Aimone Thesis 2015
Aimone Thesis 2015
Aimone Thesis 2015
FLOW PROBLEMS
by
CONNOR AIMONE
Approved by
Research Advisor: Dr. Le Xie
May 2015
Page
ABSTRACT .................................................................................................................................. 1
DEDICATION .............................................................................................................................. 2
ACKNOWLEDGEMENTS .......................................................................................................... 3
NOMENCLATURE ..................................................................................................................... 4
CHAPTER
I INTRODUCTION ............................................................................................... 5
IV CONCLUSION ................................................................................................... 33
REFERENCES ........................................................................................................................... 35
APPENDIX ................................................................................................................................. 36
ABSTRACT
Improving Microgrid Stability Through Voltage Source Converter Based DC Lines. (May 2015)
Connor Aimone
Department of Electrical Engineering
Texas A&M University
Integration of renewable power generation has emerged as a critical challenge facing electric
power systems. Increasing the penetration of such energy sources, while maintaining stability,
will require creative solutions infused with new technology. Advancements in voltage source
converters (VSC) make the device a viable option for high voltage DC (DC) transmission.
Utilization of VSC based DC lines could improve stability performances impeding large wind
farms connected to the grid by improving reactive power support. This thesis introduces a
detailed method for incorporating VSC transmission lines into AC power flow models, and
simplifies the proposed method into a combination of AC components. The simplified model
will benefit grid planning efforts by reducing computation time through smaller Jacobian
matrices, while the detailed method offers insights into the VSC’s operating condition and
analyzed through the generation of PQV capability regions. Comparing the regions for a
conventional AC connection versus a VSC-DC line shows the latter does not necessitate a
1
DEDICATION
2
ACKNOWLEDGEMENTS
Firstly, I would like to thank my parents for their unwavering support and for instilling the value
of education in me. I would also like to express my appreciation to Dr. Le Xie, who has
introduced me to the world of research and guided me through every step of this process.
Additionally, I thank Omar Urquidez for serving as my VSC expert and mentor throughout the
past year. Without the help of these people, my research experience would have been much
3
NOMENCLATURE
AC Alternating Current
DC Direct Current
I Current
R Resistance
4
CHAPTER I
INTRODUCTION
Transmission System
The first phase in the power system topology is the transmission system, which is characterized
by high voltage lines. After power is generated, the voltage is stepped up using transformers
before being transferred to the load. These lines are used to carry power to load centers, because
they minimize the amount of power lost in transfer. By having a high voltage level on the line,
the current is dropped, reducing the I2R losses. This helps to maximize efficiency as energy is
Distribution System
When the power has neared the load center the voltage is stepped down to an appropriate level
for consumer utilization. This portion of the power system is called the distribution system. Since
the voltage demands vary customer to customer, the lines must be connected to various
subsequent substations. The distribution lines typically carry power to consumers in a radial
fashion.
The growth in technology, especially in power electronics such as voltage source converters
(VSCs) and flexible AC transmission system (FACTS) devices, will improve power systems
through their controllability [1]. For example, an MIT study concerning the future of electric
5
grids concluded that the “control flexibility of voltage source converters can improve system
stability and facilitate the integration of remotely located renewable generation [1].” Naturally, a
common technical challenge in transmission and distribution systems of the future is the
models. Our specific contribution lies in how to incorporate a voltage source converter (VSC)
based DC line into power flow analysis models for both transmission and distribution systems.
This task involves accounting for the VSC’s influence on power by including its controllable
parameters into a power flow model. Additionally, a model that functions on both a transmission
and distribution systems should account for increased line losses when the voltage is dropped.
