Вы находитесь на странице: 1из 9

6/26/13

cantilever vent line vibration - mass addition? - Topic

SKF

Become a better predictive maintenance leader in just 3 days Join or Manage Your Profile Posting Boards Asset Condition Management Posts About vibration/alignment/balance cantilever vent line vibration mass addition? Page 1 2 Go New Find Notify Tools Reply Login/Join cantilever vent line vibration - mass addition?
Tw eet Like 0 posted 05 May 2010 04:13 PM 0

electricpete

We have a 1" vent line that extends straight up from a LARGE pipe in a cantilever fashion. We see varying high vibration on the pipe up to 4 ips at a subsynchronous frequency approx 1500 cpm (machine operates at 5200). I think it is flow noise. Other similar sister machine pipes vibrate at 0.5 ips or less. We added 5-7 pounds weight and magnitude came down as well as frequency. Details attached. I think we are just lowering the resonant frequency below the range where the flow-induced excitation is highest. I realize lowering resonant frequency is a more logical strategy for fixed-frequency excitation rather than broadband, but it seems to have worked in thise case. Does it make sense to permatize weights similar to this? SGFP23_VentlLineSmall.ppt (958 Kb, 91 downloads)
Posts: 5545 | Location: Texas Gulf Coast | Registered: 20 February 2005

Ron Brook

posted 05 May 2010 04:28 PM

Hide Post

EP, Go ahead and add more weight and test it. You are adding inertia which will lower the displacement, but it would appear that there is also a natural frequency involved that you are moving away from. Ron
Posts: 916 | Location: Philadelphia,PA | Registered: 18 July 2006
maintenanceforums.com/eve/forums/a/tpc/f/3751089011/m/701106666/p/1 1/9

6/26/13

cantilever vent line vibration - mass addition? - Topic

Walt Strong

posted 05 May 2010 05:28 PM

Hide Post

El'Pete, That was a good start. A true rigid weight of 10-12 lbs might give a frequency redction over 20% (now 17%) and reduce vibration levels further. More weight would be needed if it was installed closer to valve body, so keep elevation high on stub pipe, as possible. Walt
Posts: 2147 | Location: Massachusetts | Registered: 27 April 2005

electricpete

posted 05 May 2010 09:14 PM

Hide Post

Thanks. We are going to clamp weight onto the pipe cap using device shown attached. The clamp weight around 6 pounds with provisions for additional weight in 3 pound increments, following Walt's comment. Our machinist says he will figure out the exact dimensions needed to get good clamping action on the pipe cap. We didn't try stiffening. My thought is that the initial direction would be increase in vibration with small amount of stiffening (based on weight addition results), but certainly could turn back down as we get stiffen farther. There is not a lot to brace off of as you can see it sits in the air. The adjacent blue structure is removed (during outages), so it would not be good to clamp to. If I had more time, I might investigate that further but we're far enough into weight addition that I'm going to stick with it (unless there are some obvious problems with this design anyone can think of). One the thing that makes me a little uneasy is the idea of tuning away from broadband flow noise. We are now a little less than full flow. Would frequency of flow noise be expected to change with change in flow? How about change in fluid temperature? Another interesting thing is that the magnitude of the vibration wanders up and down. Would this type magnitude variation be a typical expected behavior for a resonance excited by broadband flow noise? This message has been edited. Last edited by: electricpete, 05 May 2010 09:24 PM Clamp4a.ppt (1,136 Kb, 64 downloads)
Posts: 5545 | Location: Texas Gulf Coast | Registered: 20 February 2005

Ron Brook

posted 06 May 2010 06:58 AM

Hide Post

EP, The response of a component at resonance is always dependent on the amount of energy in the forcing function. The amplification factor doesn't change. If the 'broadband' energy you describe in this pipe doesn't have a defined flow resonance, then the amplitude of the response with change with % of flow. My guess is that there is a fluid resonance and the amount of energy
maintenanceforums.com/eve/forums/a/tpc/f/3751089011/m/701106666/p/1 2/9

6/26/13

cantilever vent line vibration - mass addition? - Topic

at the resonance could be affected by small changes in flow day to day. I like the design of the weights.
Posts: 916 | Location: Philadelphia,PA | Registered: 18 July 2006

