Study On Wind Turbine Vibration Monitoring And Fault Diagnosis System

The structure of wind turbine drive system is large and complex, whose faultdetection and repair would take long time and the impact on electric power productionis also great. Therefore, it has important practical value for early warning and faultdiagnosis on the drive system of wind turbine. Based on signal analysis and artificialintelligence technology, the wind turbine drive system of vibration monitoring andfault diagnosis system was studied.Firstly, in this paper，it takes a type of1.5MW wind turbine gearbox and bearingas the study object, calculates the fault characteristic frequency of gearbox andbearing, analyses vibration signal on the high speed shaft measuring point of gearbox,extracts the characteristics of three kinds of fault signal. The result of analysis isconsistent with fault check out of the box, laid the foundation for the faultfeature of the knowledge base of expert system.Secondly, the knowledge base and reasoning machine of expert system wereconstructed. In this paper，combined with the signal processing, the differencesbetween the normal and fault parts of the kurtosis index and the spectral energydistribution, information entropy were analyzed, the knowledge base were constructedwhose monitoring parameter is kurtosis index and fault characteristic parameter isspectrum energy value and the information entry. A BP network is taken as thereasoning machine of expert system, the extracted fault characteristic parametersof gearbox vibration signal as the input samples for training and testing. The resultsshow that the inference engine can achieve the aim of fault diagnosis.Finally，based on LabVIEW programming langue, the state monitoring and faultdiagnosis software interface of wind turbine drive system was designed. Softwarefunction includes time domain waveform display, spectrum analysis, noise reduction,fault warning, fault diagnosis, data management etc.