Interpretable data-driven fault diagnosis method for data centers with composite air conditioning system

A convolutional neural network (CNN) based data-driven fault detection and diagnosis model considering temporal dependency for composite air conditioning system that is capable of cooling the high heat flux in data centers is proposed.