Getting Started with Maxwell 3D (ANSOFT)

[Fuxue Jin]

Your description is a little confusing to me. Not sure what version Maxwell you use.

In the General tab, you specify Stop time and Time step.

What you described below should be in the Save Fields tab.

You need make sure these time setup will be good for your input condition. What is the input variable frequency?

Assume everything are correct, then when you plot the Field overlay, at the left corner, the Time may show as "-1". You need to double click on that, and then select the saved time step to show the field parameter such as B or J.

Dear all

I am quite new in the use of Maxwell and looking for some good exchange of ideas and suggestions

I have built and run a 3D model of a Transient Magnetic solution type. the model is interfacing with an external circuit which I have built with maxwell circuit editor

I have setup the analysis as follow:

General:
Type: linear step
Start time 0s
Stop time 0.04s
Time step 0.002s

Save fields:
Step size 0.002s (and added to list)

when simulation finished I cannot see any result!
For instance I cannot read any J field or other values on the "results" or "Field Overlays" when I select them from the object

Can anyone help how I can see results?

Thank you in advance for any useful information / suggestion

Best regards

Valerio

 

[Valerio]

Hello Jin, thank you for the feedback, it worked... you make me see some results of the work I was beginning! :-))

I am facing now a more serious problem:

I have a geometry model made of only conductor materials that need 6 input terminals but end in only one "collector" conductor.

The exciting circuit has 3 current impulses 120° duration that sequentially, overlap on each of the 6 physical inputs on my model. In practice my geometry has 6 separate windings as input that at a certain point are connected into one output.

when I tried to create 6 windings I assigned 6 separate input terminals and 6 output terminals the latter was actually the same output terminal of the geometry.
But Maxell didn't allow me, message was something like: "it is not possible to use different windings that share the same conduction path".

the point is that I cannot separate each winding because the porpouse of the analysis is to see currents (as happened in a rectifier current arm) but keeping one end (as DC current output)

Any ideas about how I can overcome this problem?

Thank you in advance for any useful information

 

[Valerio]

Sorry little correction: I said it wrong when I was talking about 3 current impulses. I intended 6 current impulses.

Another problem: in the circuit I have set 6 ammeter to see the actual current waveshape that exite my circuit.

How can I see/check the actual current waveform at the end of the simulation??

 

[Fuxue Jin]

Let me pretend I understand what you try to do.

My first thought is that you cannot create a such 3D model, as indicated by the error message.

I would assume it will be a lot easier to just create 6 individual model representing 6 windings. Each winding has its own input/output terminals.

By now, you should be able to run the simulation assuming all setup are correct. You will have 6 individual simulation results for each winding.

Then you can right click on the "Results" on left panel, select Output Variables. Here, you can sum up all your 6 simulation results to see the combined results and then plot the final result vs time. You may need to run the simulation one more time to see the result.

Hello Jin, thank you for the feedback, it worked... you make me see some results of the work I was beginning! :-))

I am facing now a more serious problem:

I have a geometry model made of only conductor materials that need 6 input terminals but end in only one "collector" conductor.

The exciting circuit has 3 current impulses 120° duration that sequentially, overlap on each of the 6 physical inputs on my model. In practice my geometry has 6 separate windings as input that at a certain point are connected into one output.

when I tried to create 6 windings I assigned 6 separate input terminals and 6 output terminals the latter was actually the same output terminal of the geometry.
But Maxell didn't allow me, message was something like: "it is not possible to use different windings that share the same conduction path".

the point is that I cannot separate each winding because the porpouse of the analysis is to see currents (as happened in a rectifier current arm) but keeping one end (as DC current output)

Any ideas about how I can overcome this problem?

Thank you in advance for any useful information

 

[Valerio]

To go more into detail:

each of my 6 windings has 6 input terminals. Along the conduction path, each one of these windings subdivide itself in 8 different geometrical paths, resulting in 6x8 conduction paths. Then, all 6x8 conduction path (all these currents) merge together in a common busbar.

