An Assessment of Algebraic and Differential Reynolds-Stress Models for a Highly-Bent Serpentine Aircraft Intake

Serpentine ducts are commonly used as intakes for modern low-bypass aero-engines, drawing atmospheric air through to the aero-engine compressor. Because of their highly-bent geometry, serpentine ducts develop significant unsteady flow features that can, in turn, cause compressor instabilities. Quantifying the performance of a coupled intake-compressor system requires analysis of a large parameter space, which is typically undertaken with (U)RANS calculations based on linear eddy-viscosity models. In the present work, we study a highly-bent serpentine intake with three classes of turbulence model, and assess the intake as a turbulence modelling benchmark. A comparison is made between a linear eddy-viscosity model (k-ω SST with curvature correction), an explicit algebraic Reynolds-stress model (Wallin & Johansson, 2000 with curvature correction), and a differential Reynolds-stress model (the elliptic blending model of Manceau & Hanjalić, 2002). Comparisons are made between experimental wall pressure measurements and URANS calculations along the centreline, and radial cross-sections of the intake. Detailed analysis is undertaken in critical near-wall regions of the intake. Finally, based on the outcomes of the flow physics analysis, future modelling opportunities are discussed for algebraic and differential Reynolds-stress models.