Feature extraction method for fuel flow estimation

A fuel flow feature extraction method is proposed for the aero-engine condition monitoring. The method uses the fuel volume data recorded by Flight Data Recorder as input, and gets rid of the outliers and wave points in the fuel volume data by data cleaning. Then, a novel interpolation method referred to as slope constraint quadratic spline Hermite interpolation is proposed to transform discrete fuel volume data into continuous fuel volume estimation curve. Finally, the fuel flow estimation value is obtained by derivating fuel volume estimation curve with respect to time. A comparison experiment is conducted to assess the performance of fuel flow feature extraction method. The experimental result shows that the relative prediction error between measurement value and estimation value of the fuel flow is small. So, the method can be used in aero-engine condition monitoring.