Neural Network Based on Immune Algorithm for Restoration in the Power Distribution System

The power distribution system plays an important role in transmitting power to users. When the power distribution system is pressed by the typhoon, how to choose a good restoration and optimal orders appears to be particularly important to dispatchers. This paper proposes a new restoration strategy and optimal method for components' repair orders, - that is, artificial neural network (ANN) based on the immune evolutionary algorithm. The proposed restoration strategy involves in load level, line power flow, and degree of reliability. The immune evolutionary algorithm is implemented by calculating and searching for the optimal solution of affinity between the antigen and antibody, where the mean square error between actual and designed output values of the neural network is regarded as the antigen, the weight and bias of the neural network is served as the antibody. Furthermore, the neural network is trained with those optimal weights and biases to predict components' repair orders. Compared with classical back-propagation and genetic algorithm, the proposed method has shown better performance for components' repair orders.