Research On Fault Diagnosis And Fault-tolerant Control Of The Pitch System In Wind Turbines

Wind energy is one of the world’s fastest growing and most competitive renewable energy. With the increasing demand for wind energy, the capacity of single wind turbine is increasing. The harsh environments of large wind turbines, such as deserts, grasslands and the sea, have posed a serious threat to the safe operation of units and parts. Due the remote environment, it is very expensive to perform a regular maintenance, most wind turbines have to operate with faults. With the fault diagnosis and fault tolerant control method introduced, the fault can be discovered and stopped in time, and the wind turbine is able to maintain stable output power before a regular maintenance. The pitch system is an important part of wind turbines, the captured wind power can be changed by adjusting the pitch angle. When the pitch actuator fails, the dynamic characteristics of the pitch system is affected, and while the wind velocity is above the rated one, the output power is unstable. Against the background, this thesis researches the fault diagnosis and fault tolerant control method of the pitch system, and the specific content is as follows:Under the principle of simplified modeling, the pitch system and the related parts are modeled carefully. The wind model, tower model, pitch model, drive train model and generator model are constructed. According to the characteristics of pump wears, hydraulic leakage and high air content, fault models of faults are constructed. The state parameters are chosen appropriately to establish the state space model of wind turbines.By applying the impulse invariance transformation, the considered continuous time model is transformed to the difference equation and system identification model. According to the characteristics of the pitch system fault, the fault diagnosis problem is transformed into a parameter estimation issue. The multi-innovation forgetting gradient identification algorithm converges fast and takes less computation. The validity of fault diagnosis method for pitch system is verified by simulation examples.In the light of the nonlinear properties and uncertain parameters of wind turbines, an active fault tolerant control technology is proposed to solve the dynamic stability problems, which is based on the linear-parameter-varying(LPV) gain scheduling technology. The linearized model around the operating point of wind turbines and faults is converted into a convex polyhedron structure LPV model by using LPV convex decomposition method. Then, each vertex of convex polyhedron is designed to meet the desired performances by using Linear Matrix Inequalities(LMIs). All the vertex gains are synthesized to build LPV controller with convex polyhedron structure. The simulation result shows that the LPV active fault tolerant controller can be applied to the wind turbines successfully.