Torsional Analysis for Pumps according API 610

[calcmike]

Dear specialists,

in the latest API 610 11 edition the chapter for torsional analysis enlarged. In 6.9.2.5 c.) it is stated, that 1x and 2n torsional excitations exist, even in the configuration motor-coupling-pump. Especially with the use of VFD it is possible to have a natural frequency in the operating range if excitaions from 1x and 2x have to be considered.

In API 684 there are references to Wachel/Szenasi and to Corbo/Malanoski.

In Wachel/Szenasi the reason for 1x/2x excitations are turbulence and load variations. Position is where the maximum calculated amplitude is found. Can someone explain this source or give me references ?

In Corbo/Malanoski it is written from "generic excitations" from unbalance, eccentricity and misalignment (pg. 197). A position in the shaft train is not mentioned. I do not understand why this should be a source of torsional (!) excitations. Can someone explain this or give me references ? Later Corbo/Malanoski reference to Wachel/Szenasi (pg. 204) and speak from a "dearth of Information".

Does someone have experienced damages from 1x or 2x torsional excitations in the past ?
Where does the potential excitation occure ? Where is in the motor-coupling-pump-system a torsional 1x or 2x excitation source ?

Altogether I ask myself, if ever problems with torsional vibrations with 1x or 2x occured in pumps or if this 1x/2x check comes from other reasons.

Best regards
Mike

 

[rotw]

Mike I suggest you also ask the questions to some of the person you referred to here on LinkedIn.
They are answering to some questions on specific LinkedIn Group.

 

[ccfowler]

One of the sources of harmonics is the effect of the keyway on the bending characteristic of the shaft as it rotates. The amount is generally very small, but it is not zero.

Long ago, I dealt with a potentially messy shaft replacement in an induced draft fan (several hundred HP) on a modest sized old power boiler, and the replacement shaft material that had been purchased and delivered before I was involved was inadequate to work safely if weakened by cutting the normal keyways. The problem was solved by securing the hubs to the shaft with a high temperature "Loctite" locking compound and no keyways in either the shaft or the hubs. All of the other "identical" fans ran very noisily even after extensive balancing. This fan with the "odd-ball" hub attachment ran spectacularly quietly from the first start. The careful static balancing done during assembly was all that was needed. I was just trying to put together something that would work, but I was then surprised by the magnitude of the effects of eliminating the keyway induced vibrations.

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[calcmike]

Thanks for your replies,

rotaryworld: I will do that. I thought, that a lot of pump user that work with API have to deal with this topic and so I am interested in your opinion and experience about this topic.

ccfowler: I would say that this is a lateral effect, but very interesting. This is a 1x excitation, or ? Does someone know how large this excitation could be ?

 

[electricpete]

I’ve never heard of torsional resonance problem on centrifugal pump unless driven by gearbox or recip. Seems like maybe blade pass frequency would be a potential torsional excitation – does API 610 mention that?

I have seen effects mentioned from keyway. Long keyway on a belt driven shaft fan – we had 2x vib that varied with belt tension (1x smaller and also varied). My logic was the vibrations changed depending on whether the keyway was inline with the belt.

Angular asymmetries in the shaft or coupling combined with misalignment can result in transmitted torque oscillation.

As a simple example, consider universal joint under misalignment

http://www.engineersedge.com/power_transmission/universal-joint-design-calculations.htm

There the answer is simple, the shaft changes speed to accommodate the coupling as given in equation labeled Figure 2.

We have couplings that are quite a bit more complex. But there are still similar effects present.

http://ipac.kacst.edu.sa/eDoc/eBook/828.pdf

“Shaft Misalignment and Vibration - A Model” by Redmond

“More importantly, equation (13) very clearly shows that parallel misalignment produces both first (1X) and second (2X) harmonic torsional excitation forces.”

“The system equations show clearly that parallel misalignment introduces a static displacement in addition to fundamental-frequency (1X) lateral and torsional excitation components. A discrete second-harmonic (2X) torsional excitation term is also evident in the system force vector.”

