Validation of Steady Average-Passage and Mixing Plane CFD Approaches for the Performance Prediction of a Modern Gas Turbine Multistage Axial Compressor

This paper summarizes the results of a validation and calibration study for two modern Computational Fluid Dynamics programs that are capable of modeling multistage axial compressors in a multi-blade row environment. The validation test case is a modern 4-stage high pressure ratio axial compressor designed and tested by Honeywell Aerospace in the late 90’s. The two CFD programs employ two different techniques for simulating the steady three-dimensional viscous flow field in a multistage/multiblade row turbo-machine. The first code, APNASA, was developed by NASA Glenn Research Center “GRC” and applies the approach by Adamczyk [1] for solving the average-passage equations which is a time and passage-averaged version of the Reynolds Averaged Navier Stokes (RANS) equations. The second CFD code is commercially marketed by ANSYS-CFX and applies a much simpler approach, known as the mixing-plane model, for combining the relative and the stationary frames of reference in a single steady 3D viscous simulation. Results from the two CFD programs are compared against the tested compressor’s overall performance data and against measured flow profiles at the leading edge of the fourth stator. The paper also presents a turbulence modeling sensitivity study aimed at documenting the sensitivity of the prediction of the flow field of such compressors to use of different turbulence closures such as the standard K-ε model, the Wilcox K-ω model and the Shear-Stress-Transport K-ω/SST turbulence model. The paper also presents results that demonstrate the CFD prediction sensitivity to modeling the compressor’s hub leakages from the inner-banded stator cavities. Comparison to the test data, using the K-ε turbulence closure, show that APNASA provides better accuracy in predicting the absolute levels of the performance characteristics. The presented results also show that better predictions by CFX can be obtained using the K-ω and the SST turbulence models. Modeling of the hub leakage flow was found to have significant and more than expected impact on the compressor predicted overall performance. The authors recommend further validation and evaluation for the modeling of the hub leakage flow to ensure realistic predictions for turbo-machinery performance.