Can beating frequency in a srew compressor be at 22 Hz? 2

[mechanicaljw]

Hello All,

We are having a screw compressor that has 3 male lobes and 5 female flutes that runs on a 90kW, 4 pole, 50 Hz motor with synchronization speed of 1480 rpm. However, at a speed of 1350 rpm, we are experiencing a resonance frequency of 22 Hz. We are unable to pin point the source of this frequency and to tell whether it is torsional or structural? Because the driver is the female i computed the difference between 2 times line frequency, which is 45Hz and the excitation frequency and got 22.5 Hz. In our measurements we are getting 22 Hz and error of about +1 as the resonance freuqncy. So at times we are close to 23Hz. Since i know the beating frequency to be result of two frequencies interacting, i am woundering whether this is the cause of our problem? Or am I doing something wrong? Because for beating to occur from what i have read the two interacting frequencies are usually quite close to one another? What is also happening is that at this 1350 rpm where we have this resoance the torque suddenly increases. So we to figure out whether the this is as a result of the resonance?

Any thoughts would be appreciated.

Thanks!

Jimmy

 

[jeyaselvan]

Hi Jimmy,

Sorry for commenting a little late on the thread.

I have earlier worked quite a lot on a beating noise problem with a screw compressor, for which you can see more information on thread384-276533 Beating / modulating noise and vibration.

My question here is " Are you concerned about the beating noise for frequencies around beat frequency of 22Hz"? The reason I am asking this is that usally beating will be of audible concern only if it is in the range of 0.5 to 5 Hz typically. Beyond this, perceptionally we may not able to identify a beat.

Is it that your case is a typcial case of resonance or rather operating closer to a resonance?. I have earlier faced and measured a typically beat like torsional response, while I was operating closer to a torsional natural frequency of the drive train. I remember even in that case, the dynamic torque goes very high is of concern.The strange noise is common when you are operating at or close to the natural frequency.

Based on the informtation you have provided so far on the shift in natural frequency of the drive train for engine, electric motor or hydraulic motor are typical behaviour of drive train with varying torsional inertias. My suggestion would be that you do an analytical torsional analysis based on drive train inertias and stiffnesses based on lumped masses approximations and verify they reflect the same as you measured. Normally, the first few torsional modes are of concern.

Regards
Jeyaselvan

 

[mechanicaljw]

Hi jeyaselvan,

Thanks for writing and for your insight. Never mind that you responded late. It is better late than never. I have actually seen your previous posts on this subject and i read it extensively to see if i could find a clue to our problem from it. I have also taken what you suggested in you recent post. I think like you rightly said it cannot be a beating problem and we are focusing more on torsional resonance probably due to the drive train. One of our approaches has now been to numerically predict the torsional modes we are measuring and then we hope to tune the system away from torsional resonance frequency by making some changes to the shaft design. Our problem is that we have not been able to match the FEA stiffness we are predicting with that that would be needed to give us the measured torsional lst mode frequency of 40 Hz. For the point mass model we have we are using FEA to determine the shaft stiffness that we seem to be over-estimating and so we are getting a numerical 1st modal frequency of 60 Hz. Did you have to do something like this and if yes how did you estimate your shaft stiffness? We are using ANSYS STRUCTURAL to do the FEA.

The other approach we are trying is to use torsional damper to see if that would help.

Thanks!
Jimmy

 

[jeyaselvan]

Hi Jimmy,

Is it possible you can do a runup and establish your torsional resonance (possibly a speed fluc meas on flywheel ring gear)? It looks that the torsional resonance appears to be your first resonance, being lowest,and often seen in these applications. Do you have a resilent coupling in the drive train? If so, you may reduce your lumped mass systems can reduced be less than 3 or 4 for a better prediction, (ideally two would be enough) without going into the details of the stiffness of participating elements, instead you may lump their inertias. This is because your first mode will often be an coupling mode (with my experiences with engine and motor driven compressor dynamics, am an OEM designer concerned about dynamics of similar compressors) significantly influenced by the inertias and weakest link, which is often the resiient coupling with engine driven compressors. Be aware that you use the dynamic torsional stiffness,which is a strong fn of temp and load. If this is so, then you can think of a different flywheel inertia of engine (I have addressed issues like this in the past, usually you can have varying inertias upto 30%) as a remedy, instead of additional torsional damper, which your engine may already have to satisfy torsional deflection design requirements

By the way, where is the source of 1X excitation arising from with engines in the actual end application? what type of engine is it? 3,4/6 cylinder..any of the case, it is hard to have a first order excitation, unless you have severe misfiring resulting in half order excitations.

Regards
Jeyaselvan

 

[electricpete]

Is the original/standard shaft hollow?

If so, fwiw I think it reinforces the conclusion of torsional resonance (also strongly supported by the torsional campbell diagram).

