Available online at www.sciencedirect.

com

ScienceDirect
Procedia Computer Science 47 (2015) 168 – 178

Analysis of continuous power flow method, Model analysis, Linear

Regression and ANN for voltage stability assessment for different
Loading conditions
Veeranjaneyulu.Puppala a, Dr.Purna chandrarao b ,
a. Associate professor, MRIET, Secundrabad,India
b.Professor (Rtd) NIT,Warangal,India

Abstract

This Paper presents voltage stability assessment through P-V curve and bus voltage sensitivity factor. Active power margin is
used to measure the distance to instability. Analyses by using computational methods are used to obtain P-V curve of power
system. Continuation power flow starting with initial operating point and increasing load to the maximum loading point and to
reduce the size of the jacobian matrix modal analysis is adopted. A data statistics analysis tool is used to run linear regre ssion. In
this paper the linear regression method with ANN is compared with conventional methods and voltage stability is improved
within less iterations.
© 2015 The
© 2015 The Authors.
Authors. Published
PublishedbybyElsevier
ElsevierB.V.
B.V.This is an open access article under the CC BY-NC-ND license
Peer-review under responsibility of organizing committee of the Graph Algorithms, High Performance Implementations and
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Applications (ICGHIA2014).
Peer-review under responsibility of organizing committee of the Graph Algorithms, High Performance Implementations
and Applications (ICGHIA2014)
Keywords: Regression analysis; MATLAB;Votage stability;power plow,CPF;ANN; Loading conditions

1. Introduction

In general terms, voltage stability is defined as the ability of a power system to maintain steady voltages at all the
buses in the system after being subjected to a disturbance from a given initial operating condition. It depends on the
ability to maintain/restore equilibrium between load demand and load supply from the power system. Instability that
may result appears in the form of a progressive fall or rise of voltages of some buses .Voltage stability problems
mainly occur when the system is heavily stressed beyond its capability. While the disturbance leading to voltage
collapse may be initiated by a variety of causes, the main problem is the inherent weakness in the power system.
*Veeranjaneyulu Puppala Em,ail:[email protected]

1877-0509 © 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of organizing committee of the Graph Algorithms, High Performance Implementations and Applications (ICGHIA2014)
doi:10.1016/j.procs.2015.03.195
 Veeranjaneyulu Puppala and Purna Chandrarao / Procedia Computer Science 47 (2015) 168 – 178 169

Many voltage stability indices are based on the Eigen value analysis or singular value decomposition of the system
power flow Jacobian matrix. The prominent methods in voltage stability analysis are those that find system load
margin, especially when system contingency is considered P--V curve and Q-V curve are most considerable method
to find active power margin and reactive power margin. Network configurations and load distributions can also
reflect using PV curve. The linear approach between the generator reactive power reserves and voltage stability
margin is related to the system PV curves versus nodal VQ curves. Using this relationship, a systematic and practical
method for determining the online voltage stability margin is proposed in.
In most of the research work the voltage stability has been considered as static phenomenon. This is due to slow
variation of voltage over a long time observed in most of the incident until it reaches to the maximum loading point
and then it decreases rapidly to the voltage collapse. Static voltage stability can be analyzed by using bifurcation
theory. There are different types of bifurcation theory, in which saddle node bifurcation is used for static voltage
stability analysis. This paper is concentrated on the steady state aspects of voltage stability. Indeed, numerous
authors have proposed voltage stability indexes based on repeated power flow analysis. In this generally
conventional load flow is used. The main difficulty in this method is that Jacobian of NR power flow become
singular at voltage stability limit (critical point).
A power flow solution near the critical point prone to divergences and error. Singularity in the Jacobian can be
avoided by slightly reformulating the power flow equations and applying a locally parameterized continuation
technique. During the resulting ―continuation power flow, the reformulated set of equations remains well-
conditioned so that divergence and error due to a singular Jacobian are not encountered. The continuation power
flow has some disadvantage of creating the jacobian matrix so linear regression analysis with ANN is done for
voltage stability assessment.

2. 1.Linear regression

The characteristic of voltage stability are illustrated with IEEE 14-bus system. The generator produces active
power, which is transferred through a transmission line to load. The reactive power capability of the generator is
infinite. Thus the generator terminal voltage V1 is constant.

