Sign up Calendar Latest Topics Donate
 
 
 


Reply
  Author   Comment  
JohnJ

Member
Registered:
Posts: 66
Reply with quote  #1 
I collected multichannel data on a turbine (4070 rpm), Lufkin gearbox (49/113 teeth), two-stage pump (1765 rpm).  The vibration on the low-speed side increases considerably during coast down. The 1X vibration on the low-speed side of the gearbox is below 0.5 mils at full speed.  As the machine coast down to 76 rpm, there is a steady increase in vibration to 2.9 mils on one of the probes, but all the probes on the gearbox low-speed side increase in vibration during coast down.  No big change in phase to indicate a critical.

If I compensate the data, the full speed vibration increases, but the coast down data decreases as the speed decreases.  I am having trouble trusting the compensated data.  If the increase in vibration during coast down is run-out, then there must be an unbalance effect that is counteracting the run-out at full speed.  I doubt a Lufkin gearbox has that much run-out, and I doubt the shaft/gear would be out of balance.  I have attached the uncompensated and compensated bode plots for the low-speed shaft-end bearing and the uncompensated bode plot for all four probes on the low-speed side.

Opinions on whether or not to trust the compensated data?

Thanks,
John J

 
Attached Files
pdf GBLS uncomp and comp data.pdf (54.01 KB, 11 views)

John from PA

Sr. Member
Registered:
Posts: 963
Reply with quote  #2 
Keep in mind what compensation is attempting to accomplish.  If you are using 1X compensation, the idea is to vectorially subtract out the effects of the physical properties in the target area to enable a better feel for real dynamic motion.  But gearboxes are difficult to do, because one rotor has decreasing forces on it and may move from the upper half of the bearing to the lower part and load and speed drop off.  We of course don;t know anything about this gearbox so a sketch with rotations and correct hand of assembly might prove helpful.

Again, keeping in mind what 1X compensatio is doing, I would pick a point at about 400 RPM and aplply that to the data.  (400/1800)^2 is about 0.05 so the forces at 400 are about 5% of what would be expected at 1800 RPM.

By the way, what type couplings are being used?  The characteristic of high vibration at slow RPM, dropping to near "zero" at some speed, and then increasing is a characteristic of two bolted flanges that are not concentric and hence they introduce a crank effect.
JohnJ

Member
Registered:
Posts: 66
Reply with quote  #3 
John PA,

Thanks.  The couplings are flex disc.  The gear arrangement looking from the driver (turbine) pinion (49t) turning clockwise with the gear (113t) on the right.  Pressure dam bearings in the gearbox with the dam on the bottom of the pinion and on the top right of the gear. The turbine stops very rapidly-- from 4070 rpm to 0 in 48 seconds.  I was able contact a good friend of mine and he said the deceleration is probably too fast and the gears are probably climbing back on each other.  I tend to agree with that because, as I stated in the original post, I don't believe Lufkin would put out a gearbox with that much run-out or unbalance.

Thanks,
John J
John from PA

Sr. Member
Registered:
Posts: 963
Reply with quote  #4 
Flex disk coupling can rely on body bound bolts for centering but with the wrong bolts all bets are off.  Sometime when the machine is down, you might tram in the flanges to make sure you aren't introducing a crank effect.  Go to https://www.maintenance.org/fileSendAction/fcType/0/fcOid/399590942963403109/filePointer/399590942964872886/fodoid/399590942964872884/Vibration_Analysis_Pinpoints_Coupling_Problems_Maxwell.pdf for more on the crank effect.  See page 2 "off center torque tube" but keep in mind a space isn't necessary, just two flanges bolted eccentric.

As far as decel, what the driven machine being a pump, a 48 second shutdown could be normal.  Do you have past data to support that this is fast, or is it just "gut" feel?
JohnJ

Member
Registered:
Posts: 66
Reply with quote  #5 
John PA,

The friend that I contacted, who is a well experienced analyst (Category IV) felt the turbine may be decelerating too fast.  He said he hasn't analyzed many turbines under 1000 HP and this one is 580 HP and it may be that there is that much run-out accompanied with an offset unbalance during full-speed operation.  The shaft center line plots indicate the high-speed blind end bearing is running outside the bearing clearance, which is probably aggravating the problem.

Thanks,
John J
John from PA

Sr. Member
Registered:
Posts: 963
Reply with quote  #6 
Quote:
Originally Posted by JohnJ
John PA,

The shaft center line plots indicate the high-speed blind end bearing is running outside the bearing clearance, which is probably aggravating the problem.

Thanks,
John J


Make sure you are using hot shutdown data for the slow roll, both for dynamic data and the DC gap values.

Is the turbine condensing or non-condensing?
JohnJ

Member
Registered:
Posts: 66
Reply with quote  #7 
John PA,

The turbine had been running for days --- I learned that lesson several years ago the hard way. 

The X and Y signals from both turbine bearing only show about 3 to 13 degrees phase shift (flat orbits) and I found out they relieved the top of the turbine bearing because they had been overheating.  The turbine is oil bath with a slinger ring and not pressure fed.  The gearbox and pump have pressure oil going to them.

The turbine is non-condensing.

Thanks,
John J
Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.