[Akkuboy]

Hello,

I have designed a heat exchanger (closed at the top and bottom) with plates inside. It is a turbulent flow model (k-omega) with adiabatic wall conditions (Heat flux = 0 assigned in boundary condition).

Figure mesh 2 - Orthographic view of the heat exchanger (top cover and fluid domain hidden to show the solid domain and the plates)

Figure mesh 3 - Orthographic view of the heat exchanger with the fluid domain and solid domain combined)

Figure mesh 4 - Cut cross-sectional plane view

Figure 500 - Graph obtained after 500 iterations of calculations

Figure temperature profile - Temperature contour obtained in the results tab

The meshed might seem distorted in certain views, but that is just a graphics error that can be ignored. There are hexa-dominant meshes throughout with triangular meshes in some sections.

I'm facing a few issues with the same and would request anyone with any leads to help me.

1) Hexa dominant meshes have been uniformly generated throughout the model with various method sizing, edge sizing, and element sizes assigned. However, in the cross-sectional plane through the model, it is seen that the meshes are disrupted inside the model (Image: Mesh 4). If this is a problem, how can it be avoided?

2) The solution graphs follow the same pattern and tend to infinity. The calculation was forcefully stopped after 500 iterations (Computation time of around 30 minutes) in the case of the k-omega flow model. If k-epsilon or SST model is chosen, the energy graph (light blue color) rises exponentially upwards, resulting in a "floating point exception error." Which is the right model to be chosen for my problem statement, and how can the above errors be avoided?

3) The fluid inlet temperature assigned is 303K, and the calculated outlet temperature is around 316K. However, as seen in the temperature profile image, the variations throughout the model are very extreme, ranging from -90K to +5000K, which shouldn't be possible as no heat addition is taking place and the walls are adiabatic as well. So what can be the solution to this issue?

4) Similar to Q3, the pressure contour values vary from -90Pa to +428 Pa. Also, even though the assigned inlet pressure is 101325 Pa while checking for the area-weighted average of absolute pressure (Results -> Surface integrals tab) at the inlet and outlet, the inlet pressure shows a higher value of 101450 Pa, and the exit pressure displayed is of 101325 Pa. In contrast, I want an inlet pressure of 101325Pa and exit pressure < inlet pressure (owing to the pressure drop).

5) The vectors for velocity were getting plotted in the solutions tab. However, I wasn't able to simulate the same. The sweep animation tab displays "No animatable objects found (planes etc.).", so there is no option for animating the velocity vector flow that was displayed

I hope I have explained my problems clearly in this post. I apologize I couldn't upload more images due to the 5 images limit, but I can surely attach any other image if required in the successive replies to this thread. Do let me know if some more explanation is needed to elaborate on the issues faced. Thanks!

[gupta29505]

Temperature equation is not converging so check your mesh and see why it is not converging

[Akkuboy]

That's exactly my query. Why isn't it converging despite the mesh looking uniform and fine throughout? The meshing in the cross-sectional plane view looks distorted despite the outside region's connectivity. How can it be improved? And should the solid and fluid regions have connectivity with each other or do both need to have connectivity on their own (like solid with solid and fluid with fluid)?

[gupta29505]

Solid/fluid need connectivity. You can send me your ANSYS Project archive at freelancer.cae@gmail.com and I can take a look?

[Gert-Jan]

Convergence is not only influenced by the mesh, but also b the physics and transientness. So do not only look to the mesh.

To me it looks like there are multiple fundamental things wrong in the model. I would suggest to go to the fluent forum and ask for help.

Unless you are willing to pay. That is what this forum is for. Then send me a personal message.

[gupta29505]

I am sorry, my email had typo so I corrected it but if you are still struggling - please send me your ANSYS project archive at freelancer.cae@gmail.com

[myselfahmad]

I specifically asked that. Even though the mesh seems to be uniform and fine throughout, why doesn't it converge? The meshing appears deformed in the cross-sectional plane image counter despite the outside region's connectivity.