A study on one-dimensional model correction for axial-flow compressors based on measurement data

Abstract

In this study, a method for correcting a one-dimensional (1D) meanline model for axial-flow compressors using measured data and its effectiveness were described. The proposed method was evaluated for a 6-stage axial-flow compressor with variable guide vanes. In the compressor performance test, the rotating speed, the total pressure for each stage, and the total temperature at the inlet and outlet were measured. A compressor 1D meanline model was constructed using the empirical equations suggested in the existing research literature. Correction factors for the deviation angle and overall efficiency were applied to match the data measured in the performance test and the predicted values from the 1D model. Correction factors were calculated for each measurement point. The calculated correction factors were generated as a function according to the operating conditions using an artificial neural networks model. Moreover, a criterion for defining the compressor surge point in the 1D model was generated as a function of the diffusion factor and the relative Mach number of the inlet. By applying the generated functions to the existing compressor prediction model, the results at the operating point not used for model correction were compared. As a result, it was confirmed that a corrected 1D performance prediction model with high prediction accuracy can be obtained through the proposed method.