The 6 paths are the arms of a double star rectifier, with 8 paralleled components per arm. The common bus is the DC output of the rectifier.

The common bus has, geometrically, great influence on the current distribution among the 6x8 parallels, because its path is orthogonal to the 6x8 parallels. It greatly affect the current distribution of currents.

I need to estimate in which way the current decide to subdivide into these 6x8 =48 parallels, entering cyclically from each one of the 6 input arm terminals, then merging and going out trough the common path, the DC busbar.

It is important therefore that I find a way to merge the 6 arm currents in some way inside my geometry and not keeping the 6 windings separate, because I need to obtain for the final path, a DC current and not a pulse current (which is the arm current)

If there are some good ideas to replicate this geometry in maxwell I will be glad to discuss

Thank you for any suggestionsthat would come

best regards

 

[Fuxue Jin]

How is your input signal frequency? Will the current be uniformly distributed and skin effect can be ignored?

You might want to look at DC conduction solver. In DC conduction solver, you can specify 6 input terminals as Excitation, and then the final output terminal as Sink. You will see the current density J distribution by plotting the J on any plane you want to see.

For what I understand, Maxwell has no way to do what you want to do using Transient solver. In this solver, it works using winding. Winding has terminal assigned to it. If a terminal is assigned to a winding, it cannot be assigned to another winding. Your output terminal cannot be shared with multiple wingdings.

To go more into detail:

each of my 6 windings has 6 input terminals. Along the conduction path, each one of these windings subdivide itself in 8 different geometrical paths, resulting in 6x8 conduction paths. Then, all 6x8 conduction path (all these currents) merge together in a common busbar.

The 6 paths are the arms of a double star rectifier, with 8 paralleled components per arm. The common bus is the DC output of the rectifier.

The common bus has, geometrically, great influence on the current distribution among the 6x8 parallels, because its path is orthogonal to the 6x8 parallels. It greatly affect the current distribution of currents.

I need to estimate in which way the current decide to subdivide into these 6x8 =48 parallels, entering cyclically from each one of the 6 input arm terminals, then merging and going out trough the common path, the DC busbar.

It is important therefore that I find a way to merge the 6 arm currents in some way inside my geometry and not keeping the 6 windings separate, because I need to obtain for the final path, a DC current and not a pulse current (which is the arm current)

If there are some good ideas to replicate this geometry in maxwell I will be glad to discuss

Thank you for any suggestionsthat would come

best regards

 

[juvefansir]

Hi everyone

I'm working on this structure, I’m using Transient and Magnetostatic solvers. Please answer my following questions:

1-for some reasons I should split this structure to just one pie of 1/6 size, after using Boolean, which boundary conditions should I use?2-Im working with Simplorer, but in Transient and Magnetostatic solvers, there is no information about resistance of coils, how can I add resistance to the model in this kind of transient-transient Link with Simplorer?

Thanks a lot .

 

[Fuxue Jin]

Q1,

I would suggest to use the full model instead of 1/16th. The region will normally be a rectangular. When you cut the model into 1/16th, the region will be cut into the same shape because the coil current excitation surface need to meet with the region boundary. I haven't used other shape region, not sure if this will work. Assume it can be done, but It will be very difficult and confusing to use because you need to use amper-turns to the surface, not just the ampers. You need to predict the H filed direction, and assign the boundary either perpendicular or parallel to the H field. The default it tangent (parallel)

Q2,

If you specify your winding coil (conduction path) as stranded, you need to specify DCR value when you assign current excitation. If you specify your winding coil as solid, Maxwell will simulate the DCR value and the result will be used.

Not sure if this answers your questions, but it is the basic to start with. I don't use Simplorer, not sure how it works.

 

[Fuxue Jin]

Need to make correction.

Only when you assign Voltage excitation, then DCR value need to be specified or simulated. Because DCR value is needed to get current from the Voltage.

If the excitation is Current, then you cannot specify the DCR.

All above are only for Transient and the excitation is applied to a winding which includes terminals. Magnetostatic don't have this option.

