A framework development to predict remaining useful life of a gas turbine mechanical component

Power-by-the-hour is a performance based offering for delivering outstanding service to operators of civil aviation aircraft. Operators need to guarantee to minimise downtime, reduce service cost and ensure value for money which requires an innovative advanced technology for predictive maintenance. Predictability, availability and reliability of the engine offers better service for operators, and the need to estimate the expected component failure prior to failure occurrence requires a proactive approach to predict the remaining useful life of components within an assembly. This research offers a framework for component remaining useful life prediction using assembly level data. The thesis presents a critical analysis on literature identifying the Weibull method, statistical technique and data-driven methodology relating to remaining useful life prediction, which are used in this research. The AS-IS practice captures relevant information based on the investigation conducted in the aerospace industry. The analysis of maintenance cycles relates to the examination of high-level events for engine availability, whereby more communications with industry showcase a through-life performance timeline visualisation. Overhaul sequence and activities are presented to gain insights of the timeline visualisation. The thesis covers the framework development and application to gas turbine single stage assembly, repair and replacement of components in single stage assembly, and multiple stage assembly. The framework is demonstrated in aerospace engines and power generation engines. The framework developed enables and supports domain experts to quickly respond to, and prepare for maintenance and on-time delivery of spare parts. The results of the framework show the probability of failure based on a pair of error values using the corresponding Scale and Shape parameters. The probability of failure is transformed into the remaining useful life depicting a typical Weibull distribution. The resulting Weibull curves developed with three scenarios of the case shows there are components renewals, therefore, the remaining useful life of the components are established. The framework is validated and verified through a case study with three scenarios and also through expert judgement.