Advance and prospect of machine learning based fault detection and diagnosis in air conditioning systems

Abstract

Fault detection and diagnosis (FDD) are crucial aspects of maintaining efficient and energy-saving heating ventilation and air conditioning (HVAC) systems. Conditions such as inadequate maintenance, poor equipment performance, improper installation and defective control mechanisms can all contribute to a reduction in the operational efficiency of HVAC systems, resulting in unnecessary energy wastage. Machine learning serves as a potent tool in diagnosing air conditioning systems. The optimization and evaluation parameters of diagnostic methods have not been extensively studied, although the existing reviews of HVAC FDD have covered the relevant analysis of fault types and diagnostic methods. In this study, common faults in HVAC systems in recent years are analyzed and the commonly used FDD methods and their respective applications are reviewed. Optimization and evaluation parameters for diagnostic methods are also investigated. In addition, changes in important issues, diagnostic methods, and fault types over time are discussed and future research trends are analyzed. Finally, the research prospects in the field of HVAC FDD are discussed to provide a reference for future research.