面向工业能耗分析的大数据技术及其应用

Big Data Technology for Industrial Energy Consumption Analysis and Its Application

为了更加准确地对工业耗能进行分析,为节能策略提供参考,此次研究提出采用线性回归算法与支持向量机算法结合,对能耗数据进行处理,然后通过K-Means算法应用至相关的能耗分析平台的核心功能中。特此设计一项实验验证性能,实验表明,当K=3时,工作时长KL值最大,相关度较高,工人工龄KL大小在其次,相关度较高。当K=5或K=7时,工作区域节点地址维度在各个分类群组中的KL相差很小,工作时长和工人工龄两个维度在其中KL值较大,相关度较高。进一步对两个维度与工业能耗的关系进行分析,得出了相应的影响原因和因素,可见此次研究的算法对能耗的分析提升,在对节能策略的提出具有参考意义。

In order to analyse industrial energy consumption more accurately and provide reference for energy saving strategies,this study proposes to use linear regression algorithm combined with support vector machine algorithm to process the energy consumption data,and then apply it to the core functions of the relevant energy consumption analysis platform through K-Means algorithm.An experiment is hereby designed to verify the performance,which shows that when K=3,the value of working hours KL is the largest,with high correlation,and the size of workers'working age KL is in the second place,with high correlation.When K=5 or 7,the work area node address dimension has a small diference in KL across taxonomic groups,and the two dimensions of working hours and workers'working age have larger KL values and higher correlations among them.The relationship between the two dimensions and industrial energy consumption is further analysed,and the corresponding causes and factors of influence are derived,which shows that the algorithms of this research have improved the analysis of energy consumption,and have reference significance in the proposal of energy saving strategies.