Getting Started with Maxwell 3D (ANSOFT)

[siva0707]

hello frnds..
hoep you all doing good.
i am a mechanical engineer but works in electrical engineering. my work is electromechanical interactions in induction motors. so i started working in ansoft. so kindly anyone help me in the modeling of induction machine in ansoft. i am working with 3 phase 4 pole induction machine and it will me helpful for my work.

thanks
with regards
siva

 

[nautylus]

Hi,
i'm new user and I've studied ansoft maxwell for three months..
i need some help : i want to do calculate second harmonics in a magnetic material (for examp. between 2 magnetic rings center)
how can i calculate that by field calculator ??
thanks with regards

 

[gerbi]

Maybe have you any experience the distributed mode computing between computers?

I have no expierence with this. One is for sure, You'll need more than one Maxwell license to do this

i am a mechanical engineer but works in electrical engineering. my work is electromechanical interactions in induction motors. so i started working in ansoft. so kindly anyone help me in the modeling of induction machine in ansoft

Sorry to say that, but it's not that easy. If You have no expierence in EE or in Maxwell then this is really long way.. You need some professional help, not just question-answer discussion like here.

i want to do calculate second harmonics in a magnetic material (for examp. between 2 magnetic rings center)
how can i calculate that by field calculator ??

Try performing FFT on this signal. If signal is implemented in Maxwell go to Results-> Perform FFT on report.
If signal is not defined:
1) define a line on which You want to observe signal
2) go to results -> create rectangular plot
3) in left window you pick name of solution and geometry (your created line) and number of sampling points on it (pick some reasonable number, depending on mesh)
4) pick quantity to plot (some are ready to use, to have more go and create defined expression in calculator)
5) create new report and do FFT on it (Results-> Perform FFT on report). I have no big expierence with this FFT module.

This can be done in Maxwell v14 3D Transient.

 

[nautylus]

Thanx Gerbi,
i've some experience abut your suggestions (i had created a line and rect. plot etc.) but i want to see some definite equations in the center area if it works, i mean how can i load a spesific formula actually?
thanx again

 

[gerbi]

but i want to see some definite equations in the center area if it works, i mean how can i load a spesific formula actually?

Please be more specific. What equations do you want to see (in caluculator, right?) ?
If you mean FFT equations - this is not possible. To define FFT analysis other that implemented in Maxwell - you need to perform FFT in program outside Maxwell (with data outputed from Maxwell).
If you mean defining in calculator a expression to plot.. You need to work with calculator.
How to define a expression for further use ? Quick example:
Quantity->B
Scalar->ScalarX
Add->(pick a name, like Bx or so)
This viariable (Bx) is visible in calculator now (named expressions window, top of calculator) and you can plot it on defined geometry (rectangular plot, on model etc.).
Is this an answer to your question ?

 

[nautylus]

Thank you very much Gerbi..
it works really (your calc. definitions)... for the other one i guess i need to perform FFT program ouside...
thanx with regards.

 

[siva0707]

RMxprt (ANSOFT)

Hi gerbi..
i have done the modelling of three phase induction machine in RMxprt. i gave data's such as stator information, rotor information, winding information. when i export this to Maxwell 2D, i am not able to get the full model, its coming as only half model. how can i get the full model when exporting into maxwell 2d??

thnks
with regards
siva

 

[gerbi]

I have no expierence in RMxpert, but if You have half of the model and You want full.. how about duplicating it (mirror) and uniting cutted parts in Maxwell?

 

[Fuxue Jin]

For core loss, I have not done much but here is what I believe.

The core loss data set in Maxwell is used to define Cm, X and Y three parameters (after fitting the loss curve). With these three parameters, the core loss is calculated at any give frequency when the simulation runs.

For thermal, basically you need define how these three parameters change with temperature, either by a equation or data set.

Again, I have not done much but that is what I understand. However, simulation results about core loss in Maxwell is not that accurate.

Thanks for your help!

