| With the rapid development of global enconomy, energy crisis and environmentpollution have been the demanding prompt solution problem in the global; the renewableenergy sources development and utilization in recent years, especially the wind powergeneration technology have been received the high value and obtained the flying start in theevery countries of the world. However，the low utilization coefficient for wind energy of windturbine,the low wind tuebines output and the discrepancy quality of electric energy for windpower have been the ubiquity problem in operation process of every wind farms, and bringthe adverse effect for the faster development of wind power generation technology. So how toachieve the maximizeation wind turbines energy output for the specific wind energy resourcesin the fixed wind farms are the important issue in the research of wind power generationtechnology, and have the important significance for wind power industry of out country evenworldwide.In this paper, the first part revealed the basic theories of wind power generationtechnology, the wind turbines working principle and operating characteristic, support vectormachine algorithm, differential evolution algorithm, proposed the better performance leastsquare-support vector machine based on SVM algorithm and DE algorithm improvedstrategies aimed at DE algorithm shortcoming and defect, prepared for the postamble modelbuilding and problem solving.Then, due to the complicated relationship among the wind turbines, wind speed andblade pitch angle, it is difficult to build a general accurate mathematical model. This paperproposes a new method to optimize wind turbines output, firstly, using support vectormachine to fit the nonlinear relationship model between wind turbine output and wind turbineoperation parameters, and then an efficient differential evolution algorithm is designed todynamically optimize the blade pitch angle based on this model and the change of wind speedso that the wind turbine output can be maximized. To verify the feasibility and superiority ofthe proposed method, simulations have be made based on the historical operation dataobtained from Chang Ling wind farm of Poyang lake, the results demonstrate that it cansignificantly increase wind turbines power output by optimizing wind turbines blade pitchangle. The proposed method can establish an accurate dynamic relationship between windspeed and optimal blade pitch angle; it can provide scientific guidance for wind turbineoperation.Finally, aiming to the problem that wind energy utilization rate decreases due to the windturbines given tip speed ratio in factory can’t guarantee to capture the maximal wind energy;Based on the fact that wind speed follows a Weibull two-parameter distribution, This paper uses1.5MW VSCF wind turbine as the research object, The optimization goal is to maximizethe annual energy output as much as possible, to solve this model a modified DE algorithm isproposed. In comparison with the original design parameters of wind turbine, the optimizationresult shows obvious advantages, which verifies the feasibility and practicality of theproposed method. Meanwhile, this paper discusses the relationship between annual energyoutput and tip speed ratio when changing the shape parameter and scale parameter of Weibulldistribution, respectively. It can be concluded from the simulation results that the wind energyoutput is more dependent on the Weibull scale parameter than the Weibull shape parameter.This implies that the wind energy output is more dependent on the mean wind speed than thespeed distribution. |
