Maximum Wind Energy Capturing For AC-Excited Wind Power Generation System

As the world economy and industrialization rapidly develops, the world's growing demands for coal, oil and other conventional energy prompted a growing number of researches on some of the new environment-friendly renewable energy. Wind power as a new alternative energy with little environmental pollution, inexhaustible, and rich reserves has drawn increasing attention. In the wind power generation system, wind energy is captured by wind turbine. Its size depends on wind speed and angular velocity, and there is an optimal velocity value under any wind speed to capture maximum wind energy. How to quickly find the best wind turbine's velocity under the random wind speeds is one of the key issues about wind power generation efficiency, which makes a turbine gain the maximum conversion effect from wind energy.The characteristics of the wind energy, wind turbine and wind power generation system are analysed, and some existing algorithms are studied including parameter identification of wind turbine, power feedback, TSR, HCS, three points comparing method and PSO optimization methods. Because wind power is random and instable, the disadvantage of HCS and TSR algorithm for speed and stability are analysed by the comparison of different algorithm, an adaptive variable step search algorithm is proposed to capture the maximum wind power. The adaptive and variable step size strategy of this algorithm is presented in this thesis. The search speed is significantly accelerated by improving step size strategy of HCS, the search range is narrowed through the introduction of TSR of the initial estimation, and the initial angular velocity is quickly located near maximum power point by initially estimating TSR. Through the strategy of variable step size, it can reach the maximum power point more quickly and TSR is updated with the new optimal one. In this way, the initial angular velocity is closer to maximum value and the search range is much more reduced after the wind speeds change. Real-time wind speed detection is not required for this algorithm, and not dependent on the wind turbine model. The simulation results show that the algorithm has good convergence, the wind turbines can reach the maximum power point more quickly, and dynamic response of the system becomes much faster.