Fault Diagnosis Of Power Transformer With Genetic Algorithm And Neural Network

China's power system is huge in scale.The reliability and security of power system operation directly affects the lifeline of the national economy,as well as people's daily lives and property safety.In recent years,China's economic development has developed steadily and rapidly,and has also promoted the transformation of the power industry.The power system is developing in the direction of UHV,smart grid,new energy,and clean energy.The power transformer is one of the most important electrical equipments in the transmission and transformation equipment of the power system.It is also one of the electrical equipments with large capacity and high accident rate in the power system.Ensuring the safe operation of the power system will directly affects the safe and stable operation of entire power system.Detecting the potential faults of the transformer at any time and eliminating hidden dangers in time to ensure the safe operation of the transformer is an important method to improve the reliability of the power supply,which it is also a major theoretical research topic in the power system.This thesis compares commonly used power transformer fault diagnosis methods,and proposes an improved BP neural network algorithm based on genetic algorithm,which is applied in power transformer fault diagnosis.The simulation results indicate that the method proposed in this paper has a greater improvement than the traditional method in terms of fault speed and diagnostic accuracy.It can improve the safety of the power transformer and has important practical significance.This thesis analyzes the basic structure,principle and advantages and disadvantages of BP neural network and genetic algorithm,and proposes an algorithm that combines genetic algorithm and BP neural network for power transformer diagnosis,we use MATLAB software to simulate the network and the BP neural network based on the improved genetic algorithm,respectively.The influence the selection of different parameters on the training effect of the network is given.The data analysis results show that the improved BP neural network based on genetic algorithm has greater advantages in training speed and diagnostic accuracy than BP neural network,which makes up for the slow convergence speed and local minimum of objective function in BP neural network.