摘要
用电分析和预测对于工业企业的能源管理具有重要意义。现有工业用电分析与预测方法大多局限于离线本地计算,数据导入和导出等步骤仍需要人工操作,存在效率低和实时性差等问题。针对此问题,本文基于某工业企业用电数据采集系统获取的工业用电数据,在阿里云平台上实现了实时数据导入,并对影响工业用电的主要因素进行了特征分析,进而采用梯度提升决策树回归算法构建了该工业企业的工业用电预测模型,并与支持向量回归机和线性回归算法进行对比,证明了梯度提升决策树回归算法具有更好的预测效果。本文还对云平台的数据同步时间和算法计算时间进行了统计分析,证明了基于云平台的工业用电分析与预测具有更好的效率和实时性。
Power consumption analysis and forecasting are of great significance to the energy management of industrial enterprises.Existing industrial power analysis and prediction methods are mostly limited to off-line local calculations.Data import and export steps still require manual operations,and there are problems such as low efficiency and poor real-time performance.In response to this problem,based on the industrial electricity data acquired by an industrial enterprise electricity data collection system,this paper implements real-time data import on the alibaba cloud platform,and analyzes the main factors affecting the industrial electricity consumption,and then uses the gradient to enhance The decision tree regression algorithm constructed the industrial electricity prediction model of the industrial enterprise,and compared with the support vector regression machine and linear regression algorithm,which proved that the gradient lifting decision tree regression algorithm has better prediction effect.On the other hand,this paper also makes a statistical analysis of the data synchronization time and algorithm calculation time of the cloud platform,which proves that the industrial power analysis and prediction based on the cloud platform has better efficiency and real-time performance.
作者
王宏飞
周鑫
徐哲壮
张士杰
夏玉雄
Wang Hongfei;Zhou Xin;Xu Zhezhuang;Zhang Shijie;Xia Yuxiong(School of Electrical Engineering and Automation,Fuzhou University,Fuzhou 350108;Fujian Huatuo Automation Technology Co.,Ltd,Fuzhou 350003)
出处
《电气技术》
2020年第12期6-11,共6页
Electrical Engineering
基金
国家自然科学基金资助项目(61973085)。
关键词
工业用电分析
用电预测
云平台
机器学习
industrial power consumption analysis
power consumption prediction
cloud computing platform
machine learning