Sign up Calendar Latest Topics Donate
 
 
 


 
Poll Results
 
 In your own career, at what frequency have you seriously dealt with torsional vibration cases?
 Perhaps two times or less in the past 20 years 1 11%
 Around two cases every 10 years 2 22%
 Around two cases every 5 years 3 33%
 Around two cases every year 2 22%
 Another answer 1 11%
Total votes: 9   Please or sign up to vote.


Reply
  Author   Comment   Page 2 of 2      Prev   1   2
electricpete

Sr. Member
Registered:
Posts: 651
Reply with quote  #16 
Thanks, I appreciate it.  That's a lot of good info.

I found my photos from our last failure 3 years ago attached, which was a circ water pump.  Thought it might be interesting to walk through it to check whether torsional resonance may have played a role or can be ruled out.

Axial flow, 324rpm, 3-blade,  3 line shafts maybe 10' each.   

The shafts are coupled rigidly by sleeve and keyed type coupling as shown slide 1.  There are grooves in shaft where a split lock collar is inserted that engages with the sleeve.  The failure occurred at a change in diameter at one end of that groove. 

Can we rule out blade pass excitation of torsional resonance because it's axial flow construction? Maybe so, I'm not sure.  I don't have a real clear picture of what the stationary parts surrounding the impeller look like but I tend to think it's something like a conical shroud which is relatively uniform circumferentially (nothing like  sharp cutwater or diffuser vanes  on radial flow pumps). Note to self - check if we see any low level 3x and multiples in the normal / lateral vib spectrum. 

Low speed, low blade number means very low blade pass frequencyn 3x324rpm = 972cpm. But very long shaft so certainly the first torsional resonance might still be far below blade pass as I think Curran suggested.  But who knows without actually attempting a calculation/estimate. 

Now that I'm looking at it (slides 3, 4) it has the appearance I've seen described for other bending fatigue failures, with the beach marks spreading out from the keyway.  Do you think this appearance rules out torsional resonance?

ps, there was a lot more analysis done on this.  I'm not looking to re-examine all the possible causes, just wondering if we have enough to rule out torsional resonance on this one.

 
Attached Files
pdf cw12_pump_photos.pdf (2.89 MB, 15 views)

Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #17 

Jim, just like structural modes, you just have to make sure your not measuring on a node, and the chances of you measuring directly on a node are quite small.

If you are measuring torque, and you are probably measuring near the coupling, then you will be near a torque maxima for at least the first 4 modes. Even with a long, slender multistage pump shaft, the coupling will almost always have a high torsional gradient over it. If I have an instrumented coupling, I don't even really consider that I could be missing modes.

If you are measuring angular velocity with an encoder (or gearwheel, etc.) near the coupling, which is often an angular velocity node, then you are much more likely to miss a torsional natural frequency. Putting encoders at the NDE of either driver or driven ensures you are hitting an antinode, but you have to also take care of the driver moment of inertia and that it is not an order of magnitude higher than your driven, or the driver NDE will effectively have little participation in most modes.

Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #18 
Pete. That looks like textbook bending fatigue.
ivibr8

Sr. Member
Registered:
Posts: 163
Reply with quote  #19 
Thanks Curren -
Let me give your response some thought...and yes, my initial intent was to use an encoder or double-laser device.

Regards
JP
electricpete

Sr. Member
Registered:
Posts: 651
Reply with quote  #20 
Thanks Curran.  Yeah, after I dragged out the pictures and looked at them I came to that conclusion, it looks very familiar. And yet I know the guys analysing it carefully never came to any firm conclusion about it ... maybe they just didn't know the source of the bending... I'll have to ask them. 
John from PA

Sr. Member
Registered:
Posts: 963
Reply with quote  #21 
Quote:
Originally Posted by electricpete
Thanks Curran.  Yeah, after I dragged out the pictures and looked at them I came to that conclusion, it looks very familiar. And yet I know the guys analysing it carefully never came to any firm conclusion about it ... maybe they just didn't know the source of the bending... I'll have to ask them. 


e'pete, what struck me as a bit odd is the striations around the keyway.  Was any postmortem done as far as the key being fitted properly.  In practice there should be clearance at the top of the key and also some careful attention needs to be paid to the radius of the corners of the keyseat and key edge.
John from PA

Sr. Member
Registered:
Posts: 963
Reply with quote  #22 
Quote:
Originally Posted by Curran919

Jim, just like structural modes, you just have to make sure your not measuring on a node, and the chances of you measuring directly on a node are quite small.

If you are measuring torque, and you are probably measuring near the coupling, then you will be near a torque maxima for at least the first 4 modes. Even with a long, slender multistage pump shaft, the coupling will almost always have a high torsional gradient over it. If I have an instrumented coupling, I don't even really consider that I could be missing modes.

If you are measuring angular velocity with an encoder (or gearwheel, etc.) near the coupling, which is often an angular velocity node, then you are much more likely to miss a torsional natural frequency. Putting encoders at the NDE of either driver or driven ensures you are hitting an antinode, but you have to also take care of the driver moment of inertia and that it is not an order of magnitude higher than your driven, or the driver NDE will effectively have little participation in most modes.



