摘要
考虑到现代社会中气象因素影响着电力负荷的使用情况,提出了考虑气象因素的基于PCA-LM-BP的短期电力负荷预测方法。由于气象因素数据量较大,采用PCA方法对天气因素进行主元分析,选取出对负荷值影响较大的因素引入到负荷预测模型当中。由于传统的BP算法具有收敛速度慢,易陷入局部最优的缺点,采用LM算法对其进行改进,提升其预测精度。将PCA提取的主要天气因素及历史负荷数据作为LM-BP算法的输入,预测的负荷值为输出。通过算例仿真分析,分别对比BP算法,GA-BP算法,LM-BP算法的负荷预测值及误差值,可以发现LMBP预测的负荷值与实际值更接近,通过误差分析验证了文中所提方法的有效性。
Considering the use of power load affected by meteorological factors in modern society,a shortterm electric power load forecast method based on PCA-LM-BP was proposed. Due to large data volume of meteorological factors,PCA method was adopted to conduct principal component analysis of weather factors. The factors which have a great influence on load value were selected and introduced into the load forecasting model. As the traditional BP algorithm has the shortcoming of slow convergence speed and easy to fall into local optimization,LM algorithm was adopted to improve it and increased the level of its prediction accuracy. The main weather factors and historical load data extracted by PCA were used as input of LM-BP algorithm,and the predicted load value was the output. Through simulation analysis of calculation examples,the load prediction value and error value of BP algorithm,GA-BP algorithm and LMBP algorithm were compared respectively. It can be found that the load value forcasted by LM-BP is closer to the actual value,and the validity of the method proposed in this paper is verified by error analysis.
作者
张梅
李金湖
张莉娜
杨铮宇
ZHANG Mei;LI Jin-hu;ZHANG Li-na;YANG Zheng-yu(Yunnan Power Grid Co Information Center,Kunming 650217,China;State Grid Info-Telecom Greate Power Science and Technology Co.,Ltd.,Fuzhou 350003,China)
出处
《信息技术》
2019年第6期101-105,共5页
Information Technology