Yes, I found this options and I adjusted the core losses in 25°C(by the "power ferrit" core loss type).
Actually I'm using N87 ferrit core. I already can measure the core loss. The results is very nice :)
The next step what I would like to do is the "Thermal modifier".
By the core loss type I have Power ferrite. There is 3 value (Cm, X, Y). Every value have "Edit thermal modifier" option. In this window how can I adjust the loss for example 100°C?

Thanks!
Best Regards!
Bálint

 

[Fuxue Jin]

When you run transient simulation, the initial setup only use the initial mesh. It doesn't refine the mesh as it does in Magnetostatic.

Other quoestion:

The simulation is very slow. 50us transient (0.5us step) takes about 20 hour. I would like to speed up. My toroids dimension is about 50mmx16mm. What is the recomended mesh size (faster simulation/precise caltulation).

My compure:
Intel core i7
6GB 1333mHz RAM

Thanks!

 

[Fuxue Jin]

I believe distributed mode needs separate license.

Thanks gerbi!

I am going to try it!

Maybe have you any experience the distributed mode computing between computers?
I tried, but there is something error message:
"Error: Unable to contact AnsoftRSMService, Reason: ANS_CANNOT_CONNECTTO_ANSOFTRSMSERVICE
Testing Completed.
Error: Some tests have failed.

 

[klmnopq]

I need to simulate machines in maxwell

any tutorial for begainners like me?

 

[gerbi]

I need to simulate machines in maxwell

any tutorial for begainners like me?

write me your e-mail in private message, i'll send them to you

 

[homer5785]

Hey guys,

im using Maxwell 3D V14 in transient mode to simulate a pulse transformer. So imho it is useful to implement a frequency dependend magnetic permeability given in most datasheets of magnet materials. I found a how-to here:
http://www.psma.com/sites/default/f...ferrite-core-electronic-planar-transforme.pdf

But when I simulate a step response of a pulse transformer in this way the simulation does not converge very well. I think this is because of the fact that I haven't defined the permeability for up to the highest frequencies in the step. But unfortunately I get not more data from the datasheet. What does Maxwell if there are out of range data in a dataset?

Thanks in advance

homer

 

[jynth]

Hi all,

I'm a new user and i need some info on an error message i got while doing a busbar inductance simulation which says " Could not preserve critical nodes " Now where can I see in the software that shows these critical nodes so that I can fix it. Or may be in the meshing is parameters what would be the ideal settings.

I would thank anyone who can help me do a quick fix!
:)

 

[annalisemalta]

Hi,
I am also new to Maxwell and I am trying to simulate my model in 3D. however I am getting this error:

There is an error in executing 'mesh3d'

even though I created a separate mesh for each item

thanks

 

[cimopata]

Hy all!

I continued the torroid transformer simulation, but I discovered some weird things.
I measure:
-input power
-output power
-core loss
-solid loss
-standed loss
-standedR loss
I add these values but there is still have missing power.

To simplify the measurement I creat a simple O ring core and one primer coil.
The imput voltage is +/-100V square. The core is N87 ferrite adjusted B-H curve and core loss.
I attached the maxwell 15.0 file and a Excel file I calculated the power.
The comutate result is:
Pin: 3.483733733W
Pout: 0.002990193W (it is heat actuaally)
So where is my sweet 3.480743539W ? :D

Thanks!
BR.
Balint!

 

[cimopata]

And here is the excel:

 

[sourav kundu]

Hi gerbi..
i have done the modelling of three phase induction machine in RMxprt. i gave data's such as stator information, rotor information, winding information. when i export this to Maxwell 2D, i am not able to get the full model, its coming as only half model. how can i get the full model when exporting into maxwell 2d??

thnks
with regards
siva

Figured it out. Go to Maxwell 2D → Design Properties. Change the value in fractions from 2 (or as in my case 4) to 1. That's all. Good luck.

Regards
Sourav

 

[homer5785]

Has anybody an anwer for me? Ref.

