Mismatch in wind direction between precursor/successor ABL simulation

[AlexGuntner]

Hi everyone,

I’m encountering issues when running my ABL simulation using saved inflow boundary condition data. The case involves a neutral atmospheric boundary layer. I initially performed a spin-up and precursor simulation with a single level of mesh refinement. Now, I’m attempting to run the main simulation with two additional levels of refinement focused around the wind turbine wakes.

The issue is that the wind direction that I imposed in the precursor (0° at hub height = 150m) does not match the one that I have when running the successor simulation with the wind turbines, as you can see here:

Do you have any ideas on what could be wrong? I am attaching my input files as well.

3D_abl_neutral_successor_inp.txt
abl_neutral_precursor_inp.txt
Output.log

[hgopalan]

Have you observed this issue before? If not, Can we take a step back and check something?

Can you save a sampling line in the precursor and average it over time and plot the u and v components. Can you repeat the same for the successor simulations with refinement region and without any actuators and compute the temporally averaged wind speed along the same sampling line? We can first check to see if something in the refinement region is causing artificial wind turning.

[AlexGuntner]

Hi @hgopalan,

I did not observe this before. I sampled two horizontal lines at hub height (100 seconds) for the successor without wind turbines, one outside the refinements (with the same resolution as the precursor) and the other inside. As you can see, the vertical velocity component does indeed increase once it enters the refined regions. I realized that maybe diffusion.max_order = 1 might have an impact, so I switched it back to 2 as per default, but I got the same results. I don't know what else might be happening.

[hgopalan]

Are the turbulence models also the same as the simulations before or did you change the model?

[AlexGuntner]

I am using the same turbulence model for both simulations (AMD), the only difference should be in the refinement levels.

[hgopalan]

Can we try one more thing? Can you run with the older version of the AMR-Wind code which looked ok and reproduce the same line plots so that we have some reference to see what's going on?

[AlexGuntner]

That will take me some time, as my schedule is tight and I am currently re-running simulations without the effects of Coriolis force to avoid issues related to wind direction.

Which older version do you refer to? I am using AMR-Wind version :: v3.4.1-30-gaa1c63a8

[hgopalan]

I think I misunderstood. The number of refinement levels you are using now, have you used them before and the velocity profiles looked alright? This is something I wanted to check.

However, if you have not run at these refinement levels before, there is no need to run with an earlier version of AMR-Wind.

[AlexGuntner]

These are the first ABL simulations that I am running with my current setup, so I do not have any good reference values. The wind direction is correct in the precursor domain, and this remains true in the successor domain up until the flow enters the refined regions.

[hgopalan]

Not sure of your timeline but can you quickly check by replacing AMD with Smagorinsky or OneEquation?

[AlexGuntner]

Yes I should be able to do that, it might still take me a few days as I have other things running at the moment, but I will report back once I have the results.

Just to be sure, you are only referring to switching models in the successor simulation?

[hgopalan]

Yes. Run the Smagorinsky model since OneEquation needs tke in the chk file or it starts it from scratch possibly.

[AlexGuntner]

Hi @hgopalan,

I have run the simulation using the Smagorinsky model, but I got the same results - flow direction is correct in the coarser area, but once it enters the refined region there is a slight increase in $U_y$, which offsets the wind by roughly 5°.

[github-actions]

This issue is stale because it has been open 30 days with no activity.