And for the 2014 International Maintenance Conference in Daytona this year, one of the headline topics is ‘Elimination of Bearing Failures’. I‘m just wondering; How can one eliminate failure of a component that is designed to fail? Yes folks I’m here to tell you; rolling element bearings are a consumable, they are designed with a specific fatigue life and one day they will fail. Now I do concede that you can manage bearing life, ensuring you can at least get the advertised fatigue life, e.g. a reasonable service life, but you will never, never eliminate that bearing failure. If you don’t believe me, I have some swamp land just outside of Miami that you might be interested in (LOL).
To start off, 
I’m going to ask you a serious question. Just how many rolling element bearings do you have in your facility? This is a question I ask all of my clients that are interested in managing bearing life. Do you know the number? More than 10, more than 1,000 more than 100,000? I bet there are more than you are willing to count and if you think about it, there is quite a financial investment associated with those bearings. And if it were my company, I’d want to manage that investment; I want the best return I could get.
I’m going to ask you a serious question. Just how many rolling element bearings do you have in your facility? This is a question I ask all of my clients that are interested in managing bearing life. Do you know the number? More than 10, more than 1,000 more than 100,000? I bet there are more than you are willing to count and if you think about it, there is quite a financial investment associated with those bearings. And if it were my company, I’d want to manage that investment; I want the best return I could get.
Before we get too caught up in how to manage bearing life, manage our investment, we need to identify those things that cause bearing failures, those things that increase the probability of bearing failure, those things that shorten the fatigue life of a rolling element bearing. Let me share my personal list of common bearing issues.
Maintenance Concepts
Typical Causes of Bearing Failure
| |
Bearing Failure Fatigue Related
|
Bearing Failure Cage Related
|
Lack of Lubrication
|
Overuse, Overfilling Lubrication
|
Contaminated Lubrication
|
Contaminated Lubrication
|
Low Viscosity Lubrication
|
High Lubrication Viscosity
|
INSTALLATION ERROR
|
INSTALLATION ERROR
|
High Operating Temperature
|
Cyclic Torsional Loads (Chain Drive,
|
Excessive Radial Load
|
Universal, Toothed Belt, etc…, Cardan, Chordal Loading))
|
Alignment
| |
Balance
| |
Bent Shaft
| |
Shaft Diameter Tolerance
| |
High Axial Load
| |
Erratic Axial Load
| |
Age
| |
Electrical Discharge
| |
These are the typical causes of bearing failure. This is my list and I organized in a descending order of probability. Do I have a set of statistics to back me up? No I don’t. But, I do have 30 years of anecdotal experience plus I have the best reference ever developed for making this (these) assumption; I have the internet. Just pull up Google and enter ‘bearing failure’; about 4,180,000 results pop up. WOW!!! Like we stated in an earlier article No Spectrum Analyzer, Oh My!!!…………., there is a tremendous knowledge base out there, if you’re not using it you are a fool. So spend a few evenings on the internet; check out the references, make your own list of common bearing issues.
So if you cannot eliminate bearing failures, what can you do? When I give training on this subject my answer(s) follow along this line;
· Manage Lubrication
· Manage Installation
· Manage Load
· Manage Shaft Condition
· Manage Indication
You need to manage the things that are related to the cause of bearing failure. If you are just managing bearing failures, e.g. just replacing bearings as they fail and not managing the things that cause bearing failure, you are losing money, you are wasting resources. And as a last thought before we jump into more detailed discussion, since you know deep down that you will never eliminate bearing failures, you also need to manage some methodology, introduce some technology, some program that can identify indications that tell you a bearing is failing. That’s where this list comes from. This is the basis of Precision Maintenance. This is the basis of running our Maintenance Departments like a business. WOW!!! There’s a concept!!!!
So let’s get started.
Managing Lubrication……….
