基于数据挖掘的园区综合能源消耗实时监测方法

Real time monitoring method of comprehensive energy consumption in park based on data mining

对园区综合能源消耗进行实时监测对于保证电力正常供应以及节能减排具有重要意义。为此,针对当前三种基于统计学、深度学习和支持向量机的异常能耗监测方法存在的灵敏度不足的问题,提出一种基于数据挖掘的园区综合能源消耗实时监测方法。该方法首先利用中间件进行园区综合能源消耗数据采集,然后对其进行预处理,包括属性选择、数据离散化、缺失填补以及数据约简,接着将处理好的数据作为样本,利用数据挖掘中的决策树算法构建分类模型,实现园区综合能源消耗异常监测。结果表明:与当前三种传统异常能耗监测方法相比,利用基于数据挖掘的方法进行园区综合能源消耗实时监测,灵敏度更高,说明本方法能在更短的时间内监测出园区综合能源消耗异常。

Real-time monitoring of the comprehensive energy consumption of the park is of great significance to ensure the normal supply of electricity and energy conservation and emission reduction.Therefore,in view of the insufficient sensitivity of the current three abnormal energy consumption monitoring methods based on statistics,deep learning and support vector machines,a real-time monitoring method of comprehensive energy consumption in the park based on data mining is proposed.In this method,middleware is firstly used and then preprocess to collect comprehensive energy consumption data of the park,including attribute selection,data discretization,missing filling and data reduction,and then the processed data is used as samples.Using the data mining,the decision tree algorithm constructs is used to built a classification model to realize abnormal monitoring of the park’s comprehensive energy consumption.The results show that,compared with the current three traditional abnormal energy consumption monitoring methods,the data mining-based method for real-time monitoring of comprehensive energy consumption of the park has higher sensitivity,indicating that this method can monitor the comprehensive energy consumption of the park in a shorter time abnormal.