VOLTAGE STABILITY IN NIGERIA 330KV INTEGRATED 52 BUS POWER NETWORK USING PATTERN RECOGNITION TECHNIQUES: A RECURSIVE APPROACH

[Akwukwaegbu* 4(4): April, 2017] ISSN 2349-4506
Impact Factor: 2.785
Global Journal of Engineering Science and Research Management
http: // www.gjesrm.com © Global Journal of Engineering Science and Research Management
[75]
VOLTAGE STABILITY IN NIGERIA 330KV INTEGRATED 52 BUS POWER
NETWORK USING PATTERN RECOGNITION TECHNIQUES: A RECURSIVE
APPROACH
Akwukwaegbu I O*, Nosiri O. C., Agubor C. N., Opara R.O., Olubuiwe M
* Department of Electrical and Electronic Engineering, Federal University of Technology, Owerri,
Nigeria
Department of Electrical and Electronic Engineering, Federal University of Technology, Owerri,
Nigeria
Nigeria
Nigeria
Nigeria
DOI: 10.5281/zenodo.569972
KEYWORDS: Voltage Instability and Stability, PSSE, 330kV Integrated 52 bus Power System, RLSC, and
CART.
ABSTRACT
Detecting the voltage instability in advance enables remedial actions and preventive measures to cushion the effect
of the oncoming voltage collapse phenomenon in power systems. This was achieved by implementing Pattern
Recognition Techniques (PRTs) in conjunction with Power System Simulator for Engineering (PSSE) program.
It was then deployed in Nigeria 330KV Integrated 52 bus power system to actualize Regularized Least Squares
Classification (RLSC) and Classification and Regression Trees (CART) heuristic methods. The methods were
deployed for separating voltage stability and unstable cases that resulted under system contingencies and fault
conditions. Dynamic simulation, system voltage stability and unstable/instability cases results, and the channel
outputs of these voltage cases against time were realized.
INTRODUCTION
Electrical energy is an essential ingredient for the industrial and all-round development of any country. The quality
of life in any country is highly dependent on a reliable electricity supply. The frequent system collapse in the
Nigerian power sector has severally plunged the nation into darkness due to system instability in the Nigeria
electric power system. The epileptic nature of the supply has resulted to slow economic growth and dissatisfaction
among the citizenry [1, 2, 3].
To assist in overcoming the instability problems, analysis of the Nigerian electric power system becomes
necessary. Power system stability is the ability of an electric power system, for a given initial operating condition,
to regain a state of operating equilibrium after being subjected to a physical disturbance, with most system
variables bounded so that practically the entire system remains intact [4]. The stability of a particular generator
or a group of generators is of interest. A remote generator may lose synchronism without causing cascading
instability of the whole system. Similarly, stability of particular loads or load areas are also of interest [4].
Voltage stability is the ability of the power system to maintain steady acceptable voltages at all the buses in the
system under normal operating conditions and after being subjected to a disturbance [5]. Voltage instability occurs
when a disturbance, increase in load demand or change in system condition causes a progressive and uncontrolled
drop in voltage. The main factor causing instability is the inability of the power system to meet the reactive power
demand. A possible outcome of voltage instability is loss of load in an area, or tripping of transmission lines and
other elements by their protective systems leading to cascading outages. A run-down situation causing voltage

