基于气象信息因素修正的灰色短期负荷预测模型  

Short-Term Load Forecasting by Grey Model With Weather Factor-Based Correction

作　　者：焦润海[1] 苏辰隽[1] 林碧英[1] 莫瑞芳[1] 

机构地区：[1]华北电力大学控制与计算机工程学院,北京市昌平区102206

出　　处：《电网技术》2013年第3期720-725,共6页Power System Technology

基　　金：中央高校基本科研业务费专项资金资助(11MG13);河北省自然科学基金项目(F2011502038)~~

摘　　要：针对传统灰色模型应用于短期负荷预测的缺陷,改进了多角度数据组织策略,选择社会活动背景相似度较高的历史日作为相似日,建立了气象因素突变判别准则,以识别相似日中的气象突变,并采用回归法消除相似日负荷数据中的受气象敏感负荷分量。采用邻近趋势外推法和相似日替换法矫正灰色模型预测时容易产生的局部畸变点。实验结果表明,改进后的灰色模型预测精度提升明显,预测精度平均提高1%~2%,最高可达7%。该方法能同时考虑社会活动规律和气象因素对负荷的影响,克服了传统模型本身对不规律数据预测精度差的缺点。To remedy the defects in applying grey model to short-term load forecasting,the multi-angle data organization strategy is improved.Choosing historical days with higher similarity in social activity background as the similar days,a criterion of abrupt change of weather factors is established to identify the abrupt change of weather factors in similar days,and the weather-sensitive load components in load data of similar days are eliminated by regression method.The extrapolation of adjacent tendency and similar day substitution method are adopted to correct local distortion points,which are easy to appear during the forecasting by grey model.Experimental results show that using the improved grey model the load forecasting accuracy can be evidently improved by one or two percentage points in average and the top improvement of load forecasting accuracy reaches up to seven percentage points.

关 键 词：短期负荷预测 灰色模型 气象因素修正 邻近趋势外推 相似日替换 

分 类 号：TM715[电气工程—电力系统及自动化]