Re: Gearbox Bearing Life
Thanks Larry. None of it makes any sense but I appreciate your interest and effort. ■
Re: Gearbox Bearing Life
You asked:
1. What bearing do you typically specify?
a) I said that one standard (AGMA) specifies the bearings each need not be greater than L10=5,000 hours. I have talked to mfgrs. sales force. They could state no clear reason for this low value.
b) Since manufacturers normally well exceed this value, what is a reasonable value? I claim that setting the reducer assy. failure rate or Mean-Time-Between-Failures (MTBF) is an alternative. The MTBF = 5 x L10 (ie L10 represents a 10% failure in the given hours while a mean then would be 50% or 5 times).
c) I gave an example of a reducer MTBF = L10=10,000 x 5 = 50,000 hours. If you operate at full throttle for 5000 hours per year, then on average, you will fail a bearing within the reducer every 10 years.
d) Calculating Assy L10 life:
L10= 1/[ 1/L10(1st brg) + 1/L10(2nd brg) + .... 1/L10(nth brg)]
Therefore, selection of the L10 life is predicated on economic tradeoff of failure rate verses reducer cost to control failure rate.
A reducer assembly L10=10,000 hours is a reasonable value and will typically have bearings in the selection range I stated.
Maybe others can give a clearer argument.
2. MFGRs selection for reliability and lubrication properties is far more esoteric. I defer to the mfgrs.
3. If point one does not work, then no reason for 2 and 3. ■
Re: Gearbox Bearing Life
Okay Larry, I am starting to see where you are heading. Typically the L10 life we specify would relate to a single bearing and not the gearbox. Much of your discussion relates to the overall L10 life of the gearbox. I have not given that much thought so found it interesting. It would certainly be a consideration in seeking to meet the clients expectations on the reliability of a piece of equipment. I'm used to seeing a 95 % reliability being given for equipment. Is that realistic? Your analysis indicates probably not without being excessively conservative. You demonstrate that the overall reliability of the gearbox is going to be much less than the nominated 90 or 95 % by a factor of the number of bearings in the gearbox.
It would be interesting to investigate client expectations in regard to this matter. Realistically, how often would a client be comfortable with pulling out a gearbox and changing a bearing? Another approach would be to tie the reliability criteria to individual bearings. In my experience that is what typically happens on projects, at least over here. Normally an L10 life is stipulated and that L10 life value is determined with reference to the design life of the project. The L10 life for a gold plant with a design life of 5 years is typically less than the L10 life for an iron ore mine with a design life of 20 years. That is the starting point of this thread.
From discussions with manufactures, my understanding is that their typical approach in checking the bearing life of a selected gearbox is as follows. Suppose the specification nominates an L10 bearing life of 60 000 hours. That is a common number over here for a gearbox and gives about an 8 year life of continuous operation. The suppliers will then factor this life based on the conditions they think are likely to exist on site. The amount they factor the life is along the lines of the SKF bearing life calculation which predicts a service life Lnm = a(1).a(skf).L(10h).
My contention is that there should not be a significant variation between the factors used by manufacturers. The a(1) should be identical since it theoretically relates to the specified reliability. I'm not convinced they even consider this requirement. The a(skf) is a factor predominantly determined by the cleanliness of the oil and the ratio of the required lubricant kinematic viscosity and the actual kinematic viscosity at the bearing. I don't believe this should vary much between gearbox manufacturers.
The numbers I’ve typically seen used are those mentioned above ie. 2.5 to 5. The L10 life of the bearing is then factored by this amount. Providing the worst case bearing exceeds the requirement the selected gearbox is considered adequate for the bearing life criteria.
I'd like to pin down this factoring a bit more. In my view there is something not right if two gearbox manufacturers differ by a factor of two on the expected life of the same bearing selected for the same L10 life. What do you think? ■
Bearings And Rotation The- Stuff Of Conveyors
Originally posted by Derek Bishop
Okay Larry, I am starting to see where you are heading. Typically the L10 life we specify would relate to a single bearing and not the gearbox. Much of your discussion relates to the overall L10 life of the gearbox. I have not given that much thought so found it interesting. It would certainly be a consideration in seeking to meet the clients expectations on the reliability of a piece of equipment. I'm used to seeing a 95 % reliability being given for equipment. Is that realistic? Your analysis indicates probably not without being excessively conservative. You demonstrate that the overall reliability of the gearbox is going to be much less than the nominated 90 or 95 % by a factor of the number of bearings in the gearbox.
It would be interesting to investigate client expectations in regard to this matter. Realistically, how often would a client be comfortable with pulling out a gearbox and changing a bearing? Another approach would be to tie the reliability criteria to individual bearings. In my experience that is what typically happens on projects, at least over here. Normally an L10 life is stipulated and that L10 life value is determined with reference to the design life of the project. The L10 life for a gold plant with a design life of 5 years is typically less than the L10 life for an iron ore mine with a design life of 20 years. That is the starting point of this thread.
From discussions with manufactures, my understanding is that their typical approach in checking the bearing life of a selected gearbox is as follows. Suppose the specification nominates an L10 bearing life of 60 000 hours. That is a common number over here for a gearbox and gives about an 8 year life of continuous operation. The suppliers will then factor this life based on the conditions they think are likely to exist on site. The amount they factor the life is along the lines of the SKF bearing life calculation which predicts a service life Lnm = a(1).a(skf).L(10h).
My contention is that there should not be a significant variation between the factors used by manufacturers. The a(1) should be identical since it theoretically relates to the specified reliability. I'm not convinced they even consider this requirement. The a(skf) is a factor predominantly determined by the cleanliness of the oil and the ratio of the required lubricant kinematic viscosity and the actual kinematic viscosity at the bearing. I don't believe this should vary much between gearbox manufacturers.
