Subsynchronous vibration in a centrifugal pump

[scooter1]

Hi,

Does anyone have any experience of subsynchronous vibration in a centrifugal pump?

We are recording very high shaft displacement amplitudes at 0.37X once the pump achieves running speed (3600 rpm). There are no indications of oil whirl as the pump runs up and the pump is operating below its first balance critical. The pump is a horizontal multi-stage unit.

 

[sms]

Are you sure it is running below the first critical? Typically multstage pumps run above the first. If it does run above the first it may be a re excitation of the first critical due to a rub in the wear rings.

If it really runs below the first, then I would be looking for some hydraulic forcing function, like a recirculation, or cavitation. Often times cavitation will come into the spectrum as a fuzzy hump around 30 to 40 percent of running speed. You should get suction and discharge conditions and the pump curve and figure out where the pump is running...

 

[SPLIT]

Hello scooter 1:

A serius candidate to be responsable of a subsynchronous vibration in a centrifugal pump is the whirl.
I am not an ezpert in this particula subject but I can recomend some literature :
- Alford and the destabilizing forces that lead to whirl and whip-Orbit, September 1998
- The death of whirl and whip-ORBIT
- Rotodynamics ofTurbomachinery-John M. Vance ( book, AMAZON )
The two first papers can be downloaded from Internet

Regards,

SPLIT

 

[vanstoja]

Subsynchronous vibrations in a centrifugal pump are usually associated with fluid film radial bearings in either the pump or its driver. For plain sleeve fluid film bearings whirl frequencies usually range from 0.43 to 0.5X dependent on fluid viscosity but for other so-called "stabilized" types like elliptical, 2, 3 or 4-lobed or pressure dam configurations whirl frequencies can vary over 0.3 to 0.5X ranges depending on operating conditions. For pivoted pad radial bearings (PPRB), which normally will suppress fluid film bearing whirl, damage to pad pivot pins that allow circumferential displacement of the pads can negate the pad load-alignment capability and bring on bearing whirl of the kind found in the "stabilized" bearing configurations until total destruction of the bearing is reached. We had several cases of ongoing PPRB pivot damage revealed by near 0.4X vibration peaks that survived for as long as 6 years. Post-failure inspection showed that pad pivot pins had punched through their pad holders and all three pads were in the between-pad positions.

 

[electricpete]

I believe it is a very complex subject above my head and you have already gotten input from two people I respect highly sms and vanstoja.

Here is an excerpt from EPRI Pump Troubleshooting

"Subsynchronous vibration is the most damaging and unstable type of vibration that can occur in a rotating machine. Subsynchronous vibration amplitudes have been detected at frequencies ranging from 0.3 to .9 times operating speed The first and most difficult step in troubleshooting subsynchronous vibration problems is making the distinction between rotor-dynamic and hydro-dynamic, or hydraulic, induced instability. This is a very difficult task and for years hydro-dynamic induced instabilities were not considered when a subsynchronous vibration problem was investigated. Rotor-dynamics was considered to be the source of all subsynchronous vibration frequencies, resulting in many elaborate and expensive rotor modifications that did not solve the problem. When this occurred, the problem was considered a phenomenon and was left unresolved. With the help of the utilities, hydraulic modifications were made by ERCO that solved the problems and failures experienced, and the phenomena became well-understood occurrences. Frequency 3: This vibration component appears in the vicinity of 1/2 x RPM (0.3 to 0.6). It is a self-excited, bearing-induced vibration instability. It is very damaging, and if it surfaces will result in rotor destruction, often without warning. A basic requirement for this to develop is a lightly loaded journal bearing, which is the case for most centrifugal pumps, particularly for vertical applications such as reactor coolant pumps (RCP, PCP, or RRP).
Frequency 4: This vibration component appears in a wider range of frequencies, 0.35 to 0.9 x RPM. It is the result of hydraulic forces developed when operating a centrifugal pump off or away from its best efficiency point (BEP) flow. Examples are given below with distinct frequencies as low as 0.35 and as high as 0.92 x RPM.
Combination of 3 and 4: This is the most difficult case to analyze. If the vibration frequency is about 0.6 x RPM, it could be dynamic, hydraulic, or a combination of the two. Vertical pumps, such as the RCP in nuclear applications, have very lightly loaded journal bearings and, hence, are prone to bearing instability. If the hydraulic excitation is just right, it will put the pump in the Frequency 3 category."

Roughly speaking how I interpret it is that there are two broad categories
"Frequency 3" - a whirl phenomenon related to bearing parameters and loading as described by vanstoja
"Frequency 4" - hydraulic forcing fucntion similar to described by sms.

Some ways to help distinguish:

#4 - Should be sensitive to system configuration. We have had 0.35x on vertical machines and it repeatably becomes worse when we go to the left on the curve, which would seem to confirm a hydraulic forcing fucntion.

#3 - Should be sensitive to oil temperature. Vary oil cooling and temperature and look for change in vibraiton.


=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[vanstoja]

Electricpete's EPRI troubleshooting evaluation of subsynchronous vibrations in centrifugal pumps and motor drivers is comprehensive and generally correct. My only reservations would be that bearing half frequency whirl may often be undamaging (it is bearing whip involving operations above rotor bending critical speeds that are potentially destructive). Also pump hydraulic induced subsynchronous vibrations can often be below 10% of rotational speed. These hydraulic instabilities which I didn't address in my prior input, are generally, off-design phenomena involving high incidence-angle-induced stationary separation or rotating stall in either or both vaned/vaneless diffusers and impellers. In a recent paper on hydraulic instabilities (Wang,H. & Tsukamoto,H. (2003), "Experimental and Numerical Study of Unsteady Flow in a Diffuser Pump at Off-Design Conditions", ASME/J.Fluids Engineering, Vol.125, No.5, Sept., pp.767-778) identified unstable flow ranges for their test pump as follows;
Q/Q_D=<0.4 - Separating flow and stall mainly in the diffuser
Q/Q_D= 0.4 to 0.6 Separating flow and stall in both impeller and diffuser
Q/Q_D= 0.75-0.85 Separating flow and stall in the impeller
Q/Q_D=0.6-0.75 and 0.85-1.2 Stable
Their test pump had a specific speed of 3289(US units), 6 impeller blades, 11 diffuser vanes and a cutwater clearance ratio of 1.03. The above unstable flow ranges may vary somewhat for different design configurations. Airfoil blade/vane stalls usually occur at inlet flow incidence angles above 10-12 degrees. Account needs to be taken of inlet flow rotation (coswirl or counterswirl) in evaluating inlet incidence angles at various flow fractions. Rotating stall propagation speeds (ie, fractions of rotational speed) depend on the direction of stall cell rotation wrt impeller rotation and the number of stall cells involved.