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

 

[electricpete]

If so, fwiw I think it reinforces the conclusion of torsional resonance (also strongly supported by the torsional campbell diagram).

I should've clarified what I meant by "it".
What I meant was
If so, fwiw I think your experiment of changing hollow to solid shaft and examining effect on torsional oscillation magnitude and frequency reinforces the conclusion of torsional resonance (also strongly supported by the torsional campbell diagram).

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

 

[mechanicaljw]

Hi Jeyaselvan,

Thanks again for sharing and your input. It has been helpful. I will try to answer some of your questions and for those that i did not understand i will appreciate it if you can explain in some details.

1. About whether our drive train is resilient. I would say yes and we have approximated the system as a two point mass system like you said in your post. This was because in previous calculations the mode shape for 1st mode shows all masses rotating relative to prop shaft. Due to rigid body motion the 1st mode freq' was always therefore governed by K more so than I1 and I2. So for example if you doubled the inertia of the 1st stage input gear mass the effect would be negligble
So for quick rough calculations we used 2 mass + k

2. The question for us now is how to determine the stiffness. The k we have been using so far is the static k but from your previous post it must be the dynamic k?

3. Do we know the source of the 1x excitation from teh engine with actual end application? I would say the source is the input torque from the engine or the electric motor.

The type of engine: we have in the real drive a 9.6L 234kW in line 6 cylinders diesel engine with firing order 1-5-3-6-2-4, firing every 120° of crankshaft rotation. The electric motor for the test stand in the lab is a 90kW 4-pole 50 Hz driven by ABB ACS800 inverter VFD.

4. Quote "In any of the case, it is hard to have a first order excitation, unless u have severe misfiring resulting in half order excitations".. Can you please explain this further. I do not understand what you mean exactly. Are you saying that having 1x excitation like we do is because of misfiring? Not sure what misfiring here is referring to?

Thanks!
Jimmy

 

[jeyaselvan]

Hi Jimmy,

1 & 2. Yeap. You can easily tune away the coupling torsional mode by changng the stiffness of the coupling. Usually you get the dynamic torsional stiffness from your coupling manufacturer. If it is not available, you may estimate this by back calculation from your measured torisonal resonance. By the way, how are you measuring dynamic torque? is it by an inline Torque Transducer(TT) / strain gauging / encoders?. If by TT, then they might change the behaviour of your drive train. If you do a run down at full load after running the machine for sometime, you may get a pretty decent resonant frequency.

3. Regards to the excitation, the engine you have stated have third orders and harmonics as predominant excitations. I am not able to understand as to what is causing 1X excitation. I think the answer to this will take you to move forward. By the way what is your operating speed range in the end application? If the system is mounted on resilient mounts, ensure that mounted resonant frequency is away from the 22Hz. You may decouple the cooling fan and check whether the same is from the fan I have earlier had a similar issue of around 22Hz coming from the fan, but that was for linear vibrations.

4. A typical 6 cylinder would have predominant torsional orders of 3,6,9...and your airend would have 7.2,14.4,.... as given by the configuration. If you have a misfire in one of the cylinders of the engine, the lower orders do not get cancelled and then there is a high possibility of 1X excitations. Given your configuration, I do not see any reason for the source of 1X torsional excitation.BTW, do you have the 1X excitation with electric motor as well? I think tracing this help you to nail down.

5. I did also see that you are driving the female rotor ( believe this is an oil free compressor) , which is done very rarely, since your synchronising gears needs to transmit the torque as well, which is not really a better option. May I know whether you have any better reasons for this?

Regards
Jeyaselvan

 

[mechanicaljw]

Hi Jeyaselvan,

It is always great reading from you. Thanks for your support. I will try to answer some of your questions and would add some more questions.

1 & 2. The dynamic torque is being measured using string gauge on slip rings

3. I am trying to understand why the type of engine we have should have 3rd orders and harmonics as predominant excitations. Do have a citation that i can fall on to try to read more? Especially the fact mis firing could be exciting lower order modes?
The operating speed range in end application is 1000-1800 rpm. About decoupling the fan, the machine is oil cooled and so we do not have a fan but only an oil pump with a capacity of about 3L. We have not decoupled it to see whether it would affect the excitation or not.

4. I am not too sure about cancellation effects of lower torsional modes in this engine and would be grateful if you can provide me with citation that i can read to help me understand if further. With the electric motor we seem to be having the same 1x excitation in that when we run at say 1350 rpm the resonance is about the same as the rotational speed (22.5 Hz). In the case of the diesel engine for the volvo truck, we have resonance at 11/12 Hz. The 11/12 Hz seems to be excited (to varying degrees) between comp speed of 1000 - 1800 rpm. This is due to the forced frequency of the engine opposed to compressor input frequency which we normally have on electric motor tests.

