Altair RISE

A program to recognize and reward our most engaged community members

Nominate Yourself Now!

Issue with boundary conditions in Flux 2d thermal steady state analysis

I am simulating a 3 phase PM Outrunner motor (52 poles, 48 slots). I am trying thermal simualtions in flux, after all the setup as mentioned in examples and tutorials, this is the error i get upon physics check:

Also, I tried changing geometric symetry to even cyclic (which is incorrect as per no of slot and poles) but even then it results in following error:

Kindly help in resolving this issue.

Find more posts tagged with

English

Flux

v2020.0

Accepted answers

Letizia Ferrara

Hello Piyush,

As Simon wrote, boundary condition has a different meaning between magnetic and thermal physics.

Unfortunately the error is not clear, but the periodic mesh line must be assigned manually in the model if you do not have the infinite box (see attached image) and in the thermal model this box is not necessary as the boundary conditions can be imposed through exchange-type line regions.

I uploaded the working model to the support ticket you created, thanks for that.

If you still have doubts don't hesitate to contact me trough the support form.

Best regards,

Letizia

periodic mesh line.png

All comments

SimonGuicheteau

Hello,

Switching to an odd periodicity should solve the first issue.

Then the error seems to say you don't have a linked mesh. An easy fix could be to complete the infinite box again, which will set the linked mesh again.

If it does not solve it, it means you may have a slightly different geometry on both sides (lines and points). I guess it should be ok as it looks like it comes from FluxMotor initially.

So complete the infinite box again, reassign the faces to the correct region (infinite region probably), mesh the domain, and check the physics again

Piyush Verma

Hello,

Switching to an odd periodicity should solve the first issue.

Then the error seems to say you don't have a linked mesh. An easy fix could be to complete the infinite box again, which will set the linked mesh again.

If it does not solve it, it means you may have a slightly different geometry on both sides (lines and points). I guess it should be ok as it looks like it comes from FluxMotor initially.

So complete the infinite box again, reassign the faces to the correct region (infinite region probably), mesh the domain, and check the physics again

Hi thanks for your response there are few issue with that:

1. Flux thermal does not allow me to give odd periodicity, physics check pass only with even cyclic periodicity.

2. Even after even periodicity and adding infinite regions back, i get the following error:

SimonGuicheteau

Hi thanks for your response there are few issue with that:

1. Flux thermal does not allow me to give odd periodicity, physics check pass only with even cyclic periodicity.

2. Even after even periodicity and adding infinite regions back, i get the following error:

Hello,

1. Yes, in thermal, an odd periodicity does not mean anything. it needs to be cyclic/even. Otherwise, it may mean you need to have half of the machine or even the full machine if this does not suit you because of the magnets and coil potential differences. I think Even periodicity is just fine.

2. I can't really comment on this error. Please have a look at the Flux2D.report file to understand the true error.

You may need to remove mechanical sets as well as the thermal won't need to rotate. That is just a guess.

Letizia Ferrara

Hello Piyush,

As Simon wrote, boundary condition has a different meaning between magnetic and thermal physics.

I uploaded the working model to the support ticket you created, thanks for that.

If you still have doubts don't hesitate to contact me trough the support form.

Best regards,

Letizia

periodic mesh line.png