Calculate radius increase from XY nodal displacement data

[ash267]

I have XY nodal displacement results for a cylinder from FEA. I wanna calculate the radius increase for each nodes. I tried using trigonometry to calculate hypotenuse of each nodes using X and Y data. But the results does not make any sense. Values are too high.
Any ideas guys?
Thanks in advance!

 

[FEA way]

If you can access undeformed and deformed nodal coordinates in your FEA software then it would be easiest to transform them to cylindrical coordinate system. Then you can calculate radial displacement for each node.

 

[rb1957]

do mean the difference between sqrt((X+dX)^2+(Y+dY)^2) - sqrt(X^2+Y^2) where X, dX, Y, dY are dimn relative to the centre (not overall global dimns).

another day in paradise, or is paradise one day closer ?

 

[ash267]

I tried both of the above suggestions but the results does not make any sense. I have attached an image for reference.

Center coordinate system sometimes is at center of the cylinder sometimes it's not. because I have more than one cylinder. I tried trignometry and the values does not make any sense.

 

[rb1957]

how far does the center translate ? (you're looking at the translation of the point relative to a fixed center)

another day in paradise, or is paradise one day closer ?

 

[ash267]

Thanks for your reply.
let's say a 3 cylinders of an in-line engine and the center coord system is on the center of the first cylinder top. What I want is the radial deformation of all these 3 cylinders from their XYZ displacement values based on the center coord system.

 

[FEA way]

If you still have access to the FEA software and files used for this analysis then the easiest way would be to postprocess the results in the same software. Pretty much all FEA postprocessors should allow you to transform the results into user-defined cylindrical coordinate system. This way you will get the radial displacement values directly.

 

[ash267]

I tried cartesian to polar conversion but the values I get are 3 to 4 times higher than what I expected.

 

[FEA way]

Which FEA software do you use ? You may have to redefine the local cylindrical coordinate system or read different displacement component (radial displacement should be the first one).

 

[ash267]

I use hyperview to view the results

 

[ash267]

Maybe the distance I should calculate is the one shown with yellow line?

 

[ash267]

Also this is an answer I got from someone. Unfortunately he is not responding anymore.

 

[IDS]

The formula given by rb1957 (sqrt((X+dX)^2+(Y+dY)^2) - sqrt(X^2+Y^2) ) will give the change in radius from the original centre. It will also give approximately the same result as the yellow line in your diagram above.

If there has been a global movement of the cylinder, and you want the movement relative to the new centre point, you will need the coordinates of the new centre.

How do you know your results are too high?

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[ash267]

From experience and the level of distortions I am getting from calculation is around 300 microns radially where we normally see around 100-170 um.
I tried creating a new coordinate system on the top or each cylinders and resolving results in that coordinate system. The value are still in the 300 um range. Maybe I am missing the global movement

 

[rb1957]

instead of looking at a specific node, what about looking at the cross-section ? then you'll have the movement of the center.

another day in paradise, or is paradise one day closer ?

 

[ash267]

@rb1957 how do I do that?

 

[rb1957]

how have you meshing the original shape ? in an orderly manner, or just blasted nodes that the thing ?

At a minimum can you find nodes at opposite ends of a diameter ?

another day in paradise, or is paradise one day closer ?