[axry]

Hi all,

I'm a student performing an assignment designing a heat exchanger in Fluent using ANSYS, and I am encountering some difficulties in the solution stage. I'm fairly well-versed in this particular case, but I don't know Fluent well enough to understand some of the things I'm doing, so I may be missing something critical in my model.

**********

Outline: I have designed a multi-pass heat exchanger with a coolant core and a hot shell, with a wall separating the two fluids (there is no shell wall). I am trying to design this exchanger using SolidWorks, with the tube, water and oil modelled as separate bodies of the same part (see link below). For the sake of computing time, I have halved the model using an axis of symmetry (which is also modeled in ANSYS).

SolidWorks screenshot

**********
I am using Fluent to determine heat transfer in the system to ultimately determine whether or not the hot fluid outlet temperature is below 60 degrees C.

I have meshed the model to the best of my ability (see link below), but i don't know if it is too coarse/fine (if there is such a thing).

Mesh

The mass flow rate of the water means that a turbulence model is required alongside the energy equation. The turbulence model selected is detailed in the link below.

Turbulence overview


Boundary conditions are set with two counterflow mass-flow inlets, and two corresponding default pressure outlets.

When I run the solver, there are usually a few iterations which appear to follow a normal trend, but the energy residuals always rocket up eventually, causing 'Temperature limited to X' errors (see link below).

Solver limits

I have tried many methods of preventing this error, and replicating the results from a model which does work, to no avail. I suspect a meshing problem but I cannot see where the issue may lie.

I have also tried looking up the issue, and even tried the solution detailed toward the bottom of this thread: click here. However, it has not worked (although I'm unsure if I did it properly).

I know this probably won't be a simple scenario to understand, but I've tried to explain the best I can. I can upload any files that would be helpful for anyone who would like to help. I'm really tearing my hair over this as it has been a 3-day struggle up to this point!

Many thanks
Alex

[parvaz747]

Hi alex,
I have some advise:
1. Turn of the energy equation an run the problem without solving energy equation, if there is a problem it's not related to energy equation
2. reduce under relaxation factor for energy and momentum equation (0.1 at first) and increase it gradually
3. Use hybrid initialization

[axry]

Hi,
I tried running without energy equation and the solution worked. However, the energy equation is important! I also used hybrid initialisation, and this did not help the problem.

I will try step 2 when I return to my workstation tomorrow.

Thanks,
Alex

[axry]

Originally Posted by parvaz747

2. reduce under relaxation factor for energy and momentum equation (0.1 at first) and increase it gradually

I have successfully converged the model by reducing the energy URF to 0.5. Thank you!

[parvaz747]