Morton Effect in Centrifugal Compressors

[RotatingEng11]

Hi,

One of our Compressor vendor has recently conducted the Mechanical Run test on the Centrifugal Compressor. The vibration value recorded is slightly more than the limit stipulated in API (25microns).
Vendor has mentioned something about the Morton effect in the Compressors. I tried to find some write up on this but there is nothing substantial on this topic.
please let me know, if you have any idea about this.


thanks

 

[rotw]

Morton effect has to do with the thermal stability.

Consider that there are elements on the shaft line that produce loads on a rotor section ; for example coupling(s), rotor weight, thrust bearing weight, etc. These loads induce a bending effect on the rotor. The oil viscosity, bearing span and the rotor stiffness also have an impact on the bending. Because of this rotor bending, the portion of the rotor that passes through the bearing cannot be completely aligned with the bearing longitudinal axis.

So when the shaft rotates during test or operation, a hot spot (@close clearance location) and a cold spot (@large clearance location) appear on the rotor across the bearing length. This creates a thermal gradient which in turn could induce a further bending of the shaft ; similar to a system that is self-feeding energy. If a certain threshold (details of criteria skipped here) is not exceeded, the cyclic situation is stable ; but when the criteria is exceeded the system is unstable with abnormal high vibrations (high level respective to API) as a consequence.

Typically a solid compressor OEM's would screen thermal stability too along with the standard rotor-dynamic lateral analysis verification.

In general, thermal stability problems are not noticeable or significant.
But there are configurations where such type of problems can lead to serious consequences.
Also one example is when casing is drive-through ; this configuration is often non-favorable because there are couplings on both DE and NDE side.

There are plenty of publications on Morton effect via google.

 

[Strong]

There are many pages of references if you use Bing or Google searches.

Here is a reference for a gas compressor engine turbocharger that I coauthored quite a while ago:
Faulkner, H. B., Strong, W. F., and Kirk, R. G., 1997, “Thermally Induced Synchronous Instability of a Radial Inflow Overhung Turbine,” Proceedings of the ASME Design Engineering Technical Conferences, Sacramento, CA, Paper No. DETC97/VIB-4174.

Walt

 

[awhicker84]

I wonder if you couldn't try to reproduce the actual bearing loads seen in the field on the compressor at test to see if the problem goes away.

You may have to increase pressure ratios to do so.

 

[awhicker84]

I also wonder what the effect of oil viscosity, oil inlet pressure and oil inlet temperature has on the problem. I haven't read much on the subject at this point, I'm just thinking aloud. Perhaps the bearings need to be re-designed or perhaps you can get away with a change of oil properties or perhaps after reproducing the bearing loads in the field the problem goes away (best case scenario).

Good luck and keep us updated.

 

[rotw]

Some of the "weapons" to recover in terms of thermal stability / Morton effect:

- Select a different bearing length
- Reduce safety factor for coupling rating (which takes into account driver type, etc)
- Reduce coupling weight (use lighter material, provided it can still cope with the torsional)
- Change monitoring probes location
- On multi casing trains, move one impeller or more from one casing to another.

In general, it is difficult to say which combination will work out as all of theses parameters / effects are kind of connected one another.
Also thermal stability and rotor-dynamic (lateral) stability often conflict. For example, reducing bearing length would normally help in terms of thermal stability but lateral stability would suffer (e.g. less damping). Corollary is that if you could marginally improve lateral stability by other means (shunt holes, swirl brakes, etc) it would in turn give some room to improve the thermal stability.

Luckily this kind of problems requiring fine tuning of the design does not happen often except on challenging train configurations.

 

[awhicker84]

Is it true that if the rotor is well balanced and stable that the Morton Effect would still have a minimal impact on vibration?

In other words, you may acknowledge the existence of thermal bending in the rotor, but you can solve it like you are solving unbalance?

I'm mainly an Oil Free Screw guy now, so I tend to think that way lately. However, this problem makes me want to read about bearing loading vs pressure ratio for CC's. It has been a while since I've ran a CC lateral, but I recall using only gravity for journal bearing load.

 

[rotw]

My guess is that balancing the rotor perfectly at speed (virtually) would not prevent the bending from happening.
The Morton effect is due to the bending, specifically due to the fact that there is a thermal gradient and a hot spot in the shaft (throughout the bearing length) and that hot spot occurs cyclically at the same location and so the energy is cumulative ultimately leading to instability.
This is just my best guess, this is a complex problem. Specifically I do not have the expertise that would allow me to say that there is no correlation between how well the shaft is balanced and the significance of the Morton effect. I would love to learn the feedback from expert in the field.

 

[Strong]

A perfectly (nearly) balanced rotor may inhibit the onset of Morton Effect, but once it starts and the shaft gets bent, then unbalance increases. For the project I worked on (see reference) the correction was to replace the journal bearing with pressure dam bearings of custom design. Avoid/correct a rotor balance resonance or support structure resonance is another positive action.

Walt

 

[rotw]

@Strong,

quoted
Avoid/correct a rotor balance resonance or support structure resonance is another positive action.
unquoted

Its a bit of a side subject but I understand that for centrifugal compressors, in most of the cases, the support structure stiffness is usually considered infinite (or quite high in relative terms). So could you please elaborate on the support structure resonance / corrective action? I am curious to learn about that.

 

[Strong]

"Its a bit of a side subject but I understand that for centrifugal compressors, in most of the cases, the support structure stiffness is usually considered infinite (or quite high in relative terms)."

The machine I worked on was a turbocharger for a pipeline gas compressor engine. Both the engine mounting and the test stand mounting had resonance issues. Rotor critical speeds (balance resonances) can also be an issue, but not with the subject machine. Be careful about statements of infinite stiffeness!

Walt