CHT heat exchanger power flows, primary and secondary thermal power don't align

Using SimScale
1

Hi,

I am currently on evaluation of Simscale at work and i am trying a couple of typical applications we have for CFD.

Currently on a finned air cooler and its just finished the simulation., its community plan so public link here. There’s two flow regions, both internal, one is air, the other water.

The flow regions each have a single inlet and outlet only, no walls as I believe no wall is treated as adiabatic.

I set it up as follows:

Air inlet
Velocity inlet
Flow rate
Volumetric flow
0.018m^3/s
22C

Air outlet
Pressure outlet
Mean value
101325Pa

Cooling water inlet
Velocity inlet
Flow rate
mass flow
0.011kg/s
3C

Flow is counter flow, chilled water inlet is at the air outlet

I had done some excel calcs on the expected performance:

Prediction
0.018m^3/s air in the duct corresponds to 3.22m/s
Estimated heat exchange 214W
Water discharge 7.7C
Air discharge 12C

Results:
Air speed average 3.33m/s Correct
Air flow 0.0178m^3/s Correct
Air discharge 19.72C higher than expected
Water discharge 19.87C higher than expected
Water flow rate 0.0108kg/s Correct

Air inlet
image
image

Air outlet
image
image

Water inlet i am confused why the average temperature is so high when i specified the whole face to have an in flow at 3C
image
image

Water outlet
image
image

Working out the power:
Water change in energy = +716.4W
Air change in energy = -73.2W

It looks like there is a clear factor of 10 difference here but i cant see at all where it is.

I ran another simulation at 0.1kg/s water flow , the results followed the same pattern:
Air speed average 3.33m/s Correct
Air flow 0.0177m^3/s Correct
Air discharge 18.59C higher than expected
Water discharge 14.88C higher than expected
Water flow rate 0.0987kg/s Correct

Working out the power:
Water change in energy = +4957.5W
Air change in energy = -73.4W

Things are certainly going a bit wrong here, again, just not seeing where.

Hopefully somebody can see where im going wrong :slight_smile:

1 Like
2

Hi, and thanks for using the forum!

I will check the simulation and the estimations, let’s see if I can find where the devil is!

1 Like
3

Hi,

Here are some pointers on how to improve your results:

  • Improve the initial temperature condition for the different parts, to help speed up convergence. Specially for the water domain, you should use a value closer to what you expect to obtain.
  • To improve convergence, please run for more iterations, say up to 3000. Refining the mesh will also help, but I would leave that action for later.
  • As for the measuring of the resulting temperatures, please use some result control items on the inlet and outlets, instead of relying completely on the graphical post-processor:
1 Like
4

Hi,

thanks for the pointers, i had mis-understood the meaning of the initial conditions. So the dT i was getting on the water, was actually the water cooling down the copper mass i suspect?

I think now understand, so ive set:
Global temperature to 10C
Water fluid to 6C
Air fluid to 18C
Copper solid to 6C

I had wondered if i change the time step from 1s to 10s if that may also help? I am not sure if 3000s would take 3x longer to run as i have a limit on runtime under the community plan.

1 Like
5

Changing the time step will not make a difference. Bear in mind that this is an steady-state simulation, and that this ‘time’ is only a variable to control the convergence loop, and not real simulated time.

1 Like
6

Looking at Run1 & Run3 (within initial conditions) convergence seemed to be occurring reasonable well at 1e3s for Run1 and generally Run3 was showing similar levels of convergence at 3e3s. Interestingly the water T Temp continues to dive with no plateauing or indications of stabilisation, but this may be a function of the flow rates.

Good luck, an interesting project, sorry no insightful comments only to say simulation indicators for convergence (except for generally water T) showed convergence after 2e3. Oh, if you are on the Community plan and monitoring your results in real time, you can terminate the simulation if you are not happy with the expected result.

7

Hello,

thanks for your help - i have now got numbers which are more in line with what I had expected, certainly a great tool!

2 Likes