Paper Conference

Proceedings of eSim 2020: 11th Conference of IBPSA-Canada


Fault detection and diagnosis of air temperature sensors in an air handling unit using machine learning techniques

Behrad Bezyan, Radu Zmeureanu
Centre for Net-Zero Energy Buildings Studies, Department of Building, Civil and Environmental Engineering, Concordia University, Montreal, Quebec, Canada

Abstract: This paper presents the development of machine learning models for multiple fault detection and diagnosis of air temperature sensors of an air handling unit (AHU). For this purpose, air temperatures in critical points of an AHU such as mixed air temperature and temperature after the heating coil, are predicted. A compound artificial neural network (ANN) model is proposed for prediction of air temperature sensors, then the fault is detected if the differences between measured and expected values exceed the defined threshold. For fault diagnosis aspect, the Recurrent Neural network (RNN) as a deep learning model, and shallow Feedforward Neural network are developed for prediction of the air temperature value at the current time step (t) using the previous measurements of that target sensor; the faults are diagnosed if the residuals exceed the threshold. This paper uses the synthetic hourly data from the simulation of an institutional building with eQuest program as a proxy for real measurements. Models developed in this paper will be used in future work, and will be tested with real measurements for the multiple fault detection and diagnosis (MFDD) purposes.