O-ring force modeling - FEA

[dkeenan1]

I need to apply the loading from an o-ring on a part in a assembly to determine the deflection and stress in the parts due to compressing the o-ring.  The issue I am having is the force exerted by the o-ring on the part is dependent on how much the o-ring gets compressed.  I have a graph of o-ring compression force versus deflection. I am not interested in the analysis of the o-ring, just the mating part.

[dschenken]

You may just have to close the solution loop yourself - make a guess as to the load the o-ring will apply, use that load and see if the parts deflect as much as the o-ring does for that same load. If it deflects too much, lower the load. If it deflects too little, then increase the load.

It may be an approximation anyway unless the deflection curve also includes the precise area the o-ring acts on. Also consider that the force will not be uniform across the contact patch with the o-ring.

[dkeenan1]

DSCHENKEN,

 Thank you for the feedback.  I think this would work if the part had uniform deflection along the o-ring path.  In this case the deflection is variable since the part has changes in geometry / stiffness. I am not sure how to change the force applied along the path to account for the change in deflection.

[skunks]

[SweetPeasHub]

If the deflection from one side of the seal to another is varying around the seal then the springs respond with a force based on the deflection using hooks law. F=kX where k is the stiffness and X is the deflection of the spring and F is force.

[SweetPeasHub]

Keep in mind that springs add when in a parallel arrangement.  The stiffness of the o-ring that skunks simulates is the sum of the 8 springs used.  If more springs were desired to make a smoother result, the stiffness of the individual springs would still need to add up to the stiffness of the o-ring being simulated.

A similar use would be a door seal on a vehicle door in contact with the vehicle.  A seal stiffness per unit length is how this is thought of.  (N/mm  / mm)  It is implemented with discrete springs along the length where the stiffness and spacing resolves to the desired stiffness per unit length.  A non-linear stiffness - deflection curve can follow the same principals of summing and per unit length.

There can be some difficulty with the singularities at the discrete points.  Some choices to get past this might be one or more of the following(they all likely involve some trade-offs.)

1. exclude the singular geometry.
2. add small rigid surface regions at the singularity to spread the force over an area
3. limit the order of the elements and/or use generally large elements.
4. last resort is to use a actual seal in a contact model.  Prove in a separate test that the o-ring matches the force vs. deflection curve