There is clearly something I have always failed to appreciate and need some education.
Consider a model where the loads are transmitted though contacts to ground.
Contacts are stiff springs that stop parts penetrating each other (too much). Their stiffness and distribution is modified for compromise penetration.
Now, when building models we often use the methods:
.... MPA with max edge order set to 1
.... Quick check where all edges are order 3.
to debug a model as the analysis run times are quicker
Global load applied to model
Force in any ground springs used to take out minor imbalances
The question is, is there a minimum edge order before the contacts transmit the correct total force via their stiff springs?
After all, the contact pressure distribution could be nonsense but the springs must transmit the correct forces?
That is what I did. I was quite surprised that the solution doesn't settle down until order 3. Hence my question.
Contact force in my model:
Order 1 = 571N
Order 2 = 6296N
Order 3 = 7546N
There is also a mesh refinement dimension to the answer.
I don't know if the 'nature' of the contact also plays a part - cylindrical, flat, conical ...
I also played with the 'check contact force' option but saw no difference.
I was wondering what others' expectations/knowledge was.
I naively thought that as the contacts are springs then this was a force balance whether or not the contact pressure distribution was any good.
The purpose of low order mesh was to study force balance without burning too much time.
You are right,,,,, what happens at 4,5,6,7 ...?
I stopped at 3 and pursued answers for a project; the force seems to level out. Besides, we are relying on blanket 3rd order being good enough for the first pass of an SPA and it is used for 'quickcheck'. So one would hope the answer doesn;t change much
The model is not publishable but is also too large to have a blanket higher order mesh.
So I will make a separate study to understand this behaviour and will share here.
I don't properly understand yet.