Walt Strong

posted 06 May 2010 08:58 AM

Hide Post

El'Pete, An increase in flow velocity generally raises turbulence peak frequency and amplitude. Temperature change would probably have insignificant effect on structual natural frequency. I think that the add-on weight will be as good or better at higher flow/turbulence. The only negative about the weight design is that the additional weights are attached by bolts in shear, and this is not the strongest clamping direction. If you do add one or more plates to get to "magic weight" status, then weld the plates together and to one of the clamp halves. This would prevent loosening over time and reduce risk of losing parts should it be removed for any reason. Walt
Posts: 2147 | Location: Massachusetts | Registered: 27 April 2005

electricpete

posted 06 May 2010 11:08 AM

Hide Post

Thanks. Good comments. Now a crucially important aspect I have to describe this thing on paper. Everyone wants to call it a dynamic absorber, but that's not what it is. We all know it is a weight added to reduce resonant frequency. What is the short/quick catchy name for that?
Posts: 5545 | Location: Texas Gulf Coast | Registered: 20 February 2005

Walt Strong

posted 06 May 2010 11:45 AM

Hide Post

El'Pete, "Tuning Mass" "Detuning Mass" "NF Modifier" "Stubby the pacifier mass" "Magic Mass" Be aware that the valve and stub pipe may be acting like a "dynamic absorber" if the large pipe has a similar natural frequency. It would be worthwhile to inspect and/or measure vibrations at other locations in the pipe system to be sure that no new high vibration problems were created. Tip: Use thin SS shim stock to be sure there is a tight 2-point (each clamp half) contact to stub pipe. Walt
Posts: 2147 | Location: Massachusetts | Registered: 27 April 2005
maintenanceforums.com/eve/forums/a/tpc/f/3751089011/m/701106666/p/1 3/9

6/26/13

cantilever vent line vibration - mass addition? - Topic

William_C._Foiles

posted 06 May 2010 12:57 PM

Hide Post

Should the goal be to eliminate the branch pipe vibration or to just follow the pipe vibration? One can address the main pipe vibration separately. Mass can not only de-tune; it can change the mode shape. Regards, Bill
Posts: 2545 | Location: Houston, TX USA | Registered: 23 February 2005

electricpete

posted 07 May 2010 11:48 AM

Hide Post

The small pipe has enough access we could measure bottom, middle and top. Levels highest at top, lowest at bottm, in the middle between. The pattern stayed the same with our initial 5 pounds added mass. So I think it is 1st mode shape of cantilever beam and the main pipe is not vibrating excessively. LOL - I would have liked to seen my bosses face if I wrote "magic mass" in my paperwork. I like "Detuning mass" - it is very short and descriptive, and I would've used that if I'd have thought of it or read it in time, but I ended up using "vibration-reducing weight" (pretty lame). Operations wouldn't let us attach anything to the pipe cap since they have to remove it for fluid system venting during every plant startup. So we have to clamp to the pipe below the cap rather than to the cap. Also we have to keep enough clearance to operate the valves. I had to get agreement from a pipe stress guy. He wanted to restrict the total weight added to 15 pounds and ensure it was symmetric about the pipe (no moment). All of this forced the design to evolve to what is shown in attached. It has minimum 5 pounds (clamp) plus provisions to add additional weight in 1.8 pound increments. Our machinists built the device yesterday. It may be installed today or over the weekend or Monday. Will let you know results. FinalVibrationReducingWeight.ppt (1,224 Kb, 52 downloads)
Posts: 5545 | Location: Texas Gulf Coast | Registered: 20 February 2005

Steve Ciesla

posted 07 May 2010 02:42 PM

Hide Post

At least the name you picked meets the nuclear criteria for TLA (Three Letter Acronymn).
Posts: 442 | Location: Southern California | Registered: 23 February 2005

electricpete

posted 08 May 2010 09:09 PM

Hide Post

lol - good point Steve.


maintenanceforums.com/eve/forums/a/tpc/f/3751089011/m/701106666/p/1 4/9

6/26/13

cantilever vent line vibration - mass addition? - Topic

The "VRW" was installed today. I was not there, but the report is that the full 14.5 pounds was installed all at once (didnt try any intermediate mass level) . The vibration reduced to 0.20 0.23 ips on top of the pipe in both directions at a frequency of 1145 cpm. Vibration remained approx 0.20 at bottom of the pipe. I have no report of vibration level in the middle of the pipe (will check Monday to try to get an idea of the shape). Fitting this data against SDOF model: f = C / sqrt(M0+Madded) Madded = 0 pounds => we had 1555 cpm Madded = 5.5 pounds => we had 1290 cpm. From above first two datapoints we solve C=5418 cpm*lbm^0.5 , M0 = 12.14 lbm From the above we would predict: Madded = 14.5 pounds => 1050 cpm resonant freq (SDOF prediction) But we actually had Madded = 14.5 pounds => 1145 cpm resonant freq (actual measured) So a SDOF model does not fit very well accross this range of attached masses. This message has been edited. Last edited by: electricpete, 08 May 2010 09:28 PM
Posts: 5545 | Location: Texas Gulf Coast | Registered: 20 February 2005

