GEAR BOX VIBRATION MONITORING

Daer All,

Can you please assist me to set FMAX,and Resolution for Gear box vibration monitoring,

Gear Mesh Frequency (GMF) = No. of teeth x Shaft Speed

= 73 x 7950

= 580350 CPM or 9672.5 HZ

How much Fmax and Reolution?

Is 100 mv/g sensor is recommended to take vibration data by off line data collector.?

There is no online accelrometer installed on it.Only proximeter probes are installed.

No. of teeth at input shaft

73

No. of teeth at output shaft

43

Rated Input speed

7950

Rated output speed

13497

Gear Ratio

1.69

Original Post
Ralph Stewart posted:

To get 1st gearmesh, you will need an Fmax of 10,000 Hz, and at least 6400 line (if possible) if your meter and 100 mV transducer can be trusted out to  that high of an Fmax.

Thanks,

Ralph

I take note that the OP uses the term "monitoring" as opposed to "diagnostics".  For routine day to day monitoring I don't feel 6400 lines of resolution is needed for an as yet non-existent accelerometer (OP states "There is no online accelrometer installed on it." 

Further, and as William Foiles has stated, I support that the need for an accelerometer that is capable of "seeing" 4X of mesh is not needed.  API 613 states that the upper limit of the accelerometer should be 10 kHz, although in this instance I would suggest and accelerometer that is capable of 2X of mesh as being ample.  All too often there is a rule of thumb saying you need to look out to 3X of mesh but the fact of the matter is 2X is both ample and practical.  An accelerometer capable of 30 kHz (3X mesh) or 40 kHz (4X mesh) would be extremely expensive and subject to a great deal of special mounting considerations.

In addition, to the OP, if you are considering a permanent mounted accelerometer, I would suggest something that is 25 mV/g as opposed to 100 mV/g.  That will reduce the potential of transducer saturation.  In this instance, don't even think of a temporary mount like a magnet.

ALI HAIDERR posted:

Daer All,

Can you please assist me to set FMAX,and Resolution for Gear box vibration monitoring,  

Gear Mesh Frequency (GMF) = No. of teeth x Shaft Speed

= 73 x 7950

= 580350 CPM or 9672.5 HZ

How much Fmax and Reolution?

Is 100 mv/g sensor is recommended to take vibration data by off line data collector.? This does not tell me he was wanting a setup for continuous monitor, but a setup for his hand held, walk around analyzer.

There is no online accelrometer installed on it.Only proximeter probes are installed. This appears to be just a statement pointing out that this is the current setup, on his machine, prox probes.

No. of teeth at input shaft

73

No. of teeth at output shaft

43

Rated Input speed

7950

Rated output speed

13497

Gear Ratio

1.69

John of PA,

I am not sure why you are quoting my reply inside your reply to the original poster's questions.

My reply was referencing the question the original poster asked about, at the beginning of the posting: "Can you please assist me to set FMAX,and Resolution for Gear box vibration monitoring,.......

How much Fmax and Reolution?

Is 100 mv/g sensor is recommended to take vibration data by off line data collector.?".

The quota about the accelerometer: "There is no online accelrometer installed on it.Only proximeter probes are installed." was listed LAST in the posting, for what reason, I do not know, unless it was to show us that this is the way it is currently being monitored with an online system, but it sounds like, IMO, he was wanting to know if he could "monitor" (not necessarily online), but periodically monitor with an  "off line data collector" and a 100mv/g accelerometer.

I am not sure what his (the original poster) definition of monitoring is, nor what yours is, but anyway what did I say that was so stupid to warrant being, what seems to be, corrected? I did not say anything about "diagnostics", just merely giving a setup for his "off line data collector" which I read as to be what his original questions were about.

Continuous Monitoring setup does not appear to be his original questions, in my opinion.

Only my opinion and I could be totally wrong.