Therefore, a useful model for VSC based DC lines will include VSC parameters which impact
Power system stability requires that a grid keep satisfactory steady state operation under normal
operating conditions and be able to return to equilibrium after subjection to a disturbance. One of
the most crucial aspects of power system stability is voltage stability. In order for the system to
be classified as voltage stable, the bus voltages throughout the network must remain steady
within an approved range [2]. Although this aspect of stability originally only posed problems
for weak systems, as grids become increasingly congested the issue resurfaces [2]. This is
particularly an issue for the US electric grid, as the consumption of our nation’s energy by
electricity generation has grown from 14% in 1959 to 41% in 2009 [1]. Due to this large increase
6
in electricity demand, power engineers have been forced to consider voltage stability in their
Voltage instability occurs when a bus voltage enters a progressive, uncontrollable decline after
the system experiences a disturbance [2]. Typical disturbances include load fluctuations, line
outages, loss of generators, and other contingencies. The upper limit of voltage stability, also
known as the critical point, occurs at the load impedance which causes the receiving voltage to
equal the voltage drop across the transmission line. This point also represents the maximum
transmitted power. Moving beyond this limit signifies a transition into unstable operating
conditions. System response in this region is characterized by an inability to raise the receiving
end power by reducing the load impedance. This failure to control power through load
adjustment can lead to successive drops in the voltage magnitude at the load bus, also known as
voltage instability. While voltage instability is typically a local phenomenon, its effects can lead
to significant system-wide failures. If voltage instability at one location impacts the ability of
other buses to maintain acceptable voltages, the subsequent result can be low voltages
throughout a sizable portion of the grid. This cumulative effect is known as voltage collapse [2].
Several factors can be identified as causes of voltage instability; however, most issues arise from
insufficient reactive power support. This is due to the fact that increases in reactive power
demand in a system result in voltage drops. In order to prevent voltage decay, appropriate
7
The two main qualities that need to be investigated when analyzing voltage stability include the
nearness to voltage instability and the driving factors of voltage instability [3]. When
determining the proximity to an unstable voltage operating condition values such as load, real
power transfer, and available reactive power all need to be considered. Studying the mechanisms
involved in a voltage instability situation involves determining the source of the instability and
possible solutions for preventing the loss stability [3]. Classification of voltage instabilities can
broken down into two groups- problems that are accompanied by frequency instabilities and
problems in which there are insignificant changes in frequency [4]. The first class incorporates
frequency) and its field and terminal voltages [4]. The other class of voltage issues occurs with
relatively high independence from frequency. Studying these voltage problems draws upon static
and dynamic analysis. Since voltage instability typically occurs in a mid/long term time frame
(on the order of 30 seconds to tens of minutes), the validity of static methods remains for many
attributes of the problem [2]. Such analysis can reveal information regarding key contributing
factors to the source of the instability. For a more detailed analysis of the condition, dynamic
simulations can be performed. Dynamics will shed insight into the system’s response to attempts
aimed at restoring steady-state equilibrium [2]. Utilizing both of these analysis techniques in a
The growth of distributed renewable resources and new technologies available to consumers,
such as fuel cells and electric vehicles, are increasing interest in DC or mixed AC/DC
distribution systems [5]. However, the issue of voltage stability discussed above presents several
8
challenges in systems with DC lines, due to the tie between voltage level and active power [5]. In
a DC line, the concept of reactive power is nonexistent and drops in voltage are a result of line
resistance. This means the voltage level cannot be manipulated by altering reactive power;
rather, a change in voltage also implies a change in active power. Additionally at the distribution
level, lower line voltages cause the issue of voltage drop to be further exasperated.
Voltage source converter (VSC) based DC lines are emerging as a valuable option for system
operators to improve grid stability. A VSC is a power electronics device capable of acting as
both an inverter and rectifier. Spawning from improvements in semiconductor technology, its
main component is the insulated-gate bipolar transistor (IGBT), which can be turned off and on
multiple times per power frequency cycle [6]. This aspect offers significant advantages over
previously used line commutated converters, since switching control is no longer dependent on
the main current to switch off [6]. The operation of VSCs in a DC line can viewed as a black box
9
The flexible characteristics of VSCs open the door for novel applications, which can positively
influence grid stability [7]. VSCs offer both voltage and frequency control and can rapidly
changed to meet active stresses on a system by manipulating both active and reactive power [7].
