Testing Winmine Analysis with ATC Operational Error Causal Factors Data

Abstract

When less than applicable separation minima occur between two or more aircraft or between an aircraft and an obstacle, an operational error (OE) is reported and relevant information is recorded. Most analyses of these data have been descriptive. Five reported categories of OE causal factors (Data Posting, Radar Display, Communications Error, Coordination, and Position Relief Briefing) from 5,487 en route OEs reported between CY2000–2003 were analyzed using WinMine, a tool for probabilistic analysis. Communications Error was reported in 34.5% of the cases and a Radar Display factor in 65.2% of the cases. A dependency network model was constructed to show the direction and magnitude of probability between the five factors. Radar Display and Communications Error were the most strongly linked factors in this sample, suggesting that the relationship between these two factors deserves further examination. This analytic method demonstrates that probabilistic analysis of OE data may help to better understand complex relationships between factors.