discontinuous in velocity field

[秦志文]

Hello, I am doing a pipe flow at Re=18950, but I found the velocity contour shows obvious discontinuous because of grid, is there any wrong?

[Philipp Schlatter]

I guess you refer to discontinuities in the gradients of velocity; the actual velocity is continuous in the context of C0-continuity.

There is a lot of discussions on that aspect for SEM in this mailing list, and also in the Nek Users guide. In short, there is probably nothing wrong with your simulation, and the wiggles will diminish with increasing resolution.

Best, Philipp

On 2025-08-26 10:18, 秦志文 wrote:

Hello, I am doing a pipe flow at Re=18950, but I found the velocity contour shows obvious discontinuous because of grid, is there any wrong? --
You received this message because you are subscribed to the Google Groups "Nek5000" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nek5000+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/nek5000/cd74de57-32ae-4aea-972b-9a4127418050n%40googlegroups.com.

[YuHsiang Lan]

Hi,

For fun, I set up a customized ChatGPT agent last week called Nek5000 Assistant. It’s publicly available here:

  https://chatgpt.com/g/g-68a731e21f7c8191bcc684cc265a1d79-nek5000-assistant

It's not as good a s human, and some of the statements are questionable or even not correct.

However, it does the following for you, and might be a good redirect search engine.

- search nekdoc

- search google group

- search source code if couldn't find the answer.

- reduce the chance to falsely create made-up rea/par key.

Take your question as an example:

  https://chatgpt.com/share/68adb331-93a0-8009-acfa-9522e590af6a

I'm also curious how this customized version compared to general ChatGPT.

Hopefully, it's useful.

Yu-Hsiang

--