Establishment and validation of a novel composite index for energy efficiency evaluation of data center chilled water systems based on AHP and Entropy Weight Method

Abstract

Data centers typically use various energy efficiency indicators to assess performance from multiple perspectives, such as overall energy efficiency or the performance of subsystems and components. However, due to the differing focus of these indicators, such as those used to evaluate the basic equipment in chilled water systems, it is challenging to directly compare the comprehensive performance of data center chilled water systems based on the evaluation results. To address this challenge, this study proposes a comprehensive index, PCC (Performance of Composite Chilled water system in Data Centers), based on the Analytic Hierarchy Process (AHP) and the Entropy Weight Method, which integrates basic energy efficiency indicators to provide a comprehensive evaluation of the chilled water system’s energy efficiency. The basic energy efficiency indicators exhibit variability under different operating conditions, and the hybrid AHP-Entropy Weight Method balances the weights of these indicators, combining them into PCC. The proposed PCC was validated using a chilled water system simulation model built in TRNSYS. The results show that, compared to the Power Usage Effectiveness (PUE) of data centers and commonly used evaluation indicators for chilled water system, PCC is capable of evaluating the comprehensive energy efficiency of the chilled water system ’s basic equipment. Moreover, when operating conditions change, the average Relative Change Rate (RCR) of the PCC evaluation results can be up to 21 times higher than that of the PUE’s RCR, and up to 3 times higher than that of the commonly used chilled water system efficiency indicators’ RCR. The high sensitivity (RCR) of PCC can assist identify optimization potential within control strategies for data center chilled water systems.