Hey guys,

im using Maxwell 3D V14 in transient mode to simulate a pulse transformer. So imho it is useful to implement a frequency dependend magnetic permeability given in most datasheets of magnet materials. I found a how-to here:
http://www.psma.com/sites/default/fi...transforme.pdf

But when I simulate a step response of a pulse transformer in this way the simulation does not converge very well. I think this is because of the fact that I haven't defined the permeability for up to the highest frequencies in the step. But unfortunately I get not more data from the datasheet. What does Maxwell if there are out of range data in a dataset?

Thanks in advance

homer

I tried exactly the way in the publication (page 11-14) but up to now the simulation does not converge.
My magnet material (Epcos N30) is defined for frequencies up to 4MHz. I now tried with a "slower" step with a rise time of 200ns. This should avoid frequencies above 4MHz but even with this approach the simulation doesn't work.

One problem is that Maxwell is not starting without a setted Core Loss Type ("Nonlinear assignment required...") so I setted this to Power Ferrite but setted the parameters zero. Unfortunately I have no parameters for that but I think the core losses must be included in the Magnetic loss tangent.

Any ideas?

Thanks in advance

homer

 

[siva0707]

Coupling of elctrical circuit and magnetic circuit in ansys

hi guys..
i am doing analysis on 4-pole induction machines. i want to couple an electrical circuit with magnetic circuit in ansys. anybody having tutorials for that...basically am from mechanical background. i am struggling a little bit. so kindly help me for the same.

is it possible to do couple field analysis in ansoft maxwell ?

thanks
with regards
Siva

 

[siva0707]

Maxwell 2D (ANSOFT)

Hi ..

hope you are all doing good..

i need a help...

currently am working in Ansoft Maxwell,s 2D.

I created a RMXPRT model for 3 phase induction machine. then i exported to MAXWELL2D design.

its a 4-pole 3 phase induction machine. the machine has 60 slots.
it has double layered winding.
the results of the winding currents looks like non-uniform. Phase A, Phase B and Phase C currents are not unique.

kindly give me some suggestions to get uniform winding currents.

Thanks
with Regards
SIVA

 

[sophie ing]

Hi guys,
I'm working in Ansoft Maxwell 3D, Transient.
I want to simulate a sort of brushless motor: it's composed by only two coil, with the same excitation. My coil are 108 turns, with 6 layers (18 turn per layer). I have some questions:
1) Is Coil terminal of each coil set to 108 turn?
2) Are there differences in coil terminal setting between the coil of this example and a coil with the same turns and the different number of layers (obviously in this case the geometry is different)?
3) Do I have to create two winding or I can associate the same at the two coil terminals?
4) In Winding properties, how do I have to set the parameter "number of parallel branches"? What is it?
Thanks in advance

Sophie

 

[SirAskalot]

Siva: What excitation do you have? Current or voltage source? External circuit?

Sophie: How are the two coils connected, Series or parallel?
There is a option in the coil terminal dialog asking for the number of turns: 108 turns.
Do this for both coils.
Make a winding and add the two coil terminals to this winding. Then specify the excitation for this winding, e.g 100*cos(2*pi*50*Time).

If the two windings are connected in parallel I think you must write 2 in the winding properties dialog box.

 

[sophie ing]

Siva: What excitation do you have? Current or voltage source? External circuit?

Sophie: How are the two coils connected, Series or parallel?
There is a option in the coil terminal dialog asking for the number of turns: 108 turns.
Do this for both coils.
Make a winding and add the two coil terminals to this winding. Then specify the excitation for this winding, e.g 100*cos(2*pi*50*Time).

If the two windings are connected in parallel I think you must write 2 in the winding properties dialog box.

The two coil are connected in series.
Thank you

 

[sophie ing]

I have a problem in Maxwell 3D.
I tested a simple situation: I tried to calculated the magnetic field of a current loop.
In the center of the current loop, I expect the direction of the magnetic field parallel to the axis of the current loop, but in the simulation this doesn't happen (see in the picture). Why? What's wrong?

Thanks with regards
Sophie

 

[siva0707]

Hi

Greetings to you all!

