Challenges associated with generating accurate as-is building information models for existing buildings by leveraging heterogeneous data sources

Nowadays, facility management (FM) teams are facing challenges to generate accurate and semantically-rich as-is BIMs for existing buildings. Current model creation approaches, such as model generation based on point cloud data, mainly capture geometric information of a building and lack in providing additional semantic information about components and other project information. This paper provides the results of a detailed case study aimed at leveraging existing data sources (e.g., archived documents and data in FM systems) to generate accurate and semantically-rich as-is BIMs. The initial findings from the case study highlighted two main challenges associated with model generation from existing data sources: information extraction and integration. Existing information for different components is typically stored in heterogeneous data sources with various formats and quality and, hence, requires different approaches to extract information. The findings also showed that almost 40% of the component attributes investigated had conflicting values in existing sources. In order to address these challenges, formalized approaches are required to support conflict resolution, data extraction and integration.