基于云模型的短期电价预测  

Short-Term Electricity Price Forecasting Based on Cloud Model

作　　者：栗然[1] 崔天宝[1] 肖进永[2] 

机构地区：[1]华北电力大学电气与电子工程学院,河北省保定市071003 [2]华能上安电厂,河北省石家庄市050310

出　　处：《电网技术》2009年第17期185-190,共6页Power System Technology

摘　　要：现有的电价预测方法有时间序列、神经网络、小波变换等,都是对点进行预测。该文提出一种基于云模型的短期电价预测新方法。首先,介绍了云模型的概念和特点,给出基于云模型的电价和负荷数据的离散化和概念跃升过程,得到了电价和负荷的概念模型。通过极大判定法对数据集进行软划分,建立电价与负荷的布尔型数据库,然后根据给定的支持度和置信度软域值,采用基于云的关联知识挖掘算法,得到时间、负荷和电价之间的关联规则。最后,以时间、负荷的合取作为规则前件,以电价作为规则后件,建立规则发生器,根据挖掘出的规则进行预测。该文所提方法得到的预测结果是一系列不确定的离散点的集合,集合中的每一个点都可作为预测结果提供给用户,用户可以根据经验和其它信息来适当选择结果,也可以将所有点的期望值作为结果提供给用户。At present, there are various electricity price forecasting methods such as time-series, neural network, wavelet transform and so on, however these methods can merely give out the forecasted results at required point of time. In this paper, at first the concept and features of cloud model are presented, the process of the discretization of electricity price and load data and the concept zooming based on cloud model are given then a conceptual models of price and load are achieved. By means of great determination law the elastic classification of the data set are conducted and the Boolean database of electricity price and load is established; then according to given support degree and confidence level of soft domain values and by use of cloud based mining algorithm for associative knowledge, the association rules among time, load and electricity price are obtained. Finally, taking the conjunction of time and load as the premise of the rules and the price as the consequent of the rules, a rule generator is built and forecasting is proceeded by the mined rules. The forecasting results of the proposed method is a set of a series of uncertain discrete points and every point of the set can be offered to the user as forecasting result, and the user can select these results properly according to their experience and other information, and it is also possible to offer the expectation values of all points to the user as the results.

关 键 词：电价预测 云模型 数据离散化 概念跃升 不确定性推理 关联知识挖掘 前件 后件 

分 类 号：TM734[电气工程—电力系统及自动化]