Low-Pressure Compressor Near-Stall Predictions Using Unsteady CFD Methods

Abstract

Abstract Surge margin is critical for the safety of aeronautical compressors, hence predicting it early in the design process using CFD is mandatory. However, close to surge, steady-state Reynolds Averaged Navier-Stokes (RANS) simulations are proven inadequate. Unsteady techniques such as Unsteady RANS (URANS) and Large Eddy Simulation (LES) can provide more reliable predictions. Nevertheless, the accuracy of such methods are dependent on the method used to handle the rotor/stator interfaces. The most precise method, the sliding mesh, requires simulating the full annulus or a periodic sector, which can be very costly. Other techniques to reduce the domain exist, such as the phase-lagged approach or geometric blade scaling, but introduce restrictive assumptions on the flow at near-stall conditions. The objective of this paper is to investigate the near-stall flow of a low-pressure compressor using unsteady methods of varying fidelity: URANS with the phase lag assumption, URANS on a periodic sector and a high-fidelity LES on a smaller periodic sector achieved using geometric blade scaling. Results are compared to experimental measurements. An overall good agreement is found. Results show that the tip leakage vortex is not the origin of the stall on the studied configuration and a hub corner separation is initiated. LES further validates the (U)RANS flow predictions and brings additional insight on unsteady flow separations.