Research On Fault Diagnosis Method For The Pitch System Of The Wind Turbine

Wind energy is one of the world’s fastest growing energy. It is clean and renewable. However, limited by the geographical environment condition, future large scale wind farms are most likely to be built in the desert or off-shore. Though the wind resource is rich, the environmental condition around wind farms is severe. Wind turbines malfunction easily by the effect from surroundings. The harsh environments have posed a serious threat to the safe operation of the wind turbine, affect the stability of the wind turbine and even lead to system downtime. With the fault diagnosis technology applied to the wind turbine, the fault can be discovered and handled in time, and more serious consequences caused by the failure propagation can be avoided. The pitch system is an important part of the wind turbine. When the wind speed is above the rated one and below the cut out one, the pitch system of the wind turbine adjusts the captured wind power by changing the pitch angle to ensure that the wind turbine power output is a constant. When the pitch actuator fails, the dynamic characteristics of the pitch system are affected, and the output power is unstable. This thesis makes a research on the fault diagnosis method of the pitch system, and the main content is as follows:By simplifying the wind turbine, the wind model, Aerodynamics system model, tower model, pitch system model, drive train model, converter model and generator model are constructed. While the fault model is constructed according to the characteristics of pump wears, hydraulic leakage and high air content.Considering the pitch sensors biased output fault for wind turbines, a fault diagnosis method based on the multi-innovation kalman filter algorithm was proposed. According to the mechanical structure characteristics of wind turbines, the tiny displacement is produced by the force acting on the tower. And then the corresponding relationship between the change of the pitch angle and the tiny displacement was established. More accurate estimated value for the tiny displacement was achieved by using the multi-innovation kalman filter algorithm to reduce the large noise of the output information generated by sensors. The fault could be detected and the value of the pitch sensor bias could be estimated through the change of the tiny displacement.In light of the pitch sensor fixed value fault and scaling error fault of pitch systems in wind turbines, a multi-innovation observer based fault detection method is proposed. The pitch sensor fixed value fault and scaling error fault can be detected through the residual which is generated by comparing the state estimation with the output of an actual system.Aiming at the faults of pitch actuators for wind turbines, in accordance with the characteristics that the pitch system faults can lead to the change of system parameters, the pitch system model is transformed into identification model by converting into a canonical state space model. And the fault diagnosis problem is transformed into a parameter estimation issue. Then a fault diagnosis method based on the observer-based multi-innovation stochastic gradient algorithm is proposed.In this paper, a method based on the multi-innovation kalman filter algorithm, a multi-innovation observer based fault detection method, and a method based on the observer-based multi-innovation stochastic gradient algorithm are proposed for the fault diagnosis of the pitch sensor and the pitch actuator respectively. And the effectiveness of the proposed method is verified through the wind turbine benchmark model.