 

[juvefansir]

Need to make correction.

Only when you assign Voltage excitation, then DCR value need to be specified or simulated. Because DCR value is needed to get current from the Voltage.

If the excitation is Current, then you cannot specify the DCR.

All above are only for Transient and the excitation is applied to a winding which includes terminals. Magnetostatic don't have this option.

Thanks for your answer Jin :)

1-So for calculate DC_R , I should use Eddy current solver, then put the value into the excitation in transient solver ?

2-I think this program have some bugs , you know if we change the solution type to Eddy current , we don't have post processing tab , so there would be nowhere to put number of coil turns . am I right ? what do you do in this kind of coil to calculate DC_R ?

3-I think I this situation we should simulate real coil structure with physical number of turns ,instead of my easy model of coil (but it's possible for low numbers of turn), what do you think ?

thank you :)

 

[Fuxue Jin]

There are so many details for each solution type. I can only make a general comment here.

If you want to evaluate high frequency current effects, involving skin depth, eddy current and so on, you have to use real 3D model and specify the conductor as solid. This means if you have 20 turns of winding, you have to make a 20 turns 3D model.

For example, if you have a simple solenoid coil, you can draw a simple tube to represent the coil and specify the conductor as stranded. When you specify it as stranded, it means the current density is uniform, so there will be no skin depth or eddy current considered.

If you want to know the eddy effect or proximity effect between turns, you have to have a real 3D model. Maxwell will be very slow, near impossible, to finish the simulation if the winding turns are beyond 10 or 20. There are some work around on this but it will not change the fact.

 

[morteza1989]

hello everyone;
actually I have a strange problem:when I design an external circuit in maxwell the file should be saved like this:"excitation.sph" but for me it saves as"excitation.amcp".
how can I solve this problem?

 

[Fuxue Jin]

Menu Maxwell Circuit -> Export Netlist

 

[Jayaram12391]

Hi all,

I am using Ansys Maxwell for magnetostatic simulation and I have plotted the magnetic flux density (B) for the whole system - Linear generator - permanent magnet. I want to calculate the magnetic flux denisty of the air gap alone. I am not able to find how to do that.
When I plot B, it gives B values for the whole linear generator. Is there any way to export the results of B? Thanks for your reply

 

[Fuxue Jin]

B value (H, J for the same) can be plotted along a line, or at a point for specific value.

In order to see the B value along a line, you need draw a line using the line tool. You can add a line as non model so you don't have to solve the simulation again. Once you add a line, go to Results -> Create Fields Report -> Rectangular Plot, then select the line from Geometry menu on the top left. On the right, Calculator Expressions -> Mag_B.

You can export the 2D plot to a data table. Or you can plot the data table similar as you plot 2D curve.

 

[Jayaram12391]

B value (H, J for the same) can be plotted along a line, or at a point for specific value.

In order to see the B value along a line, you need draw a line using the line tool. You can add a line as non model so you don't have to solve the simulation again. Once you add a line, go to Results -> Create Fields Report -> Rectangular Plot, then select the line from Geometry menu on the top left. On the right, Calculator Expressions -> Mag_B.

You can export the 2D plot to a data table. Or you can plot the data table similar as you plot 2D curve.

Thank you very much for the reply. I want to find B for a given area. So i tried the steps given by you. There was no dropdown in geometry menu. I made an area instead of a line.
So I tried field overlays. It gave results of B, but i am not able to export the result from the field overlays.

 

[Fuxue Jin]

I believe you can only plot 2D rectangular curve, B along a line, and export the X-Y data. A plan will not show in the Geometry menu, only line or point.

You can plot B on a plane, like XY plane or any 2D plane you draw, but you cannot export that to a data table as that will be infinite number of data points. That is as far as I know.

Thank you very much for the reply. I want to find B for a given area. So i tried the steps given by you. There was no dropdown in geometry menu. I made an area instead of a line.
So I tried field overlays. It gave results of B, but i am not able to export the result from the field overlays.