I got the three phase currents as a unique pattern with phase difference.

but i got the peak to peak current value is 9A only.

my configuration of the machine is
rated voltage 415V
rated power 27.75kW
Frequency 50Hz,
Rated Speed - 1470 rpm

i need to get peak to peak current value around 60A...

what should i need to do in the maxwell design?

thanks for your help in advance!

and one more thing also..

how to calculate the power of an induction machine when the total number of strands are reduced to 3 from 4 strands? The rated output power of an existing machine is 37kW which has 4 strands and 11 turns. now it is reduced to 3 strands and 11 turns. now what will be the output power for the same machine? rated voltage is 410V, 50 Hz freq. machine has 60 slots. (basically i am from mechanical engineering, but i am working in electromechanical interactions in induction machine)

Thanks
Siva

 

[kikicooot]

Dear friends
I am student and I have exam about ansoft maxwell 3D.
I should simulate 3 phase asynchronous motor in ansoft maxwell 3D and to make several magnetostatic analysis.
Can someone help me?
I am beginner, so the best help for me is one finished project.
Please someone send me .rar file...

 

[Shantanav]

I am designing a 2D model of a three-phase core-type transformer in ANSOFT MAXWELL 14.0. While doing so I have encountered a strange problem.

Start of the problem:-

While designing the transformer, I wanted to incorporate core losses. So, I did the following steps:-

Excitations -> Set Eddy Effects -> checked the various core sections in my model (because eddy current effects occur in the transformer core)

and also

Excitations -> Set Core Loss - > checked the various core sections in my model (because core loss effects occur in the transformer core)

Symptoms :-

Model simulated (using transient solution type) and then gives the following wrong results:

1) Line currents on the primary side (star-connected) are very high (about 10 times)

2) Secondary side (delta connected) phase currents lose their phase dependence, i.e. the phase currents are no more 120 degrees apart, they all are in phase now. However, the secondary line currents are Ok though.

Additional Information :-

If I do not consider eddy current and core loss effects, then the model simulates to give correct results.

Please provide me suggestions/hints/solutions to my problem - how to incorporate core losses correctly in my model.

 

[gerbi]

Hi

Turn off eddy effect in core (turn it off in the windings as well), use only core loss calculations (based on core material loss curve P[W/kg] vs B [T] at single frequency). Set core conductivity to 0 (leave core material density, stacking factor, stacking direction unchanged - as for real core material).

Using eddy effect can give unrealistic results when done unwisely. Use eddy effect calculations only when you are considering induced losses in solid elements (tank, tank cover etc).. Applying it to windings or core (full scale model) would need very detailed design.

 

[Shantanav]

Thanks Gerbi for your response. Ok, I will shut down the settings for eddy-currents and concentrate only on the core loss settings.

So, I purchased this transformer from the manufacturer and at present I do not have the specified P (W/kg) vrs B (T). What I do have is a curve for Magnetizing inductance (Lm) VRS. Magnetizing current (Im).

Can that be helpful ?

 

[SirAskalot]

What is the purposes of your simulation? How accurate loss calculations do you need?
Hysteresis curve don't do you any good in accurate calculation of core losses. But hysteresis losses (part of core losses) are proportional to the area under the hysteresis curve.

P(T) curves at 50/60 Hz you can find at many electrical steel manufacturers. E.g sura.se

Also, since you got the transformer, you can do some measurements and find the core losses.

 

[kovibb]

Hi all

Please can you tell me how can I create equivalent circuit in MAXWELL 3D in version 15? Specially,where is "export matrix to look-up table using Equivalent Circuit Extractor" menu?

Thank you for our reply

 

[Shantanav]

Hi

Turn off eddy effect in core (turn it off in the windings as well), use only core loss calculations (based on core material loss curve P[W/kg] vs B [T] at single frequency). Set core conductivity to 0 (leave core material density, stacking factor, stacking direction unchanged - as for real core material).

Using eddy effect can give unrealistic results when done unwisely. Use eddy effect calculations only when you are considering induced losses in solid elements (tank, tank cover etc).. Applying it to windings or core (full scale model) would need very detailed design.

