A Review of Data-driven Fault Diagnosis Algorithms for Air Conditioning Systems
With the economic development and the growing demands for better living environment quality,more and more buildings have installed with central air conditioning systems.However,faults in these systems lead to energy waste,decline in indoor environmental quality,and even failure to meet the thermal comfort needs of occupants,resulting in thermal complaints.Concurrently,the increasing application of automatic control technology in central control systems,along with the surge in the quantity of sensors and voluminous data,have made it challenging for traditional white-box fault diagnosis algorithms to fulfill the requirements for air conditioning system fault diagnosis.The continuous development of artificial intelligence technology,particularly machine learning,has brought about novel approaches to fault diagnosis algorithms.This paper provides a comprehensive review of the research conducted over the past two decades on data-driven fault detection and diagnosis(FDD)applied to HVAC systems.Due to the issues of data-driven methods,such as their dependency on large-scale data and insufficient interpretability of their models,there is a growing trend to integrate physical models with data-driven methods.The aim of this integration is to enhance interpretability,improve accuracy,and reduce the dependency on large-scale data.It is expected that this will emerge as a new direction for future research on fault diagnosis,with a view to achieving more efficient and accurate diagnoses.
NETL
NSTL
万方数据
