Reliability Engineering Snapshot TM

Illustrated Case Studies in the Maintenance Reliability Engineering World of Failure Analysis, Predictive Maintenance, and Non Destructive Evaluation



Material Properties - Case Study No. 63: Drive Shaft Torsion Fatigue Versus Rotary Bending Fatigue


Torsion failures of shafts can be hard to differentiate from rotary bending failures. Here is one such case.View of Shaft Keyway - Torsional Shear Crack Initiation

The shaft (picture left) powered a large circulating pump via a belt drive. It had a large split belt sheave-wheel bolted onto it and was approximately 10 years old with over three hundred million revolutions on it. Upon replacing the sheave wheel it was noticed that the shaft had failed into two pieces. This shaft failure had gone undetected because the clamping force of the large bolted sheave-wheel kept the severed shaft pieces together. The failure is a good anatomy of a low-stress high-cycle torsion-bending fatigue failure.

Maximum torsion stresses are always at the surface of a shaft. The manner in which a crack initiates at the surface, due to torsion, can be influenced by the state of stress within the shaft due to all of the outside influences. However, keep in mind that the shear strength of a material is always half as much as its tensile strength. So it would be logical to assume that in many cases when a shaft is under torsional stress that the crack will initiate along a shear plane. The shear planes for a shaft under torsion are parallel and perpendicular to the axis of the shaft. The figure above shows six cracks initiating at the surface. Looking at the rest of the pictures, it is obvious that every crack began at the surface and progressed, initially, perpendicular to the shaft's axis. The direction soon changes toward a 45° incline with respect to the shaft axis. If the working bending stresses were greater than the torsional stresses than the cracks would have continued inward and perpendicular to the center of the shaft. Why did the direction of the crack change from perpendicular to an incline? The square shaped notch created at the bottom of the keyway is a stress riser that raises the tensile stresses by a factor of four times the normal tensile stresses anywhere else in the shaft. Therefore the shear crack travels into a region where the tensile strength of the material is exceeded before the shear strength of the material is exceeded. As a result, the crack changes direction normal to the principle tensile stresses which are at 45° (under torsion load and not bending load). Notice in the picture below how the angle becomes more pronounced as the cracks reach the notch.

View Down the Keyway

Figure 2

Close up View of Keyway Cracks

Figure 3

For a close up of crack "1" click on it

In Figure 3 the cracks numbered "1" and "3" intersect cracks "2" and "4" at "B" and "A" respectively. Cracks "1" and "3" terminate at this juncture. Cracks "2" and "4" continue on a 45° plane with respect to the shaft's axis. The regions defined by "1-B-2" and "3-A-4" no longer transmit the torsional load. The cracks running along the surface of the shaft in Figure 4 progress for awhile but eventually terminate (white dots in the picture below). Crack "2" eventually turns into crack "7"(see Figures 2 and 4) which becomes the major crack face for the entire shaft. This entire section of shaft has been taken out of action, because of the cracking, and no longer transmits any of the torque.
Close up view of Shaft OD at Crack Locations

Figure 4

End View of Shaft

One look at the end of the shaft (picture left) and it becomes quite apparent that the answer is not obvious. You can see close ups of the two top quarter sections and the lower right quarter section by clicking on them. There are chevrons almost all the way around the shaft. They are located along the outer circumference and are oriented radially inwards. This of course means that there were numerous cracks beginning at the shaft surface and traveling radially inwards. This is indicative of a rotary bending fatigue. Also quite noticeable are the beach marks that emanate from crack "1" and travel in the direction of shaft rotation. This is something that a torsional crack would do. The beach marks take a change in shape from concave (with respect to "1") to convex at line "C". This is indicative of a severe stress concentration in rotary bending. The final fracture in the lower right quadrant lines up with line "B". Again, this is indicative of a rotary bending failure that has a severe stress concentration (i.e. the keyway). And the umpire says ... The shaft failed due to rotating bending AND torsional fatigue.

Take a close look at the circumference of the shaft in the close ups. Those are chevrons, not beach marks. The beach marks are further in and they are oriented perpendicular to the chevrons! It's highly unlikely that the beach marks were the result of the circumferential cracks joining up and then travelling toward the keyway notch?! With that said, it is clear that there are two distinct loading mechanisms occurring simultaneously. One crack started in the keyway due to torsion, while the other cracks started around the shaft surface and were due to rotary bending.

So why all this trouble to determine the cause? Aren't we splitting hairs? It broke, get real, get a life!! Well, when you're involved with maintenance in a big plant it is important to know whether a failure, any failure, is the next new chronic problem. Especially when it's a 5" diameter shaft on a critical piece of equipment. Sometimes we may not get the exact cause of the failure correct, but it is equally important to know what the problem "isn't."

In this case, it isn't an overload. The tell-tale sign of that bit of information is the smooth texture of the surface that is so evident in Figure 6 below. Even the final fracture region is smooth. Overloads always have one of two things, a lot of distortion in the final fracture region, or a very granular surface appearance. The smooth texture, if magnifide under a scanning electron microscope, would probably have some faint signs of beach marks. The fracture surface is actually the torsion component of the compound failure. The rotary bending component is crack "7" which started at the shaft surface and travelled radially inwards, joining up with the torsion component at the white dot marked "7".

Close Up View of Final Fracture Location

Figure 6

The words out of my mouth were like music to the maintenance manager's ears.

"Dont' worry about it, it died of old age."


All Pictures and Text Copyright © 2000 - 2014 Adler Engineering LLC of Wyoming USA Contact Mr. Adler


Great care has been taken in the compilation of this article. However, no warranty, expressed or implied, including without limitation, warranties of merchantability or fitness for a particular purpose, are given in connection with this article or any article archived on this website. Although this information is believed to be accurate by the author, the author cannot guarantee favorable results will be obtained from the use of this article alone. This article is intended for use by persons at their sole discretion and risk. Since the conditions of product or material use are outside of the author's control, the author assumes no liability or obligation in connection with any use of this information. The author is not liable for special, indirect or consequential damages resulting from the use of this material.

No part of this article or any article archived in this website, or any part thereof, may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, without the prior written permission of the copyright holder R. H. Adler. Nothing contained in this article or any article archived in this website shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters of patent, copyright, or trademark, and nothing contained in this article or any article archived in this website, shall be construed as a defense against any alleged infringement of letters of patent, copyright, or trademark, or as a defense against any liability for such infringement.