Session: 34-04 Computing architectures and solvers

Paper Number: 153512

Numerical Design of Experiments for Repeating Low-Pressure Turbine Stages Part II: Effect of Reynolds Number on Different Blade Geometries 

The current trend towards more compact and efficient low-pressure turbine (LPT) designs can substantially benefit from advanced numerical predictive tools. The complex transitional and turbulent nature of unsteady flows seen in LPTs often demands high-order methods such as Large Eddy Simulations (LES) for accurate predictions in turbine efficiency and loss generation. Integrating high-fidelity simulations into design cycles, predominately driven by rapid Unsteady Reynolds-Averaged Navier-Stokes (URANS) calculations, requires cutting-edge numerical tools able to leverage modern high-performance computing architectures. 
In Part 2 of this paper, we present results from a multi-fidelity simulation campaign that exploits current supercomputer hardware trends for repeating LPT stages under a wide range of engine-relevant Reynolds numbers and blade designs. Three geometries with different design intents are studied, operating at an isentropic exit Mach number of 0.3 and Reynolds numbers ranging from 70,000 to 320,000. Highly resolved LES are performed using the latest GPU compute nodes at three Reynolds numbers. The data-rich results enable detailed analysis of flow phenomena, loss mechanisms, and efficiencies of different blade designs over a wide range of Reynolds numbers. 
To maximize the power and efficiency of the computing architecture of GPU nodes, state-of-the-art URANS simulations run concurrently with the LES on the co-residing, idle CPUs of the computing nodes. This approach allows for a much finer Reynolds number sweep of the chosen range using different turbulence and transition models, and cross-comparisons between LES and URANS when applicable. While overall trend lines using URANS can be largely predicted, the detailed loss mechanisms cannot be adequately represented, especially in the low Reynolds number regime, highlighting the potential for further turbulence modelling efforts.
Ultimately, this paper combines the low-cost trend predictions of URANS with the accuracy of LES to reconstruct a multi-fidelity dataset spanning the entire Reynolds number regime of interest. This involves extracting trend lines from the URANS calculations which are then corrected by high-fidelity LES results. After validation with LES data not considered in the construction of the multi-fidelity dataset, the obtained solutions prove to be superior to URANS at all operating conditions. This effectively minimizes the number of costly LES required to provide a highly accurate and fine-grained parametric sweep that can be used in industrial design cycles.

Presenting Author: Marco Rosenzweig The University of Melbourne

Presenting Author Biography: Marco Rosenzweig is a third-year Ph.D. student at the University of Melbourne. In 2022, he obtained his M.Sc. degree in Aerospace Engineering at the Technical University of Munich and has two years of industrial work experience at MTU Aero Engines AG. Marco’s research expertise is in the turbomachinery flows critical to aircraft engines with a focus on low-pressure turbines. To date his research interests have included modeling of multi-stage component interactions, the evaluation of standard industrial design methods and scale-resolving, numerical methods. His most recent work focuses on performing multi-fidelity simulations on the latest supercomputing architectures.

Authors:

Marco Rosenzweig The University of Melbourne
Melissa Kozul The University of Melbourne
Richard Sandberg The University of Melbourne
Giovanni Giannini The University of Florence
Roberto Pacciani The University of Florence
Michele Marconcini The University of Florence
Andrea Arnone The University of Florence
Ennio Spano Avio Aero
Francesco Bertini Avio Aero