A Review on Application of Soft Computing Techniques for Load Shedding in Power Systems

Abstract

In power systems, system frequency decreases if load exceeds generation and increases when power generation is greater than load demand. Load shedding is highly required in power systems to stabilize frequency in electrical power systems. Though conventional load shedding techniques are being used in many power system applications, soft computing techniques such as artificial neural networks (ANNs), fuzzy logic, genetic algorithm and particle swarm optimization have been presented by many researchers to provide optimum load shedding. However, there are many problems associated with machine learning based control techniques in real time. In this work, the utilization of soft computing algorithms and its benefits and drawbacks are discussed. A comprehensive survey is presented about the soft computing based load shedding and comparison is made among these techniques and conventional techniques. Neural network based learning rule is applied with back propagation network to minimize the error. In fuzzy based technique, load shedding position at each node was predetermined by applying certain fuzzy rules. Genetic algorithm and particle swarm optimization are robust and applied to solve many nonlinear and multi-objective problems. ANN and fuzzy logic are combined to build an adaptive neuro fuzzy inference system to provide accurate load shedding.