Evidence-Driven and Neural Network-Based Fault Monitoring and Diagnosis Technology for District Cooling and Heating System

The timely identification of the leakage and diagnosis of the specific location and degree of leakage can guarantee safe system operation of district cooling and heating systems. This study proposes an evidence-driven and neural network-based fault-monitoring and diagnosis method for district cooling and heating systems to mitigate the problem of traditional data-driven methods that rely on the quality and quantity of data, improving the robustness of the fault-monitoring and diagnosis model. The method utilizes a fault-monitoring model based on evidence-based K-nearest neighbor classifiers to monitor the system-operation status, and determines the specific leakage location and leakage amount through a neural-network-based leakage fault-diagnosis model. An actual district heating system in Chengde, Hebei Province, is used as a case study, and the results show that the accuracy of the method for fault monitoring and the diagnosis of district cooling/heating systems is 95.8%.