Gear Box

Temperature alone will not give you sufficient warning of bearing failure.

We have a couple of fans that have RTDs on the bearings for early warning. We also have a team of vibration analysts who go around the factory checking all our rotating equipment. Due to the number of equipment, they can take up to a week before they get back to a particular motor.

Most times, they detect a slight rise in vibration before the temperature rises. Sometimes, the temperature rises before they can get around to checking it again. In those cases, the bearing temperature alarm gives us sufficient warning before the bearing seizes up on us.

If, however, you can install an on-line vibration sensor on the bearing, that would provide you with better warning than an RTD or any other temperature sensor for that matter. You'd need a rate-of-rise alarm in addition to a threshold alarm though.

regards,

Vulcan

The point I wanted to make is in the condition mentioned, will temperature give at all a sign of bearing failure?

As far as my knowledge goes the temperature rise to be perceptible, the bearing is already on the point of failure.

The temperature rise in certain other cases are used more for lube failure than the bearing failure per-se

Since this case is one of sump oil lubricated gear box, this case don't arise.

Then why the RTDs ?

Any way I will be in contact with my service depts and customers to find whether in case of bearing failures any signal was generated by the RTDs ?

As far as I know, in the rare case of bearing failure, they have more relied upon the machine vibrations and things like that.

But does anybody have an idea (technically as I feel and mentioned, these are more of an ornament)

Hello Friend....

To clarify once again.... you mentioned that the Oil is high viscosity 300Cst. Can you please once again check the same.... because i doubt about this high viscosity of oil being used for Lubrication purpose in the gear boxes.( May i know the temperature at which u mentioned the viscosity to be 300cst. ) So for 300Cst ( @ 50Deg C ) of Oil to provide the lubciation properties, temperature of the oil has to be in the range of 120+ Deg C so that your viscosity comes down to the range of 15 Cst for lubrication purpose. So a RTD above 150 Dec wont be a good solution. Secondly is there only one measuring point for the temperature... There should have been Temp measuring point at the DE & NDE bearing of the gear boxes. However RTD reading should mainly be used just for comparision between two or three senros installed in the same gear box for same application....

Can you please once again look in to your data to let me understand your scenario in a better way .

Thanks & Best regards...

It is a very heavy duty and low speed gear vox. To maintain the lube film on the contact, the high viscosity oil is used. And yes it is ISO VG 320 Cst (is 320 at 40 deg C)

Th normal operating temperature of the equipment (in actual condition) may be between 40 deg C to 55 Deg C)

There are RTDs on other sides too , there is no controversy about High speed side, the question is only the LS side

about other posts -

The FFT is being done for each gear box at our test bench (while we do the trial work shop test of the gear box)

The problem is that the gear box works in a condition that putting a FFT analyser for each gear box - I am not sure will be a good idea.

The second choice could have been the customers to ask for carrying out the FFT periodically themselves as we thought, but again this aspect has been debated by them. And with a very large base, it may not be possible for us to depute our personnel and do it.

The point of siscussion is simple- Is there any use of the LS side RTDs ?

As far as I know, the RTD signal will come too late (it is bearing failure stage 3 or later)

As a part proper engineering, i want to cut off all the useless appendage, so my first target is this (will be followed by others )

So what is the answer ? does an RTD picking the temperature of TRB or CRB running at less that 100 RPM , fully immersed in oil of a few Kilo litre - give any useful information ?

sb,

I agree with you that the RTD's are not of much use for diagnosing bearing condition or to give early warning of failure in the system that you describe. However, they can be very useful as a warning of actual or imminent (final stage) bearing failure allowing tripping of the machine before there is any consequential damage to shafts or gears. For that reason alone I would prefer to have them installed.

My understanding is that this must be very low speed with that oil viscosity and submerged bearing. For this I would look at something like Shock Pulse Measurement (SPM) with the pickups (transducers) permanently installed and a hand held instrument. Good comparative measurements can give quite reliable indications of bearing condition. Vibration (FFT) analysis on slow rolling anti friction bearings deep inside gearboxes requires a lot of skill and is probably not as cost effective as something like SPM.

It is a problem of thermal balance in a transient heating.