[76]
instability occurs when load dynamics attempt to restore power consumption beyond the capability of transmission
network and the connected generation [5, 6].
A pattern recognition approach to detect the voltage instability in a power system in advance to its occurrence
was implemented. This could help the operators to take remedial actions, and thus prevent the oncoming voltage
collapse phenomenon. A pattern recognition software, Classification and Regression Trees (CART) and
Regularized Least Squares Classification (RLSC) were deployed to develop the voltage stable and voltage
unstable patterns. This was achieved by simulating voltage stable and unstable cases from a test system and then
training the pattern recognition models to come up with patterns. These patterns were later validated by testing
those using new cases which are not used for training.
Nigerian integrated 52 bus system was used as test system, it was modeled in Power System Simulator for
Engineering (PSSE), and contingencies are applied to come up with the training and testing cases. MATLAB was
used to develop RLSC algorithm and also to preprocess the data to be analyzed with CART.
METHODOLOGY
The method adopted to detect voltage collapse ahead of time consists of three phases as illustrated in the flow
charts of figures1 to 3 respectively.
1. Modeling and simulation of test system: Here the power system dynamic response is captured.
2. Data Processing and feature extraction: The system variables around the instance of a disturbance that
includes some pre-disturbance and post-disturbance variables are extracted and
3. Classification of voltage abnormality: Pattern recognition techniques phase-I of the methodology is applied
to capture the dynamic response of the system. Here, capturing the system response means recording the
system variables like bus voltage magnitudes, voltage angles, real and reactive powers, and generator rotor
angles.
Dynamic response of the system can be captured by any one of the following two methods:
a. Using the equivalent system models and dynamic simulation tools, such as:
 Model the equivalent system for dynamic simulations in power system simulation tools such as PSSE.
 Perform the dynamic simulations on the modeled system using dynamic simulation tools.
 Capture the system response for all the variables of the system.
b. Collecting data from the field using Phasor Measurement Units (PMUs):
System data are captured from the PMU installed in the substations. These PMUs will record system variables
with time stamp synchronized to Generating Plants Station (GPS) time clock
To detect voltage instability, the sequences of the first method are deployed on the Nigeria Integrated 330kV 52
bus power equivalent system as follows:
(a) Capturing test system dynamic response(phase1):Phase-I techniques have four steps to capture the test
system dynamic response and these steps are presented as follows:
(i) Test system modeled in Phase-I: Step A. Flow chart for this step is shown in figure 1. In this step, the
equivalent model data, such as, bus data and line data consisting of generator’
s output power, maximum and
minimum reactive power limit of the generator, MW and Mvar peak loads, impedance of the lines,
transmission line sizes, voltage and power ratings of the line and transformer data, and the nominal and critical
voltages of each of the buses for the test system were collected from PHCN, and modeled in power system
dynamic simulation software (PSSE). Multiple equivalent system models are created with different types of
loads like induction motor loads, composite load models, etc. Once the system models are built, a dynamic
simulation is performed on the system model for a base case (without applying any contingency). If the system
dynamic response is stable, then the tasks in the next step of Phase-I will be performed.

[77]
Figure 1: Flow chart for Phase-I: Step-A.
(ii) Phase-I: Step-B: In this step, the test system was stressed to determine critical contingencies by increasing
load or disconnecting generating plants or disconnecting transmission lines. Flow chart for Phase-I: Step-B
is shown in Figure 2. Contingency analysis was performed to determine the critical contingencies, and from
those contingencies, voltage stable and voltage unstable cases are determined.
Figure 2: Flow chart for Phase-I: Step-B.
Collect the test system
information
Build the test system in power
system simulators PSSE
Collect the test system
information
GO TO PHASE 1:B
Test System if stable?
Yes
No
Tweak the Modeling Parameters
Increase the load or disconnect
elements to stress the system
Perform contingency analysis on the
stressed test system
Determine the severe contingencies
Is contingency
voltage stable?
Classify the case as
voltage stable
voltage unstable
No Yes

