower than the carrier waveform, the switch is turned OFF by providing a gate
signal of 0. This can be depicted in Fig. 3.29. As can be seen, using the technique
of PWM, the duty ratio of 0.3 is translated to an ON time for the switch for 30% of
the time period of a switching cycle.
Figure 3.30 shows the currents iL1 and iL2 through the inductors L1 and L2 . The
currents can be seen to increase when the switch S is turned ON and conducting
and can be seen to decrease when the switch S is turned OFF and the diode D
conducts instead. Importantly, the currents rise and fall simultaneously during the
same interval. Figure 3.31 shows the output voltage. When the switch S is turned
ON, the output voltage appears to decrease which is contrary to what we expect from
our theoretical discussion. However, we have measured the output voltage with such

Fig. 3.29 Gate signals produced by PWM

98 3 Simulating Magnetically Linked Circuits

Fig. 3.30 Inductor currents iL1 and iL2

Fig. 3.31 Output voltage vo

a polarity that it is a negative voltage. Therefore, in reality, when the voltage seems
to decrease, the magnitude is increasing but in the negative sense.
The simulation of the Ćuk converter thereby shows a use case when we can wind
inductors on the same core due to which they will have a mutual inductance. The
cumulative effect of the self-inductance and the mutual inductance decreases the
current ripple. The reader is encouraged to repeat the simulation with either separate
3.9 Scalable Mathematical Models with Magnetic Coupling 99

inductors or by setting the mutual inductance M to be zero in the mathematical

model of the coupled inductors. Therefore, in such a case, winding inductors on the
same core has a double benefit—decreases the size of the converter as there is only
one core and also decreases the current ripple without having to increase the size of
the inductor.
With the simulation of the Ćuk converter, we have presented a practical case
of when magnetic coupling between two inductors is beneficial even though the
intention is not to achieve power transfer from one coil to the other. In a vast number
of cases, intentional magnetic coupling between coils is usually to achieve power
being transferred from one coil to the other, which will be discussed in the next
chapter. Before bringing this chapter to a close, we will address one other issue
in the mathematical model of coupled inductors. The model is specifically for two
coils and the currents have been obtained by solving simultaneous equations of the
flux linkages of the two coils. However, there could be any number of coils wound
on the same core resulting in many inductors being magnetically coupled. In order
to express the currents through all the coils, we would need to solve the flux linkage
equations as matrix equations. This will be discussed in the next section.

3.9 Scalable Mathematical Models with Magnetic Coupling

In the previous sections, we had expressed the mathematical model for two coupled
coils using simultaneous equations. Such a description is very convenient to analyse
and understand the phenomenon of magnetic coupling. However, in the next chapter
where we begin to model and simulate transformers, we can have many coils that
are magnetically coupled since in many cases transformers have multiple windings
with a voltage source connected to one winding supplying loads in all the others. In
such cases, using the method of manually solving simultaneous equations will not be
feasible. The sensible way of solving equations in such cases is by expressing them
as matrix equations and using matrix manipulation techniques to simplify them.
Let us start with the same case as the coupled coils with dot polarities such that
currents are entering the terminals of both coils marked by dots due to which the
fluxes produced by the currents will be in the same direction through the core.

di1 di2
v1 = r1 i1 + e1 = r1 i1 + L1 +M (3.47)
dt dt
di1 di2
v2 = r2 i2 + e2 = r2 i2 + M + L2 (3.48)
dt dt
We have merely expanded the induced emf terms e1 and e2 with the rate of changes
of flux linkages ψ1 and ψ2 , which in turn have been expressed using currents and
inductances.
100 3 Simulating Magnetically Linked Circuits

When we encounter such a system of simultaneous equations, we can begin to

express them using matrices by defining a few vectors. We can define a voltage
vector v and a current vector i as:

v1 i1
v= , i= (3.49)
v2 i2

We can express the simultaneous equations using these vectors and two more
matrices as follows:

v1 r 0 i1 L M d i1
= 1 + 1 (3.50)
v2 0 r2 i2 M L2 dt i2

We have directly written the differential equations for the currents that we wish
to compute without the intermediate computation of coil flux linkages. We can solve
these differential equations using any numerical integration technique to obtain the
currents for any sample k with respect to the voltages in the same sample k and
the currents from the previous sample k − 1. Before progressing to the step of
numerical integration, we will need to simplify the matrix equation. This is due
to the fact that in the current state, both equations are coupled and if we integrate
either equation, we end up integrating both currents. Luckily, by using matrices to
express the equation, we have at our convenience many manipulation techniques
using which the matrices can be transformed to another structure, which makes the
solution of the equations much simpler [39].
Consider this matrix that contains the self and mutual inductances:

L1 M
(3.51)
M L2

This matrix couples the derivatives of the inductor currents. If this matrix can be
converted to either one of these forms:

L1 M  0 M
(3.52)
0 L2 M  L2

we would have achieved a certain level of decoupling as now the derivative of one of
the currents can be directly integrated without any influence of the derivative of the
other current. To elaborate, in the matrix on the left di2
dt can be integrated to calculate
di1
i2 , while for the matrix on the right dt can be integrated to calculate i1 . Once the
derivative of one of the currents can be calculated, the derivative of the other current
can be calculated by substituting the previously computed derivative. Therefore, in
either of the matrix forms, a solution exists for computing all the currents.
The matrix on the left above is called an upper triangular matrix, while the matrix
on the right above is called a lower triangular matrix [39]. The former is a matrix
that has all elements below the diagonal to be zero, while the latter is a matrix in
3.9 Scalable Mathematical Models with Magnetic Coupling 101

which all the elements above the diagonal are zero. Either of these can be chosen
and the technique to produce this transformation will be the same in either case
except that the precise steps will vary. Let us choose to transform the matrix to an
upper triangular form as an example. To begin our solution, let us rewrite our matrix
equation as follows:

L 1 M d i1 r 0 i1 v
=− 1 + 1 (3.53)
M L2 dt i2 0 r2 i2 v2

We have re-arranged our equation into a left and right hand side, with the
objective of simplifying the matrix equation such that the left hand side of
T
di1 di2
,
dt dt can be computed and subsequently integrated to provide the currents
i1 and i2 . From basic mathematics, we can perform a number of operations to the
left and right hand side of the equation, which will not alter the solution as long as
the same operation is performed on both sides of the equation [39]. As an example,
we could multiply one row of all matrices on both sides of the equation by a scalar
constant (say 5) without changing the solution:

5L1 5M d i1 5r1 0 i1 5v1

=− + (3.54)
M L2 dt i2 0 r2 i2 v2

We could also add one row of every matrix to another row of every matrix of
both sides of the equation without changing the solution. For example, if we add the
first row to the second row:

5L1 5M d i1 5r1 0 i1 5v1

=− + (3.55)
M + 5L1 L2 + 5M dt i2 5r1 r2 i2 v2 + 5v1

These operations are no different from the manner in which simultaneous equations
are solved by hand. The reader is encouraged to verify this.
Instead of choosing the random value of 5 to multiply the first row with, if we
choose the value of − LM1 , the matrix equation turns out to be
     
L1 M d i1 r1 0 i1 v1
2 =− + (3.56)
0 L2 − M
L1 dt i2 r1 − Mr 1
L1 r2 i2 v2 − Mv
L1
1

In the above equation, we have retained the first row in its original form without
multiplying it by any constants but have added to the second row terms containing
the first row multiplied by the constant − LM1 . As can be observed, we have
transformed the matrix on the left hand side of the equation to an upper triangular
form. It should be noted that this transformation was not coincidental, rather
deliberate by choosing a multiplying factor that would eliminate the element in the
second row and first column for the matrix on the left hand side.
102 3 Simulating Magnetically Linked Circuits

The beauty of this technique lies in the fact that it can be applied to any matrix
that represents any system and can be formulated as an algorithm to be solved by
a computer. For example, let us suppose there were three coils wound on the same
core with self-inductances L1 , L2 and L3 . Let us suppose the mutual inductance
between coils 1 and 2 was M12 , the mutual inductance between coils 2 and 3 was
M23 and the mutual inductance between coils 1 and 3 was M13 . The new set of
equations can be represented in the matrix form as follows:
⎡ ⎤ ⎡ ⎤ ⎡ ⎤⎡ ⎤ ⎡ ⎤
L1 M12 M13 i1 r1 0 0 i1 v1
⎣M12 L2 M23 ⎦ d ⎣i2 ⎦ = − ⎣ 0 r2 0 ⎦ ⎣i2 ⎦ + ⎣v2 ⎦ (3.57)
dt
M13 M23 L3 i3 0 0 r3 i3 v3

To transform the matrix on the left hand side of the above equation to an upper
triangular form, we need to reduce three elements to zero—element (2, 1) (=M12 ),
element (3, 1) (=M13 ) and element (3, 2) (=M23 ). We can add to the second row of
all matrices on both sides of the equation the first row of the same matrix multiplied
by the product − ML12
1
. This will eliminate element (2, 1) of the left hand side matrix
while altering the elements in the second row of all matrices on both sides of the
equation. Furthermore, we can add to the third row of all matrices on both sides of
the equation the first row of the same matrix multiplied by the product − ML13 1
. This
will eliminate element (3, 1) of the left hand side matrix while altering the elements
in the third row of all matrices on both sides of the equation. We can eliminate the
element (3, 2) of the left hand side matrix by adding the second row multiplied by an
appropriate factor to the third row of all matrices on both sides of the equation. This
factor will have to be determined after the calculation carried out in the previous
step as the second row has already changed.
Though the above process may seem tedious, it is a mechanical process that can
be automated using code as follows:
L = [ [ L1 , M] , [M, L2 ] ]
R = [ [ r1 , 0 . 0 ] , [ 0 . 0 , r 2 ] ]
V = [24.0 , 0.0]
for count1 in range ( len (L ) ) :
i f not L[ count1 ] [ count1 ] :
f o r c o u n t 2 i n r a n g e ( c o u n t 1 +1 , l e n ( L ) ) :
i f L[ count2 ] [ count1 ] :
L[ count1 ] , L[ count2 ] = L[ count2 ] , L[ count1 ]
R [ c o u n t 1 ] , R[ c o u n t 2 ] = R [ c o u n t 2 ] , R [ c o u n t 1 ]
V[ c o u n t 1 ] , V[ c o u n t 2 ] = V[ c o u n t 2 ] , V[ c o u n t 1 ]
break
i f L[ count1 ] [ count1 ] :
f o r c o u n t 2 i n r a n g e ( c o u n t 1 +1 , l e n ( L ) ) :
comm_factor = L[ count2 ] [ count1 ] / L[ count1 ] [ count1 ]
for count3 in range ( len (L[ count1 ] ) ) :
L [ c o u n t 2 ] [ c o u n t 3 ] −= L [ c o u n t 1 ] [ c o u n t 3 ] ∗ c o m m _ f a c t o r
R [ c o u n t 2 ] [ c o u n t 3 ] −= R [ c o u n t 1 ] [ c o u n t 3 ] ∗ c o m m _ f a c t o r
V[ c o u n t 2 ] −= V[ c o u n t 1 ] ∗ c o m m _ f a c t o r
3.9 Scalable Mathematical Models with Magnetic Coupling 103

From the code block, it is clear how all we need to do is define the matrices and
the process of triangularization that follows is a mechanical process. The reader is
encouraged to choose larger matrices for L, R and V and verify that the process
is able to convert the matrix L to an upper triangular form in every case. The only
important factor is that the matrices L and R should be square and have the same
dimensions, while the matrix V should have the same number of rows as L and
R. Subsequent to triangularization, the differential equations can be solved in the
reverse order as follows:

dibydt = [0.0, 0.0]

for count1 in range(len(L)-1, -1, -1):
if L[count1][count1]:
k = [0.0, 0.0, 0.0, 0.0]
for k_count in range(len(k)):
for count2 in range(len(B[count1])):
k[k_count] += B[count1][count2]*coil_voltages[count2]
for count2 in range(len(A[count1])):
if (k_count==0):
k[k_count] -= A[count1][count2]*ind_current[count2]
elif count2>count1:
k[k_count] -= A[count1][count2]*(ind_current[count2] + \
dt*dibydt[count2])
if (k_count==1):
if (count2 == count1):
k[k_count] -= A[count1][count2]*(ind_current[count2] + \
dt*k[0]/2.0)
elif count2>count1:
k[k_count] -= A[count1][count2]*(ind_current[count2] + \
dt*dibydt[count2])
else:
k[k_count] -= A[count1][count2]*ind_current[count2]
if (k_count==2):
if (count2 == count1):
k[k_count] -= A[count1][count2]*(ind_current[count2] + \
dt*k[1]/2.0)
elif count2>count1:
k[k_count] -= A[count1][count2]*(ind_current[count2] + \
dt*dibydt[count2])
else:
k[k_count] -= A[count1][count2]*ind_current[count2]
if (k_count==3):
if (count2 == count1):
k[k_count] -= A[count1][count2]*(ind_current[count2] + \
dt*k[2]/2.0)
elif count2>count1:
k[k_count] -= A[count1][count2]*(ind_current[count2] + \
dt*dibydt[count2])
else:
k[k_count] -= A[count1][count2]*ind_current[count2]
104 3 Simulating Magnetically Linked Circuits

for count2 in range(len(L[count1])):

if not count2==count1:
k[k_count] -= L[count1][count2]*dibydt[count2]
k[k_count] = k[k_count] / L[count1][count1]
dibydt[count1] = (k[0] + 2*k[1] + 2*k[2] + k[3])/6.0
ind_current[count1] += dibydt[count1]*dt
else:
ind_current[count1] = V[count1]
for count2 in range(count1+1, len(L)):
ind_current[count1] −= L[count1][count2] * dibydt[count2]
for count2 in range(count1+1, len(R)):
if not count2==count1:
ind_current[count1] −= R[count1][count2] * ind_current[count2]
ind_current[count1] = ind_current[count1] / R[count1][count1]

In the above code segment, the numerical integration has been performed using
Runge–Kutta Fourth Order method, which provides greater accuracy as compared to
the simple integration by Backward Euler’s method used in the previous simulations.
The reader is encouraged to include these two code segments into the simulation of
the coupled inductors presented previously in this chapter as well as the simulation
of the Ćuk converter. In the next chapter, system equations will be expressed in
the matrix form and integrated using Runge–Kutta Fourth Order method as shown
above.

3.10 Conclusions

In this chapter, we used our knowledge of magnetic circuits gained from the previous
chapter to develop simulation models for magnetically coupled circuits. We began
with our assertion that all the basic laws of physics applicable to an inductor (single
coil wound on a core) are applicable for the case of multiple coils wound on the
same core. Such systems can also be represented by magnetic circuits comprised of
the MMFs produced by each coil and the reluctances of each limb of the core. The
emfs induced in the coils can be expressed with respect to the rate of change of flux
linkages of the coils. The chapter began with a simple explanation of the concept
of magnetic coupling between two coils wound on a rectangular core with only one
coil being energized.
As the concept of magnetic coupling forms the basis of transformer operation,
which will be dealt with in the next chapter, we examined in detail various
possibilities that can arise when multiple coils are wound on a single core. We have
expressed equivalent magnetic circuits and have performed a mathematical analysis
for different conditions. Initially, we have used only the basic laws of physics such
as Ampere’s Law, Faraday’s Law and Lenz’s Law to determine the behaviour of the
magnetically coupled coils. The scenarios helped us to understand how energy is
transferred from one coil to another magnetically coupled coil. The objective behind
3.10 Conclusions 105

these discussions was to understand how the basic laws of physics can continue to
be applied even though the behaviour of a magnetically coupled system differs from
a single inductor.
Following our discussions on understanding the concept of magnetic coupling,
we proceeded to formulate a mathematical model for a magnetically coupled
system. We introduced the concept of mutual inductance in addition to self-
inductance to express the flux linkages of a coil with respect to the currents flowing
in all coils of the magnetically coupled system. We also described how dot polarities
can be used to indicate the sense of windings of coils such that magnetically
coupled coils can be represented as simple inductors with dots on one of their
terminals instead of depicting the detailed windings on the core. Combining mutual
inductances and dot polarities, concise flux linkage equations could then be written
for any magnetically coupled system.
The chapter consolidates the theoretical discussion with a simulation of two coils
wound on the same core. The simulation has two separate cases where the dot
polarity of the coils are such that in one case, the fluxes produced by the coils are in
the same direction while in the other case, the fluxes produced by the coils are in the
opposite direction. To introduce a practical aspect to the discussion on magnetic
coupling, we described the operation of a Ćuk converter in which magnetically
coupled inductors are often used. Using the Ćuk converter, we described when it
is beneficial to wind multiple inductors on the same core. Simulations of the Ćuk
converter showed how similar patterns of currents through the inductors make it
possible to wind them on the same core, thereby decreasing the size of the converter
and also decreasing the current ripple.
To generalize the simulation of a magnetically coupled system, the simulation
model is converted to a matrix form such that any number of coils wound on a
core can be modelled. We described the advantage of a matrix representation with
the concise matrix equation that was the result. To simplify the matrix equation,
we described a process whereby the inductance matrix could be converted to an
upper triangular form and presented the code block for doing so. Subsequent to the
triangularization, the process of numerical integration was modified for the matrix
representation. Such a matrix representation results in a convenient and expandable
simulation model as will be seen in the next chapter when we simulate transformers.
In the next chapter, we will model and simulate transformers using the funda-
mentals of magnetic coupling described in this chapter. Transformers are a specific
application of coupled coils wound on a core so as to achieve the maximum possible
transfer of energy from one coil to the other. However, this chapter will help
the reader to understand the fundamental nature of magnetic coupling that makes
this possible. Moreover, this nature of magnetic coupling can also be extended to
rotating machines such as motors and generators, though this book will not cover
them.
Chapter 4
Modelling and Simulating Transformers

4.1 Introduction

In the previous chapters, we gradually built up the discussion on energy transfer

through magnetic coupling. The objective of the previous chapters was to build
a bridge between magnetism as dealt in physics and simulation models of com-
ponents such as inductors that use magnetism. The topic of magnetism is vast
and modelling the magnetic field accurately is very complex that requires special
techniques and solutions. However, by assuming uniform magnetic fields and linear
magnetic properties, we were able to formulate concise mathematical models that
captured the phenomenon of magnetic coupling to a degree acceptable for most
electrical engineering applications. Since magnetism as a concept is a bit tricky and
translating physical laws related to magnetism into electrical engineering can result
in confusion, the past two chapters introduced the modelling of magnetic circuits
with the objective of understanding the behaviour of any arbitrary magnetic circuit.
In this chapter, we finally arrive at the main course of this book—an elec-
trical machine based on the transfer of energy through magnetic coupling. The
transformer is a static machine very commonly used in electrical engineering
for interfacing different electrical systems together [10, 11, 13, 14]. There are
numerous applications of transformers, and it is impossible to cover all the different
applications in a single book [13, 14]. The objective of this chapter is to introduce
some of the most common applications of transformers that an electrical engineer
might encounter and present simulation models for these applications. As before,
even though the basic underlying concept of transformer simulations remains the
same for most applications, for an in-depth understanding of transformer operation,
this chapter examines different applications over separate sections.
Under the hood, a transformer is just a set of magnetically coupled coils.
Therefore, before reading this chapter, the reader is advised to be familiar with
all the nuances of magnetically coupled coils described in the previous chapter.
What makes a transformer a special case of magnetically coupled coils is that a

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 107
S. V. Iyer, Modeling and Python Simulation of Magnetics for Power Electronics
Applications, https://doi.org/10.1007/978-3-030-96768-0_4
108 4 Modelling and Simulating Transformers

transformer in most cases is designed to achieve maximum transfer of power from

one coil to other coils. Therefore, even though the underlying simulation model is
similar to simple magnetically coupled coils, arriving at the base model from the
specifications of a transformer as a machine will be described by representing the
equivalent circuit of a transformer. The equivalent circuit of a transformer is dealt
with during undergraduate studies [13, 14]. However, this chapter will relate the
transformer equivalent circuit to the simulation model. Moreover, the equivalent
circuit will be built up gradually as we add details to the transformer model.
During my undergraduate years, I faced a lot of confusion related to the detailed
operation of a transformer. Particularly, understanding how the turns ratio of a
transformer relates to the manner in which the winding currents of the transformer
are expressed was usually merely assumed as a simple formula. On the whole, to use
a set of formulae to figure out the operation of an equipment or machine without a
physical understanding of its working was something that I found to be a hindrance
when I became a graduate student. This is an obstacle most engineers face. The
depth to which they learn transformers and machines in general was inadequate for
application either during their professional careers or when pursuing research.
As I found the need to relearn electrical machines from a different perspective
later, I also found the motivation to learn not just through equations but also through
simulations. Simulations enable us to recreate the physical phenomena behind the
operation of a machine that enable us to understand how the different physical laws
are related [6]. Even though the basic laws of physics are known to us from high
school courses, to be able to put them together such that the operation of a machine
can be understood can be a little tricky and will need some back and forth between
these laws. This approach has been taken in Chap. 2 while introducing magnetic
circuits in the context of inductors and in Chap. 3 while extending the magnetic
circuits to understand coupled magnetic circuits.
In this chapter, following the extension of the simulation model of coupled
inductors to transformers, several sections are dedicated to describing the operation
of a transformer through simulation results [13, 14]. The concept of current
transformation and MMF equalization are explained in detail using both theory
and simulations. The chapter will present simulations of step-up and step-down
transformers and will describe the impact of turns ratio on the winding voltages
and currents. Finally, the chapter will also describe how multi-winding transformers
(with more than two windings) can be modelled and simulated. This will lay the
foundation for the next chapter, which will discuss how three-phase transformers
can be modelled and simulated.
Just like the previous chapters, this chapter will try to combine theory and
simulations by referring to the basic laws of physics while describing simulation
results. The reader might find some of the discussions repetitive. However, the
purpose of these long discussions is to provide a framework where the reader can
try to form connections between the different laws of physics to interpret simulation
results. Such an approach will be not only useful in understanding the operation
of transformers, but can later also be extended to other electrical machines such as
motors and generators. The reader is encouraged to try out the simulations described
4.2 Background 109

in the chapter and also to attempt some of the modifications that have been suggested
as exercises.

4.2 Background

Transformers are one of the most widely used electrical machines, and without
them the modern electrical grid would not be possible [21]. The different uses
of transformers and the solutions offered by them are unfortunately not fully
appreciated by most electrical engineers [13]. For a power engineer, an in-depth
understanding of transformer operation is extremely important. Quite unfortunately,
very few power engineers bother to fully understand how a transformer operates and
stick to the most basic equations that are used to mathematically express transformer
operation. As already described in the previous chapters, all that is needed to
fully understand a static magnetic component or machine are a few basic laws of
physics—Ampere’s Law, Faraday’s Law, and Lenz’s Law. In this chapter, we will
extend our knowledge of magnetic coupling described in the previous chapter to
examine in detail the working of transformers.
Transformers have numerous uses and applications. Before we dive into the
details of the operation of a transformer, let us look at some of the major uses
of transformers [13]. As already stated, the modern electrical grid would not be
possible without transformers. To elaborate on this, we need to appreciate that
modern electrical grids in most parts of the world are gigantic interconnected
systems similar to the network of roads and railways that connect different parts of
a land mass. Electrical energy is produced at generating stations that could be based
on thermal (coal, gas or nuclear), hydro-electric and nowadays even renewable
energy sources (solar, wind, etc.). Except for renewable energy generation that is
possible at the load centre itself in the form of rooftop solar for example, other
forms of energy generation are usually located in remote areas. The remote location
of generating stations is for a number of reasons—safety (for example nuclear) or
the proximity to the energy source (hydro-electric for example).
The electrical energy generated in remote generating stations needs to be
transferred to the load centres, which are cities and industrial areas that may be
several hundreds or thousands of kilometres away through transmission systems.
The generators at the generating stations produce electrical energy at a voltage
in the range of 500 V to 11 kV, and a large generation station (a thermal power
station for example) could produce thousands of megawatts of power. If such large
amounts of power were to be transmitted over long distances at the generated
voltage, the currents that will flow will result in very large ohmic losses leading
to a very inefficient transmission system. In order to transmit large amounts of
power over long distances, the voltage level needs to be increased. With the power
being the product of voltage and current, a higher voltage level will result in lower
currents and therefore lower ohmic losses. This operation of increasing the level
of voltage from the generating station to the transmission system is performed
110 4 Modelling and Simulating Transformers

using transformers. Using transformers, transmission systems have been energized

to voltage levels of several hundreds of kilovolts with the highest voltage levels
being around 800 kV.
To understand how electrical power transmission systems work, a good analogy
is with the network of roads that connect a country or region. Within a city or town,
there are usually a number of different types of roads with different speed limits.
In residential areas of the city, roads might be two lane or four lane roads with a
speed limit of 40 km/hr or 50 km/hr. There are special zones such as schools and
hospitals where the roads will have reduced speed limits of 20 km/hr or 30 km/hr
with excessive speeds being prevented by speed bumps. On the other hand, arterial
roads that connect different parts of the city together might have six lanes and
higher speed limits of 60 km/hr to 80 km/hr to enable drivers to move around faster.
Highways that connect distant parts of the city or different cities together can have
speed limits in the excess of 100 km/hr, and driving too slow will not be permitted.
The philosophy being planning the road network is to enable drivers to move around
as quickly as possible while ensuring safety of drivers and also pedestrians and
surroundings. Driving too fast in a residential area is a danger to pedestrians, and
road safety planners will implement features such as speed traps and speed cameras
to spot violators. However, in a highway, speed limits when they exist will be to
ensure safety of drivers but otherwise highways exist to facilitate rapid movement
of vehicles.
In exactly the same manner, there exist a number of different voltage levels in an
electric grid. At homes and offices, voltage levels are typically either 240 V or 120 V
(in the USA and Canada) with most appliances rated for these voltages. The reason
is that power consumed by domestic appliances is rarely larger than a few thousand
watts and at 120V or 240V, the current drawn by these appliances will be at most
a few tens of amperes. Currents larger than a few tens of amperes in a domestic
situation can be dangerous and cause fire safety issues. It is important to note that
a large number of house fires are caused due to short-circuits that produce large
currents in the order of several tens or hundreds of amperes. The voltage level at
which apartment buildings and houses are supplied however is usually much higher.
This is due to the fact that a single feeder will supply a row of houses or apartment
buildings, and therefore the overall power supplied by this feeder maybe in the range
of several thousands of watts. For example, a row of apartment buildings might be
supplied by a feeder with a voltage range of 3.3 kV to 11 kV. The higher voltage level
makes it possible to supply this augmented power at lower current levels. Similarly,
as we progress upwards in the grid with segments of the city supplied by a feeder,
the voltage levels of the feeder can be in the range of 11 kV to 33 kV or even higher.
Finally, the main transmission lines that supply the city will be at far higher voltage
levels in the range of 110 kV to 400 kV.
Therefore, comparing the electric grid and the road network, the power supplied
by a transmission line or feeder in the electric grid is equivalent to the volume or
number of vehicles passing through a highway or road in the road network. The
voltage level of a transmission line or feeder in the electric grid is equivalent to
the speed limit of a highway or a road in the road network. The current flowing
4.3 Transformer Basics 111

through a transmission line or feeder in the electric grid is equivalent to the density
of vehicles in a highway or a road in the road network. Transformers along with
a number of other equipment ensure that we are able to transfer large amounts of
power over long distances while being able to convert them from one voltage level
to another to ensure safety of the system. Without transformers, we would have just
a single voltage level across the entire grid, which would be either too low to be
totally inefficient or too high to be potentially dangerous. Therefore, one needs to
take a moment to appreciate the role played by transformers in the modern power
system.
Besides being a critical component in the modern power system for transforming
voltage levels, transformers are also used for numerous other purposes in isolating
circuits for a number of different reasons [10, 13, 14, 26]. As an example, when
designing control circuits, different parts of a controller may need to be electrically
isolated to ensure they do not interfere with each other. This is quite often the case
in a power converter with multiple power devices, with the gate driver of each
power device energized by an isolated power supply. The power supplies may need
to be isolated to ensure that there is no circulating current that can accidentally
cause a power device to conduct when it should not conduct. Transformers are also
used in certain power converter topologies such as cascaded converters to ensure
that circulating currents between modules is not possible [10, 26]. Some of these
examples will be simulated to show the role played by transformers.
The idea behind this basic background was to throw some light on the importance
of transformers in electrical engineering. It is difficult for a power engineer to not
encounter a transformer at some point of time during a project. Though there are
transformer-less appliances that are designed specifically without a transformer to
decrease the size and cost, a vast number of appliances and systems will need
transformers to function. In the next section, we will gradually ease into the
modelling of transformers from the very basic concepts of physics as we did in
the previous chapters for inductors and coupled inductors.

