[CFXer]

Hello!

I'm trying to simulate the picture's structure.

At inlet, two inlet are considered whose velocity is around 6 m/s.

And the flow is go through the cuboid core structure.

By using k-epsilon and SST turbulence model, I earned streamline.

In k-e model, its residual comes down below even 1e-7, I thought it might be converged.

But the streamline is so weird. Like picture 3, the streamline of the fluid is circulate at the center of inlet header.

In my opinion, it should collide at center like ficture 1.

I also used SST model, but its tendency was same.

Monitoring residual status, like picture 2 and 4, its residual comes down first and bouncing up and comes down again.

Before its bouncing, its streamline is collide at center, but after bouncing it becomes circulate.

I don't know what's the problem with my simulation. I think it's not well fitted to real physics.

Please anybody help me.

[CFXer]

I also attach k-epsilon residual result.

The original picture show SST result.

[ghorrocks]

k-e turbulence models are known to generate too much turbulent viscosity in stagnation regions, which is what you have where the two jets impinge. Too much viscosity means the flow is damped too much, leading to stable convergence and symmetric results. SST has a more realistic turbulence generation in impingement regions, so it will suffer this problem much less. This will lead to unstable, difficult convergence and unsymmetric results. This appears to be what you are seeing.

In short, you appear to be modelling impinging fluid jets and this is going to be very unstable. k-e models falsely model this as stable, but the SST model is more accurate in predicting it to be unstable.

I also note you ran thousands of iterations for these simulations. If this is a steady state simulation then this is probably excessive and means you have the physical time step step far too small.

[CFXer]

Thank you Glenn.

You said, does the SST model's characteristic makes its flow circulating?

I didn't attached the k-e model's result, it shows same streamline as SST shows.

Then what could you suggest to do in this simulation? other turbulence model except k-e or SST model?

or increasing timescale?

or.. just believe this simulation result?

[ghorrocks]

I recommend you do some reading into the topics I discussed in my last post so you understand what I am talking about. Have a look into the accuracy of k-e and k-omega based turbulence models in impinging jets.

Getting trustworthy CFD results is a complex task and I am not going to answer that in a forum post. Have a look at "Computational Fluid Dynamics" by Roache as it well describes the issues involved in getting accurate CFD simulations. Also the FAQ has a good link: https://www.cfd-online.com/Wiki/Ansy...publishable.3F

[CFXer]

I'm so sorry but I cannot find your post...

I think I'm not familiar enough to this site.

I have searched the whole site "impinging jets" but I cannot find your post.

Can you please leave me a link? I'm so sorry..

[ghorrocks]

I was referring to post #3 on this thread, not a post elsewhere on the forum.

And for the impinging jets thing, I was referring to a literature search. So time to go to the library, or to try your luck with Google.