Accident propensity assessment of Boeing 737 Aircraft variants using the Gompertz distribution model

Ayodele Adekunle Faiyetole

DOI: https://doi.org/10.1007/s42401-023-00202-z

2023-03-10

Aerospace Systems

Abstract:Aircraft is a socio-technical system with some unavoidable accidents due to its complexity. Although Boeing 737 Aircraft was the best-selling commercial airliner, a relatively new variant, Max 8, suffered two rapidly successive mishaps, serving as a motivation for this study. The propensity of 737 to accidents considering some predictors, is, therefore, objectified. Specifically, it examined accidents involving 737 Aircraft variants from 1970 to 2021. Cox proportional hazards regression model and Weibull distribution model were considered but Gompertz distribution models, with the best goodness-of-fit for the data, were applied to estimate Aircraft survival functions and hazard ratios, with interactions per airline, causes of accidents, and fatality rates. The study showed that the survivability of 737–100 outperformed all the other series, while Max 8 had a high hazard ratio when interacting with the airline operator factor, i.e., management or ownership. Reparameterization of Aircraft–predictor interactions show different significance levels in that airline management or ownerships contribute a long way to accidents. The high cost of Aircraft maintenance is a strong determinant of accidents. The reparameterized results further show that pilot errors significantly contribute to accidents. This investigation suggests that although two Max 8 failed with a 100% fatality rate, 737 was generally safe. The study concludes that the trend toward intuitive and accurate components diagnostics beyond the prognostic and health management approach should be invigorated. In addition, despite the rigorous certification process embarked upon by the FAA, it is suggested that intuitive human reliability analysis regarding the human–machine team, such as flight crews and pilots, and the human failure events be further entrenched in quantifying failure events.