基于网格搜索优化的主成分分析-支持向量机算法的冷水机组能耗预测  

Energy Consumption Prediction of Chillers Based on PCA-SVM Algorithm with Grid Search Optimization

作　　者：刘峥[1] 黄真银 徐成良[3] 陈焕新[3] 李昱瑾 LIU Zheng;HUANG Zhenyin;XUN Chengliang;CHEN Huanxin;LI Yujin(China-EU Institute for Clean and Renewable Energy,Huazhong University of Science and Technology,Wuhan,Hubei 430074,China;Hubei Zhuoli Control Intelligent Technology Limited Company,Yichang,Hubei 443000,China;School of Energy and Power Engineering,Huazhong University of Science and Technology,Wuhan,Hubei 430074,China)

机构地区：[1]华中科技大学中欧清洁与再生能源学院,湖北武汉430074 [2]湖北卓立集控智能技术有限公司,湖北宜昌443000 [3]华中科技大学能源与动力工程学院,湖北武汉430074

出　　处：《制冷技术》2019年第6期15-20,共6页Chinese Journal of Refrigeration Technology

基　　金：国家自然科学基金（No.51876070,No.51576074）

摘　　要：针对冷水机组能耗受多因素影响的特点,本文提出了一种基于主成分分析(Principal Component Analysis,PCA)和支持向量机(Support Vector Machine,SVM)的冷水机组能耗预测模型。采用交叉验证和网格搜索法优化支持向量机(SVM),将PCA-SVM的预测结果与优化后的SVM进行比较,结果表明:优化后的SVM模型的拟合优度较未经优化的模型提升了12.88%,建模时长较未经优化的模型缩短了80%,实现了在提升预测精度的同时节省了计算资源。In view of the energy consumption of chiller is affected by multiple factors,a chiller energy consumption prediction model is proposed based on principal component analysis(PCA)and support vector machine(SVM).Cross validation and the grid search methods are used to optimize the SVM.The results show that,compared with popular SVM method,goodness of fit of the optimized SVM model is increased by 12.88%,and modeling time of it is shortened by 80%,which predict the energy consumption of the chiller and save computing resources effectively.

关 键 词：冷水机组能耗预测 主成分分析 支持向量机 参数寻优 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TU831.3[自动化与计算机技术—控制科学与工程]