5. The female rotor is the driver and the male lobe is being driven by the female. Is this what you meant? The compressor is oil free. If this is what you meant then i guess you wanted me to give a reason for this choice? One reason i can think of is that the compact shape and size of the compressor make it ideal for mounting inside the chassis on most vehicles to enable low cost prop. shaft driving. This is the other difficulty we have. Because it is being mounted on the chassis of a truck or a tractor we are only allow to modify the compressor itself or the frame suspending the machine from the chassis of the truck or the prop shaft. We are having short distance and so have to coupe with large prop shaft angles between 11-20°. This we also think is causing our torque flctuations.

Can you kindly explain why this is not a better option, if this is what you meant?

As usual, i am grateful for your support.

Regards,
James

 

[mechanicaljw]

Hello Jeyaselvan;

I just realized that in trying to answer questions 1&2, i made a mistake. It should read "strain gauge on slip rings" and not "string gauge on slip rings". Sorry about that.

Thanks!
Jimmy

 

[jeyaselvan]

Hi Jimmy,

3. Since this is a 6 cyliner engine, the thrid torsional order and its harmonics are expected to be dominant apart from the torsional excitations from the airend, which are at the lobe meshing order and its harmonics. The lower half orders and first order excitation comes from the distortion in the Torque - crank diagram, which I have attached, for a 5 cyl engine for illustration. You may refer to text book by Nestorides ( I have a recent edition) / Den Hartog more information on torsionals.

4. Sorry, I meant the cancellation of the primary and secondary forces and moments. Torsionally, still I am still unaware of the source of 1X excitation. Even if you have a natural frequency at 11/12Hz, there need to be an excitation to excite the 1X. For laterals, your unbalance may be good enoguh. By the way, is the whole assembly on mounts?.any check on mounted resonant frequency measurements? Were you able to predict the 11/12Hz and 22Hz natural frequency from your torsional model?

Can you explain more than "This is due to the forced frequency of the engine opposed to compressor input frequency which we normally have on electric motor tests"

5. Yeah, I understand. this is will aid in packaging. Since my experienece with oil free pressure compressors (higher pressures), where in power transmission through timing gears are not often desired. That should be fine for supercharger kind of applications.

Regards
Jeyaselvan

 

[mechanicaljw]

Hello Jeyaselvan,

Thanks for writing and for providing me with more information. Appreciated. To the points we've raised:

4. Q. Is the whole assembly on mounts? Ans. No, the compressor is suspended from the side on the chassis of the truck. I am attaching a picture to give you an idea of what i mean. Not sure if the side-mounting in this case can be a source of the problem?

Our torsional model has not been able to predict the 11/12 Hz and 22 Hz. When we modified the drive with a solid hydraulic drive, the 22 Hz excitation frequecy in the torque vanished and the noise and torque peak did not exist. The resonance frequency became 40 Hz. The two point mass torsional model in this is predicting 60 Hz instead. And we suspect the problem to be coming from the stiffness determination of the shaft. We determined k using FEA simulation in ANSYS Workbench. Not sure about what we are doing wrong. I know the k should be dynamic and what FEA provides i think is static. So we want to determine the dynamic k and see what that would give us.

Quote: "Can you explain more than "This is due to the forced frequency of the engine opposed to compressor input frequency which we normally have on electric motor tests". You are right that we are not sure about this explanation. What i just meant is that the 11 Hz becomes excited when the driving frequencies of the engine excited it. But i am thinking the explanation you provided might be it but i have to try to find the books you recommended to do some more reading.

Our approach now is to predict the torsional frequency numerically and then tune the shaft away from that frequency. We are also seeking a torsional damper that we want to integrate with one of our prop shafts for it to be tested.

Thanks again for your input.

Regards,
Jimmy

 

[mechanicaljw]

Hello Jeyaselvan,

I forgot to add in my previous email that we do have PTO ratio from the diesel engine to the compressor and it lies between 1.2-1.5. The compressor speed is therefore higher. This i am sure is also playing a role in why we are having resonance at 11/12 Hz. My question is that given that our speed range is 1000-1800 Hz and we are having resonance at 11/12 Hz which is less than the lowest compressor speed, could it be that the 11/12 Hz is not a resonance frequency but an excitation frequency?

Thanks!
Jimmy

 

[mechanicaljw]

Hello Jeyaselvan,

I am looking at your response dated 11th October 2012 to this post, where you said you had a problem with a beating close to a torsional frequency and that your torque magnitude increased. I would be grateful if you can share more light on this and how you were able to resolve it. In the lastest measurement we have done on a MAN Truck Chassis we are experiencing beating around a torsional frequency of 10 Hz that seems to be present at all speeds. The source of this we believe is the engine. The question is we are also experiencing a pulsation that we are thinking is structural @ 11.7 Hz. Could it be that the 10 Hz is exciting this 11.7 Hz in the compressor or chassis and the interaction between these frequencies could be the source of the beating we are having?

