Inside/Outside diameters of hollow cylinder

[qiushi.ji]

Hello,

I want to build two hollow cylinders very precisely in geometries. They are:
1, outside diameter = 502 mm, inside diameter = 462 mm;
2, outside diameter = 462 mm, inside diameter = 85 mm

If I use Vortex Engine, it is only possible to re-calculate in %, which is not 100% precisely to the values I wanted.

Any solutions?

Cheers,
Ji

[coppelia]

For your first cylinder, why don't you specify:

• X-size: 0.502
• Inside scaling factor: 0.92032 (i.e. 462/502)

Cheers

[qiushi.ji]

Hi Admin,

1, in Bullet there is a "inside scaling" function but without "visual" effect, otherwise it is still not 100% to 462/502=0.9203187...
2, I will buy Vortex Engine in next days, so I want to use the same function in Vortex which is only precisely to 1% (i.e. 92% without decimal).

The problem I actually want to solve is that,
I'm trying to simulate a motor/rolling system with bearing. But I found that, if the diameter of axis is the same as the inside diameter of bearing, it is very difficult to verify, how is the result near to reality because the material friction should also be considered.


https://www.dropbox.com/s/q03g6llot0ab965/1.jpg

1/2 should be movable, 2/4 should be glued, 4/3 should be glued. If i simulate without Joint, the results (force/tension/stress between parts) are not stable. Before simulation "Role, Axis, Plane left/right, Gear" are groupped.

If I set up the diameters manually for a very little difference, for example:
Axis diamter(2) = 85mm,
Bearing inside diameter = 85.5mm,
and then set up the both skin thickness in Vortex = 0mm, it sees better.

I also tried to build a similar construction with very small "gaps" between cylinders. The simulation results sees much better.


https://www.dropbox.com/s/s0dyll39rxatuvz/2.jpg

Thanks for your suport,
Cheers

[coppelia]

In the primitive shape dialog, the Inside scaling factor will only display 2 decimals, but internally, your input will be remembered and applied. So if you entered 0.9203187, even if it displays 0.92, it will in fact use 0.9203187. You can veryify this by creating an auxiliary normal cylinder inside, and turning the contact point display on (with the Vortex engine).
But as you correctly mention it, there are a few other parameters to consider (such as the skin thickness)

Cheers

[qiushi.ji]

Hi Admin,

thanks for your answer. I've tried again with your suggestion and it works.
That means, with the Inside scaling factor I can define an inside diameter precisely to 0.000000(7 positions after decimal).

But still one thing I can't understand: in the real situation, it is very difficult to force an axis into a hollow cylinder, if the inside diameter of the hollow cylinder is smaller than the axis diameter. That means, you can put it in OR you can not put it in.

I've tried myself with such two examples in V-Rep (see pic). Both of the contact points are the same and the forces in Z direction are also very similar, which is in my opinion not logic.



One possible reason I could understand is, there are some parameters could be set, for example stiffness or tension of materials. That means, even the inside diameter is smaller than the axis diameter, I can still FORCE it into the hollow cylinder because of deformation and stress in both parts.

Cheers

[martin_simulation]

Hi Quishi,

Simulating an tight fitting axle in a hollow cylinder using collision is a very difficult problem And I don't recommend trying to do this.
The contact material is really hard and the error in position extremely critical. Simulating this would require a very small time step.
Such simualtion is better handled by a constraint.

Martin

[qiushi.ji]

Hi Martin!

are you the Martin that Martin? :D

Actually I also don't want to go into this problem too deep, because I just simply build the models according to our CAD files and run the simulation, but have seen some interesting results. Then I have recognized that building a bearing system in simulation is very difficult, because in real situation this "small gap" between bearing and axis also should be considered and is also very complicated.

However, I have to test the simulation results very critically - it is independent of the software ability - because it is very important to my documentation and validation. For example, now I can write in my document: "because modelling a bearing system in software is complicated and uncertain with such reasons bla bla bla, so a simplified model for actuation system would be constructed - and in this new construction, bearing system would be ignored but it still has the same function and e.g. 95% quality (validation)."

Cheers,
Ji

[martin_simulation]

Hi Qiushi,

Sounds good.

ps: yes this is the same Martin you know! I told you I'd be around!

Good week-end

Martin