Hello,
I am trying to apply a constant pressure condition on the inside of a pressurized cylinder, working with an axisymmetric mesh. I can select the node set on the inside and apply pressure, but I'm not sure what the units should be. I've also tried modelling with a distributed line and surface loads, but again, I'm not sure if I need to apply the full force, or the force on a slice of material.
Any help would be appreciated. Thanks all!
Axisymmetric Pressure Distribution
Moderatoren: ccad, mz15, auroraIco, Lehrstuhl
Re: Axisymmetric Pressure Distribution
Hello mturcios,
to apply a pressure you have to pick a surface set. Pressure is always applied to surfaces. Then you put in the value of the pressure according to your chosen units, mostly this is N/mm^2. The pressure now is calculated to node forces and applied correctly to all nodes which are in contact to your chosen surfaces with regard to the surface area.
If you use distributed line or surface loads, you need to pick node sets and type in your amount of force. This gets equally distributed over your chosen line or surface.
Then a special case (experimental): for pressure on shell edges (this is "pressure on a line"), you have to put in N/mm. Just take your pressure value and multiply it by the length of your chosen line. This works well for simple shapes but could be inprecise for complex ones...you will have to try it out.
If you have futher questions, feel free to ask again
Kind regards,
SHautsch
to apply a pressure you have to pick a surface set. Pressure is always applied to surfaces. Then you put in the value of the pressure according to your chosen units, mostly this is N/mm^2. The pressure now is calculated to node forces and applied correctly to all nodes which are in contact to your chosen surfaces with regard to the surface area.
If you use distributed line or surface loads, you need to pick node sets and type in your amount of force. This gets equally distributed over your chosen line or surface.
Then a special case (experimental): for pressure on shell edges (this is "pressure on a line"), you have to put in N/mm. Just take your pressure value and multiply it by the length of your chosen line. This works well for simple shapes but could be inprecise for complex ones...you will have to try it out.
If you have futher questions, feel free to ask again
Kind regards,
SHautsch
Re: Axisymmetric Pressure Distribution
Thanks for the help SHautch. This seems to make sense. I have an axisymmetric mesh made of a cylinder with external threads and chamfered edges on the outside. The inside is subject to an internal pressure, and I was thinking to use your N/mm multiplication solution. Since the inside of the cylinder is (initially) a straight line, the pressure is acting radially outward and your suggestion seems like an acceptable work around.
It may be useful in the future to allow pressure constraints to be applied to edges for 2D or axisymmetric cases. I know edges are not currently an entity as they are partway between nodes and elements, but for 2D, shell or axisymmetric cases, edges act the way surfaces would. Otherwise we are limited in 2D/axisymmetric cases and not able to fully use pressure BC.
On a related topic:
Would non-linear geometry help in a situation like this, since after some deformation the load is not purely radial anymore?
Regards from Canada,
Marco
It may be useful in the future to allow pressure constraints to be applied to edges for 2D or axisymmetric cases. I know edges are not currently an entity as they are partway between nodes and elements, but for 2D, shell or axisymmetric cases, edges act the way surfaces would. Otherwise we are limited in 2D/axisymmetric cases and not able to fully use pressure BC.
On a related topic:
Would non-linear geometry help in a situation like this, since after some deformation the load is not purely radial anymore?
Regards from Canada,
Marco
Re: Axisymmetric Pressure Distribution
Dear Marco,
you are welcome! It is not possible to use entities like edges in our current GUI, but you can still use them when picking node sets on an edge and using these sets for pressure or surface load BCs. Our solver then "knows" you are using edges and applies the load correctly (look at the z88i5.txt, where the element and its edge - specified by nodes - are used for surface loads or pressure...more information can be found in our theory guide on page 65 ff.).
"Non-linear geometry" only regards large displacement effects, which is achieved by applying the load incrementally (in a chosen interval) and could probably help if your part really sees great displacements. The pressure is always directed normal to the edge/surface and for axissymmetric/cylindric shapes there shouldn't be much difference in the load direction - this would only happen if your part would deform unequally.
But it's worth a try if you have the ressources (time, CPU and RAM
).
Kind regards,
SHautsch
you are welcome! It is not possible to use entities like edges in our current GUI, but you can still use them when picking node sets on an edge and using these sets for pressure or surface load BCs. Our solver then "knows" you are using edges and applies the load correctly (look at the z88i5.txt, where the element and its edge - specified by nodes - are used for surface loads or pressure...more information can be found in our theory guide on page 65 ff.).
"Non-linear geometry" only regards large displacement effects, which is achieved by applying the load incrementally (in a chosen interval) and could probably help if your part really sees great displacements. The pressure is always directed normal to the edge/surface and for axissymmetric/cylindric shapes there shouldn't be much difference in the load direction - this would only happen if your part would deform unequally.
But it's worth a try if you have the ressources (time, CPU and RAM
).Kind regards,
SHautsch