基于支持向量机理论的短期负荷预测  

Short-term Load Forecasting Based on Support Vector MachineTheory

作　　者：邹超 卢霁明 Zou Chao;Lu Jiming(Kunming Engineering Corporation Limited,Kunming 650224,China;Yunnan Energy Investment Power Design Co.,Ltd.Kunming 650000,China)

机构地区：[1]中国电建集团昆明勘测设计研究院,云南昆明650000 [2]云南能投电力设计有限公司,云南昆明650000

出　　处：《云南电力技术》2022年第1期36-39,共4页Yunnan Electric Power

摘　　要：电力系统负荷预测的关键问题在于根据预测对象的历史数据建立相应的数学模型来描述其发展规律。支持向量机理论(SVM)能较好地解决小样本、非线性、高维数和局部极小点等实际问题,并且能够用来建立较为完备的负荷预测模型。研究表明,应用SVM进行电力系统负荷预测,具有精度高、速度快等优点,显著提高了负荷预测的效果。SVM的训练相当于解决一个线性约束的二次规划问题,这有利于我们对训练过程的理解,并增强了训练的可控性。本文结合实例论述了SVM在短期负荷预测中应用的分析和实现过程。The core of power system load forecasting is to establish a corresponding mathematical model based on the historical data of the forecast object to describe its development law.The support vector machine method(SVM)can better solve practical problems such as small samples,nonlinearity,high dimensionality and local minima,and can be used to establish a relatively complete load forecasting model.The research results of many scholars show that the application of SVM for power system load forecasting has the advantages of high accuracy and fast speed,which significantly improves the effect of load forecasting.Since SVM training is equivalent to solving a linearly constrained quadratic programming problem,it is conducive to our understanding of the training process and enhances the controllability of training.This article describes the specific analysis and implementation process of SVM application in short-term load forecasting with examples.

关 键 词：负荷预测 支持向量机理论 非线性 线性约束 

分 类 号：TM74[电气工程—电力系统及自动化]