Reliability Engineering Snapshot TM

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



Machine Design - Case Study No. 120: Fracture from Flow Induced Vibration of an Axial Impeller


Sometimes forces that cannot be readily calculated or modeled are either ignored or assumed insignificant. Flow induced vibration should not be one of those parameters.

Compressor Impeller Side View

Figure 1

The impeller blade shown in Figure 1 failed in service. The section shown here is approximately 10" in length; it is 1/2" thick. The leading edge of the blade is on the left. The remains of the weld is on the bottom. The load side of the blade is shown. The front view of the impeller is shown in Figure 2 with the broken blade still attached. Rotation of the impeller in Figure 2 is clockwise.
Front View of Compressor Impeller

Figure 2


Close Up View - Impeller Blade Load Side Leading Edge

Figure 3

The broken blade shown in Figure 3 was inspected when the compressor shaft failed. The weld had fatigued along a 3" long section of the vertical heat affected zone (HAZ) of the weld. It was located about 7" in from the leading edge and was on both sides of the blade. In the region between 4" to 7" from the leading edge, the weld fatigued along the horizontal HAZ of the weld. The first 3" of fracture surface was located through the weld, on both sides.

Figure 4 clearly shows the progress of the beach marks from lower right toward the upper left. As the stress on the blade became greater, the beach marks fanned out, to the left, at a greater degree. Based upon the degree of fatigue and the small size of the shear lip, it was clear that the magnitude of the stress was very low. The unusual aspect of this failure was the fact that the fatigue started on the backside of the blade, not the front side as you might suspect.

Bottom View of Blade - Fracture Surface

Figure 4

So how does a blade get loaded on the compression side? Or should the question be rephrased as such, are there loads that put the backside of the blade in tension? If it is one thing that everyone agrees upon it is that cracks do not normally propagate in compression. So what could put the backside of a blade in tension? Possibly vibration induced by flow separation from the blade, or torsional fluctuations from erratic changes in the process flow. This fracture certainly suggests the possiblity. The failure indicates that whatever the problem cause, it was low in magnitude. However, it was not low enough and as a result it should not be ignored in design. There are methods to simulate flow induced vibration and this particular failure would warrant such an analysis. Unfortunately that won't happen. This blade went through 2.24xE10 revolutions. It may have instigated a catastrophic failure that will be showcased in a future article.


All Pictures and Text Copyright 2002 - 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.