Since wind, solar, and other renewable energy sources threaten voltage and frequency stability
due to their uncontrollable variability, VSC transmission provides a plausible solution for
counteracting some of the fluctuations in generation [8]. Additionally, when power must be
transferred over long distances, high voltage DC (HVDC) lines have proven there economic
viability over AC lines. Traditionally, current source converters have been used for conversion
between AC and DC in HVDC projects. Alternatively, VSCs provide a new option for
implementing an HVDC scheme with greater controllability, which does not depend on naturally
Power electronic conversion equipment, such as VSCs, also provide an interesting interplay
between DC lines and an AC distribution system [5]. VSCs and similar devices which are
capable of injecting reactive power could be used to manage AC voltages of connected buses [5].
In the distribution system, this characteristic of VSCs will allow for better interconnection of DC
10
CHAPTER II
In AC systems, two types of power are necessary to sustain satisfactory operation: active power
and reactive power. Controlling these values requires an understanding of the factors with the
greatest influence on both forms of power. Such relationships can be derived from the equations
The formula for net active power injection is shown below in equation (1).
(1)
Since the reactance of a transmission line is much greater than the resistance, the Bik terms will
contribute significantly more than the Gik terms. Additionally, the voltage angle difference is
typically less than 0.2618 radians (15 degrees), which means the cosine term will be about 1 and
the sine term will can be approximated as i- k. Lastly, under normal operating conditions the
voltage magnitude will be between 0.95 and 1.05 per unit. Applying these assumptions to
(2)
It can be seen from this equation that the net power injection depends heavily on the voltage
angle difference.
11
The deductions made for the active power flow derivation are also valid for the net reactive
(3)
Following the same logic from the previous derivation, equation (3) can be reduced to:
(4)
Expanding the Bik system characteristic yields additional insights into where the contributions
, for i k (5a)
(6)
The first term of equation 6 relates to the reactive power contribution of the shunt susceptance at
bus “i”, while the second term factors in the transfer of reactive power from connections.
Equation (6) also shows the strong coupling between voltage magnitude and reactive power.
VSC Operation
One of the most well practiced techniques for understanding an electric grid is steady-state
power flow analysis. Such analysis provides insight about operating points, potential
contingencies, and the impact of introducing new devices (such as VSC transmission lines) on
system conditions. A challenge for the power systems industry associated with adding innovative
technologies revolves around appropriate representation of such equipment. In the case of VSC
based HVDC lines, it is important to understand key operation features before progressing into
steady-state representation.
12
Active Power Control
VSC based DC lines, like other DC lines and traditional AC lines, is capable of transferring
active power from a generation source to loads; however, it offers a significant advantage over
its alternatives- controllability. VSCs are capable of varying the amount of active power they
supply. The internal components of a VSC are capable of switching at extremely high
load variations as they occur. Treating the converter as a black box as shown below in Figure 2, a
high level concept of active power control can be easily examined [6].
Control of the active power supplied to an AC load is achieved by regulating the phase angle of
the voltage generated on the inverter side of the VSC (Uconv). If the angle of Uconv, δ, is leading
the voltage angle of AC load, θ, then active power (Pconv) will be transferred to the AC system.
In contrast, if δ lags θ, then Pconv flows from the AC connection. This performance and control
technique follows from the strong coupling between net active power injection and voltage angle
13
VSCs are also able to supply reactive power to a connected AC load. Similar to active power, the
amount of reactive power generated is controllable with relatively fast response times. Again, the
fundamental idea behind reactive power control can be seen in Figure 3 through a black box
Reactive power control is achieved by manipulating the magnitude of the voltage produced on
the AC side of the VSC (Uconv). If Uconv is greater than the AC system voltage, then reactive
extraction of reactive power from the AC system. This scheme is consistent with the logic found
in equation (6), which outlines a heavy correlation between net reactive power injection and
voltage magnitude. Since both active and reactive power vary with the manipulation of different
characteristics, these values can be set and changed almost independently. However, it should be
noted that the logic presented for power control with VSCs might be broken when converter
angles are close to zero degrees or voltage magnitudes are numerically close to their AC
counterparts. This is due to the fact that active power is slightly influenced by voltage magnitude
14
For transmission and distribution operators, a first step towards understanding the real-time
situation is through conducting steady-state power flow analysis. Steady state analysis revolves
around acquiring a valid solution to a systems power flow equations. It takes an instantaneous
snapshot to reveal valuable information about the grids operating conditions under prescribed
loadings. In order to incorporate VSC DC lines into a power system, its steady state
The first step in modeling a VSC based DC line for steady-state analysis involves mapping DC
and AC variables to VSC power transfer equations. The following shows how to derive these
equations. Figure 1 from the “VSC Technology for Improved Power Delivery” section of chapter
one has been repeated below as a convenient reference for the nomenclature used for variable
names.