If the mass (oil) is important, heat flux is low (low friction bearings) and cooling is important (special circuit) = the temperature rate and magnitude increase are low so that in such a case detection of failures via temperature is not a good approach. Detection must be fast and at an early moment previous to final failure, this can be obtained by an FFT of the bearing noise. FAG, SKF and other suppliers have such equipment. Temperature is a good criterion for journal bearings and with not too big oil reserves. It is also a problem of the time constant of the "reaction" to an incipient failure.

I have no expertise in transmissions, but here is an electrical engineering suggestion for an alternate solution to detecting early failure of the bearings.

This suggestion assumes some computer "smarts" available to read sensor outputs.

You have stated that a simple vibration detection won't work because the overall level of ambient or normal vibration will overpower that small amount of extra vibration you get when the bearings first being to fail.

The solution to that is to convert the time domain (electrical potential vs. time) output of the vibration sensor into the frequency domain (electrical potential vs. frequency) using a Fast Fourier transform (FFT) technique - that's where the computer smarts are necessary.

If the frequency domain profile of failing bearings is significantly different than the vibration profile of the entire smooth running transmission, then you should indeed be able to flag a problem, even through the overall vibration level of the transmission is much higher than that of the failing bearings. This is because in some critical frequency range, bearing noise will pop up at levels higher than the background or ambient noise.

This ability to tune in on a subtle problem area and tune out extraneous noise is no different than being at a concert and listening for a single flute or piccolo while the rest of the orchestra is going full bore. You can do it, and so can your sensor connected to the right hardware/software.

By the way, FFT is canned software. No need to reinvent the wheel. Your effort will be in identifying the signals that represent incipient bearing failure, and writing the code to recognize it when it happens, and avoid false alarms.

You are totally right he has not to reinvent the wheel since as mentioned in the comment before yours the solution exists already and as you may notice based on an FFT analysis.

What is to be done is ONLY to get in touch with the potential suppliers who are the bearing important manufacturers SKF or FAG or others since I mentioned only European sources but I am sure also US are available.

You are on the right track with your suggestion. I had years ago already a very impressive demo of the capabilities of this kind of analysis. The change in the gap of a valve command on an engine was detected with the possibility to know which valve it was.

I misread the forum thread and thought I was the first responder. I didn't see the other posts until later.

The was a company -- I forget the name -- near Cleveland, Ohio which manufactured a bearing condition indicator, initially used in critical pumps in nuclear power stations. The bearing race was fitted with piezo-electric transducers for a very precise sonar which could measure the deformation of the race as each ball or roller passed and could also measure wear. It seems it could detect any change in the bearing long before failure. Since the speed of sound varies with temperature, it could also be a very accurate monitor of bearing temperature. Normally the temperature measurement was used to correct for thermal effects automatically without an actual read out. The electronics could be carried around and plugged in for a periodic measurement, so the cost per bearing was low if several bearings were involved.

About twenty years ago, I used the components to measure the glass removed during polishing in real time; the glass did not have to be moved or cleaned to make the measurement. The transducers were epoxied to the back of the mirror blank. The objective was to speed the manufacturing of large optics (think "Star Wars"), but we didn't get a contract to do that.

This technology (a bit advanced) we do in our test bench, quite a few companies are there - these are able to bick up the bearing conditions through vibration signatures, and provide the residual life (which is useful, though these are new equipments- to signify misassembly, damages etc) and identify the faults. But all these are on new equipments and at our works. (we do not obviously supply these to customers since they won't pay - and with current life of the bearings approx 10 years or more, it becomes difficult to convince them to pay)

Added to that, it will not be our product, so we will be de-facto marketing man for them

The RTD is not there to indicate the bearing failure but to indicate a lubrication failure. Any oil will have to be maintained at certain temperatures to be an effective lubricant. To maintain the temperature we may have to either cool (with a water or air cooler) or heat (with a heater). In both the cases it is vital to know the oil temperature either to switch on/off cooler or heater as the case may be. Or simply to know the failuer of the cooling or heating system. Hence, even if the RTD may not be a good indicator of bearing failure it definetely is a good indicator of failure of these systems. It will be foolish to get rid of RTD's from the gear box.

as mentioned in the first post itself.

a) the bearings are totally immersed in a few Kiloliters of oil

b) The oil is under agitation - by the gears

c) The gear box temperature and also the high-speed bearing temperatures as well as sump oil temperature are monitored.