Displacement Discontinuity Between Plates with Bolt Preload and Friction Using Linear Contact

[EKCP]

I have two base plates bolted together with preload and friction (see attached) set up with linear contact. When a shear load is applied to the joint there is a discontinuity in the T2 translation between the plates even though the joint should be sufficiently clamped to prevent any relative motion. I have adjusted the friction values, experimented with convergence tolerances, and search distances with no success. However setting the coefficient of friction to zero results in a larger discontinuity (0.0234in vs. 0.0179in). The surface contact is otherwise behaving as expected. Any ideas? Thanks!

 

[BlasMolero]

Hello!,
This could be a graphics magnification, please note this is linear static analysis, then displacements are small, then the deformed shape is exagerated. Use a scale 1:1 + activate "ACTUAL DEFORMATION", and you will see the real results.
Best regards,
Blas.

~~~~~~~~~~~~~~~~~~~~~~
Blas Molero Hidalgo
Ingeniero Industrial
Director

IBERISA
48011 BILBAO (SPAIN)
WEB:

http://www.iberisa.com

Blog de FEMAP & NX Nastran:

http://iberisa.wordpress.com/

 

[EKCP]

Hi Blas!

Thank you for your response. That was my initial thought so I checked the nodal displacements of two conicident nodes on each of the top and bottom plates (see attached). While not large there does exist a relative displacement ~.0004in. A similar translation occurs at the edge of the hole. The plates "should not" translate with respect to each other. However I am new at this and maybe I am missing something. Would a small movement such as this be physically justified?

Thanks again!

 

[rb1957]

why "the joint should be sufficiently clamped to prevent any relative motion" ?

how are the two pieces connected together ? anything less than a rigid constraint will strain.

Quando Omni Flunkus Moritati

 

[EKCP]

Well, the nodes I looked at are coincident but on two separate bodies. That seems to me like a discontinuity that implies relative motion of the adjacent surfaces rather than a strain gradient. That's my thought anyway.

 

[rb1957]

there's clearly some load travelling across the joint causing a tiny displacement. possibly the load isn't perfectly (and that's Perfectly) normal to the plane. look at the free body of each piece.

it's possibly due to the way you constructed the model. i've made simple test models and gotten different results depending on how i created the geometry under the model.

Quando Omni Flunkus Moritati

 

[EKCP]

The load is specifically NOT perpendicular to the plate. I applied a large enough normal load to the plates that the bolt preload "should not" be required to prevent translation. Even still this small displacement implies relative motion between the plates. In my mind two plates which are clamped together should not slide or we would have serious fatigue problems. Thus I conclude that something is not correctly set up within my model such that the FEM predicts a relative translation between the plates.

 

[rb1957]

how did you set up friction on the surface ? sounds like a -ve gapping element ... just curious, i don't use friction.

if you have clamped the plates together, and Know that your applied load is less than wahtever friction could be there, then you're satisfied that in the real world it won't slide; despite what the FEA shows you.

you know that friction is reducing the displacement, how about increasing it ?

what if you constrain a couple (3) nodes on the interface (equivalence 3 pairs of nodes) ?



Quando Omni Flunkus Moritati