Modelling uncertainty in runway safety intervention performance evaluation

The uncertainty surrounding the occurrence of runway accidents and the generally accepted proactive performance indicator for runway safety activities are rarely reported. This study presents the uncertainty associated with runway accident occurrence and evaluates the performance of runway safety activities. Runway accident hazards were modelled using the Bayesian Belief Network. This was integrated into a system dynamics stock and flow diagram. System dynamics software Vensim was used to validate the model. Data spanning ten operational years were collected from the Federal Aviation Authority, Nigeria, to estimate parameters for model implementation. Six runway accident hazards representing the parent nodes with two states each were modelled. The child node was runway accident with two states, namely YES and NO. The Bayesian probability of occurrence of runway accident was obtained. In all, 23 runway safety quantities were identified. Runway safety intervention performance measures were the average number of runway accidents caused, average number of runway accidents prevented, runway safety benefit/loss and the runway safety intervention investments' breakeven point. The study shows that the hybrid model developed can serve as a useful tool to evaluate the behaviour and performance of runway safety.