Foo - How do you model torque in FEA?

I've got a model which is basically a motor with a square armature driving a gear box with a square lug and I need to model the forces in FEA. The only problem is that the force distributes from 0-full from the middle of the side of the square lug to the edges on all 4 sides and there's no way for the ansys package I'm using now in autodesk inventor. I can apply a normal force onto the side but I end up having to build a stepped pad onto the sides of the lug and then approximating the forces onto the steps. So i'm like basically setting up a finite analysis while preparing the mesh for the actual finite analysis, which is pretty messed up.

There's gotta be an easier way than building tiny pads all the time for loading.

There's gotta be an easier way than building tiny pads all the time for loading.

I would shoot it and buy a new one.

Ah... makes sense.

Sorry...I haven't touched ANSYS in over 2 years. It seemed to me that ANSYS had a standard torque case in it's library. Why does the force increase like you describe? You can't get a reasonable approximation by placing shear forces on the faces of the lug?

I assume this is a part too complex to simply consider on it's own with an calculate the shear stress at the base...

best thread evar.

Well the lug itself is easy enough to calculate, it's not the shear it's the deformation and stresses on an adaptor around the lug, which is pretty complicated in geometry.
I guess an easier way to break it down and ask is to ask how I would model the forces exerted on a bolt interface in FEA on a plate of metal.
I'd have to create a temporary pad to simulate the forces exerted onto the plate by the bolt by grounding that pad right?

Trust me. FEA is not as accurate as people would like to think, especially on thin components were things like inclusions and processing textures will alter the response to a normalised stress field.

Cover the thing in strain guages, build it and give it some hammer. You'd be better off.

The thing is actually a big chunk of aluminium. Given the right modeling, FEA can do a pretty good job. It's often not economical to build dies for 10 different designs, even soft dies, to build prototypes to test when you can narrow it down to 1-2 before live testing. The problem with strain gages is that it doesn't always have enough resolution to capture the minor details. Strain gages are fine for simple or large geometries where you have relatively finite sampling points. A photoelastic test would be better but again, no live samples. Now if I can just figure out how to set up the analysis correctly, it should point me in the right direction.

ha ha. Didn't think you'd do it, Wong. ;)

Awww, come on guys, it's so simple. Maybe you need a refresher course. Hey! It's all ball bearings nowadays. Now you prepare that Fetzer valve with some 3-in-1 oil and some gauze pads. And I'm gonna need 'bout ten quarts of anti-freeze, preferably Prestone. No, no make that Quaker State.

Last night I slipped on my tub and fell, and dreampt up the answer to your question. Here, I drew a picture:

Last night I slipped on my tub and fell, and dreampt up the answer to your question. Here, I drew a picture:

400 watts??? Where am I going to get that kind of power?

Are you sure you need so many sample points on such a large object with essentially simple loading and unloading?

Are you sure you need so many sample points on such a large object with essentially simple loading and unloading?

The loading's simple but the geometry of the part is pretty complicated, plus we want to know what happens at all the corners and everything.

I guess an easier way to break it down and ask is to ask how I would model the forces exerted on a bolt interface in FEA on a plate of metal.

I must not be getting the geometry at all. From your description, it sounds like the load is exerted solely at bolt holes, am I right?

If so, and assuming you were able to model the part alright, I know you can place loads at the bolt holes in the appropriate directions in ANSYS. I would do a distributed (surface) load covering 180 degrees of the holes, using the varying load (0-max) option. Also make sure to put a constraint against rotating at the center of the adapter. I forget the exact terms in the ANSYS menus. It'll probably take a little bit of trigonometry on scratch paper and a lot of clicking and swearing at ANSYS's inability to select each hole. Why for the love of jove do you click on one line and it selects one 300 pixels away!?!

Then again, I'm probably just misunderstanding the problem.