Failure Interval Probabilistic Analysis for Risk-based Decisions - Concorde Crash Example

Abstract

The DC-6, DC-8, DC-10, Concorde, Boeing 787 and Boeing 737 MAX fatal crashes and nearmisses were analyzed with event interval probabilistic analysis methods. Fleet grounding decisions are the epitome of risk-based decisions, and the most important decision is the first opportunity to ground. The “first opportunity to ground” decision is retrospectively judged to be wrong if, in the immediate future, another accident or cause-and-effect findings leads to the original decision being reversed. Using only data available at the time of the significant events, the analysis examines these risk-based decisions as if it they were made at the event’s instant in time.

The event interval method identified five out of six “first opportunity to ground” decisions correctly, including the Concorde. According to these analyses, the FAA and its predecessor organizations made one correct decision out of five. Use of this method based on statistics and probability would have avoided 503 actual fatalities, plus 9.45 expected value fatalities from additional risk exposure due to flying statistically proven unreliable aircraft. In addition to the reversed decision standard for judging whether these decisions were wrong, the data show that a grounding of the DC-8 and a second grounding of the DC-6 would have been statistically appropriate — but these groundings did not occur.

A specific objective of this paper is to lead the FAA and aircraft manufacturers to using event interval probabilistic analysis in grounding decisions and air-worthiness certification. The cause-and-effect data necessary to identify issues and make corrections are often sparse or nonexistent at the time of the event. Cause-and-effect data can take days or months to acquire and analyze, but event interval timing data are simple because system performance data are available at the instant the event occurs.