A semantic model-based fault detection approach for building energy systems

In this paper, a semantic model-based approach is proposed for building energy systems fault detection. Its basic idea is to mimic the general intelligence of human experts in understanding massive amounts of operational data of various buildings, and further proposing customized fault detection solutions. A domain ontology is developed to allow computers to understand the prior knowledge of building energy systems fault detection. Classes and properties are developed to formalize all possible configurations in this domain. Semantic rules are proposed to detect the operation problems, control problems, equipment malfunction and sensor failure in building energy systems. These rules are written in abstract syntax. They can be reused in various building energy systems. For a target system, the building data are mapped to the ontology to generate a customized knowledge graph. The knowledge graph captures the physics underlying the system operations. The semantic rules are activated based on the knowledge graph to detect the faults. The proposed approach is demonstrated using the historical data from an industrial building located in Wuhan, China. The results show that the approach is powerful in providing the customized fault detection solutions for different situations. It has high levels of interpretability, reliability and automation. The knowledge graph is automatically updated with new data step by step. The semantic rules are activated if the conditions are satisfied based on the knowledge graph. The fault action mechanisms are captured based on the inference chains of the rules. Experts can find the fault reasons and take actions for commissioning.