William_C._Foiles

posted 09 May 2010 05:08 PM

Hide Post

Often one can brace the smaller pipe to the larger one. Could this not be done? When the frequency changes a great deal and when the added mass is a considerable fraction of the total mass it is possible that the mode shape has changed. The frequency reduction for this mode could be close to reaching a limit, i.e. close to the point where it would be with infinite mass (probably some room to go further if desired). Regards, Bill
Posts: 2545 | Location: Houston, TX USA | Registered: 23 February 2005

roughrider

posted 09 May 2010 05:44 PM

Hide Post

Reminds me of using a stabilizer on a bow, the longer the stabilizer the less weight required to prevent torque.
Posts: 12 | Location: alaska | Registered: 01 March 2010

Walt Strong

posted 09 May 2010 07:44 PM

Hide Post

El'Pete,
maintenanceforums.com/eve/forums/a/tpc/f/3751089011/m/701106666/p/1 5/9

6/26/13

cantilever vent line vibration - mass addition? - Topic

4 ips down to 0.2 ips is fantastic! The deviation from SDOF may be nonlinear effect of clamping the weight to the pipe as opposed to welding it. I think you had the right solution with the weight instead of bracing to the main pipe, and an infinite mass was not needed either. Paint it and file the results of the Case History"! Walt
Posts: 2147 | Location: Massachusetts | Registered: 27 April 2005

William_C._Foiles

posted 10 May 2010 07:58 AM

Hide Post

If you got a similar reduction in acceleration, did the force at the tip from inertia go down, remain the same, or go up? How about the dynamic moment at the intersection with the big pipe? Regards, Bill
Posts: 2545 | Location: Houston, TX USA | Registered: 23 February 2005

electricpete

posted 10 May 2010 10:29 AM

Hide Post

We did get a chance to take a closer look today. Vibration at bottom/middle/top in both direction was no more than 0.23 ips and no less than 0.17 ips (0.2 0.23 ips at top and 0.17 0.2 at bottom). We dont have a 2-channel analyser to take phase, but even if we did, Im not sure we would get a stable phase reading given the flow excitation. I summarized comparison of 4 configurations in attachment: Slide 1 original configuration 1555 cpm (1 4 ips) Slide 2 5.5 pounds weight (clamps added) 1290 cpm (0.6 1.4 ips) Slide 3 [NEW] 14.5 pounds permanent weight 1145 cpm (0.17 0.23 ips) Slide 4 [NEW] Sister units with different (smaller) valve style 2300 to 2700 cpm frequency. 0.2 0.45 ips among 5 sister units with smaller valve (we only have one with the larger valves... the one that originally had the 1-4 ips vibration) So, I conclude the sister valves likely have 1st resonance above the excitation frequency, and that the flow excitation falls of rapidly outside of a narrow frequency band somewhere in the neighborhood of 1550 cpm to somewhere below 2300 cpm. (I would have thunk it would be a wider range). I tend to agree with Walt that the only thing left to do is the painting. But certainly willing to discuss and learn what we can from this. Bill as I mentioned earlier there is nothing handy to easily stiffen against. Certainly would be more work to try to build something like that. What advantage would it have been to stiffen instead of adding weight?
maintenanceforums.com/eve/forums/a/tpc/f/3751089011/m/701106666/p/1 6/9

