Research On Fault Diagnosis Expert System For Wind Turbine Gearbox Based On Bayesian Network

Sophisticated, complex and highly integrated represents the developmentdirection of modern industrial equipment. In such a highly demanding industrialequipment today, safety and stable operation of the device into the normal productionof the key. As a fan gearbox during operation of key components, mounted on top ofthe fan is more narrow position, the event of failure, maintenance is very difficult,their health is directly related to the wind power production safety and economicbenefits. So the wind turbine gearbox failures quickly and accurately diagnose andcorrect handling of great practical significance and economic value. In this paper,wind turbine gearboxes for the study, especially the intelligent diagnostic expertsystem technology has been studied as a starting point for its fault diagnosis method.The main contents are as follow:First, For large wind power equipment, the causal relationship between thefailure showed uncertain mappings. Actual diagnosis process, fan failure informationis often exhibit a strong coupling and fuzziness, making uncertainty among fansshowing signs of failure and cause of the malfunction. This paper resolve thisuncertainty, combined with the advantages of Bayesian networks in dealing with theproblem of uncertainty, proposed wind turbine gearbox fault diagnosis Bayesiannetwork approach. Bayesian theory to expert systems, to effectively solve the faultdiagnosis of turbine gear boxes uncertainty problem.Second, This paper combine Bayesian networks and expert systems, focus ontwo core wind turbine gearboxes based on Bayesian network fault diagnosis expertsystem were studied, namely the establishment of wind turbine gearboxes and fangear Bayesian network knowledge base box Bayesian network inference. First,summed up the reason for wind turbine gearboxes in conditions of complex faultpatterns and the impact of, and basis for the analysis of failure mechanisms in theestablishment of a fault tree, fault tree analysis method for its failure analysis.Secondly, on the basis of fault tree analysis describes the graphic description of the Bayesian network model, probabilistic reasoning and other related theories.Describes the correlation between the fault tree analysis method and the Bayesiannetwork and introduced fault tree into fault Bayesian network steps to complete theestablishment of a Bayesian network model of wind turbine gearbox fault diagnosis,and Bayesian networks build the knowledge base. Finally, specific examples of windturbine gearbox failure describes the theory and Bayesian network inferencereasoning process.Third, Use the Microsoft Visual Studio.NET development tools to complete thefan gear box fault diagnosis based on bayesian network module design anddevelopment, to achieve the wind turbine gearbox fault analysis and diagnosis, andthrough an instance running verify the validity and accuracy of the system diagnosticresults.