基于组合模型的短期风速预测研究

摘要

Wind speed is characteristic of large stochastic volatility,which affects the wind power and the stable operation of the power grid connected with it,thus prediction of wind speeds is crucial in the integration of wind power with the grid.This paper uses Grey-Markov chain model and least squares support vector machine model to predict wind speeds,and then compares the accuracies obtained with each prediction model.Based on this study,the dynamic weight combination model and the 0-1 combination model are studied.Furthermore,analysis is made with the actually measured wind speed in a certain wind farm in China as an example.The result shows that the combination prediction model is better than the single prediction method,and has large practical values.%风速具有较大的随机波动性,影响了电网的稳定性,风速预测对于风电并网问题至关重要.本研究采用灰色-马尔可夫链(GM-Markov)与最小二乘支持向量机(LSSVM)预测模型分别对风速进行预测,比较了各单一预测模型的精度;在此基础上研究了动态权重组合模型与0-1法组合预测模型.然后以国内某风电场的实测风速数据为例进行分析,结果表明,单一预测方法时好时坏,稳定性较差,组合预测模型总体效果较好,具有较大的实用价值.