基于电量分析的配网用户用电需求量预测研究

Study of Distribution Network Users Power Consumption Demand Forecasting based on Electricity Analysis

准确预测用户电量需求对于市场竞争环境下的电网公司、工商业、居民用户来说具有重要意义。简单综述了国内外电量预测理论，包括灰色理论、人工神经网络理论等，阐述了GM（1，1）模型和BP模型预测电量需求的原理。详细介绍了电力市场电量需求预测的几种实际经常使用的方法，包括经济模型法、综合分析法、分析预测法及其他方法等。最后，以电力弹性系数法为例，基于广州市2000年~2008年的市用电量历史数据，对其2009年用电量需求进行了预测分析。同时，简要给出了提高电量预测准确率的一些措施，建议将近年来发展的机器学习算法等运用到电量需求预测中，对于“电网-用户-售电商-负荷集成商”等多主体的用电供需友好互动将具有重要的指导和参考意义。

It’s of great significance to accurately predict users’ power demand for the grid corporations, the industryand commerce enterprises and the inhabitant users in market competition environment. A simple review was made on home andabroad power forecasting theory, including grey theory, artificial neural network theory, etc. expounded the power forecastingprinciple of GM(1, 1) model and BP model. Several actual and regular used methods of electricity market power demandprediction were introduced, including the economic model method, the comprehensive analytical method, the predication parsingmethod, and other methods. Finally, set the electricity elasticity coefficient method as an example and based on the historical cityelectricity consumption data of Guangzhou from year of 2000 to 2008, its power demand in 2009 was forecasted. Meanwhile,some measures to improve power prediction accuracy were given, and suggested that the newly developed machine learningalgorithms apply in power demand prediction, which will have provide certain guidance and reference for multi-agents supplyand demand interaction, such as the grids, users, electricity sellers, and load integrators.