4.3 Transformer Basics

In this section, we will get started with the basic operating principle of a transformer
[13, 14, 27, 28]. We will leave the detailed model and simulation for a later section.
As we did in the previous chapter with coupled coils, we will gradually meander
through the concept of how a transformer comes into being and how it operates.
The previous chapter covered in detail the modelling and simulation of magnetically
coupled coils. Under the hood, a transformer is merely a set of magnetically coupled
coils. A transformer is a special case of magnetically coupled coils, which is so
designed to maximize the transfer of energy from one coil to the other while
minimizing losses and leakage. This is essential to ensure that the power system
remains efficient with minimal losses while voltage levels are being raised or
lowered using transformers.
112 4 Modelling and Simulating Transformers

Fig. 4.1 Two coils wound on

the same core
i

v2
v1

Let us start with the basic setup of two coils wound on an iron core as shown in
Fig. 4.1. A detailed discussion can be found in the previous chapter. Let us quickly
repeat some of the basic expressions so as to extend our discussion to be more
specific to transformers. The voltage v1 applied to the coil wound on the left limb
(let us call it coil 1) can be expressed as

v1 = e1 + i1 r1 (4.1)

where r1 is the winding resistance, e1 is the induced emf and i1 is the current in coil
1. This expression is merely Kirchhoff’s Voltage Law applied to coil 1. Since the
coil wound on the right limb (let us call it coil 2) is left open-circuited, no current i2
will flow through it and therefore,

v2 = e2 (4.2)

The voltage v2 across the terminals is equal to the emf e2 induced in the coil.
Using Faraday’s Law, the emf induced in the coils is equal to the rate of change
of flux linkages of the coils:

ψ1
e1 = = v1 − i1 r1 (4.3)
dt
ψ2
e2 = = v2 (4.4)
dt
If L1 and L2 are the self-inductances of coil 1 and 2, respectively, and M is the
mutual inductance between the coils, the flux linkages can be expressed as

ψ1 = L1 i1 (4.5)

ψ2 = Mi1 (4.6)

These equations are for the special case of Fig. 4.1 where coil 2 is open-circuited.
Therefore, the flux linkages of coils 1 and 2 are only due to the current flowing in
4.3 Transformer Basics 113

coil 1. The concept of mutual inductance has been described in detail in the previous
chapter, and the reader is encouraged to revisit the previous chapter if the above
expression is unclear.
Due to the current i1 flowing in coil 1, the flux produced in the core links with
coil 2 to produce flux linkages ψ2 . If the current i1 changes, the flux linkages ψ1
and ψ2 will change as well due to which emfs e1 and e2 will be induced in the coils.
The emf e2 induced in coil 2 produces a voltage v2 across the terminals of the coil.
Since coil 2 is open-circuited, the voltage produced across the coil terminals will
not result in any power consumed. But, if we do connect a load such as a simple
resistive load, a current i2 will flow through coil 2. With this, we have now achieved
transfer of energy from coil 1 to coil 2 without any electrical connection between
the coils. This phenomenon of energy transfer has been described in detail with
several different possible variations in the previous chapter. At a very basic level,
this transfer of energy is the basis of operation of a transformer.
Following this very basic discussion viewing a transformer as merely coils that
are magnetically coupled, we can now extend the discussion to the specifics of
a transformer. In the previous chapter, we had considered the coils in Fig. 4.1 to
be either connected to two independent isolated circuits or in the case of the Ćuk
converter, to different parts of the same circuit. In the case of a transformer, unless in
very specific cases, the different coils are usually connected to independent isolated
circuits. Furthermore, since the objective of a transformer is to transfer power from
one isolated circuit to another, the primary specification of a transformer is the
maximum power that can be transferred from one isolated circuit to other isolated
circuits [13, 14, 27, 28]. This is in contrast to a set of mutually coupled coils that
can be specified only with respect to their self- and mutual inductances.
Therefore, while all the equations written above involving the self- and mutual
inductances hold true for a transformer just like they did for magnetically coupled
coils, the approach for arriving at these equations for transformers is quite different.
Along with the primary specification of the transformer being the maximum power
rating, the next important specifications are the maximum rated voltages that can
be applied to each coil of the transformer [13, 14, 27, 28]. At this point, we need
to rename the coils wound on the core as the windings of the transformer as that
is the terminology that is commonly used with transformers. These windings of the
transformer may or may not be connected to an energized system. It is important to
note that more than one winding of a transformer could be energized as it is common
to use a transformer to interface two different systems at different voltages. There
may be cases where only one winding is energized since the other winding might
feed a load or another passive circuit.
Before we continue with our description, let us introduce an example with
some specifications so as to make the discussion easier to follow. Let us consider
a transformer with two windings and a maximum power rating of 10,000 Volt-
Amperes or 10,000 VA or 10 kVA. Volt-Amperes is the rating for apparent power,
which includes the active power in Watts as well as the reactive power in Volt-
Ampere-Reactive (VAR). Let us suppose that this transformer has two windings
and both windings have maximum voltage ratings of 240 V. Moreover, we can also
114 4 Modelling and Simulating Transformers

Fig. 4.2 Simple winding

i1 i2
depiction of a transformer

v1 v2

specify that this transformer should be operated at a frequency of 50 Hz or 60 Hz

(in USA or Canada). Quite a simple transformer, but to begin with, this should be
sufficient to help us understand the transformer parameters. Such a transformer can
transfer a maximum of 10 kVA of apparent power from any one winding to the other.
A maximum of 240 V can be applied to either or both windings. For example, we
could connect terminals of one winding of the transformer to the live and neutral
of a domestic power outlet that supplies 240 V, 50 Hz single-phase ac supply and
connect to the other winding a simple load resistor.
Figure 4.2 shows how a transformer can be depicted in a circuit [13, 14].
In comparison to the previous chapter, it is quite clear that the depiction of the
transformer is very similar to magnetically coupled coils. In most circuit diagrams,
a transformer contains the vertical lines between the windings that indicate the core
of the transformer and that the windings are wound on the same core. However, the
windings will not always have the dot polarity marked on their windings in which
case it is assumed that the upper terminals of the windings are marked by dots.
However, if one of the windings is wound in the opposite sense, dots marking one
of the terminals of all the windings are usually present. Simulations will make this
more clear in the later sections.
With respect to the sample specifications of the transformer, the voltages v1 and
v2 that are applied to the windings can have a maximum RMS value of 240 V. Since
the maximum power rating of the transformer is 10 kVA, the currents i1 and i2
should be limited to a maximum RMS value of 10,000/240 = 41.667 A. Applying
excessive voltage to either winding could result in saturation of the transformer
core and will destroy/burn the windings in the case of voltages much higher than
the maximum rated voltage [13, 14, 27, 28]. If the circuit connected to a winding
draws a current much greater than the maximum rated current, the ohmic losses will
increase resulting in an increased operating temperature that is detrimental to the
insulation of the wires and other parts of the transformer [13, 14, 27, 28].
In order to develop a simulation model for a transformer, we need to apply
the equations involving L1 , L2 and M. For this, we need to build a transformer
equivalent circuit [13, 14, 27, 28]. In practice, for a particular transformer, if we wish
to analyse or simulate it, we will also build the transformer equivalent circuit. If we
have a physical transformer and can perform experiments on it, there are tests such
as the short-circuit test and the open circuit test using which most of the significant
4.3 Transformer Basics 115

parameters of the equivalent circuit can be determined [13, 14, 27, 28]. However,
without conducting experiments, it is also possible to determine the approximate
parameters of the equivalent circuit by the process of estimation. This has the added
advantage of being able to examine the behaviour of the transformer for variations
in the equivalent circuit parameters. Let us develop the equivalent circuit of the
transformer step by step.
Let us suppose that the transformer of Fig. 4.2 is energized on winding 1 with a
voltage v1 applied to it. Let us suppose winding 2 is left open-circuited. We have
already applied Kirchhoff’s Voltage Law to winding 1:

v1 = e1 + i1 r1 (4.7)

If winding 1 were to have an inductance of L1 , the above equation can be rewritten

as
di1
v1 = L1 + i1 r1 (4.8)
dt
Winding 1 being a coil has an inductance L1 and a winding resistance r1 and
the voltage applied across the winding is opposed by the impedance offered by the
winding. The RMS value of the current through winding 1 can be expressed as

V1
I1 =  (4.9)
(ωL1 )2 + r12

where I1 and V1 are the RMS values of current i1 and voltage v1 , respectively, while
ω = 2πf is the angular frequency of the system.
We have chosen the maximum voltage rating of both windings of the transformer
to be 240 V. While building the equivalent circuit of the transformer, let us assume
that the voltage applied to winding 1 is equal to the maximum rated voltage of 240 V.
If an RMS value I1 is known, the impedance of winding 1 can be calculated. To be
able to assume a value for I1 , we need to ask the question, in the assumed state of
the transformer with a voltage applied to winding 1 and winding 2 open-circuited,
what happens in the transformer? Essentially, we are performing the open circuit
test on the transformer where the rated voltage is applied across a winding and the
other winding is left open-circuited.
In the previous chapter, we had discussed many conditions in which magnetically
coupled coils can be energized. We can quickly repeat the final result and the
reader is encouraged to revise the previous chapter if needed. With winding 1 alone
energized, a current i1 will flow through it and current i2 is zero since winding 2 is
open-circuited. This current i1 will establish a magnetic flux in the core. Since the
voltage v1 applied across winding 1 is an ac voltage, the core flux will be an ac flux
as well. By Faraday’s Law, emfs e1 and e2 will be induced in the two windings. The
only effect of the current i1 is to set up the magnetic flux in the core. This current
i1 is called the magnetizing current of the transformer. It is important to remember
116 4 Modelling and Simulating Transformers

that we could do exactly the same test on winding 2—apply a voltage v2 across
winding 2 with an RMS value equal to the maximum rated value and leave winding
1 open-circuited. In that case, no current i1 will flow in winding 1 and the current i2
flowing in winding 2 will also be the magnetizing current.
The transformer is a magnetic equipment as the fundamental basis of operation of
the transformer is the magnetic flux in the core. Therefore, the magnetizing current
flowing in the energized winding is absolutely essential [13, 14, 27, 28]. Since one
winding is open-circuited, the transformer is not transferring any power from one
winding to the other. The transformer is still consuming power since there is ohmic
loss in winding 1 and core losses (eddy current and hysteresis losses) since there
is a magnetic flux in the core. These losses are called the no-load losses of the
transformer [13, 14, 27, 28]. These no-load losses are an extremely small fraction
of the maximum power rating of the transformer. This is fairly obvious since we
would like the transformer to be as efficient as possible. At no load the ohmic losses
in winding 1 are usually negligible compared to the core losses.
With this background, we make an assumption and also use commonly known
data related to transformers. The magnetizing current flowing through a winding is
usually around 1–5% of the rated maximum current of the winding [13, 14, 27, 28].
If we assume the magnetizing current to be 2% of the rated maximum current, the
winding impedance can be calculated as

V1rated
(ωL1 )2 + r12 = 50 (4.10)
I1rated

We have already assumed that the voltage v1 applied across winding 1 is equal to the
maximum rated voltage. Therefore, the above expression provides the impedance
with respect to the winding ratings.
In the impedance of the transformer winding, the winding resistance will be
negligible as otherwise the no-load losses will be significant even if the current
drawn was equal to the magnetizing current. Therefore, a further approximation can
be made as follows:
V1rated
L1 ≈ 50 (4.11)
ωrated I1rated

ωrated is merely the nominal frequency of the system corresponding to 50 Hz or

60 Hz. The frequency plays a role in determining the flux density in the core, which
we will discuss later. However, without going into the details of the magnetic circuit,
we have obtained the self-inductance of winding 1. We can use this expression for
winding 2 as well:

V2rated
L2 ≈ 50 (4.12)
ωrated I2rated
4.3 Transformer Basics 117

i1
r1 r2 i2

v1 Rc1 L1 L2 Rc2 v2

Fig. 4.3 Basic transformer equivalent circuit with self-inductances

Now that we have computed the self-inductances L1 and L2 of winding 1 and 2,

respectively, we can draw a very basic equivalent circuit of the transformer as shown
in Fig. 4.3. The reader should observe that the dot polarities have been omitted
from the windings. When a transformer is depicted without dot polarities, it can
be assumed that the dots are on the upper terminals of the windings as shown in
Fig. 4.2. From Fig. 4.3, the induced emfs in the windings appear across the self-
inductances and are the difference between the applied voltages and the ohmic drop
across winding resistances r1 and r2 . We have included two more resistances—Rc1
and Rc2 . These resistances represent the core losses in the transformer. In practice,
we would estimate the core losses from the open circuit test as the no-load losses are
largely due to the core losses. However, for the simulation model, we can estimate
the core losses by assuming that the core losses are 1% of the maximum rated power
of the transformer [13, 14, 27, 28]. Therefore,
2
V1rated
Rc1 ≈ (4.13)
0.01Prated
2
V2rated
Rc2 ≈ (4.14)
0.01Prated

To complete the basic equivalent circuit of Fig. 4.3, we need values for the
winding resistances r1 and r2 . In practice, the winding resistances are determined
by performing a short-circuit test on the transformer in which one of the windings
of the transformer is short-circuited and the other winding is energized with a
voltage such that the currents flowing in the windings are equal to the rated winding
currents. In such a case, the losses in the transformer are primarily the ohmic losses
as the core losses are negligible due to the very low voltage that is needed to be
applied when one of the windings is short-circuited. For a simulation model, we can
estimate the winding resistances r1 and r2 using values that are normally found in
transformers. We can assume the winding resistances r1 and r2 to be such that if a
rated current were to be flowing through the windings, the voltage drop across the
118 4 Modelling and Simulating Transformers

winding resistances would be approximately 0.5 to 1% of the rated winding voltage

[13, 14, 27, 28]. Therefore,

V1rated
r1 ≈ 0.01 (4.15)
I1rated

V2rated
r2 ≈ 0.01 (4.16)
I2rated

With the equivalent circuit of Fig. 4.3, we can use the equations once we have
a value for the mutual inductance M. The mutual inductance can be computed as
before with respect to the self-inductances as follows:

M = k L1 L2 (4.17)

where k is the coupling factor between the transformer windings. This expression
was used in the previous chapter when modelling and simulating magnetically
coupled coils and takes into account the fact that not all of the flux generated by
the current flowing through a winding will link with the other winding. However, a
transformer is so designed as to minimize the leakage of flux produced by a winding.
Therefore, instead of choosing a value of the coupling factor k, the leakage in the
transformer windings can be better modelled in another manner.
Instead of representing the inductance of the windings by the self-inductances
L1 and L2 alone, we can separate the inductances into two parts—magnetizing
inductances Lm1 and Lm2 that produce the flux that flows through the core and links
with the other winding and leakage inductances Ll1 and Ll2 that will produce fluxes
that leak into the air [13, 14, 27, 28]. This is shown in Fig. 4.4 [13, 14, 27, 28]. The
magnetizing inductances Lm1 and Lm2 take the place of the self-inductances L1
and L2 as now these inductances are responsible for producing the core flux and
therefore the induced emfs in the windings. The leakage inductances Ll1 and Ll2
are in series with the winding resistances r1 and r2 , and the voltage drops across
them result in a decreased induced emf.
We can estimate the leakage inductances in the same way as we did for many of
the other parameters. The leakage inductances will merely decrease the induced emf

i1 r2 i2
r1 Ll1 Ll2

v1 Rc1 Lm1 Lm2 Rc2 v2

Fig. 4.4 Basic transformer equivalent circuit with magnetizing and leakage inductances
4.3 Transformer Basics 119

due to the leakage of flux. Therefore, we assume the leakage inductances to be such
that when rated currents are flowing through the windings, the voltage drop across
the leakage inductances will be 1–2% of the rated winding voltage [13, 14, 27, 28].
As a result, we arrive at the following expressions:

V1rated
Ll1 ≈ 0.02 (4.18)
ωrated I1rated

V2rated
Ll2 ≈ 0.02 (4.19)
ωrated I2rated

The mutual inductance M is merely expressed with respect to the magnetizing

inductances as follows:

M= Lm1 Lm2 (4.20)

The coupling factor k has been removed as the leakage of flux has been accounted
for separately through the leakage inductances. After all, as per definition, the
magnetizing inductances Lm1 and Lm2 produce fluxes that flow through the core
and link completely with other windings. However, there is one last complication—
we have computed the self-inductances L1 and L2 as well as the leakage inductances
Ll1 and Ll2 . We need to find expressions for the magnetizing inductances Lm1 and
Lm2 in order to compute the mutual inductance M.
The magnetizing inductances Lm1 and Lm2 can be computed rigorously using
network laws applied to the equivalent circuit of Fig. 4.4. However, a simplification
to the equivalent circuit will make the computation much simpler without much
difference to the end result. Figure 4.5 shows the simplification in which the core
loss resistors Rc1 and Rc1 are directly connected across the transformer terminals
rather than across the magnetizing inductances [13, 14, 27, 28]. This will result in
higher core losses since the core losses are due to the magnetic flux in the core
of the transformer, which is related to the induced emfs rather than the terminal
voltages. The rise in core losses will however be negligible since the voltage drop
across the winding resistances and the leakage inductances is the range of 1 to 3% of
the rated winding voltages. The simplification on the other hand brings the leakage

i1 i2
r1 Ll1 Ll2 r2

v1 Rc1 Lm1 Lm2 Rc2 v2

Fig. 4.5 Simplified transformer equivalent circuit

120 4 Modelling and Simulating Transformers

inductances in series with the magnetizing inductances. With this, we can express
the self-inductances as the sum of the leakage inductances and the magnetizing
inductances, resulting in the following expressions [13, 14, 27, 28]:

Lm1 = L1 − Ll1 (4.21)

Lm2 = L2 − Ll2 (4.22)

We now have an equivalent circuit for the transformer along with expressions to
estimate all the parameters of the equivalent circuit. Most importantly, all estimates
use only two specifications of the transformer—the maximum rated power in VA
and the maximum rated voltages of the windings. As will be shown in the next
section, the above equivalent circuit along with the parameter estimation provides
us with a very flexible yet realistic model with which a transformer can be simulated
in a number of different applications.

4.4 Simulating the Transformer

In the previous section, we described step by step how the equivalent circuit of the
transformer can be developed using the basic laws of physics. As already stated
before, under the hood, a transformer is just a set of magnetically coupled windings.
Therefore, the approach to simulation will be very similar to the approach taken
in the previous chapter. The main difference lies in being able to translate the
manufacturer specifications of the transformer into parameters of the equivalent
circuit so as to be able to solve the magnetic circuit [13, 14, 27, 28]. In this section,
we will begin with the simulation of a simple two winding transformer with both
windings having the same maximum voltage rating.
Figure 4.6 shows the circuit used for the simulation of the transformer. This
simulation can be found in the folder two_winding within chapter4_
transformers in the following link in the simulation repository:
https://github.com/opensourceelectrical/simulating-magnetics-for-power-
electronics

Fig. 4.6 Circuit for basic is Zs il

simulation of a transformer

Ll
vs
Rl
4.4 Simulating the Transformer 121

Let us consider a transformer with the same sample specifications as chosen in the
previous section—a maximum power rating of 10 kVA with both windings having
a maximum voltage rating of 240 V RMS and let us choose the operating frequency
to be 50 Hz. Let us energize winding 1 with a supply voltage vs having an RMS
value of 240 V and frequency of 50 Hz. The supply has a feeder Zs with a parasitic
impedance. Winding 2 of the transformer is supplying a load that is a resistor–
inductor load with resistance Rl and inductance Ll , which could be typical of an
industrial load such as a motor. The transformer internally has the equivalent circuit
of Fig. 4.5.
The transformer will be modelled as two voltage sources in series with resis-
tances as in the previous chapter while simulating magnetically coupled coils.
Therefore, with the external circuit of Fig. 4.6 connected to the voltage source in
series with resistance model of the transformer, all that needs to be done is to write
the control code that will model the transformer and generate the voltages to be
fed to the voltage sources. To begin with, let us compute the parameters of the
transformer equivalent circuit from the transformer specifications that are usually
provided by the manufacturer:
import math
d t = 1 . 0 e−6
VArated = 1 0 0 0 0 . 0
V1rated = 240.0
V2rated = 240.0
f r a t e d = 50.0
o m e g a _ r a t e d = 2∗ math . p i ∗ f r a t e d

Since we are simulating a 50 Hz transformer and a 50 Hz ac system, a control

time step of 1 µs is sufficiently small to ensure accuracy. When we simulate
high frequency transformers in the next chapter, this time step will need to be
decreased.
The remaining parameters of the transformer equivalent circuit can be computed
from these manufacturer specifications as described in the previous section. In the
control code:
I 1 r a t e d = VArated / V 1 r a t e d
I 2 r a t e d = VArated / V 2 r a t e d
Z1rated = V1rated / I 1 r a t e d
Z2rated = V2rated / I 2 r a t e d
L1 = 5 0 . 0 ∗ Z 1 r a t e d / o m e g a _ r a t e d
L2 = 5 0 . 0 ∗ Z 2 r a t e d / o m e g a _ r a t e d
Ll1 = 0.02 ∗ Z 1 r a t e d / omega_rated
Ll2 = 0.02 ∗ Z 2 r a t e d / omega_rated
Lm1 = L1 − L l 1
Lm2 = L2 − L l 2
M12 = math . s q r t ( Lm1 ∗ Lm2 )
r1 = 0.01 ∗ Z1rated
r2 = 0.01 ∗ Z2rated
Rc1 = V 1 r a t e d ∗ V 1 r a t e d / ( 0 . 0 1 ∗ VArated )
Rc2 = V 2 r a t e d ∗ V 2 r a t e d / ( 0 . 0 1 ∗ VArated )
122 4 Modelling and Simulating Transformers

The reader is encouraged to simulate a transformer with different equivalent circuit

parameters such as increased or decreased leakage inductances or core loss resistors.
If the reader has access to a laboratory and a physical transformer, performing open
circuit and short-circuit tests can provide more realistic values for the equivalent
circuit parameters.
The above computations result in the following parameters of the equivalent
circuit:
L1 = L2 = 0 . 9 1 6 7 3
Lm1 = Lm2 = 0 . 9 1 6 3 6
M12 = 0 . 9 1 6 3 6
Ll1 = Ll2 = 0.00036
r1 = r2 = 0.0576
Rc1 = Rc1 = 5 7 6 . 0

Let us make a few observations of these parameters. The self-inductances of the

windings are close to 1H. In the chapter on simulating inductors, we had simulated
sample inductors considering the number of turns of the coil, the area of cross-
section of the core and the mean length of the core flux path. The reader can
use the expression for the inductance to arrive at a probable number of turns N1
and N2 of the transformer windings. Moreover, the reactance of the windings will
be such that the magnetizing current drawn by the windings will be limited to
around 2% of the rated current when the rated voltages are applied at the windings.
The resistors Rc1 and Rc2 that represent the core losses of the transformer can be
included in the magnetic model. However, since they are mere resistors, these can be
connected as Variable Resistor components in the transformer model by connecting
resistors across the Voltmeters as shown in Fig. 4.7. The remaining parameters of the
equivalent circuit—the self-inductances, mutual inductance and winding resistance
will feature in the mathematical model g(V ).

V V

g(V )

Fig. 4.7 Simulation model of transformer windings

4.4 Simulating the Transformer 123

The control code for the transformer model will need some interface parameters
as was the case for inductors and coupled coils in the previous chapters. The most
important interface variable is always the Time Event variable t1. This is quite
often neglected but when designing digital controllers or solving models digitally,
the time step at which the model needs to be solved must be accurately controlled.
Next, we need to define the inputs. As with coupled coils, there are two inputs—the
measurements of the Voltmeters shown in Fig. 4.7 that we can call v1 and v2. When
we simulate multi-winding transformers, we will have more than two windings
and therefore, the measurement of applied voltage across each winding will be an
input. Finally, we can define the outputs. One set of outputs will be for the Variable
Resistor components. The Variable Resistor components in series with the voltage
source can be called res_output1 and res_output2. The Variable Resistor
components in parallel with the Voltmeters represent the core losses, and these have
already been computed as Rc1 and Rc2 and so these variables can be used to define
the outputs. The last set of outputs will be the voltages of the Controlled Voltage
sources and we can define them as vout1 and vout2.
After defining the inputs and outputs of the controller, we can define Static
Variables and Variable Storage elements. Variable Storage elements are useful to
plot internal control variables but are not essential for a working controller if the
controller is an isolated component. Variable Storage elements are also useful for
connecting different controllers, which is not the case in this simulation. Therefore,
to keep the simulation to the bare minimum, we will skip defining Variable Storage
elements. Static Variables are those control variables that need to be stored between
successive iterations of the simulation. As the magnetic model of the transformer
is merely the same as that of coupled coils, the only essential Static Variable that
needs to be defined is winding_currents. We will solve the magnetic model as
matrix equations as described in the previous chapter. Therefore, this static variable
will be initialized as a Python list as follows:
i f t _ c l o c k <= d t :
winding_currents = [0.0 , 0.0]

Following these computations and initializations, the transformer model can be

solved iteratively as follows:
i f t _ c l o c k >= t 1 :
# Model
t 1 += d t

Since the solution of the transformer mathematical model involves numerical

integration, the above conditional block is critical to ensure that the numerical
integration occurs once every 1 µs. The mathematical model of the transformer is
identical to that of coupled coils and can be repeated as follows:
124 4 Modelling and Simulating Transformers

L = [[L1, M12], [M12, L2]]

R = [[r1, 0.0], [0.0, r2]]
V = [v1, v2]
for count1 in range(len(L)):
if not L[count1][count1]:
for count2 in range(count1+1, len(L)):
if L[count2][count1]:
L[count1], L[count2] = L[count2], L[count1]
R[count1], R[count2] = R[count2], R[count1]
V[count1], V[count2] = V[count2], V[count1]
break
if L[count1][count1]:
for count2 in range(count1+1, len(L)):
comm_factor = L[count2][count1]/L[count1][count1]
for count3 in range(len(L[count1])):
L[count2][count3] -= L[count1][count3]*comm_factor
R[count2][count3] -= R[count1][count3]*comm_factor
V[count2] -= V[count1]*comm_factor
dibydt = [0.0, 0.0]
for count1 in range(len(L)-1, -1, -1):
if L[count1][count1]:
k = [0.0, 0.0, 0.0, 0.0]
for k_count in range(len(k)):
k[k_count] = V[count1]
for count2 in range(count1+1, len(L)):
k[k_count] -= L[count1][count2]*dibydt[count2]
for count2 in range(len(R)):
if k_count==0:
if count2>count1:
k[k_count] -= R[count1][count2]*(winding_currents[count2] + \
dibydt[count2]*dt)
else:
k[k_count] -= R[count1][count2]*winding_currents[count2]
if k_count==1:
if count2==count1:
k[k_count] -= R[count1][count2]*(winding_currents[count2] + \
k[0]*dt/2.0)
elif count2>count1:
k[k_count] -= R[count1][count2]*(winding_currents[count2] + \
dibydt[count2]*dt)
else:
k[k_count] -= R[count1][count2]*winding_currents[count2]
if k_count==2:
if count2==count1:
k[k_count] -= R[count1][count2]*(winding_currents[count2] + \
k[1]*dt/2.0)
elif count2>count1:
k[k_count] -= R[count1][count2]*(winding_currents[count2] + \
dibydt[count2]*dt)
else:
k[k_count] -= R[count1][count2]*winding_currents[count2]
4.4 Simulating the Transformer 125

if k_count==3:
if count2==count1:
k[k_count] -= R[count1][count2]*(winding_currents[count2] + \
k[2]*dt)
elif count2>count1:
k[k_count] -= R[count1][count2]*(winding_currents[count2] + \
dibydt[count2]*dt)
else:
k[k_count] -= R[count1][count2]*winding_currents[count2]
k[k_count] = k[k_count]/L[count1][count1]
dibydt[count1] = (k[0] + 2.0*k[1] + 2.0*k[2] + k[3])/6.0
winding_currents[count1] += dibydt[count1]*dt
else:
winding_currents[count1] = V[count1]
for count2 in range(count1+1, len(L)):
winding_currents[count1] -= L[count1][count2]*dibydt[count2]
for count2 in range(count1+1, len(R)):
if not count2==count1:
winding_currents[count1] -= R[count1][count2]*winding_currents[count2]
winding_currents[count1] = winding_currents[count1] / R[count1][count1]

vout1 = v1 - winding_currents[0]*res_output1
vout2 = v2 - winding_currents[1]*res_output2