6/26/13

cantilever vent line vibration - mass addition? - Topic

I dont think stresses from shear forces will be relevant since this resembles a long thin beam and bending stresses (from moment) are typically much greater than shear stresses (from shear force) in a long thin beam. It was our objective to reduce bending stresses from moment near the bottom (as Bill mentioned) where this vent line is welded to the main pipe which should be much larger. I tend to think we have accomplished that objective simply because we tuned obvious resonance away from the excitation, although I dont have means to calculate any stresses to provide proof. In the initial configuration we had what I believe is excitation near first resonant frequency where 1st mode shape is cantilever beam 1st modeshape with highest moment and bending stresses at the base. Im not positive what the final mode shape looks like after weight addition, but my suspicion is we cannot get anywhere near the 2nd resonant frequency based on trying some preliminary calcs. The piping is 1 schedule 160, i.e. 0.815 id and 1.35 od. The length from big pipe to tip of vent line is 23. The 1st valve is attached 8 from main pipe, 2nd valve is attached 19.5 from main pipe, and permanent 14.5 weight is attached 22 from main pipe. The valves I estimated in the neighborhood 10-30 pounds. Thats a big range (uncertaintly), but anything within that range still gives us a predicted 1st resonant frequency in the very rough ballpark of our measurements, and predicts 2nd resonant frequency way up in the neighborhood above 6000cpm. If in fact the 2nd mode is that high in frequency, then we must still be seeing a first mode shape whose shape is not tremendously different than the original and I dont think there can be enough increase in curvature to overcome the dramatic decrease in peak vibration magnitude (4.0 ips down to 0.23 ips). My preliminary calcs considered the main pipe stationary. Will try again sometime with main pipe movement allowed (forcing function)... I imagine that will have some effect on the conclusion of how high the 2nd resonant frequency is. Always open to further discussion and comments. This message has been edited. Last edited by: electricpete, 10 May 2010 11:13 AM PhotoComparison.ppt (3,934 Kb, 33 downloads)
Posts: 5545 | Location: Texas Gulf Coast | Registered: 20 February 2005

electricpete

posted 10 May 2010 07:31 PM

Hide Post

One thing you guys may have noticed... there is a gap between the clamp halves. We did not shim as Walt had suggested. It certainly doesn't look right to the eye. My first thought is to wonder whether it will increase susceptibility of the bolts to fatigue (thinking about bolt joint diagram, where compression of the joint plays an important role). My 2nd thought, maybe the stretch in the clamp halves plays a similar role for this device as joint compression does for a typical bolted joint. What do you guys think? Note - if this device should fall off, there is nothing fragile below it and also it's not a place where people would be. This message has been edited. Last edited by: electricpete, 10 May 2010 07:43 PM vrw.ppt (446 Kb, 26 downloads)
maintenanceforums.com/eve/forums/a/tpc/f/3751089011/m/701106666/p/1 7/9

6/26/13

cantilever vent line vibration - mass addition? - Topic

Posts: 5545 | Location: Texas Gulf Coast | Registered: 20 February 2005

electricpete

posted 12 May 2010 07:44 AM

Hide Post

Heres a related question. Would you consider this fix as permanent ? We require 2 barriers for the high-temperature water in the main pipe. The pipe cap counts as one barrier while installed. The pipe cap is only removed for system venting when the system is cool during plant outage/startup. So we only need one of those two valves. We could remove one of the vent valves during the next plant outage (we cant do it on-line because the system is hot and we wouldnt have isolation to cut out the valve out, short of using freeze seal). As I said, the consequence of weight falling off is not a problem. The consequence of vent pipe breaking at the bottom weld from vibration fatigue would be big problem. Any opinions on whether this weight should be considered permanent, or whether we should go to the expense/effort of cutting out the extra valve (and updating drawings) instead?
Posts: 5545 | Location: Texas Gulf Coast | Registered: 20 February 2005

electricpete

posted 13 May 2010 10:19 AM

Hide Post

Thanks Sam. What are your reasons? fwiw - I tend to agree. There is of course one selfish reason - If the thing works fine forever, no-one will ever think about the weight or my involvement in it again. But if something goes wrong many years from now (lets say the year before I retire!), they will look up the paperwork and come straight back to me. If I were to champion a request for design change and work order to cut the valve off, I would need a little more justification. I could say the flow excitation might change over time which could eventually make the weight ineffective... but if that's a concern than it applies to both solutions (the weight and cutting the valve). Although I do tend to think we would have a lot more margin above the existing flow excitation frequency band before we get to resonance with one valve cut off than we have below the existing flow excitation band before we get to resonance with the weight installed.
Posts: 5545 | Location: Texas Gulf Coast | Registered: 20 February 2005 Powered by Social Strata

Page 1 2 Reply

Become a better predictive maintenance leader in just 3 days Join or Manage Your Profile Posting Boards Asset Condition Management Posts About vibration/alignment/balance cantilever vent line vibration mass addition?
maintenanceforums.com/eve/forums/a/tpc/f/3751089011/m/701106666/p/1 8/9

6/26/13

cantilever vent line vibration - mass addition? - Topic

Contact Us | Become a better predictive maintenance leader in just 3 days | Privacy Statement | Terms of Service
C opyright 2004-2013 Reliabilityweb.com All rights reserved.

maintenanceforums.com/eve/forums/a/tpc/f/3751089011/m/701106666/p/1

9/9

Вам также может понравиться