Beating / modulating noise and vibration 2

[jeyaselvan]

I have a drive train comprising of a motor (4 Pole 3 phase induction motor, 50Hz), resilient coupling and a compressor (screw compressor- 4 male lobes / 5 female lobe combination). These compressors make a strong modulating / beating noise, which gives a perceptional feel that something is wrong with the machine. I tried measuring noise as well as vibration data on the machine. Both the data indicate strong sidebands of 1.6Hz, which we hear as beating frequency. I have attached the measured vibration data for reference.

Since this machine is a direct drive machine (no gears), the rotational inertia on the driven side (compressor) is around 30% lesser compared to the regular machines. This is the only difference I could see it from the design point of view. Or is the absence of gears in this machine is by itself does not have the flywheel kind of effect, which I am not sure.

The sidebands are there in most of the response frequencies. The response frequencies are all multiples of the male lobe frequencies, as expected in a typical positive displacement compressor.

I also did a motor current signature analysis (MCSA) measurements and observed that the data is fine, with only a 30dB of 250Hz (5th harmonic) component in addition to the line frequency of 50Hz. I suspected whether I could get some side bands related to pole passing frequency, but I didn't get it.

Kindly looking for your valuable suggestions & views in this problem.

Regards & Thanks in Advance
Jeyaselvan

 

[GregLocock]

Beating is a hard problem to solve. Often you have to make the quieter of the two noises quieter to get an improvement.


Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Tmoose]

So what you hear is a tone ~ 25X4 = 100 Hz with quickly varying amplitude?

Is 1.6 Hz also the slip frequency?

 

[jeyaselvan]

Thanks Greg & Tmoose for your feedback.

I was only trying to understand the problem as to what is causing this modulation.

My concern is whether this turns out to be a reliability problem, as my customers perceive. Is there any other measurements I can do to locate the problem?

I found very few literature / reference / case studies related to amplitude modulation (except those related to gears. I do not have gears in this machine)

I was thinking whether the torsional pulsation (being positive dispalcement type) is largely getting translated to speed pulsation, since the inertia on the driven side in this machine is lesser compared to gear driven compressors.

Any thoughts / directions ?

Tmoose.
I was looking for the slip frequency & polepass frequency in the motor current spectrum, but I couldn't find it. The sidebands around line frequencye are all 60dB less.

 

[jeyaselvan]

Tmoose

For your query on slip frequency, this is 0.6333Hz. The speed is at 1462 rpm against synchronus speed of 1500rpm ( 50 Hz 4 Pole). I am not able to find where the 1.6Hz in coming from.

In the current signature, I could see only pole pass frequencies at 2.5Hz side bands at 50 Hz cenetr frequency with amplitiude of 35dB.

Thanks
Jeyaselvan

 

[GregLocock]

Classic beating (aka heterodyne) is caused by the addition of two sine waves of almost identical frequency and amplitude. So unless you have a slip joint in your machine it is hard to understand where that might come from mechanically.

On the other hand obviously there are electrical slip angles, but the numbers don't seem to add up.

yes the literature on this stuff is lacking.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Tmoose]

I'd take some nice real time (if you can get enough resolution, etc) b*st*rd sound spectra using a radio shack SLM as a microphone on the A scale to better represent what I heard without all the derating. base

http://www.silencertalk.com/images/eqloud.jpg

It's been a while, so I forget long ago which parameters to use. Probably maybe no instrument power, velocity probe on an old Balmac swept filter (everything bur the fire alarm is down in the noise) or IRD DataPak. I'd have to play with with a signal generator for a while to get the set up even close. (Pure tone at 100 Hz loud enough to be 80 dBA on the meter, then play with the sensitivity so the spectrum overall was 80, etc, etc for a few higher frequencies. Sometimes just looking at it all linear made it clear what was really the bad guy.

 

[jeyaselvan]

Thanks Greg & Tmoose for your feedback

I suspect the problem is a kind of modulation (amplitude / speed I am not sure), since I see the sidebands of 1.6Hz in most of the system response frequencies (lobe meshing) . I understand beat as Greg said as sum of two closer sinusiods. The time data for beating and modulation looks the very same, I think.

Greg, the only joint in the system is the resilient jaw coupling which couples the motor and the compressor.

I am unable to find where the 1.6Hz sideband modulation is coming from. Could this be the load modulation itself (load being positive displacement, but the pulsation frequency quite high 4 times the motor speed).

Any possibilities with the lower load to motor inertia ratio?

