Guidelines required for Refinery piping vibration assessment 3

[Naren]

I have a an extensive task of assessing the severity of vibrating pipes in a refinery.



I require practical guidelines & tools (software/ spread sheets, etc) to do this for the following areas.



Quick screening methods. Risk classification?

 

[mot1h]

I would suggest contacting the Southwest Research Institute, in San Antonio Texas.&nbsp;&nbsp;They have determined guidelines for displacement, velocity, and acceleration vibration values (all are equivalent measures of a given vibration) as functions of frequency.&nbsp;&nbsp;I don't know if their guidelines are available for purchase or if one has to be a member.<br><br>Ask for the Physics lab.<br><br>These people not only analyze vibration but model and design fluid and mechanical systems for vibration avoidance.

 

[VibrationSpecialist]

The bottom line is the thresh hold of stress caused by the vibration not to cause fatigue failure of the pipinn.&nbsp;&nbsp;You can start with 0.50-0.75 in/sec RMS or about 100 mils pk-pk as general.

 

[mot1h]

I am concerned about the tip from Vibration Specialist. The reply starts out correctly; vibration is symptom of repetitive stress cycles which may cause fatigue failure.&nbsp;&nbsp;As with any fatigue analysis, one needs to consider the magnitude of stress, and the number or cycles.&nbsp;&nbsp;Therefore, one must look at both the magnitude of vibration, and the frequency which over time determines the cycles.<br><br>This is why I refered to SWRI.&nbsp;&nbsp;One needs a graph or mathmatical function of both magnitude and frequency to analyze vibration severity.&nbsp;&nbsp;I have such information but am not at liberty to pass it out.&nbsp;&nbsp;<b>Values of magnitude like 0.50-0.75 in/sec RMS or 100 mils pk-pk are completely misleading.&nbsp;&nbsp;As an example, at 1.5 hz, .7 in/sec peak can lead to almost immediate failure</b>.&nbsp;&nbsp;The same amplitude of vibration at 300 hz is an acceptabel design value (fatigue is not likely even under continuous service at this vibration level and frequency).<br><br>The 100 mils peak-peak guideline is not applicable to piping.&nbsp;&nbsp;Maybe it is okay for shafts; I wouldn't know off hand.&nbsp;&nbsp;For piping, <b>100 mils p-p should not be tolerated at any frequency</b>.&nbsp;&nbsp;It would be expected to cause almost immediate failure at any frequency above 3 hz!&nbsp;&nbsp;Some companies use 10 mils p-p as a criteria but as I said this is way oversimplified.&nbsp;&nbsp;At about 2.5 hz, this would be a design level but at 210 hz, it could lead to almost immediate failure.<br><br>I strongly urge you to find a reference that provides the complete analysis of both magnitude and frequency.<br><br>You will note that I used units of peak and Vibration Specialist used RMS (root mean square).&nbsp;&nbsp;It has been a while since I have converted these but it is just a matter of math.&nbsp;&nbsp;I did not look up the conversion but I believe that .7 in/sec peak would be equivalent to .5 in/sec RMS.&nbsp;&nbsp;What matters is that you measure in the same units as are used by the criteria you apply, or that you convert properly to the criteria units if they are not the same.<br><br>Be thourough or your analysis could lead to disaster.<br>

 

[VibrationSpecialist]

As stated before, the bottom line is the stress that could cause the faliure.&nbsp;&nbsp;One need to determine the dominant frequency and the vibration amplitude at that frequency.&nbsp;&nbsp;The stress concentration factor could be obtained from published liratures depending on the piping lay out (mode shape should be determined in the vibration analysis).&nbsp;&nbsp;Then the piping material should be put into the picture to calculate whether the amplitude could cause an imminant failure or could be withstand.&nbsp;&nbsp;&nbsp;The value I gave of 0.5-0.75 in/sec RMS is really acceptable in most crude oil process.&nbsp;&nbsp;In fact this is my company guidlines that we gathered many statistical information in order to reach this guidlaine.

 

[mot1h]