[78]
(iii) Phase-I : Step-C : After determining the critical contingencies, dynamic simulations are performed on
voltage stable and voltage unstable cases in Phase-I: Step-C using the power system dynamic simulation
software. Flow chart for Phase-I: Step-C is shown in Figure 3. Initially, the base case will be prepared for
dynamic simulations, then dynamic simulations are performed on the contingencies found in Phase-I: Step-
B and system parameters are captured. Voltage stability will be determined from the dynamic simulations
performed on each contingency.
Figure 3: Flow chart for Phase-I: Step-C.
(iv) Phase-1: Step-D: In this step, system parameters or variables such as bus voltage magnitudes, voltage angles,
real and reactive powers, and generator rotor angles are collected, modeled and simulated in PSSE software.
(b) Data Processing and Feature Extraction (Phase-II): Phase-II has two stages to pre-process the captured
system dynamic response for Phase-III. Simulation parameters are extracted from the captured dynamic response
of the system around the disturbance. This extracted system variables have a few seconds of the system dynamic
response for pre-disturbance and post-disturbance data.
(c) Classification of Voltage Abnormality (Phase-III): The last phase of the method uses the pattern recognition
techniques to predict the voltage stability from the power system dynamic response. Flowchart for Phase-III is
shown in figure 4. Algorithmic and statistical pattern recognition techniques are programmed in this phase. These
programmed methods are trained using the training samples extracted from Phase-II and a classifier to predict
voltage stability was developed from the pattern recognition methods. This classifier was tested with the test
samples extracted from Phase-II. If prediction using the classifier is accurate and efficient, then the prominent
features and prominent system variables leading to this stability prediction are determined. If the developed
classifier from pattern recognition method is not accurate, then problems relating to this defective classifier, either
the pattern recognition method or the training samples used to train the pattern recognition method, are determined
and corrected. In the last stage of Phase-III, the accurate pattern recognition methods to predict voltage collapse
are determined.
3.1Pattern Recognition Procedure Formulation
Pattern recognition techniques such as, CART and RLSC are used to develop the voltage stable and unstable
waveforms, as shown figure 4.
Prepare the base cases for dynamic
simulation
Run the dynamic simulation scripts
on the base case to apply
disturbance
Collect the simulation data and
determine the stability
Is contingency
voltage stable?
voltage stable
voltage unstable
No Yes

[79]
Figure 4: Flow chart of pattern recognition model (Phase III)
Mathematical formulations involved in RLSC and CART are presented as follows:
(a) RLSC Mathematical Formulation:
RLSC is a learning method that obtains solution for binary classification via Tikhonov regularization in a
Reproducing Kernel Hilbert Space using the square loss function [7]. Let’s assume X and Y are two sets of random
variables, and the training set for pattern classification is S = (x1, y1), …., (xn, yn) and it satisfies xi ϵ Rn
and yi ϵ {-
1,1}. For all i, the main goal is to learn a function f(x) while minimizing the probability of error described by the
expression given in equation (1).
(1)
This can be found by solving the Empirical Risk Management (ERM) problem of finding the function in Hk
(Hypothesis space) which minimizes
Hk = (2)
Minimizing, the hypothesis space Ңk, for a fixed positive parameter γ, the regularized functional M2 is computed
as follows:
(3)
Where ‖Ϝ‖2
k is the norm in Ңk showing the
Reproducing Kernel Hilbert Space (RKHS), defined by the kernel .The solution for Tikhonov regularization
problem can be solved by Representer Theorem, given by
(4)

[80]
The first solution being the kernel matrix (K) is constructed from the training set S, as follows:
njiijkK  ,1)( (5)
),( jiij xxkk  (6)
The next step involves the computation of the vector coefficients c = (c1, c2,….cn)T
by solving the system of linear
equations as follows:
(7)
Where y = (y1, y2….yn)T
, I is the identity matrix of dimension n and finally, the classifier is:
(8)
The sign of this function )(xf decides the class (+1 or −1) for x and its magnitude is the confidence in this
prediction.
(b)Data Mining - CART:
Data mining is the process of extracting knowledge from data. The goal is to extract rules or knowledge from
regularity patterns exhibited by the data. Decision Trees (DTs) is a method used for Data Mining [8]. CART is
the methodology used to build the DTs.
3.2 Data Collection and Inputs:
The data used in this research were collected from Power Holding Company of Nigeria (PHCN) [9, 10, 11].
Table 1: Generating stations currently in operation in Nigeria.
S/N STATION STATE TURBINE INSTALLED
CAPACITY
(MW)
AVAILABLE
CAPACITY
(MW)
1 Kanji Niger Turbine 760 259
2 Jebba Niger Hydro 504 352
3 Shiroro Niger Hydro 600 402
4 Egbin Lagos Steam 1320 900
5* Trans-Amadi Rivers Gas 100 57.3
6* A.E.S (Egbin) Lagos Gas 250 211.8
7 Sapele Delta Gas 1020 170
8* Ibom Akwa-
Ibom
Gas 155 25.3
9* Okpai (Agip) Delta Gas 900 221
10 Afam I-V Rivers Gas 726 60
11* Afam VI(Shell) Rivers Gas 650 520
12 Delta Delta Gas 912 281
13 Geregu Kogi Gas 414 120
14* Omoku Rivers Gas 150 53