I've attached a screen shot of an undamped Holzer analysis (Excel spreadsheet) of a cooling tower drive.

What Curran is referring to as far as a node you can see in the diagram at the right, where the torsional mode shape crosses the "0" axis at about station 9 which is the output shaft of the cooling tower.  The fan is a very large inertia and essentially the entire balance of the machine "beats" against the fan for the 1-node TNF.  That nodal point represents a great place for strain gauges, since the stress due to the vibratory torque would be high.  It would not represent a good point for measurement of torsional amplitude.

When I do a field torsional analysis, I like to precede it with an analytical study to show the various torsional mode shapes, at least out to the 3-node TNF.  I use that information to decide the best places for the measurement variables which might include torsional amplitude and/or torsional stress.

Untitled.jpg 
  



Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #23 
I've never done anything torsional on a fan. Interesting to see that they have the opposite 'problem' as pumps, where the driven MoI >> driver MoI.
John from PA

Sr. Member
Registered:
Posts: 963
Reply with quote  #24 
Quote:
Originally Posted by Curran919
I've never done anything torsional on a fan. Interesting to see that they have the opposite 'problem' as pumps, where the driven MoI >> driver MoI.


Driven MoI >> driver MoI is somewhat of a specialized case as this was a cooling tower.  Without looking it up, I’d guess from memory that the blades were probably on the order of 20 to 25 feet diameter and the hub of the fan was on the order of 6 to 8 feet.

When driven inertia is high compared to driver, that introduces the possibility of a unique situation if the driver is an induction motor.  Huge torques can be generated at switch closure and in the case of this cooling tower, which was started “across the line” these torques destroyed the gearbox, actually numerous gearboxes.  Ultimately the end user went to a soft start on the motors.


John from PA

Sr. Member
Registered:
Posts: 963
Reply with quote  #25 
Quote:
Originally Posted by Curran919
From a conversation I had with a colleague earlier this year, apparently shaft failure in vertical pumps (VS) tends to be lateral vibration related. The torsional system is essentially the line shaft is so flexible, that there isn't that much incidence of torsional modes in the bandwidth of interest. It is less common for a vertical pump to get torsional analysis than a BB pump of comparable power/speed.

As for failure, I think you just end up with the typical 45deg spiral break, as can be seen in John's 2nd link.


Very basic but a good article on vertical pump torsional vibration...https://www.peerlesspump.com/documents/tibs/TIB-25_ANALYSIS-OF-TORSIONAL-VIBRATIONS-IN-VERT-CENTRIFUGAL-PUMPS.pdf
electricpete

Sr. Member
Registered:
Posts: 651
Reply with quote  #26 
Quote:
e'pete, what struck me as a bit odd is the striations around the keyway.  Was any postmortem done as far as the key being fitted properly.  In practice there should be clearance at the top of the key and also some careful attention needs to be paid to the radius of the corners of the keyseat and key edge

Yes, those are reasonable questions.  To me the beach marks spreading away from the keyway might indicate that was the location where it started.   I was involved at the very beginning of the investigation... was assigned to go out at take pictures of everything and put it together in that powerpoint because the pump engineer was unavailable at that time.  But then the pump engineer took over the investigation completely, they got pump OEM involved, there are some reports somewhere.   To my memory they didn't have any smoking gun and we didn't identify any single significant corrective action from the investigation, just a bunch of minor tweaks.   But I am working from home these days so it's a little tougher to ask the pump engineer than it is when I'm at work where he sits right next to me. I'll ask him what they thought about the keyway when we are both back at work... which is tentatively planned for mid July ;-(

Quote:
Very basic but a good article on vertical pump torsional vibration

I notice they implicate a right angle gear drive as making a train much more susceptible to torsional vibration.  Do you have any idea why that is?
John from PA

Sr. Member
Registered:
Posts: 963
Reply with quote  #27 
Quote:
Originally Posted by electricpete


I notice they implicate a right angle gear drive as making a train much more susceptible to torsional vibration.  Do you have any idea why that is?


Actually the choice of wording is

Quote:
Systems consisting of an internal combustion engine, a right angle gear drive, and a pump are much more susceptible to torsional vibrations than those consisting only of a motor and a pump.


I would agree with that statement, key being “internal combustion engine”.  In my career I don‘t think I’ve ever seen an TV issue with a motor driven installation but I’ve seen many issues on diesel driver installations.
electricpete

Sr. Member
Registered:
Posts: 651
Reply with quote  #28 
Thanks - That makes sense, in an article limited to vertical pumps, any diesel engine  (horizontal) is going to need a right angle gearbox to drive the  pump.
John from PA

Sr. Member
Registered:
Posts: 963
Reply with quote  #29 
Quote:
Originally Posted by electricpete
Thanks - That makes sense, in an article limited to vertical pumps, any diesel engine  (horizontal) is going to need a right angle gearbox to drive the  pump.


Just speaking generically, the use of a gearbox of almost any configuration introduces a mechanism for the conversion of torsional energy into lateral vibration.  
Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #30 
Quote:
Originally Posted by John from PA


Just speaking generically, the use of a gearbox of almost any configuration introduces a mechanism for the conversion of torsional energy into lateral vibration.  


Yeh, this is it. Gearboxes make a modeler's life hell.
Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.