At least that’s the theory….


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

 

[electricpete]

I’ve never heard of torsional resonance problem on centrifugal pump unless driven by gearbox or recip or vfd

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

 

[electricpete]

There's a lot of math in that Redmond article and I'd encourage you to read it. I studied it once for a few hours and don't think I got it. Certainly don't remember anything now.

Here's my simple thinking torque transmitted through about parallel misaligned shafts. Represent the torque on the first shaft as two forces pointing opposite directions at points 180 degrees away, and same radius from the centerline. It is a pure torque, with no radial component. But if we translate those two forces to the other shaft at the same location, they are no longer equidistant from the new center. So the torque from one shaft created a a radial component of force on the other shaft. If the torque interacts with radial forces, then the reaction forces from the radial forces will interact from the force.

That's my simplistic view. It might be wrong, but it makes me happy.


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

 

[electricpete]

Whoops. I scrambled a few words...

Here's my simple thinking about torque transmitted through about parallel misaligned shafts. Represent the torque on the first shaft as two forces pointing opposite directions at points 180 degrees away, and same radius from the centerline. It is a pure torque, with no radial component. But if we translate those two forces to the other shaft at the same location, they are no longer equidistant from the new center. So the torque from one shaft created a a radial component of force on the other shaft. If the torque interacts with radial forces, then the reaction forces from the radial forces will interact from the torqueforce.

That's my simplistic view. It might be wrong, but it makes me happy.

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

 

[electricpete]

Some other ideas:
1 - if you have unbalance on horizontal shaft, the action of gravity on that unbalance reinforces the torque during one half of revolution and opposes during the other half of the revolution, so it adds a rotational speed variation in torque.
2 - what if shaft is orbiting within the casing at rotational speed. So it gets nearer the cutwater once per revolution for example. Wouldn't the torque vary as it repositions within the casing.

These may be small effects, but if torsional natural frequency resonance is at running speed and torsional damping is very light, perhaps they could have significant impact.

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

 

[ccfowler]

I'm sure the phenomenon (particularly its magnitude) is very specific to the actual geometry. In the case where I discovered the surprising effect, it was a forward curved fan with the shaft suspended between two bearings. The fan wheel was several feet in diameter with axial inlet flows from both sides. The fan wheel was attached to the shaft by two "spiders" each attached to its own hub "glued" to the shaft. Everything was very symmetrical except that the driven end of the shaft extended enough for its coupling hub which did have a conventional keyway. Under the circumstances, it quickly became apparent that the asymmetry of bending introduced by the cutting of keyways in the shaft was the controlling factor. The span between the bearings was several feet, and the significant inherent bending was the reason that the weakening introduced by keyways could not be tolerated by the available shaft material. The strong characteristic vibration observed at all of the other similar fans was at 1x shaft speed.

I no longer remember the exact dimensions, but I recall the fan wheel diameter being somewhere around 6 ft or a bit more, and the span between the bearings was probably on the order of 10 ft or a bit more. as I recall, the hub locations would have been somewhere in the neighborhood of 1/3 and 2/3 of the bearing span. They may have been a bit closer to the center of the span than the exact 1/3, 2/3 proportions, but the locations were exactly symmetrical within the span.

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[GaTechTheron]

I believe electricpete is correct. 1x & 2x torsional excitations are a result of rotor lateral vibration which causes twisting/movement of the shaft as a result. Think of this this way, when the rotor has an unbalance causing a vibration, the rotor is slightly deflecting. This requires energy, or rather it takes slightly more energy to spin a vibrating shaft. Energy is absorbed by the vibration being transfered into the bearings & bearing lubricant. The affect of energy absorbtion affects the torque required to drive the shaft, therefore a torque pulsation is induced on the rotor shaft.

Not sure if I am right, but this is how I have always understood it.

Note that I have also heard of a 1x electrical exictation frequency as well, but this is not in my realm.