High vibration in vertical turbine pump 3

[Prajesh Mistry]

Hi,
We have a newly commissioned vertical turbine pump of KBL make which is having the details as below:
Power: 180 kW, Flow: 2500 m3/hr, Disch. head: 20 mtr, Service: Raw water, motor RPM: 984
Problem:
The pump is having high vibration of approx. 30 mm/s at motor NDE, However we did the following steps one by one to eliminate the source of vibration but stills the vibration persist.
1. Foundation strengthening 2. Blue matching of motor with motor stool & 3. we also did pump overhauling (based on hard rotation of pump) to check the bush clearances & found clearances are very less however after increasing the clearances to normal clearance still the vibration remains high.
- One of our observation is the pump structure frequency is around 900 which is very close to pump running speed. May be this is the cause of high vibration? If yes how can we eliminate the same?

 

[Tmoose]

"newly commissioned" ......... "pump structure frequency is around 900"

Is the "high vibration" primarily at 1X rotating frequency? Or is it at a "blade pass" frequency? Or some other discreet frequency" Or, an "overall" reading and the combination of several smaller vibrations?
Is the vibration much higher in one direction? For instance pependicular to the discharge piping, or in the directoin of large cut outs in the "can" support between motor and pump.
As always, detailed frequency information is required for any meaningful analysis and thus effective repair.


Yes, a structural resonant frequency close to running speed can amplify normal 1X vibration, such as imbalance.
Vertical pumps OFTEN have resonant frequencies too close to running speed. In my experience the problem is often the various structural details from the baseplate up is deficient (flexible or weak).

What does the purchase contract mention along the lines of "no natural frequencies, reed frequency, etc within XX% of running speed " ?
If so, the manufacturer has an obligation to "fix it."
The fact you have undertaken so many mitigating steps suggests the manufacturer has abandoned you.
The mitigating steps you have undertaken also gives the manufacturer an "out", at least on paper, to "claim" no responsibility.

Useful ways to reduce 1X vibration.
1 - If there is a rigid threaded coupling below the motor, check the runout of the shaft. IF it is greater than ~ .002" the there is a problem how the shaft faces butt together. Something is keeping the faces from mating flush, thus creating a mechanical "crank" that will shake the motor at 1X.
2 - Trim balancing with weights at the very top of the motor ( one way clutch, etc ) often reduces 1X vibration for long periods of time.
3 - "DE-tuning" the system/structure by softening or stiffening can move the resonance far enough away from running speed ( assuming the vibration is primarily 1X rotation). This can range from softening by adding shims between faying surfaces at each mounting bolt, to any of a variety of stiffeners.
4 - Adding a tuned absorber near the top of the motor.

 

[mfgenggear]

I am taking a stab at this.
Pumps are not my expertise.
How ever rotating shaft assemblies is.
Did the machined parts get dynamic balanced?
Is there certifications?

 

[desertfox]

Hi

There is no mention of alignment checks or procedures between the pump and motor shaf,is everything aligned correctly?

“Do not worry about your problems with mathematics, I assure you mine are far greater.” Albert Einstein

 

[mfgenggear]

So some causes of out of balance which cause vibration.
When machined parts are pushed to extreme
Tolerances there will be vibration.

Shafts out of concentricity to each other
Will cause out of balance.

Run-out of all components needs to run less than .0005. any casted parts. Will require balancing.

Check for damaged bearing. Indicate with a mag Chuck any precision components accessible.

 

[BrianE22]

One of our observation is the pump structure frequency is around 900 which is very close to pump running speed

I think you are confusing RPM with Hertz.

 

[Strong]

You need to measure natural frequencies in orthogonal directions in horizontal plane (90-degrees apart) at top of motor. If natural frequency is close to 1xSS in direction with high vibration (as stated), and the natural frequency in direction that is perpendicular to high vibration direction has a higher frequency, then adding mass to top (or near top) of motor should be considered. The basic concept is to avoid exciting the natural frequency in the other direction when adding mass (as mentioned) or increasing stiffness of motor support.

I have achieved vibration reduction by improving shaft alignment and by rotating/indexing the motor and pump shafts at the coupling (trial and error.

Other corrective actions that can work well:
Eliminate poor grout under sole plate
Trim balance motor (as mentioned)
Add weight to motor (as mentioned)
Stiffen motor support and/or pump case/column
Tuned mass damper on motor (by others)
Spring plates (Belville washer style) between motor base and support (by others on job that did not work well)

Can you provide photos of motor and support?

Walt

 

[mfgenggear]

This only my observation and a rhetorical comment.

any time I see work arounds like this were hardware has to be repaired, by adding fixes is a problem.
it seems to be a continuous issue with vibration harmonics. a fundamental issue with the OEM design.
and it seems to be intermittent. I under stand it is a complicated design, however seems it would be
conducive for the OEM to prevent this in the first place. why would a customer who paid top dollar
have to go thru this.

 

[Strong]

"I under stand it is a complicated design, however seems it would be
conducive for the OEM to prevent this in the first place. why would a customer who paid top dollar
have to go thru this."

Typically assigning responsibility to a problem could be more complicated than the machine design for machine in power plants and other industrial plants. One or more entities may be responsible for the problem:
1) OEM for design and manufacture -- possibly different for driver and driven components
2) Architect/engineer for foundation and piping design
3) Construction and subcontractors for machine installation
4) Plant operations and maintenance personnel

It is generally best to accurately identify the problem, and possibly one or more corrective actions, before the arguments for responsibility begin. It is also very possible that the customer did not pay "top dollar". Occasionally a temporary fix can be installed quickly that allows a machine to operate while a permanent fix is being developed and financial responsibility is assigned.

