Using cluster and multivariate analyses to appraise the operating performance of a chiller system serving an institutional building

F. W. Yu, K. T. Chan

Research output: Contribution to journalArticlepeer-review

21 Citations (Scopus)

Abstract

This study explains how to appraise the operating performance of a chiller system using cluster and multivariate analyses. Cluster analysis is a statistical tool used to identify groups of individuals similar to each other but different from individuals in other groups. This serves the purpose of classifying what operating condition constitutes a high or low system coefficient of performance (COP). A case study has been carried out on a system with five sets of chillers, pumps and cooling towers. Each operating condition to be clustered involves the following variables: the load of each operating chiller, the total system load, the number of primary chilled water pumps running, the number of condenser water pumps running and the number of cooling towers operating. The two-step cluster analysis was applied to group around 6800 sets of operating conditions into three clusters with distinct operating characteristics. For each clustered data, a multivariate analysis was made to examine the degree of correlation between the system coefficient of performance (COP), the operating variables and typical weather parameters. The significance of this study is to demonstrate a systematic method to ascertain if a chiller system operates under optimal control with maximum COP and to estimate COP improvements if operating problems are rectified.

Original languageEnglish
Pages (from-to)104-113
Number of pages10
JournalEnergy and Buildings
Volume44
Issue number1
DOIs
Publication statusPublished - Jan 2012

Keywords

  • Multiple linear regression
  • Two-step cluster
  • Water-cooled chiller

Fingerprint

Dive into the research topics of 'Using cluster and multivariate analyses to appraise the operating performance of a chiller system serving an institutional building'. Together they form a unique fingerprint.

Cite this