Modeling Of Wind Turbine Power Characteristic Curves And Wind Energy Assessment Considering Wind Speed And Direction

With the development of domestic and international economy,the demand for energy is increasing dramatically.Among them,wind energy occupies an important position as a renewable and clean energy source.However,compared with traditional fossil energy sources,wind energy has more uncertainties.Accordingly,an accurate wind energy assessment method is imperative for large-scale wind grid connectivity and wind farm site selection studies.The actual power characteristics curves of the wind turbines are constructed by removing outliers from the measured wind field data to provide power predictions for wind grid connectivity and wind energy loss reduction,and setting up joint wind speed-wind direction probability density distribution for wind energy assessment.The main contents of this paper are as follows:(1)Based on the measured wind speed-power data of wind farms,a segmentation modeling method applying Bessel curves is proposed.First,the characteristics of the Bessel curve and the modeling strategy are highlighted,and then compared with the commonly used higher-order polynomial model,and the proposed method is proved to be more reasonable using the evaluation index of the model.(2)The modeling was carried out based on the Logistic function with the wind speed-power measured data,and at the same time,the overfitting situation due to the number of model parameters was solved by using the Akaike information criterion(AIC)and Bayesian information criterion(BIC)as the evaluation index.First,the Logistic function modeling process and parameter evaluation methods are described in detail.In addition,for the problem that the least squares method is more sensitive to the initial value,a combination of genetic algorithm and least squares method is proposed to solve the dependence on the initial value in the process of solving the parameters by least squares method.Finally,the actual power at different wind speeds is predicted according to the established model and the accurate parameter evaluation method.It is found that the prediction accuracy of the model is high,which shows the validity and reasonableness of the proposed method.(3)A parameter evaluation method of the Weibull wind speed distribution model combining simulated annealing algorithm(an intelligent optimization algorithm)with the traditional least squares method is proposed,which is based on the global optimization search capability of the simulated annealing algorithm to find the optimal initial value and endowed with the least squares method for the parameter evaluation of the Weibull distribution model.Firstly,the wind speed data after eliminating the abnormal data are used to model the wind speed with a single Weibull distribution(two-parameter Weibull distribution and three-parameter Weibull distribution),then the model parameters are evaluated using this method,and finally the wind energy characteristic indexes are evaluated based on the wind speed distribution model calculation and analysis.In order to further improve the accuracy of the wind speed distribution model,a mixed Weibull distribution model was introduced to model the wind speed,and the same method was used to evaluate the parameters of the mixed distribution,and finally the optimal wind speed distribution model was selected according to the root mean square error and correlation coefficient,and the annual energy production was calculated and analyzed with the wind speed-power characteristic curve model,and it was found that the annual energy production value was close to the actual annual energy production value,it is demonstrated that the proposed method is reasonable and accurate in evaluating the parameters of the Weibull distribution wind speed model.(4)A joint wind speed-wind direction distribution model with mixed Weibull wind speed distribution and higher-order von Mises wind direction distribution is established based on the Angle-Linear(AL)distribution model.In order to make the wind distribution model more accurate,the mixed von Mises distribution model is first introduced to model the wind direction,and then the probability density function of the optimal mixed von Mises wind distribution model is selected from the model evaluation index,and the joint wind speed-wind direction probability density function is formed by associating with the optimal wind speed distribution probability density function.Finally,the equivalent wind speed and active wind power density are introduced to evaluate the wind energy again.