Development of a fault diagnosis method for heating systems using neural networks

Abstract

The application of artificial neural networks (ANNs) for developing a fault diagnosis (FD) method in complex heating systems is presented in this paper. The six operating modes with faults used to develop this FD method came from the results of a detailed investigation in cooperation with heating system maintenance experts and are among the most important operating faults for this type of system. Because a daily diagnosis is generally sufficient, the ANNs have been developed using the daily values obtained by a preprocessing of the numerical simulation data. This paper presents the first step of the method development. It demonstrates the feasibility of using ANNs for fault diagnosis of a specific heating, ventilating, and air-conditioning (HVAC) system provided training data representative of the behavior of the system with and without faults are available. The next step will consist of developing a generic method that requires less training data.