Power transfer through the DC line can be described using equation (7) below.
(7)
15
VSC/AC System Power Exchange:
As mentioned in the “VSC Operation” section of this chapter, power transfer between the AC
system and the VSC depends heavily on two variables- Uconv and δ. The active power from the
(8)
Equation (9) shows the reactive power relationship between the VSC and the connected AC
system.
(9)
The DC variables must be linked to the VSC variables to relate the DC power flow to the power
exchanged between the VSC and the AC system. The following equation establishes this
connection.
(10)
These equations introduce two new parameters (C and kλ), which depend on the VSC’s
operation. C is a constant associated with VSC’s configuration and number of levels. For a two-
level VSC in a three phase configuration, the value of C equates to . The kλ term is a
controllable parameter that represents a ratio of Uconv to its maximum possible value, and it can
take any value between 0 and 1. Substituting equation (10) into equation into equations (8) and
(9) yields a set of power equations that depend on DC, AC, and VSC variables.
(11)
(12)
16
Equations (11) and (12) illustrate that the power flow through a VSC based DC transmission line
varies with the DC line voltage (Ud), the AC system voltage (UL), the voltage ratio (kλ), and the
phase angle difference (δ) between the VSC output voltage and the AC system voltage.
Including a VSC transmission line into a power flow model requires several initial assumptions.
Equations (11) and (12) contain four variables that determine the power flow; therefore, two of
the variables must be “set” (i.e. assume a prescribed value) to reach a solution. Setting these
values depends on the VSC control scheme implemented. For example, the VSC could be used
to control the AC bus voltage and DC line voltage, which means that UL and Ud would be
known. Additionally, the real and reactive power must be assumed or solved for prior to finding
the unknown variables. Typically, the desired real power is known. Determining the reactive
power depends on the selected control scheme. For AC voltage control, the reactive power
should be allowed to change, meaning that it must be found in a similar manner to solving for the
reactive power produced by a generator. In other words, the AC system power flow equations
should be solved first to determine the reactive power the VSC must supply (or absorb).
Alternatively, if the AC bus voltage is not controlled, then the desired reactive power for the
VSC must be known. A series of equations is provided below as a general guide for solving a
(13)
Where Pac is given by equation (1) and Pvsc is given by equation (11).
(14)
Where Qac is given by equation (3) and Qvsc is given by equation (12).
17
(15)
The system in (15) can be broken up into AC and VSC equations, which can be solved
separately. The next four equations can be used to find the AC solution.
(16)
(17)
Where Pvsc and Qvsc are known, or there is AC voltage control and the bus is treated as a PV bus.
(18)
The Newton-Raphson iteration for the AC system is presented below in equation (19). It should
(19)
After finding the AC system’s solution, Qvsc can be calculated (if it was not already known)
(20)
Again, Newton’s method can be used to find the solution to the system of equations.
(21)
18
After reaching convergence in equation (21), all system variables (xac and xvsc) will be known
The complexity of the steady state model for VSC-DC can be reduced if the values of for the
VSC parameters do not need to be known. The next two sections introduce concepts for
simplifying the VSC transmission line down to AC components that mimic the VSC’s impact on
the AC system.
