Studies show that buildings are able to attain energy savings of up to 30% by rectifying faulty HVAC systems using fault detection and diagnostic methods. In this study, we propose an automated top-down model-based fault detection method that can detect abnormal electricity consumption of building ventilation systems. This is achieved by employing a dynamic building model of the highly efficient OU44 teaching building at the University of Southern Denmark in Odense, allowing for identification and quantification of sub-optimal performance in real-time. The methodology is based on a statistical definition of abnormal telecommunicational operation that employs an approximation of the Chernoff bound proposed by Cheung et al. We develop and implement the Chernoff bound method for building ventilation systems and propose a simple approach of defining threshold limits based on a percentage of the model predicted performance of the ventilation systems. Each time the observed data deviates from the threshold limits, a suspicion is then evoked. The urgency of each suspicion period is classified into three levels of urgency (high, medium and low) depending on the probability of abnormality calculated from the Chernoff approximation, thereby aiding building managers in resource delegation matters. A FDD analysis for a time period of two months is conducted using the proposed method. The results show some high urgency classified suspicion periods within the investigated period, which are manually diagnosed by utilising room level sensor data. This led to the observation of a faulty occupancy counter. The proposed method shows promising outlook within the field of automated fault detection methods in terms of highlighting periods with abnormal behaviour. (C) 2018 Elsevier B.V. All rights reserved.