International Conference on Innovations in Power and Advanced Computing Technologies [i-PACT2017]

A Review on Current Control Techniques for Inverter

for Three Phase Grid Connected Renewable Sources
Priyanka Paikray1 Ritesh Dash2
M.Tech Student Research Scholar
School Of Electrical Engineering School Of Electrical Engineering
KIIT, University

Dr. Sarat Chandra Swain3 Dr. P.C Panda4

Associate Professor Professor,
School Of Electrical Engineering School Of Electrical Engineering
KIIT, University KIIT, University

Abstract- Renewable based power generation system and their (PCC) through inner current control loop and to extract the
grid interconnection throughout the world. Due to large maximum amount of power from the renewable generator like
penetration of renewable sources into the grid, maintenance of PV, Wind and small Hydroelectric plant (SHEp) [3]. In
power quality, grid integrity and protection against islanding comparison to the traditional open loop control algorithm the
detection became a critical issue. One of the most important current control PWM algorithm are many advantages like
aspects in the renewable grid interconnection is the adoption and
protection against injection of peak current, highly dynamic
current control techniques for the inverter. These current
controlled inverters must meet the grid criteria and international characteristic in response to variable load demand,
standard while interconnecting to grid. This paper presents a compensation against DC link voltage and change in grid
review on different types of current control techniques that has parameters such as resistance and reactance [4]. Provision of
been adopted in the industry as per the best practices and state of frequency and voltage support is also an essential parameter of
technology. Various techniques and their control algorithms were inverter control action. This refers to stable operation of
discussed in two categories, like linear controller and non-linear controlled inverter both under steady state and transient
controller. New current control techniques such as Artificial condition of operation. Steady state refers to the change in
Neural Network, Fuzzy logic Controllers were also discussed in characteristics of the controller with low variation in the grid
the paper. Performance analysis of different types of controller
parameters [5,6]. Whereas, the dynamic performance evaluates
were also discussed in this paper for better understanding of the
characteristics of the system. MATLAB/SIMULINK based test how fast a controller may respond to the change in grid
was carried out to the performance of the system. parameters. Some international standards such as IEEE 1547
must be maintained while interconnection the inverter with the
Keywords- Current Source Inverter (CSI), Distributed grid [7,8 ]. Development in the current control strategy is still
Power Generation System (DPGS), Grid Interconnection, Low under the progress however, the new controlled strategies such
Voltage Ride through Capability, Total Harmonic Distortion (THD), as Neural Network, adopted neural network and fuzzy logic
Voltage Source Inverter (VSI).
based controller are some of the digital areas which needs to be
I. INTRODUCTION exploded. Micro Electronics manufacturing system is generally
used for manufacturing the solar cell and hence the efficiency
Most of the power electronic based voltage source PWM
for the polycrystalline cell is about 10-15% and for mono
techniques uses an internal current controller loop. It is used to
crystalline system it is about 9-13% [9]. Commercially
maintain required output at the terminal level from where the
available solar cells are generally of monocrystalline type or
power is to be evacuated [1]. Hence, the performance of the
polycrystalline type [10].
converter depends upon the characteristics of the controller [2].
Renewable energy based distributed power generation requires In this paper, a brief comparison among the different types
some special attention while designing the controller has to of controllers and their characteristics were presented starting
perform three number of task normally maintaining the require from the traditional control method such as bang-bang
DC link voltage at the input side through after control loop, controller (Hysteresis controller), PI controller as discussed
meeting the grid parameter at the point of common coupling above. The basic concept and their characteristics approaches

978-1-5090-5682-8 /17/$31.00 ©2017 IEEE

1
 International Conference on Innovations in Power and Advanced Computing Technologies [i-PACT2017]
were discussed above. The basic concept and their To force the AC current in the load side, it is required to
characteristics approaches were discussed under second part. maintain the DC source at the input side. Converter forces to
Third part discussed about linear controller followed by the inject the current into the grid point the grid voltage is less than
forth part stating non-linear controller and their characteristics. the DC link voltage [17]. However, when the grid voltage
The fifth section contains the detailed characteristics of the reaches at its maximum value the converter into a six step
converter and their discussion part six comprise of conclusion. operation mode where the current controlled is not able to force
the command current. Hence it is required to maintain sufficient
II. BASIC CONCEPTS & CHARACTERISTICS DC link at the input [18].

