ANN/ DT approach for security evaluation and preventive control of power systems

  • Authors:

  • M. K. Shah;K. R. Niazi;C. M. Arora

  • Affiliations:

  • Department of Electrical Engineering, Malaviya National Institute of Technology, Jaipur, India;Department of Electrical Engineering, Malaviya National Institute of Technology, Jaipur, India;Department of Electrical Engineering, Malaviya National Institute of Technology, Jaipur, India

  • Venue:
  • CIMMACS'07 Proceedings of the 6th WSEAS international conference on Computational intelligence, man-machine systems and cybernetics
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Security evaluation is becoming a major concern in the real time operation of modern power systems which are large and complex. Neural Networks have shown great promise as a means of predicting security of large power systems. But because of their black box type nature, neural network based security classifiers are not able to provide information about preventive control if required following a contingency. The Decision Tree based classifiers, on the other hand, are known for their interpretability and therefore can be used to design preventive control strategy. However, the DT approach suffers from the drawback that as the size of power systems grows, the complexity of DT classifier becomes high. Building of complex DT classifier requires a prohibitively large learning set unless some compromise is made on the accuracy. This paper presents a hybrid approach for on-line security evaluation and preventive control of power systems, which combines ANN and DT approaches to exploit their potential while suppressing their drawbacks. It applies an ANN based classifier for security evaluation of power systems. If an operating state of power system is found to be insecure, a DT classifier is applied to infer preventive control measures. The accuracy of the DT classifier does not significantly affect the overall performance of this approach. The method has been applied on an IEEE power system and the results obtained are promising.