基于PSO和曲线重迭算法的扩展短期负荷预测  

Extended Short-term Load Forecasting Based on PSO and Curves Overlap Algorithm

作　　者：张玲玲[1] 杨明玉[1] 张华彬[1] 

机构地区：[1]华北电力大学电气与电子工程学院,河北保定071003

出　　处：《电力科学与工程》2014年第7期32-36,共5页Electric Power Science and Engineering

摘　　要：精确的电力系统扩展短期负荷预测,有利于改善短期负荷预测效果和制定科学合理的滚动发电计划。根据历史负荷数据的内在规律性,提出了一种基于粒子群优化改进曲线重迭算法的扩展短期负荷预测方法。该方法由若干个同日类型日形成相关负荷集,并结合粒子群优化算法(PSO)的全局寻优能力对传统曲线重迭法中的参数进行了优化,有效克服了传统曲线重迭算法中依据经验选定参数的盲目性。研究结果表明,该预测方法较传统曲线重迭预测法有更高的预测精度。Accurate extended short-term load forecasting is both helpful to improve the short-term load forecasting results and useful to plan rolling generating schedule scientifically and properly. Based on the inherent regularity of the historical load data, this paper proposes a modified curves overlap algorithm based on particle swarm optimization （PSO）. Forming a related load data set consisting of several data of the same data type, this method optimizes its parameters by the global optimizing ability of PSO, and effectively over comes the blindness of the parameters selecting in traditional algorithm effectively. The results show that the proposed method has higher prediction accuracy than the traditional one.

关 键 词：扩展短期负荷预测 粒子群优化 改进曲线重迭算法 

分 类 号：TM714[电气工程—电力系统及自动化]