|
[Sponsors] |
March 7, 2021, 12:32 |
Coupled analysis with Peltier module
|
#1 |
New Member
Valaki Valami
Join Date: Sep 2018
Posts: 14
Rep Power: 7 |
I would like to modell coupled system with Peltier module.
The Peltier is modelled with Mechanic, other parts with Fluent. Hot side of the module is on heatsink, it has a fixed wall temperature of 280K. Cold side of the module is on PCB with air, enclosed with metal walls. See the attached picture. After some coupling iteration I get the known error temperature limited to 1.000000e+00 in 57967 cells on zone 17 in domain 1. So I started debug the problem and I found out that after ~10 coupling iterations the Total heat transfer rate (found on Fluent->Monitors->Report Plots->Flux Report->Total heat transfer rate->Chose the Peltier surface) is gets very big some 10 W, see attached picture. I don't know why that problem appears. Can anyone have any idea? Thank you. HeatTransfer.jpg SchematicSketch.png |
|
March 9, 2021, 03:20 |
|
#2 |
New Member
Valaki Valami
Join Date: Sep 2018
Posts: 14
Rep Power: 7 |
I further investigated the problem and found a solution of reducing the Under Relaxation Factor to 0.05 or even smaller value.
My theory is that the Fluent cannot handle the arising heat shortage (from the Thermal-Electric solver) on the cold side of the Peltier because the thin (0.8mm) FR4-epoxy and air there, have a very very small amount of heat capacity and conductivity (850J/Kg*K; 0.29W/mK for FR4). Is that a correct approach of solving that problem? 0.05 for Under Relaxation Factor seems to me very small value. Is there possible another solution? if I put a thin (0.1mm) thermal paste layer (with high thermal conductivity) between Peltier and FR4-epoxy can help? Interface.jpg |
|
March 9, 2021, 13:39 |
|
#3 |
New Member
Valaki Valami
Join Date: Sep 2018
Posts: 14
Rep Power: 7 |
On the first approach, I reduced the Under Relaxation Factor until the solution doesn't converge. I earn convergency at value 0.045 with 1000 iterations which is so so much.
One of my ideas is to put a thermal paste layer between the Peltier modules (which I solve in Thermal-Electric solver) and the PCB (which I solved in Fluent). With that approach, the solution started to converge at Under Relaxation Factor 0.2 with about 300 iterations which is much less than in the previous solution. See the attached pictures. THTR.jpg THTRwithThermalPaste.jpg |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Generating Stack Trace while solving | johny_walker | OpenFOAM Running, Solving & CFD | 1 | January 29, 2021 03:48 |
InterFOAM water channel case Error | mkjmalik | OpenFOAM Running, Solving & CFD | 4 | August 23, 2020 11:32 |
ANSYS WORKBENCH Transfer result of one analysis to a new analysis as preload | ingjuanm90 | ANSYS | 0 | July 26, 2016 14:04 |
Can someone PLEASE document the development version installation | bernd | OpenFOAM Installation | 76 | November 14, 2008 21:51 |
Coupled 1D/3D STAR-CD Training | CD adapco Group Marketing | Siemens | 1 | November 13, 2002 15:48 |