Application And Research On Neural Networks And Expert System In Power System Transmission Equipment Protection

With the expansion of the scale of power system in our country,the requirement of system supply power quality is higher,and the security and stability of power supply system must be guaranteed,which makes the importance of transmission line fault diagnosis and relay protection in safe and stable power system operation growing.In this paper,the fault diagnosis and distance protection of power transmission lines are studied by using the BP neural network and expert system in artificial intelligence control technology.The main research on the neural network,the expert system as the assistant.Part only in the original expert system based on expert system to join Prolog program language for different power system short circuit fault using facts,rules,and asked to distinguish,to enhance the processing ability of expert system in fault diagnosis.Firstly,the research background of transmission line fault diagnosis and relay protection and the development of transmission line relay protection and artificial intelligence control system are introduced.Secondly,the basic principle of the traditional distance protection of power system and the required relay components are expounded,and the defects of traditional distance protection are analyzed.This paper describes the basic fault type of transmission line,analyzes the characteristics of each fault type and the process of fault phase discrimination.Thirdly,analysis of basic principle and characteristics of neural network expert system,introduce the BP algorithm and improved,on this basis,the design used in transmission line fault diagnosis,and distance protection of neural network and expert system.Using Prolog to distinguish different short circuit in power system according to facts,patterns and inquiry,enhance the solving capacity of the fault diagnosis of the expert system.Finally,the application of neural network and expert system in transmission line fault diagnosis and distance protection is analyzed and the single phase grounding fault,two phase ground fault,and two phase short circuit fault and three phase short circuit fault model are set and analyzed using EMTP software.Using the data obtained from the simulation measurement,with the help of BP neural network training function in matlab,some data are trained.In the building process of neural network fault diagnosis,according to factors of iterations and target error,from contents of 3.1.3,using normalized conjugate gradient method,variables,traingdx and Fletcher-Reeves to test.Through the comparison of training result from different training function,the practicability of neural network in transmission line fault diagnosis.Fletcher-Reeves is the most suitable BP modified algorithm;Then analysis the simulation results of neural network in the distance protection,step tangent trainoss quasi-newton method and normalized conjugate gradient method trainscg compared with other algorithms,can better meet the power system relay protection sensitivity,rapidity,accuracy requirements...