Optimization Of Bp Neural Network Based On Cloud Genetic Algorithm For Fault Location Of Distribution Network

When the grounding fault occurs in the distribution network, the voltage of the system will be greatly reduced, which will bring a great impact on the social production and people’s life. How to ensure the safety and reliability of the distribution network has become the subject of the power industry practitioners have to face. However, with the increase in electricity species,the voltage level of distribution network continues to rise, the structure of the distribution network is more and more complex. Although the State Grid Corporation has increased input costs to upgrade the grid, but due to the aging of the line, human error or natural disasters and other reasons for the failure of the distribution network is not possible. Therefore, how to improve the speed and effectiveness with diagnostic grid failure occurs has been a hot research topic.In this paper, the method of fault diagnosis of distribution network is summarized, and it is found that there are some problems in the application of BP neural network in the fault diagnosis of distribution network. Although BP neural network has the advantages of artificial intelligence, it is generally able to meet the needs of fault diagnosis after full study. However, it may not be fast enough in the diagnosis process, the diagnostic accuracy is not able to meet the requirements, but also may sink into local optima.Aiming at the problems of BP neural network in the fault location in distribution network, this paper will used in global search, good application effect, genetic algorithm and BP neural network combination, through the change of BP neural network model in the initial weights and threshold method to improve the effect. At the same time, the genetic algorithm is improved in this paper. In the genetic algorithm, the cloud theory which can effectively deal with the fuzzy information is introduced to form the cloud genetic algorithm, so that the effect of the genetic algorithm itself is improved.In order to verify the effectiveness of the improved BP neural network algorithm. In this paper, the standard BP neural network is applied to a simple distribution network system,which is simulated by the simulation software of MATLAB, the validity of fault diagnosis in distribution network is explained. And then the BP neural network improved by cloud genetic algorithm is applied in the same distribution network system, through the comparison of the two training curves, and the final diagnosis of actual output value, can be found although there is no big difference between the accuracy of diagnostic results of the two methods can meet the requirements, but the rate of the diagnosis of the improved BP neural network to be much faster than the rate of the standard BP neural network. Then by the with a BP neural network model and enhance the learning samples, and through MATLAB simulation found that study because of the increase in the number of samples, the diagnostic efficacy of the standard BP neural network to than improved BP neural network a lot worse. Therefore, it can be concluded that the improved BP neural network by the cloud genetic algorithm is very effective. It is able to improve the speed and accuracy of BP neural network in the application of fault diagnosis in distribution network.