Toward artificially intelligent cloud-based building information modelling for collaborative multidisciplinary design

The technological tools people use for designing buildings have progressed from drawings to descriptive ge-ometry, and from computer-aided drafting and design (CAD) to building information modelling (BIM). Yet despite their use of state-of-the-art BIM technology, the multidisciplinary teams that design modern buildings still face numerous challenges. Building models lack sufficient semantic content to properly express design intent, concurrent design is difficult due to the need for operators to maintain model consistency and integrity manually, managing design variations is cumbersome due to the packaging of information in files, and collab-oration requires making-do with imperfect interoperability between application software. In response, we pro-pose a 'Cloud BIM' (CBIM) approach to building modelling that seeks to automate maintenance of consistency across federated discipline-specific models by enriching models with semantic information that encapsulates design intent. The approach requires a new ontology to represent knowledge about the relationships between building model objects within and across disciplines. Discipline-specific building models are stored together with their data schema in knowledge graphs, and linked using objects and relationships from the CBIM ontology. The links are established using artificially intelligent semantic enrichment methods that recognize patterns of loca-tion, geometry, topology and more. Software methods that operate along CBIM relationship chains can detect inconsistencies that arise across disciplines and act to inform users, propose meaningful corrections, and apply them if approved. Future CBIM systems may provide designers with the functionality for collaborative multi-disciplinary design by maintaining model consistency and managing versioning at the object level.