A hybrid SEM-FCM approach for data-driven fatigue risk management in construction

As a national economic cornerstone, the construction industry faces frequent accidents mainly caused by operator fatigue-induced unsafe behaviours. Existing research focuses on single fatigue dimensions, lacking exploration of multidimensional fatigue interactions and their association with safety attitudes; traditional methods are also subjective and unable to quantify complex nonlinear relationships. This study analysed 301 excavator operator surveys via structural equation modelling, finding muscular fatigue shows the strongest direct association with unsafe behaviours, and cognitive fatigue exhibits the largest total association, with safety attitudes partially mediating the pathways between all three fatigue dimensions and unsafe behaviours. Fuzzy C-Means clustering identified high and low-risk groups: the high-risk cluster, characterised by high muscular-cognitive fatigue and negative safety attitudes, exhibits 72.12% higher unsafe behaviours. This research provides scientific evidence and operational tools for construction enterprises to formulate differentiated fatigue risk management measures, including optimising work-rest schedules and fostering positive safety attitudes.