[81]
15* Omotosho Ondo Gas 304 88.3
16* Olorunshogo
Phase I
Ogun Gas 100 54.3
17* Olorunshogo
Phase II
Ogun Gas 200 105.5
Total Power 9,065 3,855.5
Note: Generating stations marked* are the completed and functional independent power generation already in the
Grid
Table 2: Buses for both existing and integrated 330kV power system project
S/N BUSES S/N BUSES S/N BUSES
1 Shiroro 21 New Haven
South
41 Yola
2 Afam 22 Makurdi 42 Gwagwalada
3 Ikot-Ekpene 23 Benin-Kebbi 43 Sakete
4 Port-Harcourt 24 Kanji 44 Ikot-Abasi
5 Aiyede 25 Oshogbo 45 Jalingo
6 Ikeja West 26 Onitsha 46 Kaduna
7 Papalanto 27 Benin North 47 Jebba GS
8 Aja 28 Omotosho 48 Kano
9 Egbin PS 29 Eyaen 49 Katampe
10 Ajaokuta 30 Calabar 50 Okapi
11 Benin 31 Alagbon 51 Jebba
12 Geregu 32 Damaturu 52 AES
13 Lokoja 33 Gombe
14 Akangba 34 Maiduguri
15 Sapele 35 Egbema
16 Aladja 36 Omoku
17 Delta PS 37 Owerri
18 Alaoji 38 Erunkan
19 Alaide 39 Ganmo
20 New Haven 40 Jos
Table 3: Transmission Line data of the 52-bus networks
S/N
Transmission line Line Impedance
B (pu)
From Bus To Bus R (pu) X (pu)
1 49 1 0.0029 0.0205 0.308
2 3 18 0.009 0.007 0.104
3 3 3 0.0155 0.0172 0.104
4 3 4 0.006 0.007 0.104
5 19 25 0.0291 0.0349 0.437
6 19 6 0.0341 0.0416 0.521
7 19 7 0.0291 0.0349 0.437
8 8 9 0.0155 0.0172 0.257
9 8 31 0.006 0.007 0.257
10 10 11 0.0126 0.0139 0.208
11 10 12 0.0155 0.0172 0.257
12 10 13 0.0155 0.0172 0.257
13 14 6 0.0155 0.0172 0.065

[82]
14 16 15 0.016 0.019 0.239
15 18 37 0.006 0.007 0.308
16 16 17 0.016 0.019 0.239
17 16 26 0.035 0.0419 0.524
18 16 3 0.0155 0.0172 0.257
19 19 21 0.006 0.0007 0.308
20 19 22 0.0205 0.0246 0.308
21 23 24 0.0786 0.0942 1.178
22 11 6 0.0705 0.0779 1.162
23 11 15 0.0126 0.0139 0.208
24 11 17 0.016 0.019 0.239
25 11 25 0.0636 0.0763 0.954
26 11 26 0.0347 0.0416 0.521
27 11 27 0.049 0.056 0.208
28 11 9 0.016 0.019 0.239
29 11 28 0.016 0.019 0.365
30 27 29 0.0126 0.0139 0.208
31 30 3 0.0126 0.0139 0.208
32 32 33 0.0786 0.0942 1.178
33 32 34 0.0786 0.0942 1.178
34 35 36 0.0126 0.0139 0.208
35 35 37 0.0126 0.0139 0.208
36 9 6 0.0155 0.0172 0.257
37 9 38 0.016 0.019 0.239
38 38 6 0.016 0.019 0.239
39 33 25 0.016 0.019 0.239
40 33 51 0.0341 0.0416 0.239
41 33 40 0.067 0.081 1.01
42 33 41 0.0245 0.0292 1.01
43 42 13 0.0156 0.0172 0.257
44 42 1 0.0155 0.0172 0.257
45 6 25 0.0341 0.0416 0.521
46 6 28 0.024 0.0292 0.365
47 6 7 0.0398 0.0477 0.597
48 6 43 0.0398 0.0477 0.521
49 44 3 0.0155 0.0172 0.257
50 51 25 0.0398 0.0477 0.597
51 45 41 0.0126 0.0139 0.208
52 51 47 0.002 0.0022 0.033
53 51 24 0.0205 0.0246 0.308
54 51 1 0.062 0.0702 0.927
55 40 46 0.049 0.0599 0.927
56 40 22 0.002 0.0022 0.308
57 46 48 0.058 0.0699 0.874
58 46 1 0.0249 0.0292 0.364
59 46 1 0.0205 0.0246 0.308
60 20 26 0.024 0.0292 0.365
61 20 21 0.0205 0.0246 0.308
62 50 26 0.006 0.007 0.104
63 26 37 0.006 0.007 0.104
64 3 21 0.0205 0.0246 0.257

