Application of Non-Parametric Statistical Methods to Reliability Database Analysis

An extensive reliability and maintenance data base has been compiled for a gas turbine engine used for propulsion of a fleet of military aircraft, covering a full eight years of operational service. Non-parametric statistical tools were applied to analyze the engine flight line removal data set contained in this data base. This analysis was directed towards eliciting information useful in managing maintenance and design change activity directed to enhancing aircraft reliability and availability.

Intensive preventive maintenance, high levels of “censoring” due to the phased buildup of the fleet of aircraft and the modular maintenance policy applied to this military engine introduced significant conceptual and practical challenges that required innovations in data treatment and analysis. The resultant analysis toolset is defined and some results described.

The results show the tools' utility and suitability for the purpose intended, the visualization and quantification of the hazard function for classes of unplanned engine removals in the context of intensive preventive maintenance. These non-parametric approaches can assist in forecasting future removal rates and causes as a function of the usage of engines scheduled for inspection, and assess the effectiveness of preventive maintenance.