|Was it a torsion fatigue failure, torsional overload, or did the Machine Shop mechanics do a bad job of overhauling the gearbox?
This shaft came off of a double reduction gearbox. It was the output shaft, and drove the pinion gear that turned a large rotary kiln. It failed while in service. The pictures to the left, and directly below, clearly indicate a torsional failure mode. The crack is at a 450 angle with respect to the shaft length. This type of crack orientation usually happens in torsion. The crack begins on the left hand side of the keyway tip and progresses inwards. There are numerous beach marks and the surface is quite smooth. The crack continues until there is very little shaft material left. At this point in time the shaft is overloaded as evidenced by the high profile shear lip. The load in this location is not really torsional anymore, and the principal direction of the stress has changed. It is more of a tearing action and less of a twisting action. The fact that the shear lip is rough indicates brittle failure. This meant that the magnitude of stress was greater than the material could handle, and therefore the failure occurred quickly. If the failure was due to a process upset that created a torsional overload, the appearance of the roughened shear lip would be spread over a far greater portion of the shaft surface. Production operators did not abuse this piece of equipment.
Did the Machine Shop mechanics do a bad job overhauling this piece of equipment? That makes for good water-cooler "get-no-where" chit chat. It's kind of like asking the question "Who won the Presidential debate? Bush or Gore." In that debate everyone saw what they wanted to see; most "party" people had already made up their minds before anything was said. So too with people's perception of machine shop mechanics. The mechanics took the hit for this one until a failure analysis was completed.
For all of you quality control buffs, yes, the Machine Shop mechanics could have probably found the crack before they finished overhauling it by simply using dye-penetrant on the shaft. The crack would have shown up quite nicely. But let's have a reality check here. You have to ask yourself, what is the cost of this extra inspection versus the benefit? How often does this type of failure happen? Are you going to dye check every single shaft that comes into the shop? If not, then how do you determine what not to check? How is that new inspection program going to delay other repairs, and how much are those delays going to cost in lost production and profits? When a Machine Shop foremen is up to his rear-end in alligators, it's hard to ask that question. It's a lot easier to justify this kind of inspection if lives are on the line, or millions of dollars. When it's not that obvious, it's not going to happen because it simply isn't worth it. In ten years time one shaft had failed in this manner, while in those same ten years, several hundred machines out of the same area had been overhauled.
To conclude, the cause of failure was death by natural causes, or torsional fatigue. The shaft was about seven years old and had approximately two billion revolutions on it.
|To see the close ups, click on the dots at the ends of the lines. For point "A", click to the left and to the right, as there are two pictures for that one point.
The picture for "C" shows how the crack initiated at the tip of the keyway and not at the sharp interior corner of the keyway. The beach marks in this picture are oriented perpendicular to the crack path. Looking closely at the cut keyway corner, there are no beach marks that begin at this sharp interior corner and then radiate outwards; this is the failure mode one would normally expect to see. However, there was clearance between the key and keyway. As the gear loaded up, the key did not stay in alignment with the keyway, and it cocked. Only the tip of the key was loaded, therefore, only the outer diameter of the shaft was initially stressed.
The picture for "B" shows a transition zone between the obvious smooth appearance of "C" versus the coarse chevrons to the right of "A". It closely resembles what one might call a tilt and twist boundary region. There is a mix of both features. In the close up picture for "B", these features are between two large beach marks whose ends are highlighted with arrows. Since the shaft was in torsion there was a twisting action from the torsion. This is no big surprise, and one would expect to see this somewhere on the shaft. There are many cracks initiating on different planes because of the twisting action. As the numerous cracks on different planes propagate away from the twist boundary and to the right, they tend to join up and make bigger crack planes. The final appearance is like that shown in the picture to the left, the stair step look of chevrons. Close up, it looks like a river pattern with many tiny tributaries that eventually make those big stair steps.
The pictures for "A" show twist boundaries and stair stepping chevrons. The picture to the left of point "A" is right at a twist boundary region. In that picture, the twist boundary is between two beach marks (highlighted with two red dots). In the picture to the right of point "A" the larger chevrons are propagating independent of each other. The red dots indicate a crack arrest point for each chevron. The chevron "steps" are pulled away from each other because of the torsional loading. Each crack traveled underneath the preceding crack, and then sheared at the vertical portion of the step.
The picture for "D" shows how the crack path wrapped around the back corner of the keyway. The manner in which the beach marks wrap around the corner indicates there was little stress in this region. This makes sense. If the direction of rotation were unknown, this type of beach mark configuration would be used to determine the direction of rotation. If the back side of the keyway was loaded, it would have looked more like the beach marks in "C".
|All Pictures and Text Copyright © 2000 - 2016 Contact Mr. Adler Adler Engineering LLC of Wyoming USA|
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.