Thanks,

Ralph

Ralph Stewart posted:
 
 ...but it sounds like, IMO, he was wanting to know if he could "monitor" (not necessarily online), but periodically monitor with an  "off line data collector" and a 100mv/g accelerometer.

 

If indeed this is the case, in summary, monitoring with an off line data collector and 100 mV/g accelerometer, you should have stated the potential problems.  For one, any accelerometer that would "see" a 1X mesh of 9672.5 Hz would have to be permanently mounted.  You just would not be able to have a temporary mounted transducer deliver reliable data at that frequency.  In addition, the need for 6400 lines of resolution just doesn't exist in this instance.  The gear generated sidebands would be 9672.5 ± n132.5 (Hz) and the pinion generated sidebands would be 9672.5 ± n224.95 (Hz).  Why would I need 6400 lines of resolution with sideband spacing of 132.5 Hz (gear) or 224.95 Hz (pinion)?  10000/6400 = 1.56 Hz.

Good question and good discussion.

I second the comments given on the limitations of the hardware:

1. Typical general purpose accelerometers have limitations on Fmax, usually 10 kHz or less. Readings above this limit are not reliable especially if you analyze the data without a history with consistent measurements. 

2. Any frequencies say higher than 5 kHz need a very good mounting method for the sensor (e.g. stud). Despite the accelerometer and analyzer capabilities, the mounting is very detrimental in such cases.

3. vibration components at very higher frequencies are usually low amplitudes and hence would require low a scale factor to be measured. If trends will be used to detect anomalies, 100 mV/g may help but it will be a bit late to sense the initial changes.

Regards- Ali M. Al-Shurafa

ALI HAIDERR posted:

Thanks for your kind input.

I just want to set up offline data collector periodically routine with 100mv/g sensor.We dont have 10 or 25 mv/g sensor.

Thanks Ralph Stewart,You always come forwad to assist me.

Regards

 

At the frequency under discussion, 9672.5 hz, it will be essential to permanently mount the accelerometer.

Ralph Stewart,Ali M. Al-Shurafa,

Data taken by 100 mv/g sensor,with Fmax 10KHZ and LOR 6400.

Dominant vibration peaks are 97594 CPM,165281CPM and 426750 CPM

Is below data is Okay for further trending?

It is train of (Motor+Voith+LP compressor+GB+HP compressor)

Velocity Spectrum

Velocity TWF

Acceleration Spectrum

Acceleration TWF

 

Attachments

Photos (4)

How do you know I didn’t  There is a great big real world out there that some live in that must get by with what they have to work with. Just because you have not blessed it in one of your committees you always condemn it. Bill I suggest you try to be helpful with the fast knowledge you have and be less arrogant. 

 

Barry posted:

  I agree with Ralph. I collected a lot of gear box data with a 100 mV/g acc on a two rail magnet (the square ones that CSI had back  then) and always had good trends and I don’t remember a bad or missed call on a gearbox. 

...and how many of those gearboxes had a mesh frequency of 9672.5 hz?  

ALI HAIDERR posted:

Ralph Stewart,Ali M. Al-Shurafa,

Data taken by 100 mv/g sensor,with Fmax 10KHZ and LOR 6400.

Dominant vibration peaks are 97594 CPM,165281CPM and 426750 CPM

Is below data is Okay for further trending? It might be, (it looks fairly good) but can you tell us where the data was actually taken? An actual drawing or an actual photo of the real machine might clarify a few things. Such as, where is the "97594 CPM,165281CPM and 426750 CPM" might be coming from? The numbers appear to low to be coming from the gearbox gearmesh in the original question posting. How many data points are there between the motor and where this data was taken, AND ALSO AFTER WHERE THESE DATA POINTS WERE TAKEN?

Could these numbers, 97594, 165281 and 426750 be from something other than a gearmesh, like maybe some type of vanepass or from??????

What, if anything, is the present concern with the High Speed Compressor gearbox? Does it have something going on that might be a problem or are you just wanting to get data from another source other than the prox probes?

Can you help us out a little more?

