search

NREL / SOWFA

Other
130 stars 113 forks source link

Problems occur when running example.ABL. Flatterrain. Neutral #94

Open gaoxiangand opened 1 year ago

gaoxiangand commented 1 year ago

When I use example. ABL. Flatterrain In the neutral case, it was found that the generated turbulent inflow of the wind farm was less than I wanted, and the recovery of the wind turbine wake was also slower than I wanted. Therefore, I want to achieve high turbulence (12.8%) inflow and fast wake recovery(As shown in the figure below) by increasing the surface roughness to 1 in the case of neutral atmosphere; 图片 However, I find that increasing the roughness does not seem to have much effect on the inflow turbulence and wake recovery. I am not sure whether it is the problem of my grid or settings, or whether the developed case (example. ABL. Flatterrain. Neutral) cannot achieve high turbulence and fast wake recovery under neutral conditions. The grid I use is mainly two-layer encryption, with a resolution of dx20-10 and a calculation domain of 2000 800 800. 图片 图片

ewquon commented 1 year ago

I don't understand what you mean by "two-layer encryption, with a resolution of dx20-10". I would use 10-m resolution throughout the ABL on a larger domain. Have you looked at the flow field to make sure the flow appears physical (e.g., the periodic boundaries aren't creating spurious flow features due to your domain size)? Have you verified that the turbulence statistics have fully developed?

gaoxiangand commented 1 year ago

Hello, when I say two-layer encryption, the resolution DX20-10 means the two resolutions used in the whole computing domain. The resolution used near the wind turbine is 10, and the resolution used outside the computing domain is 20.

I'm sorry, but I don't understand what you said. I would like to ask how you confirmed that the turbulence statistics have been fully developed. And how do you observe that the flow field looks physical?

ewquon commented 1 year ago

Since your qwall is (0 0 0), your solution should reach an equilibrium over time. Check your planarAverages output, your velocity variances (and therefore TKE) at all heights below your capping inversion should converge after several hours of simulation time. Otherwise, I would not say that your turbulence field is fully developed and I would not put turbines in that flow.

You can also check instantaneous flowfield slices through your domain to qualitatively verify that there is developed turbulence.