Coupling transient torque design - how critical?

[jeyaselvan]

This is primarily concerned to the coupling selection / design. Usually, we do the static torque design estimations (take some safety factor!)and free torsional analysis to ensure that we are away from opertaing in resonance. The coupling manufacturers tyically gives the peak torque as twice the continuous torque!

How critical is the transient torsional analysis requirements for the couplings?. The drives can be either engine or motor and the compressor/pump on the driven side.

Thanks in advance for you contribution towards this query.

Jeyaselvan

 

[ColonelSanders83]

First of all you are only looking at two of the three critical torque for a coupling torque analysis.

1) Peak torque (low number of events for life of coupling usually at startup, usually allows stresses in the coupling approaching yield)
2) Running torque ( A static torque the coupling is designed to in conjunction with three below)
3) oscillating torque (critical for application being driven by piston engines or driving recip compressors)

The oscillating torque is almost always the driver with coupling selection as the oscillating torque dictates the fatigue requirements of the coupling. This is the reason for the running torque being so much lower than the peak torque. In order to have a coupling last a transient torsional analysis is rather critical. To get accurate values a transient dynamic analysis of the system should be performed. I can recommend a specialist if you need one.

 

[ColonelSanders83]

I should also state the dynamic analysis is required in order to capure the interaction of all the distibuted masses (pistons, connectin rods, shafts, armetures) and springs (shafts and the coupling) to acurately calculate the peak, running and oscillating torques for the specific system you are sizing a coupling for.

 

[gwolf2]

I have had my fingers burned with piston engine transient torque vs mean torque calculated from power and rpm. In my case it was a friend's single cylinder two stroke with very little flywheel. The transient torque was 10x the mean torque and broke the prop shaft by fatigue quite quickly. Lucky it was just an aeromodel test on the workbench!

 

[GregLocock]

The factor of two probably comes from (a) experience and/or (b) the force multiplication seen in a suddenly applied load to a single degree of freedom system.

In my experience industrial drive units are rather conservatively designed, and it is possible to ignore the torque ratings to some extent, at the cost of being entirely responsible when they fail in service. So if you can take that risk, go for it. However, expect to be able to justify your decision.



Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[jeyaselvan]

Thanks for all of your feedback

We normally do only a static design & free torsional analysis for constant speed applications. A forced torsional response analysis is carried out for engine driven applications which has a range of operating speed. Since it is not possible to avoid opertaing at resonance, vibratory torque & power loss in the coupling are estimated and ennured to within the manufacturer's limits.

But the above are all steady state analysis. I am trying to decide on whether it would be necessary to carry out transient analysis. My observations are based on

i) For cases, I have seen with transient torques with motor applications, if we are able to push the coupling mode frequency to 20% away from the 1X & 2X motor excitations,
the estimated peak transient torque is less than or close to 2, most of the times.

ii) All the cases, the inertias of either sides of the coupling are fairly equal for the coupling mode. ( I haved some where the transient analysis is critical when the inertias are largely different on either sides)

Greg, is it a good reason for me to skip the transient analysis?

Jeyaselvan

 

[GregLocock]

How can I say-I don't know whether you are designing spaceships or swimming pool pumps?



Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[jeyaselvan]

Sorry Greg about that..As said in my first mail, they are typical industrial drive units...industrial compressors & pumps.

I was just trying to get a feel from the experts in field on my observations.

 

[ColonelSanders83]

What level of power are you talking about?

A coupling for sized for a 150 HP electric motor coupled to a pump is a far different aninmal than a 1000+ HP Caterpillar diesel coupled to a recip compressor.

 

[ColonelSanders83]

Another question to ask yourself is the cost impact of downtime. The units we typically evaluated had downtime losses of $20,000 to $200,000 a day. Compared to how much the company stood to loose, the time and money spent on a transient dynamic analysis to properly size the coupling was chump change (remember the coupling is the weak link, push comes to shove it will fail if not done properly, and oversizing risks equipment damage).

 

[jeyaselvan]

Thanks ColonelSanders for your feedback.

I am talking about power ranges from 15 HP to a maximum of 300HP, mostly the industrial compressors..I really wanted to have the coupling (commonly resilient type) as the weak link (as suggested), but at the same time do not want that to fail under normal machine opertaing conditions, steady state as well as the usual operational transients (start up & downs).

 

[ColonelSanders83]

For this type and size of equipment I would rely heavily on the coupling vendor for the requirements to size the coupling. If they say to do a transient dynamic analysis then I would do one. If they recomend a static analysis will be ok then go with that. The smaller your unit gets the cheaper the coupling should become comepared to one time engineering costs. At the end of the day you need to decide if you want to know whether the coupling will operate trouble free before it gets to the field. People will be coming to you if it fails every six months for a particular application.

 

[jeyaselvan]

Thanks to all for your feeback on this.