Regards
Jeyaselvan

 

[Strong]

You could calculate the compressor bearing fault frequencies. There are two shaft speeds and possibly different bearings, so that there could be a 1.6 Hz difference. Does the compressor have timing gears? If yes, then calculate all of the gear fault frequencies including repeat (common factor) frequency. Does the compressor have a load-modulating valve? If yes, then perhaps there is a control issue.

Walt

 

[GregLocock]

Are there two shaft speeds?

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[MikeyP]

A customer recently saw the same thing when testing a piece of rotary equipment (exactly the same sideband frequency). They swapped the driving motor for a different one and the problem went away.

Now correlation doesn't equal causation but they seemed pretty convinced it was a motor speed variation. I would have thought that the runout tolerances on a screw compressor (especially a contacting one) would be pretty tight and unlikely to cause serious out-of-balance so the motor seems the most likely to me.

M

--
Dr Michael F Platten

 

[jeyaselvan]

Thanks Walt, Greg & Dr.Michael for your response.

Walt : This compressor does not have timing gears & the modulating valve. This is an oil injected compressor, in which the male rotor (4lobes)drives the female rotor (5 flutes). I will check for bearing frequencies.

Greg : Yes, the motor drives the male rotor, which in turn drives the female rotor. The female rotor runs at 4/5 times the speed of the male rotor / motor.

I have some new information from my latest test. When I couple the compressor to a slightly larger motor (larger GD^2, 30% higher), the modulation frequency gets dropped from 1.6Hz to 1.0Hz and audible irritation of modulating noise has also come down(not fully, but tolerable). I am not able to explain this.

Since this compressor is a direct drive machine without gears as stated in my first post, I was suspecting whether the inertia ratios on the drive and driven side has anything to do with this.

Regards
Jeyaselvan

 

[electricpete]

I really don't have a clue what is causing your oscillation or whether it represents any problem. Nevertheless some comments...

We have an interesting datapoint that the frequency of vibration went down upon changing the motor. We note there is a change in inertia, but we have to ask whether anything else changed.

Are you positive this is not pole pass frequency? How was running speed established (how good was the frequency resolution). I'm not sure why pole pass frequency modulation would show up on 2 different motors, but if it did the frequencies could be different even for same load level (since nameplate speed varies slightly).

In your original post you started asking about torsional oscillation. Certainly decrease in frequency with increase of inertia would be consistent with torsional resonance. And we know positive displacement type machines can excite torsional resonance.
BUT:
1 – the torsional excitation would typically be a much higher frequency.... what possible torsional excitation exists at 1.6hz on this machine?
2 – I would also expect to see sidebands around 1x in current signature.

If you look at this thread you will see example of a belt-driven recip pump that I believe experiences torsional oscillation:
thread237-249262
When current and voltage are plotted together, there was evidence of change in power factor angle over time varying at the frequency that would be torsional oscillation frequency. In fact this particular torsional oscillation resulted in power reversals (points in the waveform where power factor dropped below zero and power was delievered from motor back to power system). Later at a LEMUG meeting I saw a very similar case study.

Some more comments about torsional resonance:
If we had 2 simple rigid masses (like motor and compressor) connected by flexible element (coupling), then the radian torsional resonant frequency is:
w = sqrt(Kcoupling/Jeq)
where Jeq = Jmotor*Jcomp/(Jmotor+Jcomp)
You might check the change in frequency against this formula for kicks.

BUT unfortunately, the dynamics of an induction motor are much more complicated than modelling it as a simple mass, even if we include torque associated with "torque speed curve" which suggests simple damping. The problem is that the torque speed curve is a steady state concept and does not apply to dynamic analysis. Krause lays out a very detailed model for induction motor dynamic analysis that is widely used. It results in 5x5 linearized state space matrix that gives 5 eigenvalues (2 complex pairs and a real value).

Attached are Krause's summary of predicted eigenvalues for 4 "typical" motors of speed 1800rpm. Horsepower ratings range from 3 hp to 2250 hp. There is assumed connected a load inertia equal to the motor inertia. The imaginary part of the complex eigenvalues would be the radian oscillation frequency... of course have to divide by 2*pi to get hz. The lowest of any of these motors in any of these conditions is in the neighborhood of 18 rad./sec (around 3hz) which is almost double or triple of your 1.6 and 1.2 hz. All of the others are quite a bit higher with next highest being around 40 rad/sec or 12hz.

Now it could be that your motor has higher inertia or significantly different parameters, but I sort of doubt it (for reasons discussed above torsional resonance seems not to fit). But if you care to post complete nameplate data and inertias I would be glad to compute the eigenvalues for you using Krause's method.


