基于NSGA-II算法最优组合的风电功率预测的研究

Research on Wind Power Prediction Based on Optimal Combination of NSGA-II Algorithm

风电场的出力是一个受风速波动性和各种气象条件影响的复杂过程,风电功率预测的准确性可以大大提高电力系统调度运行的效率,维持发、输、用电之间功率的平衡。针对于此,对风电场进行功率预测时,建立了表征风电功率波动的平稳性指标,考虑到风电的波动性越小,预测精度就越高,引入了带精英策略的快速非支配排序遗传算法(NSGA-Ⅱ),以此来求取各个风力发电机组的最优组合,使得组合后的风电出力更加平稳,波动更小,得到了一组pareto最优解集。然后对pareto解集中的所有组合的风力发电机组,利用BP神经网络进行功率预测,预测精度最高的解就是最优的组合。通过仿真验证,证明该方法的有效性和合理性。并将所得到的结果与经典的风电功率预测方法—小波预测和粒子群优化的BP神经网络(PSOBP)预测进行对比分析,证明了所提方法的优越性。

The output of wind farm is a complex process affected by the fluctuation of wind speed and various meteorological conditions.The accuracy of wind power prediction can greatly improve the efficiency of power system dispatching and operation,and maintain the balance of power between power generation,transmission and consumption.In view of this,a stability index is established to characterize the fluctuation of wind power when predicting the power of wind farms.Considering that the smaller the fluctuation of wind power,the higher the prediction accuracy,a fast non-dominated sorting genetic algorithm Ⅱ(NSGA-Ⅱ)with elite strategy is introduced to find the optimal combination of each wind turbine,so that the combined wind power output is more stable and the fluctuation is smaller,and a set of pareto optimal solution set is obtained.Then,the BP neural network is used to predict the power of all combined wind turbines in pareto solution set,and the solution with the highest prediction accuracy is the optimal combination.The validity and rationality of the method are proved by simulation,and the simulation results are compared with that of classical wind power prediction methods-wavelet prediction and PSOBP prediction,which proves the superiority of the proposed method.