search

su2code / SU2

SU2: An Open-Source Suite for Multiphysics Simulation and Design
https://su2code.github.io
Other
1.3k stars 833 forks source link

Add wall functions #399

Closed petebachant closed 4 years ago

petebachant commented 7 years ago

I saw on the projects tab that there is a wall functions implementation 80% complete, but stalled. I was looking to help out a little on this project. What's left to be done?

talbring commented 7 years ago

Hi Pete,

we appreciate your help on that. I just updated the branch feature_Wall_Functions to be up-to-date with develop. The main routines dealing with the wall functions are in solver_direct_mean.cpp and solver_direct_turbulent.cpp. Maybe @economon can tell you what exactly is missing there.

Tim

juanjosealonso commented 7 years ago

Pete, Tim,

I have had a long standing interest in getting this working and would be happy to help. The (partial) implementation that is there is already 3-4 years old, so it will have to be updated, but we can discuss some of the details of the implementation and how it can be done both more efficiently and for both compressible formulations and adaptive wall functions.

We can start a thread separately (jjalonso@stanford.edumailto:jjalonso@stanford.edu) and perhaps have a telecon to kick things off. My availability for this improves after the next 2 weeks.

Best,

Juan

On Jun 7, 2017, at 1:19 PM, Tim Albring notifications@github.com<mailto:notifications@github.com> wrote:

Hi Pete,

we appreciate your help on that. I just updated the branch feature_Wall_Functionshttps://github.com/su2code/SU2/tree/feature_Wall_Functions to be up-to-date with develop. The main routines dealing with the wall functions are in solver_direct_mean.cpphttps://github.com/su2code/SU2/blob/feature_Wall_Functions/SU2_CFD/src/solver_direct_mean.cpp#L16612 and solver_direct_turbulent.cpphttps://github.com/su2code/SU2/blob/feature_Wall_Functions/SU2_CFD/src/solver_direct_turbulent.cpp#L2569. Maybe @economonhttps://github.com/economon can tell you what exactly is missing there.

Tim

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/su2code/SU2/issues/399#issuecomment-306912607, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ADpSxOvHCOSUFKDoizjyvrHbgyXgZ-vVks5sBwW8gaJpZM4NvG6w.

petebachant commented 7 years ago

We might as well keep the discussion going up here on GitHub in case anyone else wants to jump in. @talbring, @economon any information on what's left to be done would be very helpful!

EduardoMolina commented 7 years ago

Hi all,

I am also very interested in this topic. Please let me know.

Best

Eduardo

On 7 Jun 2017, at 18:59, Pete Bachant notifications@github.com wrote:

We might as well keep the discussion going up here on GitHub in case anyone else wants to jump in. @talbring https://github.com/talbring, @economon https://github.com/economon any information on what's left to be done would be very helpful!

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-306967818, or mute the thread https://github.com/notifications/unsubscribe-auth/AJVmCWNF7ouj2f2rIchhEKj1mqK4A3iVks5sB0eKgaJpZM4NvG6w.

economon commented 7 years ago

Hi all,

Yes, it would be wonderful to get this feature completed. I had started this a couple of years ago, but never quite got around to finishing (too many fun things to work on!). I was following this paper, which should be a straightforward implementation for us, as FUN3D has a very similar structure:

Jan-Renee Carlson, Veer Vatsa, and Jeffery A. White. "Node-Centered Wall Function Models for the Unstructured FLow Code Fun3D", 22nd AIAA Computational Fluid Dynamics Conference, AIAA AVIATION Forum, AIAA 2015-2758, https://doi.org/10.2514/6.2015-2758

The key changes can be found in the following locations:

Let's keep the discussion going, Tom

juanjosealonso commented 7 years ago

Thanks Tom.

And, as mentioned before, let’s have a bit of a discussion as to which wall functions formulations would make most sense to implement. I am pretty sure that we can create a generic implementation that can then be used for standard wall functions, but also for adaptive wall functions (see Medic, G., Kalitzin, G., et al., “Adaptive Wall Functions with Applications”, 2006), and even compressible/incompressible formulations.

The adaptive wall function trick is quite useful as it makes the wall function implementation more robust to the varying y+ of the first cell of the surface of the geometry and lessens the burden during the grid generation stage. Formulations for S-A and k-w/SST are available.

If anyone else is interested in participating, please let us know.

