simple coil with core
Altair EmployeeHello,
I started to work with Flux. I modeled an unmeshed coil with an iron core inside. I used the application ‘’magneto static 3D’’ and the 'automatic formulations'. The dimensions of the coil are Lcoil=0.4m (length), Rcoil=0.05m (radius) und N=1600 (number of turns). The dimensions of the core are L=0.4m, R=0.02m, myr=1200 (reltive permeability of the material). In Flux the material is described by the magnetic property 'linear isotropic'. The current through the coil is I=0.5A. To verify my solutions I made a calculation for the magnetic flux density and the magnetic field in the middle of the coil with the simple formula: H=I*N/sqrt(Lcoil2+(2*Rcoil)2)=1940.285A/m and B=myr*my0*H=2.926T. But with my simulation model I get the values H=110.58A/m an B=0.167T. These are high differences. Does somebody know what my fault could be? I checked all input Parameter and my calculation. When I modeled the same coil without a core the calculation accorded with the simulation.
Thanking you in advance.
Answers
-
Hello,
If you use non-meshed coil, you should check that the mesh on the external part of the coil is refined enough. Usually non-meshed coils are used when the flux density is centered in magnetic materials. If most of the flux is in air, you should either improve the mesh around the non-mesehd coils, or use meshed coils.
I hope it will help.
0 -
Hello,
Thank you for your quick respond.
I tried both of the alternatives: An unmeshed coil with a refined mesh arround it and a meshed coil. On both of them I become nearly the same result. But it doesn't accord with the calculated result above.
Did you ever tried to model a simple case, where you can calculate the right result on your own?
I don't know if there is a thing I forgot to consider in my calculation or if I forgot something important in my model. But I tried this simulation like a 100 times and every time it doesn't accord and I get the wrong result like above.
I also tried a parameterized study where myr is variable between 1 to 1200. When we take a look on the formula for H on the first post, then H always has to be the same value if we change nothing on the coil (because the material is only described by myr , so it can't have losses or anything else). Is this assumption right? But in my parameterized study H falls with increasing myr. Do you understand why?
Thank you very much.
0 -
Hi,
We need to look at your Flux file to see the error or defect.
Please contact the support Team (flux_support_emea@altair.com)
Regards
Yann
0 -
Hello,
You can try from the supervisor to run the following example. It is an Helmholtz coil, where we use non-meshed coil, and we compute magnetic field and compare with analytical formula. To run the example on your computer, you can follow the next steps:
- open Flux supervisor
- select the [3D] tab (on top)
- select [Open Examples] (on the left)
- in the [Exemple Tree] (in the middle), select [Helmholtz coil] in the [Application Notes] part
- if you double click on [readme] it will open a pdf file explaining step by step how to build the project
- if you click on [Helmholtz coil] (after the blue button), and double click on [Postprocessing], it will create a new folder, and will create a full Flux example
I hope it will help and show that we have compared Flux to analytical formula with non meshed coils.
0 -
hello,
thanks for your respond. Unfortunately my example tree is empty. So I looked in the directory 'FluxDocExamples_12.13' after Helmholtz coil, but I can't find it. Do you know the path to this example?
0 -
Hello,
You seem to have an old Flux version (Flux 12.1 instead of Flux 2018, the current one). I can propose you one version of the Helmholtz coil example, and it seems the python files are from 12.1. So it should be working in your case. You should follow the next steps to activate it:
- unzip the file in your working directory
- open the flux supervisor
- go in the working directory, and select [HelmholtzCoil/HelmholtzCoil3D_Case1] as working directory
- select [Python scripts] (on the left)
- click on [Main.py] (on the right)
- click on [Run the selected script] (on bottom left )
It should run Flux and create the full project until post-processing. When unzipping the archive, you will also see pdf files with explanations.
I hope it will help.
0 -
Hello lombard,
Thank you very much for this extensive explanation. I understand this really good.
My Problem is when I model sth. with an core inside, like my simple coil. I don't understand how flux calculates, when there is a material inside the coil. Because without the material, I get according results to my calculations ( like in your example with the Halmholtz coil or in my example with an normal cylindrical coil). As soon as I model an core inside the coil the result don't accord to my calculations. My mateial has the relative permeability myr=1200.
