基于数据挖掘技术的地区电网负荷特性分析与预测

Load Characteristics Analysis and Prediction of a Regional Power Grid Based on Data Mining

针对不同地区的电网用电负荷特性,基于一种数据挖掘技术建立了地区电网负荷特性分析与预测模型。首先用模糊聚类分析对负荷影响因素约简,消除负荷影响因素之间的强相关性,并可将众多的影响因素按其特点进行归类;其次依托灰靶理论原理,以最大负荷状态模式构造标准模式序列并建立灰靶,对影响因素类进行灰靶变换,得出影响因素指标相对标准模式的靶心度;最后将靶心度转化为灰靶贡献度,将其平均值作为各类影响因素对最大负荷影响的权重系数,形成最大负荷预测模型。结合地区电网历年最大负荷和影响因素实际数据,验证了此组合模型有较好的实用性和较高的精确性。

According to the load characteristics of power grid in different regions,the analysis and prediction model of regional power grid load characteristics is established based on a data mining technology.Firstly,fuzzy cluster analysis is used to reduce load influencing factors,eliminate the strong correlation between load influencing factors,and clasvsify many influencing factors according to their characteristics.Secondly,based on the principle of grey target theory,the standard model sequence is constructed with the maximum load state model and the grey target is established.The grey target transformation is carried out on the influencing factors,and the approaching degree of the influencing factors relative to the standard model is obtained.Finally,the approaching degree is converted into the gray target contribution degree,and its mean value is taken as the weight coefficient of various influencing factors on the maximum load,forming the maximum load prediction model.Combined with the actual data of the maximum load and influencing factors of the regional power grid over the years,the practicability and accuracy of the combined model are verified.