Abstract:
Every architectural and building engineering design in the AEC/FM (Architectural, Engineering, Construction, and Facility Management) domain is subject to conformance with a diverse range of statutory requirements and standards before the official consent for construction may be granted by the relevant authority. There is an enormous amount of information to process and share among the stakeholders for this to occur. Historically, building information has been exchanged in the domain using paper-based drawings and written specifications. Auditing a building design for compliance, for example, has been a manual undertaking against a voluminous corpus of paperbased legal texts that are written in natural language intended for human interpretation. This paper-based practice is resource intensive, inefficient, and error-prone, which has sometimes resulted in costly remediations of problem buildings and even loss of lives. It is also a factor for declining productivity in the domain. There are now methods of sharing digital building information via building information modelling (BIM), which promotes better collaboration and provides more opportunities for automation in the compliant building design process. However, there has not been a successful standard automated compliance audit implementation to date despite considerable research effort over the years. This research has explored a practical approach where human designers retain their natural role of making compliant design decisions. Machines are only given the task of doing what they do best, which is to execute repetitive instructions and procedures consistently and accurately. The findings of the research suggest that this human-machine partnership approach allows conventional compliant design procedures (CDPs) to be formalised as executable workflows to guide the automated audit of BIM models against regulatory knowledge models (RKMs). An important criterion in the approach is that CDP and RKM representations must be manageable by the domain experts equipped with only basic computing skills. A software framework (ARCABIM) has been developed to illustrate the approach using a test building model for evaluation. The potential of interfacing with external computational tools required to audit against qualitative performance-based criteria is described. A fire compliance model (FCM) representing a BIM model view for the compliant fire safety design of buildings has been developed and used in a case study to demonstrate the capabilities of the approach in conjunction with an RKM representing a performance-based fire safety design method of the New Zealand Building Code.