Automated Diagnosis for Heating Systems in Buildings

Estimations of the energy consumption of buildings range from 25 to 40%. Approaches like predictive control can reduce energy consumption by up to 27%. However, it is also well-known that due to a lack of maintenance activities, e.g., heating devices perform at a lowered efficiency level. Therefore, there is a need for fault diagnosis, especially in the building sector, to reduce energy consumption further. This work introduces different diagnosis approaches, ranging from data-centric to knowledge-based approaches. We introduce the foundations and explain the advantages and disadvantages of the various diagnosis methodologies, including the application of machine learning and model-based and qualitative reasoning. In addition, we discuss using different diagnosis methods on a building's heating system and compare the obtained outcome of the approaches. The work aims to provide a profound introduction to diagnosis and its methods, focusing on the application domain of building automation.