The purpose of the current controlled technique is to force Based on the above performance characteristics the basic
the injected AC load current to follow the reference current or requirement of a current control techniques can be summarized
signal. Triggering pulse of the current controller can be as (1) Maintenance of sufficient DC link at the input side. (2)
evaluated by comparing the command reference current with No phase and amplitude error at the point of tracking. (3)
that of the instantaneous value of the controller current. So Constant switching frequency for the entire operation of the
current controlled techniques usually compensates the converter. (4) Implementation of power filters for regulating
evaluated error along with the required modulation index for harmonics.
the gates.
In addition to the above basic criteria the dynamic response
Characteristics of the converter can be measured through of the current controller can be verified through dead time, rise
the modulation index, switching and ripple frequency, phase time, settling time, and over shoot factors.
interference effect and DC link of the converter [11].
III. LINEAR CONTROLLERS
Voltage source converter usually produces two different Linear controller consist of conventional voltage type
level for each phase in a three phase in a three phase grid PWM converters or modulation. Linear controller schemes has
connected system. This two level has no effects on the two separated control action like compensation against current
performance of the inverter if hard switching is allowed with error and voltage modulation constant. These linear controllers
no mutual constraints [12, 13]. However during higher are suitable for maintaining constant switching frequency,
modulation index it produces mutual interferences among the optimum switching patterns and for maintaining better DC link
three phase system and their by creating high order harmonics. at the input of the controller. In this paper linear controller
This requires additional filters for maintaining phase comprising stationary PI controller, synchronous PI controller,
synchronisation. If the modulation index becomes greater than PR controller, State feedback and predictive dead beat
one then the two groups may be separated from each other. controllers are discussed.
However, it injects higher order harmonics which can be
compensated through STATCOM, shunt active power filter and (a). Stationary PI Controller:
active power filters [14].
Stationary PI controller uses three PI error compensator
Increase in modulation index increases the voltage to produce three voltage commands for a three phase sinusoidal
harmonics, hence the average current also deviates from the PWM. It uses a comparison techniques where it compares a
instantaneous in a large extent. This ripple current also depends triangular wave with three phase sinusoidal signal to generate
upon the duration of modulation period, the supply voltage, the control signal for inverter switches. One of the
load side parameters and AC voltage [15]. In order to maintain disadvantages of this controller is that the switching time and
a constant ripple current it is required to vary the modulation its operation is generally affected due to feedback operation of
frequency which is again controlled by the modulation index. the current ripple.
This future requires user filter to maintain the required
The integral part present in the PI controller minimises the
switching frequency.
current error and the proportional part is related to the amount
Phase interference is produced due to change in phase of ripple present in the controller. This controller suffers from
angle of inverter and the load between which power is being tracking of amplitude and phase error. Therefore in order to
transfer from inverter to load. This arises when the reduce the error the slope of the command voltage must be
interconnection between inverter neutral point and load neutral maintained below the triangular slope. Presence of multiple
point is not established [16]. This leads to variation in phase crossing of triangular wave also created ripple current error.
current which ultimately depends upon phase voltage to which Therefore to avoid the above problem additional tracking
it is connected and other phases in the vicinity.

2
 International Conference on Innovations in Power and Advanced Computing Technologies [i-PACT2017]
method such as phase locked loop is also used to track the phase controller can be described as follows. The fig-2 & 3 shows the
angle and frequency. current waveform when the renewable generator started.