VSC as Injection/Extraction
If the VSC is operating without controlling the AC bus voltage, then the VSC can be treated as
an active power extraction/injection at its connection points. At the sending side of the link, the
VSC extracts active power from the bus like a load. On the receiving side, active power is
injected into the bus. Thus, it acts as a constant power source supplying the amount of active
power that was extracted from the sending side, minus a small line loss. On both ends of the
VSC transmission line, reactive power can be removed or added independent of the reactive
power operation at the other end and any active power settings. This means that the VSC-DC
line can be represented as only a constant PQ “load” on the sending side and a constant PQ
“generator” on the receiving side. Since DC lines have low resistance, the amount of power lost
during transmission is typically negligible relative to the total amount of power sent. However, if
neglecting this loss is not permissible, it can be easily calculated using the DC line voltages and
added to the “load”. The Matlab Simulink model below (constructed using the Power Systems
19
Figure 4: VSC based DC Line (From Bus 2 to Bus 3)
VSC as PV Bus
VSCs are capable of supplying reactive power to an AC system, and the amount can be varied
quickly. An appropriate controller design could take advantage of this attribute. Since voltage
magnitude is strongly coupled to reactive power, a VSC could be used to supply or drain the
necessary reactive power needed to maintain the desired bus voltage. By operating in this fashion
and providing constant active power support, both buses connected to the VSC transmission line
would experience voltage regulation. This means that the buses can be classified as PV buses in
a power flow model, as long as the reactive power requirement is within the VSC’s operating
limits. An example of a three bus system with reactive power controlled VSCs is implemented in
Figure 5.
20
Steady State Voltage Stability
Measuring the level of voltage stability in the steady state time frame is accomplished through
the use of PV and QV curves. These curves relate the voltage magnitude at a bus to various
active and reactive power contributions. They can be used to check a particular buses proximity
to voltage instability.
PV Curve Analysis
A PV curve maps specific active power loads to voltages. It can be constructed using a series of
iterations which increment in small steps until a power flow solution no longer exists. Each
increment should only increase the active power (i.e. the reactive power must be kept the same).
The upper portion of the PV curve represents stable operating conditions, which is characterized
by a decline in voltage magnitude as the power demand increases. The critical point occurs at the
right side of the curve, where the upper and lower portions meet. If the system moves into the
21
lower part of the graph, then the voltage magnitude becomes unstable. Operation in the unstable
area means that the power systems attempts to raise the voltage will fail, resulting in a possible
collapse. It is also worth noting that the load flow Jacobian becomes singular for the lower part
of the curve. This implies that eigenvalue analysis can also be used to investigate voltage
stability.
Construction of PV curves for buses connected to VSCs dictates use of either the full detailed
method without AC voltage control or the extraction/injection method. This is due to the
constant reactive power supply and demand requirement for PV curves. When the VSC is
operating without any reactive power control, this demand is not a problem. Steady state voltage
stability for implementations with reactive power control, however, should not be neglected.
Therefore, to account for the range of reactive power a controlled VSC can supply, at least two
PV curves should be created using stable operating points. One of the curves should represent the
lower reactive power limit for the VSC, while the other should correspond to the upper bound.
The result will be a PV “region” that outlines all of the possible active power and voltage
magnitude combinations. For a more accuracy regarding the critical side of the region, additional
22
CHAPTER III
The test system presented in Figure 7 represents a simple two bus system, for the purpose of
testing the VSC steady state model derived in the previous chapter. The variable values the
proposed model converges to should satisfy each of the VSC power flow equations and follow
Table 1 outlines all of the system parameters. The VSC control scheme used for the simulation is
AC voltage control; therefore, the voltage at bus 2 is known. Additionally, it is assumed that the
R Xc Pvsc2 Ud2 V2 V1 P2 Q2
23
Following the procedure in the “VSC Steady State Representation” section of Chapter II, the first
step is to solve the AC system. Since there is no AC transmission line found in Figure 7, the
active and reactive AC power flows are both zero. The next step is to calculate the unknown
VSC-DC variables using the Newton-Raphson method. A Matlab function and script were
written to perform this action. (The code can be found in the Appendix for reference.) The results
Ud1 δ1 k1 Ud2 δ2 k2
The validity of the converged values in Table 2 can be confirmed by plugging them into
equations (8) through (10), and verifying the power demands are met without violating the VSC
The active and reactive power transfer between bus 2 and its VSC connection (Pvsc2 and Qvsc2)
show that load (P2 and Q2) is satisfied. It is also important to check if the converter output
24
voltage (Uconv) and angle (δ) correspond with understandings of VSC operation. The value of
Uconv2 is greater than V2, implying that reactive power is being transferred to bus 2. Additionally,
δ2 is negative, indicating that it is leading the bus 2 voltage angle. Therefore, active power is
being transmitted to bus 2. Since bus 2 is a load bus, active and reactive power should indeed be
moving from the VSC to the AC system. A similar analysis can be performed on bus 1 and the
VSC connected to it. In this case δ1 is positive, which means that the VSC is receiving active
power. Pvsc2 is positive, confirming this assertion. The reactive power demand by the VSC
connected to bus 1 was set to zero to show the special case, when the operation logic fails.
