基于数据挖掘的智能电网需求预测

Smart Grid Demand Forecasting Based on Data Mining

随着电力需求的日益增加,获得能源效率也变得越来越重要。因此,开发准确的需求预测方法对于通过有效的系统运行确保能源效率至关重要。为此提出一种基于数据挖掘技术的需求预测方法。该方法是k-均值聚类、贝叶斯分类和ARIMA的结合。以前的大多数研究都试图从供应侧管理的角度来解决这一问题,但提出的预测模型适用于消费侧。实验中,对某市的实际荷载剖面进行案例研究。结果表明,提出的方法的最小错误率接近0.3,可以满足实际需求。

With the increasing demand for electricity,obtaining energy efficiency becomes more and more important.Therefore,the development of accurate demand forecasting methods is very important to ensure energy efficiency through effective system operation.A demand forecasting method based on data mining technology is proposed.This method combines k-means clustering,Bayesian classification and ARIMA.Most previous studies have tried to solve this problem from the perspective of supply side management,but the prediction models proposed are suitable for the consumer side.In the experiment,the actual load profile of a city is studied.The results show that the minimum error rate of the proposed method is nearly 0.3,which can meet the actual needs.