Fig. 4.8 Transformer winding voltages

The results of the simulation are shown in Figs. 4.8 and 4.9. Since, we are
simulating a transformer with the same voltage ratings on both windings (also
called a 1:1 transformer), the voltages at both windings in Fig. 4.8 appear to be
overlapping. The voltage across the load will be lesser than the voltage at the
source by the ohmic drop across the winding resistances and leakage inductances.
Figure 4.9 shows the currents in the transformer windings that at first glance appear
to have a few anomalies. The currents are not overlapping in the same manner as
126 4 Modelling and Simulating Transformers

Fig. 4.9 Transformer winding currents

the winding voltages in Fig. 4.8. This is due to the fact that the primary winding
(winding 1) that is energized by the voltage source draws a magnetizing current that
is around 2% of the maximum rated current of the winding. Since only one winding
is energized by the transformer, the magnetizing current will be drawn from winding
1, resulting in the current in winding 1 being noticeably larger than the current in
winding 2. There is another anomaly that seems like an error but in reality can be
explained—the current in winding 1 appears to have a dc offset.
As already stated before, as per Faraday’s Law, the induced emfs e1 and e2 in
the two windings will be directly proportional to the rate of change of core flux.
Therefore, the core flux will be directly proportional to the integral of the induced
emfs. Since, the voltage drops across the winding resistance and leakage inductance
is a mere 2–3% of the rated winding voltage, the induced emfs can be approximated
to the applied voltages. Therefore, the following equation can be written:

φ ∝ v1 dt (4.23)

The integral of a function will produce an integration offset, which will depend
on the nature of the function being integrated and the initial conditions (at t = 0).
The voltage source from Fig. 4.8 can be seen to be sine waveform. As we know,
the integral of a sine waveform is the sum of a cosine waveform and an integration
offset as follows:

Vm
φ ∝ Vm sin ωtdt = − cos ωt + C (4.24)
ω
The integration offset can be determined by applying initial conditions. If we assume
the transformer to be not energized at time t = 0, the flux φ in the core is zero at
4.4 Simulating the Transformer 127

t = 0. The flux in the core cannot abruptly increase with the application of a voltage
and therefore, the flux will gradually increase. This implies
Vm Vm
0=− + C ⇒ C = (4.25)
ω ω
The above equation implies that the core flux has a dc offset. This might seem
confusing as we had said that the flux cannot change abruptly upon application of
a voltage and will gradually increase. But, the dc offset merely ensures that the
waveform of the core flux starts as zero as core flux is expressed as the sum of a
cosine waveform and a dc offset.
Vm Vm
φ∝− cos ωt + (4.26)
ω ω
The cosine waveform alone would have resulted in the core flux abruptly increasing
to the negative maximum value immediately upon the application of the voltage.
The dc offset on the other hand ensures that despite being a cosine waveform, the
flux must conform to the physical constraint of not changing abruptly.
Therefore, the core flux has a dc offset. The magnetizing current drawn by
winding 1 is responsible for maintaining this core flux. The core flux is directly
proportional to the current in winding 1 since the core flux is merely the ampere
turns of winding 1 against the reluctance of the flux path. This in turn implies that the
magnetizing current has a dc offset. It is this dc offset that is evident in Fig. 4.9. The
next question that would follow would be—how is it acceptable that a transformer
connected to an ac circuit is drawing a current that has a dc component? This dc
offset in the magnetizing current is only temporary and will decay with the time
constant of the circuit. The time constant of winding 1 is

L1 0.91673
τ= = = 15.92 s (4.27)
r1 0.0576

The settling time of a circuit is 4 times the time constant of the circuit, which implies
that the dc offset will disappear in 63.7 s.
To verify this effect, we can repeat the simulation while choosing the applied
voltage at winding 1 to be a cosine waveform instead of a sine waveform as follows:
√
vs = 2 × 240 cos 100π t (4.28)

In this case, the core flux being approximately the integral of the applied voltage
can be expressed with a zero integration offset as merely (Fig. 4.10):

Vm
φ∝ sin ωt + 0 (4.29)
ω
Figure 4.11 shows the winding currents when the source voltage vs still has a
magnitude of 240 V RMS but is a cosine waveform of 50 Hz frequency as shown in
128 4 Modelling and Simulating Transformers

Fig. 4.10 Transformer winding voltages with a cosine source voltage

Fig. 4.11 Transformer winding currents with a cosine source voltage

Fig. 4.10. The currents do not have a dc offset, though there is a one cycle transient
which is due to the time constant of the load connected at winding 2. As can be seen,
current i1 in winding 1 is still noticeably larger than current i2 in winding 2, due to
the magnetizing current being drawn by winding 1. The reader is encouraged to try
out different initial phase angles to the source voltage waveform to examine the dc
offset in the current in winding 1.
4.5 Understanding Transformer Winding Currents 129

There are several variations to this basic simulation that the reader can attempt.
As an example, the reader can energize winding 2 instead of winding 1 and observe
that the results change only to the extent that now the magnetizing current is drawn
by winding 2. On the other hand, there are several nuances to the operation of a
transformer that most often confuses a beginner. In the next few sections, we will
gradually introduce several more simulations that will dive deeper into the operation
of the transformer. The reader is advised to review the basic discussion section of
the previous chapter on magnetically coupled coils, as the next few sections will use
those concepts repeatedly.

4.5 Understanding Transformer Winding Currents

In the previous section, we had started off with a basic simulation of a two winding
transformer. We had energized one of the windings and had connected a load to the
other winding. The simulation results showed the load being supplied by a voltage
approximately equal to that applied to the other winding and the load current being
supplied by the source in addition to a magnetizing current. This simple simulation
got us started to understanding transformer operation. However, to fully understand
how the current in one winding affects the current in the other, it would be interesting
to dig a little deeper into the working of the transformer using simulations [13, 14,
27, 28]. We can continue with the same simulation of a two winding transformer as
in the previous section, and just alter some details as necessary.
To begin with, let us increase the load resistor Rl in Fig. 4.6 to a very large value
of around 10 k such that the transformer is now operating at a very light load
and therefore approximately no load. The currents in the transformer windings are
shown in the simulation results of Fig. 4.12. Even though the load has such a large
resistance that the current drawn by the load should be negligible, the currents in
both windings appear to have non-negligible values. At no load, the transformer will
draw currents to supply the core losses in the transformer that have been represented
by the two resistors Rc1 and Rc2 [13, 14, 27, 28]. The current il in winding 2 shown
in Fig. 4.12 is the current drawn by the load Rl and Ll . The current is on the other
hand is the current drawn by the loss resistors Rc1 , Rc2 , load current il and also the
magnetizing current of the transformer [13, 14, 27, 28].
It might be a bit confusing to understand the different currents. Let us compare
the simulation results of Fig. 4.12 with the simulation results of Fig. 4.13, which
show the winding currents i1 and i2 (note the direction of i2 ) as depicted in the
equivalent circuit of Fig. 4.5. The winding currents are clearly different from the
source and load currents. The load current il in Fig. 4.12 is the current drawn by the
load Rl -Ll alone and is negligible. The current i2 is the current drawn by the loss
resistor Rc2 and the load current. This explains how the current i2 in Fig. 4.13 is
larger than the current il in Fig. 4.12. The current i1 on the other hand is the sum of
the currents drawn by the loss resistor Rc2 , the load current il and the magnetizing
130 4 Modelling and Simulating Transformers

Fig. 4.12 Transformer source and load currents at no load

Fig. 4.13 Transformer winding currents at no load

current. All the currents can be found on the different branches of the simplified
transformer equivalent circuit of Fig. 4.5 except for the magnetizing current.
Let us now talk about this mysterious magnetizing current that appears in i1 and
is . The magnetizing current is the current drawn by the transformer to maintain
the magnetic flux in the core. Since the magnetic flux in the transformer core is
4.5 Understanding Transformer Winding Currents 131

fundamental to its operation, the magnetizing current is also fundamental to the

operation of the transformer [13, 14, 27, 28]. An analogy to the magnetizing current
can be made to the altitude an airplane needs to reach after it takes off. A normal
commercial airplane will probably rise to an altitude of 20,000 feet from sea level
and maintain that height until it reaches its destination. Whether the airplane is
booked to capacity with passengers or is almost empty, the airplane still needs to
rise to this height. This is the intrinsic cost of any commercial aerospace flight.
Similarly, when a transformer is energized and is able to serve a load on one or more
of its windings, it needs to draw a magnetizing current to establish the magnetic flux
in its core that makes it possible to serve loads on multiple windings. Let us now
examine in depth, how this magnetizing current comes into being.
Even when the transformer is operating at no load, it will continue to draw a
magnetizing current. This magnetizing current is essential to set up the core flux of
the transformer. When any one winding of the transformer is energized, a core flux
will be established. This can be deduced by the combination of Kirchhoff’s Voltage
Law and Faraday’s Law to winding 1:

dφ
v1 = N1 + i1 r1 (4.30)
dt
where the flux linkage ψ1 of winding 1 is N1 φ with N1 being the number of turns
of winding 1. The core flux is therefore [13, 14, 27, 28]:

1
φ= (v1 − i1 r1 )dt (4.31)
N1

The core flux links with winding 2 as well inducing an emf in it and therefore
producing a voltage across the terminals of winding 2. In this no-load case, the
current drawn by the load connected to winding 2 is approximately zero. However,
in order that the flux φ flows in the transformer core, a current must be drawn by one
or both of the windings. This follows from our discussion of magnetic circuits in the
previous chapters, where one coil or multiple coils wound on a core can be solved
as a magnetic circuit with the ampere turns of the coils being the sources against the
reluctance offered by the core. In this case, if we assume a very simple core (such
as a rectangular core) [13, 14, 27, 28]:

N1 i1 + N2 i2
φ= (4.32)
ρ

where ρ is the reluctance of the core:

l
ρ= (4.33)
μ0 μr Ac

The reason why the net ampere turns in the magnetic circuit have been chosen
as the sum of the ampere turns of the two windings is due to the sense in which
132 4 Modelling and Simulating Transformers

the windings have been wound and the directions we have chosen for the currents
i1 and i2 in Fig. 4.5. For this, the reader is advised to review the previous chapter
where a detailed explanation is provided of the nature of induced emfs due to the
sense of windings and also of the dot polarity convention. Since, for the transformer
of Fig. 4.6, the dot polarities have been omitted, this implies that the dots can be
placed at the upper terminals of the two windings.
In this particular simulation case, since a passive load is connected to winding
2, the current i2 will be such that it decreases the net ampere turns, and therefore
demagnetizes the core. With winding 1 being the only winding being energized by
a source, the current i1 flowing in winding 1 is the current that sets up the magnetic
flux in the core. If we neglect the currents drawn by the loss resistors Rc1 and Rc2 ,
we could express the net ampere turns as

N1 is − N2 il
φ≈ (4.34)
ρ

Given the direction of the supply/source current is and the load current il in Fig. 4.6,
it is fairly obvious that the current drawn by the load will demagnetize the core.
However, since in this no-load condition, the current il is negligible, we can further
make the approximation:

N1 is
φ≈ (4.35)
ρ

What can we conclude from the above discussion? When the transformer is
operating at close to no load, the transformer draws a current from the supply due to
the core losses, which have been represented by resistors Rc1 and Rc2 . Besides the
core loss component, the transformer will also draw a current called the magnetizing
current that is necessary to maintain the flux in the core. In the no-load case, these
are the only currents that are flowing in the transformer with respect to the simplified
equivalent circuit of Fig. 4.5 [13, 14, 27, 28].
Let us now bring back the load connected to winding 2 by restoring Rl =10 .
The simulation results have been presented in the previous section, so let us try
to analyse the condition in this section. In this load condition, the three currents
from the no-load condition will continue to flow—the core loss currents and the
magnetizing current [13, 14, 27, 28]. In addition, the current il drawn by the load
will no longer be negligible but will be quite significant. So, therefore, if we return
back to the approximate expression for flux of (4.34), it would appear that the load
current would significantly demagnetize the core leaving the core flux at a very low
value.
This is quite often a source of confusion for students and it was for me as well
when I was an undergraduate. However, we must remember that when winding 1
of the transformer is energized with a supply, the flux in the core is determined by
(4.31) [13, 14, 27, 28]. Though (4.31) contains a voltage drop term dependent on the
current i1 , with the winding resistance r1 being negligible, the flux will therefore not
4.5 Understanding Transformer Winding Currents 133

change significantly even if the current i1 changes from the minimal value equal to
the magnetizing current to the maximum value equal to the rated current of winding
1. Therefore, the core flux is determined by the magnitude of the supply voltage
with which the transformer is energized [13, 14, 27, 28].
Equations (4.31) and (4.34) appear to be contradictory. Equation (4.31) describes
a flux that will change minimally with changes in the winding current. Equa-
tion (4.34) on the other hand states that the flux will decrease as the load current
il increases. Both equations are correct and we merely need to understand them
as cause and effect [13, 14, 27, 28]. Equation (4.31) is the primary equation as the
supply voltage is the independent input that energizes the transformer, and therefore,
with this expression, we can arrive at the core flux. Equation (4.34) merely states
that the difference of the ampere turns of the windings will produce the core flux.
If the flux φ in the core will be unaffected but the load current il has a
demagnetizing effect, this points to only one possibility. The demagnetizing effect
of the load current il is cancelled out by another current that has a magnetizing effect
[13, 14, 27, 28]. In the no-load case, only the magnetizing current was flowing in
winding 1 of the transformer. However, to nullify the demagnetizing effect of il , a
current must flow in winding 1 such that φ is unaltered. Let us write this expression
as

N1 im + N1 il − N2 il
φ≈ (4.36)
ρ

In the above expression for flux, the ampere turns N1 im produced the core flux in
the no-load case. Therefore, it follows that the second and third terms must be equal
and cancel each other out resulting in [13, 14, 27, 28]

N1 il = N2 il (4.37)

The current il is the load current flowing in winding 2 referred to winding 1. Any
current flowing in any winding can be referred to another winding by using the
constraint that the referred current must produce the same ampere turns as the
original current [13, 14, 27, 28].
As a result, for the load current il flowing in winding 2, the current supplied by
the source in winding 1 will be

is = im + il (4.38)

This follows from simple energy conservation as the power consumed by the load
must be provided by the source in addition to the power lost in the transformer. The
transformer is merely a machine that transfers electrical energy from one winding
to another through a magnetic field thereby allowing the windings to be electrically
isolated. The transformer does not produce any energy, and therefore the energy
supplied at the load cannot be greater than the energy supplied at the source.
134 4 Modelling and Simulating Transformers


The referred current il leads us to another term that is extremely important for
transformers [13, 14, 27, 28]:

 N2
il = il (4.39)
N1

The ratio N 2
N1 is called the turns ratio of the transformer and is the ratio of the number
of turns of one winding with respect to the number of turns of the other [13, 14, 27,
28]. This turns ratio is also used to express the terminal voltage of one winding with
respect to the other.
We have already written the expression for the terminal voltage using Kirchhoff’s
Voltage Law and Faraday’s Law:

dφ dφ
v1 = N1 − i1 r1 ≈ N1 (4.40)
dt dt
dφ dφ
v2 = N2 − i2 r2 ≈ N2 (4.41)
dt dt
The above expressions can be rearranged and equated as
v1 v2
= (4.42)
N1 N2

To produce an expression for the voltages with respect to the turns ratio as follows:

N1
v1 = v2 (4.43)
N2

The above expression can be used to determine the turns ratio of the transformer
as the maximum rated voltages of the windings are specified by the manufacturer
[13, 14, 27, 28]. Therefore,

N1 v1rated
= (4.44)
N2 v2rated

With the knowledge of the turns ratio from the manufacturer specifications, we can
use the turns ratio to refer voltages and currents from one winding to another. This
provides a very convenient tool for mathematical analysis.
In the circuit topology of Fig. 4.6 chosen for the above simulations, only one
winding (winding 1) has been energized by an ac voltage source while the other
winding (winding 2) had a passive load connected to its terminals. In such a case,
the flow of power is fairly obvious—from the voltage source in winding 1 to the
load in winding 2. Moreover, the magnetizing current is drawn by winding 1 from
4.5 Understanding Transformer Winding Currents 135

is1 Zs1 is2 Zs2

vs1 vs2

Fig. 4.14 Circuit topology with more than one voltage source

the voltage source to maintain the core flux as it is the only energy source connected
to the transformer. In many cases, more than one winding of the transformer will
be energized as the transformer is quite often used for transferring power from one
circuit to another. Figure 4.14 shows the circuit of the same transformer as simulated
so far but with a voltage source connected to each winding.
For the simulation, let us assume the feeder impedances Zs1 and Zs2 to be equal
and with a nominal value of

rs1 = rs2 = 0.02 , Ls1 = Ls2 = 0.0001H

While choosing the voltage sources for the simulation, let us choose them to have
the same phase angle but introduce a very small difference in their RMS magnitudes
of 2V as follows:
√
vs1 = 240 2 cos(100π t)
√
vs2 = 242 2 cos(100π t)

Figure 4.15 shows the currents is1 and is2 flowing between the two voltage
sources. Since we have chosen the waveforms of the voltage sources to be cosines,
we do not see a dc offset and this makes interpreting the results a little easier.
In Fig. 4.15, the source voltage vs1 has been superimposed while being scaled
down by a factor of 15 to better understand the nature of currents is1 and is2 . By
comparing with the directions of currents is1 and is2 in Fig. 4.14, and knowing that
the magnitude of vs2 is greater than vs1 , it is clear that power is flowing from vs2 to
vs1 . This is due to the fact that the currents is1 and is2 are leading the voltage vs1 ,
which implies that the currents are flowing in the reverse direction. Furthermore,
it can be observed that the magnitude of is2 is greater than is1 as now winding 2
is the winding that is being primarily energized and therefore winding 2 draws the
magnetizing current from vs2 .
136 4 Modelling and Simulating Transformers

Fig. 4.15 Source and load currents with dual voltage sources

The above simulation with a voltage source connected to each winding can be
interpreted analytically in exactly the same way we did for the no load and the load
simulations. We can now write expressions for both windings using Kirchhoff’s
Voltage Law and Faraday’s Law as follows:

dφ
v1 = N1 + i1 r1 (4.45)
dt
dφ
v2 = N2 + i2 r2 (4.46)
dt
From the above expressions, it is very clear that the transformer core flux will be
determined by the voltage source that has the larger magnitude. Not only will the
magnitude of the core flux be determined by the voltage source that has the larger
magnitude, but also the direction of the core flux as was described in the previous
chapter [13, 14, 27, 28].
When a transformer is used to interface two different energized circuits in such a
manner, it is very similar to connecting voltage sources in series with the transformer
winding resistances and leakage inductances as interfacing impedances. In the above
case, since both windings had the same rated maximum voltages and therefore equal
number of turns, it was fairly simple to analyse the results and label one of the
voltage sources to be the “stronger” voltage source [13, 14, 27, 28]. However, the
most common use of a transformer is to be the interface between two systems at a
very different voltage as is required in the power system. This will be discussed in
the next section where we will consider transformers with a turns ratio other than 1.
4.6 Transformer Turns Ratio 137

In this section, we examined in detail how currents flow in the windings of

the transformer. In order to analyse simulation results, we used only the basic
laws of physics. The analysis helps to understand how the transformer operates
as a magnetic machine in comparison to merely solving the equivalent circuit. The
analysis has been simple due to the presence of only two windings in the transformer
and both windings having identical maximum voltage ratings. In the next section,
we will simulate and analyse step-up and step-down transformers.

4.6 Transformer Turns Ratio

In the previous sections, we had simulated a two winding transformer where both
windings had identical maximum voltage ratings. In this section, let us get started
with the practical transformer where the windings can have very different maximum
voltage ratings. As an example, let us consider transformers to have voltage ratings
of 240 V/11 kV or 240 V/24 V—this first type usually is a distribution transformer
that steps down the voltage from a distribution feeder and supplies an apartment
building or house while the second type is usually a control transformer, which one
can use to generate a power supply of 5 V, 9 V or 12 V [13, 14]. The analysis in this
section will be very similar to that carried out in the previous section and we will be
using the very same basic laws of physics.
To simulate the 240 V/11 kV transformer, we can use the same circuit as
Fig. 4.6 with winding 1 energized by a 240V RMS, 50Hz voltage source
while winding 2 is supplying a load. This simulation can be found in the
folder step_up_transformer within chapter4_transformers in the
following link in the simulation repository:
https://github.com/opensourceelectrical/simulating-magnetics-for-power-
electronics
In most cases, such step-up transformers are usually three-phase transformers,
which will be covered in the next chapter. However, for now, to simulate a step-up
transformer, we could choose any voltage ratings for winding 1 and winding 2.
Since this is a step-up transformer that could potentially connect to a high voltage
distribution feeder, let us choose a higher maximum power rating of 100 kVA. Since
the load comprising of Rl and Ll will be connected on winding 2 having a maximum
voltage rating of 11 kV, let us choose Rl = 1000 and Ll = 30H. As for the feeder
Zs connecting the source voltage to the winding 1 terminals, we can continue with
the parameters in the previous sections of Ls = 0.0001H and rs = 0.02.
To change the transformer magnetic model of the previous section, the only
change that needs to be made is with respect to the parameters defined at the head
of the file:
import math
d t = 1 . 0 e−6
VArated = 1 0 0 0 0 0 . 0
V1rated = 240.0
138 4 Modelling and Simulating Transformers

V2rated = 11000.0
f r a t e d = 50.0
o m e g a _ r a t e d = 2∗ math . p i ∗ f r a t e d

Once these parameters have been changed, the remaining computations of the
transformer equivalent circuit parameters will use these parameters, and therefore
no changes need to be made in those computations unless we wish to change our
estimates for any of the parameters.
With the above transformer specifications, the parameters of the transformer
equivalent circuit will be
L1 = 0 . 0 9 1 6 7 , L2 = 1 9 2 . 5 7 7
Lm1 = 0 . 0 9 1 6 3 , Lm2 = 1 9 2 . 5
M12 = 4 . 2
L l 1 = 3 6 . 6 7 e −6 , L l 2 = 0 . 0 7 7
r1 = 0.00576 , r2 = 12.1
Rc1 = 5 7 . 6 , Rc1 = 1210000

A quick look at the above parameters gives us an idea of how different the two
systems are that are being interfaced by the transformer. As an example, the winding
resistance of the 240 V winding 1 is a mere 576 m while the winding resistance of
the 11 kV winding 2 is 12.1 . This is equivalent to comparing the tail of a mouse
and an elephant!
Besides the parameters of the transformer equivalent circuit being drastically
different for the two windings, we must also remember that the transformer is
merely an energy conversion device and does not generate any energy. Therefore,
the power available at winding 1 will differ from the power available at winding 2 by
the losses of the transformer [13, 14, 27, 28]. The rated currents of the two windings
are
I 1 r a t e d = 416.67 , I 2 r a t e d = 9.09

The maximum rated current of winding 1 (the low voltage winding) is much larger
than the maximum current rating of winding 2 (the high voltage winding).
This might at first appear confusing as one would expect the high voltage system
to have a larger current rating. However, once again, it is important to remember that
the transformer is merely the interface between two systems at a different voltage
[13, 14, 27, 28]. If this transformer were supplying power to apartment buildings
from a 11 kV distribution feeder, there would be such a transformer for every
building in a street or an area. Therefore, each 240 V winding of the transformer
would be supplying the building itself while the 11 kV winding is merely one branch
of the distribution feeder. The distribution feeder may be supplying numerous
apartment buildings due to which the total current in the feeder could also be in the
range of hundreds of amperes. The total power capacity of the distribution feeder
could be in the range of several hundreds of kVAs or even a few MVAs (mega Volt-
Amperes). The transformer is merely supplying one branch from this distribution
feeder to an apartment building that we have chosen to have a maximum power
rating of 100 kVA.
4.6 Transformer Turns Ratio 139

The advantage of building the transformer simulation model in the manner

described in the previous section is that it results in a mathematical model that is
easily modifiable as the system changes. Once we change the maximum power
rating and maximum voltage ratings of the windings, the rest of the simulation
model—the matrix equation, the triangularization and the numerical integration—
can remain the same. In the next section, we will examine how we can simulate a
transformer with more than two windings by a simple modification. The resultant
transformer model is extremely flexible and can be used to simulate some fairly
complex transformers as will be demonstrated in the next chapter.
Figures 4.16 and 4.17 show the simulation results. The results show how
drastically different the quantities on the two different windings are. To be able to
better interpret the results, let us scale one of the quantities in each plot so that
the waveforms can be compared. Figure 4.18 shows the voltages with the load
voltage scaled down by a factor of 10 and Fig. 4.19 show the currents with the
source current scaled down by a factor of 10. Due to the difference in the windings
of the transformer, it is no longer a simple matter to determine which winding is
supplying the magnetizing current. However, we could undertake a similar exercise
as in the previous section to further analyse the results.
To begin with, once the maximum rated voltage of the windings has been
specified, we can establish the turns ratio of the transformer:

N2 v2rated 11000
= = = 45.83 (4.47)
N1 v1rated 240

Fig. 4.16 Step up transformer source and load voltages

140 4 Modelling and Simulating Transformers

Fig. 4.17 Step up transformer source and load currents

Fig. 4.18 Step up transformer source and load scaled voltages

The turns ratio of a transformer can be defined either as N 2 N1

N1 or N2 [13, 14, 27, 28].
Either is acceptable and in this case, we have chosen the former as it results in a
more convenient number. All that we must remember is that the ratio of the winding
turns is equal to the ratio of the voltage ratings of the windings.
As already described in the previous section, the significance of the turns ratio
of the transformer lies in the ratio of the induced emfs of the windings, which are
4.6 Transformer Turns Ratio 141

Fig. 4.19 Step up transformer source and load scaled currents

equal to the rate of change of flux linkages of the windings. Assuming a simple core
construction where the flux linking the windings is the same, the flux linkage of a
winding is equal to the product of the flux and the number of turns of the winding:

ψ1 = N1 φ (4.48)

ψ2 = N2 φ (4.49)

The induced emfs of the windings can be expressed as

dψ1 dφ
e1 = = N1 (4.50)
dt dt
dψ2 dφ
e2 = = N2 (4.51)
dt dt
Which implies that for a given core flux φ, the induced emfs will also follow the
transformer turns ratio. Since the winding voltages differ from the induced emfs
by the voltage drop across the winding resistance, the winding voltages are also
constrained by the turns ratio:

dφ
v1 = e1 + i1 r1 ≈ N1 (4.52)
dt
dφ
v2 = e2 + i2 r2 ≈ N2 (4.53)
dt
142 4 Modelling and Simulating Transformers

As in the previous section, these few equations alone can be used to understand
the transformer winding voltages. In this specific simulation, winding 1 is energized,
which leads to the core flux being determined as in the previous section:

1
φ= (v1 − i1 r1 ) dt (4.54)
N1

With this core flux linking winding 2, the emf induced in the winding is

dφ N2 N2
e2 = N2 = e1 = (v1 − i1 r1 ) (4.55)
dt N1 N1

The induced emf in winding 2 produces a terminal voltage:

