Fault Detection and Diagnostics in Ventilation Units Using Linear Regression Virtual Sensors

Buildings represent a significant portion of global energy consumption. Ventilation units are one of the largest components in buildings systems and are responsible for large part of energy consumption. Ventilation units are complex components, often customized for the specific building. Their faults impact buildings' energy efficiency and occupancy comfort. In order to ensure their correct operation, proper Fault Detection and Diagnostics methods must be applied. Hardware redundancy, an effective approach to detect faults, leads to increased costs and space requirements. We propose to exploit physical relations inside the unit to create virtual sensors from other sensors' readings, introducing redundancy in the system. We create linear regression models for three sensors using other sensors related through physical laws as inputs. We use two different measures to detect when a virtual sensor deviates from the actual one: coefficient of determination and acceptable range. We test our method on a real building at the University of Southern Denmark. Our method detects a fault in temperature sensor, where its readings have an abnormal trend and fall outside acceptable ranae for one day.