QUALITY ASSESSMENT OF POWER SYSTEM USING ARTIFICIAL NEURAL
NETWORK
S. GUPTA & J. SHUKLA
Professor, Department of Electrical, NIT, Raipur, Chhattisgarh, India
ABSTRACT
In recent years Artificial Intelligence has been proposed as an alternative tool to provide quick solutions to certain
difficult power system problems. With intelligent systems, such as ANN, the available information regarding the power
system can be stored and retrieved as a part of new solution process. ANN computing may require larger time for off line
training but is capable of giving instant response for a given condition and hence, is suitable for on-line applications.
This paper describes ANN base technique for obtaining the quality of interconnected power system. Desired level
of voltage has been maintained using back propagation algorithm base network, which is trained to represent the system as
ANN model. The well-trained model can use to predict the quality and the losses of the power system. This trained model
can also predict the behavior of the network as load increase.
KEYWORDS: Artificial Intelligence, ANN, Alternative Tool, Off Line Training, Quality of Interconnected Power
System
INTRODUCTION
There has been a remarkable progress in the development of software and hardware for the design and analysis of
power systems. However, much still depends on judgment of human experts. Experienced planning and design personnel
make efficient and viable decisions on the basis of their comprehensive experience and the knowledge of prevailing
circumstances.
This paper mainly gives the description of the concept of Artificial Neural Network in the field of assessment of
power system. The IEEE-30 bus model has been taken for the analysis of voltage, phase angle and losses variation.
Simulated model has been trained for the production of quality of load flow.
Artificial Intelligence based Demand Side Management has been developed by I.E.Hopley et al [1] and
A.S.Zadgaokkar [2] to support energy management in power system during abnormal situations. Hoyong Kim et al [3] and
E.A. Mohamed [4] suggested Artificial Neural Network based feeder reconfiguration for loss reduction in distribution
systems.
ANN Models have been suggested to give direct and accurate results nearly instantly for the sample IEEE 30-Bus
power systems. The study includes bus voltage distribution and phase angle.
ANN FOR LOAD FLOW SOLUTION
The conventional solution techniques use series mode of iterative calculations that makes it more time consuming.
The strong parallel processing capability of Artificial Neural Network is therefore applied to solve the load flow problem.
The ANN technique is totally different from the conventional techniques. The ANN model does not at all represent the
inherent intricate mathematical non-linear relationships among the input-output parameters constituting the power system
International Journal of Electrical and
Electronics Engineering Research (IJEEER)
ISSN(P): 2250-155X; ISSN(E): 2278-943X
Vol. 3, Issue 5, Dec 2013, 39-44
© TJPRC Pvt. Ltd.