Prediction of the unsteady turbulent flow in an axial compressor stage. Part 1: Comparison of unsteady RANS and LES with experiments

Abstract

a b s t r a c tA better understanding of turbulent unsteady flows is a necessary step towards a breakthrough in the design of modern gas turbine components. With the increase in computing power, LES emerges as a promising method to improve both knowledge of complex physics and reliability of flow solver predictions. However, there is still a lack of evidences in the literature that LES is applicable for turbomachinery at conditions relevant to industrial applications. In that context, the objective of the present work is to investigate the capability of LES to predict the turbulent flow in a stage of an axial compressor and compare the results with unsteady RANS data and experiments. The compressor operates at industrial relevant conditions, with Mach and Reynolds numbers equal to M ¼ 0:5 and Re ¼ 7 Â 10 5 , respectively. This paper presents the numerical method and a comparison of URANS and LES results to experimental data. A particular care is brought to estimate the results sensitivity to grid refinement (LES) and to turbulence and transition modelling (URANS). The comparison to experiments shows that LES better predicts time-dependent quantities than URANS, especially close to the casing. However, both URANS and LES fail to accurately estimate the compressor performance (efficiency and pressure ratio).

Large Eddy Simulation of Pulsed Blowing in a Supersonic Compressor Cascade

“…1. The mesh is refined near the walls with a standard of y + number to be less than 1 and the streamwise and the pitchwise mesh nodes are also controlled to be fine thus the x + number meets the requirements of LES (Gourdain, 2015). Meanwhile the mesh grid size in the streamwise is refined to ensure a better capture of the SWBLI features.…”

“…The energetic larger scale motions are resolved directly and only smaller scale fluctuations are modeled under LES calculations. Thus the LES results could reserve more details of flow field fluctuations compared with unsteady Reynolds averaged Navier-Stokes equations (URANS) (Gourdain, 2015).…”

Investigation of Speed Matching Affecting Contrarotating Fan’s Performance Using Wireless Sensor Network including Big Data and Numerical Simulation

“…No-slip and adiabatic conditions are applied to all wall surfaces; the total temperature and total pressure of standard atmosphere as well as axial inflow are specified as the inlet boundary condition, while the static pressure is specified at the outlet assuming radial equilibrium. The stage method [30] is used to process the interface.…”

Extended integral wall-model for large-eddy simulations of compressible wall-bounded turbulent flows

“…In the frame of this work, the aim is to develop a wallmodel capable of taking into account the physical phenomena encountered in turbomachinery systems. Usually, in turbomachinery flows, Mach numbers are up to 1.5, [39][40][41][42][43] Reynolds numbers 44,45 between 10 5 and 10 7 , and temperature gradients in the order of ten to hundred Kelvin. 46 As pointed out by Tyacke and Tucker,47 if the flow in low-pressure turbines can be simulated with LES due to their relatively low-Reynolds number, the flows in the other components require a wallmodeling in order to reduce computational requirements at high Reynolds numbers.…”