[83]
Table 4: Load data.
3.3 Design and Simulation of Nigeria 330kV Integrated 52 bus Power Network Using PSSE:
Power System Simulator for Engineering (PSSE) is composed of a comprehensive set of programs for studies of
power system transmission network and generation performance in both steady-state and dynamic conditions. The
proposed method of determining voltage stability using Pattern Recognition was tested on Nigeria 330kV Integrated
52 bus power Network of figure 5. This network was designed in PSSE in a SIMULINK draft environment, as
shown in figure 6. The dynamics simulation procedure was followed, and voltage stable and unstable cases were
simulated by varying the load levels of table 4. The pattern recognition techniques were applied to extract the
knowledge in the simulated data.
A portion of Nigeria 330kV integrated 52-bus power network of figure 6 was chosen for voltage stability. The
transmission line branch from bus 6 (Ikeja West TS) to bus 9 (Egbin PS) was chosen for the study. The power flow
was solved for the base loading of the network. Initially, N-1 contingencies were applied to the power network,
which handles them effectively. This power network was stressed by increasing the real and reactive loads in steps
and contingencies are applied for each loading level to observe the stability of the system. For each simulation, the
following steps were adopted:
1. New loading condition was created by updating the real and reactive loads.
2. Power flow was solved.
3. The dynamic simulation was run for 10 seconds, without any contingency.
4. Contingency was applied after 10 seconds.
5. The simulation was continued till 100 seconds.
Bus No. Load Data Bus No. Load Data
P(MW) Q(Mvar) P(MW) Q(Mvar)
1 0.00 0.00 27 0.00 0.00
2 40.00 -10.00 28 0.00 0.00
3 0.00 0.00 29 120.00 80.00
4 140.00 30.00 30 130.00 -78.00
5 90.00 30.00 31 0.00 0.00
6 160.00 70.00 32 200.00 67.00
7 0.00 0.00 33 0.00 0.00
8 130.00 70.00 34 0.00 0.00
9 300.00 90.00 35 0.00 0.00
10 210.00 40.00 36 0.00 0.00
11 0.00 0.00 37 0.00 0.00
12 50.00 -20.00 38 0,00 0.00
13 100.00 -30.00 39 0.00 0.00
14 120.00 60.00 40 0.00 0.00
15 500.00 50.00 41 0.00 0.00
16 250.00 43.00 42 0.00 0.00
17 70.00 38.00 43 0.00 0.00
18 0.00 0.00 44 0.00 0.00
19 200.00 55.00 45 0.00 0.00
20 150.00 35.00 46 0.00 0.00
21 0.00 0.00 47 0.00 0.00
22 0.00 0.00 48 0.00 0.00
23 300.00 45.00 49 0.00 0.00
24 0.00 0.00 50 0.00 0.00
25 100.00 58.00 51 0.00 0.00
26 0.00 0.00 52 0.00 0.00

[84]
A python automation file was used for this purpose. The bus number was specified to accommodate load change
and the amount of load (P and Q) for each run.
The bus at which the load is to be increased was determined. The load (P and Q) was then increased in steps and
for each step, a specific contingency was applied and the stability was observed. Once an unstable case is
encountered, the details are documented. Once the contingencies to be applied for that bus are completed, the
analysis continued with the next bus.
Figure 5: Nigeria 330kV Integrated 52 bus Power Network one line diagram

