考虑温度累积效应下基于LS-SVMR电力负荷预测研究  

Study on Power Load Forecasting Based on LS-SVMR Considering Cumulative Effect of Temperature

作　　者：缪智伟 方睿 MIAO Zhiwei;FANG Rui(Department of Mathematics,Shantou University,Shantou 515063,Guangdong,China)

机构地区：[1]汕头大学数学系,广东汕头515063

出　　处：《汕头大学学报（自然科学版）》2023年第3期42-51,共10页Journal of Shantou University：Natural Science Edition

摘　　要：基于广东省某地区2018—2022年每日最大负荷数据及同期该地区日气象要素资料,发现最高气温对最大负荷的影响具有累积效应,影响温度类型效益的因素主要是预测日最大气温以及持续高温的天数;文章建立了气温累积效应的日最高气温修正公式,并利用实例验证了最高气温累积效应对最大电力负荷的影响.面对96个时点负荷数据复杂时序性和非线性的特性,构建了一种基于最小二乘支持向量机(LS-SVMR)网络电负荷最大值的预测模型,该方法考虑了对负荷有影响的节假日与工作日、天气、温度等相关因素,将修正后的日最高气温及最大电力负荷作为输入层,应用基于遗传算法优化后的最小二乘支持向量机对最大电力负荷进行预测.模型预测结果表明:本文的模型预测精度比传统BP、RBF神经网络负荷预测方法,具有更高的预测精度,预测结果能更好地为电力调度及安全运行提供参考依据.Based on the daily maximum load data of a region in Guangdong Province from 2018 to 2022 and the daily meteorological elements data in the same period,it is found that the maximum temperature has a cumulative effect on the maximum load.It is found that the factors affecting the benefit of temperature type are mainly the predicted daily maximum temperature and the number of days of continuous high temperature.The daily maximum temperature correction formula for the cumulative effect of temperature is established,and the influence of the cumulative effect of maximum temperature on the maximum power load is verified by an example.Facing the complex scheduling 96 point load data and the characteristics of nonlinear,a kind of network maximum electrical load forecasting model based on least squares support vector machines(LS-SVMR)is built.The impact load of holidays,working days,weather,temperature and other factors is considered.Daily maximum temperature and maximum power load are revised as input layer.The least squares support vector machine(LS-SVMR)optimized based on genetic algorithm is used to predict the maximum power load.The model prediction results show that the prediction accuracy of the proposed model is higher than that of the traditional BP and RBF neural network load prediction methods,and the prediction results can provide a reference for power scheduling and safe operation.

关 键 词：短期负荷预测 遗传算法 最小二乘支持向量机 电负荷温度累积效应 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TM715[自动化与计算机技术—控制科学与工程]