According to the idealized relationship derived in the previous chapter, for zero reactive power
transfer Uconv1 should be set equal to V1; however, the actual value for Uconv1 found in the
simulation is slightly greater than V1. This is necessary to offset the effect of δ1 on reactive
power transmission between bus 1 and the connected VSC. From inspection it is clear that the
variables in Table 2 converged to values that satisfy the power flow equations.
A two bus mixed AC/VSC-DC system will be used to compare the accuracy of the simplified
“VSC as PV bus” model presented in Chapter II with the detailed model used in the simulation
above. Figure 8 below depicts the test system, with all known parameters listed in Table 4.
25
Figure 8: Mixed AC/VSC-DC Test System
Table 4. Two Bus, Mixed AC/VSC-DC System Parameters (With AC Voltage Control)
System AC Paramters
Xc Pvsc2 Ud2
A Matlab function was written to handle a two bus AC system. This code can be used in
conjunction with the detailed, VSC model code used for the simulation in the previous section.
(The code for the two bus AC function can be found in the appendix.) The power flow solution
26
Table 5. Mixed AC/VSC-DC Power Flow Results (Detailed, With AC Voltage Control)
The simplified PV bus model introduced in the previous chapter represents the VSC transmission
line as an AC load and PV generator. For the model to be useful, it should reach the same power
flow results as the detailed model. The Power Systems Analysis Toolbox was used to construct
27
Table 6. Mixed AC/VSC-DC Power Flow Results (Simplified PV Bus)
Comparing Table 5 and Table 6 indicates that for VSC transmission using an AC voltage control
scheme, the simplified PV bus model yields the same power flow solutions. The disadvantage of
using the simplified method lies in the lack of information about the VSC parameters. In the
If the VSC is not being used to control the AC voltage at its connected bus, then the control
scheme will maintain a constant active and reactive power transfer. In this case, the “VSC as a
PQ bus” simplified model introduced in Chapter II can be used as an alternative to the detailed
model. This option can be verified as a reliable option for solving the steady state power flow
problem by cross-referencing the results with the solution provided by the detailed model. The
test system set-up will be the same as the one shown in Figure 8. The system parameters can be
found in Table 7.
28
Table 7. Two Bus, Mixed AC/VSC-DC System Parameters (No AC Voltage Control)
System AC Paramters
The detailed and simplified models can be solved with the same procedures used to verify the PV
bus simplified model. The solutions for both methods can be found in the tables below, as well
29
Figure 10: Simplified Model (VSC as PQ bus)
Evaluating the results found in tables 8 and 9, reveals that the same power flow solution is found
using both methods. This shows that the simplified model provides an accurate alternative to the
detailed model, when the VSC parameters (Ud, kλ, and δ) are not required.