Best,

Juan

On Jun 8, 2017, at 5:11 PM, Thomas D. Economon notifications@github.com<mailto:notifications@github.com> wrote:

Hi all,

Yes, it would be wonderful to get this feature completed. I had started this a couple of years ago, but never quite got around to finishing (too many fun things to work on!). I was following this paper, which should be a straightforward implementation for us, as FUN3D has a very similar structure:

Jan-Renee Carlson, Veer Vatsa, and Jeffery A. White. "Node-Centered Wall Function Models for the Unstructured FLow Code Fun3D", 22nd AIAA Computational Fluid Dynamics Conference, AIAA AVIATION Forum, AIAA 2015-2758, https://doi.org/10.2514/6.2015-2758

The key changes can be found in the following locations:

Let's keep the discussion going, Tom

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/su2code/SU2/issues/399#issuecomment-307260814, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ADpSxEMtW6EqpjBFJOTo8mRqxMFvvf4uks5sCI2zgaJpZM4NvG6w.

vdweide commented 7 years ago

Juan, all,

The adaptive wall function formulation as described in the paper of Medic et al. has two serious issues. First, the convergence stalls after three orders of magnitude or so and second it only works with an upwind discretization of the inviscid fluxes. For some reason JST type of schemes gave rather poor results. In order to have a reliable implementation, both issues need to be addressed. Unfortunately, we never did.

Edwin

erangit commented 7 years ago

Hi all

May I just add to Edwin comment: some of you have seen the tests that I made with SUmb (structured mesh code) using the adaptive WF: Results using the JST scheme were definitely off, while upwind +WF converged reasonably to wall-resolved results. Naturally this conclusion is strongly coupled with the specific discretization and implementation so it might not apply to a SU2 implementation. I suggest to keep working on the Carlson formulation (due to FUN3D-SU2 similarities mentioned above) and hopefully extend it to adaptive WF as a second step. I'll be happy to be on board of his effort. Best regards, Eran

On Fri, Jun 9, 2017 at 7:49 AM, Edwin van der Weide < notifications@github.com> wrote:

Juan, all,

The adaptive wall function formulation as described in the paper of Medic et al. has two serious issues. First, the convergence stalls after three orders of magnitude or so and second it only works with an upwind discretization of the inviscid fluxes. For some reason JST type of schemes gave rather poor results. In order to have a reliable implementation, both issues need to be addressed. Unfortunately, we never did.

Edwin

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-307294686, or mute the thread https://github.com/notifications/unsubscribe-auth/APKNmwNgNVmXDYW_tqtQzvpDvxGe4OeLks5sCM7ugaJpZM4NvG6w .

petebachant commented 7 years ago

Any ideas on a good test case(s) to ensure the implementation is correct? Maybe the RANS flat plate case at one Re that puts us in the wall-resolved regime, one with y+ ~10, and one with y+ ~ 50? I haven't taken a look at that particular case--just wondering if it might work.

vdweide commented 6 years ago

How do you specify that wall functions have to be used? Is there a general boolean flag, e.g. UseWallFunctions, or is it organized per marker? Is there an example of a cfg file how this is done? The reason for asking is that for WM-LES for the high order solver we need the same info and it would make sense to do the specification identically, if possible.

economon commented 6 years ago

@petebachant: the paper that I linked earlier on this thread had some fairly simple examples similar to the one you propose.. different flat plates and so on with varying y^+. You can likely reuse the grid(s) in the TestCases repo.

@vdweide: it was very simple in this first implementation. It checks for a single boolean for whether or not to apply wall functions (WALL_FUNCTIONS= YES or NO, https://github.com/su2code/SU2/blob/43b8a4015c4b09b01d78e4243d32011c663c2b70/SU2_CFD/src/solver_direct_mean.cpp#L15339), and then assumes they should be applied to all no-slip walls (heat flux or isothermal, https://github.com/su2code/SU2/blob/43b8a4015c4b09b01d78e4243d32011c663c2b70/SU2_CFD/src/solver_direct_mean.cpp#L16653). This could of course be made more specific for individual markers using a config structure like MARKER_MONITORING to list them, or by making a dedicated BC type for it potentially.

juanjosealonso commented 6 years ago

All,

Regarding the wall function specification, if we are going to settle on a standard way that could work for both the FV and DG-FEM solvers, it might be good to think about some modifiers that can later allow for a generality of approaches.

We can make a distinction between wall functions and wall models, but this is subject to discussion…in essence both are performing similar duties, but people should express their opinions.

With that said, how about a boolean such as USE_WALL_FUNCTIONS / USE_WALL_MODEL (YES or NO) with an optional argument WALL_FUNCTION_TYPE / WALL_MODEL_TYPE that can taken one of many pre-specified values that can be added as these options are developed and tested? Certainly options like STANDARD_WALL_FUNCTION, ADAPTIVE_WALL_FUNCTION, SCALABLE_WALL_FUNCTION, compressible and incompressible versions, and even EQUILIBRIUM_WALL_MODEL and NONEQUILIBRIUM_WALL_MODEL are things that are likely to be in the code in the near future.