(b) Synchronous PI Controller: ª kis º

« k p + 0 »
Controlling the DC variables in a grid interconnection s + w2
2

system is the most critical challenge. Industrial drives

GP R (αβ ) =« » ……….(2)
« 0 k »
consisting of precise motor and load requires variable supply k p + 2 is 2
«¬ s +w »¼
system. In such cases d-q control strategies is best suited. d-q
control scheme transfer the abc reference frame into d-q control
signal. This can be achieved in a two stage conversion system
such as step-1 transfers the signal from abc to Įȕ and the second Width of resonance frequency bond can be adjusted through Ki,
step transfer the signal from Įȕ to d-q co-ordinates system. Fig- like smaller Ki results in narrow bond whereas larger Ki results
1 shows the synchronous PI controller. Based on the for large bond.
requirement of active power, DC link can be adjusted to
generate the reference current. Similarly, reactive power x 10
4
Generator-1
demand can be adjusted on set to zero depending upon 1
operation. B-Phase
G-Phase
Gain of the PI controller based on d-q co-ordinate system 0.5
R-Phase
can be adjusted through equation 1 and the current control of

Current
synchronous PI controller as shown in fig-1.
0
ª ki º
«k p + s 0 » -0.5
GPI ( dq ) =« » ……………..(1)
« 0 ki »
kp +
«¬ s »¼ -1
Time
0 0.02 0.04 0.06 0.08 0.1
Where Kp and Ki represents the potential and integral gain of
the controller.
Fig-2: Current Wave Form at Renewable Generator
4
x 10 Generator-2
1
B-Phase
G-Phase
0.5
R-Phase
Current

-0.5

Fig-1: Synchronous PI Controller

Time
-1
(C) PR Controller: 0 0.02 0.04 0.06 0.08 0.1

It is the advance version of PI controller. It does not

require any steady state magnitude or phase error to calculate
the sinusoidal reference current. Advantages of PR controller is Fig-3: Current Wave Form at Renewable Generator
that a large gain around the resonance frequency can be (D) Predictive and Deadbeat Controller:
achieved with the integral controller. The gain of the PR

3
 International Conference on Innovations in Power and Advanced Computing Technologies [i-PACT2017]
This controller forecast the error at the beginning of each x 10
7
Load-1
sampling period. The error is being forecasted based on the 5

current error vector and AC load side parameters. Thus the

Active & Reactive Power

4
voltage vector to be generated by PWM controller is generated
to minimize the forecasted error. When the voltage vector is so 3

assumed to nullify the error at the end of the sample period the 2
predictive controller is often called a deadbeat controller. Here
in fig-4 PLL used because it supply the information of the grid 1
voltage to grid current.
0

-1
0 0.02 0.04 0.06 0.08 0.1
Time

Fig-5: Real & Reactive Power Exchange

Fig-4: Dead Beat Controller with PLL Tracking

IV. NON LINERA CONTROLLER

Non-linear controller has many advantages over the linear
controller in the form of robustness, independent in the load
parameter variation, good dynamics and dynamic property in
tracking the error. Non-linear controllers are of many types
such as Hysteresis controller, Optimized controller, Neural
Network fuzzy controller and DM controller. Fig-6: Hysteresis Controller
(a) Hysteresis Controller:
(b) Optimised Controller:
It compares the reference current and instantaneous
grid current to produce the gating signal for the inverter. The It is based on the space vector analysis of Hysteresis
band of Hysteresis controller defined by minimum and Controller. The boundary of the current error area in the case of
maximum error. This further allows the current to be kept under independent controller with equal tolerance band in each of
the lower and upper limit. Instead of many advantages of the three phases makes a regular hexagon. Based on the area of
converter, it suffers from variation in the load parameter hexagon, optimisation is applies to minimise the voltage vector.
changes and AC voltages. Inherent randomness present in the (c) Neural Networks:
controller limit cycle requires some precise protection system.
The hysteresis controller is shown in fig-6 is contain PLL for
the reference frame generation.

