Data Driven Diagnosis Method for Wind Turbine Faults Under Model Mismatch
Under the condition of mismatch between control model and actual data of horizontal axis wind turbine system,any operation failure event of the system will cause the system to deviate from its target and cause performance deterioration.There-fore,a data-driven fault diagnosis strategy for wind turbine under model mismatch is proposed to achieve excellent fault diagno-sis performance under the conditions of model reality mismatch and modeling error effect.Firstly,the benchmark model of the horizontal axis wind turbine system is established,and the system fault scenarios and fault sensitivity analysis are introduced.Then,the fault estimator of the system is constructed using data-driven diagnosis strategy,and Takagi Sugeno fuzzy(TS FS)di-agnosis strategy and neural network nonlinear automatic regression exogenous input(NARX NN)diagnosis strategy are pro-posed,Finally,the proposed data-driven diagnosis strategy for wind turbine faults is verified by simulation and HIL hardware in the loop experiment,and compared with the existing nonlinear geometry adaptive filter(NLGA-AF),fuzzy recursive identifi-cation(RFS)and sliding mode observer(SMO).The experimental results show that the proposed diagnosis scheme can better deal with the conditions with model reality mismatch and modeling error effects.And maintain excellent robustness and reliability.
万方数据
维普
