Effect of region on force calculation

I implemented the same example as 'Instantaneus Forces on Busbars in Maxwell 2D' in Ansys Maxwell Field Simulator v15 users guide. I realized that when I change the size of the region in the example, the force results also change. What size should I do in order to get the right force result? What is the rule of this? How can I find out the correct region size?

Could you please help Thanks.

note:Example is on eddy current solver.

Best Answer

  • nchodenchode Forum Coordinator
    Accepted Answer

    I suggest you run a parametric starting from 2r to 20r region size ..

    At some regions, maybe at 5r, hopefully, you will not see the difference in the result after that.

Answers

  • nchodenchode Forum Coordinator

    Hi,

    Are you reducing or increasing the region size?

    As long as the region is sufficient enough to cover the whole field distribution you should get the same result for anything Region size above that region size.

    But if you are reducing the computational region such that it is not fully enclosing the whole magnetic field you will get a difference in results.

    You can plot the fields and check if all the dominant filed lines are enclosed in the region to figure this out.


    Regards

    Navya

  • Hi nchode,

    Force change based on region y size;

    +Y&-Y: 50...F=154kN

    +Y&-Y: 200...F=51kN

    +Y&-Y: 300..F=33kN

    +Y&-Y: 500...F=17kN

    +Y&-Y: 800...F=9600kN

    +Y&-Y: 1000...F=9986kN

    +Y&-Y: 1500...F=12515kN

    +Y&-Y: 2000...F=13615kN

    +Y&-Y: 2500...F=15600kN

  • nchodenchode Forum Coordinator

    @johnfides 

    I tried doing a similar instantaneous force setup using the eddy current solver. I did not find any change in the force magnitude with an increase in the region size.

    Can you please confirm if you got the same result as shown in the tutorial if you follow the tutorial?

    If it is not, please share the comparison screenshots of the tutorial with your result.

    Also, confirm which version of AEDT are you using?


    Regards

    Navya

  • johnfidesjohnfides Member
    edited January 6

    Hi @nchode

    I am sharing the picture of the model I am working with in the attachment. The model I am working with is slightly different from the model in the example, but I stimulate and analyze in the same way. I use ansys maxwell 15.


  • nchodenchode Forum Coordinator

    Hi johnfides 

    Can you provide more details regarding your model?

    How much is the current excitation? How is your setup? how are you assigning force parameter? boundary conditions etc.,

    Force parameter should never vary with an increase in the region size.

    Without knowing the specifics of your model it is difficult to assist.

    I suggest you do this exercise starting with what is exactly in the tutorial and increase the conductor number in steps of one to understand the problem.


    Regards

    Navya

  • johnfidesjohnfides Member
    edited January 6

    Hi,

    In my model, I have actually defined a force parameter for only 2 groups, not for all sections as below. I select each section one by one and apply current. my current is 9kA. my setup;(eddy current solver)

    max number of passes: 20

    percent error: 0.01

    refinement per pass: 30

    min number of passes: 2

    min converged passes: 1

    solver: 50 Hz

    nonlinear residual: 0.0001

    freq sweep;

    start: 10 Hz

    stop: 1000Hz

    step: 10 Hz

    I assigned boundary as a vector potantial value:0, Phase:0


  • nchodenchode Forum Coordinator

    Hi,

    can you first confirm whether you are getting the same result as the tutorial with a single conductor?

    And is the result varying with region increase in that case?

    Just want to make sure what you and I got is the same.

    As I said, I did not face this issue with region change for a single conductor.

    I do not have Maxwell 15 installed, that very old version.

    You can also try installing a latest version of AEDT (2020 R2)

  • Hi,

    Is it normally like this, as the region increases, the force decreases, and then the force value remains constant, no matter how much the region increases.?

    Thanks,

  • nchodenchode Forum Coordinator
    edited January 6

    The problem region is the region where Maxwell calculates the field solution. It is just enough that Region size (solution space) be chosen such that the respective region walls are not close to field sources(current or permanent magnets). This allows us to assume the field is a tangent at the region limits. The region must be given sufficient room for fringing.

    For example, if the magnetic fields are well-contained, as in motor, the padding can be 20-50%. However, if the magnetic fields are not well-contained, such as two parallel current-carrying conductors in air, the padding may need to be 1000% or more.

    *If the field results change significantly when the region size is changed, the region is not large enough.  

    If you have this create region option in Maxwell 15 use this option and pad all the directions to around 1000%. then you should not see this issue.

    You can check by increasing padding all the way to 2000% - you should not see any difference in result.




  • Hi,

    I will try what you say. Thanks for your patience and help.

    I read this in ansys maxwell help;

    Typical boundary conditions used in eddy current problems include magnetic field tangent (the default, natural boundary condition which is automatically applied on all surfaces of the problem space -the surfaces of the geometry entity containing the model inside it). This default boundary condition can be overwritten if other boundary conditions are applied on exterior surfaces of the solution space. The default boundary condition confines the magnetic field to the solution space and thus this boundary must be placed at some distance from the sources of the problem to avoid over-constraining the fields by placing the boundaries to close to the model objects. While it is dif­ficult to provide "recipes" with general validity on the placement of the boundaries of problems, a good rule of thumb says that if a model can be imagined as being contained in a sphere of radius R, the boundaries can be placed at a 4-5 radii R from the imaginary center of the model.

    According to this explanation, is the radius definition that I will create for the region that I show in the picture below?


  • nchodenchode Forum Coordinator
    Accepted Answer

    I suggest you run a parametric starting from 2r to 20r region size ..

    At some regions, maybe at 5r, hopefully, you will not see the difference in the result after that.

Sign In or Register to comment.