Study of Method on Genetic BP Networks for Vibration Fault Diagnosis of Turbogenerator

摘要

Considering the complexity and relevance of fault diagnosis for turbo-generator, the paper makes diagnosis by adopting improved genetic BP network algorithm. In order to solve the problems of slow network learning and tendency of minimum point in BP algorithm, the structure and specific parameter of BP network optimized by genetic algorithm were used in the discussion of a model integrated with adaptive genetic neural algorithm, which was applied in the fault identification of turbo-generator. Experimental data shows that the algorithm is characterized by high convergence rate, effective vibration fault diagnosis for turbo-generator and relative high reliability and practicability.