Development and Validation of a Comprehensive Hybrid Causal Model forSafety Assessment and Management of Aviation Systems

摘要

The United States Federal Aviation Administration (FAA) has initiated the development of a causalrnrisk model of commercial air transport in support of the System Approach for Safety Oversight (SASO)rnprogram. The model uses the so-called Hybrid Causal Logic (HCL) methodology which combinesrnEvent Sequence Diagrams (ESD), fault trees (FT) and Bayesian Belief Networks (BBN). The model isrnhierarchically structured: it includes 31 generic ESDs covering accident scenarios in various phases ofrnflight at the top-level of the model * , supported by numerous fault trees and BBNs to represent deeperrncauses of the ESD events. BBNs are used to extend the causal chain of events to potential human andrnorganizational roots. Probabilities of the HCL models are obtained from extensive review andrnclassification of commercial aviation accident/incident databases. A number of FAA operationsrnresearch analysts, principal inspectors, and other experts actively participated in review and validationrnof the model development. A dedicated software prototype (IRIS) has been developed thatrnsimultaneously supports model development and model application.