The numbers I’ve typically seen used are those mentioned above ie. 2.5 to 5. The L10 life of the bearing is then factored by this amount. Providing the worst case bearing exceeds the requirement the selected gearbox is considered adequate for the bearing life criteria.
I'd like to pin down this factoring a bit more. In my view there is something not right if two gearbox manufacturers differ by a factor of two on the expected life of the same bearing selected for the same L10 life. What do you think?
99% percent of this goes to the bearing type and its design.
Ball bearings single and double row
Cup and cone tapered roller bearings
single and double row barrel roller bearings
single and double row tapered barrel roller bearings
And my favorite: Timken box car bearings
With a conveyor drive reduction gear box you are dealing with a motor running at a rated RPM delivering power to a gearbox and in turn reducing the revolutions per minute via the gear train in the gear box to a usable RPM to drive the chains gears etc. or the hydraulic drive all the while being lubbricated by a hypoid gear oil
or a synthetic lubricant such as Mobil 629-I dont even know if that oil is being synthesized anymore:^)
Ball bearings can and are used for very high speed applications such as used in motor cars, Lorries, locomotives, and engines of all types hence the use of same carrying the driving input shaft of the gear box an having it reduced by the gear train ratios in the reducer to drive the conveyor through a rubber or oil coupler.
Most gear boxes are and have splash lubricated bearings and they are essentially starved at start up every time so this plays a role in their life during actual conditons.
Not having the drive motor lined up properly at installation or reinstallation during a repair does not help much as it puts huge amounts of strain on the bearings components outer and inner race ways which in turn put pressure on the bearings themselves rdcung their carrying ability.
These bearings are making thousand of revolutions per minute depending on the application and being starved of oil at start up does not help much and they travel millions of miles linearly as they operate at a high RPM
Low rpm sealed bearings fair much better but the open bearings are less expensive to make and offer for sale.
So it is a case of getting what you pay for versus the providers costs of production and retail cost.
Perhaps you should consider rebuilding drive units in house and buying the strongest bearings possible for the application such as the barrel roller type providing the size of the bearing can be matched-do not forget that all bearings are metric and are either a press fit or heated compresssion fit on a shaft- prior to installation using a bearing heater allow for easier installation when you install them.
Some folks freeze shafts in liquid nitrogen or oxygen to allow for easier installation of boom pins on front end loaders as an example since the new bronze bushings are so tight.
Bearings installed in reducer gearbox machined castings are press fit as a rule as the casting are generally to large to heat efficiently.
Putting roller bearings in a freezer for several hours allows for easier installation in a press fit installation also.
The same goes for small shafts that need to have bearings installed in a press fit application.
A reducer gear box will have as many as eight bearings or more in it depending upon whether it is a right angle drive also.
Oil seal life also plays a huge factor in this as well-bad bearing=radial shaft loading=seal failure=gearbox failure= why does it go bad now istead of waiting for the maintenance window?
lzaharis ■
Re: Gearbox Bearing Life
There are standards that manufacturers use such as AGMA. These standards are quite lax when it comes to bearing selection. Their minimum criteria is a bearing L10 life of 5000 hours at SF=1.0.
This is not acceptable to most 24 hour per day operations. Often we see the engineering company place a 50,000-100,000 hours on the bearing. This is also not realistic. What should the goal be?
If you step back and apply an L10 life to the reducer assembly, then the manufacturer can economically select shafts and bearings to fit the purpose. To this end we specify a reducer with a minimum overall L10 rating of 10,000 hours. This means the reducer will have a 50% chance of failure in 10 years, or two in 5 years and so on.
This approach means, by example, a 2 stage reducer, which has 6 bearings, will each have a equal bearing L10 life = 60,000 hours. However, the bearing L10 hour rating will likely be higher for the small gears and less for the large gears such as: 100,000 for the smallest reduction stage, 60,000 for the intermediate stage, and 40,000 for the largest stage equal 10,000 hours for the overall assembly. Thee are a lot of assumptions in this analysis. It is an example. ■
Re: Gearbox Bearing Life
When specifying bearings for a gearbox L10 should not be used for life in hours as it is actually the life in revolutions, as per Australian standards AS 3890 and AS 2729. Should use B10 life. (also called L10h)
A specification should say that gearbox shall have minimum life (B10) of 50,000 hours based on maximum demand power.
Note that B10 is enough to say that bearing life is unadjusted.
If the adjusting factors a2 and a3 can be used, then a B10a would be stated.
The standard unadjusted factor a1 is 90% that bearing will acieve life required.
So assuming suppliers follow correct design / selections and remove interpretation it is necessary to use B10. Other wise to fully specify what is required for the bearings it is neccessary to state the bearing life required, the reliability life adjustment factor (a1) ( ie betwen 90 an 99 chance of run required hours) and loads to be used (motor power, demand power, average load). If bearing life is to be adjusted for material and lubrication factor(a23) it also also necessary to say so.
So strongly suggest that fo all standard gearbox duties, the bearing adjustment fators equal 1, and always request B10 (or L10h) ■
Gearbox Bearing Life
I understand that the life of bearings in gearboxes can become critical in gearbox selection.
1. What bearing life do you typically specify?
2. The L10 life of bearings is typically factored by the suppliers to suit the prevailing conditions. For example, for SKF bearings, the L10 life is adjusted for reliability and lubrication properties. These factors normally combine to increase the calculated bearing life. Any suggestions on typical factors to use here? I've heard of the L10 bearing life being increased by anywhere between a factor of 2.5 and 10.
3. Do you have differing criteria for different power ratings of gearboxes.
Thanks. ■