Automated design for safety performance evaluation system using accident probability and risk cost analytics: A case study of industrial construction

Design for Safety (DfS) has been widely recognized as an effective strategy for reducing construction accidents; however, its practical implementation remains limited due to the lack of quantitative and repeatable evaluation methods at the design stage. This study proposes an Automated Design for Safety Performance Evaluation System (DfS-APES) that enables document-based, quantitative comparison of safety performance between AS-IS and TO-BE design alternatives. The proposed system integrates resource-based exposure data, baseline accident statistics, and standardized loss cost assumptions to estimate expected fatalities, expected injuries, and expected risk cost associated with design options. The methodology was implemented as a web-based system and validated through a case study involving ten construction work packages from an industrial project. The results demonstrate that the system can consistently distinguish safety performance differences attributable to design changes, identify both safety improvements and potential regressions, and highlight high-impact construction activities through contribution analysis. In particular, labor-intensive and temporary works were found to be dominant drivers of risk cost reduction. The study demonstrates the feasibility of automated, quantitative DfS performance evaluation using commonly available design documents and provides a practical decision-support tool for designers and project owners to support proactive safety management at the early design stage.