Joule heating

DaveStuppleDaveStupple
Trying some initial simulations to model heating in a resistive filament. I have tried with solid hex elements, but the resistor option is not available. Tried with line elements (both with and without finite section thickness), I gave them a resistance, I get potential and current solutions, but adding power density to the solution gives a red entry in the tree: Compatible with element types and materials: unknown. What am I missing? Thanks.

Comments

  • VictorVictor
    Power density in the solution is currently only available for solid and shell elements. You should be able to extrude the line mesh into a single element wide strip of narrow shells.
  • DaveStuppleDaveStupple
    Works fine now. I was confused that the resistor button was not available, but I just used the conductivity property. Applied this to a trivial problem of filament temperatures, mainly because I wanted to see if I could make the helical mesh. It would be a nice touch if I could adapt the theory for two parallel plates radiating to each other, so that I can approximate the effects of adjacent coils heating each other. I don't think it would change much, but I haven't done the sums.
  • prop_designprop_design
    what did you mesh that with
  • barminbarmin
    That kind of mesh can be obtained by Salome for example :)
  • prop_designprop_design
    thanks I'll have to give it a try. does it do automatic hex dominant meshing
  • VictorVictor
    CCX has the capability of modelling radiation between surfaces. Last I checked, it was a bit complicated to set up by hand so it's something I want to add to the UI eventually. It could be that you have to copy the generate *RADIATE keyword block and add the CAVITY parameter.
  • DaveStuppleDaveStupple
    edited April 2019
    Thanks Victor, I'll take a look at the dreaded CCX manual.

    Prop design, I know not much about Salome. I have it installed and have had a go, but I found it quite confusing. I believe it does do hex dominant. It looks like a useful tool.

    These lovely helical meshes were done in Mecway 11, and were pretty straightforward. I'll try to remember what I did. To create the helix I used >Mesh tools >Create >Curve generator >Cirlce. Enter two equal radii for the overall diameter of the helix, then OK, which will return you to the Curve generator window with the parametric equations to generate a circle. Change the Final p entry from 2*pi to, for example, 20*pi to give 10 turns, then in the Z box enter some small value *p to move the helix along the z axis, such as 0.003*p. Increase the Number of elements figure to give a sufficiently smooth curve. Hit OK and you should have your helix made of line elements (subject to a bit of tweaking of the figures). Next >Mesh tools >Create >Plate mesh and enter the radius of the wire, and -if you like- adjust the meshing parameters (with a live preview, I think). This gives you the cross section of the wire. I moved this cross-section so that its centre was on the end of the helix, then rotated so that it was oriented normal to the end of the helix. Merge the two nodes at the centre with the Merge nodes tool. Next use the sweep tool: for the profile select your cross section, and for the path select the helix (I had these as named selections to make this easier). Hit OK and you should have have your helix. I think you can avoid having to move the cross-section to the helix by linking the helix end node and plate centre node with a line element, but you will have to delete the elements that the line elements create as they sweep.

    I think this should work, but if I have misremembered let me know and I will work through it. Pretty impressive built-in meshing tools, I think.


