Research On Fault Damage Mechanisms And Diagnostic Methods For High-speed Shafting Of Wind Turbines

In recent years,China has developed rapidly in the field of wind power generation,and its installed capacity and scale have been greatly improved.However,after years of installation and operation of the wind turbine,high-speed shaft components in the transmission chain failed frequently.How to reduce the sudden accident rate and downtime of high-speed shaft components of wind turbine drive chain,thereby reducing maintenance costs and improving economic benefits,has become an urgent problem to be solved in wind power investment construction,operation and maintenance.The fault damage mechanism and automatic diagnosis method for high-speed shaft components of wind turbines are studied.The research based on fault damage mechanism provides a theoretical basis for the improvement of high-speed shafting technology of wind turbines.Finally,an automatic fault diagnosis system is constructed.The test results show that the diagnostic system can significantly improve the fault diagnosis accuracy of wind turbines with high shafting.The main research contents are listed as follows:(1)Literature investigation and field test for the main faults of high-speed shafting of wind turbine drive chain are carried out,and classification is conducted based on the effects of fault failure.Typical faults that have great influence on equipment operation as the research object of fault damage mechanism are found out.The fault phenomenon and fault performance characteristics,damage internal factors and external factors are proposed,and the solution is proposed and verified.(2)Based on the fault performance characteristics of the typical fault damage mechanism of high-speed shafting of wind power transmission chain,the sensor and the installation technology,signal transmission and processing technology under certain conditions are studied.The time and frequency domain feature extraction technology of low frequency vibration is studied and automatic extraction of fault features such as unbalance,misalignment and looseness is realized.The fault impact feature extraction method based on generalized resonance and resonance demodulation technology is studied.Temperature and oil metal particle processing methods are studied and automatic extraction of temperature and oil metal particle signal characteristics is realized.Synchronous extraction of data characteristics of operating conditions such as speed,wind speed and power is studied.The operating condition data characteristics are extracted and used as auxiliary feature information for fault diagnosis.(3)The fault diagnosis automatic diagnosis comprehensive decision model using multi-physical information fusion is constructed by “multi-point verification method” based on probabilistic neural network,multi-feature weighted fault diagnosis method based on case and expert experience,multi-source information fusion and improved DS evidence fault diagnosis method.The diagnosis result of the automatic diagnosis system constructed by the model through multiple single physical quantity monitoring case data and working condition data collected in the field is verified.Results show that the diagnosis system can significantly improve the fault diagnosis accuracy.