Dear Vibration Specialist and Naren:<br><br>This is unfortunate.&nbsp;&nbsp;Naren needs advice and we are failing to reach a consensus.&nbsp;&nbsp;&quot;The bottom line&quot; is to avoid disaster and injury.&nbsp;&nbsp;That is my concern and I do not want to criticize other contributors.&nbsp;&nbsp;Nonetheless, some of your advice was oversimplified and I had hoped that when I pointed this out, that you would agree to it.&nbsp;&nbsp;Could you please just amend your advice with a specific frequency of applicability or maybe a physical application (e.g., 1800 rpm pump).<br><br>Since this did not happen, my concern is with Naren or any other reader of your recommendations.&nbsp;&nbsp;I ask you, Vibration Specialist,&nbsp;&nbsp;to address my specific comments:<br><br>1) Do you agree that the amplitudes you gave as acceptance limits, can lead to almost immediate failure at many realistic frequencies?<br>2) Do you agree that one has to consider both vibration frequency (fatigue failure cycles factor) and amplitude (fatigue failure stress factor) when evaluating vibration severity?<br><br>If you don't, please restudy your published literature.&nbsp;&nbsp;What you describe is a cookbook method of accounting for frequency, stress, and material allowable stress.&nbsp;&nbsp;The trouble is that you have recommended criteria that ignores all of the other factors, and only considers velocity or displacement amplitude.&nbsp;&nbsp;You cannot do that!&nbsp;&nbsp;Life is not always simple.<br><br>One must do a frequency analysis and get the individual frequencies and amplitudes, not just the &quot;dominant&quot; one.&nbsp;&nbsp;There is typically a fundamental and many harmonics involved.&nbsp;&nbsp;The one with the highest amplitude is not necessarily the most critical.<br><br>I am interested in the source of concentration factors.&nbsp;&nbsp;This sounds like a good refinement of what I have used.&nbsp;&nbsp;If you will share this, I highly recommend that Naren apply such concentration factor in his analysis.<br><br>Considering material allowable stress is also a refinement of the criteria I have used.&nbsp;&nbsp;Vibration severity graphs that I have used were undoubtedly for grade B, carbon steel pipe (70,000 ultimate tensile strength, and a specific ductility).&nbsp;&nbsp;This is a good illustration of what I am getting at.&nbsp;&nbsp;Even what I recommended has hidden assumptions.&nbsp;&nbsp;We need to check the applicability of our assumptions to each application.&nbsp;&nbsp;That is why I suggested that Naren needs to consult a nationally recognized specialist; not just you or I.<br><br>The source of your criteria reveals its shortcomings.&nbsp;&nbsp;Statistical analysis is only as good as the controls of the experiment or study, and similarity of the study to future applications.&nbsp;&nbsp;Your study is necessarily limited to the vibrations and configurations included in your study.&nbsp;&nbsp;Your criteria are probably just fine for your existing piping.&nbsp;&nbsp;Look out if you ever apply it to a different frequency!<br><br>Sincerely,<br><br>mot1h<br>

 

[VibrationSpecialist]

Dear mot1h & Naren :<br>First I would like to thank mot1h for his willingness to share his experience with others and to deliver tips that would not cause any misleading.&nbsp;&nbsp;&nbsp;As stated by mont1h that a full picture should be looked at and not to overestimate or simplify the situation.&nbsp;&nbsp;As you may know that vibration parameters (amplitude, frequency, phase and shape) need to be evaluated in diagnostics process.&nbsp;&nbsp;The amplitude types (displacement, velocity, acceleration) are related and&nbsp;&nbsp;a mathematical relation can be derived to convert from one type to another for the sinusoidal wave form.&nbsp;&nbsp;From the relation you can see:<br>velocity is proportional to speed times displacement and Acceleration is proportional to speed square.<br>The 100 mils pk-pk limit will cross the 0.5 in rms and a frequency of about 2.25 HZ.&nbsp;&nbsp;That means for a frequency above 2.25 Hz the limit should be 0.5 in/sec rms.&nbsp;&nbsp;I did not mention before that we have 1 g as the acceleration limit since this is will cross the velocity limit at about 122 Hz which is considered high frequency for piping random vibration.<br>Back again to mot1h questions:<br>1) Yes.&nbsp;&nbsp;Using the 100 mil pk-pk limit alone is not what intended.&nbsp;&nbsp;From the explanation above, the 100 mil pk-pk limit should be for frequencies below 2.25 Hz.&nbsp;&nbsp;For this range, the 100 mils is adequate.<br>2) Yes.&nbsp;&nbsp;Definitely.&nbsp;&nbsp;You need to look to the big picture.&nbsp;&nbsp;There is no limit on what you want to do perfectly and invest in finding the exact numbers. But on the other hand you need to be practical and time and cost conscious.&nbsp;&nbsp;When we collected the statistical data, we invested many hours of collecting these data.&nbsp;&nbsp;We relate it to failure to determine the appropriate level.&nbsp;&nbsp;As motioned in Naren in his first posted question that he works in a Refinery.&nbsp;&nbsp;&nbsp;The data we have gathered covered four major refineries and seven gas plants.&nbsp;&nbsp;&nbsp;I believe this what gave us the confidence in our data.&nbsp;&nbsp;But mot1h and others can still argue that these statistics are as good as our data collection process.&nbsp;&nbsp;&nbsp;I posted my first tip to give a simple answer to his request. If Naren are going to measure the vibration (we did not discuss how proper piping vibration measurement should be carried out !!!), and found that some of the piping vibrate more that 100 mils pk-pk or 0.5 in/sec rms then he should be very concern.&nbsp;&nbsp;If not then he should know (at least by now ) that further assessment should be carried out.<br><br>There are a lot of important tips when dealing with low frequency measurement.&nbsp;&nbsp;<br><br>Since this forum is text only discussion, I do not know to post some attachment for you review.&nbsp;&nbsp;If reference are requested on the concentration factors, please feel free to mail you request.<br><br>Sincerely,<br>

 

[Naren]

Colleagues,

I have not had time to respond to any of the discussion. I do apologise, however I did find it interesting & informative.