Hello Gerbi,

Thanks for your suggestion. Yes, I followed your steps. Actually, I created a new material. Firstly, I got a BH curve for hot-rolled low-carbon steel from the Sawhney machine design book and incorporated that into the material (since my transformer's material is hot-rolled low-carbon steel). Next up, I got the P (W/kg) versus B (T) for that material from the same book and incorporated it as well. I set the material density, bulk conductivity etc to optimum values.

The transformer model now simulates and gives correct results.

Thanks,
Shantanav

 

[Shantanav]

What is the purposes of your simulation? How accurate loss calculations do you need?
Hysteresis curve don't do you any good in accurate calculation of core losses. But hysteresis losses (part of core losses) are proportional to the area under the hysteresis curve.

P(T) curves at 50/60 Hz you can find at many electrical steel manufacturers. E.g sura.se

Also, since you got the transformer, you can do some measurements and find the core losses.

Hello Sir Askalot,

Purpose of my simulation - Trying to simulate the transformer under healthy condition and get current waveforms as close as possible to my actual lab-transformer. So, I am trying to model the core as similar as possible to the actual specimen I have.

Yes, I incorporated the P versus B curve for my core material and the model simulates now.

Shantanav

 

[firdausa]

hello guys,
i want to model a switched reluctance motor in ansoft maxwell with double layered winding, but i am not being able to find the option for layers of winding. please help me with a solution for it if anyone is dealing with switched reluctance motor in ansoft maxwell.thanks
Firdausa

 

[Fuxue Jin]

In Maxwell, most of time you don't draw the real winding (turns, layers and so on). You draw a 3D object and specify the cross section as stranded and define how many turns.

hello guys,
i want to model a switched reluctance motor in ansoft maxwell with double layered winding, but i am not being able to find the option for layers of winding. please help me with a solution for it if anyone is dealing with switched reluctance motor in ansoft maxwell.


thanks
Firdausa

 

[kovibb]

Help with genetic algorythm

Hi everbody,

I have created project for my thesis and everything is already set up.The problem is that my Maxwell version sent me back message error about "setting up first generation" when i start genetic algorithm in optimetrics analysis. I hear that can be bug. So I want to ask for help, I sent you my project,you try to run it on your machine and sent me result back with all project. Please,Is here someone who would do it?

Thank a lot.

 

[arial947]

Hello,

I have worked a good deal with Magnetostatic solutions in Maxwell; however, I now have to design an electromagnetic system that requires a transient setup. I have a single coil electromagnet with a ferrite core that I need to operate at a frequency of 50 kHz. I set up the meshes, including the skin depth, as well as an external circuit with a sinusoidal current source. My solution time step is therefore 1e-6 and the stop time is 1e-4. I specified that I want the field values saved at the same points by adding them to the list under 'Save Fields', but after I run the simulation, I do not see how to view these results. Does anyone know the best way to view the fields calculated at these time steps?

Thanks,

Arielle

 

[gerbi]

Hi Arielle,

after the solution is complete (for a specific time moment) it can be viewed by double clicking on "Time" button in your main window (in left bottom corner, just under your model or Viev/Set Solution Context). There you can choose solution to view (and when you will select one you can update your Fields Overlays to this specific time moment).

 

[jerrard]

Hello,

I am working on modeling in Maxwell an inductive charger for electric vehicles. The whole concept is quite similar to a transformer, though instead of simple windings or coils, litz wire is used. Would it be possible to design litz wire like a normal coil? Would it be all right if I drew a circular or rectangular conductor with a specified diameter? And if so what would be the size of the cross section of the conductor?

 

[SirAskalot]

If I state that: A solid circular or rectangular conductor would be a good approximation to a litz wire, IF you specify that eddy-currents are NOT induced in the conductor. Would you agree? And can you give a reasoning behind the statement?

As for the cross section, it depends on the effects you want to model. In reality a coil, and especially a litz wire would give a "copper fill factor" of less than 1, due to insulation. If you are unsure if the radius and fill factor has a effect on the solution you can try using different radii and see if there is a difference in your "answer".

