Quote:
Originally Posted by Adam.garcia
Just for curiosity's sake, are there any resources on the internet that neatly describe how to effectively use FEA on a model? I've tinkered around with it a bit in Solidworks, but it's hard to estimate the applied forces that will be experienced by this gearbox.
|
I haven't found very much either (a reason that I'm very much looking forward to my finite element analysis class at college next year). But you can get some pretty good results through free body diagrams. For example, if you know the applied torque at a CIM motor's pinion, you can calculate the force experienced at each bolt. The same goes for gears; think about how they'd move if there was no material present, and what kind of reaction forces the sideplate is providing at the bearings. Be methodical about it, and you can get a pretty accurate picture of most forces generated internally by the gearbox.
Finite element analysis is quite sensitive to incorrectly set constraints and forces. My stress analysis professor gave us some advice on how to verify that a finite element model is set up correctly. Simplify the geometry of your part (in this case, probably a non-pocketed plate), and perform classical stress analysis. Model the same simplified geometry, set it up for finite element analysis, and verify that your results are in line with the hand-done analysis. If they are, your forces and boundary conditions are accurate. You now know that as long as these remain the same, your finite element model will now provide accurate results, no matter how complicated the part geometry. Add in the pocketing, and rerun the analysis, and you should get accurate results.
The more difficult forces to estimate are those created by impulse forces (ie, the bending moment experienced from the shock of a robot flying off the bump and hitting the ground). When in doubt, overestimate. Unknowns like this is why everything is engineered with a saftey factor. Remember that this is an application where you're shaving mere fractions of pounds...it's probably best to be on the safe side with such a critical part of the robot.
Another thing to remember is that finite element analysis is good not just for showing material displacement under loading, but also stress flow. Even if your force
magnitudes are off, if you get the force
distribution correct, you will get a pretty accurate picture of where high stress and low stress areas in your gearbox are. You can then beef up the areas that show up as stress concentrations, and feel safe with minimal material where there isn't much stress.
Another thing to consider in gearbox pocketing: I'm suprised that
this style of pocketing hasn't caught on much more. Leaving a very thin "skin" in the pockets dramatically improves strength at a miniscule weight cost.