 

[Jayaram12391]

I believe you can only plot 2D rectangular curve, B along a line, and export the X-Y data. A plan will not show in the Geometry menu, only line or point.

You can plot B on a plane, like XY plane or any 2D plane you draw, but you cannot export that to a data table as that will be infinite number of data points. That is as far as I know.

Thanks very much Fuxue Jin. That makes sense to me. Is it possible to find the average of B in an area?

 

[HamidMazaheri]

Hi every buddy
I want to create a maxwell 2D design from RMxprt.
My model is a three phase induction machine with broken bars. So it is not symmetrical because of broken bars in rotor cage. When creating a maxwell 2D design, it considers the symmetry of machine and models a quarter of it. How can I eliminate symmetry when creating 2D model?
Thanks

 

[Jayaram12391]

Hi every buddy
I want to create a maxwell 2D design from RMxprt.
My model is a three phase induction machine with broken bars. So it is not symmetrical because of broken bars in rotor cage. When creating a maxwell 2D design, it considers the symmetry of machine and models a quarter of it. How can I eliminate symmetry when creating 2D model?
Thanks

First create the model in RMxprt without broken rotor bars. After that convert the RMxprt design to maxwell 2D and then modify the model according to your requirement

 

[HamidMazaheri]

Thanks Jayaram. So I prefer not to create the model from RMxprt. Because that seems to be difficult to modify the 2-D model, because some parts of the model is deleted due to symmetry.
I created a 2-D model from the beginning. So I have some questions:
1- As we should set the length of objects zero, how does maxwell understand the length of them?
2- In maxwell 3-D when assigning a winding we should create a cross section of the coils from Modeler>Surface>Section and associate it with an excitation, what is the process in 2-D?

 

[Jayaram12391]

Thanks Jayaram. So I prefer not to create the model from RMxprt. Because that seems to be difficult to modify the 2-D model, because some parts of the model is deleted due to symmetry.
I created a 2-D model from the beginning. So I have some questions:
1- As we should set the length of objects zero, how does maxwell understand the length of them?
If you are asking about depth, you can check Maxwell2D---> Design settings ----> Model depth
2- In maxwell 3-D when assigning a winding we should create a cross section of the coils from Modeler>Surface>Section and associate it with an excitation, what is the process in 2-D?
1. In 2-D, first you can make a rectangle for the slots in induction machine.
2. Then make those to be made of copper material as it contains coils.
3. Using excitations tab, make each of the slot as a coil with certain number of conductors according to the design and give positive or negative.
4. Then add three winding A, B, C and include the appropriate coils to the windings.
Hope this gives an idea.

 

[HamidMazaheri]

Thank you very much. I did whatever you said.
I have some other questions:
1. For squirrel cage of an induction machine, I should short circuit the bars in rotor. How can I do that?
2. How can I assign solving Region? (As you know in 3-D we draw a rectangle and sweep it). When I draw a line in 2-D and sweep it, it disappears.

 

[hrdlickavaclav]

Hello,
Is it possible to make a point in vector field and show coordinates of vector in this point?
Thank you very much

 

[HamidMazaheri]

Thank you very much. I did whatever you said.
I have some other questions:
1. For squirrel cage of an induction machine, I should short circuit the bars in rotor. How can I do that?
2. How can I assign solving Region? (As you know in 3-D we draw a rectangle and sweep it). When I draw a line in 2-D and sweep it, it disappears.

I really appreciate it if anybody could help me. And I have another question:
When validating my model I get this error in message manager window:> At least one of the following boundaries needs to be specified: Vector potential, Balloon or (odd)Symmetry.
How can I deal with this error?

 

[carmatic]

Hello everyone, I am trying to use Maxwell 3D to simulate a rotating magnetic device , where 2 magnetically linked sections are rotating in opposite directions
However, the program only allows one 'band' object for simulating rotating parts
Is there a way to overcome this problem?

