Research On BP Neural Network Optimization Algorithm For Power Grid Fault Diagnosis

With the rapid development of economy,the development of China's electric power industry in the direction of greater capacity and higher level of automation.The requirements for the stability of the power network are getting higher and higher,and the faults in power grid operation bring huge losses to the power supply companies and the users.In order to reduce the fault time of the grid,while enhancing the reliability of power supply,after a power grid breakdown,We should be find the fault location in time and accurately,isolation fault components,Eliminate the hidden trouble,As to improve the safety and reliability of the system,Take appropriate measures to restore the power grid operation at the same time.BP neural network is one of the most widely used neural network models,It has good self-learning ability and adaptability and generalization ability,But the BP neural network algorithm is based on gradient method,In the process of computing,it is easy to fall into the local minimum,At the same time,when the number of learning samples is relatively large,and the relationship between input and output is more complex,the network will appear slow convergence speed,convergence precision is not high,and even do not converge.GSO has the ability of global optimization,using the GSO to optimize the initial weight and threshold of BP neural network and BP neural network can avoid falling into local minimum problem,Effectively improve the accuracy of fault diagnosis.However,the traditional algorithm has its own shortcomings,in the process of global optimization,it is easy to get the local optimal solution,and the problem of premature.In order to improve the global searching ability of the GSO,the problem of premature convergence is overcome.We introduce a variable step size based on traditional GSO,when the GSO individual did not find the best individual in the perception range,by moving a step,thus changing the neighborhood,find the local optima in the new range.The global optimization ability of the algorithm is improved.In this paper,we propose an improved algorithm to optimize BP neural network.When the individual is not found in the sensing range,a moving step is updated by introducing a variable step size factor.When the individual is not found in the sensing range,a moving step is updated by introducing a variable step size factor,Thus,the local search ability of the algorithm is improved,and the global optimization ability is improved.The combined algorithm is applied to the fault diagnosis of the power grid area,and a good diagnosis effect is obtained.