Thoughts?

Juan

On Sep 5, 2017, at 2:27 PM, Thomas D. Economon notifications@github.com<mailto:notifications@github.com> wrote:

@petebachanthttps://github.com/petebachant: the paper that I linked earlier on this thread had some fairly simple examples similar to the one you propose.. different flat plates and so on with varying y^+. You can likely reuse the grid(s) in the TestCases repo.

@vdweidehttps://github.com/vdweide: it was very simple in this first implementation. It checks for a single boolean for whether or not to apply wall functions (WALL_FUNCTIONS= YES or NO, https://github.com/su2code/SU2/blob/43b8a4015c4b09b01d78e4243d32011c663c2b70/SU2_CFD/src/solver_direct_mean.cpp#L15339), and then assumes they should be applied to all no-slip walls (heat flux or isothermal, https://github.com/su2code/SU2/blob/43b8a4015c4b09b01d78e4243d32011c663c2b70/SU2_CFD/src/solver_direct_mean.cpp#L16653). This could of course be made more specific for individual markers using a config structure like MARKER_MONITORING to list them, or by making a dedicated BC type for it potentially.

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/su2code/SU2/issues/399#issuecomment-327308587, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ADpSxJUhyEBSuSzHV1a7BZM_Frxbtb5sks5sfbzUgaJpZM4NvG6w.

vdweide commented 6 years ago

Juan,

I think that having the flexibility to apply wall functions on only a subset of the viscous wall boundaries would be useful. E.g., if you are interested in a wing, you can apply integration to the wall there, but apply wall functions on the fuselage. My preference would be to have a dedicated BC type for that, i.e. Tom's second suggestion.

Apart from this, we can have the different wall model types as a single input parameter, i.e. you apply the same wall model type for all viscous boundaries. Or would you like to have the flexibility to specify this as well per individual marker?

Edwin

juanjosealonso commented 6 years ago

Edwin,

I think it is a good idea to be able to have some level of control over the application of the wall functions on a marker-by-marker basis: as you say, it would be fairly common to have a wing marker where integration to the wall makes sense and a fuselage marker where wall functions (or even an inviscid BC) is the right thing to do. So let’s indeed do something like what Tom suggests (in the spirit of MARKER_MONITORING) so that one can control which markers get wall functions applied. If a list of markers is not specified in the config file, the default behavior should be that all no-slip walls get wall function BCs. If a list of markers is specified, then those and only those markers listed get wall function Bfs (with the others getting integration to the wall treatment).

Now, what you point out in the second half of your message is the ultimate in flexibility: not only specifying wall functions in a subset of the no-slip wall markers, but also allowing for different kinds of wall functions to be used in those markers. I imagine that this would be a seldom used capability…but someone out there might have the need.

I guess it does not seem to me that the coding of the most flexible logic is that much more complicated, so we may as well do that. If you guys think it is too complicated, the advertised options in the config file can allow some simpler behavior.

My two cents,

Juan

On Sep 5, 2017, at 9:58 PM, Edwin van der Weide notifications@github.com<mailto:notifications@github.com> wrote:

Juan,

I think that having the flexibility to apply wall functions on only a subset of the viscous wall boundaries would be useful. E.g., if you are interested in a wing, you can apply integration to the wall there, but apply wall functions on the fuselage. My preference would be to have a dedicated BC type for that, i.e. Tom's second suggestion.

Apart from this, we can have the different wall model types as a single input parameter, i.e. you apply the same wall model type for all viscous boundaries. Or would you like to have the flexibility to specify this as well per individual marker?

Edwin

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/su2code/SU2/issues/399#issuecomment-327374728, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ADpSxOEhKLW_U0n9PDoz5m6cJoCScV3_ks5sfiZdgaJpZM4NvG6w.

vdweide commented 6 years ago

What about generalizing the MARKER_MONITORING approach a bit, where the markers have an argument what type of wall function to apply, i.e.

MARKER_WALL_FUNCTIONS= (Wing, STANDARD_WALL_FUNCTION, Fuselage, ADAPTIVE_WALL_FUNCTION)

Would that be an option? For clarity we could actually add NO_WALL_FUNCTION as well, or do you think that is not necessary?

Edwin

juanjosealonso commented 6 years ago

I am ok with that idea. Tom, does it create the expectation that other lists would be handled the same way? If not, let’s go with it.

