Perpustakaan Digital ITB

The safety level of aviation is continuously improving. Although the methods that have been commonly used today have provided a big picture of how aviation safety is carried out, demands for methods to predict the probability of aircraft accidents arise. In addition, each airline is also required to define its own Acceptable Level of Safety Performance (ALoSP). This thesis discusses the relationship between the contributing factors that are formed by a model that incorporates aircraft behavior for runway overrun accident. Aircraft modeling includes aerodynamics, gravity, propulsion, and braking model. The analysis will be carried out using incident-free flight data to determine the probability for flights to experience runway overruns. Contributing factors are then extracted from the data to obtain the probability distribution. Then these contributing factors are propagated into the model to calculate the incident probability. This thesis uses Markov Chain Monte Carlo and Subset Simulation to find the incident probability. The results of the subset simulation produce a probability value of 2.45-E07 for runway overrun events. After the incident probability value has been obtained, sensitivity analysis is conducted to determine the most influential contributing factor to the occurrence of runway overrun. The use of brakes, the mass of the aircraft, and the height of the aircraft when crossing the threshold were the three main factors that contributed the most. From these results, recommendations are given to airlines to improve operational safety by setting a target height of the aircraft when crossing a threshold of 50 ft.