Abstract
Construction sites are considered high-risk working environments. Safety risk analysis of construction sites is one of the important aspects of construction safety management. However, existing construction safety risk analysis methods typically face overreliance on subjective experience and cannot reflect the real-time risk level of construction projects. Crucial information about construction accidents implies a complex semantic network that can obtain objective quantitative data and then serve the quantitative analysis of construction safety risk. Therefore, a knowledge-graph-improved dynamic risk analysis method for behavior-based safety (BBS) management on construction sites is proposed. Specifically, this study quantifies the risks and consequences of unsafe behaviors in construction by conducting graph topology analysis based on historical accident data, improves the grey clustering model, and calculates the construction site risk. Thus, the proposed method combines expert experience and objective historical data to give more objective and realistic results. The results of the application of this method on one project show that the improved dynamic analysis method can utilize the historical accident data to achieve a more reasonable overall risk grading for BBS on construction sites; at the same time, the method can determine real-time key BBS indicators; further, the construction safety management measures for key BBS indicators can effectively reduce current BBS risks of a construction project.