Transformer is an indispensable important equipment in power system

its safe and stable operation is very important for the whole system. In order to find out and eliminate the faults of transformer in time

this paper proposes a fault diagnosis method of power equipment based on Improved BP neural network. By comparing the improved methods of additional parameters

optimizing learning factors and hidden layer number and traditional BP neural network algorithm

we can find that the convergence speed has been greatly improved and the local minimum problem has been solved. The simulation results show the feasibility of the improved method.