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

 

[electricpete]

Correction in bold:
2 – I would also expect to see sidebands around line frequency in current signature.

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

 

[electricpete]

Do you know the inertia of the driven machine?

Without knowing the torsional spring constant of coupling, we could still apply rigid-body model:
w = sqrt(Kcoupling/Jeq)
where Jeq = (Jmotor*Jcomp)/(Jmotor+Jcomp)

Let J1 and w1 correspond to motor 1 and J2 and w2 correspond to motor 2. Then

w1/w2 = Jeq2/Jeq1 = [J2/J1]*[(J1+Jcomp)/(J2+Jcomp)]

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

 

[jeyaselvan]

Thanks electripete for your feedback.

Regarding polepass frequency, I estimated that to be at 2.5Hz and was able to see that in the current signature with magnitude less than 60dB when compared to the 50Hz (line frequency). Typically we would expect to see that in case of motor rotor bar (and this is a new motor) related issues, I understand from your earlier posts in Engtips & maintenance forums. What I get in vibration (and noise) signature is a pretty clear 1.6Hz sidebands(measurements with 0.16Hz resolution). Meanwhile, I will collect information for Krause model for the detailed induction motor dynamic analysis and thanks again for offering me to carry out the analysis.
Make : Siemens
Power:15kW,Voltage:400+/-10%, Current:28A,Freq:50Hz
Eff:92%, Rated RPM:1478,pf:0.83 ,Class F
Brg - DE: 6310C3, Brg - NDE:6309C3

I have estimated the torsional natural frequency (from a 4DOF model) of the system as 65.3 Hz, which has good safety margin from the excitations. With the larger size motor, the couping torsional frequency is at 62.4Hz. If I need to change the torsional characteristics of the system significantly, I have to change the coupling stiffness. I am working on this as well.

My intention of changing the motor (in turn inertia) was to see whether this has a flywheel effect, since this compressor (direct drive) does not have gears compared to the conventional ones.

I do suspect misalignment as a concern, since the motor is coupled through an adapter ring (with spigots on either sides) to the compressor. This component has not undergone any dynamic designs and also has some history of previous manufacturing issues with even 500micron concentricity errors. The drive train is assumed to be self aligning (assumed?)with the housing on motor & compressor coupled through adapter.

I have a paper (attached),which cities experimental results of modulations arising out of misalignments. The paper says for his range of misalignments, the modulation frequency varies from 0.5Hz to 3Hz. This is new information to me, since I was of the view that misalignments induces 2X or higher excitation orders as well as phase vraitions depending on nature of mislaignments. I am not aware of any mathematical models predicting modulations arising out of misalignments. Kindly for your suggestions.

Regards
Jeyaselvan

 

[jeyaselvan]

paper on misalignments (experiment based)

 

[electricpete]

I am not aware of any mathematical models predicting modulations arising out of misalignments. Kindly for your suggestions.

Just a thought - in my mind there would have to be something altering the alignment state at the modulating frequency (let’s say 2hz). For example with a very soft mount (isolating type with springs or rubber) maybe the casing is bouncing at that 2hz at one extreme of the bounce the stress alters the alignment state. We have a set of large fans on spring isolators where bump test gives around 3hz and that same 3hz frequency also shows up as sidebands around the running speed vibration.

That is good motor data. You don’t happen to have 50% load data? (current, power factor, efficiency... any 2 of these 3)


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

 

[electricpete]

For the Krause eigenvalues:
The 50% data would be nice (to get better estimate) but not necessary.
Motor inertia would be nice (but not necessary...I have some thumbrules to estimate it).
What is necessary would be some idea of the inertia of the compressor or means to estimate it.

Also the Krause model focuses on electromagnetic side without much detail on mechanical side... therefore assumes rigid connection of the motor inertia and load inertia. I'll show you what I come up with (once compressor inertia is known) and you can decide if it is worth to try to modify the model for more detail on the mechanical side.

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

 

[electricpete]

We have a set of large fans on spring isolators where bump test gives around 3hz and that same 3hz frequency also shows up as sidebands around the running speed vibration.

Attached is more details on this. The sidebands are approximately 3.75hz = 225cpm as shown on slide 2. I don't know the cause, but the bump test of the machine while shut down shows the exact same frequency (slide 1). I think it is somehow interelated. In this case the fan is actually overhung off of the motor shaft via a rigid coupling (no bearings inside that fan casing).

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