Prefabricated buildings have gained significant traction worldwide due to their efficiency, quality control, and sustainability advantages. However, the construction phase remains characterised by diverse construction activities and dynamically evolving risks, many of which are interdependent and difficult to predict using conventional, static risk assessment methods. Existing studies primarily focus on isolated risk categories and lack a unified framework capable of representing complex risk relationships while supporting probabilistic, real-time inference. To address this gap, this study proposes an ontology-based Bayesian network (OBN) model for dynamic and intelligent risk assessment in prefabricated building construction. The proposed framework systematically integrates objective risk knowledge derived from construction standards, prior literature, and a structured WBS–RBS mapping mechanism with subjective knowledge obtained from expert judgement and large-scale questionnaire perceptions. Ontology modelling is employed to formalise risk concepts and relationships, ensuring semantic consistency and machine-interpretable representation, while a Bayesian network enables probabilistic inference and real-time risk updating under uncertainty. Risk indicators were identified and refined through literature review and expert validation, followed by quantitative data collection from 406 industry practitioners involved in prefabricated construction projects. The survey results demonstrate excellent reliability, with Cronbach’s Alpha exceeding 0.98, providing a robust basis for probabilistic modelling. The Bayesian network supports forward and reverse inference, sensitivity analysis, and optimal causal chain identification, enabling dynamic risk prediction and prioritisation. To enhance practical applicability, the framework is operationalised through the development of a prototype decision-support system, the Prefabricated Construction–Risk Management System (PC-RMS), which links risk factors to observable on-site indicators and corresponding control actions. The results confirm that the proposed OBN framework offers a transparent, adaptive, and practically deployable solution for intelligent risk management in prefabricated building construction.
Chenya et al. (Fri,) studied this question.