N2
v2 ≈ e2 − il r2 = (v1 − i1 r1 ) − il r2 (4.56)
N1

It is important to note that we have used the load current il in the expression for
v2 rather than the internal winding 2 current i2 as the direction of current i2 is the
opposite of il (Fig. 4.5). With these equations, it is fairly clear how a voltage applied
to the terminals of winding 1 eventually produces a voltage across winding 2, which
has been multiplied by the turns ratio N 2
N1 . This is exactly the behaviour we observe
in Figs. 4.16 and 4.18 where the load voltage is a multiple of the source voltage and
both winding voltages have approximately the same phase angle. In this simulation,
only a passive load was connected to winding 2. However, we could think of several
variations to the simulation. First, we could connect a source voltage to winding 2
and a passive load to winding 1. Second, we could connect voltage sources to both
winding 1 and winding 2.
The first case is fairly simple. All we need to keep in mind is that the voltage
source connected to winding 2 must remain within the maximum rating of 11 kV.
The load connected to winding 1 must be such that the current drawn by the load
must be within the rated maximum current for the transformed voltage N 1
N2 v2 that
appears across the terminals of winding 1. We can progress through the equations in
reverse—starting from the terminal voltage expression for winding 2, the core flux
expression, the expression for induced emf in winding 1 and finally the expression
for terminal voltage in winding 1 [13, 14, 27, 28]. The reader is encouraged to try
out this simulation.
The second case is a bit more interesting. If there were voltage sources connected
to both windings, we could think of it as a case when a generator is being interfaced
to the high voltage transmission line that is already energized by other generators.
At the terminals of the two windings, we can write the following expressions:

dφ
v1 = N1 + i1 r1 (4.57)
dt
dφ
v2 = N2 + i2 r2 (4.58)
dt
4.6 Transformer Turns Ratio 143

We had examined a similar case in the previous section except that the trans-
former had the same number of turns in each winding. In this section, the
transformer windings have different number of turns. Therefore, to determine which
winding behaves as a source and which winding behaves as a load, we need to take
into account not only the voltage at the winding but also the number of turns of the
winding [13, 14, 27, 28]. As an example, let us suppose that Nv11 was larger than Nv22 .
This would result in the flux in the core flowing in such a manner that winding 1 was
magnetizing the core and winding 2 would be demagnetizing the core as winding 1
becomes the “stronger” winding. As a result, in the above equations, the current i1
would be positive while the current i2 will be negative. Power flows from winding
1 to winding 2. If a generator interfaced to the power system would need to inject
power into the transmission system, the voltage applied to the low voltage winding
would merely have to be such that the ratio of voltage and number of turns of the
winding makes the winding appear as a stronger winding.
A note of caution must be exercised at this point. Solving the above voltage
equations directly is being done in the simulation as we express the voltage
equations as matrix equations and triangularize them. However, due to the derivative
of the flux, we cannot equate the above equations to two voltage sources connected
together in series with an impedance. The analogy to voltage sources interconnected
by an impedance is relevant when we consider a magnetic circuit where the ampere
turns of coils/windings and the reluctance of the core on which they are wound can
be solved to determine the core flux [13, 14, 27, 28]. However, a direct analogy to
the voltages applied to the coils/windings will not be correct. We can, however, use
the equations to estimate the nature of power flow and understand the basic physical
operation of the transformer.
Now that we have had a discussion on the winding voltages of the transformer,
we can now direct our attention to the winding currents of the transformer. As with
the case of the winding voltages, the winding currents can also be understood in the
same manner as we did in the previous section. To begin with, let us start with the
magnetizing current. Every transformer will need to draw a magnetizing current in
order to maintain the flux in the core. In order to simulate the condition where the
winding currents are approximately equal to the magnetizing current, we can use
the same strategy as in the previous section—make the load resistance very large
so as to make the load current negligible. If winding 1 is energized and the load is
connected to winding 2, we can increase the load resistance from Rl = 1000 to
Rl = 1, 000, 000 .
Figure 4.20 shows the no-load source and load currents of the transformer. The
load current il is negligible as we have increased the load resistance to 1 M. The
source current is supplies the loss in resistors Rc1 and Rc2 and the magnetizing
current im of the transformer. The no-load current has a dc offset as we have chosen
the source voltage to have a sine waveform. The source current is has a peak of
around 30A, which at first might seem unusual. Taking into account the dc offset
in the source current, the ac peak of the source current is around 15 A, which
corresponds to an RMS value of around 10.6 A. However, the rated current in the
primary winding has an RMS value of 416.67 A, which implies that the no-load
144 4 Modelling and Simulating Transformers

Fig. 4.20 No-load current of the transformer

current is merely 2.5% of the rated current. Since the self-inductance of transformer
winding is so designed that the current drawn will be 2% of the rated current, these
numbers are only expected. The reader is encouraged to repeat the simulation while
energizing winding 2 with a 11 kV RMS voltage source and connecting a large
resistance in winding 1. The no-load current should also be around 2–2.5% of the
rated current (9.09 A) of winding 2.
If we restore the load resistor connected to winding 2 back to its normal value of
1000 , the currents will be as shown in Figs. 4.17 and 4.19. The current il drawn
by the load will result in a winding current i2 . From Figs. 4.6 and 4.5, this winding
current will have a demagnetizing effect on the core flux. This has been discussed
in detail in the previous chapter while describing dot polarities and the magnetic
coupling. However, as in the previous section, the core flux is primarily determined
by the voltage v1 that energizes winding 1 of the transformer:
 
1 1
φ= (v1 − i1 r1 ) dt ≈ v1 dt (4.59)
N1 N1

The above approximation implies that the flux will not change significantly with
the winding current i1 but will remain roughly unaffected by current flowing in the
winding. Just like in the previous section, the core flux can be expressed as

N1 is − N2 il
φ= (4.60)
R
4.6 Transformer Turns Ratio 145

where R is the reluctance of the core. In the previous no-load case, with the source
current being approximately equal to the magnetizing current:

N1 im
φ≈ (4.61)
R
If the load current despite having a demagnetizing effect on the core does not
change the core flux, this would imply that the demagnetizing effect of the load
current is cancelled by another component in the source current:

N1 im + N1 il − N2 il
φ= (4.62)
R

In the above expression, the current il is the load current il referred to winding
1. As discussed in the previous section, a referred current is merely the current in a
winding being referred to another winding such that the ampere turns of the referred
current is the same as the original current [13, 14, 27, 28]. Therefore,

N1 il = N2 il (4.63)

In the previous section, we had arrived at this result. However, since both
windings had identical number of turns, in the previous section, the referred load

current il was the same as il . In this simulation case, with a turns ratio of N
N1 =
2


45.83, the referred load current il will be 45.83 times the load current il . As a result,
even though the load current is merely around 9A due to the values of Rl = 1000
and Ll = 3H, the referred load current is approximately 450 A. From Figs. 4.17 and
4.19, the source current can now be explained.
To complete the discussion on transformers with a turns ratio, let us also
simulate a step-down transformer where the maximum rated voltage of winding
2 is lower than that of winding 1. Though such transformers can be used for a wide
number of applications, let us consider a step-down transformer used in control
applications quite often to produce power supplies for control circuits used in
power converters [10, 26]. Let us suppose that winding 1 of the transformer is
energized by an ac supply of 240 V, 50 Hz. As an example, let us suppose that
we wish to use this transformer to produce a 5 V power supply. In such a case,
winding 2 of the transformer can have a maximum rating of 12 V. To this winding,
we could connect a simple diode rectifier that converts the ac voltage at the winding
terminals to an uncontrolled dc to be further regulated by any dc–dc converter
such as a buck converter, flyback converter or a forward converter. At this stage,
we will not simulate the power electronics but will simulate just the transformer.
This simulation can be found in the folder step_down_transformer within
chapter4_transformers in the following link in the simulation repository:
https://github.com/opensourceelectrical/simulating-magnetics-for-power-
electronics
146 4 Modelling and Simulating Transformers

As with the simulation of the step-up transformer, let us define the parameters of
the step-down transformer. Since this step-down transformer will be used for control
purposes, we do not need a large power rating. A maximum power rating of 100 VA
should be sufficient for most control circuits used in power electronic applications.
We can use the circuit topology of Fig. 4.6. The feeder impedance that connects the
source voltage to winding 1 can remain at the same value as the previous simulation
at Rs = 0.02, Ls = 0.0001H. The load connected to winding 2 needs to be such
a value that power drawn by the load will be less than 100 VA of a terminal voltage
of 12 V RMS. We can choose the load impedance to be Rl = 1, Ll = 0.003H.
As with the simulation of the step-up transformer, the only change that needs to
be made in the transformer model are the parameters defined at the head of the file:
import math
d t = 1 . 0 e−6
VArated = 1 0 0 . 0
V1rated = 240.0
V2rated = 12.0
f r a t e d = 50.0
o m e g a _ r a t e d = 2∗ math . p i ∗ f r a t e d

The calculation of the equivalent circuit parameters, the simplification of the matrix
equations and the numerical integration will remain the same.
Figures 4.21 and 4.22 show the simulation results with the step-down trans-
former. As can be seen from Fig. 4.21, the voltage v2 available at the winding 2
is a fraction of the voltage v1 available at winding 1. On the other hand, the current
drawn by the load il is referred to winding 1 and can be calculated as

 N2
il = il (4.64)
N1

Fig. 4.21 Step down transformer source and load voltages

4.7 Multi-winding Transformers 147

Fig. 4.22 Step down transformer source and load currents

As the transformer is a step-down transformer, N 2

N1 < 1, and therefore, the source
current is is a mere fraction of the load current il . The analysis carried out for the
step-up transformer can be repeated for the step-down transformer as well. This
exercise will be left to the reader.
In this section, we simulated transformers where the windings did not have the
same maximum voltage rating and therefore had turns ratios different from 1. In
a vast number of applications, transformers are used to either step-up the voltage
from a generator or step-down the voltage from a feeder to a load. Therefore,
understanding the operation of step-up and step-down transformers from the basic
laws of physics will make this very common application of the transformer clearer.
Until now, the transformers simulated always had two windings. However, in a
number of cases, transformers can have more than two windings. In such cases,
a single supply can be used to feed loads at very different voltages. This will be
discussed in the next section.

4.7 Multi-winding Transformers

In the past few sections, we had simulated a number of transformers and analysed
them in detail. The purpose of these simulations was to show how the basic laws of
physics can be used to both simulate transformers and understand their behaviour.
In this section, we will simulate a transformer with more than two windings. Such
multi-winding transformers are commonly used when designing power supplies for
control circuits as most control circuits need multiple voltages such as 5 V, 9 V, 12 V,
etc. [13, 14, 27, 28]. An example of a flyback converter will be described in the next
148 4 Modelling and Simulating Transformers

Fig. 4.23 Transformer with i2

four windings

v2
i1
i3

v1 v3

v4

i4

chapter where a single source can be used to produce a number of isolated power
supplies. Besides power supplies, simulating multi-winding transformers is essential
for simulating three-phase transformers, which form the bulk of transformers used
in the power system [13, 14, 27, 28].
The simulation of a multi-winding transformer is a mere extension of the
simulation of a two winding transformer. Figure 4.23 shows a transformer with four
windings. It is important to note that the arrangement of the windings in Fig. 4.23
is not significant. We could have arranged the four windings in any manner—all in
a vertical line, two on the left and two on the right, etc. Physically, these are coils
wound on a core and the exact construction details might be very different. This is
due to the fact that the main objective in a transformer is to achieve maximum power
transfer between windings. The schematic of Fig. 4.23 is merely for the convenience
of analysing a transformer in a circuit.
We could use Kirchhoff’s Voltage Law for each winding:

dψ1
v 1 − i1 r 1 − =0 (4.65)
dt
dψ2
v2 − i2 r2 − =0 (4.66)
dt
dψ3
v3 − i3 r3 − =0 (4.67)
dt
dψ4
v4 − i4 r4 − =0 (4.68)
dt
4.7 Multi-winding Transformers 149

From Fig. 4.23, the fourth winding has a dot on the lower terminal in contrast with
the other windings. However, the KVL equation can still be written in the same
manner due to the directions of current i4 and voltage v4 .
The flux linkages of the windings can be expressed in terms of the self-
inductances of the windings and the mutual inductances between windings as
follows:

ψ1 = L1 i1 + M12 i2 + M13 i3 + M14 i4 (4.69)

ψ2 = M12 i1 + L2 i2 + M23 i3 + M24 i4 (4.70)

ψ3 = M13 i1 + M23 i2 + L3 i3 + M34 i4 (4.71)

ψ4 = M14 i1 + M24 i2 + M34 i3 + L4 i4 (4.72)

In the previous simulation, we had a single mutual inductance for the windings,
which was defined as the flux linkage produced in a winding due to the current
flowing in another winding. Therefore, we can define mutual inductances for every
pair of windings x and y and we can continue to use the expression:

Mxy = k Lx Ly (4.73)

In the above expression k is the coupling factor between the two windings. This
coupling factor does not take into consideration the leakage of flux as we will be
computing a separate leakage inductance that is a part of the self-inductance. The
coupling factor is merely an indication of the fraction of the flux produced by a
winding that links with another winding. In three-phase transformers, which will
be discussed in the next chapter, windings may be wound on separate limbs of
the transformer, due to which only a fraction of the flux produced by the winding
may link with another winding. At this point, we can assume a simple rectangular
core construction where all the flux produced by a winding links completely with
the other windings, due to which the coupling factor k is equal to 1 for all mutual
inductances.
We could rewrite the voltage equations as a matrix equation as follows:
⎡ ⎤ ⎡ ⎤⎡ ⎤ ⎡ ⎤ ⎡ ⎤
v1 r1 0 0 0 i1 L1 M12 M13 M14 i1
⎢v2 ⎥ ⎢ 0 r2 0 0 ⎥ ⎢i2 ⎥ ⎢M12 M24 ⎥ ⎢i 2 ⎥
⎢ ⎥=⎢ ⎥⎢ ⎥ ⎢ L2 M23 ⎥ d ⎢ ⎥
⎣v3 ⎦ ⎣ 0 0 r3 0 ⎦ ⎣i3 ⎦ + ⎣M13 M23 L3 M34 ⎦ dt ⎣i 3 ⎦ (4.74)
v4 0 0 0 r4 i4 M14 M24 M34 L4 i4

If we were to compare this matrix equation with the matrix equation for the two
winding transformer, it is fairly clear that the model is merely an extension of the
two winding transformer with larger matrices. Moreover, the inductance matrix has
a fairly well-defined construction—symmetric with the diagonal elements being
150 4 Modelling and Simulating Transformers

V V

g(V )

V V

Fig. 4.24 Simulation model of the four winding transformer

the self-inductances and the off-diagonal elements being the mutual inductances.
With this matrix equation, we can simulate fairly complex transformers with several
windings without much alterations to the mathematical model and its solution.
In the mathematical model, we will be solving the above matrix equation and
the resultant currents would then be represented by the usual combination of
controllable voltage sources in series with variable resistors. Figure 4.24 shows
how each transformer winding is modelled as a block comprising of a controllable
voltage source in series with a variable resistance and the winding terminal voltage
measured by a Voltmeter. The model receives as inputs the winding terminal
voltages v1 , v2 , v3 , v4 and produces signals for the windings currents i1 , i2 , i3 ,
i4 . These winding currents are achieved by adjusting the voltage outputs of the
controllable voltage sources with respect to the measured winding terminal voltages
and the series resistances.
Let us describe the details in simulating a multi-winding transformer by
performing a simulation of a four winding transformer as shown in Fig. 4.23. This
simulation can be found in the folder multiwinding_transformer within
chapter4_transformers in the following link in the simulation repository:
https://github.com/opensourceelectrical/simulating-magnetics-for-power-
electronics
4.7 Multi-winding Transformers 151

Let us consider the following maximum rated voltages—240 V RMS for winding 1,
240 V RMS for winding 2, 480 V RMS for winding 3 and 120 V RMS for winding
4. Therefore, we have a 1:1 turns ratio between winding 1 and 2, 1:2 turns ratio
between winding 1 and 3, and 2:1 turns ratio between winding 1 and winding 4.
Let us energize winding 1 with a 240 V, 50 Hz ac single-phase supply and connect
loads to all the other windings. Therefore, the objective of such a transformer
would be to supply different loads at different voltage levels. Since we have three
loads altogether, let us choose the maximum power rating of this transformer to be
30 kVA—three times the two winding transformer simulated before.
We can define these specifications of the transformer at the head of the Python
file as follows:
import math
d t = 1 . 0 e−6
VArated = 3 0 0 0 0 . 0
V1rated = 240.0
V2rated = 240.0
V3rated = 480.0
V4rated = 120.0
f r a t e d = 50.0
o m e g a _ r a t e d = 2∗ math . p i ∗ f r a t e d

We have merely added the ratings for the third and fourth winding. We can
proceed to calculate the maximum current ratings and the rated impedance for all
the windings:
I1rated = VArated / V1rated
I2rated = VArated / V2rated
I3rated = VArated / V3rated
I4rated = VArated / V4rated
Z1rated = V1rated / I1rated
Z2rated = V2rated / I2rated
Z3rated = V3rated / I3rated
Z4rated = V4rated / I4rated

From the above computations, it needs to be emphasized that the maximum current
rating of each winding is calculated assuming that the winding is going to be
transferring the maximum rated power. In our specific application, that is not the
case, as only winding 1 is supplying the maximum rated power while all other
windings are supplying a load that is only a fraction of the total power rating of
the transformer. However, when a transformer is designed and simulated, we do not
usually differentiate between windings as source and load windings. Any winding
could be energized and supplying loads connected to the other windings. Therefore,
the maximum rated current for each winding is calculated assuming that the total
maximum rated power of the transformer will be supplied by the winding.
With these base computations, the parameters of the transformer equivalent
circuit can be computed as follows:
L1 = 5 0 . 0 ∗ Z 1 r a t e d / o m e g a _ r a t e d
L2 = 5 0 . 0 ∗ Z 2 r a t e d / o m e g a _ r a t e d
L3 = 5 0 . 0 ∗ Z 3 r a t e d / o m e g a _ r a t e d
152 4 Modelling and Simulating Transformers

L4 = 5 0 . 0 ∗ Z 4 r a t e d / o m e g a _ r a t e d
Ll1 = 0.02 ∗ Z 1 r a t e d / omega_rated
Ll2 = 0.02 ∗ Z 2 r a t e d / omega_rated
Ll3 = 0.02 ∗ Z 3 r a t e d / omega_rated
Ll4 = 0.02 ∗ Z 4 r a t e d / omega_rated
Lm1 = L1 − L l 1
Lm2 = L2 − L l 2
Lm3 = L3 − L l 3
Lm4 = L4 − L l 4
M12 = math . s q r t ( Lm1 ∗ Lm2 )
M13 = math . s q r t ( Lm1 ∗ Lm3 )
M14 = math . s q r t ( Lm1 ∗ Lm4 )
M23 = math . s q r t ( Lm2 ∗ Lm3 )
M24 = math . s q r t ( Lm2 ∗ Lm4 )
M34 = math . s q r t ( Lm3 ∗ Lm4 )
r1 = 0.01 ∗ Z1rated
r2 = 0.01 ∗ Z2rated
r3 = 0.01 ∗ Z3rated
r4 = 0.01 ∗ Z4rated
Rc1 = V 1 r a t e d ∗ V 1 r a t e d / ( 0 . 0 1 ∗ VArated )
Rc2 = V 2 r a t e d ∗ V 2 r a t e d / ( 0 . 0 1 ∗ VArated )
Rc3 = V 3 r a t e d ∗ V 3 r a t e d / ( 0 . 0 1 ∗ VArated )
Rc4 = V 4 r a t e d ∗ V 4 r a t e d / ( 0 . 0 1 ∗ VArated )
res_output1 = 100.0 ∗ Z1rated
res_output2 = 100.0 ∗ Z2rated
res_output3 = 100.0 ∗ Z3rated
res_output4 = 100.0 ∗ Z4rated

With the parameters of the transformer equivalent circuit, the matrices in the
matrix equation can be written as follows:
winding_currents = [0.0 , 0.0 , 0.0 , 0.0]
L = [
[ L1 , M12 , M13 , M14 ] ,
[ M12 , L2 , M23 , M24 ] ,
[ M13 , M23 , L3 , M34 ] ,
[ M14 , M24 , M34 , L4 ]
]
R = [
[ r1 , 0 . 0 , 0 . 0 , 0 . 0 ] ,
[ 0 . 0 , r2 , 0 . 0 , 0 . 0 ] ,
[ 0 . 0 , 0 . 0 , r3 , 0 . 0 ] ,
[ 0 . 0 , 0.0 , 0.0 , r4 ]
]
V = [ v1 , v2 , v3 , v4 ]

After the definition of the matrices of the equation, the code for simplification
and numerical integration are exactly the same as before. After the winding currents
have been updated, the output voltages can be computed as
vout1 = v1 − winding_currents [0]∗ res_output1
vout2 = v2 − winding_currents [1]∗ res_output2
vout3 = v3 − winding_currents [2]∗ res_output3
vout4 = v4 − winding_currents [3]∗ res_output4
4.7 Multi-winding Transformers 153

Fig. 4.25 Transformer winding voltages

Fig. 4.26 Transformer winding currents

Figures 4.25 and 4.26 show the simulation results of the four winding trans-
former. The voltages v1 and v2 at windings 1 and 2 are approximately the same
as both windings have the same turns ratio. The voltage v3 at winding 3 has a
magnitude that is twice that of v1 as the turns ratio between winding 3 and winding
1 is 2:1. The voltage v4 at winding 4 has a magnitude that is half of that of v1 as
the turns ratio between winding 4 and 1 is 1:2. Moreover, the voltage v4 is phase
154 4 Modelling and Simulating Transformers

shifted by 180◦ with respect to v1 due to the dot polarity of winding 4 being the
opposite of the other windings. The currents in Fig. 4.26 can also be explained with
respect to the turns ratios between the transformer windings. The current is supplied
  
by the load will be the sum of the transferred currents il1 , il2 , il3 drawn by the loads
connected in windings 2, 3 and 4.
Before we complete this section and also the chapter, let us optimize the
mathematical model of the transformer. Though the code that we have been using is
functional, it is inefficient as we are performing several computations repeatedly in
every simulation instant when they need to be performed only once at the beginning
of the simulation. The transformer model is largely decided by construction as the
equivalent circuit of a transformer will not change significantly. As the transformer
ages, the parameters will change since component parameters do change with age.
However, these changes will not be very significant and therefore, assuming the
transformer equivalent circuit to remain the same is not very unreasonable and will
simplify the model to a great extent. This simulation can be found in the folder
simplified_transformer_model within chapter4_transformers
in the following link in the simulation repository:
https://github.com/opensourceelectrical/simulating-magnetics-for-power-
electronics
We have been simplifying and integrating the matrix equation of (4.74). Let us
make a modification to the equation as follows:
⎡ ⎤⎡ ⎤ ⎡ ⎤⎡ ⎤ ⎡ ⎤ ⎡ ⎤
1 0 0 0 v1 r1 0 0 0 i1 L1 M12 M13 M14 i1
⎢0 0⎥ ⎢v2 ⎥ ⎢ 0 r2 0 0 ⎥ ⎢i2 ⎥ ⎢M12 M24 ⎥ ⎢i 2 ⎥
⎢ 1 0 ⎥⎢ ⎥ = ⎢ ⎥⎢ ⎥ + ⎢ L2 M23 ⎥ d ⎢ ⎥
⎣0 0 1 0⎦ ⎣v3 ⎦ ⎣ 0 0 r3 0 ⎦ ⎣i3 ⎦ ⎣M13 M23 L3 M34 ⎦ dt ⎣i 3 ⎦
0 0 0 1 v4 0 0 0 r4 i4 M14 M24 M34 L4 i4
(4.75)

We have multiplied the vector of winding voltages by an identity matrix and there-
fore, the equation remains unchanged. However, by doing so, the row operations in
the process of triangularization will now be performed on the identity matrix rather
than the vector of winding voltages:
B = [
[1.0 , 0.0 , 0.0 , 0.0] ,
[0.0 , 1.0 , 0.0 , 0.0] ,
[0.0 , 0.0 , 1.0 , 0.0] ,
[0.0 , 0.0 , 0.0 , 1.0]
]
for count1 in range ( len (L ) ) :
i f not L[ count1 ] [ count1 ] :
f o r c o u n t 2 i n r a n g e ( c o u n t 1 +1 , l e n ( L ) ) :
i f L[ count2 ] [ count1 ] :
L[ count1 ] , L[ count2 ] = L[ count2 ] , L[ count1 ]
R[ c o u n t 1 ] , R[ c o u n t 2 ] = R [ c o u n t 2 ] , R[ c o u n t 1 ]
B[ c o u n t 1 ] , B[ c o u n t 2 ] = B [ c o u n t 2 ] , B[ c o u n t 1 ]
break
4.7 Multi-winding Transformers 155

i f L[ count1 ] [ count1 ] :
f o r c o u n t 2 i n r a n g e ( c o u n t 1 +1 , l e n ( L ) ) :
comm_factor = L[ count2 ] [ count1 ] / L[ count1 ] [ count1 ]
for count3 in range ( len (L[ count1 ] ) ) :
L [ c o u n t 2 ] [ c o u n t 3 ] −= L [ c o u n t 1 ] [ c o u n t 3 ] ∗ c o m m _ f a c t o r
R[ c o u n t 2 ] [ c o u n t 3 ] −= R [ c o u n t 1 ] [ c o u n t 3 ] ∗ c o m m _ f a c t o r
B[ c o u n t 2 ] [ c o u n t 3 ] −= B [ c o u n t 1 ] [ c o u n t 3 ] ∗ c o m m _ f a c t o r

The matrices L, R, B are constants as we are assuming the parameters of the

transformer equivalent circuit to be constants that do not change after design.
Therefore, once the matrix equation has been simplified, the matrices in the
simplified equation will also remain constant. Therefore, the simplification needs to
be performed only once in the simulation, thereby reducing the computation burden
significantly. At each simulation instant, the only change that needs to be made is
that we must compute the initial value during integration, rather than initializing it
to the voltage:
# k [ k _ c o u n t ] = V[ c o u n t 1 ]
k [ k_count ] = 0.0
f o r c o u n t 2 i n r a n g e ( l e n (B ) ) :
k [ k _ c o u n t ] += B[ c o u n t 1 ] [ c o u n t 2 ] ∗V[ c o u n t 2 ]
.....
.....
# w i n d i n g _ c u r r e n t s [ c o u n t 1 ] = V[ c o u n t 1 ]
winding_currents [ count1 ] = 0.0
f o r c o u n t 2 i n r a n g e ( l e n (B ) ) :
w i n d i n g _ c u r r e n t s [ c o u n t 1 ] += B[ c o u n t 1 ] [ c o u n t 2 ] ∗V[ c o u n t 2 ]

With these changes made to ensure that the solution is unaffected, we need to
ensure that the simplification of the matrices occurs only once in the simulation.
For this, we can define a static variable init_simulation and provide an initial
value of 1, which signifies that initially the parameters need to be computed. We can
now perform all initial computations within this conditional check:
i f i n i t _ s i m u l a t i o n > 0:
I 1 r a t e d = VArated / V 1 r a t e d
I 2 r a t e d = VArated / V 2 r a t e d
I 3 r a t e d = VArated / V 3 r a t e d
I 4 r a t e d = VArated / V 4 r a t e d
Z1rated = V1rated / I 1 r a t e d
Z2rated = V2rated / I 2 r a t e d
Z3rated = V3rated / I 3 r a t e d
Z4rated = V4rated / I 4 r a t e d
L1 = 5 0 . 0 ∗ Z 1 r a t e d / o m e g a _ r a t e d
L2 = 5 0 . 0 ∗ Z 2 r a t e d / o m e g a _ r a t e d
L3 = 5 0 . 0 ∗ Z 3 r a t e d / o m e g a _ r a t e d
L4 = 5 0 . 0 ∗ Z 4 r a t e d / o m e g a _ r a t e d
Ll1 = 0.02 ∗ Z 1 r a t e d / omega_rated
Ll2 = 0.02 ∗ Z 2 r a t e d / omega_rated
Ll3 = 0.02 ∗ Z 3 r a t e d / omega_rated
Ll4 = 0.02 ∗ Z 4 r a t e d / omega_rated
Lm1 = L1 − L l 1
Lm2 = L2 − L l 2
156 4 Modelling and Simulating Transformers