Fig. 1. Line diagram of IEEE 14 bus test system

V2 =√((V12-2QX) ±√( V14-4QX V12-4PX2))/2

Pi = ΣViVjYij cos(δi – δijYij)

Qi = ΣViVjYij sin(δi – δijYij)
For i,j =1 to n
170 Veeranjaneyulu Puppala and Purna Chandrarao / Procedia Computer Science 47 (2015) 168 – 178

Simulation model

Synchronous Synchronous
Generator1 Generator
C ompensator C ompensator

Bus 2

C
A

A
B
Fourteen Bus system

C
A
B
A

B
C

C
Bus 5 Bus 6
Bus 4

C
Bus 3

C
A
B
C
A
B
b

b
a

a
c

b
a

c
A

B A

C B

Iab
Va

Iab

Iab

Iab

Iab

Iab
Bus 1

Va

Va

Va

Va

Va
b
Ba

Iab
A

C
Bus 3 Bus 4 Bus 5 Bus 6 Bus 7

Va

b
Ba

Ba

Ba

Ba

Ba
c

c
A

A
C

C
Bus 2

b
Ba

c
A

C
Bus 1 Bus 7

A A A A A A A A A A
B B B B B B B B B B
C C C C C C C C C C

A A
B B
C C

A A
B B A A
C C B B
C C
A A
B B
C C

A A A A
B B B B
C C C C

A A
B B
C C

A A
B B
C C

A A
B B
C C

A
B
C
A
B
C
C o ntinuo us

C
A
B
C
A
B
po we rgui

A A A A A A
B B B B B B
C C C C C C

A A
B B
C C
Iab
Va

Bus 8
b
Ba

Iab

Iab

Iab

Iab

Iab

Iab
A

Va

Va

Va

Va

Va

Va
Bus 9 Bus 10 Bus 11 Bus 12 Bus 13 Bus 14
b

b
Ba

Ba

Ba

Ba

Ba

Ba
c

c
A

A
C

C
A

B
C

C
b
a

B
C

C
A

B
C

Bus 12 Bus 13 Bus 14

Bus 11
Bus 9 Bus 10
Bus 8
A

Synchronous
C ompensator

Fig2. Simulation model for IEEE 14-bus system

(a) (b)
 Veeranjaneyulu Puppala and Purna Chandrarao / Procedia Computer Science 47 (2015) 168 – 178 171

(c)

Fig 3. Results of bus voltages across each bus (example 1,3,7 taken)

Fig 4. Vbus formation Fig 5. Line data of ieee 14-bus system

2.2. Modal Analysis:

Fig.6 Eigen values of the reduced Jacobian matrix against load multiplication factor, K
172 Veeranjaneyulu Puppala and Purna Chandrarao / Procedia Computer Science 47 (2015) 168 – 178

Table 1. Transmission lines data (r, x and b in pu on 100mva base) for the 14-bus test system

End buses R X B/2

1-2 0.01938 0.05917 0.0264
2-4 0.05811 0.17632 0.0170
12-13 0.22092 0.19988 0
13-14 0.17093 0.34802 0

Table 2: transformer data (r, x in pu on 100 mva base) for the 14-bus test system

End buses R X
3-8 0.0671 0.17173
7-9 0 0.11001
6-7 0 0.2522

Table 3: Shunt capacitor(r, x in pu on 100 mva base) for the 14-bus test system
End buses MVAR(pu)
4 0.191
5 0.016

Table 4: Base case load data (pu on 100 mva base) for the 14-bus test system
Bus P(MW) QMVAR(pu)
3 0.217 0.127
4 0.942 0.191
7 0.112 0.075
8 0.050 0
9 0.295 0.166
10 0.09 0.058
11 0.035 0.018
12 0.061 0.016
13 0.135 0.058
14 0.1499 0.050

Table 5: Base case generator data (pu on 100 mva base) for the 14-bus test system
Bus V(pu)
1 1.06
2 1.045

Table 6: Eigen values of reduced Jacobian matrix (Pu on 100 MVA base) for the 14-bus test system
K E1 E2 E3 E4
1.124 0.1861 0.3190 0.1361 0.5786

Table 7: Transformer data for different load levels (pu on 100 mva base) for the 14-bus test system
End buses R X Tap
setting
1
0.978
9-10 0.03181 0.08450
0.969
0.932
 Veeranjaneyulu Puppala and Purna Chandrarao / Procedia Computer Science 47 (2015) 168 – 178 173

Table 8: Load data for different load levels (pu on 100 mva base) for the 14-bus test system
Bus P(pu) Q(pu) Load level
4 0.942 0.191 1
0.931 0.185 0.978
0.928 0.189 0.969
0.940 0.172 0.932
5 0.478 0.197 1
0.435 0.192 0.978
0.241 0.188 0.969
0.448 0.179 0.932
9 0.295 0.166 1
0.285 0.164 0.978
0.274 0.158 0.969
0.286 0.149 0.932

Table .9: generator data for different load levels (pu on 100 mva base) for the 14-bus test system

Bus P(pu) Voltage(pu) Load level

4 0.035 1.06 1

0.061 0.978
0.135 0.969
0.1499 0.932
5 0.050 1.045 1

0.295 0.978
0.09 0.969
0.217 0.932
9 0.242 1.01 1

0.112 0.978
0.235 0.969
0.241 0.932

Table 10: load voltages and reactive power outputs of generator 2 and 3 at load level 1 (pu on 100 mva base) for the 14 -bus test system

