Most articles are a lot of hand-waving.
Here is the announcement from Oak Ridge National Lab with some detail.
https://www.olcf.ornl.gov/2026/01/27/frontier-provides-high-fidelity-insights-into-turbine-aerothermal-performance/
Researchers at the University of Melbourne in Australia collaborated with GE Aerospace. By the way, my first job after the service was at GE Aerospace.
Article at University of Melbourne.
https://eng.unimelb.edu.au/ingenium/engineers-keep-cool-on-jet-engines,-thanks-to-frontier-supercomputerengineers-keep-cool-on-jet-engines,-thanks-to-frontier-supercomputer
And they all reference a paper by ASME which you can have for a mere $25, UNLESS you know Usonian, who found the paper (or one almost the same) for free at the University of Melbourne.
https://people.eng.unimelb.edu.au/imarusic/publications/Edited%20Papers%202023/High-fidelity%20computational%20study%20of%20roughness%20effects%20on%20high%20pressure_Int%20J%20Heat%20Fluid%20Flow.pdf
High-fidelity computational study of roughness effects on high pressure turbine performance and heat transfer
Thomas O. Jelly ∗, Massimiliano Nardini, Marco Rosenzweig, John Leggett, Ivan Marusic, Richard D. Sandberg
Department of Mechanical Engineering, University of Melbourne, Victoria 3010, Australia
ARTICLE INFO
Keywords:
Roughness Heat transfer Turbulence
ABSTRACT
While blade surface roughness arising from in-service wear and/or the manufacturing process greatly affects aero-thermal performance, the detailed underlying physical mechanisms remain far from fully understood. In this study, a series of highly-resolved Large-Eddy Simulations of compressible flow past a high-pressure turbine vane with systematically varied levels of blade surface roughness have been performed, along with a smooth-blade simulation at matched flow conditions for comparison. Three non-dimensional roughness amplitudes have been investigated, namely, 𝑘s ∕𝑐 = {1.0, 2.0, 3.0} × 10−3 , where 𝑘𝑠 is an equivalent value of Nikuradse’s sandgrain roughness for an irregular, multi-scale near-Gaussian height distribution, and 𝑐 is the axial blade chord. All simulations have been performed at an axial chord Reynolds number of 0.59 × 106 and a Mach number of 0.9, based on the exit conditions of the reference smooth vane, and with synthetic inflow turbulence to mimic unsteady, three-dimensional disturbances from an upstream combustion chamber. The present investigation highlights the profound impact that blade surface roughness can have upon boundary- layer transition mechanisms, wall shear stress and blade surface heat flux, as well as the levels of turbulence kinetic energy and total pressure losses incurred in the wake. While blade surface roughness leads to major aero-thermal differences between the suction-side of the smooth and rough vanes, the pressure-side surface remains relatively unaffected — even for the largest roughness amplitude investigated here.
https://doi.org/10.1016/j.ijheatfluidflow.2023.109134
Received 31 January 2023; Received in revised form 6 March 2023; Accepted 14 March 2023
Available online 31 March 2023
0142-727X/© 2023 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license (
http://creativecommons.org/licenses/by/4.0/).
The paper is dated 2023, but seems to describe the study well.
and you can find out how well for $25.
Or not.
= new reply since forum marked as read