WMLES

The work from the workshops in 2022 and 2024 is continuing during monthly telecons.

Monthly videoconferences

Participants meet on the first Monday of every month at 10am US Eastern Time. The Zoom link for these calls is https://umd.zoom.us/j/4487374796. The best way to stay up to date is to (a) check this website regularly and (b) subscribe to the Google calendar.

Notes from telecons (continued from the old page):

• 02/02/2026
  • Decided to choose 1-2 test cases by April, and then have our first comparison of results at Scitech 2027.
  • Possible case #1: low-speed aerodynamics. Could be A-airfoil at lower AoA and possibly lower Re to make transition more difficult to predict. Could be airfoil with wavy leading edge and spanwise periodicity to get 3D effects, at multiple high AoAs. Interested: Ivan, Johan, Timofey, Prahladh, Christoph, Dan.
  • Possible case #2: SBLI. Could be a ramject (ONERA DNS), CUBRC M=6 flared cone (possibly without modeling variable gamma etc). Interested: Christoph, Ivan, Johan.
• 01/12/2026 (at Scitech)
  • A few of us met at Scitech and discussed the working group going forward. The plan is to meet virtually in February to have a broader discussion — date and time to be announced (it will be moved from the usual first Monday of the month slot). The BEST way to stay updated is to subscribe to the Google calendar, with the link available above!
• 12/01/2025
  • Agreed to have 1 high-speed case (simplified ramjet?) and 1 airfoil case (with some type of 3D feature, like viscous side-walls, wing tip, spanwise-varying chord, single slat bracket, or similar?).
• 11/03/2025
  • Decided to have next workshop at Scitech 2027.
  • Decided to have two test cases, one low-speed airfoil and one high-speed case with SBLI.
  • There’s a desire to add some minor 3D effects to our test cases, to take a small step towards real applications (while still remaining more academic than the high-lift and drag prediction workshops). This has to be balanced against the desire to enable excessive grid refinements.
  • There’s also a desire to run at multiple flow conditions, for example, at predicting a lift curve near stall rather than just a single condition.
• 10/06/2025
  • Xiang, Daluar and Timofey had prepared summaries of the Beverli hill, juncture flow, and 30P/30N cases. Some comments during the dicsussions:
    • Beverli hill has a thick incoming BL and highly curved edges, so the separation location is not that hard to predict, meaning the main challenge is the post-separation flow which LES should be good at. On the other hand, it could be challenging and interesting to see if WMLES produces asymmetric solutions for the symmetric geometries.
    • Juncture flow requires larger grid sizes compared to the other cases, and is more difficult to mesh. It provides an opportunity to work on leading edge and transition problems. It has large regions of attached BLs that increase the cost without offering much room for interesting tests.
    • 30P/30N has been done at a large range of Reynolds numbers. Would offer simpler geometry than the juncture flow and be a logical next step from the A-airfoil. Would require transition sensors etc.
  • We deferred any decision until the November meeting, hoping that most people can attend then.
• 09/08/2025 (moved one week due to US Labor Day):
  • Discussed the need for a new test case that captures everyone’s interests. Assigned as homework for different participants to read up on and present (next time) the following cases: Beverli hill (Xiang), 30P/30N (Timofey), Juncture flow (Johan), VR12 rotor (Christoph), and CRM-WB (Xiang).
• 08/04/2025 — cancelled due to vacations
• 07/07/2025:
  • Discussed the intricacies of how to implement the BCs for WMLES in explicit/implicit codes.
• 06/02/2025:
  • Stefan Domino described his group’s work on fire and isothermal simulations of cylinders in cross flow near the ground. They successfully predicted the drag crisis (along with demonstrated mesh convergence) when using a dynamic LES model with an equilibrium wall-model and off-wall exchange-location. Static models and local velocity sampling did not demonstrate mesh convergence, with predictivity of C_D and C_p suffering.
  • Discussed the possibility of looking at stall cells on airfoils in the future. This could be on aerospace-relevant airfoils or wind turbines. Stefan Domino pointed to two relevant studies for the latter: https://pubs.aip.org/aip/pof/article/33/8/085106/1080219/Investigation-of-the-three-dimensional-flow-past-a and https://pubs.aip.org/aip/pof/article-abstract/26/4/045101/314505/Study-of-a-stall-cell-using-stereo-particle-image?redirectedFrom=fulltext.
• 05/05/2025 — cancelled (sorry about cancelling at the last minute — please subscribe to the Google calendar using the link above to get these notifications).
• 04/07/2025:
  • Daluar Hussain presented his ongoing numerical experiments with different leading edge treatments for the A-airfoil.
• 03/03/2025:
  • Prakyash described his PhD work on using WMLES for noise predictions from wind turbine blades.
  • Spalart presented his views on the state of WMLES and how it lags that of DES in terms of maturity.
  • Advertised the NASA High Fidelity CFD workshop on May 20-22, 2025.
• 02/03/2025:
  • Discussed what test cases to focus on going forward. 4 groups are still working on the A-airfoil to some degree, either actively running cases or using it occasionally to test things. Only 1 group is running simulations at this time of the smooth ramp.
