Development of an aircraft performance risk assessment model

Abstract

This paper presents the development of a risk assessment model for aircraft performance. The model consists of an automated set of tools that makes use of data from flight data recorders (FDRs) and quick access recorders (QARs) to evaluate performance data from routine flights in order to identify precursor events that indicate a risk of incidents and accidents. This model takes a hybrid approach to modeling that allows for the combination of different modeling techniques. It combines the use of exceeding information from flight recorder data with information about contextual factors that are not directly available from flight data recorders. Expert opinion is incorporated into the model through the use of knowledge-based rules. The model provides a framework that addresses consequence, probability of occurrence, and severity, similar to that used for aircraft certification.