Evaluation and Classification of Accident-Inducing and Risk Propagation in Airport Apron Networks

Airport apron operations involve complex interactions among personnel, equipment, and environmental factors, posing significant safety challenges. To address these risks, this study establishes an innovative dual evaluation system combining static and dynamic methods to assess accident-inducing potential and risk propagation capability. Based on accident investigation reports from 2015 to 2022, a risk network with 82 nodes and 244 edges was constructed. Two static indicators—Target Oriented Centrality (TOC) and Accident Neighbor Risk Rate (ANRR)—were used to evaluate accident-inducing potential, while risk propagation capability was dynamically simulated using the Susceptible-Infectious-Recovered-Susceptible (SIRS) model, with Average Peak Infection Proportion (APIP) and Infection Cycle (APIC) as key metrics. Risk nodes were categorized into four clusters using spectral clustering, which was selected after comparing multiple algorithms with the Calinski-Harabasz (CH) index. Nodes were prioritized into three tiers based on accident-inducing potential and risk propagation capabilities. Results show that Tier 1 nodes (personnel and equipment) exhibit the highest accident-inducing potential, while Tier 2 nodes (management) play significant roles in risk propagation. Tier 3 nodes (environment) have limited direct impacts but may pose long-term risks. This system provides actionable insights to prioritize safety interventions, offering a robust framework for enhancing apron safety.