[85]
RESULTS AND DISCUSSION
The results obtained in this work showed the training cases of the RLSC and CART, voltage stable and unstable
cases of 17 generators of Nigeria 330kV Integrated 52 bus Power Network.
Table 5: Training cases result of the RLSC algorithm and CART.
Case
#
Bus
No.
P (MW) Q
(MW)
Contingency Stability
1 4 600 350 Line 6-11 Stable (+1)
2 4 550 300 Line 3-4 Stable (+1)
3 4 600 350 Line 3-4 Stable (+1)
4 4 700 450 Line 4-14 Stable (+1)
5 7 233.8 84 Line 8-9 Stable (+1)
6 7 233.8 84 Line 5-8 Stable (+1)
Figure 6: Simulink draft of the integrated 52 power bus network using PSSE

[86]
7 7 400 350 Line 5-8 Stable (+1)
8 7 400 250 Line 6-11 Stable (+1)
9 7 233.8 84 Line 6-11 Stable (+1)
10 8 522 176 Line 8-9 Stable (+1)
11 8 600 300 Line 8-9 Stable (+1)
12 4 700 400 Line 3-4 Unstable (-
1)
13 4 750 450 Line 6-11 Unstable (-
1)
14 4 750 500 Line 4-14 Unstable (-
1)
15 7 400 250 Line 8-9 Unstable (-
1)
16 7 500 260 Line 5-8 Unstable (-
1)
17 7 450 250 Line 6-11 Unstable (-
1)
Table 5 showed the training cases result of the RLSC algorithm and CART. Five more cases were simulated in
order to test the pattern recognition model developed from the training cases. The loading conditions and the
contingencies applied for each testing case are shown in Table 6.
Figure 7: Voltage magnitude of training case 3 showing stable voltage (+1).
The voltage magnitude plot for case 3 showing stable voltage (+1) is shown in Figure 7. The load at bus 4 was
increased and line 3 to 4 was disconnected after 10 seconds of dynamic simulation. Due to the loss of line 3 to 4,
the voltage at bus 4 drops considerably as shown in the plot. This was due to increase in reactive power demand.
The terminal voltages of generators 2 and 3 are restored by the excitation control. The increased reactive power
demand was supplied by these two generators.
The Under Load Tap Changer (ULTC) connected at generator 8 between bus 6 and bus 31 operates after 20
seconds and the voltage at controlled bus, which is bus 6 was increased. This result showed an increase in the
voltage level at bus 7. The voltage magnitude settles near 0.955pu as shown in figure 7.

[87]
Figure 8: Voltage magnitude of Training Case 12 showing unstable voltage(-1).
The voltage magnitude plot for case 12 showing unstable voltage (-1) is shown in Figure 8.The load at bus 4 was
increased, which stresses the system. After 10 seconds of simulation, the line3 to 4 was disconnected. This result
showed a considerable decrease in voltage level at bus 4 due to the increase in reactive power demand. The
terminal voltages at generators 2 and 3 are restored by the action of excitation control. The extra reactive power
was supplied by these two generators. The Under Load Tap Changers (ULTCs) connected at generators 2 and 3
restored the voltages at the controlled buses (bus 6 and bus 9). With each tap change, the line I2
X and I2
R losses
increase, which in turn decrease the voltage magnitude.
Table 6: Voltage stability result of the testing cases.
Case # Bus P (MW) Q (MW) Contingency Stability
1 8 522 176 Line 5-8 Stable (+1)
2 8 700 450 Line 5-8 Stable (+1)
3 8 522 176 Line 6-11 Stable (+1)
4 8 650 350 Line 8-9 Unstable (-1)
5 8 700 650 Line 5-8 Unstable (-1)
Figure 9: Voltage magnitude of testing Figure 10: Voltage magnitude of testing
Case 1 showing stable voltage(+1). case 5 showing unstable voltage (-1).
The 17 training cases shown in Table 5 are used to train the RLSC algorithm. The data used was voltage magnitude
from 0-15 seconds, the contingency was applied after 10 seconds. The developed model was used to predict the
stability of the testing cases shown in table 8. Graphs showing the voltage magnitude testing case 1 for stable
voltage(+1) and testing case 5 for unstable voltage(-1) as shown in table 8 are presented in figures 9 and 10
respectively.