Cheers,

Juan

On Sep 5, 2017, at 11:40 PM, Edwin van der Weide notifications@github.com<mailto:notifications@github.com> wrote:

What about generalizing the MARKER_MONITORING approach a bit, where the markers have an argument what type of wall function to apply, i.e.

MARKER_WALL_FUNCTIONS= (Wing, STANDARD_WALL_FUNCTION, Fuselage, ADAPTIVE_WALL_FUNCTION)

Would that be an option? For clarity we could actually add NO_WALL_FUNCTION as well, or do you think that is not necessary?

Edwin

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/su2code/SU2/issues/399#issuecomment-327390022, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ADpSxGxuS9AT-f9TcIhY7GxMrSlT9ACUks5sfj55gaJpZM4NvG6w.

vdweide commented 6 years ago

Juan,

I don't think that is an issue. I'll create a PR in develop for the input data, such that it can be merged into both the FVM and FEM solvers.

Edwin

GomerOfDoom commented 6 years ago

Hi guys,

Somehow these ended up in my Clutter folder, so I missed them. Of course, I’m fine with whatever you guys think is best.

My thoughts, prior to reading this chain, were that we would use separate boundary conditions for each surface. I.e. we would have different markers for each wall modeled BC. For example:

MARKER_WM_1D_EQUIL=(body, thickness, other params, etc)

MARKER_WM_1D_NON_EQUIL=(body, thickness, other params, etc)

MARKER_WM_3D_NON_EQUIL=(body, thickness, other params, etc)

There definitely could be cases where we might apply one type of wall model to one surface, but another to a different surface.

Ooops… telecon is starting. I’ll talk to you all in a minute.

-Paul

From: juanjosealonso [mailto:notifications@github.com] Sent: Wednesday, September 6, 2017 7:32 AM To: su2code/SU2 SU2@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: Re: [su2code/SU2] Add wall functions (#399)

I am ok with that idea. Tom, does it create the expectation that other lists would be handled the same way? If not, let’s go with it.

Cheers,

Juan

On Sep 5, 2017, at 11:40 PM, Edwin van der Weide notifications@github.com<mailto:notifications@github.com> wrote:

What about generalizing the MARKER_MONITORING approach a bit, where the markers have an argument what type of wall function to apply, i.e.

MARKER_WALL_FUNCTIONS= (Wing, STANDARD_WALL_FUNCTION, Fuselage, ADAPTIVE_WALL_FUNCTION)

Would that be an option? For clarity we could actually add NO_WALL_FUNCTION as well, or do you think that is not necessary?

Edwin

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/su2code/SU2/issues/399#issuecomment-327390022, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ADpSxGxuS9AT-f9TcIhY7GxMrSlT9ACUks5sfj55gaJpZM4NvG6w.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/su2code/SU2/issues/399#issuecomment-327502096, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AGnNSppFttPzZQE7j5w6Xjpj-3OivlMYks5sfqzMgaJpZM4NvG6w.

economon commented 6 years ago

In case others are interested, I was going through old papers, and I found the following one that was very helpful for the preliminary version of wall functions for the FV solver (feature_Wall_Functions branch):

Nichols, R. H., Nelson, C. C., "Wall Function Boundary Conditions Including Heat Transfer and Compressibility," AIAA JOURNAL, Vol. 42, No. 6, June 2004.

vdweide commented 6 years ago

The PR about the input specification for wall functions is closed for now, but it does not mean that we cannot continue the discussion. The question is how to handle the necessary input parameters for the different wall models. First option would be to add them as additional arguments when the type of wall model is specified. However, this could lead to a rather messy input. A different option would be to have a separate input line for each wall function marker.

Edwin

stale[bot] commented 4 years ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. If this is still a relevant issue please comment on it to restart the discussion. Thank you for your contributions.

monika1387 commented 4 years ago

This wall function implementation is still a problem? Because I am running my case using this wall function Boundary Marker, but its failing without giving any error in error output file. Its just stopped after running for max 30mins. I am running Standard Wall function marker for RANS with SST model. Can someone advise on this? Thank you in advance for help!

pcarruscag commented 4 years ago

I gave the wall function routines a cleanup in #975 and I think something is missing, we compute the shear stresses but they are not considered in the BC's of the NS solver.

monika1387 commented 4 years ago

Sure, now make sense why it is showing issues, Thank you for the confirmation!

On Wed, May 27, 2020 at 5:22 AM Pedro Gomes notifications@github.com wrote:

I gave the wall function routines a cleanup in #975 https://github.com/su2code/SU2/pull/975 and I think something is missing, we compute the shear stresses but they are not considered in the BC's of the NS solver.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634539255, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALJ3OXAF4VX5CFIJVJFLIDTRTTLVBANCNFSM4DN4N2YA .

