Causal modeling using Bayesian belief nets for integrated safety at airports

Abstract

A study funded by the Federal Aviation Administration via the Dutch Ministry of Transport has explored the possible structure of a causal model for aviation safety. The study is initiated by the regulator and the parties concerned at the airport to improve the modeling of risk in order to better understand the effects on the level of risk of different influencing factors. Existing modeling gives no insight into how the many factors under control of air traffic control, the airport or the airlines, play their part in the control of risk. The objective of the new modeling approach is to allow such assessment. This article describes the overall approach proposed for the modeling and discusses the requirements and problems for full implementation. It is illustrated with two case studies considering the aspect of missed approach and the effect of flight-crew fatigue on performance.