Arcabouço conceitual de tecnologias semânticas para automação da certificação LEED

Abstract

The evaluation of sustainable building projects aims to measure the sustainability performance of built environments and is important to improve the environment in increasing deterioration. However, most assessment tasks are still performed manually, consuming time and increasing the risk of errors. In civil construction, BIM (Building Information Modeling) technology is used to simulate a project in its life cycle. Faced with the increasing use of this technology, the opportunity arises to use your data and relate it to other data sets, to analyze different aspects of the project. The use of semantic technologies is an alternative to integrate data and make inferences in the decision-making process. In an integrated workflow, mediated by ontologies, mapping between building elements and other data can aid in expert analysis. This work presents the development of a conceptual framework, supported by semantic technologies, to establish the relationship between BIM data and other data for automating assessment for LEED certification (Leadership in Energy and Environmental Design). The process for implementing the artifact applies the Design Science Reseach and begins with obtaining the file in IFC format (Industry Foundation Classes) of a project developed in BIM for the development of a simplified ontology. The ontology of this work contemplates the general requirements of the conceptual modeling language OntoUML, based on the UFO ontology (Unified Foundational Ontology). The SEBIM ontology seeks to model more complex relationships, contemplating classes to organize the materials and zones of a built environment, as well as the criteria to be scored. For the integration with data external to the built environment, open data about the surroundings and means of transport are obtained. These tabular data sets contain geometric coordinates of areas and points in the city, following the OGC (Open Geospatial Consortium) standard. Once the relevant datasets for this research have been selected, the data are mapped into ontologies and semantically annotated, following the Semantic Data Dictionary (SDD) approach for conversion into RDF (Resource Description Framework). The RDF graph is generated containing the formalization of the data and favoring the integration for querying about city locations by spatial data. The datasets are inserted into a triplestore to perform inferences in SPARQL, generated by converting the criteria for evaluating the certification into queries and inserting the new inferred triples in the dataset. A final consultation is carried out to present the result of the inferences to promote new knowledge to the specialists. The artifact developed is verified with experiments and validated according to the competence questions referring to the certification criteria. The results showed that the application of these technologies promotes the semantic extension of the constructive elements in BIM, facilitates the integration with other knowledge bases and conceptually organizes the data to better retrieve information. The developed inferences can be considered in other studies that seek to relate and integrate data to discover new correlations and create new knowledge in data-based applications.