Fault diagnosis, whole building faut, Bayesian network
Buildings consume more than 40% of primary energy in the U.S. and 57% of the energy usage in commercial and residential buildings are consumed by the heating, ventilation and air conditioning (HVAC) system.Malfunctioning sensors, components, and control systems, as well as degrading systems in HVAC and lighting systems are main reasons for energy waste and unsatisfactory indoor environment. In HVAC systems, faults occur in one component or equipmentcan also cause abnormality in other subsystems because of the coupling among different subsystems. Therefore, whole building level fault diagnosis methods is critical to locate fault root cause and isolate the fault. Bayesian network (BN) is a prevalent toolin fault diagnosis which can deal withprobabilistic reasoning of uncertainty. In this paper, a two-layer Bayesian network which consists of fault layer and fault symptom layer is developed to diagnose whole building HVAC system fault. Weather information based Pattern Matching (WPM) method which was employed in fault detection was also used to create baseline data and generate LEAK probability. BAS data from a campus building are collected to evaluate the effectiveness of the proposed method.