I will appreciate a copy of the doucuments mentioned by &quot;Vibration specialist&quot;... refinery piping data & concentration factors.
Other documents by others will be welcome as well.
My e-mail is as follows:
naren.sukaih@sapref.com

Thanks

 

[Franko]

For fairly comprehensive reports on machinery piping vibration and noise, contact Texas A&M for their Turbomachinery Symposium proceedings - for abstracts see

http://turbolab.tamu.edu/.

Tutorial in 1999 was by Engineering Dynamics Inc. and paper in 1997 was by Southwest Research Institute. Both companies can assist with general guidelines for piping vibration and specific problems.

 

[ralphsare]

i just log in. the severity chart prepared by SwRI is excellent but limited in application. Normally what we do is take the vibration reading (amplitude and frquency) and plot the mode shape then work with piping engr to evaluate stresses and possible fix either by rerouting or fix add'l support.
The other excellent reference is the Piping Design Handbook piblished by Puillman.
Recently, articles regarding the evalaution of piping vibration have appeared in the October 2000issue of the Hydro Carbon Processing mag

 

[electricpete]

Go to reliabilitydirect.com. Look for their section on Application notes. Select the App note on piping vibration.
(Please note I am not associated with reliabilitydirect.com).

 

I am looking at SwRI's chart for &quot;vibration induced stresses and allowable vibrations for piping spans&quot; 100 mil pp is certainly an over simplification, but it does work for a large number of pipe sizes and configurations. The formula they are using for allowable stress is based on a stress constant, that is a function of the piping configuration (u bend, elbow, etc.) and the diameter over the length squared. Interesting that frequency does not come into this equation. The allowable vibration is then a function of the allowable stress.

 

[dgallagher]

Here is a link to paper I wrote on piping vibration about 10 years ago:

http://www.reliabilitydirect.com/appnotes/pv.html

 

Guys. You have been discussing the symptoms. What about the cause of the vibraiton. The suction piping of two recycle gas compressors (reciprocating type) started to vibrate with in the 0.75 in/sec. However, I am really worry about this vibration. It goes back and fouth. There is no changes in the compressor perforamce nor in the mechanical side.

Could you please comment.

 

[Franko]

Sounds like resonance of the piping or the gas as &quot;it goes back and forth&quot;. Small changes can greatly change response of piping for centrifugal compresors Elliott Co. supplies; the same should be true for recips . If speed and performance conditions are constant, ambient conditions could be enough to change support stiffness. You may need a consultant such as those described on Aug. 22 pasted below:
&quot;For fairly comprehensive reports on machinery piping vibration and noise, contact Texas A&M for their Turbomachinery Symposium proceedings - for abstracts see

http://turbolab.tamu.edu/.

Tutorial in 1999 was by Engineering Dynamics Inc. and paper in 1997 was by Southwest Research Institute. Both companies can assist with general guidelines for piping vibration and specific problems. &quot;
Note: SWRI -

http://www.swri.com/

EDI -

http://www.engdyn.com/

 

[electricpete]

The exact same topic is being discussed at another forum right now. See:

http://www.reliability-magazine.com/ubb2000/ubb/Forum2/HTML/000981.html

 

[sms]

To Jeddah Refinery:

If this is a problem that has developed on a system that has not been a problem in the past, you might look at the pulsation bottles on the suction of the machine. Two common problems are either liquid build up in the suction bottle, or damage to suction bottle internals. Either will affect the performance of the suction bottle in reducing piping pulsations, and ultimately piping vibration. If the suction bottle is a knock out, make sure the drains are working right. If you can take a shut down, open the inspection ports on the suction bottle and make sure the baffles and choke tubes are still there and have not collapsed.

 

[ronfrend]

To Jeddah Refinery
I had a similar problem on a Solar Centrifugal Compressor in Oman. The problem changed with the time of day - worse at noon until about 5pm. Problem was that the sun shade had been removed for a project expansion and the piping was expanding - changing the natural frequency because the stiffness changed. Put the shade back and the problem went away.

Another thing to look at is if the volume flow has changed. This would affect the exciting force frequency which MAY be coincident with the piping natural frequency - hence resonance.