Three-spool turbofan pass-off test data analysis using an optimization-based diagnostic technique

Production engine pass-off testing is a compulsory technique adopted to ensure that each engine meets the required performance criteria before entering into service. Gas turbine performance analysis greatly supports this process and substantial economic benefits can be achieved if an effective and efficient analysis is attained. This paper presents the use of an integrated method to enable engine health assessment using real pass-off test data of production engines obtained over a year. The proposed method is based on a well-established diagnostic technique enhanced for a highly-complex problem of a three-spool turbofan engine. It makes use of a modified optimization algorithm for the evaluation of the overall engine performance in the presence of component degradation, as well as, sensor noise and bias. The developed method is validated using simulated data extracted from a representative adapted engine performance model. The results demonstrate that the method is successful for 82% of the fault scenarios considered. Next, the pass-off test data are analyzed in two stages. Initially, correlation and trend analyses are conducted using the available measurements to obtain diagnostic information from the raw data. Subsequently, the method is utilized to predict the condition of 264 production turbofan engines undergoing a compulsory pass-off test