Wind Turbine Gearbox Evaluation And Diagnosis System Based On Grey Cloud Model And BP Neural Network

In recent years,wind energy has become one of the fastest growing renewable energy sources in the world.China attaches great importance to the development of wind power energy.According to the government's plan,China's wind turbine assembly machine capacity will reach 30 GW by 2020.The maintenance and repair work of wind turbines is an important challenge in the development of wind power technology.As a key part of the wind turbine,the wind turbine gearbox has a failure rate of 40% to 50% under severe conditions.If the gearbox status can be accurately and timely evaluated before the fault occurs,and potential faults can be discovered in advance,so that a reasonable maintenance strategy can be arranged in time to prevent the abnormal deterioration of the gearbox.Accurate and fast fault diagnosis can effectively improve the efficiency of gearbox maintenance,reduce downtime,and reduce economic and labor costs.In this paper,the state evaluation and fault diagnosis of wind turbine gearboxes are deeply studied.Aiming at the problem of operational state assessment,this paper proposes an assessment method based on grey cloud model and cloud gravity theory.In this paper,the wind power gearbox evaluation system is established,and the grey cloud model is established for each index.The cloud gravity center theory evaluation is used to further improve the accuracy of the evaluation.The example shows that the evaluation method proposed in this paper can solve the problems of incomplete information and insufficient subjective influence of decision makers in traditional methods.Aiming at the diagnosis of gear faults such as pitting corrosion,gear tooth wear and broken teeth in wind turbine gearbox,this paper proposes a fault diagnosis method based on EEMD wavelet threshold denoising and Cuckoo Search to optimize BP neural network.The pre-processing method combining EEMD decomposition and wavelet threshold denoising can suppress noise interference in the original vibration signal.The Cuckoo Search-optimized BP neural network can accurately and quickly diagnose the pre-processed signals.The simulation results show that the proposed diagnosis method has good accuracy and timeliness.Combined with state evaluation and fault diagnosis methods,this paper developed a wind power gearbox evaluation and diagnosis system.The system has the functions of storing state data,finding historical state information and providing maintenance and repair strategies,which is an effective integration of the state evaluation and fault diagnosis methods.