-- Thank you,

Monika Chauhan

Graduate Research Assistant, Doctoral Program

Aerospace and Ocean Engineering Dept, Virginia Tech,Blacksburg, VA 24061 Cell# 540-998-5012

juanjosealonso commented 4 years ago

Monika,

There is a chance we could work together on this…Eduardo Molina (copied) co-advised a short project last year on wall functions which made some progress but did not finalize an implementation. It may be a better starting point. Eduardo, can you comment where that code went?

I am also copying Matt S. who is interested in wall function implementations in SU2. Perhaps we can get everyone on the same page to get these working in the baseline state…and then improve them in various ways?

Best,

Juan

On May 27, 2020, at 9:20 AM, monika1387 notifications@github.com wrote:

Sure, now make sense why it is showing issues, Thank you for the confirmation!

On Wed, May 27, 2020 at 5:22 AM Pedro Gomes notifications@github.com wrote:

I gave the wall function routines a cleanup in #975 https://github.com/su2code/SU2/pull/975 and I think something is missing, we compute the shear stresses but they are not considered in the BC's of the NS solver.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634539255, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALJ3OXAF4VX5CFIJVJFLIDTRTTLVBANCNFSM4DN4N2YA .

-- Thank you,

Monika Chauhan

Graduate Research Assistant, Doctoral Program

Aerospace and Ocean Engineering Dept, Virginia Tech,Blacksburg, VA 24061 Cell# 540-998-5012 — You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634777816, or unsubscribe https://github.com/notifications/unsubscribe-auth/AA5FFRFBZC767ZAD7HMXWQ3RTU4VPANCNFSM4DN4N2YA.

monika1387 commented 4 years ago

Sure Juan! I would like to. Because I was running on RANS but I am working on LES to make this wall functions work which is not working either. Please let us know how we can support and work to get this working for LES simulation (wall models in SU2).

On Wed, May 27, 2020 at 12:29 PM Juan J. Alonso notifications@github.com wrote:

Monika,

There is a chance we could work together on this…Eduardo Molina (copied) co-advised a short project last year on wall functions which made some progress but did not finalize an implementation. It may be a better starting point. Eduardo, can you comment where that code went?

I am also copying Matt S. who is interested in wall function implementations in SU2. Perhaps we can get everyone on the same page to get these working in the baseline state…and then improve them in various ways?

Best,

Juan

On May 27, 2020, at 9:20 AM, monika1387 notifications@github.com wrote:

Sure, now make sense why it is showing issues, Thank you for the confirmation!

On Wed, May 27, 2020 at 5:22 AM Pedro Gomes notifications@github.com wrote:

I gave the wall function routines a cleanup in #975 https://github.com/su2code/SU2/pull/975 and I think something is missing, we compute the shear stresses but they are not considered in the BC's of the NS solver.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634539255, or unsubscribe < https://github.com/notifications/unsubscribe-auth/ALJ3OXAF4VX5CFIJVJFLIDTRTTLVBANCNFSM4DN4N2YA

.

-- Thank you,

Monika Chauhan

Graduate Research Assistant, Doctoral Program

Aerospace and Ocean Engineering Dept, Virginia Tech,Blacksburg, VA 24061 Cell# 540-998-5012 — You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/su2code/SU2/issues/399#issuecomment-634777816>, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AA5FFRFBZC767ZAD7HMXWQ3RTU4VPANCNFSM4DN4N2YA .

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634782860, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALJ3OXD2JNVVJZGRX6A3LDLRTU5VXANCNFSM4DN4N2YA .

-- Thank you,

Monika Chauhan

Graduate Research Assistant, Doctoral Program

Aerospace and Ocean Engineering Dept, Virginia Tech,Blacksburg, VA 24061 Cell# 540-998-5012

EduardoMolina commented 4 years ago

Hi Monika,

We are in the V&V process of a new wall model LES implementation. Also, I took the opportunity to start fixing the RANS wall function since it will share the same code structure with the WMLES part, i.e. replace the wall shear stress in the NS fluxes.

We didn’t extensively test the RANS wall functions, only very limited test cases. It will be great if you could run your own cases using the following branch:

https://github.com/su2code/SU2/tree/feature_WallModelLES https://github.com/su2code/SU2/tree/feature_WallModelLES

Please let me know if you need additional help and looking forward to hear from you.

Best,

Eduardo

On 27 May 2020, at 11:30, monika1387 notifications@github.com wrote:

Sure Juan! I would like to. Because I was running on RANS but I am working on LES to make this wall functions work which is not working either. Please let us know how we can support and work to get this working for LES simulation (wall models in SU2).

