13 Replies Latest reply on Dec 10, 2014 3:10 PM by Chris Michalski

    Problem with Flowsimulation

    Henning Wagner

      Hello at all!

      I am new in the forum and i am a student who uses this Software for the firt time.


      I have got a problem with my Flowsimulation results.

      If I difine all materials, e.g. aluminium and PTFE, the fluid won't become hotter during the simulation.

      Fluid temperature is 20 °C at the inlet and 20 °C at the outlet. It seems that the simulation doesn't change the fluid temperature and only the material temp.


      If i don't difine the materials it is working better.The fluid temperature rises from 20°C to 26 °C.


      Which method ist the right one?

      I want to simulate a liquid cooler. The results should regard the fluid temperature and the solid temperature.


      Can someone help?


      Thanks in advance!


        • Re: Problem with Flowsimulation
          Chris Michalski

          Have you started by going through the tutorials?  They are the first place you should look for answers and knowledge on how to setup flow simulations.

          Assuming you have already worked through the tutorials and understand the basics of the software, you can post a screen shot of the geometry as well as the feature tree of how your model is setup to help locate your problem.

            • Re: Problem with Flowsimulation
              Henning Wagner

              Hello Chris,


              thanks for the fast response.


              I have done all tutorials a could find.


              Here are the screenshots of my model and the feature tree:




              If I define the materials i am getting wired results:


              Here is my fluid subspace:


              I am sorry, but it is the german version.


              Best regards,


                • Re: Problem with Flowsimulation
                  Chris Michalski

                  So you have the default material as aluminum.  Which parts did you add PTFE to in the version that did not alter the outlet temperature?

                  Have you looked at the mesh to verify that your internal serpentine features are being resolved?  Without a local initial mesh or high refinement they are likely lost (although that would be the same for both materials).

                  And the green blocks (I presume aluminum) are conducting the heat into the system.

                  Did you follow the convergence of the goals?  Or did you leave it as the default of 4 travels maximum before stopping?  I see that you have most of your goals suppressed in the image.

                    • Re: Problem with Flowsimulation
                      Henning Wagner

                      The "red" inlet is PTFE.


                      I have forgotton the adjust the mesh.

                      Now I have activated the feature to adjust the mesh in small spaces.

                      Maybe it is getting better.


                      If not, how can I define a mesh in the serpentines/channels?


                      Or should i define a very small mesh, e.g. 1mm x 1mm ? The channels are 2mm deep and 3mm wide.

                        • Re: Problem with Flowsimulation
                          Chris Michalski

                          Because of the size of the aluminum block you want to use a local mesh.  Otherwise your entire part will have a very fine mesh and it will take very long to solve. 

                          Create a solid body that is slightly larger than the area with your serpentine.

                          When you create the local initial mesh, select this body and apply the settings.  Then either in the local mesh settings or through Component Control, disable these solid bodies.  In this way Flow will refine the cells inside of that body, but then ignores that body so that it does not affect the flow.  (if you do not disable the part it will remain a solid and prevent fluid flow).


                          So the PTFE inserts are purely to restrict flow and increase the contact time/distance between the fluid and the green aluminum plates, correct?


                          If it is acceptable as a start I would simplify to a single heat sink (1/3 of your system) and assume 1/3 of the water flow rate.  This way you can make the mesh much finer without greatly increasing the time to solve.  Assuming you have sufficient flow restriction that your manifold produces uniform flow through the 3 serpentines this may be a valid simplification.

                            • Re: Problem with Flowsimulation
                              Henning Wagner

                              Hi Chris,


                              thanks for your response.


                              Now i have created a little cube above the serpentines and adjusted the mesh with the local mesh settings.



                              It seems to be much better. But the first results are not very good. My liquid temperature rises only a little bit.


                              I haven't got the Flow Simulation Add-In installed on my Notebook, so i will post a picture of the mesh later.


                              You are right with the PTFE part.


                              Thanks for the idea to simulate only a part of it. I will give it a try.

                    • Re: Problem with Flowsimulation
                      Jared Conway

                      are you sure the PTFE materials aren't basically an insulator?


                      if your default material is aluminum, if you don't apply any materials, everything is aluminum


                      if you then go and apply ptfe to some components and the temperature no longer changes, it could be because they are insulated.


                      to get some confidence with the software, i'd recommend making some much smaller simulations like just a tube to get comfortable

                        • Re: Problem with Flowsimulation
                          Henning Wagner

                          Hello Jared,


                          thanks for the response.


                          I will define all materials in the next simulation.

                          I have tried smaller simulations and it worked fine.

                          Only one problem appears, the temperature of the liquid didn't rise in some cases.

                          I will post a picture later.

                          Maybe the flowrate was to high?!


                          Do i have to define the wall conditions between the solid materials and the fluid?

                            • Re: Problem with Flowsimulation
                              Chris Michalski

                              The beauty of Flow is that you don't have to define the properties where solids and fluid meet, it will calculate the heat transfer coefficient and such based on the materials in contact, velocity, temperature, etc.

                              My guess would be that the PTFE part made the mesh a mix of solid and fluid cells (mostly partial cells) so the heat transfer was reduced to a blend of aluminum and PTFE based on their ratio.

                              I might start by making an alternate version of the PTFE where you eliminate the vertical serpentine paths.  Keep the horizontals so you have multiple parallel flow paths, but eliminate the sinuations so that it is easier for Flow to accurately recreate the PTFE as a solid.

                              In this case you also could have tried skipping making an extra body for the local mesh and just using the PTFE so that any cell that touches it gets refined.

                                • Re: Problem with Flowsimulation
                                  Henning Wagner


                                  here is my mesh after the optimisation:



                                  It seems to be very good.

                                  After 233 iterations i can see a rising of the liquid temperature.

                                  But now i have got ~500000 fluid cells, is this a normal value?


                                  Best regards,


                                    • Re: Problem with Flowsimulation
                                      Chris Michalski

                                      That looks reasonable for the mesh.  When you look from the face I would assume you can see each of serpentine walls.

                                      500k fluid isn't unusual.  I typically try to refine my mesh so that my total is around 2M cells.  For a first approximation I might dumb things down (total <1M cells) to get a rough solution in a shorter time to make sure all of your inputs and settings are correct, then refine the mesh for accuracy.

                                      This is part of why i suggested cutting your model into 3rds and assuming uniform flow - it would drastically reduce the convergence time.

                                      I often determine the location of maximum interest and have Flow maintain a plot of that section.  In this case I would make a plane halfway between the PTFE and the aluminum.  Then you could watch the temperature of the water as it flows along that face.