冷水机组回归模型的分析与评价  

Analysis and Valuation of Chiller Plant Regression Model

作　　者：王群 WANG Qun(Shanghai Tansuo Energy Service Limited Company)

机构地区：[1]上海碳索能源服务股份有限公司

出　　处：《上海节能》2020年第7期734-746,共13页Shanghai Energy Saving

摘　　要：以某项目的数据对6种冷水机组模型(Simple Linear (SL)Regression Model、Bi-quadratic (BQ)Regression Model、Multivariate Polynomial (MP) Regression model、Gordon-NG Universal (GNU) Model、Gordon-Ng Simplified (GNS) Model、Lee’s Simplified (LS) Model)的预测精度进行分析与评价,并讨论了当数据量减少时模型预测精度的变化,结果发现GNU模型的性能更优秀,可以作为冷水机组故障诊断的模型。The author analyzes and values prediction accuracy of 6 chiller plant models(Simple Linear(SL) Regression Model、Bi-quadratic(BQ) Regression Model、Multivariate Polynomial(MP) Regression model、Gordon-NG Universal(GNU) Model、Gordon-Ng Simplified(GNS) Model、Lee’s Simplified(LS)Model) in some project and discusses prediction accuracy changes while data reduces.The results show that GNU model has better performance as chiller plant failure diagnosis model.

关 键 词：冷水机组 回归模型 预测精度 GNU模型 

分 类 号：TB65[一般工业技术—制冷工程]