2
$\begingroup$

I want to model a DC motor in order to eventually do some simulations (temperature/cooling related) and I'm wondering what is the best way to model the motor windings? I've seen a few video's on how others have done it but these only model 1 layer of windings onto the stator in the video's I've found so far (here is one of the videos: https://www.youtube.com/watch?v=j_7m2-e9MFw )

This is the reference image for the motor I am trying to model: enter image description here I'll only need to model one of the windings (1 of 36 if i counted correctly) since I can use a circular pattern.

EDIT: I could perhaps model a simplified version of the coils for the simulation (for example using rings of copper instead of spirals) if it would be significantly easier to model and give results that are not too far off. Or should I perhaps model sleeves of copper (of the approximate size/mass of the copper windings) instead of the individual windings ? Would this still give me a somewhat accurate result?

Background: I want to design a liquid cooling system for this motor and want to simulate the effectiveness of several different designs using Solidworks.

The windings have multiple layers (and the ones in the YouTube videos I found only have 1 layer). How would I go about creating the 'multilayered spiral shape' I need in order to create the winding model using sweep command? Or is sweep not the best option for this model? if so what modeling method would you recommend?

The first layer of the 'spiral' will go from left to right, the next layer from right to left, the following from left to right again and so on. The same way as the wire in this following video is wound around a spool : https://www.youtube.com/watch?v=j_7m2-e9MFw

My current plan of approach is:

1)

Draw a 3d sketch of the first (bottom layer, the one that is wound directly against the stator) layer of winding(creating the 3d sketch in the same way as in the video ( https://www.youtube.com/watch?v=GBNAnqolQ-I )).

It should look something like this:

enter image description here image source: https://www.youtube.com/watch?v=GBNAnqolQ-I (at 12:46)

2)

Add a fillet (round the corners) to all of the corners of this 3d sketch.

Should look similar to this:

enter image description here

3)

Copy the 3d sketch and scale up the copy such that it's new location matches the location of the second winding layer (I'll have to scale up the sketch in 2 of 3 axis only since the width of the winding doesn't change depending on which winding layer I'm modeling)

4)

Mirror the 3d sketch such that the new winding layer is wound in the opposite direction as the previous layer (from left to right instead of from right to left and visa versa)

5)

Repeat this process until the desired amount of layers is achieved

6)

Connect the several copies (aka layers) of the initial 3d sketch together to create a single 3d sketch out of them

7)

Use sweep feature along 3d sketch to create a 3d model of the windings

There are a couple of issues with this method though... :

Should I ever decide to change the number of winding layers or thickness of the winding wire or radius of the stator (the part that the wire wraps around) I will have to do a lot of 'cleanup work' to the winding model in order to not get intersection of the windings with each other (if i were to decrease stator radius without decreasing wire thickness) or get way too much spacing in between the wires (if i were to for example reduce wire thickness without changing the stator's 3d model manually (stator = the thing the wire wraps around).

The most optimal solution would be one where the amount of 'turns' in each layer would be automatically adjusted based on the wire thickness, the width of the spiral itself and the predetermined spacing between individual wires. And what would be even better is if I were to change the wire thickness.. all the layers would 'shift' upwards when increasing wire thickness and downward when decreasing wire thickness (since the same amount of layers with thicker wire results in more overall thickness).

Perhaps this is not possible though and I should accept that I'll have to remodel/retouch the winding whenever I change any of these parameters.

I hope that explanation was somewhat clear. If not please let me know and I'll try to clarify.

Is there something I can improve to my method of approach or an entirely different approach which would be better? What approach would you yourself use if you were to model this ?

Any input is greatly appreciated! Thank you!

$\endgroup$
5
  • 1
    $\begingroup$ Have a look at "Insert - Curve - Through Points" or "Through Reference Points". I would define coordinates for the wire path externally e.g. using MS Excel, set up with adjustable parameters. If you want to change the model, simply edit the feature, and re-link to the updated coordinates. $\endgroup$ – Jonathan R Swift Feb 24 '20 at 13:23
  • 2
    $\begingroup$ Can we assume that you already know the heat load from the coils? If so, couldn't you just create a simplified coil that is the same mass as your intended winding and apply the load? The simplified geometry would make it much easier to run a simulation. $\endgroup$ – Brice Edward Furr Feb 25 '20 at 18:14
  • $\begingroup$ @BriceEdwardFurr I don't know the exact heat load but should be able to make a pretty good estimate. Yes that's definitely something I am considering, if the results of a simulation on such a simplified model would be relatively close (withing 5-10% ) to the result on a more detailed model it might be the best option indeed. $\endgroup$ – Maarten -Monica for president Feb 26 '20 at 23:30
  • $\begingroup$ @JonathanRSwift thank you for the suggestions. It sounds like that approach would be a good solution. I could probably use linear patterns (in two directions) with the spacing and amount of instances defined in the Excel sheet . I'll try it soon :) $\endgroup$ – Maarten -Monica for president Feb 26 '20 at 23:31
  • $\begingroup$ I’m with you, @Brice Edward Fir $\endgroup$ – Jonathan R Swift Feb 27 '20 at 6:41
-1
$\begingroup$

I'm can't offer much advice on the thermal analytics part - don't have the specific knowledge. However, your modeling problem has a simple solution. You can define variables in the Solidworks model, and use them (or equations derived from them) in your drawings. Take a look here: https://help.solidworks.com/2018/English/SolidWorks/sldworks/HIDD_EQUATION_MANAGER.htm?id=4e1fdbad62f645edaff8e62eb3c9684c#Pg0

$\endgroup$
1
  • 1
    $\begingroup$ The use of equations is certainly going to be involved here, but it's not a simple solution - how would you suggest driving the 3D sketch via equations? It would need to be built up from multiple patterned entities, with the number of instances equation driven... But then how to link up the multiple layers of winding? Your answer would be better as a comment, since it helps point in the right direction but doesn't actually provide a solution. $\endgroup$ – Jonathan R Swift Feb 24 '20 at 13:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.