When working with components in the Stress Analysis Environment, there is a tendency for new users to take the converted contacts that form the intended ‘perfect world’ geometry, which often result in some very ‘kind’ results. The reality is that the ‘real world’ is never quite that nice, and nothing ever fits together that well. Today, I’d like to ask you to think carefully about what will happen when forces are appled, and use that to your advantage when applying constraints, contacts, and forces.
Inventor offers the ability to create contact pairs automatically by converting the existing assembly constraints. In the original release of the Stress Analysis environment, this worked very poorly, however I have found it quite effective in 2013. The downside to these is that Inventor will use the default contact (in the simulation setup) to guide the conversion. In almost every case, one type of contact is NOT SUFFICIENT. Some adjustments need to be made.
I this example, we have an INVAR rod, and a hand actuated clamp that holds the rod’s support legs. The clamp’s job is to act as a simple low stress holder to keep the legs from flopping about. Notice the gaps circled in Figure 1 above. These are allowances for hard rubber pads that will keep the bracket from scratching the rod, and allow additional friction.
Reviewing the other images, you should notice that a downward force has been applied to the aft, upper edge of the plate, representing the effect of a 200 pound man standing on the bracket. I wanted to prove that while many scenarios with basic bonded constraints are (barely) tolerant of this treatment, I suggest the reality is otherwise. The reason for the difference is in the simulation setup.
When any force is applied to the bracket, some form of take-up is going to be made at the rubber pads. This will show up in the form of compression. Eventually the gap will be reduced to nil and the aluminum bracket will meet the (really expensive) aluminum rod. Additionally, not only will the gap compress to nil, but because of the compression on one side or another, the bracket will rotate, and two sets of edges will meet the rod.
Additional contacts need to be added to keep the reduce the assembly DOF to 0, at the point where the bracket comes to rest on the bar. If we add a bonded constraint to the front or rear of the bar, the constraint will convert all the force into moment at the joint and show completely erroneous and overly stable results. That is why I like Shrink Fit for situations like this.
Adjust the Contacts
I this example I used Sliding/No Separation as the default contact method. All the constraints were converted to Sliding/No Separation. What I want to do is edit certain contacts as needed.
Add 2 sets of Shrink Fit/Sliding contacts to the upper, forward edges of the clamp jaw where it will meet the forward surfaces of the bar.
In Figure 3 notice the highlighted features: 1) the edge of the clamp and 2) the surface of the bar.
Next, add 2 sets of Shrink Fit/No Sliding to the rear SURFACES of the bar, where it will inevitably meet the bottom EDGE of the clamp. I need the components to remain in place here without a bonded contact.
The Shrink Fit/No Sliding contact is like a Sliding contact, with infinite friction.
Last I need to adjust 2 sets of Sliding/No Separation that lie between the inner surfaces of the clamp jaw and the outer surfaces of the bar. I will change these to Separation/No Sliding.
What this will accomplish is to allow the bar jaws to stretch outward under the strain. If we leave the constraints with the ‘No Separation’ option in place, the jaws will never leave the bar.
Notice in the image below how the bracket has rotated after the new contacts were applied. Remember, there is no motion is a static analysis. We need to guide the setup to how the problem should be measured.
If you look at the relation of the force with the bracket, you can see the new contacts much more closely relate to the inevitable rotation point.
In this example the jaws displaced 0.029 inches over approximately 0.75 inches of length. This indicates that the bracket would not open enough to permit the bar to slip free. HOWEVER…
The Safety Factor is reported to be absolutely insufficient, and indicates that permanent deformation will occur at all contact points.
Based on the lack of permanent deformation in the remainder of the bracket edges, it is possible that the bracket may not slip free under the strain, but will be permanently twisted at the top edge, ruining the bracket.
One thing to keep in mind is that we often have an expectation of what will occur, and that we are simply trying to prove or disprove a theory. Review that expectation thoroughly, and ensure that you are preparing your simulation properly.
Similar evaluations with differing bonded faces and edges were attempted, delivering an array of results, in some cases exceeding a safety factor of 1. None of those are consistent with the expectations of how the bracket will shift and the how a great portion of the stress will be directed at a very small, aluminum feature.