Research into the development of fault detection anddiagnosis (FDD) methods as applied to heating, ventilating, air-conditioning, and refrigerating equipment has been ongoing for over a decade, and several papers have been published. However, studies specifically related to large chillers have been few despite the importance of such equipment. The objective of this paper is to propose and illustrate a simple model-based FDD method for medium to large chillers that uses sensors available in most cooling plants, allows tuning of specific thresholds so as to attain the desired compromise between robustness and sensitivity, and has the potential to be automated and implemented online. Lacking actual field data of chillers operated under fault-free and faulty operation, the performance of the proposed FDD method is demonstrated with data generated from tests on alaboratory chiller in the framework of an earlier research study. This proposed FDD scheme is based on five important characteristic features (that have been identified from 15 variables evaluated) that allow six process faults to be identified (although two cannot be done uniquely). Since large chillers are not prepackaged as is unitary equipment, their design and assembly allow selecting different subsystems for a stipulated maximum cooling capacity. In such a case, selection of specific variables needed for FDD will be impacted by different faults, depending on several aspects unique to the chiller installation: the type of chiller load control (thermostatic expansion valve or inlet guide vane), specific choice of the relative heat exchanger size, and the load fraction at which the chiller is usually operated. Hence, it is likely that some customization would be needed to adapt the FDD thresholds and association rules for each chiller installation. Different options and challenges relevant to practical implementation of the proposed FDD method are discussed.

Units: SI