A new performance adaptation method for aero gas turbine engines based on large amounts of measured data

Multiple unexpected uncertainty factors can occur when measuring gas turbine engine data, and the quality of the measured data can directly affect the accuracy of gas turbine engine models during performance adaptation. In the present study, a new performance adaptation method for aero gas turbine engines is proposed to improve prediction accuracy, by effectively processing a large amount of measured data. Adaptation factors were obtained to match the engine model and the measured data of every single operating point. These adaptation factors were then used to adjust the compressor performance, bleed air flow, engine thrust, and exhaust gas temperature. A data clustering technique was employed to exclude physically non-reasonable data points from the time series adaptation factors. The correlations for the adaptation factors were generated by using selected centroids from the clustered data, then the correlations were applied to the engine simulation. As a result, the values in the adapted engine model were in good agreement with transient measurement data. This confirms that the proposed performance adaptation method can be used to generate accurate gas turbine engine models using time series measurement data.