On Wed, May 27, 2020 at 12:29 PM Juan J. Alonso notifications@github.com wrote:

Monika,

There is a chance we could work together on this…Eduardo Molina (copied) co-advised a short project last year on wall functions which made some progress but did not finalize an implementation. It may be a better starting point. Eduardo, can you comment where that code went?

I am also copying Matt S. who is interested in wall function implementations in SU2. Perhaps we can get everyone on the same page to get these working in the baseline state…and then improve them in various ways?

Best,

Juan

On May 27, 2020, at 9:20 AM, monika1387 notifications@github.com wrote:

Sure, now make sense why it is showing issues, Thank you for the confirmation!

On Wed, May 27, 2020 at 5:22 AM Pedro Gomes notifications@github.com wrote:

I gave the wall function routines a cleanup in #975 https://github.com/su2code/SU2/pull/975 and I think something is missing, we compute the shear stresses but they are not considered in the BC's of the NS solver.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634539255, or unsubscribe < https://github.com/notifications/unsubscribe-auth/ALJ3OXAF4VX5CFIJVJFLIDTRTTLVBANCNFSM4DN4N2YA

.

-- Thank you,

Monika Chauhan

Graduate Research Assistant, Doctoral Program

Aerospace and Ocean Engineering Dept, Virginia Tech,Blacksburg, VA 24061 Cell# 540-998-5012 — You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/su2code/SU2/issues/399#issuecomment-634777816>, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AA5FFRFBZC767ZAD7HMXWQ3RTU4VPANCNFSM4DN4N2YA .

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634782860, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALJ3OXD2JNVVJZGRX6A3LDLRTU5VXANCNFSM4DN4N2YA .

-- Thank you,

Monika Chauhan

Graduate Research Assistant, Doctoral Program

Aerospace and Ocean Engineering Dept, Virginia Tech,Blacksburg, VA 24061 Cell# 540-998-5012 — You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634857692, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACKWMCOUWWOQDBVQOXS3IDTRTVL3VANCNFSM4DN4N2YA.

monika1387 commented 4 years ago

Hello Eduardo,

Sure! I will run my case and will let you know. Thank you for the information!

Monika Chauhan Sent from my iPhone

On 27-May-2020, at 8:32 PM, Eduardo Molina notifications@github.com wrote:

 Hi Monika,

We are in the V&V process of a new wall model LES implementation. Also, I took the opportunity to start fixing the RANS wall function since it will share the same code structure with the WMLES part, i.e. replace the wall shear stress in the NS fluxes.

We didn’t extensively test the RANS wall functions, only very limited test cases. It will be great if you could run your own cases using the following branch:

https://github.com/su2code/SU2/tree/feature_WallModelLES https://github.com/su2code/SU2/tree/feature_WallModelLES

Please let me know if you need additional help and looking forward to hear from you.

Best,

Eduardo

On 27 May 2020, at 11:30, monika1387 notifications@github.com wrote:

Sure Juan! I would like to. Because I was running on RANS but I am working on LES to make this wall functions work which is not working either. Please let us know how we can support and work to get this working for LES simulation (wall models in SU2).

On Wed, May 27, 2020 at 12:29 PM Juan J. Alonso notifications@github.com wrote:

Monika,

There is a chance we could work together on this…Eduardo Molina (copied) co-advised a short project last year on wall functions which made some progress but did not finalize an implementation. It may be a better starting point. Eduardo, can you comment where that code went?

I am also copying Matt S. who is interested in wall function implementations in SU2. Perhaps we can get everyone on the same page to get these working in the baseline state…and then improve them in various ways?

Best,

Juan

On May 27, 2020, at 9:20 AM, monika1387 notifications@github.com wrote:

Sure, now make sense why it is showing issues, Thank you for the confirmation!

On Wed, May 27, 2020 at 5:22 AM Pedro Gomes notifications@github.com wrote:

I gave the wall function routines a cleanup in #975 https://github.com/su2code/SU2/pull/975 and I think something is missing, we compute the shear stresses but they are not considered in the BC's of the NS solver.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634539255, or unsubscribe < https://github.com/notifications/unsubscribe-auth/ALJ3OXAF4VX5CFIJVJFLIDTRTTLVBANCNFSM4DN4N2YA

.

-- Thank you,

Monika Chauhan

Graduate Research Assistant, Doctoral Program

Aerospace and Ocean Engineering Dept, Virginia Tech,Blacksburg, VA 24061 Cell# 540-998-5012 — You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/su2code/SU2/issues/399#issuecomment-634777816>, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AA5FFRFBZC767ZAD7HMXWQ3RTU4VPANCNFSM4DN4N2YA .

