AN APPLICATION OF KRIGING MODELING FOR CHILLER FAULT DETECTION AND DIAGNOSIS STRATEGY

Abstract

An accurate parameter model plays an essential role in enhancing the accuracy rate of the chiller fault detection and diagnosis strategy. Thus, Kriging as a parameter model is proposed in this study to capture operating characteristics for a chiller system. Then, a strategy includes the Kriging modeling, ttest method and a diagnosis rule which is based on the most sensitive feature parameter related fault, is combined in this study. The proposed chiller FDD strategy is tested and validated using the chiller experimental data of ASHRAE RP-1043. Results of this study show that the approximation ability of KRG model achieves high accuracy and this proposed method is robust for fault detection and diagnosis in chiller systems.