Force or Pressure 4

[Kesh78]

Load on the supporting structure should be considered as pressure or Force?

 

[civilperson]

Yes, force for a specific member and pressure for an area which then leads to a force.

 

[Kesh78]

thanks for the quick reply. Attached is the image of structure i need to analyze for stress and displacement. Red arrows are load - should these applied as force or pressure

 

[rb1957]

if you want pressures are forces distributed over (= applied to) areas ... both can be applied to structures.

eg the weight of an object can be repreented as a point force at the CG or as a pressure applied to an area about the CG (so that the total force (the integral of pressure over area) has the same line of action (ie through the CG).

clear as mud ??

 

[GBor]

I weigh...well, since we're in etherland, let's say 180 pounds. I can place a load of 180 pounds at the center of the seat that I'm sitting in, or I can place a pressure of 180/A over the sitting surface that has an area of A.

The force will probably due if I'm looking at the stress in the chair legs, but if I'm sitting in a desk chair, the bending stress on the seat will not be seen without the pressure load approach (or some distributed force loading).

 

[GregLocock]

".. the strains that can be produced in a body by the application, to a small part of its surface, of a system of forces statically equivalent to zero force and zero couple, are of negligible magnitude at distances which are large compared with the linear dimensions of the part."


So using superimposition it makes no difference whetehr you apply a force or a pressure, remote from the loading point.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[crisb]

Your model looks like it made of 3D elements. I find its best to use pressures for 3D elements, as forces will produce very high local stresses.

 

[rb1957]

pressure loads are just distributed to the corner nodes ...

there should be little difference between a pressure load in the upper surface, points loads at a set of upper surface nodes (all of them, some of them), or a single point force at the point of application with RBE3s to the surface of your model (i'd suggest that you don't use this method untill you understand how it works).

 

[johnhors]

"pressure loads are just distributed to the corner nodes ..."

Pressures on an element face are distributed to the nodes of that face (including mid-side nodes if present) via the element shape functions, as any other form of distribution will not be equivalent to the specified pressure and will create the false very high local stresses that Crisb referred to.

Nastran RBE3 elements cannot correctly emulate a uniform or smoothly varying pressure dustribution either.

If the structure in the vicinity of the load application is of interest then you'd be best to follow Crisb's advice.

 

[jetmaker]

For the most part, applying the load as a pressure or as distributed loads should produce minor variations in the results, especially if looking at regions remote to the point of load application.

I believe that a pressure distribution adjusts the loads at each node to provide a balanced loading. What I mean is that if you have an axially loaded bar, and you apply a pressure to the end face, the bar will expand uniformly such that the end face remains plane. Now if you were to apply the pressure as discrete forces, all of equal value, then you would see more deflection around the free edges and less in the middle as the face dishes. If the other end of the bar is a sufficient distance from the applied face, then both methods should produce the same stresses at that end.

jetmaker

 

[johnhors]

No Jetmaker, that is not the way to apply pressure as a set of discrete point forces. There is only one correct way, as I said above.

 

[jetmaker]

john,

I agree with you, if you want to simulate the pressure correctly, then you do as you have stated. What I was refering to was that if you apply the loads evenly to each node, as some people do, then the effect will not be the same as a pressure applied (except when remote to the applied loading).

jetmaker

 

[johnhors]

"apply the loads evenly to each node as some people do" - yes I've seen Patran users do this on numerous occasions, blissfully unaware that it was an incorrect procedure. It makes me wonder about the level of knowledge of the authors of the code.

 

[051174]

sigma = F / A
[N/mm^2]=[N]/[mm^2]

 

[cvan42]

Like crisb I use pressure heavily. Another aspect to think of, is forces are typically assigned in x,y,z coordinates, and pressures are applied perpendicular to the surface. So, if you have a rod/plate bending or buckling, a pressure load could induce it quicker than force load.