Half Shaft Headaches

[denn]

So I was under the impression that I had the right equations established for the maximum torque at the shafts our car would produce, but now that I am trying to apply it to some FEA I am getting a little confused.

The Max torque that our shafts should see making a few rounded assumptions would be about say 500N.m. This is the case if you multiply our engine torque by primary reduction, by first gear, and by the f/r sprocket ratio. Then divide the moment by two fot the two shafts. Is this a safe assumption, not considering inirtia effects, or using the load transfer method for the time being? Or is this method bad practice?

Next we wanted to confirm a model in Cosmos by hand calculating a maximum torque for a given solid shaft diameter (say ~20mm) and known TAUmax for 4340norm. Our thoughts were that this should show an FOS~1 (right at the failure point) But the maximum torque that can be applied from hand calulations, fails the shaft twice over in FEA.

Any suggestions....

[denn]

So I was under the impression that I had the right equations established for the maximum torque at the shafts our car would produce, but now that I am trying to apply it to some FEA I am getting a little confused.

The Max torque that our shafts should see making a few rounded assumptions would be about say 500N.m. This is the case if you multiply our engine torque by primary reduction, by first gear, and by the f/r sprocket ratio. Then divide the moment by two fot the two shafts. Is this a safe assumption, not considering inirtia effects, or using the load transfer method for the time being? Or is this method bad practice?

Next we wanted to confirm a model in Cosmos by hand calculating a maximum torque for a given solid shaft diameter (say ~20mm) and known TAUmax for 4340norm. Our thoughts were that this should show an FOS~1 (right at the failure point) But the maximum torque that can be applied from hand calulations, fails the shaft twice over in FEA.

Any suggestions....

[billywight]

What element size are you using? To get realistic results for a halfshaft in tosion you need pretty small elements which results in a lengthy computational time for the computer... Do a search (using the find feature) for more posts on this. I seem to recall a thesis paper posted recently about a program written for FSAE half-shaft analysis. It wasn't FEA, just an in depth hand calc method on the computer.

[denn]

I have searched but to no real avail, so I thought I might bring up the topic again.

If you could point me in the direction of this paper, or anyone cares to comment, it would be much appreciated.

[billywight]

Here it is:

http://www.mech.uq.edu.au/ugthesis/2002/FENNING_Mark.pdf

[Buckingham]

Does the stress the FEA shows in the middle of the shaft (an area of uniform geometry) match the stress values you get from going hand calcs on a solid bar / tube?

[denn]

Hey Billy, thanks for the thesis, I can apporeciate the work that went into his vBasic Program http://fsae.com/groupee_common/emoticons/icon_smile.gif

Haidinger, lets take this back to the beginning, becase I don't want to get lost in iterations if Im way off already.

note: t=tau

To verify my half shaft model in COSMOS, I have essentialy a 20mm dia. shaft with a mock CV interface on each end. One fixed one with an applied torsional load. My thought is rearanging t=Tr/J to T=tJ/r and putting in a known J,r, and t will output the Tmax. Applying this Tmax in our model should show a FOS~1.

However I am questing the method from which we have calculated tau. Machinerys Handbook suggestes maximum shear stress for a material is "not more than 30% of the elastic limit in tension, but not more than 18% the UTS." Having a data sheet of your 4340 with appropriate treatments, is this a realistic way to find tau, or is there a better practice Im forgetting?

Thanks again guys.

[Jersey Tom]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Then divide the moment by two fot the two shafts </div></BLOCKQUOTE>

If you're running a limited slip differential you will want to look into the effect of torque distribution in your shafts.

[KU_Racing]

Have you taken into account a clutch-drop situation with hot, sticky tires?? The inertia of the engine/drivetrain can make for a very large impact load on any elatic part of the drivetrain, i.e. the halfshafts

[Conor]

This is what I did, and it may make your life a little easier:

I took the full weight of our vehicle last year (745 which was huge) and distributed all the weight over the rear two wheels because I wasn't sure of the load transfer on acceleration. Then I calculated the torque as a result of the friction force and used that to find my tau max. Then I applied a safety factor of 2 and compared it to the values in the Machinery Handbook (there was a relation for shear strength based on the tensile stength). We were going to use 4130 (when Carroll Smith says its bad to use 4130 for axles, he's talking about 500hp cars, not fsae cars) but we decided we were cutting it kind of close and switched to 4340 instead. You may argue that we overbuilt our axles, but I'd rather not have a shaft fail and rip apart the rear of the car while its flailing around. That would be an even bigger headache than saving a couple ounces on an axle. Hope this helps.

[denn]

Thanks Guys,

The issue originally was not so much for calculating our actual cars conditions, but more so for validating a simple model in COSMOS. Upon closer inspection, and changing a few settings we realized that the shaft FOS was actually =1, the failure was occuring at an exposed spline region.

The issue was the scale was so large on the colour spectrum, that everything from a FOS of 0.3-4 looked the same. So basically we are kicking our own asses, and can further go onto analyze our cars actual output.

Now we move onto obtaining a solid data sheet for a HT4340 based on Rc data.

Thanks Again Gents.