 

[jerrard]

SirAskalot thanks a lot for your answer.
I agree that if I specify that eddy-currents are not induced in the conductor, a better approximation of a litz wire can be achieved. Considering the radius, I will try using different radii and see how my answer is affected.
Yet i am unsure of how to model the litz wire considering the number of turns: For my application I need a litz wire of two turns for the "primary" of my charger (for a transformer the litz wire would be a coil of two turns) carrying a current of 200A at 20kHz. I also need 16 turns of litz wire for the "secondary".
Would it be OK to define the conductor as stranded and specify the number of conductors as 2 for the "primary" and 16 for the "secondary"? Or would it be a better idea to draw two connected conductors for the primary and 16 conductors for the secondary and define them as solid?

 

[SirAskalot]

It again depends on what effects you want to model, but most times its not necessary to draw each turn.(most times are when eddy currents like proximity effect, skin effect etc. are neglected in the coil/winding) So I would agree with you regarding defining the conductor as stranded, and use a multiplier like 2 and 16 (number of turns or similar in a dialog box).

The reason for neglecting these effect in coils are due to the complexity of drawing and meshing such small elements. Analytic formulas may be used instead for calculating the neglected phenomena, like resistive losses etc.

 

[jerrard]

It again depends on what effects you want to model, but most times its not necessary to draw each turn.(most times are when eddy currents like proximity effect, skin effect etc. are neglected in the coil/winding) So I would agree with you regarding defining the conductor as stranded, and use a multiplier like 2 and 16 (number of turns or similar in a dialog box).

The reason for neglecting these effect in coils are due to the complexity of drawing and meshing such small elements. Analytic formulas may be used instead for calculating the neglected phenomena, like resistive losses etc.

Although I could obtain a value about the output voltage similar to the one expected (the radius of the conductor does not seem to affect this value), the self-inductance of the two coils as well as the mutual inductance is different from the ones expected. Is it possible to accurately measure the inductance of a litz wire using the method described above...?

 

[ajax20]

Royerconverter - inductive coupling of to aircoils

hi guys!

i try to calculate the efficiency of a Royer converter with two air coils (wireless power transfer) and the effects of steal or iron near to the coils. I have got a transmission frequency of 100 khz.

Which solver is the right? Magnetostaic or Eddy Current? (without steal, the coupling coefficient is nearly the same, but with a steal core it is half that.)

are the core losses included in the calculated inductive coupling coefficient?

 

[eddyrooney87]

How to establish motion on two different material

Hi Guys,

Need your help here. I intend to simulate linear motion on my linear actuator design. In my design, there are few multiple objects move along together with different type of material. In order for the design to work, all moving objects must be included in one-all inclusive object which I do not know how to do it. For your information, my design moving object include of a aluminum cylindrical shaft and multiple sets of permanent magnet ring. In reality, both objects are attach together.

1.The problem is how do I attach both together in simulation so they can be an one-all object?
2. Is there any way for me to simulate linear motion for both object in single motion?

Appreciate your quick feedback on this.

Thanks & regards,
Eddy

 

[pkelecy]

To do this you need to add a "band object" that encloses the moving objects in your model, and then assign a motion setup for that band object. If you search the help, it will explain how to do this. If the model is 2D, the translational motion will need to be approximated by a large radius rotational motion, as motion setup for 2D translation is broken in Maxwell since v12.

Good luck!

 

[eddyrooney87]

Hi pkelecy,

Many thanks for your feedback. As for the band object, I already did it. The band cover all moving objects. I have no issue with "band object". My concern is more towards how you combine different object into a single-motion simulation? I did read the tutorial but there is no clue on having two different objects to moving together under one single-motion. If I unite both objects, it will going to have a single type of material which I did not want. I need those two objects with different material to moving together. Any solutions on my problem? Will be really grateful if you can help.

Thanks & regards,
Eddy

 

[gerbi]

Eddy,
this is why you create box to which you assign motion. Any object inside it will be affected by motion setup. (i.e.: rotor of synchronous generator - there is a shaft + coils, but you create only one cylinder which contains all moving objects - and they move simultaneously).