Lm3 = L3 − L l 3
Lm4 = L4 − L l 4
M12 = math . s q r t ( Lm1 ∗ Lm2 )
M13 = math . s q r t ( Lm1 ∗ Lm3 )
M14 = math . s q r t ( Lm1 ∗ Lm4 )
M23 = math . s q r t ( Lm2 ∗ Lm3 )
M24 = math . s q r t ( Lm2 ∗ Lm4 )
M34 = math . s q r t ( Lm3 ∗ Lm4 )
r1 = 0.01 ∗ Z1rated
r2 = 0.01 ∗ Z2rated
r3 = 0.01 ∗ Z3rated
r4 = 0.01 ∗ Z4rated
Rc1 = V 1 r a t e d ∗ V 1 r a t e d / ( 0 . 0 1 ∗ VArated )
Rc2 = V 2 r a t e d ∗ V 2 r a t e d / ( 0 . 0 1 ∗ VArated )
Rc3 = V 3 r a t e d ∗ V 3 r a t e d / ( 0 . 0 1 ∗ VArated )
Rc4 = V 4 r a t e d ∗ V 4 r a t e d / ( 0 . 0 1 ∗ VArated )
res_output1 = 100.0 ∗ Z1rated
res_output2 = 100.0 ∗ Z2rated
res_output3 = 100.0 ∗ Z3rated
res_output4 = 100.0 ∗ Z4rated
winding_currents = [0.0 , 0.0 , 0.0 , 0.0]
L = [
[ L1 , M12 , M13 , M14 ] ,
[ M12 , L2 , M23 , M24 ] ,
[ M13 , M23 , L3 , M34 ] ,
[ M14 , M24 , M34 , L4 ]
]
R = [
[ r1 , 0 . 0 , 0 . 0 , 0 . 0 ] ,
[ 0 . 0 , r2 , 0 . 0 , 0 . 0 ] ,
[ 0 . 0 , 0 . 0 , r3 , 0 . 0 ] ,
[ 0 . 0 , 0.0 , 0.0 , r4 ]
]
B = [
[1.0 , 0.0 , 0.0 , 0.0] ,
[0.0 , 1.0 , 0.0 , 0.0] ,
[0.0 , 0.0 , 1.0 , 0.0] ,
[0.0 , 0.0 , 0.0 , 1.0]
]
for count1 in range ( len (L ) ) :
i f not L[ count1 ] [ count1 ] :
f o r c o u n t 2 i n r a n g e ( c o u n t 1 +1 , l e n ( L ) ) :
i f L[ count2 ] [ count1 ] :
L[ count1 ] , L[ count2 ] = L[ count2 ] , L[ count1 ]
R [ c o u n t 1 ] , R [ c o u n t 2 ] = R[ c o u n t 2 ] , R[ c o u n t 1 ]
B [ c o u n t 1 ] , B [ c o u n t 2 ] = B[ c o u n t 2 ] , B[ c o u n t 1 ]
break
i f L[ count1 ] [ count1 ] :
f o r c o u n t 2 i n r a n g e ( c o u n t 1 +1 , l e n ( L ) ) :
comm_factor = L[ count2 ] [ count1 ] / L[ count1 ] [ count1 ]
for count3 in range ( len (L[ count1 ] ) ) :
L [ c o u n t 2 ] [ c o u n t 3 ] −= \
L[ count1 ] [ count3 ]∗ comm_factor
R [ c o u n t 2 ] [ c o u n t 3 ] −= \
4.8 Conclusions 157

R[ c o u n t 1 ] [ c o u n t 3 ] ∗ c o m m _ f a c t o r
B [ c o u n t 2 ] [ c o u n t 3 ] −= \
B[ c o u n t 1 ] [ c o u n t 3 ] ∗ c o m m _ f a c t o r

i n i t _ s i m u l a t i o n = −1

The computation of the parameters of the transformer equivalent circuit will

now be performed only once at the very first simulation instant at t = 0
when init_simulation is equal to 1. After computing the parameters and
simplifying the matrices, we are setting the value of init_simulation to -1
to ensure that the computations are not performed subsequently. With this change,
however, many of the variables will not exist after the first simulation instant. This is
due to fact that in Python, a variable is created when it is assigned a value for the very
first time. Therefore, as an example, the variable I1rated will be created in the
very first simulation instant and assigned a value equal to VArated / V1rated.
In subsequent iterations, when the conditional block will not be executed, I1rated
will no longer be created and therefore will not appear anywhere in the control file.
This is a problem for certain variables such as the matrices L, R and B as these are
used in the numerical integration and therefore, must exist and retain the computed
value for the entire simulation. In order to ensure that they retain their computed
values, they can be defined as static variables and assigned dummy initial values of
0. In Python, when a variable is defined a value but redefined as another value, the
old object is destroyed and a new object is created. Therefore, defining L, R and B as
static variables with initial values of 0 but later computing them as lists will imply
that for the rest of the simulation, they will retain their values as lists.
In this section, we have shown how the mathematical model can be expanded
fairly easily to simulate transformers with several windings. Most computations
related to the parameters of the equivalent circuit are merely replicates of the
basic computations for the base two winding transformer. The final matrices in
the matrix equation have a well-defined structure—matrix R is a diagonal matrix
with the diagonal elements being the winding resistances and B is an identity
matrix. Matrix L is a symmetric matrix with the diagonal elements being the
self-inductances of the windings and the off-diagonal elements being the mutual
inductances between the windings. The concluding simplification to the model by
minimizing the computations by performing initialization only once at the beginning
of the simulation helps to improve the speed of the simulation. This simulation
model of the transformer will be used in the next chapter to simulate a few chosen
advanced applications of transformers.

4.8 Conclusions

In this chapter, we have extended the simulation models presented in the previous
chapters to a simulation model of a transformer. In Chap. 2, we converted the basic
laws of physics into a very basic simulation model of the inductor, which is one of
158 4 Modelling and Simulating Transformers

the simplest magnetic components in electrical engineering. In Chap. 3, we extended

the basic magnetic model of an inductor to simulate magnetically coupled inductors.
A transformer, in its most fundamental form, is just a set of magnetically coupled
inductors. Therefore, the underlying simulation model of magnetically coupled
coils presented in the previous chapter is applicable for transformers as well. The
difference lies in the transformer having a different set of specifications since it is an
electrical machine quite often designed, fabricated and manufactured for a particular
application.
The chapter begins with an overview of transformers, their importance in modern
electrical engineering and some of the major applications of transformers. The
chapter describes how the transformer as a machine can be translated into an
equivalent circuit. The parameters of this equivalent circuit can either be determined
through laboratory experiments such as the open circuit test and the short-circuit
test, or can be estimated as shown in the chapter. Though this might result
in approximate values of the parameters of the transformer equivalent circuit,
reasonable estimates can produce a fairly usable simulation model of a transformer.
Furthermore, the process of estimating the transformer equivalent circuit has an
added advantage—we can potentially change these estimates to determine how a
transformer behaves as it might change due to ageing.
Once the equivalent circuit of the transformer is determined, it is clear that
this equivalent circuit can be thought of as coupled inductors. Therefore, the
parameters of the equivalent circuit can be transferred to the simulation model of
coupled inductors already derived in the previous chapter. Simulations of several
transformers are presented such that in different cases, we have not only translated
the specifications of the transformer to the simulation model, but we have also
examined the impact of the transformer specifications on the nature of parameters.
This is particularly the case of step-up and step-down transformers, in which it is
clear that the electrical systems on the different windings of the transformer are
totally different. This provides us with an insight into the role of transformers in the
modern power system, where transformers interface the generation, transmission
and distribution systems.
Using simulations, we have delved deep into the operation of transformers. We
use simulations to examine the no-load operation of the transformer where we
differentiate between the core loss components of current and the magnetizing
current of the transformer. We use theory and simulations to describe why a
transformer needs to draw a magnetizing current at all times. We also examined how
the core flux of the transformer is determined primarily by the energizing voltage
and to a much smaller extent by the winding currents. Furthermore, we also examine
the impact of the winding currents on the core flux and how winding currents can be
transformed using the turns ratio of the transformer. We emphasize on the fact that
besides the magnetizing current of the transformer, any other current flowing in any
other winding will result in a transformed current flowing in the other winding due
to the necessity for the MMFs of the windings to be balanced.
To emphasize on how the simulation models used are flexible and scalable, we
have simulated a number of different type of transformers. We have shown how
4.8 Conclusions 159

simulating step-up and step-down transformers needs nothing more than modified
transformer specifications in the simulation model with all other computations
remaining unaffected. Furthermore, we have also simulated multi-winding trans-
formers and have shown how a transformer with several windings requires merely
larger system matrices with most of these matrices being intuitive and having a
well-defined structure. In the previous chapter, we had already expressed the final
differential equation of magnetically coupled systems as a matrix equation that
allows us to implement a repeatable algorithm to simplify the matrices and integrate
them numerically.
Though we have simulated a number of different transformers in this chapter, in
the next chapter we will use these simulation models to simulate some transformers
that occur often in electrical engineering. A few examples of these are three-phase
transformers that form the bulk of the modern power system. In addition, we will
also examine the application of high frequency transformers in a power electronics
application. We will consider a power supply consisting of a flyback converter
and will show through simulations, how even taking into consideration the highly
non-linear behaviour of power electronic converters, the underlying principle of
transformer operation remains the same.
Chapter 5
Applications of Transformers

5.1 Introduction

In the previous chapter, we had presented simulation models for transformers and
analysed their operation through simulation results. However, the transformers stud-
ied in detail were two winding transformers that are typically used for single-phase
applications. The major application of transformers is as three-phase transformers in
the power system [13, 14, 21]. Though the simulation models of the previous chapter
can be extended to simulate three-phase transformers, there are a few aspects of
three-phase transformers that are not immediately obvious. In this chapter, we will
cover the simulation of three-phase transformers to make the material in this book
more complete.
The simulation of three-phase transformers is an extension of the simulation of
multi-winding transformers already described in the previous chapter. However, to
simulate any three-phase machine, whether a transformer or a generator, a certain
detail of construction needs to be known or assumed [27–31]. In this chapter, we will
introduce the concept of three-phase systems and how the windings of a three-phase
transformer can be connected. Moreover, we will also examine how the windings
of the transformer can be wound on the core and how this will affect the simulation
model of the transformer.
We will simulate different three-phase transformers having different connection
types to cover some of the most popularly used transformers in the power system
[27–31]. In all the simulations, the underlying simulation model is the same as the
basic simulation model of the transformer presented in the previous chapter, except
that the specifications of the transformer will change. We will simulate the star–
star and delta–star transformers. We will also discuss how distribution systems that
need a neutral wire to support single-phase appliances can be fed by a delta–star
transformer.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 161
S. V. Iyer, Modeling and Python Simulation of Magnetics for Power Electronics
Applications, https://doi.org/10.1007/978-3-030-96768-0_5
162 5 Applications of Transformers

In addition to simulating three-phase transformers, this chapter will introduce

high frequency transformers used in power electronic applications [10, 11, 40, 41].
This is of great relevance to a power electronics engineer, as a vast number of dc–
dc converters contain transformers and provide isolated outputs. We will begin the
discussion with the effect of frequency on the transformer construction. The notable
differences in a high frequency transformer are decreased size and the use of ferrite
cores as opposed to laminated iron cores [15].
We will examine a very popular converter used in power supplies—the flyback
converter [19, 20]. The flyback converter contains a high frequency transformer and
we will use the basic laws of physics to understand how the transformer behaves
during converter operation. Even though the operation of the power converter
forces the transformer to operate in a manner drastically different from regular grid
frequency transformers, we will show how the basic laws of physics still hold true
to describe the operation of the converter. We will simulate a flyback converter with
three isolated outputs using a four winding high frequency transformer to examine
the operation through simulation results.
This chapter will not present any new theory or discussion related to transform-
ers, but will merely use the discussions already put forth in the previous chapter.
The reader is advised to be well-versed with transformer simulations and applying
the basic laws of physics to understand transformer operation. As with the previous
chapters, there will be numerous references of simulations in this chapter. The reader
is advised to try out the simulations and also attempt modifications as suggested.

5.2 Three-Phase Systems

In the domestic (home and office) power system, equipment and appliances are
usually single-phase and rated for either 120 V, 60 Hz (USA, Canada) or 240 V,
50 Hz. Outside the home and office, a vast majority of equipment and appliances
are usually three-phase as multi-phase systems allow for greater energy density—
more power at relatively lower size as compared to the equivalent single-phase
equipment [1, 21]. Therefore, as the power rating of generators, motors and other
loads increases, it is usual for these loads to be designed to receive three-phase ac
supply. Most of the power system is also three phase except for the feeders that
feed homes and offices. The transformers that interconnect different segments of
the power system are also usually three-phase transformers. Before we get to three-
phase transformers, let us talk about three-phase systems in general.
Let us start by how a three-phase voltage is generated. The three phases are three
windings wound on the generator stator or rotor. These windings are wound in such
a manner that the voltages produced in the windings will such that they will be phase
shifted with respect to each other [1]. For a multi-phase machine with x phases, the
voltages will have a phase shift of 360◦ /x. For a three-phase machine, this phase
shift will therefore be 120◦ . Let us name the generated output voltages of the three-
5.2 Three-Phase Systems 163

phase machine va , vb , vc , and assume that they have RMS magnitudes of 240 V and
frequencies of 50 Hz. We could express the three-phase voltages as follows [1]:

√
va = 240 2 cos (100π t)

√ 2π
vb = 240 2 cos 100π t −
3 (5.1)

√ 4π
vc = 240 2 cos 100π t −
3

Figure 5.1 shows the three-phase voltage waveforms for a few cycles. As can be
seen, all three voltages are waveforms that have the same peak as they have the same
RMS magnitudes, but are phase shifted by 120◦ . This is one of the critical aspects
of three-phase (or multi-phase) systems—electrical quantities such as voltages and
currents that have the same magnitudes in all phases and are phase shifted from
each other by 120◦ . In such cases, the three-phase system is called a balanced system
[1, 21]. In practical cases, three-phase systems can and will be unbalanced, either by
having unequal magnitudes in the three phases or the phase shift between them may
not be 120◦ [1, 21]. There is vast theory in power systems studies on three-phase
systems and also on handling unbalanced systems [21]. However, let us begin with
balanced three-phase systems as our goal is to simulate three-phase transformers.
One of the most popular methods of analysing three-phase systems is by
expressing them as phasors [1, 21]. A phasor is a rotating vector with a magnitude

Fig. 5.1 Sample three-phase voltages

164 5 Applications of Transformers

ω
vc
va

vb

Fig. 5.2 Three-phase voltages as phasors

vc √
3 vc
2 Vm
√
3
4 Vm
1
va 4 Vm va
1
2 Vm
√
3
√
3 4 Vm
vb 2 Vm vb

Fig. 5.3 The use of phasors for computations

and a frequency of rotation. This concept is best described with a diagram such
as Fig. 5.2. Figure 5.2 shows the three-phase voltages as phasors rotating at an
angular speed of ω = 2πf . The magnitude of the phasors is equal to the RMS
magnitudes of the voltages. At any given instant, the phase angle of a phasor is
the angle made by the phasor with the positive horizontal axis being measured in
the counter-clockwise sense as shown for the phasor va . If the voltages are cosine
waveforms as assumed in our case, the projections of the phasors on the horizontal
axis will be their instantaneous values as shown in Fig. 5.2. If on the other hand, the
voltages were assumed to be sine waveforms, the projections of the phasors on the
vertical axis will be their instantaneous values. It can be observed from Fig. 5.2 that
the phasor vb lags behind phasor va by 120◦ and phasor vc lags behind phasor vb
by 120◦ . The reader is encouraged to verify that these phasors would result in the
voltage waveforms that have been assumed.
This phasor representation is applicable for all ac quantities such as three-phase
currents. Such a phasor representation of three-phase ac quantities is extremely
popular as we can use it not only for visualization but also for performing
basic computations. An example of phasor computations can be demonstrated by
performing a couple of sample computations such as va + vb + vc and va − v2b − v2c .
These computations are shown in Fig. 5.3. Each resultant phasor can be resolved
into projections on the horizontal and vertical axes, and subsequently the arithmetic
operation yields the projections of the resultant phasor on the same horizontal and
vertical axes [1, 21]. For convenience, the phasor va has been chosen along the
5.2 Three-Phase Systems 165

horizontal axis and therefore has a zero projection on the vertical axis. The other
phasors vb and vc have projections on the horizontal and vertical axes as shown by
the dashed lines. It is quite clear that va + vb + vc = 0 and va − v2b − v2c = 32 Vm with
√
Vm = 240 2. The reader is encouraged to verify the computations on their own.
With this basic overview of three-phase systems, we can now begin to talk about
three-phase transformers [13]. A three-phase transformer will interface two or more
three-phase systems of either the same or different voltage levels. As in the previous
chapter, we will begin with the simple example of two three-phase systems being
interconnected by a transformer. Therefore, the transformer will have a primary
consisting of three windings that will be connected to the a, b, c terminals of a three-
phase supply and will have a secondary consisting of three windings that will be
connected to the a, b, c terminals of another three-phase supply. As in the previous
chapter, we could have only one of the three-phase systems being energized or both
of the three-phase systems being energized. The next question to be asked is how
will the windings of the transformer be connected and what are the connections
available in the three-phase systems being interfaced.
There are several ways in which the three phases of a transformer’s windings can
be connected, but a few are the most widely used—the star (or wye) with neutral,
the isolated star (or wye) and the delta connection [13]. Figure 5.4 shows these
connections for the three different cases. The neutral terminal is a fourth terminal in
the case of the star winding with the common interconnection node being brought
out. Such a terminal is extremely useful in domestic distribution systems as each
phase terminal along with the neutral forms a single-phase supply as is available
in the power outlets of our homes and offices [1, 13]. However, in transmission
systems, the neutral terminal and wire are not convenient as this would result in
increased costs since we now need an extra wire. Therefore, transmission systems
are usually devoid of the neutral wire. This results in three-wire and four-wire

Star (Y) with neutral Star (Y) Delta (Δ)

a a a

b b b

c c c

Fig. 5.4 Star and delta connection of transformer windings

166 5 Applications of Transformers

systems with the difference being the absence or presence of the neutral wire. To
achieve a four-wire system with a neutral wire, the star connection with the neutral
terminal is usually the preferred option. To achieve a three-wire system, though
both the isolated star and the delta are possible, the delta connection quite often has
advantages over the star.
With these transformer connections, it is clear that the voltages across a winding
may not be the same as the voltages between two terminals. In Fig. 5.4, for star
connected windings, the voltage between terminals a and b will not be the same as
the voltage across the phase a or phase b winding. The voltage between terminals a
and b can be expressed as

vab = van − vbn (5.2)

In the above equation, the voltage vab is called the line voltage as it is the voltage
available between the terminals (or lines) a and b, while voltages van and vbn are
called the phase voltages as these are the voltages across the windings connected
between a terminal and the neutral thereby forming a phase [1].
We can define three line voltages as follows:

vab = van − vbn

vbc = vbn − vcn (5.3)
vca = vcn − van

We can use the phasor computation method similar to Fig. 5.3, to compute these line
voltages. The reader is encouraged to perform these computations
√ as an exercise.
The line voltage vab will have a magnitude that is 3 times the phase voltage
van and will lead the phase voltage by 30◦ [1]. Similar relations exist between
vbc and vbn , and vca and vcn . These line voltages can be written as the following
trigonometric equations:
√ √  π
vab = 240 3 2 cos 100π t +
6

√ √ 2π π
vbc = 240 3 2 cos 100π t − + (5.4)
3 6

√ √ 4π π
vca = 240 3 2 cos 100π t − +
3 6
5.2 Three-Phase Systems 167

On the other hand, for delta connected windings such as that shown in Fig. 5.4,
the line voltages are the same as the phase voltages. With reference to the delta
connected winding in Fig. 5.4, the line voltages can be written as [1]
√
vab = va = 240 2 cos (100π t)

√ 2π
vbc = vb = 240 2 cos 100π t −
3 (5.5)

√ 4π
vca = vc = 240 2 cos 100π t −
3

Along with the definition of line voltages for three-phase systems, there also
exists the definition of line currents as the currents flowing in or out of the terminals
of the three-phase system. The currents in the windings can be expressed as ia , ib ,
ic and are called the phase currents in a three-phase system. The currents flowing
into or out of the a, b and c terminals of the three-phase system can be called iLa ,
iLb , iLc , with the subscript L indicating that the current is flowing in the line. For a
star connected system as shown in Fig. 5.4, the line currents are equal to the phase
currents [1]:

iLa = ia
iLb = ib (5.6)
iLc = ic

On the other hand, for the delta connected system shown in Fig. 5.4, the line currents
are differences of the phase currents as follows:

iLa = ia − ib
iLb = ib − ic (5.7)
iLc = ic − ia

Phasor computations can be performed for the line currents in delta connected
systems.
When we wish to provide the voltage specifications of a three-phase system, the
convention is to specify the RMS line voltages [1]. As an example, if we specify
the voltage of a three-phase domestic distribution system to be 400 V, we imply that
the RMS line voltage is 400 V. As we have seen before, the relationship between the
line voltages and the phase voltages can differ depending on whether the three-phase
system is a star connected or a delta connected system. Therefore, along with the
specification of the voltage level, it is always customary to specify the connection
type of the three-phase system. As an example, a domestic distribution system could
be a three-phase star with neutral having a voltage level of 400 V. This implies that
the line voltage has an RMS value of 400 V and the phase voltage (between a line
168 5 Applications of Transformers

terminal and the neutral) has an RMS value of 230 V. On the other hand, a 11 kV
delta connected three-phase system will have both line and phase voltages to have
the same RMS value of 11 kV.
With the above discussion, we are now ready to modify the specifications of the
single-phase transformers of the previous chapter and begin modelling three-phase
transformers. If the reason for having several different types of connection of three-
phase systems is confusing at this point, the simulations in the later sections will
show their advantages and therefore, their applications. Before going further to the
next section, the reader must be clear about the difference between line voltage and
phase voltage in a three-phase system. The reader is encouraged to use the phasor
computation method to arrive at phasors for line voltages with respect to phase
voltages for both star and delta connected three-phase systems.

5.3 Simulation Model of a Three-Phase Transformer

In the previous section, we have provided a basic overview of three-phase systems,

which was essential in order to describe some of the winding connections of
three-phase transformers. Besides the star and delta connections, there are a
few other types that have their uses in special applications [13, 27, 28]. In this
section, we will begin simulating three-phase transformers while choosing a few
connection possibilities to cover some of the broad three-phase transformer types.
High power rating transformers with several windings will need to some extent a
basic understanding of the construction of transformers in order to arrive at the
parameters of the transformer equivalent circuit. Subsequently, we will extend the
transformer models developed in the previous chapter, and it will be clear, that
as with multi-winding transformers, simulating a three-phase transformer involves
merely extending the two winding transformer model.
Let us start by simulating a 500 kVA transformer with the primary and the
secondary being star connected set of windings. This simulation can be found
in the folder transformer_star_star within chapter5_transformer_
applications in the following link in the simulation repository: https://github.
com/opensourceelectrical/simulating-magnetics-for-power-electronics
Such a three-phase transformer can be called a star–star (Y-Y) transformer to
indicate the method in which the windings are connected. Let us connect a three-
phase voltage source to the primary and a three-phase resistive-inductive load to the
secondary. Figure 5.5 shows the topology of the circuit that will be simulated. Let
the voltage levels on the primary and the secondary of the transformer be 415 V and
11 kV, respectively. From our discussion in the previous section, these voltage levels
being three-phase voltage levels are the RMS magnitudes of the line voltages. Each
phase of the voltage source in the primary can have an RMS magnitude of 240 V.
5.3 Simulation Model of a Three-Phase Transformer 169

vsa isa ila Ll Rl

Zs a a

vsb isb ilb

b b

vsc isc ilc

c c

Fig. 5.5 Circuit for simulation of a star–star transformer

The voltage sources can be assumed to form a balanced three-phase supply and have
the following expressions:
√
va = 240 2 sin (100π t)

√ 2π
vb = 240 2 sin 100π t −
3 (5.8)

√ 4π
vc = 240 2 sin 100π t −
3

The three-phase transformer that we are going to simulate will have six windings
numbered 1–6. Each winding will be modelled by the voltage source in series
with resistance and with the voltage across the terminals measured by a Voltmeter.
We much choose the indexes to denote a particular winding so as to compute the
parameters of the equivalent circuit. Let us choose the odd numbers 1, 3, and 5 to
denote the primary windings of the transformer and the even numbers 2, 4, and 6
to denote the secondary windings of the transformer. Therefore, windings 1 and 2
will represent phase a of the transformer with 1 being the primary phase a winding
and the 2 being the secondary phase a winding. Windings 3 and 4 will represent
phase b of the transformer with 3 being the primary phase b winding and the 4
being the secondary phase b winding. Windings 5 and 6 will represent phase c of the
transformer with 5 being the primary phase c winding and the 6 being the secondary
phase c winding. The reader is free to choose another convention instead of this.
With this convention, we can begin with the transformer specifications:
import math
d t = 1 . 0 e−6
V r a t e d _ p r i m a r y = 4 1 5 . 0 / math . s q r t ( 3 . 0 )
V r a t e d _ s e c o n d a r y = 1 1 0 0 0 . 0 / math . s q r t ( 3 . 0 )
# Phase a − W i n d i n g s 1 and 2
# Phase b − W i n d i n g s 3 and 4
# Phase c − W i n d i n g s 5 and 6
VArated = 5 0 0 0 0 0 . 0 / 3
170 5 Applications of Transformers

# Primary windings
V1rated = Vrated_primary
V3rated = Vrated_primary
V5rated = Vrated_primary
# Secondary windings
V2rated = Vrated_secondary
V4rated = Vrated_secondary
V6rated = Vrated_secondary
f r a t e d = 50.0
o m e g a _ r a t e d = 2∗ math . p i ∗ f r a t e d

In the above block, Vrated_primary and Vrated_secondary are the phase

voltage ratings of the transformer windings. These are chosen to be √1 of the
3
RMS values of the line voltages as both primary and secondary of the transformer
are connected in star. This step is quite often skipped and results in errors. It is
important to remember that in a three-phase system, any voltage level specified is
the RMS magnitude of the line voltage and not the phase voltage. When computing
the parameters of the transformer, we must compute all parameters per-phase, and
therefore, the line voltage needs to be translated to the phase voltage for the chosen
connection type. Besides converting voltage ratings from line voltage levels to phase
voltage levels, we need to compute the per-phase maximum rated power by dividing
the total maximum power rating by a factor of 3.
We can use the power and the voltage ratings to compute the rated impedance
and therefore, the self-, leakage and magnetizing inductances of each winding:
I 1 r a t e d = VArated / V 1 r a t e d
I 2 r a t e d = VArated / V 2 r a t e d
I 3 r a t e d = VArated / V 3 r a t e d
I 4 r a t e d = VArated / V 4 r a t e d
I 5 r a t e d = VArated / V 5 r a t e d
I 6 r a t e d = VArated / V 6 r a t e d
Z1rated = V1rated / I 1 r a t e d
Z2rated = V2rated / I 2 r a t e d
Z3rated = V3rated / I 3 r a t e d
Z4rated = V4rated / I 4 r a t e d
Z5rated = V5rated / I 5 r a t e d
Z6rated = V6rated / I 6 r a t e d
L1 = 5 0 . 0 ∗ Z 1 r a t e d / o m e g a _ r a t e d
L2 = 5 0 . 0 ∗ Z 2 r a t e d / o m e g a _ r a t e d
L3 = 5 0 . 0 ∗ Z 3 r a t e d / o m e g a _ r a t e d
L4 = 5 0 . 0 ∗ Z 4 r a t e d / o m e g a _ r a t e d
L5 = 5 0 . 0 ∗ Z 5 r a t e d / o m e g a _ r a t e d
L6 = 5 0 . 0 ∗ Z 6 r a t e d / o m e g a _ r a t e d
Ll1 = 0.02 ∗ Z 1 r a t e d / omega_rated
Ll2 = 0.02 ∗ Z 2 r a t e d / omega_rated
Ll3 = 0.02 ∗ Z 3 r a t e d / omega_rated
Ll4 = 0.02 ∗ Z 4 r a t e d / omega_rated
Ll5 = 0.02 ∗ Z 5 r a t e d / omega_rated
Ll6 = 0.02 ∗ Z 6 r a t e d / omega_rated
Lm1 = L1 − L l 1
Lm2 = L2 − L l 2
Lm3 = L3 − L l 3
5.3 Simulation Model of a Three-Phase Transformer 171

