Wind Turbine Gearbox Fault Detection Based On Model

With the increasing electricity generation capacity of wind turbine, gearbox sizes dramatically increase. Wind farms are being explored from onshore to offshore, which drives the operational costs to rise significantly. Therefore, developing fault detection algorithm for wind turbine gearbox is becoming important. Due to the variable-speed nature of wind turbines operation characteristic, there is challenge on using classical method to develop fault detection algorithm. Typical wind turbine gearbox condition monitoring is based on vibration signals, which is effective to detect failures with high frequency signal range. But it may not be effective on low speed components which have low frequency signal characteristic of different failure modes. Therefore, this paper presents research results of model based wind turbine gearbox fault detection, analyses the fault physical mechanism from the perspective of the system, explores the fault characteristics of different gearbox failure modes, and demonstrates the relationships between the degraded efficiency of wind turbine gearbox, oil temperature and control strategies. On this basis, further use of SCADA system which captures the low-frequency signals for fault feature extraction, demonstration and the development of new detection algorithm.The first part of the paper will be to introduce the various types of wind turbine components, systems and sub-systems and SCADA data system.Secondly, the simulation model is developed based on MATLAB/Simulink software. Typical wind turbine principles which include four parts:1) stochastic wind characteristics, 2) rotor aerodynamics,3) drive train transmission efficiency,4) generator simplification and 5) control strategy.Then, thermodynamic process of gearbox lubrication system is studied and the energy loss model which relates the transmission efficiency to oil temperature variation is coupled to the system model; by introducing fault to the drive train the failure characteristics are obtained by comparing the signal variation pattern between healthy and fault circumstances; besides of that, do oil leakage failure characteristics prediction.Finally, build two control schemes models below the rated wind speed-speed control and torque control. Analyze the effects of different control strategies based on the simulation model with the same efficiency attenuation for wind turbines. Through the above research, this paper summarizes the characteristics of the wind turbine gearbox failures in different failure modes and provides a theoretical foundation for future smart wind turbine condition monitoring system development.