Towards a causal model for air transport safety - an ongoing research project

摘要

The development of the Netherlands international airport Schiphol has been the subject of fierce political debate for several decades. One of the considerations has been the safety of the population living around the airport, the density of which has been and still is growing. The concern about this third party risk has been enhanced by the crash in 1993 of a 747-200 cargo plane into a suburb of Amsterdam, the capital of the Netherlands. In the debate about the acceptability of the risks associated with the air traffic above The Netherlands extensive use has been made of statistical models relating the movement of airplanes to the risks on the ground. Although these models are adequate for the debate and for the physical planning round the airport, the need has arisen to gain a more thorough understanding of the accident genesis in air traffic, with the ultimate aim of improving the safety situation in air traffic in general and around Schiphol in particular. To this aim a research effort has started in order to develop causal models for air traffic risks in the expectation that these will ultimately give the insight needed. The current paper describes the structure of a 3 year project aimed at developing a full operational causal model, which can be used for comparing alternatives for strengthening safety measures, for finding causes of incidents and accidents and for quantification of the probability of adverse events in the aviation industry. It discusses the obstacles to be overcome in the project, and can therefore be seen as a review of the state of the art of causal modelling at this moment in aviation. The project will result in a working causal model containing the minimum of sub-system descriptions, data gathering and software development which are necessary to obtain results that are meaningful in order to gain insight into the working of the safety system of air transport and to obtain results that allow the development of improvement strategies. The model will be delivered in the form of a computer assisted decision tool supported by reports on the underlying technology and data.