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634782860, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALJ3OXD2JNVVJZGRX6A3LDLRTU5VXANCNFSM4DN4N2YA .

-- Thank you,

Monika Chauhan

Graduate Research Assistant, Doctoral Program

Aerospace and Ocean Engineering Dept, Virginia Tech,Blacksburg, VA 24061 Cell# 540-998-5012 — You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634857692, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACKWMCOUWWOQDBVQOXS3IDTRTVL3VANCNFSM4DN4N2YA.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

monika1387 commented 4 years ago

Hi Edurado,

We are running LES with wall functions and just wondering if there is a mistake on lines 1322 to 1325 of Common/src/adt_structure.cpp . It is currently < parCoor[0] = -0.5(parCoor[0] + parCoor[1] + parCoor[2] + 1.0); < parCoor[1] = 0.5(parCoor[0] + 1.0); < parCoor[2] = 0.5(parCoor[1] + 1.0); < parCoor[3] = 0.5(parCoor[2] + 1.0);

I believe it should be

weightsInterpol[0] = -0.5*(parCoor[0] + parCoor[1] + parCoor[2] +

1.0); weightsInterpol[1] = 0.5(parCoor[0] + 1.0); weightsInterpol[2] = 0.5(parCoor[1] + 1.0); weightsInterpol[3] = 0.5*(parCoor[2] + 1.0); Can you please confirm if this is correct?

Thank you Monika

On Wed, May 27, 2020 at 8:32 PM Eduardo Molina notifications@github.com wrote:

Hi Monika,

We are in the V&V process of a new wall model LES implementation. Also, I took the opportunity to start fixing the RANS wall function since it will share the same code structure with the WMLES part, i.e. replace the wall shear stress in the NS fluxes.

We didn’t extensively test the RANS wall functions, only very limited test cases. It will be great if you could run your own cases using the following branch:

https://github.com/su2code/SU2/tree/feature_WallModelLES < https://github.com/su2code/SU2/tree/feature_WallModelLES>

Please let me know if you need additional help and looking forward to hear from you.

Best,

Eduardo

On 27 May 2020, at 11:30, monika1387 notifications@github.com wrote:

Sure Juan! I would like to. Because I was running on RANS but I am working on LES to make this wall functions work which is not working either. Please let us know how we can support and work to get this working for LES simulation (wall models in SU2).

On Wed, May 27, 2020 at 12:29 PM Juan J. Alonso < notifications@github.com> wrote:

Monika,

There is a chance we could work together on this…Eduardo Molina (copied) co-advised a short project last year on wall functions which made some progress but did not finalize an implementation. It may be a better starting point. Eduardo, can you comment where that code went?

I am also copying Matt S. who is interested in wall function implementations in SU2. Perhaps we can get everyone on the same page to get these working in the baseline state…and then improve them in various ways?

Best,

Juan

On May 27, 2020, at 9:20 AM, monika1387 notifications@github.com wrote:

Sure, now make sense why it is showing issues, Thank you for the confirmation!

On Wed, May 27, 2020 at 5:22 AM Pedro Gomes < notifications@github.com> wrote:

I gave the wall function routines a cleanup in #975 https://github.com/su2code/SU2/pull/975 and I think something is missing, we compute the shear stresses but they are not considered in the BC's of the NS solver.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634539255, or unsubscribe <

https://github.com/notifications/unsubscribe-auth/ALJ3OXAF4VX5CFIJVJFLIDTRTTLVBANCNFSM4DN4N2YA

.

-- Thank you,

Monika Chauhan

Graduate Research Assistant, Doctoral Program

Aerospace and Ocean Engineering Dept, Virginia Tech,Blacksburg, VA 24061 Cell# 540-998-5012 — You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/su2code/SU2/issues/399#issuecomment-634777816>, or unsubscribe <

https://github.com/notifications/unsubscribe-auth/AA5FFRFBZC767ZAD7HMXWQ3RTU4VPANCNFSM4DN4N2YA

.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634782860, or unsubscribe < https://github.com/notifications/unsubscribe-auth/ALJ3OXD2JNVVJZGRX6A3LDLRTU5VXANCNFSM4DN4N2YA

.

-- Thank you,

Monika Chauhan

Graduate Research Assistant, Doctoral Program

Aerospace and Ocean Engineering Dept, Virginia Tech,Blacksburg, VA 24061 Cell# 540-998-5012 — You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/su2code/SU2/issues/399#issuecomment-634857692>, or unsubscribe < https://github.com/notifications/unsubscribe-auth/ACKWMCOUWWOQDBVQOXS3IDTRTVL3VANCNFSM4DN4N2YA .

