Unraveling hierarchical penetration mechanisms and coupling relationships of safety risks in major transportation...

The continuous expansion and increasing complexity of major transportation infrastructure construction have led to frequent safety accidents during construction. To address the limitations of traditional methods in safety risk management and to reveal the hierarchical penetration mechanisms and coupling relationships of risk factors, this study established a new paradigm for understanding safety risks in transportation infrastructure construction by integrating text mining and complex network theory, based on 244 construction safety accidents that occurred in China between 2010 and 2023. First, text mining was used to extract 101 key features related to safety accidents in major transportation infrastructure construction from raw data, and these features were utilized to develop an improved Human Factors Analysis and Classification System (HFACS) model. Subsequently, a co-occurrence matrix was employed to quantify the frequency and strength of associations among factors, leading to the construction of a Safety Risk Network (SRN) with risk factors as nodes. By calculating the Multi-Feature Gravity Model (MCGM) value and the comprehensive degree C(i) for cross-validation, the following findings were revealed:① Inadequate safety training, though widely connected, have a weak global impact;② Explicit factors such as equipment defects are easily identifiable and are predominantly influenced by supervisory-level factors;③ Safety management in construction should focus on addressing systemic deficiencies at the organizational and supervisory levels rather than merely rectifying operational-level issues. Finally, a hierarchical blocking and proactive defense strategy was developed using the Bow-Tie Model, providing a quantitative decision-making basis for enhancing the intelligence of safety management in major transportation engineering projects.