Coupling Analysis of Chemical Accident Causation Factors in the Framework of Risk‐Structural Importance Controllability

ABSTRACT This study proposes a three‐dimensional framework—assessing risk, structural importance and controllability—to analyse the multidimensional causes of chemical accidents. We constructed a causal factor system via text mining of accident reports. Risk was quantified using a random forest algorithm, structural importance evaluated through social network analysis (SNA), and controllability assessed via fuzzy comprehensive evaluation. These dimensions were integrated using an entropy weight method to generate objective composite scores. Results identify five critical factors requiring urgent intervention: education and training, entity accountability, hidden hazard investigations, illegal production and management systems. These factors rank highest due to their combined high risk, structural centrality and controllability. Conversely, five factors showed lower priority for immediate systemic change: labour protection equipment, engineering supply quality, outsourced construction teams, equipment operation and noncompliant third‐party institutions. These ranked lowest, exhibiting low risk, structural importance and controllability. This framework offers a systematic approach for prioritizing safety interventions in the chemical industry.