30
For a conventional AC transmission line, there is a one-to-one relationship between a bus’s
loading and its voltage magnitude. In other words, plotting stable AC bus voltages as a function
of active and reactive power yields a surface that outlines all of the possible loading and voltage
(PQV) combinations. The figure below shows the shape of the PQV capability region at the
One advantage of using VSC transmission is its ability to maintain steady AC bus voltages for
multiple different loadings. Therefore a change in loading does not necessitate a change in
voltage. Mapping loadings to achievable AC bus voltages produces a shape with volume. An
example PQV capability region for the receiving end of a two bus system with a VSC-DC line is
31
Figure 12: PQV Capability Region for VSC-DC Line
Figure 11 and Figure 12 show only a portion of the complete capability regions for their
respective power delivery schemes; however, both provide enough information to visualize the
relationship between achievable loading and voltage combinations. (The Matlab code used to
32
CHAPTER IV
CONCLUSION
The detailed model introduced in Chapter II provides information about controllable parameters
for the VSC, as well as a viable solution to the steady state power flow problem. Calculating the
parameter values offers valuable insight into VSC’s operating condition. Particularly, a detailed
model can be used to generate a capability region for a particular VSC line. Knowledge of a
solution’s location within this region will let system operators discern proximity to unattainable
loading conditions. The detailed model also allows the impact of the AC voltage (UL), DC line
voltage (Ud), VSC output voltage ratio (kλ), and VSC output angle (δ) to be investigated. This
can be helpful for improving intuition about the abilities of a VSC based DC line and developing
The simplified models presented in the second chapter offer a less computationally expensive
alternative to the detailed model. The detailed model requires Newton-Raphson formulations for
both the AC and VSC-DC lines, doubling the number of Jacobian matrices that need to be
inverted. Since the simplified models represents the VSC transmission line using pre-existing
AC system concepts (PV generators and PQ loads), the power flow solution can be found
utilizing conventional methods. In the case of the “VSC as a PV bus” model, the VSC-DC line
actually reduces the size of the Jacobian matrix if it is connected to a load bus. This reduces the
matrix inversion time less than both the detailed model and in a purely AC line system. These
33
models are valuable as a time saving alternative to detailed model, provided the VSC system
parameters do not need to be known. This characteristic is valuable when large numbers of
power flow simulations are being run, for example during long term system planning.
VSC Flexibility
The flexibility of VSC-DC is one of its most important advantages over alternative power
delivery schemes. The PQV capability region for a bus with a VSC line connection illustrates
this concept. The VSC allows the power demand to vary without necessitating an adverse affect
on the bus voltage. This property is a result of the fast response time that characterizes the power
electronics used in a VSC, as well as the strong decoupling between active and reactive power.
VSC transmission could prove a viable option for connecting renewable resources to the grid.
Since renewable generation, such as wind or solar, offers limited reactive power support and has
high variability, VSC-DC with its flexibility to quickly change its control parameters offers a
possible answer to the challenges associated with incorporating these energy sources in the
existing grid.
Future Work
Future work for the steady state model includes writing Matlab code that can be integrated with
the existing MATPOWER program. This will introduce the model to be utilized by others
interested in power system operation and VSC based DC lines. Additionally, detailed models for
dynamic time frames also need to be investigated in order to fully understand the effects and
34
REFERENCES
[1] J.G. Kassakian and R.Schmalensee. (2011). The Future of the Electric Grid.
Massachusetts Institute of Technology. Cambridge. [Online]. Available:
http://mitei.mit.edu/system/files/Electric_Grid_Full_Report.pdf
[2] P. Kundur. “Voltage Stability” in Power System Stability and Control. New York:
McGraw-Hill, 1994, pp. 959-1021.
[3] B. Gao, G.K. Morison, and P. Kundur, “Voltage Stability Evaluation Using Modal
Analysis,” IEEE Trans., vol. PWRS-7, no. 4, pp. 1529-1542, Nov. 1992.
[5] Hammerstrom, D.J., "AC Versus DC Distribution SystemsDid We Get it Right?," Power
Engineering Society General Meeting, 2007. IEEE , vol., no., pp.1,5, 24-28, June 2007.
[7] S. Johansson et al. “Power System Stability Benefits with VSC DC-Transmission
Systems.” [Online]. Available: http://www.transform.ru/articles/pdf/sigre/b4-204.pdf
[8] J. Pan et al. “AC Grid with Embedded VSC-HV DC for Secure and Efficient Power
Delivery,” in IEEE Energy 2030, Atlanta, 2008.
35
APPENDIX
function [ Pvsc_i, Qvsc_i, Vd_i, d_i, k_i, Pvsc_r, Qvsc_r, Vd_r, d_r, k_r ] =
VSC_PF( P2, Q2, V2, V1, Vd_i, Xc, R)
%The purpose of this function is to preform a NR iteration on a 2 bus
%system and return the variables for a VSC-DC line
% This function assumes that the VSC is operating to keep the AC voltage
% constant. Vd, k, and d represent the DC variables. It is assumed that
% the DC voltage on the inverter side (Vd_i) is known as well as the AC
% voltage at the connected bus. Additionally, this function operates on
% a committed real power delivery to the load. It solves for the unknown
% DC variables on the inverter side (k_i and d_i), then solves for the
% DC variables on the rectifier side (Vd_r, k_r, and d_r).
