Zonal Detached-Eddy Simulation Applied to the Tip-Clearance Flow in an Axial Compressor

Abstract

This paper presents the study of the tip-clearance flow in an axial compressor using several turbulence-modeling approaches: Reynolds-averaged Navier–Stokes, unsteady Reynolds-averaged Navier–Stokes, and zonal detached-eddy simulation. Rotor simulations are carried out on the same mesh based on the zonal-detached-eddy-simulation requirements, and compared to experimental data. Zonal detached-eddy simulation brings significant improvements over Reynolds-averaged Navier–Stokes and unsteady Reynolds-averaged Navier–Stokes approaches to reach accurate integral values, especially the total-pressure-gradient prediction near the casing, as well as to capture complex flow patterns near the casing. Whereas the unsteady Reynolds-averaged Navier–Stokes method predicts only tip-leakage and induced vortices, the zonal-detached-eddy-simulation approach is able to reveal more complex phenomena as the tip-leakage vortex, the contra-rotative induced vortex, and numerous secondary vortices are captured. Because of their interaction with the tip-leakage vortex, secondary vortices roll up around it. This leads to the generation of smaller vortices. The interaction between the tip-leakage vortices and the incoming stator wake, the latter being made up of the blade wake itself along with a tip vortex and a hub vortex, is evaluated, and a tip-leakage-vortex-flutter phenomenon is detailed. A spectral analysis reveals a major difference between unsteady Reynolds-averaged Navier–Stokes and zonal detached-eddy simulation concerning the tip-leakage-vortex disruption. The unsteady Reynolds-averaged Navier–Stokes approach dissipates the tip-leakage vortex across the weak shock present at the blade tip, and thus, captures only the first three harmonics of the blade-passing frequency. On the contrary, with the zonal-detached-eddy-simulation method, the first 10 harmonics are well captured, and the tip-leakage vortex further develops beyond the shock leading to a better prediction of velocity and pressure fluctuations.