Box-type substation fault prediction strategy based on GA-BP network

Abstract：Box-type substations are widely used in industry, commerce, urban power transmission and distribution system.They play an important role in the power system.It is very significant to the fault diagnosis of box-type substations.By studying the internal structure and working principle of the box-type substation, and by analyzing the fault and the fault characteristics of the box-type substation, a fault diagnosis network model combined with genetic algorithm and BP network was proposed to train the system data.The BP network was optimized by using the optimal global search function to avoid the disadvantages of BP algorithm learning, so that the model had good convergence and adaptability.The simulation results showed that the network had a good recognition effect and it was very promising in the fault prediction of box-type substation.

Keyword：box-type transformer substation;fault diagnosis;genetic algorithm;global optimization;BP network;

Received：2019-01-23