[ags]

Hello all,

I am new to the forums and was hoping someone could offer some help with a model I have been trying to run. I have some basic CFD experience, but am certainly still a novice, and have been working at what I think should be a relatively straightforward simulation, but have been encountering some problems.

I am trying to model water flowing through a duct of constant cross section, with a known, constant flow rate. Flow enters through the inlet (at a known, specified temperature), and passes through a short, well insulated (no heat enters or leaves through the walls) section of the duct. It then passes through a heated section, where there is constant heat flux through the duct walls. The flow then exits the heated section, and passes through a short, well insulated (no heat in or out through the walls) section before flowing out the outlet.

My CFD modelling is not meant to be especially precise, and is mainly being done as a sanity check to get a sense of what temperature profiles to expect, so I have kept the geometry and meshing very simple. The duct is long and straight with a constant square cross section. The mesh is equally basic, with only one division along each edge of the cross-section perimeter, and divisions every centimeter along the length of the duct (image attached as mesh.jpg).

The first 10 cm of the duct is the well insulated entrance region (entrance_section.jpg), the next 120 cm are the heated section (heated_section.jpg), and the final 10 cm are the well insulated exit section (exit_section.jpg).

I set the inlet boundary condition to "mass-flow inlet", with the known mass flow and temperature. I set all the walls as walls with slip by manually specifying 0 sheer stress, because I am mainly interested in the temperature results, so a simple uniform flow is fine for my purposes. The walls of the entrance and exit regions are set to have 0 heat flux, and the heated region is set to have a constant heat flux through the walls. The outlet boundary condition is set to "outflow".

I enabled the energy equation, and employed laminar viscous flow modelling.

When I run the calculation, the results are puzzling. The flow enters at the entrance temperature, as expected, and heats up as it passes through the heated section, but what is puzzling is that it cools down as it passes through the exit section, returning to essentially the same temperature as at the entrance. A temperature contour image is attached as contour.png.

This doesn't seem correct to me. As the flow leaves the heated section, it is passing through another insulated section, and I don't see how it would be cooling down. It seems as if a temperature is being enforced at the outlet boundary, but I was not given any option to specify one when setting the "outflow" boundary condition. I feel as if I should need to specify a temperature at this boundary, since heat conduction is described by a second order DE, so I should need a boundary condition at both boundaries but there was no opportunity to input one.

I have tried imposing several other types of boundary conditions at the outlet:
- Mass-flow-outlet: I input the same flow rate as the input, but there is no option to specify an exit temperature, and the results look the same as the earlier case.
- pressure-outlet: This does give the opportunity to enter a "backflow total temperature" but no matter what value I enter the results once again look the same.

Can anyone see something that I may be doing wrong? Or is my interpretation of the results incorrect?

Any help would be much appreciated!

[LoGaL]

So the section of the pipe is made of only one cell? Seems so from the image

[ags]

Originally Posted by LoGaL

So the section of the pipe is made of only one cell? Seems so from the image

Correct. I may make full 2D or 3D model at some point, but for now I'm really only trying to get a sense of the 1 dimensional temperature profile.

[LoGaL]

No man, that is not how it works, CFD is always 2D or 3D, you will never get anything good out of that pseudo-1D thing. The solver is trying to solve 3D equations on that. Just forget about “gaining the feeling with 1D” and run a 2D or 3D simulation with a decent amount of elements. Especially a 2D computation will be fast enough (maybe 5 min calculation time) to gain the feeling. To obtain 1D temperature profiles just take section averages of the quantities