Application of black-box models to HVAC systems for fault detection

Abstract

This paper describes the application of black-box models for fault detection and diagnosis (FDD) in heating, ventilat-ing, and air-conditioning (HVAC) systems. In this study, mul-tiple-input/single-output (MISO) ARX models and artificial neural network (ANN) models are used. The ARX models are examined for different processes and subprocesses and compared with each other. Two types of models are established--system models and component mod-els. In the case of system models, the HVAC system as a whole is regarded as a black box insteati of as a collection of compo-nent models. With the component model type, the components of the HVAC system are regarded as separate black boxes