Integrating a Knowledge-based Recommendation System into a BIM Workflow for Energy Efficient Facility Management
The presentation relates to an ongoing research project that aims at developing a recommendation system for energy-efficient operation of buildings by the means of BIM and semantic web technologies. So far, many developments and research have led to enhanced building automation systems embedding data analytics algorithms and performing energy-optimized building systems control. The presented approach tries to bring a complementary added value to such systems by the means of semantic modeling and knowledge reuse. It aims at providing an additional analysis layer to traditional algorithmic and model-predictive approaches in which a semantic analysis and interpretation of the operational state of a building is executed. For that purpose, it implements a knowledge base of energy conservation measures that prescribe building operation actions and handlings that a building user or a facility manager may execute for saving energy. The system consumes for a part building data gathered during its operation through a monitoring system. For another part, it relies on information contained in initial BIM-compliant building design models. The resulting software application might be used in the future as an add-on to existing building management systems (BMS). Modern BMS systems play a major role in today's race for the reduction of energy use and greenhouse gas emission. They are able to handle a huge amount of data that are analyzed for supervising, controlling and benchmarking buildings. BMS data are mainly gained through sensors and meters that provide information about e.g. the operational state of technical equipment, indoor temperature or energy consumption. Because of its highly time-dependent nature, this kind of information can be categorized as dynamic data about a building in contrast to static data which represent the building and its technical systems as they are i.e. as built physical entities. This latter kind of information encompasses data about the energy system components, their technical characteristics and their layout in the building. Even if numerous dynamic data are produced during building operation there is no much use of building static information created during its design. In view of that, the proposed methodology aims at closing this informational gaps between building design and operation by making reuse of initial design models serialized in IFC. The proposed methodology relies on the relationships existing between dynamic and static data which are necessary to perform a semantic analysis of the building. In existing BMS those relationships are semantically poor and only contained partially in the backend data model of the BMS, like a relational database in most cases. The methodology follows then data integration steps that use initial building information models and extend them with operation data as well as a knowledge model. The resulting semantic building information model is then used for interpreting building energy system behaviors and identifying best energy conservation measures. More specifically, an energy system ontology is introduced that supports reuse of knowledge for optimized building operation.