Contingency V5 V6 QG3 QG2

Without outage, fixed tap 0.96 1.11 290 -83
Without outage, LTC active 0.99 1.08 227 144
Line outage, fixed tap 0.91 1.00 200 224
Line outage, LTC active 1.01 1.09 243 146

Table 11: load voltages and reactive power outputs of generator 2 and 3 at load level 2 (pu on 100 mva base) for the 14 -bus test system

Contingency V5 V6 QG3 QG2

Without outage, fixed tap 1.03 1.11 290 -83
Without outage, LTC active 0.99 1.08 227 164

Table.12: load voltages and reactive power outputs of generator 2 and 3 at load level 3 (pu on 100 mva base) for the 14 -bus test system

Contingency V5 V6 QG3 QG2

Without outage, fixed tap 1.02 1.11 401 -81
Without outage, LTC active 0.98 1.07 700 249
174 Veeranjaneyulu Puppala and Purna Chandrarao / Procedia Computer Science 47 (2015) 168 – 178

Fig 7 Voltages in magnitude and angle, Eigen values after modal analysis applied

Fig 8 Power generation versus bus number Fig 9 Power load versus bus number
Table 13. Active and reactive power at various buses

Fig 10 Reactive power versus bus number

2.3. Continuation Power Flow Result (With Gen. Reactive Limit Constraint) and Linear Regression:
 Veeranjaneyulu Puppala and Purna Chandrarao / Procedia Computer Science 47 (2015) 168 – 178 175

K=0.5 K= 0.9

K=1 K=1.1

Fig 11. Forming the linear regression equation

Fig 12. Linear regression (Basic fitting) Analysis(K=1.146) Fig 13. Data statistics of mean, standard deviation
176 Veeranjaneyulu Puppala and Purna Chandrarao / Procedia Computer Science 47 (2015) 168 – 178

Fig 14. Data statistics for input Fig 15. Data statistics for output

Fig 16. Individual voltages variation (linear regression)

Fig 17. Modified power with linear regression

 Veeranjaneyulu Puppala and Purna Chandrarao / Procedia Computer Science 47 (2015) 168 – 178 177

2.4. Neural network and regression analysis for MLP estimation:

Fig 18. Only data statistics and linear regression Fig 19.With linear regression for MLP estimation

Step by step procedure:

With NN tool box

Selecting data in NN

Selecting the no. of hidden layers and network size

Training the Network

Fig 20. Results after training

178 Veeranjaneyulu Puppala and Purna Chandrarao / Procedia Computer Science 47 (2015) 168 – 178

Fig 21. Comparing results with ANN and Linear regression

3. CONCLUSION

Above all results shows that voltage stability margin can be found easily by linear regression with
ANN. And P-V curve and max. Loading point can access. Only collapse point is not enough for voltage
stability assessment .So, using tangent vector sensitivity analysis can be done. From voltage sensitivity factor
weakest bus can identify. The Weakest bus identification is done by without excessive calculation. Placement
of reactive power sources such as Fact devices, capacitor bank is known. This result is same accurate as to
find Bus participation factor using QV modal analysis, continuation load flow and etc. This linear regression
method and neural networks method is more accurate and simple for Voltage stability analysis.

4. REFERENCES

1. Venkataramana Ajjarapu ―Computational Techniques for Voltage Stability Assessment and Control E-
Book—Library of Congress Control Number: 2006926216, Iowa State University, Department of
Electrical and Computer Engineering. 1122 Coover Hall, Ames Iowa 50011, U.S.A.
2 .Varun Togiti ―Pattern Recognition of Power System Voltage Stability using Statistical and Algorithmic
Methods University of New Orleans ScholarWorks@UNO University of New Orleans . 5-18-2012.
3.B. Gao, G. K. Morison, and P. Kundur, ―Voltage stability evaluation using modal analysis, IEEE Trans. on
Power Systems, vol. 7, no. 4,pp. 1529–1542, Nov. 1992.
4. P. A. Lof, T. Smed, G. Anderson, and D. J. Hill, ―Fast calculation of a voltage stability index,IEEE Trans.
on Power Systems, vol. 7, no. 1,pp. 54–64, Feb.1992..
5. L. Bao, Z. Huang, and W. Xu, ―On-line voltage stability monitoring using var reserves, IEEE Trans.
Power Syst., vol. 18, no. 4, pp.1461–1469, Nov. 2003.
6.Satish Joshi , ―A Thesis on Voltage stability and contingency selection studies in electrical power system,
Department of electrical engineering. Indian institute of technology Kanpur . December 1995.
7.P. Kundur, ―Power System Stability and Control McGraw-Hilll, 1994.
8.J. Paserba, V. Ajjarapu, G. Andersson, A. Bose, C. Canizares, N. Hatziargyriou, D. Hill, A. Stankovic, C.
Taylor, T. Van Cutsem, V. Vittal
9.P.Kundur, "Definition and Classification of Power System Stability," IEEE Trans. on Power Syst., vol. 19,
no. 2, pp. 1387-1401, May 2004.