  • MartinBMartinB
    edited April 2019
    In reply to an earlier post, Salome Meca is great for modelling and making hex dominant (or tet) meshes. Look for code-aster-windows if on windows..Export as UNV to continue in Mecway, or try your hand with the code aster solver (not for the faint of heart :)
    But I'm impressed you've meshed this with Mecway!
  • prop_designprop_design
    Thanks guys. Prior to your replies I tried Salome. I also tried gmsh again. I have tried that several times and I just can't stand it. Salome was equally terrible. The windows download is ridiculously huge. It takes up a lot of hard drive space once it finally unzips and the gui was awful. So I promptly deleted it. I prefer Netgen and Mecway. But it would be nice to have auto hex meshing. Cool that you made that model in Mecway. It's a nice looking model.
  • SergioSergio
    Even if Salome use Netgen, there are some tools not available on Netgen or Mecway for meshing. I agree that takes several steps to acomplish a simple task that could be done easily on a comercial program, but there is nothing better in the open source world. In my case I use a lot for extructured hexa meshes, combined with Mecway to made the extrusions/revolutions or create groups.
  • tk1537tk1537
    I have also used Salome to create hex meshes on STEP solids, then import the mesh into Mecway for contraining and solving with good results. You have to fiddle a lot with settings, partitions, 2D and 1D mesh controls (for 3D) and submeshes for regions, but nice hex meshing is possible. Gmsh is saying that they plan to incorporate Hextreme into v5.0 when released, which is supposed to produce 1 step solid hex meshing. We'll see.
  • prop_designprop_design
    I can't figure out the Gmsh interface. I would love auto-hex meshing though. That would make it worth trying to learn.
  • mhoulroydmhoulroyd
    Hello everyone,
    I am trying to set up a Joule heating simulation to display maximum radial heat in the center of a small brass wire. I have worked out the math, and know the answer, I want to learn how to simulate it. I have read the manual but still my simulations generate values which do not make sense. I've searched for online examples and tutorials but haven't had success.
    Here's what I'm trying to do::
    Starting with a small brass rod 1mm dia. x 5mm long, I am applying 0.42A. Ta=ambient 20C , h=10, K=109 Tmax should be approximately 50C
    Using temperature dependent material, setting parameters t using DC Circuit, switching to Thermal transient, I've tried thermal steady state too. My results display a radial heating, and the center is the maxT, but the values make no sense. I am obviously setting this simulation wrong. Any advice would be helpful and appreciated. If I can get my simulation to work, I will gladly share it to help other novices like myself. thanks for reading. - Kind Regards, Marty H
  • disladisla
    Starting with a small brass rod 1mm dia. x 5mm long, I am applying 0.42A. Ta=ambient 20C , h=10, K=109 Tmax should be approximately 50C


    ¿Could you detail the Brass material properties units you are using?
  • mhoulroydmhoulroyd
    Density 8730 kg/m3
    Youngs Modulus 112 GPa
    Thermal Conductivity 109 W/m3
    Poisson ratio .311
    Thank you for helping me
  • disladisla
    edited May 15
    Thermal Conductivity units are W/m/K
    What about your electrical properties.?
    How did you end up in this value of "Tmax should be approximately 50C".
  • mhoulroydmhoulroyd
    yes, I made a typo on the thermal conductivity units.
    Heat flow: -5.67W
    heat flux: -6.71e+6
    Resistance 2.74e-2 ohm
    Qin 4.83e-3 W

    My math values were calculated on a spreadsheet and when inputted example
    joule heating problems match answers.
  • disladisla
    You need to put all your electrical and thermal material properties right into place first.
    Clean and with their units.
    You started the post saying something about h=10 which now seems that has disappeared, You are mixing material properties with results.
    Be sure you understand the difference between Electrical Resistivity and Electrical Resistance. They are only the same for Length/Section =1
    Please work it out a little and if you do not advance come later.



  • disladisla
    edited May 16
    I have work Joule Effect on Mecway based on an example problem very similar to yours (more complete) that has analytical solution to compare.
    It solves the nude and insulated wire. It could eventually help to validate your Setup and Procedure. Read the assumption closely.
    The problem and its solution are free access. Theory and Solution are described in the well known book :

    Fundamentals Of Heat And Mass Transfer Frank P Incropera
    6th-Edition. (Be sure both Theory and problem books are 6th edition. If not Example reference number will not match)
    PROBLEM 3.42

    Both can be downloaded free here. First check if your custom spreadsheet numbers also match the solution.

    https://archive.org/details/FundamentalsOfHeatAndMassTransferFrankPIncropera
    https://archive.org/details/SolutionManualFundamentalsOfHeatAndMassTransfer6thEdition/page/n7/mode/2up?view=theater

    I can help you based on that problem data.
  • disladisla
    I completed the exercise with very good agreement.
    I don't do electrical FEA at all so probably my first part of the analysis to obtain the internal heat generation could be improved or speed up.
    Regarding the thermal analysis I had to do it with thermal transient because Radiation is not available in Steady State.
    I have used the same iterative process as in the book to find the heat transfer coefficient Wire/Air. Both, naked and insulated solutions have converged to the expected surface temperatures.
    Caution with the units and properties (Resistivity/Resistance , Conductivity/Conductance)
    Anyone identifying potential issues with this approach or having suggestions to improve it please comment.

    Questions:

    -The electrical analysis doesn’t give the current value. It shows in red. Don’t know what’s Wrong. ?¿?

    I hope it can serve as reference example.





Sign In or Register to comment.

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Sign In Register