Ralph

It is train of (Motor+Voith+LP compressor+GB+HP compressor)

 

Your anticipated frequencies are summarized below.  Like Walt has said, the source of the components at 97594 CPM, 165281 CPM and 426750 CPM may be from something else as they are unlikely to be from the gearbox as you've presented it.

By the way, I see your comment "It is train of Motor+Voith+LP compressor+GB+HP compressor".  Are your speeds correct?  How do you get from typical motor speed to 7950 RPM?

 

Attachments

Photos (1)

Based on your "pictorial" that shows the machinery arrangement, I think you have a Voith KGS drive that provides an increase in motor speed through a two-stage gearbox.  There may or may not also be a Voith fluid drive integral to the unit.  The Voith as pictured in your pictorial shows the 1st stage as speed increasing (gear is the driver) and the 2nd stage as speed decreasing (pinion is driver).  This is just based on the relative sizes (of the gears) as shown but I wouldn't trust a simple pictorial.  The pictorial I've attached below shows both stages as speed increasing.  In any event through some means, maybe you need to contact Voith, you will have to establish the tooth combination of the two gearsets in the KGS drive to fully understand the source of these frequencies.  In addition, as others have mentioned you should attempt to get detailed compressor information.  You obviously have a lot going on and the potential for vibration to be transmitted. 

I have also attached the page from the Voith marketing materials that discusses the KGS drive. 

Voith_KGS

Attachments

Photos (1)
Files (1)

John

Motor Speed is 1500RPM

then voith speed is upto 7900 attached with LP compressor.Voith speed varies between 7500RPM to 7900RPM only during operation

Then there is GB between LP and HP compressor and out put GB speed  is 13400RPM(Approx) which is the speed of HP compressor also.

Voith,Gearbox are equiped with proximity sensors,my concern is to monitor gear or voith faults with my off line vibration meter as there is no online accelrometer is installed.I have only 100 mv/g senseor. 

 

Thanks for your input.

...and I will state again, you need permanently mounted accelerometers on the Voith unit, even perhaps two and at least one on what you are referring to as the gearbox.  You could use some form of connection panel for taking readings, but the signal characteristics, mainly the high frequencies, dictate that the transducers be permanently mounted.  The accelerometers can be either 25 mV/g or 100 mV/g although I would lean toward 25 mV/g on the gearbox that has the 9672 Hz mesh frequency to avoid transducer saturation.

ALI HAIDERR posted:

Ralph Stewart,Ali M. Al-Shurafa,

Data taken by 100 mv/g sensor,with Fmax 10KHZ and LOR 6400.

Dominant vibration peaks are 97594 CPM,165281CPM and 426750 CPM

Is below data is Okay for further trending?

Ali,

 

I assume you mean that is the data you collected with the general purpose sensor reliable to give an early warning for gear faults? If this is all you can have, simply use it. It will give you reasonable detection on low frequency faults like misalignment but it may not be accurate for higher frequency faults. So, you get something but there is a risk for missing some faults. If I were you, I'll let the owner know about the limitation and ask for his/her concurrence on the outcomes.

 

Many times we use what we have. Most of the time we can find supporting ways to increase the detectability like oil analysis, temperature measurements, frequent operator rounds etc. Do you need all of these? Maybe. I usually ask for more, if the risk of failure is more.

 

I hope my comments help a little.

 

Regards- Ali M. Al-Shurafa 

ALI HAIDERR posted:

Dear Ralph,

No .I didn't take.

Regards

Ali

Thanks for the reply.

I was sort of hoping you had so we could compare it with the 1st data taken with just a hand held transducer.

Could you send me, or post on the forum, the acceleration spectrum you recently posted, except instead of having an "autoscaled" vertical Gs scale, but have it with a manually controlled vertical scale of 1 Gs so as to see if there is anything located near the gearmesh calculated frequency of 9672.5 Hz? Just my curiosity.

Thanks and Have a Great Day,

Ralph

Add Reply

×
×
×
×