Research on predictive control and fault diagnosis of wind turbine based on MLD

Wind turbine is a mechanical-electrical-hydraulic integration complex dynamic systems, which consists of the mechanical, electrical and hydraulic servo mechanism, with nonlinear and strong coupling characteristics; Random wind makes the wind turbine shows some characters such as randomness, hybrid and so on. The accidents of the wind turbine become more frequent both at home and abroad; From bearing wear, gear teeth broke and shaft eccentricity to the wind turbine burning, blade fall off and tower collapse are likely to cause wind turbine damage and even cause social security accidents. For the stochastic and hybrid dynamic characteristics of wind turbine, this paper proposed that using mixed logical dynamic (MLD) modeling method, make wind turbine maintaining stop in the cut-on and cut-out situation, optimization power capture when partial load and mixed logical dynamic system modeling when full load; Based on wind turbine´s hybrid system model of full working condition, using multi-parameter predictive controller (MPT) method, realized the full process of mixed logical dynamic prediction control of wind turbine´s variable pitch and variable speed constant frequency power and torque control. At the same time, this paper realized the fault diagnosis of the pitch sensor fault and pitch actuator faults by using the extended Kalman filter (EKF) based on the mixed logical dynamic modeling and controlling. The experimental results show that mixed logical dynamic modeling and control method of the wind turbine based on MLD has good control performance and the EKF fault diagnosis can quickly and correctly diagnose pitch sensor fault and pitch actuator faults of the wind turbine based on hybrid system model.