Fault Diagnostic Method for Micro-grid Based on Wavelet SOM Neural Network and Multi Agent System

Abstract

Abstract. Fault diagnosis using traditional neural network method requires training samples to train the neural network. The micro grid system has the characteristics of flexible operation mode and variety of topology structure, the fault diagnosis method of it by using neural network has the problem of poor adaptability and requires a large number of training samples. A fault diagnosis method combining multi agent system with wavelet som neural network is proposed. The wavelet som neural network can judge the reason of fault and the multi agent system judge the location of the fault. A micro grid simulation system is established based on PSCAD. The simulation results prove the feasibility of the fault diagnosis method based on multi agent system and wavelet som neural network. the wavelet som neural network only need typical fault training samples, this method will not affected by fault location, fault time and other factors, it has good adaptability to the change of the topology structure of micro grid system.

Based on the Digital Platform of Road Construction Machinery and Equipment Maintenance Management Methods

“…Although this model is less risky and difficult, it also incurs higher service costs. [8][9] (3) Combination mode. When maintaining and detecting faults of highway electromechanical equipment, it is necessary to choose a reasonable implementation mode according to the specific situation of electromechanical equipment and its own operation and maintenance capacity.…”