Do you understand what my problem is?
0 -
Hello,
I guess we need your Flux project to understand what is going on. Please zip the directory *.FLU (and maybe remove the mesh). I just wonder how you compute analytically flux density with magnetic material (it is usually quite difficult without FEA tool).
0 -
Hello,
I calculated it with the formula from the Amperes theorem. But the Support sais this is impossible. But in university we've done it with this formula: H=I*N/sqrt(Lcoil2+(2*Rcoil)2) and B=myr*my0*H.
This is my model:
When you look at the botom of this page :http://hyperphysics.phy-astr.gsu.edu/hbase/magnetic/solenoid.html
There is also a calculator which results are approximately the same.
0 -
Hello,
I have seen what the support has tried to explain you. I will do it differently. When applying the Ampere's law, we assume that the magnetic field is only going through the iron part (and is constant in it). then we have: H.L = N.I with
H magnetic field (B=Mu0.MuR.H) (B flux density
L length of the iron part
N number of turns
I current in one turn
Then we have : B=Mu0.MuR.N.I/L
B=4E-7.Pi.1200.1600.0.5/0.4=3T
When I try with your initial device, I have B=0.167T
Then I have modified the device and model only the iron plus a small volume of air around non-meshed coil. In this case the magnetic field is nearly constant in the iron part (following the initial hypothesis of Ampere's law) and equal to 2.935T (to be compared to 3T found out analytically)
In the first case, all of the flux is going in the air (the flux leakage), leading to the low value of magnetic field (and different from 'analytical' result).
I hope this case will convince you:
- that Flux is giving right result
- that it is not easy to compute 'manually' the magnetic field (you must be sure of your hypothesis)
You will find attached my model, already meshed and solved (it is not too big).
I hope it will help.
0 -
Hello,
Thank you ever so much!
Unfortunatly I have an old Version of Flux and can't open or import your model. But I tried it on my own and the magnetic flux density increase. But only to 0.5 T.
Is there any possibility to open your project whereas it is modeled in a newer version? Or may you can send me only your geometry or your values for the iron, coil and infinite box?
Is my assumption right that flux always computes magnetic circuits? And in my old case the circuit goes through a large space of air and now only through a little part of air and mostly through the iron core ?
0 -
Hello,
You will find the file in version Flux 12.3 (it seems you have it). there is a warning when opening it, but at least you can see what I have done.
Basically, I have reduced the radius of the coil to 21, and add a volume of air around the non meshed coil (radius 22). I have defined boudary condition to allow the flux to go out of the cylinder (on top and bottom). And when there is no boundary condition, by default it is set to parallel flux (flux is parallel to the faces). This is why the flux becomes constant in the iron part.
I hope it will work on your side.
0 -
Thank you so much !
It works
/emoticons/default_smile.png' srcset='/emoticons/smile@2x.png 2x' title=':)' width='20' />.
And is there any possibility to look at the magnetic flux through the iron core? (Phi=B*A=3.668 mWb). Or what is the magnetic flux I can calculate with 'FluxCoilConductor' ?
0 -
Hello,
The flux can be directly computed on the coil as you have done (this is the flux going through the coil). If you want to have it on the iron, one possible way is to define a sensor, which will integrate flux density on an existing face (top or bottom one).
0 -
Very Interesting discussion. I find myself in same boat as that of his user. Although I understand most of the discussion, Theres one question pertaining to non meshed coil. How do we know the length of the coil when we use non meshed coil? In this example I see that the length of core is taken as length of coil. However core can be of any length and it doesn't really say anything about turn density.
Please kindly help me understand.
0 -
Very Interesting discussion. I find myself in same boat as that of his user. Although I understand most of the discussion, Theres one question pertaining to non meshed coil. How do we know the length of the coil when we use non meshed coil? In this example I see that the length of core is taken as length of coil. However core can be of any length and it doesn't really say anything about turn density.
Please kindly help me understand.
Hi,
In principle the length of the non-meshed coil depends on the length of the device (iron length of the device). If you are using a non-mesh coil it is good to use geometric parameters (especially the length of the device). So, if you modify a parameter, the dimensions of the non-meshed coils also follow.
Hope it will help.
Best regards.
1