Fig-7 Neural Network based CC

4
 International Conference on Innovations in Power and Advanced Computing Technologies [i-PACT2017]
Neural Network controller is the most advance version of x 10
5

20
non-linear controllers. Parallel processing, robustness and Real B Phase
Real G Phase

learning ability present in the controller makes it suitable for Real R Phase
Reactive B Phase

high speed operation, calculation of error signal and online 15 Reactive G Phase
Reactive R Phase

optimisation are neutralised through parallel processing.

Training each layer of the controller was carried out through 10

back propagation algorithm. No further training of the layer is

Voltage
allowed during the controller operation and how exactly neural 5
network exactly work that is shown in fig-7. Hence the
performance of the controller largely depends upon the quality
of training data and the amount of data. Online training of 0

neural data is allowed where it is required to compensate the

large variation in load parameter. However for fast operation of -5
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1

online training random with change must be applied. time

6
x 10
Fig:-10 Real & Reactive Power Flow
B-Phase
G-Phase
The fig-8 & 9 shows the grid level voltage during single
5
R Phase

4
phase & three phase fault through neural network and these
3 faults occurred after integration of renewable source to the grid
2
side. Fig-10 shows general real & reactive power flow in the
system.
Voltage

0
V. CONCLUSION
-1
A comprehensive review of different types of current
control techniques has been presented in this paper. Grid
-2
interconnection issues of renewable sources were also
-3
addressed in this paper. A fair discussion among the different
-4
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
types of controller were also discussed with their results. From
Time
the literature and result it may be concluded that, PI controller
typically uses d-q reference frame of operation and PR
Fig:- 8 Grid level voltage during single phase fault
controller uses Įȕ reference frame of control action. PR
through ANN
controller has very high dynamic performance with flexibility
10
x 10
4
for selective harmonic compensation. All the controllers were
B Phase
G Phase
tested under different grid fault condition. However, Deadbeat
8 R Phase converter shows good performance under single phase too
ground fault condition which is more common among the
6
distribution lines. However, from overall performance view PR
4 controller shows high dynamic characteristics and harmonic
compensation capacity, especially lower order harmonics as
Voltage

2
compared to others.

VI. ACKNOWLEDGEMENT
0

-2 The author would like to thank School of Electrical

Engineering, KIIT University for providing the necessary
-4
laboratory facility throughout the entire experimental work.
-6
0 0.01 0.02 0.03 0.04 0.05
Time
0.06 0.07 0.08 0.09 0.1
VII. REFERENCE
[1]. IEEE Standard for Interconnecting Distributed Resources
Fig:-9 Grid level voltage during three phase fault with Electric Power Systems - Amendment 1," in IEEE Std
through ANN 1547a-2014 (Amendment to IEEE Std 1547-2003) , vol.,
no., pp.1-16, May 21 2014

