A Comprehensive Safety Risk Evaluation Method for Low-Altitude Flights

Abstract

To scientifically identify and assess the safety risks of low-altitude flights, this study adopted the “human-machine-environment- management” framework from safety systems engineering theory. A total of 23 risk factors affecting safe operations were initially identified and extracted, forming a safety evaluation index system for low-altitude flight projects. Subsequently, the Likelihood-Exposure-Consequence (LEC) risk evaluation model was applied to these 23 risk factors using a “qualitative-quantitative-qualitative” approach to extract 10 key risks for low-altitude flights. The weight values of these key risk indicators were calculated using a game-theory-based combination weighting method. Finally, a safety risk evaluation method integrating the LEC model, game theory, and a backpropagation neural network was developed using deep learning algorithms. Using the Fuxi Mountain paragliding project in Zhengzhou, China, as a case study, the safety risk level of the low-altitude flight project was assessed, and the feasibility of the evaluation model was validated. Results show that the maximum safety risk level for the project is 63.632, while the minimum is 58.543, categorizing it as relatively unsafe. Key influencing factors include pilot psychological quality, safety and protective equipment, enterprise supervision, and low-altitude training exercises. The findings provide a method support for safety management and emergency decision-making in low-altitude flight projects.