Assisting automated fault detection and diagnostics in building automation through semantic description of functions and process data

Abstract

Building automation (BA) systems ensure comfort and safety of the occupants while optimizing energy consumption. Malfunctions of electro-mechanical equipment or the associated control function result not only in incorrect operation, but also wastage of energy. Automated fault detection and diagnostics (AFDD) of processes like heating, ventilation, and air-conditioning (HVAC) is a challenging task due to the diversity of the applications and inter-dependencies between the equipment. Semantic description of the building and its equipment can help AFDD agents to better understand the system and continuously examine its operation. However, AFDD using only the structural description of the building topology and the contained equipment without relating it to the control functions requires manual verification of the root cause. In this work, we added a model of the control functionality to the semantic information of the structural and topological description and represented it using W3C Web of Things Thing Descriptions. Based on these descriptions, we have demonstrated in a case study of a building installation that fault detection rules which have access to the semantic description of underlying control functions can be formulated more precisely and thereby avoid false positives.