大规模风电并网电力系统经济调度中风电场出力的短期预测模型  

Short-term Wind Power Output Forecasting Model for Economic Dispatch of Power System Incorporating Large-scale Wind Farm

作　　者：袁铁江[1] 晁勤[1] 李义岩[1] 吐尔逊.伊不拉音 

机构地区：[1]新疆大学电气工程学院,新疆维吾尔自治区乌鲁木齐市830008

出　　处：《中国电机工程学报》2010年第13期23-27,共5页Proceedings of the CSEE

基　　金：国家自然科学基金项目(50667002);新疆青年自然科学基金(2009211B04);新疆高校科研计划重点项目(XJEDU2009I04)~~

摘　　要：考虑到大规模风电并网电力系统经济调度中,对风电场出力的短期预测在时间尺度和精度尺度方面的要求,以传统的反传播神经网络(back propagation artificial neural network,BP-ANN)作为预测手段的基础,建立了风电场短期出力预预测模型。考虑到历史的预测误差与未来预测误差间的映射关系,利用传统的BP-ANN预测技术对未来的预测误差进行预测。通过算例仿真发现,误差预测变化趋势能跟踪预预测的误差变化,基于此并考虑到经济调度对风电场出力预测精度的要求,建立了对风电场出力短期预预测进行修正的风电场出力短期预测模型,进一步的算例仿真表明了该模型的有效性。This paper illustrates a wind power forecasting model based on back propagation artificial neural network （BP-ANN）, considering the requirement of power output pre-forecasting of wind farm in economic dispatch both in time and accuarcy. Considering the relation of historical prediction error and future prediction error, an error forecasting model is developed based on artificial neural network （ANN）. Simulation results show that the prediction error can follow the variation tendency of the error of pre-forecasting model, so considering the requirement of economical dispatch to output power prediction accuracy, a wind farm power prediction model is constructed to modify the short-term wind output power prediction. The further experimental results show that the prediction-modification model is effective.

关 键 词：风电 反传播神经网络 误差 预测 短期 

分 类 号：TM73[电气工程—电力系统及自动化]