Application of Extended Cox Regression Model to Time-On-Wing Data of Aircraft Repairables

Abstract

Global and local aviation traffic is growing while economic and performance pressures on the industry are increasing. As a consequence, airlines try to maximise their fleet utilization. Airline operators and Maintenance, Repair and Overhaul (MRO) providers therefore require as much insight as possible in factors affecting component reliability and availability. Reliability analysis in literature rarely considers the existence of relations between explanatory variables and time-based component reliability, and includes strict assumptions on independence of events and underlying distributions. This disregards the complex nature of aircraft operations, where the probability of an event may be influenced by various operational and maintenance factors. This paper develops new insights from operational and maintenance data about the impact of operating environment and ageing of components and fleet on reliability of the components by incorporating these factors in an extension of the Cox regression model. Examination of results obtained from analysing historical data of a set of three components with respect to installations and removals indicate that the natural environment at the hub airport, maintenance history of components, the age of the aircraft on which the component is installed and different modification designs are useful significant predictors of the time-on-wing duration of the component.