FEA simulation of a Linear Rail

[lakeroe]

Hello,

within a complex simulation i need linear rails . For that reason i made an extrem simplified modell:

Front view:

The lower plane of the linear rail is fixed in all directions:

The distance between the two parts is 0,3mm and the contact type is "gliding":

I put a sideward force of 200N at the linear block:

I would expect the linear block glides along the rail and falls off at the end. But it moves almost not at all (max. 1,6 x 10^-4 mm):

The DOF-View also shows an error:

So what am i doing wrong respectively what's the correct approach to simulate such linear rails ?
The used modell (.f3d) is attached.

Thanks in advance and best regards,
lakeroe

[whunter]

You're trying to solve a transient problem with the static stress solver, it cannot solve that type of problem.

Normally, static stress solvers assume that your displacements are small so that it can be modelled as a linear problem. Your problem will have a large displacement, and basically the solver cannot handle that type of problem.

See this page for more info:

Learning | Simulation Types

http://help.autodesk.com/view/fusion360/ENU/?guid=GUID-3B4834A5-C954-44BE-893E-0C482560A887

Looks like you want to perform an Event Simulation (available in Ultimate edition as a tech preview).

What are you trying to solve with the simulation? Stress or acceleration or both? Acceleration you can work out by hand.

--

whunter

[lakeroe]

Thanks for the quick response.

I'm trying to simulate a whole gantry milling machine similar to this one (it's a german page though):
http://cncwerk.de/forum/viewtopic.php?f=3&t=493

I'm mainly interested in displacement, so I think the static stress solver should be okay.
My main problem right now is how to setup the simulation to get rid of the error in the "DOF-View" !

I think the error is related to the contact type.

Do you have any further suggestions ?

Best regards,
lakeroe

[whunter]

I would model whatever structure is underneath the rail too. With your current setup you won't see any deflection due to the 'hard' constraint just below the rail, in other words, you need to allow the rail to deflect. Try and model it as accurately as possible but keep it simple so that you can verify results by hand with simple calculations and slowly build the complexity of the simulation. It's the best way to get confidence in the results.

I think you can suppress Rigid Body modes in F360, but not sure (there should be an option), else you need to add a "soft constraint" to stop the slide from moving. The solver algorithm has probably recognised that there's a rigid body mode and it's warning you about it.

At a later stage, I would try to do a non-linear (NL) static analysis, but only to capture the effects of contact and local displacement as this will affect boundary conditions at the interface. NL static analysis comes in three (or four) flavours:

1) NL Material (when the material yields)

2) NL Geometry (e.g., large displacements, or locally different displacements)

3) NL Boundary (e.g., contacts that change)

4) Combinations of all three, so 7 possible combinations (M & G & B & MG & MB & GB & MGB).

You can spend (and waste) hours on NL analysis if you're not careful, so do most of your work in the linear domain initially. Mostly these types of problems can be dealt with by means of simple hand calcs (or use Mathcad or Swift Calcs).

--

whunter