Automated Calibration of Compressor Loss and Deviation Correlations

Abstract

Throughout preliminary design of multistage axial compressors performance prediction is based on fast 2D throughflow codes incorporating models for loss, deviation and secondary flow effects. To achieve high solution quality these models have to be calibrated on experimental or simulation data. Particularly 2D cascade loss and deviation correlations are of major importance. Correlations available in literature are mostly based on data from early cascade wind tunnel tests of NACA-65 and DCA profile families. In general, calibrations for state-of-the-art airfoil shapes are not published. This paper describes a methodology to re-calibrate parameters of well-known empirical loss and deviation correlations to CFD simulations of compressor cascades. Correlations describing choke incidence, on- and off-design loss and deviation are presented, resolving dependencies on inlet flow angle, Mach number, Reynolds number and MVDR. Based on a set of loss and deviation polars, nonlinear least-squares problems are solved to determine the empirical parameters. By including shock losses and choking incidence, the model is suited for subsonic as well as supersonic inlet conditions. Results of the calibration process are shown for two state-of-the-art compressor cascades. The overall model is evaluated on a sub- and a transonic compressor as well as on a transonic fan comparing throughflow results with 3D CFD and experimental data. A high level of automation makes the methodology applicable to arbitrary profile families or in combination with programs for parametric airfoil design.