Qvsc_i = (V2*(V2-(sqrt(6)/pi)*k_i*Vd_i*cos(d_i)))/Xc;
L21 = (sqrt(6)/pi)*(k_i*Vd_i*sin(d_i)/Xc);
L22 = (-sqrt(6)/pi)*(Vd_i*cos(d_i)/Xc);
36
%Do NR until satisfactory convergence of unknown variables
while dd_r > 1e-6 || dk_r > 1e-6
Pvsc_r = (sqrt(6)/pi)*(k_r*Vd_r*V1*sin(d_r)/Xc);
L11 = (sqrt(6)/pi)*(k_r*Vd_r*V1*cos(d_r)/Xc);
L12 = (sqrt(6)/pi)*(Vd_r*V1*sin(d_r)/Xc);
Qvsc_r = (V1*(V1-(sqrt(6)/pi)*k_r*Vd_r*cos(d_r)))/Xc;
L21 = (sqrt(6)/pi)*(k_r*Vd_r*sin(d_r)/Xc);
L22 = (-sqrt(6)/pi)*(Vd_r*cos(d_r)/Xc);
function [ P2, Q2, V2, Theta2, P1, Q1, V1, Theta1, dTheta2 ] = ACPF_2bus( P2,
V2, V1, Theta1, G12, B12, G11, B11, G22, B22 )
%This function calculates the AC power flow for a 2 bus system with PV bus.
% The fucntion returns all of the AC parameters in a 2 bus system
dTheta2 = 1;
Theta2 = 0;
%Find Theta2
while abs(dTheta2) > 1e-6
P2x = V2*V1*(G12*cos(Theta2-Theta1)+B12*sin(Theta2-Theta1))+(V2^2)*G22;
J = V2*V1*(-G12*sin(Theta2-Theta1)+B12*cos(Theta2-Theta1));
dP2 = P2-P2x;
dTheta2 = inv(J)*dP2;
Theta2 = Theta2+dTheta2;
end
%Find Q2
Q2 = V2*V1*(G12*sin(Theta2-Theta1)-B12*cos(Theta2-Theta1))+(V2^2)*(-B22);
%Find P1
P1 = (V1^2)*(G11)+V1*V2*(G12*cos(Theta1-Theta2)+B12*sin(Theta1-Theta2));
%Find Q1
Q1 = (V1^2)*(-B11)+V1*V2*(G12*sin(Theta1-Theta2)-B12*cos(Theta1-Theta2));
end
37
Matlab Code for Generating 2-Bus AC PQV Region
F = 1+((Zload./Zline).^2)+2*(Zload./Zline).*cosd(Zln_angle-PF_angle);
Pr = (Zload./F).*((Vs/Zline)^2).*cosd(PF_angle);
end
F = 1+((Zload./Zline).^2)+2*(Zload./Zline).*cosd(Zln_angle-PF_angle);
Qr = (Zload./F).*((Vs/Zline)^2).*sind(PF_angle);
end
F = 1+((Zload./Zline).^2)+2*(Zload./Zline).*cosd(Zln_angle-PF_angle);
Vr = ((1./F).^(.5)).*(Zload./Zline)*Vs;
end
38
%vary by load impedence magnitude.
Pr = zeros(length(Phi),length(Zld));
i = 1;
for i = 1:140
PF_ang = Phi(i);
Pr_i = ActivePwr(Vs, Zln, Alpha, Zld, PF_ang);
Pr(i,:) = Pr_i;
end
Pvsc = (sqrt(6)/pi)*k*Vd*Vi*sind(d)/X;
End
Qvsc = (Vi*(Vi-(sqrt(6)/pi)*k*Vd*cosd(d)))/X;
end
39
Script to Generate VSC PQV Region
40