Discussion Groups

Simulation Mechanical and Multiphysics

Active Contributor
42 Posts
0 Kudos
Registered: ‎08-18-2013

Internal Pressure of 0.5MPA in Midplane Meshed Cylinder

250 Views, 2 Replies
09-21-2013 07:05 AM

I've got a DN500 pipe with 90^ elbow. Wall thickness is 9.525mm. Attached pipe on the elbow is 2.5m on each side. Material is carbon steel. I fixed one side of the pipe. Applied an internal nomal pressure of 0.5Mpa. I used midplane meshing but the the result doesn't seem to match the solid meshing result (result is almost 50 times off). The solid meshing result compares well with other FE software and more acceptable. The Von Mises result is 180 Mpa. But the mid-plane meshed element is way above this. Is there a technique in inputting internal pressure in mid-plane meshed element?.....Or it is not capable of handling internal pressure formulation?....


Your help is  greatly appreciated...Many thanks in advance...Cheers!!!!

*Expert Elite*
595 Posts
127 Kudos
Registered: ‎08-30-2012

Re: Internal Pressure of 0.5MPA in Midplane Meshed Cylinder

09-21-2013 11:34 AM in reply to: tibor121774

My guess is that you did not setup the plate model input properly.


Since a brick element has real physical volume, applying a pressure to surface "A" uniquely defines the direction of the pressure. A plate element does not have any volume; it is just like a piece of paper. If I were to ask you to apply a pressure to a piece of paper, do you apply it to the front or the back? Why? Well, the software is even worse! Bear in mind that it does not know that a "tube" should have pressure on the "inside" or "outside". All it knows is to apply the pressure to either the "top" or "bottom" of every element, so you are the source to tell the software what the "top" side means. (Specifically, a positive pressure is applied to the bottom side of a plate element, where the bottom side is the side facing the element normal coordinate entered in the Element Definition. See the wiki help page "Plate Elements".) 


Elbows are particularly tricky. If you display the loads and boundary conditions in the Results environment, you can see if the pressure arrows are in a consistent direction (all outward, for example) or mixed directions (some in, some out). Depending on the geometry of the elbow, there may be a small area where a single element normal coordinate will work for the entire elbow. If the diameter of the elbow is small compared to the radius of curvature, you may need to split the elbow into multiple surfaces or parts to get the proper element normal coordinate input. 

John Holtz, PE
Mechanical Engineer
Pittsburgh, PA

Simulation Mechanical user since Dec 1997
Active Contributor
42 Posts
0 Kudos
Registered: ‎08-18-2013

Re: Internal Pressure of 0.5MPA in Midplane Meshed Cylinder

09-21-2013 07:19 PM in reply to: AstroJohnPE

Thanks John for your reply. Actually if it is just elbow with no trunnion welded to it, a solid brick/tets element would suffice. I'm getting good comparison with the other FEA software which uses shell elements. The only difference is when there is a trunnion support welded to the elbow. The difference now varies up to 50 Mpa between solid and shell elements. Bending moments and forces can be easily modelled in plate/shell elements. It's really the internal pressure that's a little bit tricky.


Your reply is a good start for me to dig more on how to properly simulate the internal pressure in a plate/shell element. Does anybody have attempted this type of example in Autodesk Simulation?. I would really appreciate if someone can share their file to easily grasp the method. Please note that it's the only the plate/shell element example. The solid brick/tets elements can be easily modelled as John pointed out the solid has a volume and the plate/shell is just like a piece of paper.


Cheers to all!!!!!