ANSYS - Tube Bending and Measured Ovality

[MECHENGTOM]

Hi All,

Quick background on the matter:

1. I have conducted a non-linear FEA (using ANSYS) on a super duplex stainless steel tube (Tube dimensions: 15.875mm ID and 1.00mm WT, Poisson's: 0.3, Young's Modulus = 200 GPa) representing the lateral crushing of a tube between two flat plates.
2. I concluded that a force of 52.5 kN/m produced an ovality of 3% within this specific tube. The value was validated against Roark's formulas for stress and strain - Table 9.2 formulas for circular rings.
3. The plan was to use the figure of 52.5 kN/m to calculate a radius in which the tube could be bent and not exceed the 3% ovality - methodology attached.
4. The calculated radius, for the case of the previously mentioned tube was 0.193m. Tension, i.e. force to bend the tube was calculated at ~ 10 kN.

FEA Setup Overview:

- Static Structural
- Tube dimensions as per above. Bending profile radius as calculated above @ 0.193m. Tube guide to prevent sliding of tube across the bending profile.
- Tube non-linear material profile verified from previous FEA - same material model used for the above. 2 layers of mesh across WT with 60 divisions. Large deflections turned on.
- Frictionless contacts between tube guide to tube and tube to bending profile. Bonded contact between tube end and flat face.
- Joint rotation of 120 degrees applied to bending profile. Back tension applied to = force to bend tube @ ~ 10kN.
- 2 off commands entered into ANSYS solution which allows the recording of XYZ dimensions between 2 nodes (one for horizontal and one for vertical displacement). This allowed me to find the minimum and max ODs.

Results:

Results from the simulation are producing an ovality of 21% - in comparison to the target of 3%, this is way off. Can anyone see if there's anything fundamentally wrong with the method?

 

[mfgenggear]

Hi All,

Quick background on the matter:

1. I have conducted a non-linear FEA (using ANSYS) on a super duplex stainless steel tube (Tube dimensions: 15.875mm ID and 1.00mm WT, Poisson's: 0.3, Young's Modulus = 200 GPa) representing the lateral crushing of a tube between two flat plates.
2. I concluded that a force of 52.5 kN/m produced an ovality of 3% within this specific tube. The value was validated against Roark's formulas for stress and strain - Table 9.2 formulas for circular rings.
3. The plan was to use the figure of 52.5 kN/m to calculate a radius in which the tube could be bent and not exceed the 3% ovality - methodology attached.
4. The calculated radius, for the case of the previously mentioned tube was 0.193m. Tension, i.e. force to bend the tube was calculated at ~ 10 kN.

FEA Setup Overview:

- Static Structural
- Tube dimensions as per above. Bending profile radius as calculated above @ 0.193m. Tube guide to prevent sliding of tube across the bending profile.
- Tube non-linear material profile verified from previous FEA - same material model used for the above. 2 layers of mesh across WT with 60 divisions. Large deflections turned on.
- Frictionless contacts between tube guide to tube and tube to bending profile. Bonded contact between tube end and flat face.
- Joint rotation of 120 degrees applied to bending profile. Back tension applied to = force to bend tube @ ~ 10kN.
- 2 off commands entered into ANSYS solution which allows the recording of XYZ dimensions between 2 nodes (one for horizontal and one for vertical displacement). This allowed me to find the minimum and max ODs.

Results:

Results from the simulation are producing an ovality of 21% - in comparison to the target of 3%, this is way off. Can anyone see if there's anything fundamentally wrong with the method?

Suppliers have tricks of the trade they don't share. As to have a competitive edge.
While can have high cost I would pay to run test.
And see the results.
Is this for actual buy or for practice.
If this an actual buy disscus with sales engineer
With high experience with precise tube bending
Not saying your results was not accurate. 3% seems
Tight.

 

[JStephen]

My impression- you are confusing two different effects.
A tube flattens when squished between two flat plates due to applied force.
I think (but can't necessarily prove) that a tube subjected to pure moment would tend to flatten as well, without any applied force.
So when you wrap the tube around a cylinder, you are assuming that the flattening is due to that contact force, when in fact, most of it would happen without the contact force.
I guess a quick check on this idea would be to apply moments to each end of a long tube, while holding the ends round, and see if the tube flattens in the middle without any contact force.
I would think this might be addressed in some of the Strength of Materials books (Timoshenko's is pretty good for stuff like this), but don't have any handy here to check.