POTENTIAL APPLICATIONS OF ARTIFICIAL INTELLIGENCE FOR FACILITIES MANAGEMENT

Abstract

Facilities Management (FM) contains repetitive activities which could potentially be replaced by technology. Building Information Modelling (BIM) for FM has become a ubiquitous topic of research in recent years. Advances in BIM technology are achieving a certain level of maturity in Singapore, but the FMs seem to not be adept in operating and maintaining BIM models. The lack of skills in the technology is the main barrier of why BIM has not taken off in FM like it is promised in research studies.

Artificial Intelligence (AI) as a tool to improve workflows for FM is unheard of. However, AI for other industries has been implemented successfully AI has gained much recognition all around the world. AI has been utilised to replace time-intensive activities and error prone tasks to achieve higher productivity and efficiency.

This research examines precursory literature review on FM, BIM-enabled FM and AI, analyses the applicability and feasibility of AI functions to replace manual, time-consuming FM work processes and peruses the main barrier of BIM implementation in FM to justify the need for AI introduction to be user-friendly.

A feasibility study was carried out and backed by experimental results and surveys with local FM professionals to establish the potential of the two proposed functions (AI Image Recognition & AI Chat Bot) for FM. This research aims to allow AI to be implemented in FM without the need for staff to be equipped with programming knowledge or skills. An Image Recognition Model was successfully trained to achieve a 99.48% percent to identify the presence of cracks on wall surfaces using dependencies which require no prior programming knowledge.