Learning Forum

[fatih29]

I am making solution for plate impingement jet flow. As the mesh number increases, the average convection heat transfer coefficient on the plate, the average temperature and the average Nu number do not change. But for a single point at the stopping point, the local Nu number decreases as the mesh increases.

I define a point on the plate right in the middle. I get the local heat flux and temperature value of this point using facets average. Then I calculate the number Nu.

What could be the reason for this to happen at the stopping point? Am I going the wrong way to calculate the local Nu number?

Thanks for your help.

[Rob]

Why would the value change as the mesh resolution is increased? Are you checking an average or min/max value?

[fatih29]

I put a point in the center of the plate where the fluid hits. For this point, I calculate the local Nu number by taking the convection heat transfer coefficient and the temperature of that point from ansys. This value changes as the mesh increases. I didn't know what to do.

[Rob]

Which will be changing as you resolve the mesh and near wall flow. As you refine it should change less and eventually reach a stable value: in other words a mesh dependency study. Basically as you refine the mesh you're better resolving the thermal boundary layer, so temperature will change and so will HTC as the flow is also better resolved.
The other common oversight when calculating HTC and Nu is the reference temperature. It's very common to not pick an equivalent value to the experiment so the values vary significantly even when the heat flux ties up well.

[fatih29]

Understood. I'm solving impingement jet stream. I am using SST transition as turbulence model. My y+ value is around max 2.1, min 0.12 . I wonder if the y+ value changes this situation. I've tried 1, 2, and 3 million meshes. Should I decrease the y plus a little while increasing the mesh? Average h and temperature on the plate do not change as the mesh increases. Doesn't it show independence from mesh in this case? The residuals are converging. At the same time, the average friction coefficient of the 6 points on the plate converges. But for the single point data I get for the local Nu number, h changes as the mesh increases. I got the reference value correctly when calculating h. I have never faced such a situation before. Thank you for your help.

[Rob]

Check the contours around the point. I've seen cases where the monitor was just in the "wrong" place and picked up on very subtle changes in the results. The y+ value sounds fine. I would question using the k-w model for jet impingement, I'd be looking at using RSM for this application if you're after really accurate results.

[YasserSelima]

Does the free surface oscillate?
I believe the nature of the problem is transient. This temperature will keep oscillating at the stagnation point ... try transient simulation and plot the temperature Vs. time. Then you can use time averaged value.

[fatih29]

Understood. Thank you. I will implement your recommendations.

[fatih29]

I didn't think you said that. I solved only steady state the problem. I'll try the time dependent as well.
I don't understand the part you mean free surface. If you mean the impact surface. The mean values on the impact surface do not change. Also, the average values I get for 6 points do not change.

[Rob]

Assuming it's a single phase system there is no free surface, is thinking you're looking at a water jet in air or similar.

[YasserSelima]

In the past.I have conducted experiments of impinging water jet. I observed that the water free surface was actually oscillating. I didn't do numerical simulation though. It was in air as mentioned.