  • Some ideas for cases going forward: the Beverly Hill from Virginia Tech (3D separation), high-Mach channels, channels with either spanwise acceleration or wall motion, or infinite swept bump or swept separation bubble. Participants who start taking an interest in these or other cases are encouraged to present their findings/opinions to the group so we can eventually identify a case that multiple groups are interested in.
• 01/08/2025 (in person at Scitech):
  • Decided to move forward in the working group format, and the larger High Fidelity CFD Verification group decided to do the same. We will not do weekend workshops going forward, but instead create a special session every other year or so at Scitech in order to present our findings to the world.
• 12/02/2024:
  • Daluar showed some new results for the A-airfoil with a coarser grid and two different wall-normal resolutions in the leading edge region. With a coarser resolution the “transition point” is moved back quite a bit, but with a refined wall-normal resolution the “transition point” is captured well. Ongoing work is to analyze the solution around the leading edge more carefully to understand the results.
  • Dan Garmann shared preliminary updated smooth ramp results, suggesting a large influence on the size of the separation bubble from the exact nature of the boundary conditions, specifically whether non-reflecting or not.
• 11/04/2024:
  • Daluar Hussain has preliminary results with a transition sensor for the A-airfoil, seems to get transition at the right location. The hypothesis is that the nature of the flow (strong FPG first, then APG) forces this.
  • Siavash Toosi remarked that the cases with erroneously too thick boundary layer (due to errors at the leading edge) have a boundary layer with a larger-than-otherwise Rotta-Clauser parameter in the remainder of the airfoil, since the thickness increases while the wall stress decreases. His point being that this would make the boundary layer grow even more quickly and thus never recover from the initial error.
• 10/07/2024:
  • No updates.
• 09/03/2024 (moved to Tuesday due to Labor Day):
  • Timofey is experimenting with a transition sensor for the A-airfoil, seem to be getting transition in the right place. Will have more complete results next month.
  • Matt and Dan are also experimenting with transition for the A-airfoil, and are also re-running the smooth ramp case with different subgrid modeling treatment.
  • Johan has added his and Ivan’s smooth ramp results to the website.
• 08/05/2024 (cancelled due to vacations)
• 07/01/2024:
  • Discussed the A-airfoil. Johan showed additional numerical experiments aimed at the laminar region, with none really solving the problem. Dan showed numerical experiments where they tried different exchange location heights, and found different behaviors in the laminar region but low sensitivity in the turbulent region.
  • AFRL will re-visit the smooth ramp case with their new wall-model implementation and will also try different exchange locations for that case.
  • It was suggested that people try the A-airfoil with the WMLES grid but without the wall-model just to see what that looks like.
• 06/03/2024: (cancelled due to scheduling conflicts)
• 05/06/2024:
  • Johan presented numerical experiments for the A-airfoil where the wall shear stress was computed from Xfoil for the first 10% of the chord; this produced much better agreement with the reference WRLES.
  • Bala presented his ongoing simulations of the A-airfoil; he will start using the same grid as the UMD group and try with that.
  • David Flad presented his experiments on the smooth ramp problem where they get strange results when running with numerical dissipation but much better results without it.
  • David and Dan Garmann have both observed that their codes can produce fully laminar solutions when using the first grid point for the wall-model input but turbulent solutions when using the second or higher.
• 04/01/2024:
  • Timofey Mukha described his airfoil simulations and his a priori study of wall-model accuracy. An important point of the latter was how, for this particular case, choosing \( h_{\rm wm}(x) \approx 0.1 \delta_{99}(x) \) results in very good a priori accuracy due to luck.
  • Johan presented some numerical experiments of the smooth ramp problem where he increased the WENO fraction in the incoming boundary layer to produce different boundary layer states. Adding 1% WENO did not change much, but adding 4% WENO drastically reduced the separation bubble size. Not clear how to interpret these findings, though.
• 03/04/2024:
  • We agreed to slightly change the format going forward. The monthly telecons will continue (at noon US Eastern time on the first Monday of every month), but we will try to make these discussions more technical. The results comparisons will be updated on an ongoing basis rather than only once every two years. We will then meet formally or informally at every AIAA conference (Scitech and Aviation) for in-person discussions. In other words, we will function like an AIAA Working Group, which is essentially how this all started out.
  • ZJ Wang, Christoph Brehm, Dan Garmann and Timofey Mukha will re-run the airfoil case with identical wall-modeling exchange location (likely \(h_{\rm wm}/c = 0.001 \)) and without any special handling of the laminar/transitional region. The idea is to compare the different codes with identical, even if imperfect, settings.
  • Dan Garmann is taking the lead in putting together a paper for Aviation 2024 summarizing the January workshop. This paper won’t include updated results, but instead will provide a snapshot of the blind predictions at the workshop.