Lm4 = L4 − L l 4
Lm5 = L5 − L l 5
Lm6 = L6 − L l 6

All we have done is replicate the computation for each winding. For computing
the self- and leakage inductances of a winding, we do not need any association
or linkage with any other winding. Therefore, these computations are independent
for each winding. The next step would be to compute the mutual inductances. At
this point, we need to either know or make assumptions about the transformer
connection.
The construction of transformers is a very detailed domain and transformers can
be wound in various different ways according to the need of the application [13,
14, 28]. However, in most applications that need transformers, the purpose of the
transformer is to achieve the maximum possible transfer of power with the minimal
possible size of the transformer. In terms of the size of the transformer, the core of
the transformer usually contributes the most to the bulk of the transformer. Cores
used for transformers have different shapes and use different materials to decrease
magnetic losses such as hysteresis and eddy current losses. In the previous chapter,
we had assumed the transformer core to be a simple rectangular core. In the chapter
on inductors and coupled inductors, we had also considered the three-limbed core.
In a similar manner, it is possible to have multi-limbed cores with a number of
different limbs on which windings can be wound.
In most practical transformers, if the purpose is to transfer power from one
winding to another, it would be obvious that they be wound in such a manner
that flux produced by one winding would link with the other winding with the
minimal possible leakage. On the other hand, it is also essential to ensure that there
is sufficient isolation between the windings to ensure that the two windings remain
completely isolated in the electrical sense. Therefore, in the case of transformers
with large turns ratio, the windings would also need to be sufficiently isolated. This
introduces a constraint on the space needed for a winding. Here lies the challenge
in transformer design—achieving maximum efficiency at lowest possible size while
ensuring acceptable electrical isolation.
With this background, let us choose a sample construction for our three-phase
transformer. Let us assume that the transformer core consists of three limbs as
shown in Fig. 5.6 [13, 14]. The primary and the secondary windings of a phase
are wound on one limb. Figure 5.6 is just a simplified schematic, and the actual
construction will be much more complicated taking into account the isolation
between the primary and the secondary. By assuming such a construction, we are
using only the basic logic in transformer construction—maximizing the coupling
between windings of a phase while allowing for sufficient space for windings and
ensuring electrical isolation. As an example, the windings of phase a − 1 and 2 are
wound on the same limb ensuring maximum coupling between them and therefore
minimal leakage of flux. By winding all three phases on the same core, we are
maximizing the use of the iron core and therefore, such a three-phase transformer
will be less bulky as compared to three single-phase transformers.
172 5 Applications of Transformers

Fig. 5.6 Three-phase

transformer wound on a
three-limbed core 1 3 5

1 3 5
2 4 6

2 4 6

If the three-phase transformer had been constructed as three single-phase

transformers, there would have been no magnetic coupling between the windings of
different phases. However, with all three phases wound on the same core, there will
be some coupling between the windings of different phases [27]. From Fig. 5.6, the
primary and the secondary windings of a phase have been wound on the same limb,
and therefore all the flux generated by one of the windings will completely link with
the other except for the leakage flux. However, the flux generated by a winding of
one of the phases will not completely link with the winding of another phase wound
on another limb. The reader is encouraged to refer the previous chapters, where
the magnetic circuit has been solved to determine the core flux. If we assume the
reluctance of each limb of the core to be the same, the flux generated by a winding
in one limb will divide equally into the other two limbs.
The mutual inductance between two windings can be expressed as

Mxy = k Lx Ly (5.9)

with k being the proportion of the flux generated by a winding that links with the
other winding. It is important to remember that k does not include the leakage flux as
leakage is accounted by the leakage inductance of the winding. Therefore, between
winding 1 and 2 of phase a, the coupling factor is k = 1 while between winding
1 of phase a and winding 3 of phase b, the coupling factor is k = 0.5. We are
making the assumption that all three limbs of the core have the same reluctances. If
we choose to calculate the reluctance of each limb with precise dimensions, the left
and the right limbs might have a slightly larger reluctance than the central limb. In
such a case, the coupling factor k will not be 0.5 but will be larger than 0.5 if we
consider the mutual inductance between a winding on one of the extreme limbs and
the central limb; but it will be smaller than 0.5 if we consider the mutual inductance
between windings on the extreme limbs. For simplicity, we will assume all three
limbs to have the same reluctance leading to a convenient value of k = 0.5.
The above discussion on transformer construction was essential to understand
how to compute the mutual inductance between the windings of the transformer.
In the case of high power transformers with several windings, it might be a bit
confusing to determine the extent of magnetic coupling between the windings. All
5.3 Simulation Model of a Three-Phase Transformer 173

that needs to be known is the basic construction of the transformer core and manner
in which the windings are wound on the core. The coupling factor between two
windings is merely the ratio of the flux produced by a winding that links with the
other. This coupling factor can be determined by knowledge of the reluctances of
the core segments and solving the magnetic circuit with the reluctances as series–
parallel combinations. However, for a complete simulation model, it is important
that we take into account all mutual inductances. It is tempting to assume that
the windings between phases are not magnetically coupled. However, such an
assumption is equivalent to simulating three single-phase transformers with separate
cores for each phase.
With the above background, we can now compute the mutual inductances
between the windings as
# Mutual i n d u c t a n c e w i t h i n w i ndi ngs o f a phase
M12 = math . s q r t ( Lm1 ∗ Lm2 )
M34 = math . s q r t ( Lm3 ∗ Lm4 )
M56 = math . s q r t ( Lm5 ∗ Lm6 )
# Mutual i n d u c t a n c e between phases
# H a l f o f t h e f l u x g e n e r a t e d i n one l i m b
# l i n k s with a winding in another limb
k_factor = 0.5
M13 = k _ f a c t o r ∗ math . s q r t ( Lm1 ∗ Lm3 )
M14 = k _ f a c t o r ∗ math . s q r t ( Lm1 ∗ Lm4 )
M15 = k _ f a c t o r ∗ math . s q r t ( Lm1 ∗ Lm5 )
M16 = k _ f a c t o r ∗ math . s q r t ( Lm1 ∗ Lm6 )
M23 = k _ f a c t o r ∗ math . s q r t ( Lm2 ∗ Lm3 )
M24 = k _ f a c t o r ∗ math . s q r t ( Lm2 ∗ Lm4 )
M25 = k _ f a c t o r ∗ math . s q r t ( Lm2 ∗ Lm5 )
M26 = k _ f a c t o r ∗ math . s q r t ( Lm2 ∗ Lm6 )
M35 = k _ f a c t o r ∗ math . s q r t ( Lm3 ∗ Lm5 )
M36 = k _ f a c t o r ∗ math . s q r t ( Lm3 ∗ Lm6 )
M45 = k _ f a c t o r ∗ math . s q r t ( Lm4 ∗ Lm5 )
M46 = k _ f a c t o r ∗ math . s q r t ( Lm4 ∗ Lm6 )
With the self-, leakage and mutual inductances computed, we can compute the
series resistance, core loss resistance and the interfacing resistance for each winding:
r1 = 0.01 ∗ Z1rated
r2 = 0.01 ∗ Z2rated
r3 = 0.01 ∗ Z3rated
r4 = 0.01 ∗ Z4rated
r5 = 0.01 ∗ Z5rated
r6 = 0.01 ∗ Z6rated
Rc1 = V 1 r a t e d ∗ V 1 r a t e d / ( 0 . 0 1 ∗ VArated )
Rc2 = V 2 r a t e d ∗ V 2 r a t e d / ( 0 . 0 1 ∗ VArated )
Rc3 = V 3 r a t e d ∗ V 3 r a t e d / ( 0 . 0 1 ∗ VArated )
Rc4 = V 4 r a t e d ∗ V 4 r a t e d / ( 0 . 0 1 ∗ VArated )
Rc5 = V 5 r a t e d ∗ V 5 r a t e d / ( 0 . 0 1 ∗ VArated )
Rc6 = V 6 r a t e d ∗ V 6 r a t e d / ( 0 . 0 1 ∗ VArated )
res_output1 = 100.0 ∗ Z1rated
res_output2 = 100.0 ∗ Z2rated
res_output3 = 100.0 ∗ Z3rated
res_output4 = 100.0 ∗ Z4rated
174 5 Applications of Transformers

res_output5 = 100.0 ∗ Z5rated

res_output6 = 100.0 ∗ Z6rated

All the parameters of the transformer equivalent circuit have now been computed.
We can now define the matrices of the matrix equation, which will be simplified and
solved using numerical integration:
L = [
[ L1 , M12 , M13 , M14 , M15 , M16 ] ,
[ M12 , L2 , M23 , M24 , M25 , M26 ] ,
[ M13 , M23 , L3 , M34 , M35 , M36 ] ,
[ M14 , M24 , M34 , L4 , M45 , M46 ] ,
[ M15 , M25 , M35 , M45 , L5 , M56 ] ,
[ M16 , M26 , M36 , M46 , M56 , L6 ]
]
R = [
[ r1 , 0 . 0 , 0 . 0 , 0 . 0 , 0 . 0 , 0 . 0 ] ,
[ 0 . 0 , r2 , 0 . 0 , 0 . 0 , 0 . 0 , 0 . 0 ] ,
[ 0 . 0 , 0 . 0 , r3 , 0 . 0 , 0 . 0 , 0 . 0 ] ,
[ 0 . 0 , 0 . 0 , 0 . 0 , r4 , 0 . 0 , 0 . 0 ] ,
[ 0 . 0 , 0 . 0 , 0 . 0 , 0 . 0 , r5 , 0 . 0 ] ,
[ 0 . 0 , 0.0 , 0.0 , 0.0 , 0.0 , r6 ]
]
B = [
[1.0 , 0.0 , 0.0 , 0.0 , 0.0 , 0.0] ,
[0.0 , 1.0 , 0.0 , 0.0 , 0.0 , 0.0] ,
[0.0 , 0.0 , 1.0 , 0.0 , 0.0 , 0.0] ,
[0.0 , 0.0 , 0.0 , 1.0 , 0.0 , 0.0] ,
[0.0 , 0.0 , 0.0 , 0.0 , 1.0 , 0.0] ,
[0.0 , 0.0 , 0.0 , 0.0 , 0.0 , 1.0]
]

As can be seen from all the computations above, most of them are merely
repetitions. Only the computation of the mutual inductances need some knowledge
of the transformer construction. Subsequently, construction of the matrices of the
matrix equation that represents the mathematical model of the transformer is a
mechanical process that can be easily extended for large transformers with many
windings. The matrix R is a diagonal matrix with the elements of the diagonal
being the winding resistances while the matrix B is an identity matrix. The
inductance matrix L is a well structured matrix with the diagonal elements being
the self-inductance of the winding and the off-diagonal elements being the mutual
inductances between two windings. Once these matrices have been defined, their
simplification so as to convert L to an upper triangular form and the numerical
integration is the same as before.
Once the winding currents have been computed by numerical integration, the
voltages of the controllable voltage sources that are used to represent each winding
can then be computed as
vout1 = v1 − winding_currents [0]∗ res_output1
vout2 = v2 − winding_currents [1]∗ res_output2
vout3 = v3 − winding_currents [2]∗ res_output3
vout4 = v4 − winding_currents [3]∗ res_output4
5.4 Simulation Results of Three-Phase Transformers 175

v o u t 5 = v5 − w i n d i n g _ c u r r e n t s [ 4 ] ∗ r e s _ o u t p u t 5
v o u t 6 = v6 − w i n d i n g _ c u r r e n t s [ 5 ] ∗ r e s _ o u t p u t 6

In this section, we have demonstrated how a three-phase transformer can be

simulated by merely extending the basic transformer model. Furthermore, a fairly
robust simulation model can be generated by taking into account some basic
aspects of the transformer construction that will affect the coupling between the
transformer windings. However, the transformer can be simulated using only the
basic specifications provided by the manufacturer—the maximum rated power in
VA, the maximum line voltage rating of each interface of the transformer and
the connection type of the transformer windings. The reader can either refer to
construction details provided by the manufacturer to compute parameters such as
the mutual inductances between windings or can make reasonable assumptions. In
the next section, let us examine the simulation results of the above model and further
discuss how with a few minor modifications, three-phase transformers with other
winding connections can also be simulated.

5.4 Simulation Results of Three-Phase Transformers

In the previous section, we examined in detail how a simulation model for a three-
phase transformer can be developed with only the basic manufacturer specifications.
In this section, we will examine the results of the simulation model in the previous
section while also examining how transformers with different winding connections
can be simulated. As will be seen, the modifications are surprisingly minor allowing
us to use this model for simulating a wide number of three-phase transformers used
in the modern power system.
With the transformer model developed in the previous section, we can simulate
the circuit of Fig. 5.5. Figures 5.7 and 5.8 show the line voltages in the primary and
secondary of the transformer. The peaks of the voltages can be seen to correspond
to RMS values of 415 V for the primary and 11 kV for the secondary. The voltage
sources connected to the primary have RMS magnitudes of 240 V. It is due to the
star √
connection that the line voltages applied to the primary of the transformer
are 3 times the RMS magnitude of the voltage sources connected in the phases.
By computing Vrated_primary and Vrated_secondary to be √1 of the
3
line voltages, the rated voltages of the windings are maintained equal to the RMS
magnitudes of the phase voltages instead of the line voltages. As stated in the
previous section, this computation is very important as any three-phase system
voltage magnitude is usually specified as the RMS value of the line voltage, but to
compute transformer equivalent circuit parameters, the rated voltage of the windings
must be calculated as phase voltages.
Figures 5.9 and 5.10 show the primary and secondary phase voltages, respec-
tively. In contrast to line voltages in Figs. 5.7 and 5.8, the RMS magnitudes of the
phase voltages are 240 V for the primary and 6.35 kV for the secondary. The reader
176 5 Applications of Transformers

Fig. 5.7 Transformer primary line voltages

Fig. 5.8 Transformer secondary line voltages

is encouraged to plot the phase voltage vsa and the line voltage vsab together to
verify that vsab leads vsa by a phase angle of 30◦ . Figures 5.11 and 5.12 show the
primary and secondary currents, respectively. The primary currents have a dc offset
due to the magnetizing currents having a dc offset with the primary voltages being
sine waveforms.
The reader is encouraged to apply different voltages to the primary of the
transformer and verify the secondary voltage that is produced is according to the
turns ratio specified. The three-phase transformer that has been simulated is a
5.4 Simulation Results of Three-Phase Transformers 177

Fig. 5.9 Transformer primary phase voltages

Fig. 5.10 Transformer secondary phase voltages

star–star transformer, where both the primary and the secondary windings of the
transformer are connected in star. Moreover, we have been dealing with balanced
three-phase systems where the RMS magnitude of all three-phase voltages is the
same, phase b lags behind phase a by 120◦ and phase c lags behind phase b by
120◦ . In domestic distribution systems, where each phase of a three-phase system
might be supplying loads in different parts of a house or an office, the chances
are that the three-phase system will be unbalanced. This is due to the fact that the
178 5 Applications of Transformers

Fig. 5.11 Transformer primary currents

Fig. 5.12 Transformer secondary currents

currents drawn by unbalanced loads will in turn produce unbalanced voltages due
to the unbalanced voltage drops across feeders.
A very popular connection strategy with respect to three-phase transformers
is when either the primary or the secondary is connected in star and the other
is connected in delta [13]. These transformers are called star–delta or delta–star
5.4 Simulation Results of Three-Phase Transformers 179

transformers. These transformers are very popular when one of the subsystems
needs to be a four-wire system with a neutral but the other need not [1]. An example
is of a distribution transformer that supplies a house or building at 415 V from
a 11 kV or a 22 kV distribution feeder. The three-phase supply at the house will
need to have a neutral connection as domestic power supplies are single-phase and
therefore, most appliances will be connected between a phase and a neutral. The
415 V three-phase supply will need to be star connected with a neutral. However,
the three-phase system on the 11 kV or 22 kV does not need a neutral connection
as that feeder is merely supplying power and is not feeding any loads that need a
neutral. In such a case, connecting the 11 kV or 22 kV feeder in star with a neutral
would unnecessarily require a four-wire feeder. If instead, we choose to connect the
11 kV or 22 kV windings of the transformer in delta, we would have a three-wire
feeder, which would decrease the cost.
A four-wire system is required in certain cases such as while supplying loads
that will need a neutral connection. However, in such cases, where the proportion
of single-phase loads is high, a four-wire system is usually unbalanced. A current
will flow in the neutral wire besides also the three-phase wires. This flow of current
cannot be avoided as that is the requirement of the load. The question then arises—
how would four currents map to three currents in a delta–star transformer, with the
delta connected windings not having a neutral wire similar to the star connected
windings? The current in the neutral on the star connected windings will circulate
within the delta connected windings [13]. The line currents on the delta connected
side will be devoid of this neutral current. This might seem a little confusing and
will be demonstrated through a simulation soon.
Figure 5.13 shows a circuit for simulating a three-phase delta–star transformer.
This simulation can be found in the folder transformer_delta_star within
chapter5_transformer_applications in the following link in the sim-
ulation repository: https://github.com/opensourceelectrical/simulating-magnetics-
for-power-electronics
The primary windings of the transformer are connected in delta and supplied
by a three-phase voltage source. The secondary of the transformer is connected

vsa ila
Zs a a Ll Rl

isa
vsb ilb
b b

isb
vsc ilc
c c

isc iln

Fig. 5.13 Circuit for simulation of a delta–star transformer

180 5 Applications of Transformers

in star and is supplying a three-phase load that is connected in star. To begin

with, let us consider a balanced load, and therefore, the neutral connection on the
secondary carries no current. Soon, we will introduce an unbalance in the load and
repeat the simulation. Let us choose the transformer to have a maximum power
rating of 50 kVA. Let us choose the primary (delta) of the transformer to have a
maximum voltage rating of 11 kV and the secondary (star) of the transformer to
have a maximum voltage rating of 415 V. With these basic specifications, let us
simulate the circuit of Fig. 5.13.
We can get started with translating the transformer power and voltage rating to
the simulation model of the transformer:
import math
d t = 1 . 0 e−6
Vrated_primary = 11000.0
V r a t e d _ s e c o n d a r y = 4 1 5 . 0 / math . s q r t ( 3 . 0 )
# Phase a − W i n d i n g s 1 and 2
# Phase b − W i n d i n g s 3 and 4
# Phase c − W i n d i n g s 5 and 6
VArated = 5 0 0 0 0 . 0 / 3 # V o l t −Amperes
# Primary windings
V1rated = Vrated_primary
V3rated = Vrated_primary
V5rated = Vrated_primary
# Secondary windings
V2rated = Vrated_secondary
V4rated = Vrated_secondary
V6rated = Vrated_secondary
f r a t e d = 50.0 # Hz
o m e g a _ r a t e d = 2∗ math . p i ∗ f r a t e d

The per-phase voltage rating of the primary is the same as the voltage level of the
three-phase system because the primary windings of the transformer are connected
in delta due to which the phase and line voltages have the same RMS magnitude. On
the other hand, the per-phase voltage rating of the secondary is √1 times the voltage
3
level of the three-phase system because the secondary windings are connected in
star. Just as in the previous simulation, we need to compute the maximum per-
phase power rating by dividing the total three-phase maximum power rating by 3.
The above changes are the only changes that need to be made to the simulation
model. All other computations will remain the same as the computation of all other
parameters are made on a per-phase basis using these basic power and voltage
ratings.
Figures 5.14 and 5.15 show the primary and the secondary line voltages of the
three-phase transformer. The peaks of the waveforms correspond to the line voltage
ratings of the primary and the secondary of the transformer. Figures 5.16 and 5.17
show the primary and secondary line currents of the three-phase transformer. As
in the previous simulation, the currents in the primary have dc offsets due to the
magnetizing currents drawn by the primary winding. The reader is encouraged to
plot the waveforms of the other measured quantities such as the phase voltages and
phase currents.
5.4 Simulation Results of Three-Phase Transformers 181

Fig. 5.14 Transformer primary line voltages

Fig. 5.15 Transformer secondary line voltages

The operation of the delta–star transformer is most interesting when we consider

the case of the load on the star connected windings being unbalanced. Let us make
the load impedance in phase b to be 1.5 times the load impedance in phase a, and
the load impedance in phase c to be 2 times the load impedance in phase a. We
have now introduced a fairly severe unbalance between the three phases of the load.
182 5 Applications of Transformers

Fig. 5.16 Transformer primary line currents

Fig. 5.17 Transformer secondary line currents

5.4 Simulation Results of Three-Phase Transformers 183

Under these unbalanced conditions, the current iln will now be a non-negligible
value from the expression:

iln = − (ila + ilb + ilc ) (5.10)

Since the secondary of the transformer is connected in star, both the phase
and line currents in the secondary will be unbalanced and will have a component
iln
3 , which is called the zero sequence component. The zero sequence component
2 iln
can only flow in the secondary transformer windings if a referred current N
N1 3
(N 2
N1 —transformer turns ratio) can flow in the primary windings as well [1]. This
is fundamental to the operation of any transformer and follows from the theory
described in the previous chapter. If the voltage applied to the primary windings
is approximately constant, the core flux will remain approximately constant as
well. Therefore, any current flowing in the secondary winding due to a load being
connected will have a demagnetizing effect. This demagnetizing effect has to be
neutralized by a current flowing in the primary winding that has a nullifying
magnetizing effect.
If the secondary windings of the transformer were to be connected in star with
a neutral wire, that alone does not imply that a current can flow in the neutral wire
due to the load being unbalanced [1]. If a current flows in the neutral wire, a zero
sequence component must also flow in the three phases of the secondary. And this
can only be possible if a referred zero sequence component flows in the primary
windings as well. Not all transformer connections will permit this. For example, the
star–star transformer of Fig. 5.5 that has no neutral wire in the primary or secondary
windings cannot support a zero sequence current. Even if the secondary windings
of Fig. 5.5 had a neutral wire that was connected to the neutral of the load, such
a connection would not result in the flow of a neutral current. This is due to the
fact that the neutral current will require zero sequence components to flow in the
secondary windings as well. With the primary of the transformer not having a neutral
wire, referred zero sequence components cannot flow in the primary windings.
In the case of the delta–star transformer in Fig. 5.13, a very interesting case
follows. Due to the primary of the transformer being connected in delta, the three-
phase system on the primary cannot have a neutral wire. Therefore, it is easy to
jump to the conclusion that a delta–star transformer will not support zero sequence
currents just like a star–star transformer without any neutral on the primary.
However, the zero sequence currents can circulate within the delta connection as
shown in Fig. 5.18. Due to the nature of connection of the primary windings, a zero
sequence component can flow as

kiin = iwa + iiwb + iwc (5.11)

The currents in each winding iwa , iwb , iwc will have zero sequence components
equal to i3ln . However, the line currents isa , isb , isc will not have these zero sequence
184 5 Applications of Transformers

isa
a
iwc iwa
kiln
isb
b
c iwb
isc

Fig. 5.18 Zero sequence component circulating in delta connected windings

Fig. 5.19 Transformer secondary line currents—unbalanced

components as the zero sequence components in the windings will cancel out as
follows:

isa = iwa − iwc

isb = iwb − iwa (5.12)
isc = iwc − iwb

Therefore, a delta–star (or a star–delta) transformer is a very useful transformer

when one three-phase system must support an unbalanced system but the other
three-phase system need not. Let us examine this with a modified simulation.
Figure 5.19 shows the simulation results with the load currents in the star con-
nected secondary being unbalanced sine waveforms of different RMS magnitudes.
5.4 Simulation Results of Three-Phase Transformers 185

Fig. 5.20 Secondary neutral current and circulating current in the primary

The neutral current in the secondary can be measured using the component Amme-
ter_loadneutral. However, the circulating current in the delta connected primary
cannot be directly measured by an Ammeter as this circulating current is due to
the zero sequence component present in the winding currents. We can calculate this
circulating current by computing in the simulation model the sum of the currents
measured by Ammeter_T1W1, Ammeter_T1W3 and Ammeter_T1W5. We will
need to configure the control file with three additional inputs that are connected to
the measurements of Ammeter_T1W1, Ammeter_T1W3 and Ammeter_T1W5 and
designated by the input variables iprim1, iprim2 and iprim3, respectively.
Figure 5.20 shows the current in the secondary neutral along with the circulating
current i in the delta connected primary expressed as

i = iwa + iwb + iwc (5.13)

The circulating current in the delta connected primary is much smaller than the
neutral current in the secondary due to the turns ratio of the transformer. To compare
the circulating current with the neutral current, the circulating current is multiplied
by a value of 45 close to the turns ratio of the transformer 45.83. This is so that the
waveforms are not exactly superimposed but are close enough to visually verify that
they are equal. Furthermore, since the neutral current is the negative sum of the three
secondary line currents, the circulating current is inverted. From Fig. 5.20, it is very
clear that the neutral current in the secondary is circulating in the delta connected
primary windings.
186 5 Applications of Transformers

In this section, we have examined in detail the simulation of two different three-
phase transformers. It is clear that the transformer model developed in the previous
chapter is extremely flexible and can be modified to simulate different types of
transformers. Besides the mathematical model, the only other change that needs
to be made is the physical connection in the circuit schematics between transformer
windings. The reader is encouraged to try simulating other combinations of three-
phase transformers such as the star–delta transformer or the star–star transformer
with and without a neutral wire.

5.5 High Frequency Transformers

Until now, all the transformers that have been simulated have been line frequency
(50 Hz or 60 Hz) transformers. These transformers are critical in the power system
and enable efficient bulk power transmission over long distances by stepping up
the voltage to hundreds of kilovolts. With the proliferation of portable electronic
appliances such as laptops, mobiles, tablets, etc., the power supplies for these
have been steadily getting smaller and lighter. Within these power supplies are
transformers that would have been completely unnoticed unless the power supplies
are taken apart for repair [10, 26]. This application of the transformer is also
extremely interesting, especially for a power electronics engineer. In this section,
we will introduce the concept of these power supply transformers and describe how
they are different from line frequency transformers.
Before we describe transformers used in power supply applications that contain
power electronic converters, let us describe the impact of high frequencies on
transformers [1, 13]. When a winding of a transformer is energized with an ac
voltage of some frequency f , a magnetic flux will flow in the core of the transformer
that will also be an ac quantity of the same frequency f . Let us express this core
flux as the following sinusoid:

φ = φm sin 2πf t (5.14)

With the core flux being an ac quantity, emfs will be induced in all the windings
of the transformer, which will be directly proportional to the rate of the change of
core flux:
dφ
ex = Nx = 2π Nx φm f cos 2πf t (5.15)
dt
From the above expression, for a given winding with Nx turns, the product φm f
contains the variables that impact the magnitude of the induced emf as frequency
changes. For a fixed magnitude φm of core flux, the induced emf will increase as the
frequency increases.
5.5 High Frequency Transformers 187

Let us consider the simple case of a transformer being energized by a single

supply whose voltage magnitude is equal to the maximum rated voltage of the
winding being energized. As we have already described in the previous chapter,
the emf induced in this winding will be expressed as

ex = v − ix rx ≈ v (5.16)

As it is a fair approximation to neglect the voltage drop across the winding

resistance, the induced emf is approximately equal to the voltage applied across
the winding.
From the above equations, it follows that the peak Vm of the applied voltage can
be expressed as

Vm ≈ 2π Nx φm f (5.17)

Subsequently, we can ignore the constants and the number of turns Nx of the
winding as a construction constant, to result in the following expression:

Vm ∝ φm f (5.18)

This final proportionality expression describes the impact of frequency on a

transformer. As the frequency f increases, the magnitude of the core flux φm will
decrease. For a particular construction of the core and a given cross-sectional area
Ac of the core, the flux density B will decrease as a result of the increase in
frequency. In Chap. 2, we had described the B-H curve of magnetic materials. Such
B-H curves are available for most commonly used magnetic materials besides also
being available from the manufacturer supplying the core or the magnetic material
for the core. For every magnetic material, there exists for a particular temperature
and other conditions, a value of flux density B beyond which the core saturates.
If the flux in the core is such that the flux density B exceeds this saturation value,
the resultant core saturation will cause increased losses and raised temperatures,
which in turn can cause the magnetic properties of the core to deteriorate. However,
it is also advisable to allow the maximum flux in the core to attain a value such
that the flux density will be slightly lower than the saturation value. In this manner,
the transformer core is fully utilized resulting in an optimized design [13]. Due to
increased frequency, the core flux decreases in magnitude. Instead of allowing the
flux density in the core to drop to a very low value, the size of the core can be
decreased such that the maximum flux density will continue to be slightly lower
than the saturation value of flux. Therefore, with increased frequency, the size of the
core decreases resulting in smaller transformers and thereby smaller power supplies
[10, 26].
There are a few other effects of high frequency operation that require high
frequency transformers to be very different in construction with respect to line
frequency transformers. The ac flux in the transformer core results in an induced
emf in the windings of the transformer. However, besides the windings, emfs are
188 5 Applications of Transformers

induced in the core itself as core materials such as iron are conductors. This is due
to the fact that Faraday’s Law is applicable to any conductor that is subjected to a
changing magnetic field and this includes the core and not just the windings. Due
to the emfs induced in the core, currents called eddy currents flow in the core [13].
These currents have circular paths and flow in planes perpendicular to the direction
of the magnetic flux. The direction of these currents can be deciphered by using
Lenz’s Law—they will be in such a direction that they oppose the cause (changing
flux) that produces them.
These eddy currents are quite obviously undesirable as there is no benefit in
currents flowing in the core—only the voltages at the windings are useful. These
eddy currents result in ohmic losses in the core and result in increased core
temperatures. In our simulation model, we have captured eddy current losses using
the core loss resistor Rc . To minimize the heating effect of these eddy currents,
the iron core is divided into sheets or laminations that are coated with insulating
materials. Therefore, instead of using a solid block of iron, the transformer core is
comprised of a number of iron laminations insulated from one another with a thin
layer of insulation [13]. This increases the resistance for eddy currents to flow and
therefore decreases eddy current losses.
Another major core loss component in transformers is the hysteresis loss [13].
In Chap. 2, we have described the B-H curve. We have described how the ideal
linear B-H curve does not exist in practical cores as when a core is magnetized, it
tends to retain its magnetism. This implies that as the magnetic field strength H
increases, the flux density B will not increase linearly but will increase at a lower
rate. Similarly, when H decreases, B will not follow linearly as the core will retain
some of its magnetism and therefore will decrease at a lower rate. This retention of
the core produces a loss that is called as the hysteresis loss. In the simulation model,
the hysteresis loss was included in the loss resistor Rc .
Since eddy currents are produced due to the changing magnetic flux in the
core, as the frequency of the supply increases, the eddy currents will increase in
magnitude and so will eddy current losses [13]. The same applies for hysteresis
loss as well [13]. With increased frequency, the loss resulting from the retention of
the core increases. As we progress to very high frequency operation in thousands
of Hertz, eddy current and hysteresis losses will be unacceptably high even in iron
cores that are laminated. For this reason, high frequency transformers use cores
made from ferrite that is a ceramic composed of iron oxide and several other
metals [15]. Ferrite cores have magnetic properties similar to iron cores but are
not conductive, which results in decreased eddy current losses at high frequencies.
Ferrite cores are available in numerous shapes (circular, rectangular, C-shaped,
E-shaped, etc.) and sizes for ready use in winding inductors or transformers [15].
In addition, a number of high frequency inductors and transformers of standard
ratings are available for standard power supply applications. A power electronics
engineer can either purchase a ready-made inductor or transformer, or can use ferrite
cores available in different shapes and construct custom-designed transformers and
inductors. All concepts in design for high frequency inductors and transformers are
exactly the same as for line frequency inductors and transformers. An engineer need
5.6 Flyback Converter 189

only remember three laws—Kirchhoff’s Voltage Law, Faraday’s Law and Lenz’s
law to translate the windings and the core to a magnetic circuit.
In this section, we have described the basic differences between high frequency
transformers and grid frequency transformers. The major advantage of a high
frequency transformer is the decreased size due to the lower flux magnitude, which
results in a smaller core. On the other hand, high frequency operation needs the
transformer core to be composed of a special material such as ferrite to minimize
core losses that are proportional to frequency. In the next section, we will describe
the operation of the high frequency transformer using a very popular converter
called the flyback converter.

