[geod]

Hi all,

I am currently working on a simple pipe-pipe heat exchanger with hot water flowing through the inner pipe and cold water flowing through the outer pipe. The mesh for my pipes was created using ICEM CFD and I generated it as a structured grid. I am required to obtain results for Nusselt number, friction factor and average heat transfer coefficient at a range of Reynolds number.

I have applied all the settings for solving the problem and run the calculation. So after completing 400+ iterations my scaled residuals are shown in the picture below. It seems to have converged except for continuity being unstable.

I have a couple queries regarding my project and would really appreciate it if anyone could lend a hand!

1. Will this unstable continuity affect my results majorly and if so how can I solve the problem?

2. In regards to obtaining results at a range of different reynolds number, the velocity of the fluid moving through the pipes should vary from 0.0321 to 3.4096 m/s, can I obtain results for the varying velocities or will I have to produce separate simulations for every step of increased velocity (and in turn reynolds number)?

convergence.jpg

Any help would be really appreciated. Cheers.

[ghost82]

Hi,
I think your simulation is not converged, not only the continuity but also other residuals. As you can see you have still varying number of faces with reversed flow.
You can read here:
http://www.cfd-online.com/Forums/flu...nvergence.html

First post by Centurion explain well how you can evaluate if your solution is converged.

[geod]

Originally Posted by ghost82

Hi,
I think your simulation is not converged, not only the continuity but also other residuals. As you can see you have still varying number of faces with reversed flow.
You can read here:
http://www.cfd-online.com/Forums/flu...nvergence.html

First post by Centurion explain well how you can evaluate if your solution is converged.

Hi ghost,

Thanks for the information. I think my mesh is of good quality and I have attempted the calculation with first order discretization, and then also reduced under-relaxtion factors after unsuccessful convergence, however I still don't seem to be able to reach a point of convergence. I am almost certain that the input parameters and solution I have posed are ok as I am using values that have been used before for this instance and gained results.

Is there something I am missing here?

Here is a screen of the residuals now:

resids.jpg

Thanks

[geod]

Started again and now the residuals looking much nicer however could do with continuity reducing to e-04, hopefully will come with more iterations. Could anyone advise if these values are ok for convergence and should I be able to optain results? Please really struggling to progress with this!

Thanks

scaledresids.jpg