Effects of Keyway on Rotating Shaft

There are a number of good case studies among the papers presented at the Pump Users and Turbo-machinery Symposia over the years. They can be researched at:


We have used FEA to analyze specific issues related to shafts with keyways. In one instance it drove us to change from a square bottom keyway to a round bottom keyway. In several other cases, it drove us to use polygon fits or hydraulically mounted interference fits and eliminate the keyways entirely.

Johnny Pellin

Penguineer:
They are stress raiser and a significant discontinuity in the surface and the structure of the shaft. To that extent, most anything you read on stress raisers and discontinuities and the problems they cause with cyclic loading would pretty much be applicable. Any sudden change in shaft shape, size, stiffness, etc. can cause stress concentrations which tend to lead to problems and/or failures. Much of what we know about the basics of Strength of Materials and Theory of Elasticity really hasn’t changed much in many years, the same old Engineering Mechanics principles still apply, so I wouldn’t be surprised that an earlier paper would cover some of the basic facts of the matter. Today, we just FEA the hell out of all problems, and then aren’t sure how to model the problem with such-n-such software or how to interpret the results. As you suggest the torsional shear stress, which wraps around the surface of the shaft may not very by a great deal, but it still produces high stress concentrations at the corners of the key slot. If the key slot is in a high bending stress region, the bending normal stresses vary from max. tension to max. compression with each revolution of the shaft, and the bending stiffness of the shaft varies too and those bending normal stresses flowing along the length of the shaft and around all the sharp corners in the keyway, and can cause problems. Then, you also have the shear stresses from the bending loads, it acts essentially across the shaft section, but varies in direction with each rev. This all adds up to cause a pretty complex combined stress situation right at a significant stress raiser.

Bear in mind that ideally the keyway is only an angular locator, ideally the torque is transmitted by friction between the inner bore of the pulley, and the general surface of the shaft. Sadly this solution is rarely implemented these days.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

Hi

Try this

Thank you all for your responses.

JJPellin - I searched through the database and could not find a paper covering this topic specifically.

dhengr - I appreciate the input and completely agree with your summary. I was hoping (naively perhaps) that there were established stress concentration factors that I could apply to the safety factor calculations that I am using. Unfortunately all of my text references only account for scenarios with a stepped or grooved shaft; not so much for keyways. We could FEA this but I was thinking that might be unnecessary given the nature of the problem and knowing that these have been engineered decades before FEA was widely used.

desertfox - The link you sent me has the exact same formula I found in the 1909 paper. That gives me more confidence in this approach. The efficiency of the shaft as the strength ratio of the keyed shaft to an unkeyed shaft - typically 0.75 to 0.80. I'm wondering if this efficiency value can be substituted into the safety factor equation from the standard handbook for mechanical engineers (see attached, eq. 9). Perhaps the efficiency could be inserted in front of the endurance stress (sigma e) to reduce that value by 25%. Also note in the attachment, I believe the "2" values after the parenthesis are supposed to be superscripts, though they appear to be subscripts. This is an old, 7th edition of the book, it's probably been corrected since then.

I just realized that I should also provide the efficiency calculation.

e = 1 - 0.2(w/d) - 1.1(h/d), where:

e = shaft strength factor (efficiency)
w = width of keyway
d = diameter of shaft
h = depth of keyway = thickness of key / 2

I have also included the modified calculation I mentioned in the previous post with the newly added efficiency value noted in red. It makes sense to me to apply the efficiency value to the endurance stress to effectively reduce the life, but it's just a thought at this point.

I'd expect a machine design text book to have info like this for "sled runner" and end milled key seats.

There is a section in Peterson's Stress Concentration Factors, 2nd edition: Chapter 5 MISCELLANEOUS DESIGN ELEMENTS, Section 5.2 SHAFT WITH KEYSEAT.

I appears that Peterson's Stress Concentration Factors is the superior reference literature for this topic. I'm thinking that I should be able to acquire the stress concentration factor from that source and directly apply it to the safety factor equation in my previous post.

Thanks!