An improved genetic algorithm based on fuzzy inference theory and its application in distribution network fault location

Abstract

An improved genetic algorithm based on fuzzy inference theory is proposed in this paper, which is applied in faulty location of distribution network. This algorithm improves the problem of faulty operation caused by the distorted signal transmission in traditional faulty location method and the problem of premature convergence and slow convergence speed of standard genetic algorithm. Firstly, a mathematical model for fault location of distribution network is established by the switching function which contains multivariate equality constraints. On this basis, the improved genetic algorithm adopts optimal individual reserve strategy in selection operation and combines adaptive strategy with fuzzy inference theory to calculate cross-over operator and mutation operator, which improves the ability of search to avoid premature convergence. Finally, simulated annealing algorithm is integrated in the algorithm to accelerate the convergence speed. The simulation examples proposed in the paper verify the accuracy and effectiveness of the improved genetic algorithm applied in fault location of distribution network.