 

[zand1000]

hi everyone . I 've worked with maxwell since 2011 for industrial motors and generators design . I'm phd student of electrical engineering and I 'm trying to make more money from the skills that I ve earned . If anyone want to to simulate any academic or industrial project with MAXWELL I will be glad to be able to help him or her . feel free to ask any question by my email. my email is zand6565@gmail.com and you can see some of my simulations in my blog : http://electechno.blogfa.com/post/2

 

[gatopardos]

Hello, I have modeled a motor that needs a rectangular wave as input instead of a sine wave, you can input math functions directly into the winding's current tab so I can put a sine wave like I_max*sin(omega*time+phase) but I need this to be a rectangular wave , I guess there must be a way to do that without building a full circuit in simplorer using maxwell tricks but how is it possible?

 

[gatopardos]

For two objects rotating create two new coordinate system that pass through the C.M. of the rotating bodies and then add motion band for each body.

There is a special function in maxwell called pwl_periodic(ds1,time) that let's you define datasets, it's like a modulation for sine waves but maybe it can be used for rectangular waves.

 

[HamidMazaheri]

Hello every buddy.
I've modeled a linear induction motor in Maxwell3D. Selecting the "Translation Motion Type" in "Motion Setup", I face the following set of errors:
Maxwell3DDesign1 (Transient)
[error] Invalid moving objects. All moving objects inside band must form single connected region. (2:36:21 PM Nov 14, 2016)
[error] Motion setup (geometry) is not valid. (2:36:21 PM Nov 14, 2016)

These errors occur only in translation motion.
I'll be grateful if anyone helps me.

 

[Fuxue Jin]

For two objects rotating create two new coordinate system that pass through the C.M. of the rotating bodies and then add motion band for each body.

There is a special function in maxwell called pwl_periodic(ds1,time) that let's you define datasets, it's like a modulation for sine waves but maybe it can be used for rectangular waves.

Use the formula you mentioned above, and define the dataset "ds1". It cannot be a perfect rectangular wave form, you have to define a short period of time for rising and falling, such as 0.01s, depending on your frequency of the wave form. The dataset is more like, (0,0) (0.99,0) (1,1) (1.99,1) (2, 0)

 

[Fuxue Jin]

Hello every buddy.
I've modeled a linear induction motor in Maxwell3D. Selecting the "Translation Motion Type" in "Motion Setup", I face the following set of errors:
Maxwell3DDesign1 (Transient)
[error] Invalid moving objects. All moving objects inside band must form single connected region. (2:36:21 PM Nov 14, 2016)
[error] Motion setup (geometry) is not valid. (2:36:21 PM Nov 14, 2016)

These errors occur only in translation motion.
I'll be grateful if anyone helps me.

I don't have access to Maxwell and couldn't recall any details, but such kind of error sometimes indicates that due to the movement of your object, they are out of defined region. Try animation to see the movement, and see if all objects are still within the region after moving. I could be wrong though.

 

[gatopardos]

Use the formula you mentioned above, and define the dataset "ds1". It cannot be a perfect rectangular wave form, you have to define a short period of time for rising and falling, such as 0.01s, depending on your frequency of the wave form. The dataset is more like, (0,0) (0.99,0) (1,1) (1.99,1) (2, 0)

There isn't a perfect rectangular wave in nature either, the rise/fall time can be 0.00001 s or much less which is how I set it up.

 

[gatopardos]

I don't have access to Maxwell and couldn't recall any details, but such kind of error sometimes indicates that due to the movement of your object, they are out of defined region. Try animation to see the movement, and see if all objects are still within the region after moving. I could be wrong though.

This happens when you set multiple moving objects too.

 

[HamidMazaheri]

This happens when you set multiple moving objects too.

Could you please describe more. I have attached a view of my model to make it more clarified. In this model, for example when I set the motion setup type in "Rotation" around Y-axis, there are no faults. And I don't understand the meaning of the error: "[error] Invalid moving objects. All moving objects inside band must form single connected region". How can the all moving objects inside band form single connected region?