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-635018259, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALJ3OXH6UGKNEB2TL73OOPLRTWWK3ANCNFSM4DN4N2YA .

-- Thank you,

Monika Chauhan

Graduate Research Assistant, Doctoral Program

Aerospace and Ocean Engineering Dept, Virginia Tech,Blacksburg, VA 24061 Cell# 540-998-5012

monika1387 commented 4 years ago

Hi Eduardo,

So far LES_FEM worked pretty well for us with few minor changes at our end. Thank you for all your support and work in this. But I do have one query, is anytime the SU2 team worked on averaging mean flow data (or even perturbations) to avoid huge memory issues with large LES dumped files in SU2? I mean we are doing averaging but for a large number of files locally offline is challenging so we are wondering if you have anything like that during the simulation SU2 can dump averaged mean flow for a number of iterations? Please advise if any work or anything closer worked upon which we can use or take it from there. Thank you in advance for your help!

On Wed, May 27, 2020 at 8:32 PM Eduardo Molina notifications@github.com wrote:

Hi Monika,

We are in the V&V process of a new wall model LES implementation. Also, I took the opportunity to start fixing the RANS wall function since it will share the same code structure with the WMLES part, i.e. replace the wall shear stress in the NS fluxes.

We didn’t extensively test the RANS wall functions, only very limited test cases. It will be great if you could run your own cases using the following branch:

https://github.com/su2code/SU2/tree/feature_WallModelLES < https://github.com/su2code/SU2/tree/feature_WallModelLES>

Please let me know if you need additional help and looking forward to hear from you.

Best,

Eduardo

On 27 May 2020, at 11:30, monika1387 notifications@github.com wrote:

Sure Juan! I would like to. Because I was running on RANS but I am working on LES to make this wall functions work which is not working either. Please let us know how we can support and work to get this working for LES simulation (wall models in SU2).

On Wed, May 27, 2020 at 12:29 PM Juan J. Alonso < notifications@github.com> wrote:

Monika,

There is a chance we could work together on this…Eduardo Molina (copied) co-advised a short project last year on wall functions which made some progress but did not finalize an implementation. It may be a better starting point. Eduardo, can you comment where that code went?

I am also copying Matt S. who is interested in wall function implementations in SU2. Perhaps we can get everyone on the same page to get these working in the baseline state…and then improve them in various ways?

Best,

Juan

On May 27, 2020, at 9:20 AM, monika1387 notifications@github.com wrote:

Sure, now make sense why it is showing issues, Thank you for the confirmation!

On Wed, May 27, 2020 at 5:22 AM Pedro Gomes < notifications@github.com> wrote:

I gave the wall function routines a cleanup in #975 https://github.com/su2code/SU2/pull/975 and I think something is missing, we compute the shear stresses but they are not considered in the BC's of the NS solver.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634539255, or unsubscribe <

https://github.com/notifications/unsubscribe-auth/ALJ3OXAF4VX5CFIJVJFLIDTRTTLVBANCNFSM4DN4N2YA

.

-- Thank you,

Monika Chauhan

Graduate Research Assistant, Doctoral Program

Aerospace and Ocean Engineering Dept, Virginia Tech,Blacksburg, VA 24061 Cell# 540-998-5012 — You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/su2code/SU2/issues/399#issuecomment-634777816>, or unsubscribe <

https://github.com/notifications/unsubscribe-auth/AA5FFRFBZC767ZAD7HMXWQ3RTU4VPANCNFSM4DN4N2YA

.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-634782860, or unsubscribe < https://github.com/notifications/unsubscribe-auth/ALJ3OXD2JNVVJZGRX6A3LDLRTU5VXANCNFSM4DN4N2YA

.

-- Thank you,

Monika Chauhan

Graduate Research Assistant, Doctoral Program

Aerospace and Ocean Engineering Dept, Virginia Tech,Blacksburg, VA 24061 Cell# 540-998-5012 — You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/su2code/SU2/issues/399#issuecomment-634857692>, or unsubscribe < https://github.com/notifications/unsubscribe-auth/ACKWMCOUWWOQDBVQOXS3IDTRTVL3VANCNFSM4DN4N2YA .

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/su2code/SU2/issues/399#issuecomment-635018259, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALJ3OXH6UGKNEB2TL73OOPLRTWWK3ANCNFSM4DN4N2YA .

-- Thank you,

Monika Chauhan

Graduate Research Assistant, Doctoral Program

Aerospace and Ocean Engineering Dept, Virginia Tech,Blacksburg, VA 24061 Cell# 540-998-5012