摘要
气象因素是短期负荷预测重要的影响因素。为提高预测精度,研究了一种基于气温累积效应和灰色关联度的支持向量机拓展算法——最小二乘支持向量机(least squares support vector machine,LSSVM)。通过相关性分析得到与日平均负荷相关程度较大的气象因素。在此基础上,结合气温累积效应采用灰色关联方法对历史日进行分析,选取与待预测日关联度较大的历史日作为相似日,并对LSSVM模型进行训练和预测。实际应用表明,使用所提出的预测模型和数据处理方法能够得到更加精确的预测结果。
Meteorological factor is an important factor which affects short-term load forecasting. To improve forecasting accuracy, this paper developed an extension algorithm of the support vector machine based on accumulative effect of temperature and grey relational degree, namely least squares support vector machine(LSSVM). Firstly, meteorological factor highly related to average daily load was obtained through correlation analysis. Then, on the basis of accumulative effect of temperature, grey correlation method was adopted to analyze historical days, the historical day having higher correlation with the day to be forecasted was taken as similar day, and the LSSVM model was trained and forecasted. Actual application showed that the proposed forecasting model and data processing method could obtain a highly accurate result.
引文
[1] 康重庆,周安石,王鹏,等.短期负荷预测中实时气象因素的影响分析及其处理策略[J].电网技术,2006,30(7):5-10.
[2] 刘文颖,门德月,梁纪峰,等.基于灰色关联度与LSSVM组合的月度负荷预测[J].电网技术,2012,36(8):228-232.
[3] 龚文龙.基于最小二乘支持向量机的短期负荷预测[D].湖南:湖南大学,2014.
[4] 黎灿兵,杨鹏,刘玮,等.短期负荷预测中考虑夏季气温累积效应的方法[J].电力系统自动化,2009,33(9):96-99.
[5] 刘嘉龙,李小燕,刘思捷,等.考虑气温累积效应的短期负荷预测[J].华北电力大学学报,2013,40(1):49-54.
[6] 李笋,王超,张桂林,等.基于支持向量回归的短期负荷预测[J].山东大学学报(工学报),2017,47(6):8-12.