5.6 Flyback Converter

In the previous section, we described some of the broad differences between grid
frequency transformers and high frequency transformers. The philosophy behind
the operation of a high frequency transformer is the same as that for a grid fre-
quency transformer. The difference lies in the application, whereby high frequency
transformers are largely used in conjunction with power electronic converters due
to which they are energized by a switched voltage in contrast to grid frequency
transformers that are energized by a voltage that has roughly a sine waveform
[10, 26]. Depending on the power electronic converter, the type of switched voltage
a high frequency transformer is subjected to can vary drastically. In this section, we
will examine a very popularly used converter—the flyback converter [19, 20].
Figure 5.21 shows the topology of the flyback converter [19, 20]. What makes
a flyback converter such an interesting power converter is, that at first glance, one
might think such a converter will never work. The flyback converter consists of a dc
voltage source such as a battery energizing the primary winding of a high frequency
transformer through a power device S such as a MOSFET. The gate signals to

Fig. 5.21 Topology of a ip

flyback converter
is D

+
Vin vo C RL
−

S
190 5 Applications of Transformers

the power device can be turned on and off such that the voltage applied across
the primary winding is a switched voltage. It is extremely important to notice the
dot polarities of the transformer windings—for the primary the dot is on the upper
terminal while for the secondary the dot is on the lower terminal. The secondary
of the transformer feeds a load RL connected across an output capacitor C through
a Diode D. It is again extremely important to notice that the anode of the diode is
connected to the terminal of the secondary that does not have a dot.
As with all power converters, in order to regulate the output voltage produced,
the conduction and non-conduction time intervals of the power device S need to
be regulated. The literature of control strategies proposed for the flyback converter
is vast [19, 20] and the simulation covered in this chapter will not include closed-
loop control. Since our purpose is merely to understand the operation of the high
frequency transformer, all we need is to implement an open-loop switching strategy.
We already covered an example of a Ćuk converter in Chap. 3 where the power
device was operated at a constant switching frequency during which it was provided
gate signals for a fraction of the switching time period such that the device was
conducting and for the remaining part of the switching time period, gate signals
are removed and the device does not conduct. This is one of the simplest switching
strategies and can be implemented as a Pulse Width Modulation (PWM) scheme
with a constant frequency carrier waveform and a duty ratio. There have been
many variations proposed to this scheme, and the reader is encouraged to try and
implement them.
Figure 5.22 shows the operation of the flyback converter through the waveforms
of the three main quantities—the current in the primary winding, the current in the
secondary winding and the output voltage. Let us assume for simplicity that the
input voltage Vin is constant. In Fig. 5.22, the time interval for which the power

Fig. 5.22 Working of a T

flyback controller
ip tON tOF F

is

vo
5.6 Flyback Converter 191

device is conducting is marked as tON and the time interval when non-conducting
is marked as tOF F . When the power device is conducting, the input voltage appears
across the primary winding of the transformer. As with any transformer, an induced
emf will be produced in the winding that will be equal and opposite to the applied
voltage. Therefore, the induced emf will have a polarity such that the upper terminal
of the primary has a positive polarity with respect to the lower terminal.
Due to the dot polarities of the transformer windings, the emf induced in
the secondary terminal will be such that the voltage at the terminal with the
corresponding dot (lower terminal) will be positive with respect to the upper
terminal. The reader should use the theory presented in Chaps. 3 and 4 to verify
this concept. As a result, the diode will be reverse biased and will not conduct.
Since only one winding of the transformer is conducting, the current drawn by
the primary winding is the magnetizing current alone, which increases as shown
in Fig. 5.22. During this interval (tON ), the current is in the secondary is zero. The
output voltage vo across capacitor C will decrease as it discharges through the load
resistor RL .
When the gate signals to the power device S are withdrawn and the power device
stops conducting, we are now in the interval marked by tOF F . One might think
that nothing will happen during this interval as the input voltage Vin has been
disconnected from the transformer. However, we must remember that in the previous
interval tON , due to the flow of the magnetizing current, a flux will flow in the
transformer core and energy is stored in the magnetic field of the transformer. This
stored energy will try to find a way to continue to flow. The reason—Faraday’s Law
states that any change in the magnetic field associated with a conductor (winding)
will produce an induced emf. Therefore, if the power device has stopped conducting,
and the primary is no longer energized by the input voltage Vin , the core flux will
fall due to the loss of the energy source. This will result in emfs induced in both the
primary and the secondary.
The next question—what will be the nature or polarity of the induced emf? For
this, we use Lenz’s Law—the emfs induced will be such so as to oppose the cause
that produced them. The cause in this case is the falling core flux, which implies
that the emfs will attempt to restore the core flux. The reader is again advised to
review the theory behind dot polarity conventions. The dots on the terminals of the
transformer indicate that currents flowing in the two windings will have the same
effect if they were both to enter (or both to leave) the terminals with the dots. During
the period tON of conduction, the current ip was flowing in the primary winding
such that it entered the primary at the upper terminal marked by the dot. During the
period, the magnetic flux rose in the transformer core as the magnetizing current ip
increased. This implies, that if the core flux were to be made to increase, a current
must flow in either winding such that it enters one of the dots.
With the power device S being made to stop conducting by withdrawing the gate
signals, a current cannot flow in the primary. This implies the only way to restore
the core flux is for a current to flow in the secondary winding. If the induced emf in
the secondary winding is such that the upper terminal of the secondary that does not
have a dot has a positive voltage with respect to the lower terminal of the secondary,
192 5 Applications of Transformers

the diode will be forward biased if the induced emf is greater than the output voltage
vo . If the diode is forward biased and begins to conduct, a current is will flow in the
secondary charging the capacitor C. This current is will flow into the lower terminal
of the secondary marked by the dot thereby opposing the fall in core flux. Since the
output capacitor C is being charged by the secondary current is , the output voltage
vo will increase. As the energy stored in the magnetic field is dissipating during this
interval, the secondary current begins at a non-zero value and decreases, and the
energy in the field dissipates.
As can be seen from the operation of the flyback converter described above,
there are stark differences with respect to the normal operation of a transformer
even though the basic laws of physics governing the operation of the transformer
are the same. When the primary of the transformer is energized by the input voltage
during tON , the Kirchhoff’s Voltage Law expressed at the primary winding is as
follows:

Vin = ep + ip rp (5.19)

where rp is the primary winding resistance and ep is the primary winding induced
emf.
Faraday’s Law applies to the high frequency transformer as well, and therefore
induced emf can be expressed as the rate of change of primary winding flux
linkages:

dφ
Vin = Np + ip rp (5.20)
dt
From the above expression, the transformer core flux will increase as it is approxi-
mately equal to the integral of the applied voltage. Subsequently, the primary current
ip will increase so as to provide the ampere turns necessary for the core flux to
increase.
All the above expressions were very similar to the case of the grid frequency
transformer. However, there is now a drastic difference between the flyback
transformer and a normal grid frequency transformer. When the primary is energized
and the core flux increases, the load is disconnected from the secondary winding.
This is due to the dot polarity of the windings and the placement of the diode that
causes it to be reverse biased. In a regular grid frequency transformer, the secondary
usually supplies the load and therefore, a current flows in the secondary winding.
This current in the secondary winding has a demagnetizing effect on the core flux
and results in a current component drawn by the primary winding such so as to
restore the core flux. In the case of the flyback transformer, with no current flowing
in the secondary, the primary current and the core flux increase without any relation
to the secondary.
5.6 Flyback Converter 193

This might almost seem like a violation, as until now, we had described how
turns ratio can be used to equate the primary and the secondary currents of the
transformer:

Np ip = Ns is (5.21)

When the power device is conducting and is = 0, it might seem like the above
equation is violated. However, we must remember that a transformer must draw
a magnetizing current in order to maintain the magnetic field in the core. This
magnetizing current is drawn by the winding that is energized and is unrelated to
the current flowing in any other winding. In the case of the flyback transformer, the
primary current is the magnetizing current of the transformer and is needed for the
core flux to increase.
There is one more riddle in the operation of the flyback converter. How is it
that the induced emf in the secondary winding attains a polarity such so as to
cause the diode to conduct during the tOF F interval? We must remember that, emfs
are induced whenever the core flux changes. Therefore, an emf is induced in the
secondary winding during the tON period as well when the core flux is increasing,
and the magnitude of the secondary winding induced emf can be expressed as

dφ
es = Ns (5.22)
dt
During the tON period, the induced emf in the primary winding is such so as to
oppose the increase in the core flux. Since the dot polarities of the windings are
chosen in a reverse fashion, the induced emf in the secondary winding will have such
a polarity that it will attempt to produce a secondary current that will also oppose the
increasing core flux. As per our theory of dot polarities, for the secondary current to
oppose the increase in core flux, it would have to be in a direction opposite to that
shown in Fig. 5.21. Due to the presence of the diode, such a current cannot flow.
Therefore, the induced emf in the secondary winding though present has no effect.
During the tOF F period, when the primary is no longer energized by the source
Vin , the core flux will tend to decrease. The resultant induced emf in both the
primary and the secondary windings will now be exactly the reverse of what they
were during the tON period as now the induced emfs will such so as to oppose the
decreasing core flux. During this period, it is the induced emf in the primary that has
no effect while the induced emf in the secondary winding will cause the diode to be
forward biased. Once again, with only the current in one winding being non-zero,
it is the magnetizing current that is flowing in the secondary. Therefore, in reality,
the flyback converter is a very clever design that uses a transformer to store energy
in the magnetic field from the input source in one interval and dissipate this stored
energy in the load in the other interval.
In this section, we described the operation of a flyback converter using the
basic laws of physics as was done in all the cases before. The transformer used
in a flyback converter behaves very differently from a regular grid frequency
194 5 Applications of Transformers

transformer. However, every state of operation of the flyback converter and the
transformer complies with the basic laws of physics, and therefore, a flyback
transformer under the hood is just another transformer. In the next section, we will
present the simulation of a flyback converter.

5.7 Simulation of a Flyback Converter

In the previous section, we described the operation of the flyback converter. The
reader will find several arguments on the internet about whether the transformer
used in the flyback converter is indeed a transformer. This is due to the fact that
the flyback converter has a pulsed transfer of power through the transformer unlike
most other transformer applications, where a load is continuously served by the
transformer. From the previous section, it is clear that with all the laws of physics
that are applicable to the flyback transformer, we can continue to model and simulate
the transformer for the flyback converter using the simulation model developed
so far. In this section, we will describe the changes needed to simulate a flyback
converter and present the simulation results.
Figure 5.23 shows the topology of the flyback converter that will be simu-
lated. This simulation can be found in the folder flyback_converter within
chapter5_transformer_applications in the following link in the sim-

Fig. 5.23 Flyback converter ip is1 D1

with multiple outputs

RL1
C1
Vin

S is2 D2
RL2
C2

is3 D3
RL3
C3
5.7 Simulation of a Flyback Converter 195

ulation repository: https://github.com/opensourceelectrical/simulating-magnetics-

for-power-electronics
One of the advantages of the flyback converter is that due to the necessity of the
transformer, it is also possible to use a transformer with several secondary windings
to produce multiple outputs. Such a topology has a few very important benefits.
Since we have several windings, we can have different turns ratios between the
secondary windings and the primary windings so as to produce outputs at different
levels. This is extremely useful when a control circuit requires multiple power
supplies such as 3.3 V, 5 V, 9 V, etc. Another benefit of having several windings is
that each output is isolated from the other. This is very important in gate drivers used
for power electronic converters, when the gate drivers for the different power devices
in the converter need isolated power supplies to ensure that there is no interference
between them that could cause a false conduction state of a power device.
In the topology of the flyback converter seen in Fig. 5.23, it is important to note
that we are maintaining the reverse dot polarities for all three secondary windings.
This is to ensure that none of the diodes D1 , D2 or D3 are forward biased and
conducting when the power device S in the primary is conducting. Let us suppose
that the input voltage Vin were the output voltage of a 24 V battery. Let us suppose
we wish to produce output voltages across capacitors C1 , C2 and C3 to be 5 V, 3.3 V
and 12 V, respectively. We must now choose the rated voltages of the windings of
the transformer. It is important to remember that the power device S is conducting
for a period of time during a switching time period. During this period, the primary
winding is energized by the battery. When the power device S is not conducting, the
secondary windings are transferring power to the outputs. The duty ratio d of the
power device is the ratio tON
Ts , which will always be less than 1.
There is a vast amount of literature related to various possible control and
operation strategies of the flyback converter [19, 20]. In this simulation, we will
not delve into the details of the various ways in which the flyback converter can be
operated as our focus is to examine the behaviour of the high frequency transformer
in the converter. Therefore, we will choose a fairly simple mode of operation. We
will not implement a closed-loop control scheme whereby one of the output voltages
can be regulated to a desired reference. We will choose a fixed duty ratio d and
implement an open-loop control scheme. Let us choose a duty ratio of 0.25 such
that the power device is conducting only for 25% of the switching time period.
Since we have chosen this value of duty ratio, let us choose the rated voltages of the
primary winding of the transformer to be 24 V and the secondary windings of the
transformer to be 15 V, 9 V and 36 V, respectively. Due to the open-loop operation
of the converter, the output voltages will only be close to the values that are desired.
To define the specifications of the transformer in the flyback converter, we need
to specify the maximum power rating of the transformer. Since we have three
isolated outputs that would be used for control purposes, we could choose the
196 5 Applications of Transformers

maximum power rating of the transformer to be 500 VA. The main specifications
of the transformer in the simulation model can be
import math
d t = 1 . 0 e−8
VArated = 5 0 0 . 0
V1rated = 24.0
V2rated = 15.0
V3rated = 9.0
V4rated = 36.0
f r a t e d = 10000.0
o m e g a _ r a t e d = 2∗ math . p i ∗ f r a t e d

We have chosen the rated frequency of the transformer to be 10 kHz as we

will choose this frequency to be the switching frequency of the converter. Flyback
converters usually have a switching frequency range between 10 kHz and 50 kHz.
We have also chosen the time step dt of the control file to be 10 ns as we are
simulating a high frequency transformer instead of a grid frequency transformer.
This decrease in the control time step can be seen to be essential when we examine
the parameters of the transformer equivalent circuit. As an example, the parameters
of winding 1 are listed:
L1 = 0 . 0 0 0 9 1 6 7
Lm1 = 0 . 0 0 0 9 1 6 3
M12 = 0 . 0 0 0 5 7 2 7
L l 1 = 3 6 6 . 7 e−9
r1 = 0.01152
Rc1 = 1 1 5 . 2
L
Due to the decreased values of inductances, the time constants ( R ) of all the
branches in the transformer sub-circuit have also decreased. In order to simulate a
circuit with much smaller time constants in a stable manner, the integration time step
of the simulation also needs to be decreased. By decreasing the control time step to
10 ns, we can force the integration time step to be 10 ns as well. Additionally, we can
also decrease the integration time step in the simulation parameters to 10 ns along
with the data storage time step so that the plots of simulation waveforms can capture
all events in the simulation. Once these have been defined, all other computations of
the parameters of the equivalent circuit, the formulation of the matrix equation and
the numerical integration will remain the same as the previous simulations.
To complete the simulation, we need to choose the output bus capacitors C1 ,
C2 and C3 and the load resistors RL1 , RL2 and RL3 . The purpose of these output
bus capacitors are to produce a stable voltage for the load. In order to do so, we
should calculate the maximum possible ripple current in the secondary windings
and the largest possible load so as to determine the ripple in the output voltage
[19, 20]. However, since this is not a detailed converter design, we can choose the
output bus capacitors to be 100 µF each. The load resistors will typically resemble
the current drawn by the particular control circuit that is supplied by the particular
output. Again, for the sake of simplicity, we can choose all three load resistors to be
10  each.
5.7 Simulation of a Flyback Converter 197

Fig. 5.24 Output voltages of the flyback converter

Figure 5.24 shows the output voltages vo1 , vo2 , and vo3 across the capacitors
C1 , C2 and C3 in Fig. 5.23. After initial transients, the output voltages settle
approximately to 5 V, 3.3 V and 12 V. The output voltages can be seen to have
a ripple component. This ripple component can be decreased by increasing the
value of the capacitors C1 , C2 , and C3 . The reader is encouraged to zoom into the
simulation result and examine the fall of the output voltages when the power device
S is conducting and the rise of the output voltages when S stops conducting.
Figure 5.25 shows the current ip in the primary winding and the currents is1 , is2
and is3 in the secondary windings for one switching time period. When the power
device is conducting, ip is increasing while the currents is1 , is2 and is3 are zero.
When the power device stops conducting, ip is zero while the currents is1 , is2 and
is3 are flowing. The currents fall to zero at different rates as the output voltages
in the three windings are different, and therefore, the load currents supplied by the
windings are different. The currents is1 and is2 fall to zero before the power device
starts conducting at the beginning of the next switching time period. The current is3
however continues to flow until the end of the switching time period. As a result,
the current ip is seen to start from a non-zero value as in the beginning of the next
switching time period; the transformer core flux has not completely fallen to zero.
This mode of operation is called the continuous conduction mode. When all three
secondary currents fall to zero before the start of the next switching time period, the
converter is said to operate in discontinuous conduction mode.
Figure 5.26 shows the flux linkages of the primary winding for the same time
interval as in Fig. 5.25. The flux linkage increases when the power device is
conducting, as the transformer is being magnetized by the input voltage. The flux
198 5 Applications of Transformers

Fig. 5.25 Primary and secondary winding currents

Fig. 5.26 Flux linkages of the primary winding

5.8 Conclusions 199

linkage decreases when the power device stops conducting, as the energy in the
magnetic field is transferred to the loads in the secondary windings. It should be
noted that this is the flux linkage of the primary winding, which is expressed as

ψ1 = L1 ip + M12 is1 + M13 is2 + M14 is3 = Np φ (5.23)

With this simulation of the flyback converter, we conclude on the simulation

of high frequency transformers. Though the behaviour of transformers varies with
the operation strategy of power converters, it is always possible to use the basic
laws of physics. Subsequently, one can use the simulation model of the transformer
with basic modifications in the specifications of the transformer. The reader is
encouraged to use the simulation model presented for simulating other power
electronic converters such as the forward converter or the push-pull converter.
With this simulation, we can also conclude this chapter on the applications of
transformers in practical engineering systems.

5.8 Conclusions

The use of transformers in electrical engineering is too vast and varied to cover in a
single book. However, a power engineer will deal with both single-phase and three-
phase transformers as he or she works on projects. The purpose of this chapter was
to provide a set of simulation examples that a power engineer can use and further
extend as required by the application. In this chapter, we introduced the concept
of three-phase systems and how a three-phase transformer can be conceived. A
simple solution to achieving a three-phase transformer is to use three single-phase
transformers. However, in this chapter, we have provided along with the basics
of three-phase systems, a glimpse into the possible construction of a three-phase
transformer and how that can be translated to a simulation model.
Three-phase systems differ widely in their application, and this also requires
three-phase transformers to have varied construction and connection arrangements.
In this chapter, we have introduced some of the most popular connection arrange-
ments of three-phase transformer windings—the star and delta connections. We
have also described the significance of these connections with respect to the
requirements of the three-phase system. We have described how the star connection
provides a neutral wire and terminal, which are essential in distribution systems for
domestic applications where single-phase appliances are connected between a phase
and neutral. Furthermore, we have described how three-phase domestic distribution
systems can be unbalanced due to the demand of loads divided into the three phases.
The simulation models in this chapter are mere extensions of the simulation
models of transformers already presented in the previous chapter. The manner in
which the simulation models have been used with minimal modifications points to
how flexible and scalable the models are. By translating a potential construction
of a three-phase transformer to the change required in the simulation model, we
200 5 Applications of Transformers

have provided the reader with an approach that he or she can take if a special
transformer needs to be modelled. This will be of use if the reader wishes to simulate
transformers for applications such as High Voltage DC (HVDC) transmission
systems where the transformers can be significantly modified.
The simulation results presented in the chapter describe not only the basic
performance of three-phase transformers by taking the example of star–star and
delta–star transformers, but also enable us to understand some of the trickier
concepts in three-phase systems. We had considered the simulation of an unbalanced
domestic distribution system fed by a delta–star transformer. We had examined
how the zero sequence component in the star connected transformer windings will
circulate within the delta connected windings. This approach will enable the reader
to attempt to understand the circulating currents that can appear in systems with
multiple transformer windings that may occur in applications such as cascaded
converters.
With this chapter, we will bring the main contents of this book to an end. The
purpose of these chapters is to enable the reader to not only simulate magnetic
components, but also to dissect them and understand the phenomena that enable
their operation. Though transformers are considered basic machines since they do
not rotate, systems that contain transformers can be fairly complex and without
“looking under the hood,” an engineer may not fully understand every mode of
operation of the system. Therefore, with the repeated use of the basic laws of
physics, an engineer can simulate and analyse systems that he or she may encounter
in their professional careers. The advantage of returning to the basic laws of physics
is that an engineer need only to remember very basic laws that he or she has learned
as a high-school student to approach any engineering system.
Chapter 6
Conclusions

6.1 Usefulness of the Book

As we conclude this book, it is worthwhile to reflect on what this book might have
achieved and also potential pit-falls that a reader might want to avoid. Writing this
book has helped me to re-learn many of the basic concepts of electromagnetism. At
the same time, it has also made me aware of how vast this domain is, and how foolish
it would be for any single book to even attempt to completely cover it. Therefore,
this section will talk about how a reader can use the contents of the book, and cases
where it might be advisable for a reader to not use the book.
In the introduction of the book, I had narrated my own experiences and why I
decided to write this book. I had never bothered to learn about the inner working
or design of magnetic components until I was faced with the task of approving a
transformer design submitted to me. It was then that I realized that as an electrical
engineer, I could not afford to not understand how magnetic components are
designed. In order to understand how they are designed, I needed to understand
how they work. To understand how magnetic components work, all one has to do is
revise the basic laws of magnetism, which one learns in high school. In reality, this
entire book has been merely a revision of high school physics.
In this book, I had used a few of the basic laws of physics related to electromag-
netism. And here, it is important to emphasize the word “few.” Magnetism is a vast
domain and there are numerous laws to understand the finer detail of magnetic fields.
This book has not considered the complexity of magnetic fields. The magnetic fields
have been assumed to be uniform in most cases. This is rarely ever the case. Non-
uniformity in magnetic fields and effects such as fringing occur fairly commonly
in every magnetic component. However, these effects will not drastically change
our results, especially since we have considered static magnetic components. In the
case of rotating magnetic machines, non-linearity in the magnetic field can result in
increased vibration, which might not be acceptable.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 201
S. V. Iyer, Modeling and Python Simulation of Magnetics for Power Electronics
Applications, https://doi.org/10.1007/978-3-030-96768-0_6
202 6 Conclusions

The reader must note that this book has not dealt with the advanced topics in
electromagnetism. If a reader does need an accurate map of the magnetic field in his
or her design or analysis, the reader should seek specialized software that can map
the magnetic field of an electromagnet in detail. Such software is used in machine
design by researchers who are working on improvements to motors and generators.
Such applications are quite common in the domain of electric vehicles and wind
turbines, where we are constantly pushing the boundaries in terms of increasing the
range of vehicles or increasing the capacity of offshore wind turbine generators.
One of the objectives of this book is to demystify the role of magnetism in the
operation of magnetic components that electrical engineers use on a regular basis.
This could be the case of a power supply designer that uses high frequency inductors
and transformers, or a power system engineer that might be working at a substation
with a high voltage transformer. The purpose of this book is to go beyond the very
basic equations that most electrical engineers use to analyse circuits. Though the
equations and equivalent circuits that are commonly used are extremely useful,
mathematical equations alone do not help in understanding the inner workings of
any system. For that purpose, this book uses a theory-and-simulation approach. In
this book, the basic laws of physics are described and these are then compared with
simulation results.
It is also fairly important to emphasize what this book does not suggest a reader
to do. The book does not recommend a reader to build simulation models from
scratch for every purpose. Simulations are a great way to learn, and in some cases,
when simulation models do not exist, we are forced to build our own. However,
building simulation models from scratch when such models exist is simply re-
inventing the wheel. Not only is that a waste of time, it is important to remember
that when a scientific group releases a simulation model, a great deal of research
has been invested in building that model. As already stated, physical phenomena
surrounding any particular application may be quite complex and in order to build
a mathematical model, one may need to use several laws and relationships. As
an example, if one wishes to simulate the impact of saturation on a transformer,
the effect of saturation may be fairly complex and to accurately simulate it, it is
advisable to look for specialized models. To attempt to build a custom model might
not only be an unnecessary waste of time, but may also yield inaccurate results that
could completely derail a project.
To conclude, this book should help an electrical engineer understand how
magnetic components work. It should also provide an insight into how to combine
the basic laws of physics with electrical engineering, to produce functional simu-
lation models. However, do not get too enthusiastic and attempt to build complex
simulation models using basic laws. If one needs to model complexity, it is best to
place our faith in the work of experts. Please do remember—moderation is always
the best approach, whether it is for drinking alcohol or simulations.
6.2 A Summary of the Course Contents 203

6.2 A Summary of the Course Contents

In this section, I will describe the main highlights of each chapter of this book. I will
also describe what are the main learnings from each chapter and how one should
perceive the lessons learned from every chapter. This book has dealt with a very
narrow topic in electrical engineering, and most texts on basic electrical engineering
would normally dedicate just a chapter to magnetics.
In Chap. 1, I have provided an introduction that described the need for the book
and the motivation for writing this book. This book was on a list of topics that I
had marked off when I was working in industry as one topic that needed detailed
documentation, and more importantly, written and presented in a manner that is
understandable to a young engineer. While writing this book, I kept in the back of
my mind my own experiences and frustrations when trying to understand magnetics.
During my interaction with other electrical engineers, my own experiences were not
isolated or even rare. Since the purpose of the book was to build an understanding
of magnetics from the basic laws of physics, the introduction covers these basic
laws that will be used in the book. The laws are Kirchhoff’s circuit laws, Ampere’s
Law, Faraday’s Law and Lenz’s Law. One would learn these laws in high school.
The purpose of describing them in the introduction was to provide a quick reference
for the reader, and also to specifically describe how these laws are applicable in
electrical engineering.
In Chap. 2, I have started with analysis and simulation of inductors. As we start
with a simple magnetic component, it is much easier to understand how to apply the
basic laws of physics in understanding how inductors behave and how we can build a
simulation model for an inductor. We introduce the concept of a magnetic circuit and
describe how a magnetic circuit is analogous to an electric circuit. We examined a
few sample inductor cores and represented the magnetic circuits for each one, while
also describing how these can be solved. With this theory, we presented a simulation
model of an inductor. We simulated several inductors, and in each simulation, we
gradually increased the complexity of the inductor construction. We analysed the
simulation results to understand how the inductor behaved at a deeper level than
merely the L didt equation. Most importantly, in this chapter, we had built that bridge
between the world of magnetism and electrical engineering.
In Chap. 3, we had begun with the concept of magnetic coupling. We had shown
with the simple example of two coils wound on the same core, how the flux produced
by one coil links the other, therefore, laying the foundation of a magnetic link
between two circuits that had no electrical link. We had examined several cases of
these two coils to fully understand the nature of magnetic coupling, while expressing
each case mathematically in order to explicitly describe how each case adheres to
the basic laws of physics. We introduced the concept of mutual inductance and the
dot polarity convention to depict the nature of magnetic coupling. We then extended
the simulation model of an inductor to include a magnetically coupled inductor.
Using simulations, we examined the nature of magnetic coupling and analysed the
results to compare them with our theoretical discussions.
204 6 Conclusions

In Chap. 4, we began with the discussion on transformers by describing some

of the broad applications of transformers in electrical engineering. We stated that
transformers are merely magnetically coupled coils that have been so designed
to achieve the maximum possible transfer of power from one winding to another.
We described how we can translate the machine specifications of a transformer
to an equivalent circuit. We also described how the parameters of the equivalent
circuit can be estimated. Subsequently, this equivalent circuit can be expressed as
magnetically coupled coils, and therefore, the simulation model in Chap. 3 can be
used. We simulated a few simple transformers and analysed the results to examine
the working of the transformer in greater detail, by examining the magnetizing
current, and the effect of turns ratio on the transformation of voltage and current.
We had shown how the transformer model is flexible and scalable, and can be used
to simulate a transformer with several windings.
In Chap. 5, we had described how the simulation model of the transformer
presented in Chap. 4 can be used in practical scenarios. We had simulated three-
phase transformers that would be useful for a power engineer working in distribution
systems, and a high frequency transformer in a flyback converter that would be
useful for a power electronics engineer working in power supply design. We
show how the simulation models in Chap. 4 can be easily extended to simulate a
three-phase transformer whether the windings are connected in star or delta. We
simulated a star–star transformer and a delta–star transformer. We also simulated
an unbalanced four-wire load fed by a delta–star transformer to describe how the
concept of circulating zero sequence components can be simulated as well. By
simulating a high frequency transformer for a flyback converter, we show how the
simulation of a high frequency transformer is merely a change in the transformer
specifications. By describing in detail the working of the flyback converter, we
show how even in the context of a non-linear power converter, the operation of
the transformer can still continue to be interpreted by the basic laws of physics.

