基于IDE-LSSVM的风电场短期风速预测  

Short-term Prediction of Wind Speed for Wind Farm Based on IDE-LSSVM Model

作　　者：张妍[1] 王东风[1] 韩璞[1] 

机构地区：[1]华北电力大学自动化系,河北保定071003

出　　处：《系统仿真学报》2017年第7期1561-1571,共11页Journal of System Simulation

基　　金：中央高校基本科研业务费专项资金(2014MS139)

摘　　要：为进一步提高风电场短期风速预测精度,提出一种改进的差分进化算法优化的最小二乘支持向量机短期风速预测模型。在改进的差分进化算法中综合了两种变异操作算子,改进了变异因子和交叉概率因子,使其根据进化代数自适应变化,保证了进化初期算法的全局搜索能力和种群多样性,提高了进化算法末期局部搜索精度和收敛速度。把改进的差分进化算法用于最小二乘支持向量机的参数寻优,提高了模型的预测精度,并在河北某风电场的真实历史数据集上建立风速预测模型,仿真实验验证了方法的有效性。To improve the prediction accuracy of short-term wind speed for wind farm, an improved differential evolution algorithm was applied to optimize the parameters of least squares support vector machine. Two mutation operators were integrated, and the scale factor and crossover probability factor were changed gradually to adapt to the evolutionary generations. The good global search ability and population diversity in early stage of evolution were ensured, therefore the local search accuracy and the convergence speed in the late stage were enhanced. The forecasting performance of the least squares support vector machine optimized by IDE was improved. Simulation experiments on the historical wind speed data sets in a wind farm of Hebei province show that the proposed model is effective.

关 键 词：风电场 短期风速预测 最小二乘支持向量机 改进差分进化算法 

分 类 号：TM614[电气工程—电力系统及自动化]