Master thesis and internship[BR]- Master's thesis : Turbulence modeling in a linear low-pressure compressor cascade configuration[BR]- Integration internship

Abstract

A wide range of turbulence models were tested following the standard CFD practices of Safran Aero Boosters in a linear low-pressure compressor cascade. The models evaluated include k-ε Launder-Sharma, k-ω, BSL, Spalart-Allmaras, k-l and k-kl models. The effects of a vorticity-based production term, QCR correction, SST correction, rotation correction, a production limiter and the Menter-Langtry transition model were also examined. Two inlet Mach numbers, 0.45 and 0.6, were considered, with incidence angles ranging from 0° to 12°. Only the results for the Mach 0.6 configuration are discussed, as the Mach 0.45 cases led to similar conclusions. Model incorporating production or viscosity limiters (KW06, Menter-Langtry models, BSL with SST correction and production limiter) predicted thinner upstream endwall boundary layer, stronger cross-flow and larger corner separation structure. These results deviate notably from experimental results. In contrast, two-equation models such as BSL, KL and KKL offered better agreement at low and mid incidences, though none accurately captured high-incidence flow. Major discrepancies in the leading edge region between models were observed. Foremostly turbulence production in the stagnation region which was found to critically influence midspan losses. A vorticity-based production term reduced these losses but slightly increased corner separation. The SA-QCR model showed promising improvements in capturing separation and loss predictions. Finally, models using vorticity-based formulations or SST corrections experienced notable convergence issues, limiting their practical applicability.