基于CFA优化最小二乘支持向量机的短期风速预测  

Short-Term Wind Speed Forecasting Based on CFA-LSSVM Model

作　　者：郑丰[1] 方必武 

机构地区：[1]中国南方电网超高压输电公司天生桥局,贵州兴义562400 [2]武汉大学电气工程学院,湖北武汉430072

出　　处：《陕西电力》2015年第6期15-19,共5页Shanxi Electric Power

基　　金：国家科技支撑计划资助(2015BAA01B01)

摘　　要：准确预测风速对风电规模化并网至关重要。为提高短期风速预测精度,提出一种改进的萤火虫算法优化最小二乘支持向量机的风速预测模型,利用自适应惯性权重和混沌搜索机制提高基本萤火虫算法的全局收敛能力。分别采用IPSO-LSSVM和CFA-LSSVM预测模型对历史时序数据进行提前1 h风速预测,通过与实测数据对比进行误差分析,结果表明CFA-LSSVM模型具有更高的预测精度。Accurate wind speed prediction is of key importance for large scale wind power connected to the grid. To improve the short-term wind speed forecast accuracy, an improved firefly algorithm based least squares support vector machine wind speed prediction model is proposed, adaptive inertia weight and chaos search mechanism have been used to improve the global convergence ability of the basic firefly algorithm. IPS0-LSSVM and CFA-LSSVM are used respectively to forecast wind speed an hour ahead through historical time series data. Through analyzing the prediction error compared with the measured data, the resuhs demonstrate that the CFA-LSSVM has a higher prediction accuracy.

关 键 词：短期风速预测 混沌萤火虫算法 最小二乘支持向量机 

分 类 号：TM614[电气工程—电力系统及自动化]