Walt

 

[mfgenggear]

Hi Walt

yes it seem that the OEM would have minimum requirements for the installation, but yes thanks for that explanation.
I know that large mfg machinery, the depth, width & length of the foundation is very important.
I suspect same for the installation of the pumps.

 

[Prajesh Mistry]

Thanks for replying.

However i want to throw more light on the issue as below:
1. The dominant frequency is 1X with side bands of 80 CPM. Vibration is fluctuating between 15 to 30 mm/s.
2. The vibration is higher in one direction(along the discharge line) only.
3. During overhauling of pump, impeller was balanced @G2.5.
4. After loosening the discharge flange, natural frequency came down to 720 CPM from 900 CPM.


Can you elaborate the de-tuning method?
What are the methods to increase/decrease stiffness & mass to the said system. (Photos attached for reference.)

 

[desertfox]

Still no mention of alignment checks which is fundamental to a pump running without vibration, rather than getting to deeply involved in the theory resonant natural frequencies etc why not check the basics of the installation first because until certain things are ruled out everything is just conjecture.


“Do not worry about your problems with mathematics, I assure you mine are far greater.” Albert Einstein

 

[electricpete]

2. The vibration is higher in one direction(along the discharge line) only.

That fits resonance perfectly.

1. The dominant frequency is 1X with side bands of 80 CPM. Vibration is fluctuating between 15 to 30 mm/s.

That deserves more investigation.
How high are the sidebands relative to the main peak?
Based on the 50% fluctuation in magnitude I expect they are pretty high.
I'm assuming your vibration isn't being influenced by another pump nearby at slightly different speed.

What is the exact running speed? If this is a 6-pole motor running at 50hz supply (1000cpm sync speed) at a speed of 986.7, then the pole pass frequency would be 6*(1000-986.7)~80cpm. In general that suggests either rotor degradation or dynamic eccentricity. Current signature analysis can help discriminate.


=====================================
(2B)+(2B)' ?

 

[Strong]

I agree with E-Pete about the current analysis. If an electrical fault is not indicated, then I would suspect a loose (or very worn) bearing. I have seen the pole pass do this a few times.

Walt

 

[Prajesh Mistry]

But the vibration reading on motor solo run is very below, so this eliminate the motor issue.

 

[Tmoose]

"But the vibration reading on motor solo run is very below, so this eliminate the motor issue."

Was this mentioned before ? If not, why the heck not ?

Please describe the "solo" test. Specifically whether or not it was mounted in place on on the spool, etc just as in service.

==========

Not singling out Prajesh, but I really think that a standard questionnaire needs to be created to be completely filled out and attached to new posts. It will be at least one page long. Probably two. And at least 3 clear pictures and 2 drawings must accompany each Original Post. I think history has shown Relying on OPs to provide a reasonable amount of background information is a "capital mistake."

 

[electricpete]

Not singling out Prajesh, but I really think that a standard questionnaire needs to be created to be completely filled out and attached to new posts. It will be at least one page long. Probably two. And at least 3 clear pictures and 2 drawings must accompany each Original Post. I think history has shown Relying on OPs to provide a reasonable amount of background information is a "capital mistake."

Good point. Even the legendary Tmoose can't read minds!

But the vibration reading on motor solo run is very below, so this eliminate the motor issue.

Not at all. Many motor problems don't show on uncoupled run. Rotor bar problem vibration can be load dependent and only show under load (hence why current signature analysis is only conducted at high loads). Rotor bow causing dynamic eccentricity can be temperature dependent (so it may not show up on uncoupled run and may take awhile to show up loaded).

You never answered my question about speed, along with Tmoose's questions.



=====================================
(2B)+(2B)' ?

 

[mfgenggear]

Check for damaged bearing. Indicate with a mag Chuck any precision components accessible. [highlight #EF2929][/highlight]

 

[Strong]

I do not know if one or more of these videos show the best method for the instrument you have and for the machine/structure situation, but they are a good starting point:

Learn about Operating Deflection Shape Vibration Test
Search: ods vibration test video

https://www.bing.com/search?q=ods v...-24&sk=&cvid=DF51CF66C1FF4B3EABDAB913B0CFB0FC

Learn about Impact Vibration Test for Natural Frequencies and Vibration Modes
Search: impact vibration test video

https://www.bing.com/search?q=impac...-27&sk=&cvid=FE731D2B119F4E59BB85737E169C7316

Conduct impact test top of motor inline with discharge and perpendicular to discharge, both in horizontal plane. Repeat test with discharge pipe loose, if possible.

Conduct ODS test with Reference accelerometer (or use shaft tachometer for 1xSS phase) at top of motor and Rover accelerometer at the following measurement points in horizontal plane (inline and perpendicular to discharge):
1) Motor top or NDE bearing housing
2) Motor DE bearing housing
3) Motor Base (on flange)
4) Top of motor support pedestal/stool (on flange)
5) Bottom of motor support pedestal/stool (on flange)
6) Sole plate (remove grout, as necessary)
7) On top of concrete foundation/floor (use steel block/plate to hole accelerometer steady)

These two vibration tests will indicate the two natural frequencies near 1x shaft speed and how the structure is deflecting by elevation. Excel SS can be used to plot the ODS vibration mode shapes for the two data sets.

Provide one or more machine photos or drawings with this data, then one or more reasonable modifications can be suggested.

Walt

 

[mfgenggear]

nice post Walt, eye opener