Fault diagnosis of air handling system based on enhanced long short-term memory network

HVAC air handling systems have strong dynamic time-varying and batch-dynamic characteristics. In order to effectively diagnose the detected fault patterns, this paper constructs a fault diagnosis mode based on enhanced long short-term memory (LSTM) network, which can efficiently identify the sparse and slow features of the fault data. A case study based on the ASHRAE research project RP-1312 experimental dataset shows that the proposed method has a significant improvement in identifying air handling system faults compared with the related fault identification methods.