Risk Evaluation Method for Aviation Unsafe Incidents Based on Cloud Model

In the pursuit of advancing aviation safety, the utilization of quick access recorder (QAR) data has become increasingly critical. This study presents a refined risk assessment model that integrates cloud model theory with QAR data to address inconsistencies in data standards and enhance the robustness of risk evaluation methodologies. The proposed model employs entropy weighting to determine the relative significance of various indices and utilizes the [Formula: see text] rule to systematically classify flight risk levels. Focusing on runway excursions—a prevalent category of safety incidents—this research examines two primary scenarios: veer-off and overrun. By incorporating four key indices across two dimensions and applying similarity calculations, the model effectively quantifies the risk associated with runway excursions. The results indicate that the majority of analyzed flights fell within the “Very good” and “Good” safety categories, demonstrating strong alignment with actual flight data and validating the model’s accuracy. Furthermore, its adaptability enables application to other critical safety incidents, such as hard landings and tail strikes, thereby offering a comprehensive tool for enhancing aviation risk management practices.