5
 International Conference on Innovations in Power and Advanced Computing Technologies [i-PACT2017]
[2]. IEEE standard conformance test procedures for equipment IEEE 9th International Conference on, Coimbatore, 2015,
interconnecting distributed resources with electric power pp. 1-7.
systems - amendment 1," in IEEE Std 1547.1a-2015 [14]. J. Zhang, H. Huang and Y. B. Zhang, "The comparative
(Amendment to IEEE Std 1547.1-2005) , vol., no., pp.1-27, analysis of foreign smart grid top-level roadmaps," Power
May 1 2015. System Technology (POWERCON), 2014 International
[3]. A. Al-Amoudi and L. Zhang, "Application of radial basis Conference on, Chengdu, 2014, pp. 1908-1913
function networks for solar-array modelling and maximum [15]. S. C. Swain, R. Dash, S. M. Ali and A. K. Mohanta,
power-point prediction," in IEE Proceedings - Generation, "Performance evaluation of photovoltaic system based on
solar cell modelling," Circuit, Power and Computing
Transmission and Distribution, vol. 147, no. 5, pp. 310-
Technologies (ICCPCT), 2015 International Conference
316, Sep 2000.
on, Nagercoil, 2015, pp. 1-6.
[4]. N. Nikmehr and S. Najafi Ravadanegh, "Heuristic [16]. Siddhartha Panda, Sarat Chandra Swain, Srikanta
probabilistic power flow algorithm for microgrids Mohapatra, “Design and Analysis of Bacteria Foraging
operation and planning," in IET Generation, Transmission Optimized TCSC based controller for Power System
& Distribution, vol. 9, no. 11, pp. 985-995, 8 6 2015. Stability Improvement”, International Journal of Data
Analysis Techniques and Strategies (IJDATS),
[5]. L. Sandrolini, M. Artioli, U. Reggiani, Numerical method INDERSCIENCE Publishers, Geneva, Switzerland. Doi:
for the extraction of photovoltaic module double-diode 10.1504/IJDATS.2014.066602., Volume: 6, Issue: 4, Dec-
model parameters through cluster analysis, Applied 2014, Pages: 384-406
Energy, Volume 87, Issue 2, February 2010, Pages 442- [17]. R. Dash, S. C. Sahoo, S. M. Ali and A. K. Mohanta,
451, ISSN 0306-2619 "Performance evaluation and comparision of a grid
[6]. J. M. Aller, A. Bueno and T. Paga, "Power system analysis connected photovoltaic system based on solar cell
using space-vector transformation," in IEEE Transactions modelling:-Part-I," 2015 International Conference on
on Power Systems, vol. 17, no. 4, pp. 957-965, Nov 2002. Electrical, Electronics, Signals, Communication and
[7]. K. K. Tse, B. M. T. Ho, H. S. H. Chung and S. Y. R. Hui, Optimization (EESCO), Visakhapatnam, 2015, pp. 1-6.
"A comparative study of maximum-power-point trackers [18]. S. C. Swain, R. Dash, S. M. Ali and A. K. Mohanta,
for photovoltaic panels using switching-frequency "Performance evaluation of photovoltaic system based on
modulation scheme," in IEEE Transactions on Industrial solar cell modelling," 2015 International Conference on
Electronics, vol. 51, no. 2, pp. 410-418, April 2004. Circuits, Power and Computing Technologies [ICCPCT-
[8]. T. T. Madangombe, K. A. Folly and P. Pillay, "Standards 2015], Nagercoil, 2015, pp. 1-6.
and Technical Guidelines for the Interconnection of
Renewable Energy Sources into the Grid (Distribution
Networks)," Power Engineering Society Conference and
Exposition in Africa, 2007. PowerAfrica '07. IEEE,
Johannesburg, 2007, pp. 1-7.
[9]. M. Singh, V. Khadkikar, A. Chandra and R. K. Varma,
"Grid Interconnection of Renewable Energy Sources at the
Distribution Level With Power-Quality Improvement
Features," in IEEE Transactions on Power Delivery, vol.
26, no. 1, pp. 307-315, Jan. 2011.
[10]. A. B. Shitole, H. M. Suryawanshi and S. Sathyan,
"Comparative evaluation of synchronization techniques for
grid interconnection of renewable energy
sources," Industrial Electronics Society, IECON 2015 -
41st Annual Conference of the IEEE, Yokohama, 2015, pp.
001436-001441.
[11]. L. Hadjidemetriou, E. Kyriakides and F. Blaabjerg, "A grid
side converter current controller for accurate current
injection under normal and fault ride through
operation," Industrial Electronics Society, IECON 2013 -
39th Annual Conference of the IEEE, Vienna, 2013, pp.
1454-1459.
[12]. B. Lundstrom, M. Shirazi, M. Coddington and B.
Kroposki, "An Advanced Platform for Development and
Evaluation of Grid Interconnection Systems Using
Hardware-in-the-Loop: Part III -- Grid Interconnection
System Evaluator," Green Technologies Conference, 2013
IEEE, Denver, CO, 2013, pp. 392-399.
[13]. D. S. Saragar and C. L. Bhattar, "Modelling and designing
of LCLR filter for power quality improvement in
microgrid," Intelligent Systems and Control (ISCO), 2015