Thanks in advance for your input.

Jimmy

 

[jeyaselvan]

Hi Jimmy..

Yeap.. that was almost close to 10yrs back..it was on a test rig with motor(thru VFD) & compressor through a disc pack bibbys coupling on either sides of the torque transducer, wherein we had that isssue. I managed to pull back a snap shot of the measured data..the units are as measured in Volts..you could see the mesaured torque beating at the speed closer to resonance.. In that case, the main concern was not beating, but the abnormal behaviour of torque transducer (strain gauge based) operating at resonance.(we even had mechanical failures of the torque transducer instrumented shaft, a couple of times, until I diagnose the problem). With stiffness increase of the weakest element in the link, we moved it outside our operating speeds range. That was the first time I was exposed to torsional resonance!

In your case, just check your beat freq is indicating your assessment. Slips under direct drive combination could end in beating . May be you could some run up / down studies to see any obvious resonances are in your frequency range of concern? If you could instrument for torsionals, that would be great.If not, you may use conventional lateral measurements, since you often expect some indicators due to torsional/lateral coupling, if you are lucky.

Regards
Jeyaselvan

 

[mechanicaljw]

Hello Jeyaselvan;
As usual, thanks for the input and the attachment. What we are experiencing is similar to the plot you sent to me. Does it mean you had a modulation frequency? In our case we are thinking that there is this 10 Hz modulation that we cannot show where it is coming from, but suspecting that it would be torsional. In the lateral measurements, we had a frequency close to it (11.7 Hz) and when that mode was suppressed (by bracing to increase stiffness) the frequency and peaks remained about the same (10 Hz noise was still being heard). Leading us to think that it cannot be a resonant frequency. We do not have instruments for torsional, unfortunately.

The other observation is that the noise coming from the compressor itself is at a much higher frequency but for each of the peaks at the higher frequencies we have a side band of 10 Hz.

I wanted you to share any thoughts you might have based on my input above.

In one of your posts you indicated that if we use conventional lateral measurements we can expect indicators due to torsional/lateral coupling. Can you elaborate more?

Thanks!
Jimmy

 

[jeyaselvan]

Hi Jimmy,

That attachment was of a beating. The beating was between 90Hz (torsional excitation frequency) and 97Hz(torsional natural frequency).
(the data attached is more illustrative)

The beating pattern could be due to the existence of a) two closer torsional excitation frequencies or torsional natural frequencies and b)if you have an excitation frequency closer to a torsional natural frequency frequency. May be if you can change the speed by 10-20%, then you can observe for changes in beat frequency. If this is so, you may have type (b) beat, and if this is not, you may have type (a) beat. Then you can investigate your system with more detail / refinement.

May in you case, since lobe meshng is dominat, suggest looking for natural frequencies close to your lobe meshing frequency (atleast two orders)..Hope this helps

Reagrding your query on torsional / lateral coupling, since in most drive systems with gears / rotors , there is every chance that torsional resonance can be captured in housing vibration measurements. The torsional dynamic amplifications often induces lateral forces, and hence be captured in housing rup up measurements. Sometime back I was lucky enough to capture a torsional resonance induced gear rattle due to gear mesh excitations around 2000Hz through normal housing vibrations (this case in axial directoin, the gears being helical), which was subsequently verified analytically through a Campell diagram for torsionals and experimentally through angular velocity measurements.

Regards
Jeyaselvan

 

[mechanicaljw]

Hello Jeyaselvan,
And thanks for the prompt response to my previous post. Appreciated. I wanted to say that we are about 90% certain of the source of our problem. We have a modulation frequency in our torque signal (10 Hz on the MAN truck and about 12/13 Hz) on the test stand (electric or diesel). In all the tests that has been carried out for different screw compressors these frequencies are always present. So we have concluded that this modulation is driving the response of our compressors. In one case, where the synchronization gears are located at the opposite side of the air end (same side as driving gears) the response is less harsh and quieter when compared to the case where the sny gears are located at the air end. Even though on the test on the deisel engine we also see a torque peak fluctuation for the case with syn gears on opposite end of air end, the noise is still lower. From a torsional model we realized that there is a large difference in the rotor deflection modes of these two configurations. The former is about 838 Hz and the latter about 297 Hz. From the mode shapes we are thinking the sny shafts seems to be problematic because in the latter we have to drive through 2 additional shafts and so we are thinking it has more compliance. The sny shafts diameters are also different. Smaller in the latter than in the former configuration, which we think might be compounding the issue.

The lower mode of the latter is making us to believe that they are being excited by the lobe passing freuqency?

The idea now is to modify the case where we have sny gears on the air end and see what that would do. So the question is whether you have any previous experience with working with these kind of configurations and whether you faced such issues?

Thanks!
Jimmy