Hi guys,

I'm posting the circular plots from the second stage bearings of 4 similar Air Compressors.. The reason for this post is that these circular plots interested me in shape and the disimilarity with one another.I think there should be some reason between each plot and i'm really interseted if someone can give any comment on this...

Circular_Plots.doc (60 Kb, 128 downloads) Circular Plots

I've seen some pretty interesting shapes when you look at them on the cirle plot. The top one looks like it wants to be a 5X but I guess the last two events petered out. Where's the 5X in a compressor? The bottom one looks like the standard 4X I see on a lot of compressors.

The labels are confusing to me. Can you clarify what the plot represents?

Derived from housing accleration?

Is this polar plot of x acceleration and y accleration?
Or is this polar plot of x acceleratio autocorrelation wavevform and y acceleration autocorrelation waveform?

It looks like you have a 4X turning speed periodic event in each compressor, the plots are of the autocorrelated waveform so all non periodic energy has been removed. You are just seeing the periodic energy. The first plot looks like the 4X event is slightly truncated every 4th time and the last plot looks like there may be some 8X turning speed periodicity also. What are the amplitudes in the regular waveform? When you look at autocorrelated data it will just have a +1 to -1 amplitude value which just indicated how much of the waveform energy is being generated by the frequency represented by the period. I assume you have turning speed set right since the circular plot is based on one complete revolution per circle generated. If turning speed is not set correctly it is worthless.

Modern Art or useless information?

Walt

Walt:
Do you want them framed for your office?

I've heard circular plots were designed mainly for rolls, but IMHO another application they could be efficiently used for is visualization of ramdomness severity in the TWF ( as oppose to autocorrelation factor).

Of course, as Jim has mentioned, setting up correct 1x is crucial. Since there is always some error involved in this, I wonder how it will affect the shape of a circular plot (will have to play with it).

Circular plots are normally used to find roll profiles when some covering defects may be suspected (in cases of out of circularity-oval shaped, barring on the surface and the like).
I don’t see what a circular plot can provide in the case of screw compressors (but I think it may be just one compressor with a combination of 4/6 lobes and a second one with 5/6). The 4/6 combination is normally used in air compressors, but I have measured ammonia screw compressors where the combination is 5/6.

This is a CSI Plotdata feature that not many want to use. But it has a lot of good applications. The circular shape is not necessary. If you can just see the waveform on a time scale can see the same things.

If you take the waveform and insert turning speed phase reference information and then make the circular plot, one turn equal to 360 degreees, this is what you get.

I do not find many turns useful. But I use the time average waveform and will then see how similar each impacts from blades or teeth are. I will easily detect for instance erosion damage from cavitation in a hydro machine, or a destroyed (pitted or partly cracked) tooth on a gear wheel and so on.

If the waveform´s signal is not related to the choosen speed, the curve will be just a noisy blurr. If it is perfectly repeatable, there will be a sharper picture, but it will be at best of the time averaging is first made over several hundred revs and just one or two turns are shown.

I forgot the case using it to verify the proper spacing as well as pitch of blading in Kaplan turbines. Fine tuning the blades to be eaxactly the same around can save a few percent of efficiency and that can easily finance the whole measuring system in one go.

Thanks for the informations guys...

Pete,
These are circular plots of radial acceleration autocorrelation waveform from the second stage bearing housings of 4 similar compressors!!!

These compressors are of same speed and was set as 43263rpm at the second stage. So 1x is correct, i believe.. And one more thing is that we have a problem in the Compressor #1 where the compressor is giving a vibration alarm during sutdown but was running smooth during the normal operation. i couldn't find any difference in the spectrums when i compared. But look at the circular plots!! The first one is fairly different from the others,ryt?

In certain machines i found the plot going away from the circle when dealing with a balancing problem. Also the plot shows some difference when it was misalignment. Maybe in future circular plots will become a useful tool in regular analysis...

HI Jenish,
I use circular plot for single reduction gearing. It helps me find discrepancies synchronous with a specific gear. Speed has to be set very precisely but the results are often very impressive. It is a visual mean of finding repeatable pattern in the waveform. I do exercise some care when other source of synchronous signal may be present. I send you four graph of the same point with different Fmax and resolution. I usually try to get between a waveform that cover five to 8 turn of the analyse part. In the case of your compressor, seem to be four lobes; it looks like there is a small flaw on one of the lobe or maybe an uneven air discharge from one lobe? I would use that info carefully, see if it is repetitive and what about the other compressors? Are they showing the same pattern? It is a good idea to use it on a compressor; I will see what is coming out of my compressors.

Best regard, Marcel

cicular_plot.xls (1,061 Kb, 55 downloads)

Hi Marcel,
Itz nice!! Thanks for your circular plots.. I also know someone who is using these circular plots for gear analysis. He is using it even in complicated gear trains...

About my compressor, itz a screw type and that is the only thing i know about it. Uneven air discharge maybe a problem.ryt? I have to check the data from other compressors..

quote:

I will easily detect for instance erosion damage from cavitation in a hydro machine, or a destroyed (pitted or partly cracked) tooth on a gear wheel and so on.

Hi Arne,

I would like to know how useful these plots are in detecting cavitation.. For me,i'm still depending primarily on the sound from the pumps for cavitation problems.

Guys,
I'm attaching the circular plot of an impeller having huge dust deposition and also from a fairly new impeller replaced. The dust deposition in that impeller causes(along with loose structure) the equipment to vibration at a level of 72mm/s which came down to 9mm/s with the new impeller.

This message has been edited. Last edited by: Jenish,

Unbalance.doc (37 Kb, 25 downloads) Imbalance in Polar Plot

A circular waveform sample which was taken from a Steam Turbine bearing with using accelerometer. Looks like one stage has a problem because of Turbine Blades.

Steam_Turbine_circular_waveform.doc (39 Kb, 31 downloads)