May 15, 2014 -- Last month, my family had the misadventure of an unexpected visit to the ER. My eldest wanted to find out who would win a fight: the ground or him. The ground won, which it usually does. And my son? He's entirely healed, thankfully. My takeaway? Well, the hospital has these really cool visitor badge clips. While waiting, I had plenty of time to examine this little plastic wonder (and plenty of time trying not to worry about my son).
Here's what I learned from observation and from analysis. The badge and clip:
Close-up of the injection molded plastic clip. Material? I guessed it to belong to the polyethylene family.
For the purposes of analysis there are a few features I took out of the model, almost always a good and necessary process. For example, there are small ridges and teeth as well as injection molding features that don't pertain to the analysis. For that matter, all I will be testing is the force to compress the clip open (as I observed that it was fairly large in order to get it clipped to the edge of my slacks).
Originally thinking this was a simple extrusion, I saw that the supporting rib that undergoes bending is actually a couple of millimeters less wide than the part. The reason for my concern? I was hoping to invoke a 2D plane stress solution with SolidWorks Simulation. In the end, I did decide to assume constant width for just that purpose. However, I initially evaluated a model using all solid tetrahedral elements to gauge the part's initial response.
Now, I don't know about you but I don't usually carry along strain gauges or load cells with me. That said, I instead initially provided a very small displacement on to the clip. Additional assumptions are, again, that the material is a PE and that it obeys Hooke's Law. But, I also know something by looking at the finished clip in its open state. Like this:
From this image we can see the areas where contact may occur within the structure. We also observe that the part deflects rather largely to obtain the shape in the image. We can continue to assume that this problem is linear due to assume light loading; but we'll quickly find... it is not.
I began with only 0.1mm of displacement in a single direction to compress the part. It behaved linearly so I upped the displacement to 1 mm, then 4mm, and finally 5mm. Measuring the part, 5mm was about the needed displacement in order to open the clip. But see the behavior with a traditional linear static analysis (with 3D contact):
This was evaluated also as a 2D plane stress study. From the results of both analyses, a sharp crinkle occurs at the end of the applied displacement. And using Simulation to measure the reaction force at the restraint, over 200N is observed. So what happens when the applied load is prescribed as a force and not a displacement? Ugly things, my friends. See for yourself:
Why does this occur? Because the load must maintain its prescribed direction. It does not update its direction in a linear, static study. Our results are invalid although the assumed applied load is very likely to be close to what's required. This model's behavior dictates a non-linear solution despite the material's most-likely linear properties. (the joys of some polymers)
Again, we start our tests with prescribed displacements. Similar behavior is observed with 5mm of displacement.
A force is needed, not a displacement. (the same crinkling is observed) Again, applying our 200+N load as well as what we have learned by observing the actual part, that loads need to update with direction, and the linear behavior from earlier we know to use contacts where needed and other study-specific options. We're also assuming, for this go-around, no friction. The time-stepping aspect of non-linear analysis is perhaps one of the more important ideas the user must gain understanding of. (shameless plug: we're offering one-day courses now for our FE tools)
The behavior is more realistic and from here one may wish to begin optimizing this clip's shape. Why? Well, although it is a commodity part we may wish to try to use less plastic in the molding process. Or, we may wish to lessen the required force by moving the support rib or changing the webbing's thickness. All very do-able within SolidWorks Simulation. Furthermore, if we were to take out material then we could use SolidWorks Plastics to confirm, very quickly, how the cavity of the tool would fill and where we could find time and cost savings within the molding process. For example, using a generic LDPE and a gate location almost identical to the location on the actual part, I get the following 1.5 second fill:
Potential sink above and below the supporting rib, which may result in a less strong part.
Higher shear stress remaining in the region where the user would apply force to open the clip. This too
could lead to localized warp within the part and hence, an initial deformed shape (noted some camber in the actual physical parts) that could lend itself to localized buckling as the clip is loaded. This, in turn, could make the clip unable to open under load at all. Hence, a failed design and lots of expensive scrap. (and expensive tool changes or wasted time troubleshooting the modeling process parameters)
So, while sitting in a hospital ER waiting room, I'd like to think that I solved a small problem the world might potentially have with plastic visitor badge clips. But more importantly, I hope you have seen that non-linearities are everywhere. That they are solvable using the fine FE tools provided by SolidWorks. And that design problems translate to manufacturing problems. Problems that can be analyzed with SolidWorks Plastics such that they can be solved by you: the engineer.
For more from Chris Schaefer, read his article Hidden Gems of SolidWorks to learn about Sensors.
By: Chris Schaefer, Simulation Specialist