Turbofan Condition Monitoring using Evolutionary Algorithm based Gas Path Analysis

Abstract

In this thesis, a hybrid Gas Path Analysis (GPA) tool is developed for next-generation turbofan engine condition monitoring purposes at KLM Engine Services. The main drawback of these new engines is that fewer gas path sensors are installed. However, this is compensated by a greater quantity in-flight data, including information on bleed valves, active clearance control systems and variable geometry positions. With the availability of this data optimal near steady-state operating points can be selected and a Multiple Operating Point Analysis (MOPA) can be implemented. Then, an Evolutionary Algorithm (EA) optimization approach is combined with the non-linear GPA program GSP in order to predict engine component health parameter deviations. Using this method it is possible to track fan, LPC, HPC, HPT and LPT deterioration. The tool has been verified with simulated data and validated using on-wing data from the General Electric GEnx-1B engine.