6.3 The Difference in Approach

If you are reading this conclusion chapter after reading the rest of the book, it would
be fairly clear that the approach to writing this book is completely different from
that of a conventional text or for that matter other technical books. This approach
in writing this book, and also used in previous books, is in fact my philosophy
behind writing books and teaching engineering as an online teacher. In this section,
I would like to describe why I chose this approach and what I hope to achieve that
the conventional approach might not be able to.
As already stated, what motivated me to write this book has been my frustration
with most of conventional books in helping young engineers understand basic
concepts. In reality, most textbooks are written as accompanying material for
students who are enrolled at a university. It is assumed that a teacher would help
a student through the process of raising doubts, and solving assignments. However,
6.3 The Difference in Approach 205

as we are well aware, there has been an acute shortage of teachers all over the
world. The increasing opportunities in engineering have resulted in fewer qualified
engineers choosing to become teachers. For that matter, I too had declined teaching
positions following my PhD and chose to remain in industry. Therefore, to assume
that all engineering students will receive the guidance of a qualified and motivated
teacher is now no longer true. Books nowadays will have to be written such that
students can learn independently, rather than need a teacher.
Which brings us to the question—how do we learn? This question falls more
under the domain of psychology and neuroscience than engineering. However,
almost all of us can attest to the fact that our best learning experiences have
been either visually, or through a series of questions and answers or by a process
of experimentation. What do we mean by a visual learning experience? A visual
learning experience is when we perceive a physical phenomenon as occurring right
before our eyes. When trying to learn topics such as electricity and magnetism, the
greatest challenge is that these are not phenomena that can be seen unless there is
a mechanical output as in the case of a motor. Therefore, to understand magnetism,
we need to visualize the magnetic field.
In this book, by understanding the magnetic field produced by an inductor,
a coupled set of coils or a transformer, we are trying to visualize the magnetic
field produced. In a conventional text, one would expect merely equations or at
the most equivalent circuits. However, most of us would not directly be able to
form the connection between a circuit or an equation and the underlying physical
phenomenon. Furthermore, when one considers different scenarios and possibilities,
that helps in understanding a phenomenon. In this book, whether it is for a mere
inductor or coupled inductors or transformers, the book describes many different
possibilities that help to improve the learning experience for the reader.
Simulations are being rapidly adopted nowadays as a learning tool, and this book
utilizes simulations to a great extent. In this book, we go back and forth between
simulations and theory. We simulate several different variations of a scenario and use
our theory to interpret the results. Moreover, by creating our own simulation models,
we can make modifications and attempt to simulate special conditions. This “poking
around” is greatly beneficial to a student, especially since a simulation carries no risk
of a circuit being physically damaged.
Finally, the target audience of this book is not merely students of electrical
engineering, but also practising engineers, who have regular jobs in the power
industry. Therefore, this book has attempted to break the topic of magnetism in
electrical engineering in such a manner that it can be read and understood by
someone who is not a full-time student. With the minimal use of equations, and
simulations and discussions to describe concepts, this book attempts to project
magnetism in an easily understandable manner to a varied audience.
206 6 Conclusions

6.4 Scope for Future Work

To conclude this book, I would like to talk about the road ahead in this last section
of the last chapter. This will be my third book ever since I decided to become a
teacher/author. As a teacher and an author, I draw from my own experiences from
when I was an undergraduate student, a graduate student and a researcher employed
in a company. I teach and write about topics that I found I needed to learn about and
struggled to understand. In the words of the great Richard Feynman—“if you want
to master something, teach it!”
Becoming a teacher has forced me to become a student again. Writing this book
has made me go back and relearn concepts that I had either long forgotten or had
taken for granted. Writing an equation over and over again had made me stop
thinking—why is this equation the way it is? Could it not be another way? And
if not, why not? In this book, while trying to explain magnetism, I have used this
questioning approach—of examining different scenarios and uncovering potential
contradictions. In most cases, it is always possible through a process of questioning
and seeking for answers, to apply basic laws and arrive at the final results.
For the reader, I would suggest that if you find the contents of this book
interesting, look at the internet homepage of the project Python Power Electronics:
https://www.pythonpowerelectronics.com/
A reader can find vast resources on various topics in power electronics. This
project is ongoing and continuous, and I add new material almost every week. Most
importantly, most of the content related to this project is free and open source in
order to encourage community engagement and support. The objective is to achieve
the level of collaboration and growth that is present in the software community
through open source content and good documentation.
References

1. V. Valkenburgh, V. Valkenburgh, and Nooger & Neville, Basic Electricity. Prompt

Publications, 1992, no. Bd. 1–5. [Online]. Available: https://books.google.de/books?id=
vmg1UKsTntAC
2. C. R. Paul, Fundamentals of Circuit Analysis, 1st ed. John Wiley, 2000.
3. S. J. Chapman, Electric Machinery Fundamentals, 5th ed. McGraw-Hill Education, 2011.
4. A. E. Fitzgerald, C. Kingsley, and S. D. Umans, Electric Machinery, 6th ed. McGraw-Hill,
2005.
5. D. A. Lowther and P. P. Silvester, Computer-Aided Design in Magnetics, 1st ed.
Springer-Verlag Berlin Heidelberg, 1986.
6. P. Smith and D. Pollard, “The role of computer simulations in engineering education,”
Computers & Education, vol. 10, no. 3, pp. 335–340, 1986. [Online]. Available: https://www.
sciencedirect.com/science/article/pii/0360131586900023
7. A. Azemi and E. Yaz, “Pspice and matlab in undergraduate and graduate electrical engineering
courses,” in Proceedings of 1994 IEEE Frontiers in Education Conference - FIE ’94, 1994,
pp. 456–459.
8. C. Canizares and Z. Faur, “Advantages and disadvantages of using various computer tools in
electrical engineering courses,” IEEE Transactions on Education, vol. 40, no. 3, pp. 166–171,
1997.
9. D. Pederson, “A historical review of circuit simulation,” IEEE Transactions on Circuits and
Systems, vol. 31, no. 1, pp. 103–111, 1984.
10. W. Hurley and W. Wölfe, Transformers and Inductors for Power Electronics: Theory, Design
and Applications, 1st ed. Wiley, 2013.
11. A. Witulski, “Introduction to modeling of transformers and coupled inductors,” IEEE
Transactions on Power Electronics, vol. 10, no. 3, pp. 349–357, 1995.
12. G. Zhu, B. A. McDonald, and K. Wang, “Modeling and analysis of coupled inductors in power
converters,” IEEE Transactions on Power Electronics, vol. 26, no. 5, pp. 1355–1363, 2011.
13. J. Winders, Power Transformers: Principles and Applications, 1st ed. CRC Press, 2002.
14. E. I. Amoiralis, M. A. Tsili, and A. G. Kladas, “Transformer design and optimization: A
literature survey,” IEEE Transactions on Power Delivery, vol. 24, no. 4, pp. 1999–2024, 2009.
15. R. Petkov, “Optimum design of a high-power, high-frequency transformer,” IEEE Transactions
on Power Electronics, vol. 11, no. 1, pp. 33–42, 1996.
16. M. F. Barnothy, Biological Effects of Magnetic Fields, 1st ed. Springer US, 1964.
17. L. Anderson, “Biological effects of magnetic fields: laboratory studies,” in Proceedings of
the 20th Annual International Conference of the IEEE Engineering in Medicine and Biology
Society, vol. 6, 1998, pp. 2791–2797.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 207
S. V. Iyer, Modeling and Python Simulation of Magnetics for Power Electronics
Applications, https://doi.org/10.1007/978-3-030-96768-0
208 References

18. C. R. Paul, Introduction to Electromagnetic Compatibility, 2nd ed. John Wiley, 2005.
19. R. Erickson, M. Madigan, and S. Singer, “Design of a simple high-power-factor rectifier
based on the flyback converter,” in Fifth Annual Proceedings on Applied Power Electronics
Conference and Exposition, 1990, pp. 792–801.
20. K.-W. Ma and Y.-S. Lee, “An integrated flyback converter for dc uninterruptible power
supply,” IEEE Transactions on Power Electronics, vol. 11, no. 2, pp. 318–327, 1996.
21. D. Kothari and I. Nagrath, Power System Engineering, ser. McGraw Hill Education. Tata
McGraw-Hill, 2008. [Online]. Available: https://books.google.de/books?id=yPPYGD88UR0C
22. R. A. Serway, J. W. Jewett, and V. Peroomian, Physics for Scientists and Engineers, 9th ed.
Brooks and Cole, 2012.
23. A. Cunningham and N. Jardine, Romanticism and the Sciences. Cambridge University Press,
1990.
24. R. Becker, Electromagnetic Fields and Interactions. Dover Publications, 1982.
25. H. A. Rowland, “On the general equations of electro-magnetic action, with application to a
new theory of magnetic attractions, and to the theory of the magnetic rotation of the plane of
polarization of light,” American Journal of Mathematics, vol. 3, no. 1, pp. 89–96, 1880.
26. V. C. Valchev and A. V. den Bossche, Inductors and Transformers for Power Electronics,
1st ed. CRC Press, 2005.
27. B.-K. Chen and B.-S. Guo, “Three phase models of specially connected transformers,” IEEE
Transactions on Power Delivery, vol. 11, no. 1, pp. 323–330, 1996.
28. A. D. Theocharis, J. Milias-Argitis, and T. Zacharias, “Three-phase transformer model
including magnetic hysteresis and eddy currents effects,” IEEE Transactions on Power
Delivery, vol. 24, no. 3, pp. 1284–1294, 2009.
29. D. Dolinar, J. Pihler, and B. Grcar, “Dynamic model of a three-phase power transformer,”
IEEE Transactions on Power Delivery, vol. 8, no. 4, pp. 1811–1819, 1993.
30. S. Chimklai and J. Marti, “Simplified three-phase transformer model for electromagnetic
transient studies,” IEEE Transactions on Power Delivery, vol. 10, no. 3, pp. 1316–1325, 1995.
31. J. Pedra, L. Sainz, F. Corcoles, R. Lopez, and M. Salichs, “Pspice computer model of
a nonlinear three-phase three-legged transformer,” IEEE Transactions on Power Delivery,
vol. 19, no. 1, pp. 200–207, 2004.
32. J. Grainger and W. Stevenson, Power Systems Analysis, 1st ed. McGraw-Hill Education, 1994.
33. A. I. Zverev, Handbook of Filter Synthesis. Wiley-Interscience, 2005.
34. F. J. Taylor and A. Williams, Electronic Filter Design Handbook, 4th ed. McGraw-Hill
Education, 2006.
35. N. Mohan, T. M. Undeland, and W. P. Robbins, Power Electronics: Converters, Applications,
and Design, 3rd ed. Wiley, 2002.
36. K. Atkinson, An Introduction to Numerical Analysis, 2nd ed. Wiley, 1991.
37. M. Brkovic and S. Cuk, “Input current shaper using cuk converter,” in [Proceedings]
Fourteenth International Telecommunications Energy Conference - INTELEC ’92, 1992, pp.
532–539.
38. E. Santi and S. Cuk, “Comparison and design of three coupled inductor structures,” in
Proceedings of IECON’94 - 20th Annual Conference of IEEE Industrial Electronics, vol. 1,
1994, pp. 262–267 vol.1.
39. S. Axler, Linear Algebra Done Right, 3rd ed. Springer International Publishing, 2015.
40. A. Morched, L. Marti, and J. Ottevangers, “A high frequency transformer model for the emtp,”
IEEE Transactions on Power Delivery, vol. 8, no. 3, pp. 1615–1626, 1993.
41. E. Dallago, G. Sassone, and G. Venchi, “High-frequency power transformer model for circuit
simulation,” IEEE Transactions on Power Electronics, vol. 12, no. 4, pp. 664–670, 1997.
Index

A saturation, 34, 35
Ampere–Maxwell law, 11 saturation zone, 37
Ampere’s Law, 12, 13, 24, 56, 203 without saturation, 36
Applications of transformers
power converter forces, 162
power electronic applications, 162 C
power supplies, 162 Circuit topology, 134, 135
power system, 161 Closed-loop mechanism, 20
simulation model, 161 ControlledVoltageSource, 43
single-phase applications, 161 Coupled coils
star–star and delta–star transformers, 161 dot polarity, 81
three-phase systems (see Three-phase flux linkage equations, 82
systems) inductors, 81
three-phase transformers, 161 limb coil, 82
mutual inductances, 80
sense of winding, 81, 83
B set of inductors, 83
Backward Euler integration method, 44 simulation model, 80
Backward Euler’s method, 104 wound, 82
B–H relationship Ćuk converter
coercive magnetic field strength, 36 carrier waveform, 97
curve-fitting techniques, 35 circuit equations, 93
drawback, 33 component, 96
electromagnet, 35 conventional buck-boost converter, 93
flux density, 35 diode D, 94
ideal linear characteristics, 35 electrical engineers, 93
linear relationship, 34 flux linkage equation, 95
magnetic circuits, 33 gate signals, 97
magnetic field, 36 independent isolated electric circuits, 93
magnetic flux equation, 33 inductor currents, 97, 98
magnetic materials, 35 matrix equations, 99
power electronic applications, 34 mutual inductance, 98, 99
retentive flux density, 36 output voltage, 97, 98

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 209
S. V. Iyer, Modeling and Python Simulation of Magnetics for Power Electronics
Applications, https://doi.org/10.1007/978-3-030-96768-0
210 Index

Ćuk converter (cont.) Flyback converter, 204

power supply, 93 Faraday’s Law, 192
self-inductance, 98 grid frequency transformer, 189, 192–194
switch S, 94 Kirchhoff’s Voltage Law, 192
transfer energy, 93 Lenz’s Law, 191
magnetic field, 193
magnetic flux, 191
D MOSFET, 189
Delta–star transformer, 204 open-loop switching strategy, 190
power converters, 190
E power device, 191
Electrical engineer, 202 power electronic converters, 189
Electrical engineering, 203–205 primary and the secondary currents, 193
appliances and systems, 3–4 primary winding, 193
basic laws of physics quantities, 190
analysis of simulation, 7 secondary winding, 193
conductor, 12, 13 simulation
electromagnetic induction, 15 advantages, 195
electromagnetics, 11 behaviour of transformers, 199
electromotive force (emf), 15 closed-loop control scheme, 195
flux lines, 14, 17 control and operation strategies, 195
flux linkage, 17 flux linkages, 197, 198
magnetic field, 11, 13, 16, 17 load resistors, 196
magnetic flux, 16 multiple outputs, 194
mathematical equations, 6 open-loop control scheme, 195
mathematical expressions, 7 output voltages, 197
parameter, 7 power electronic converters, 195
static magnetic components, 7 primary winding, 197
consumer appliances, 6 push-pull converter, 199
dc–dc converters, 4 secondary windings, 197
electrical machine, 22 specifications, 196
electric motors, 5 topology, 195
electric vehicles (EVs), 5 transformer applications, 194
flyback converter, 21, 22 sine waveform, 189
inductor, 21 topology, 189
industries, 5 transformer windings, 190, 191
magnetic field, 4
mobile phones, 4
simulation models, 19–21 H
smart phones, 4 High frequency transformers
static magnetic components, 6 B-H curve, 187
three-phase transformers, 21 eddy currents, 188
transmission lines, 6 electronic appliances, 186
Electricity, 205 Faraday’s Law, 188
Electric vehicles (EVs), 5 flux, 186
Electromagnetic Compatibility (EMC), 4 frequency, 187
Electromagnetism, 201, 202 magnetic circuit, 189
Equivalent circuit, 168, 174, 196 power electronics, 186, 188

F I
Faraday’s Law, 16–18, 25, 26, 37, 45, 49, 52, Inductor, 203
59, 63, 64, 112, 115, 126, 131, 203 air gap, 50, 51
Finite element methods, 27 coil wound, 25
Index 211

control file, 42, 53 Kirchhoff’s Voltage Laws (KVL), 7, 9–11, 27,

ControlledVoltageSource element, 41 43, 45, 63, 64, 67, 115, 131
current, 51, 54
electrical circuit, 41
electrical component, 37 L
electrical engineering, 23 Leakage inductances, 2, 118–120
equivalent magnetic circuit, 53 Lenz’s Law, 18, 26, 37, 49, 59, 64, 67, 68, 203
Faraday’s Law, 53
flux density, 56
fluxes, 52, 54, 55 M
flux linkages, 26, 51 Magnetic circuit
hand-wound inductors, 56 air gap, 30–31
in-built inductor model, 40 basic laws of physics, 59, 104
integral equation, 43 circuit energization, 87, 89
integration, 41 circuit topology, 84, 89, 90
magnetic circuit modelling, 55 coil currents, 87, 88, 90–92
magnetic circuits, 44, 56 (see also Magnetic coil-induced emfs, 87, 88, 90, 91
circuit) coil terminal voltages, 87, 90
magnetic component, 50 coil winding resistance, 89
magnetic flux, 25 coil wound, 30
magnetism, 55 coupled inductors, 60
manufacturers, 23 design, 47
mass-manufactured component, 24 electrical engineering, 59
mathematical model, 39 electric circuits, 33
power electronics applications, 23 electric current, 31
power electronics courses, 24 equivalent circuit, 85
segments, 54 Faraday’s Law, 47
series resistance, 41 flux, 48
simulation model, 24, 38 flux linkages, 45, 86, 104
solid state devices, 23 induced emf produced, 49
test circuit, 39 inductor coil, 59
transformers, 56, 57 iron core, 32
two-terminal component, 37 linear magnetic relationship, 45
VariableResistor element, 41 magnetic component, 84
VariableStorage elements, 54 magnetic coupling, 59, 92 (see also
voltage, 51 Magnetic coupling)
voltage source, 40 magnetic flux, 30
Voltmeter measures, 38 multi-limb core inductor, 46
Inductor model mutual inductance, 60, 84
current-carrying conductor, 28 Output Variables, 85
Faraday’s Law, 27 parameter definitions, 85
magnetic characteristics, 28 power electronic converter, 59
magnetic flux, 30 Python code, 46
magnetic properties of elements, 27 self-inductances, 60, 85, 89
magnetism, 29 simulating coupled coils, 78–80
network equations, 27 simulation, 84
permeability, 28–29 static variable, 47
power electronics, 29 supply voltage, 90, 92
transformer operation, 104
transformers, 48
value of flux density, 50
K VariableResistor elements, 85
Kirchhoff’s Current Laws (KCL), 7, 8, 11, 27, variables, 86
203 VariableStorage elements, 48
212 Index

Magnetic components, 201, 202 KVL equation, 149

Magnetic coupling, 203 mathematical model, 150, 157
basic laws of physics, 65, 72 matrix equation, 154, 155
B–H curve, 72 parameters, 151, 152
coil flux linkages, 100 simulations, 147, 153, 158
coil wound, 62 specifications, 151
core flux, 64, 65, 72 voltage equations, 149
coupled coils, 66, 99 voltages, 151–152
dc voltage, 71 windings, 148
dump resistor, 67, 69 Mutual inductance
electric circuit, 60 basic laws of physics, 72
energy, 68 coil, flux linkage, 76
Faraday’s Law, 60–62, 67, 70 coupled inductors, 72
flux, 60 coupling factor, 78
flux linkages, 61, 65, 67 flux, 75
induced emf, 63–65, 68, 99 flux linkages, 73, 75
Lenz’s Law, 60–62, 69 magnetically coupled coils, 73
magnetic fields, 65 magnetic circuit, 74
matrix, 100–102 magnetic fields, 74
mechanical process, 102–104 mathematical modelling, 74
relative permeability, 70 property of inductance, 73
R-L-C circuit, 66 Right Hand Thumb Rule, 74
self and mutual inductances, 100 self-inductances, 75, 77, 78
series impedance, 70 transfer of energy, 76
simulation model, 65
simultaneous equations, 99
transformed current, 69 N
two coils wound, 61 Newton’s Third Law of Motion, 18
winding resistance, 62 No-load current, 143
Magnetic field, 201
Magnetic field strength (MMF), 34
P
Magnetism, 107, 205
Power engineering, 18, 19
basic network laws, 1
Power supply designer, 202
dc–ac converter, 2
Pulse Width Modulation (PWM), 96, 190
electrical engineering, 2, 3 (see also
Python Power Electronics, 20, 39, 42, 45, 206
Electrical engineering)
electrical engineers, 1, 3
electrical machines, 1 R
magnetic components, 1, 2 Resistance–inductance (R–L), 23
magnetic fields, 1, 3 Resistance–inductance–capacitance (R–L–C),
power community, 1 23
R&D department, 2 Right Hand Thumb Rule, 11–14, 24, 25, 28,
software, 1–2 56, 61, 68
transformer component, 2 Runge–Kutta Fourth Order method, 44, 104
transformer design, 2–3
Magnetizing inductances, 120, 170
Magneto motive force (MMF), 15, 28, 33, 56 S
Multi-winding transformers Simplified transformer equivalent circuit, 119
in circuit, 147–148 Simulation models, 107, 202
computations, 151, 157 Sports Utility Vehicle (SUV), 5
currents, 153 Star–star transformer, 204
equivalent circuit, 158 Static variables, 123
inductance, 149 Step-down transformer, 147
Kirchhoff’s Voltage Law, 148–149 Step-up transformers, 147
Index 213

Step up transformer source star and delta connections, 168

and load currents, 139–141 star–delta/delta–star, 178–179
and load voltages, 139–141 star–star transformer, 168, 169, 183
three-limbed core, 171, 172
three-phase supply, 179
T transformer construction, 172, 174, 175
Theory-and-simulation approach, 202 transformer equivalent circuit, 168
Three-phase systems transformer power and voltage rating,
advantage, 200 180
balanced system, 163 transformer specifications, 169–170
basic laws of physics, 200 voltage sources, 169
computations, 164 Voltmeter, 169
delta-star transformers, 200 winding resistances, 174
equipment and appliances, 162 zero sequence component, 183–184
generator stator/rotor, 162 single-phase appliances, 199
horizontal and vertical axes, 164–165 single-phase transformers, 168
line voltage, 166 star–star transformers, 200
multi-phase machine, 162 three-wire and four-wire systems, 165, 166
neutral terminal, 165 transformer windings, 165, 166
neutral wire, 165 transmission systems, 165
phase currents, 167 voltages, 163
phase voltages, 167 Transformers, 205
phasors, 163, 164 appliances and systems, 111
RMS line voltages, 167–168 circuit, 114
RMS magnitudes, 163 coils, 112
simulation model coupling factor, 118
basic manufacturer specifications, 175 electrical machines, 109
basic power and voltage ratings, 180 electrical power transmission systems, 110
construction of transformers, 171 electric grid, 110
coupling factor, 173 equivalent circuit, 117–118, 120
delta–star transformer, 179, 181, 183 inductance, 113
electrical sense, 171 parameters, 114, 137
four-wire system, 179 principle, 111
isolation, 171 renewable energy generation, 109
leakage flux, 172 RMS value, 115
mutual inductance, 172 simulating
neutral current, 185 circuit, 120
numerical integration, 174 control code, 123
phases wound, 172 core flux, 127
power and the voltage ratings, 170–171 cosine waveform, 127
primary currents, 176, 178 integration offset, 126
primary line currents, 180, 182 mathematical model, 123
primary line voltages, 175–177, 180, parameters, 121, 122
181 results, 126
primary winding, 183 storage elements, 123
reader, 186 transformer windings, 125
RMS magnitudes, 168, 175, 177 voltage sources, 121
secondary currents, 176, 178 windings, 122, 127
secondary line currents, 180, 182, 184 simulation results, 139
secondary line voltages, 175–177, 180, specification, 113
181 step-up transformers, 137
secondary windings, 183 circuit, 108
self-, leakage and mutual inductances, uses and applications, 109
173–174 voltage and current, 109
214 Index

Transformers (cont.) Transmission systems, 165

voltage levels, 110, 111 Turns ratio, 134, 136
winding, 116
Transformer simulation model, 139
Transformer winding currents, 126 U
ampere, 133 Universal Serial Bus (USB), 4
circuit topology, 134 User-defined component, 20
core flux, 132, 133
energy conservation, 133
load resistor, 129
V
magnetizing current, 134
VariableResistor, 43
magnitude, 135, 136
Voltmeters, 123
at no load, 130, 132
simulation, 129
source and load currents, 130
voltage sources, 136 W
Transformer winding voltages, 125, 128 Winding voltages, 141, 142