Applications of evolutionary algorithms and neural network for congestion management in power systems

Abstract

In power Systems, under deregulated environment, transmission line congestion has become more intensified and recurrent, as compared to conventional regulated power System. One of the most practiced and an obvious technique of Congestion Management (CM) is rescheduling the power outputs of generators in the System. In this work, an efficient method for managing the congestion is presented for a day ahead electricity market using generation rescheduling as corrective measures. The objective of the study is to find the optimal generation values that minimize the total cost for congestion management based on the bids submitted by Generation Companies (GENCOs). The CM problem is solved using various evolutionary methods like Evolutionary Programming (EP), Real Coded Genetic Algorithm (RCGA), Particle Swarm Optimization (PSO), Hybrid Particle Swarm Optimization (HPSO) and Differential Evolution (DE). Their performances are tested on three test Systems, namely WSCC Nine Bus, IEEE30 Bus and IEEE118 Bus Systems. Through the test results, it is proved that the proposed DE yields best solutions than the other approaches and retains consistency and accuracy even for a large System. However, to minimize the computational time for large System, Cascade Neural Network (CNN) is proposed for congestion management in real time applications. For training the proposed CNN, two training methods Back Propagation (BP) and Radial Basis Function (RBF) are applied and the network is tested with the above-said test Systems. From the test results, it is shown that RBF learning method is found to be faster than BPN method, and also the classification accuracy is improved in RBF. Moreover, angular distance based clustering method is applied for selecting the input features of IEEE118 Bus System in order to reduce the dimension of neural network structure. The proposed methods can be applied for congestion management in off-line and on-line mode as well. newline