[88]
CONCLUSION
The current Nigeria power grid is highly stressed due to the increased demand in electrical power without
considerable addition of transmission lines. This brings a situation where the hazards of voltage collapse are highly
possible to occur. This paper presents pattern recognition techniques, such as, Regularized Least Squares
algorithm (RLSC) and Classification and Regression Trees (CART) for detecting voltage collapse ahead of time
in power systems. This helps the operators to take remedial actions and eventually prevents the system from
collapsing.
The work introduced the concept of the pattern recognition techniques in conjunction with Power System
Simulator for Engineering (PSSE) and applied these techniques in voltage stability analysis for control of the
Nigeria 330kV Integrated 52 bus Power Network. The method predicted correctly the voltage stability and
instability cases for +1 and -1. The captured dynamics of the system within 10 seconds after a contingency was
used to determine and detect the patterns. The three phases involved in achieving the voltage stability and
instability cases of the test power network were presented.
The results of the dynamic simulations of the Nigerian integrated 52 bus test system and the system voltage
stability for those simulations were obtained. The voltage stable and unstable dynamic simulation cases were
divided into training and test set cases to validate the proposed methodology. CART and RLSC pattern recognition
algorithms were trained using the training set and the trained algorithm was tested using the test set of the dynamic
simulations.
REFERENCES
1. Galiana, F. D., “Load Flow Feasibility and the Voltage Collapse Problem”, IEEE Proceedings of 23rd
Conference on Control and Design, pp. 485-487, Dec. 1984.
2. Tamura, Y., Mori, H. and Iwamoto, S., “Relationship between Voltage Instability and Multiple Load
Flow Solutions in Electric Power System”, IEEE Trans. on PAS, no. 5, pp. 1115-1125, May 1983.
3. Yorion, N. , “An Investigation of Voltage Instability Problem, Transactions on Power Systems”, vol. 7,
no. 2, pp. 600-611, May 1992.
4. Haque, M. H. and Pothula, U.M. R. “Evaluation of Dynamic Voltage Stability of a Power System”,
International Conference on Power System Technology - POWERCON, Nov. 2004.
5. Kundur, P., “Power System Stability and Control, Surrey”, British Columbia: McGraw-Hill, Inc, pp. 42-
56, Sept.1995.
6. Jain, A. K. R., Duin, R.P. W and Mao, J. “Statistical Pattern Recognition: A Review”, IEEE Transactions
on Pattern Recognition and Machine Intelligence, vol. 22, no. 1, pp. 4-37, Jan. 2000.
7. Wadsworth International Group, “Classification and Regression Trees”, Belmont, Calif, 1984.
8. Munong, C., “Dynamic simulation of Electric Machinery using Matlab/Simulink”, School of Electrical
and Computer Engineering Purdue University West Lafayette, Indiana, Prentice Hall PTR, 1998.
9. Power Holding Company of Nigeria (PHCN), “Annual Report on Generation Profile of Nigeria”, 2016.
10. Power Holding Company of Nigeria (PHCN), “Transysco Daily Logbook on Power and Voltage
Readings at Various Transmission Station”, 2016.
11. National Control Centre, Generation and Transmission Grid Operations, “Annual Technical Report”,
2016.

VOLTAGE STABILITY IN NIGERIA 330KV INTEGRATED 52 BUS POWER NETWORK USING PATTERN RECOGNITION TECHNIQUES: A RECURSIVE APPROACH

More Related Content

What's hot

Similar to VOLTAGE STABILITY IN NIGERIA 330KV INTEGRATED 52 BUS POWER NETWORK USING PATTERN RECOGNITION TECHNIQUES: A RECURSIVE APPROACH

More from Onyebuchi nosiri

Recently uploaded

VOLTAGE STABILITY IN NIGERIA 330KV INTEGRATED 52 BUS POWER NETWORK USING PATTERN RECOGNITION TECHNIQUES: A RECURSIVE APPROACH