Reliability Engineering Snapshot TM

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



Vibration Case Study No. 17 - Axial Screw Compressor Cracked Case at Journal Bearing

Rotary Lobe Screw Compressor; intenal view with lobes removedIn vibration analysis sometimes it's not so much knowing what the problem is, but rather that the signature just doesn't look right.

This first article is dedicated to all those new vibration analysts out there that are taking vibration readings for the first time and come across their first situation where the famous words are uttered: "What is that? I don't know but it just doesn't look right." This one's for you...

The picture to the left shows the inside of a rotary lobe screw compressor from the "suction side" of the compressor, and looking toward the "discharge end" of the compressor. There are two shafts which fit inside, each one shaped like a screw, and both messing together; the way in which they mess conveys the gas stream from the suction side to the discharge end and increases the pressure of the gas stream along the way. The higher pressure gas stream discharges through the bottom middle hole in this picture. The top two holes on either side are where the discharge end journal bearings are located for the two shafts.

This rotary lobe screw compressor had just been overhauled for its second time after having been taken out of service due to a very high increase in vibration (see Figure 2). The overhaul went smoothly without a hitch. When the unit was started back up and put on-line the vibration dropped back down to a normal level. By vibration level standards, the problem had been solved because the vibration was back down to its historically low and typical level. However, the vibration frequency spectrum was all wrong, and the waveform pattern was all wrong (see Figure 3). It's moments such as these that makes one wonder whether the data is right or whether the vibration analyzer is wrong. What was intimidating to say the least was the thought of telling the maintenance manager that something was not right in a machine that was painstakingly overhauled; and then not even being able to tell him what that "something" was. So at this moment of truth I remembered a phrase that a man by the name of Art Crawford once said at a conference presentation, he said "The little black box never lies." The little black box being the vibration analyzer.

This is the time known as "the leap of faith," because you're all alone; and there will be hell to pay if you're wrong. It is at a point such as this that one had to step back and really look at the data. The first question asked was: "What's not wrong with the compressor?" We knew what a bad journal bearing looked like, what a rub looked like, what a thrust bearing going bad looked like; none of these things were happening. So if they weren't going bad, then what was? Well......." I DON'T KNOW, BUT IT JUST AIN'T RIGHT!" So the compressor was taken off-line and disassembled.

Close Up of Compressor Internal With MapThe journal bearings looked good, the thrust bearings looked good, the rotor lobes looked good, the balance drums looked good; everything looked good. Well of course they all looked good, none of that was showing up on the vibration diagnostics! It was really looking like that "little black box" finally blew the call. The compressor body was placed into the parts washer prior to reassembling the whole thing over again, it looked as though crow was going to be served for dinner. When the compressor came out of the parts washer and the mechanics were getting ready to insert the journal bearings, they noticed two small cracks. The cracks were located at the thin section of the compressor body (see figure). This small detail was almost overlooked because dirt and grease had hidden the damage. The cracked sections of the compressor body eliminated the press fit which held the journal bearings in place, and therefore the bearings were actually loose and not providing any support to the two rotary screws. If the unit had been reassembled and placed back into service it is very likely that the same problem would have persisted.
A typical vibration signature for the male drive rotary lobe is shown at the left. This signature is for a good running compressor, and is unique in that the four lobes create multiple harmonics throughout the frequency range. The vibration software easily highlights where all of the harmonics are located with the red squares. The time waveform is also unique in that there is basically four sinusoidal waves per revolution. The black vertical lines in the lower graph indicate each revolution of the drive lobe.
The compressor had been recently overhauled and when it was put back on-line there was some difficulty with water and dirt contamination of the lubrication system during start-up. It was obvious that the compressor rebuild had been ruined and that it would have to be overhauled a second time. A follow up failure analysis had revealed that a critical filter had ruptured and allowed all sorts of debris into the system; subsequent flushing of the system had only introduced additional contaminates and water . The journal bearings had been wiped. The red square in the top graph indicates the 4X fundamental frequency of the male drive screw, note the significant difference to Figure 1. The waveform has completely changed too.
The compressor was rebuilt and put on-line for a second time. Although the vibration level had returned to normal levels the signature did not look right. Note how the harmonics of four times running speed (4X) are gone. The dominating frequency in the spectrum at the top is two times running speed (2X). The waveform at the bottom is also dramatically different, showing two sinusoidal waves per revolution of the male drive lobe. Armchair quarterback vibration analysts will of course say that these signatures are signs of "classic looseness," but it wasn't quite that obvious under the circumstances. There are times when you just aren't going to believe what you see, and you're just going to say something like "it just doesn't look right." And that remark, although scary, is good enough.

When the compressor was opened up for inspection it revealed that the two discharge end journal bearings were loose in the compressor. The compressor body had cracked where the journal bearings were pressed in; thus allowing the two bearings, one on either rotary lobe, to literally rattle inside the compressor (see above close up figure of discharge end of compressor, click on where you think the cracks are located, and if you're correct a close up of the crack will appear).

A new compressor body was used, and the compressor was rebuilt for the third time. This time the vibration signature returned to its normal configuration.










All Pictures and Text Copyright © 1998 - 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.

Crack Crack in Casing