Generating Flight Operations Quality Assurance (foqa) data from the X-Plane Simulation

Analysis of Flight Operations Quality Assurance (FOQA) data from millions of flights from at least 10 U.S. airlines has enabled researchers, using data mining and machine learning algorithms, to identify abnormal scenarios that maybe precursors to accidents and incidents. Innovation of these algorithms is hindered by restricted access to the FOQA data to comply with confidentiality, proprietary and security policies. This paper describes a method to generate larges sets of FOQA-like data for development and testing of machine learning and data mining algorithms. This data is not restricted and can be made publicly available. The data is generated using a C++ plug-in to the X-Plane Simulation. The plug-in manipulates the simulation set-up configuration, the pilot commands, and coordinates a Monte Carlo shell to run multiple runs. An example for generating data for CAT III ILS approach for KSFO Runway 28L and 28R is described. Implications of this capability, limitations and future work are discussed.