The number one, el numero uno, the primary cause of bearing failure in today’s industrial environment is lubrication. Yes, having a sufficient quantity and sufficient quality of oil or grease in the bearing is the most important thing you can do to keep bearings rolling and your equipment operational. And keeping that oil or grease free of contaminants is the second most important thing you can do. And by far the biggest contaminant we see today is………….water.
So now the question is; Do you have a Documented Lubrication Program at your facility? And if you do, is it effective? Almost every company has some sort of developed process for managing lubrication; maybe. If you don’t have a documented lubrication program or process, stop reading this article and go develop one. Where I have completed a gap analysis on maintenance practices at DOE and DOD facilities, international and domestic manufactures, power and water treatment utilities, what I have discovered is that even though they have a documented program, a developed process, it is not followed. I personally have seen everything from color coding, equipment specific SOP’s, general specifications, training, PM’s, outsourcing to chanting and praying as far as lubrication programs or policy. However, where the program(s) fail, there is no program monitoring; there is no audit for compliance. If you do not routinely challenge a program, it eventually will be forgotten. Remember my last rant, Read the Friking Instructions, if you do not routinely challenge a program, Aliteracy sets in. Your lubrication program controls quantity and quality of the product you pump into your bearings to keep them healthy. If you are not managing that program, you are not managing your finical assets.
Contamination, specifically water contamination is by far the most ignored issue out there. And the insidious part, it generally occurs at a level where it is impractical or uneconomical to sample. Water wash down, exposure to large temperature swings, humid and moist environments, are things that need to be addresses in your bearing lubrication management process. Seals, if you have oil leaks at your seals, you have moisture leaks into your bearings; if the seals can’t keep the oil in, I can pretty much guarantee they’re not keeping contaminants out.
Here’s a tidbit I picked up many years ago, I’m just not sure where. The assumption is you can achieve 100% of the advertised bearing life if you maintain less than 100 parts per million (PPM, .01%) water in your lubricant. At 200 PPM (.02%) your bearing life is reduced by 50% and at 400 PPM (.04%), bearing life is reduced to less than 25%.
Bottom line, when you manage lubrication, when you define your process, you need to include keeping water (and other contaminants) out.
Managing Installation……….
Installation of bearings is one of my favorite topics. If we are outside of a shop and completing a field installation, bearing replacement is very difficult to complete ‘per the book’, e.g. Read the Fricking Instructions. If you can get the equipment to a bench, to a shop, do it. If you can’t, bring the shop to the equipment. Bring What??????? Yep that’s correct, set up a portable bench, bring in a pallet with your tools, establish a boundary. A tent that establishes a defined work area works best. You also need to ensure the work is PLANNED, not in the scheduling sense, but in the technical scene. New gaskets, seals, nuts and bolts, shaft sleeves, the correct tools, etc… are available, is a planned evolution. It shows someone has Read the Fricking Instructions. In your temporary shop at the job site there had better be four documents; the Work Order with detailed Instructions, the equipment O&M manual and the bearing manufacturer’s installation instructions. Oh yes, you had also better include your lubrication program instructions.
Field bearing removal is a brutal process, it is very barbaric. I’ve seen old bearings fall off, pulled off, beat off, burned off, ground off and if I use enough four letter words, talked off. This is a very messy proposition. Grease, oil, chunks of metal, old seals, process media, liquid wrench, WD40, etc… is everywhere. And believe it or not, in our haste in getting the bearing off, this is generally where the first error(s) is introduced; the shaft is damaged. More on this under the heading Manage Shaft Condition. If we contain our mess to our make shift shop, it will be easier to clean prior to shifting to an installation mode. The area needs to be sanitary prior to starting the bearing installation process. One last item on removal; we want to salvage as much of the old bearing as practical. We need to analyze the failure. We need to be able to update our list of Typical Bearing Failures.
One of the most important steps in bearing installation is acquiring the correct tools. I cringe when I see a small frame pump rebuild and no bearing heater. I want to scream every time I see a hammer come out, brass rod or not. I want to pull what’s left of my hair out when I don’t see a set of micrometers, dial indicators, a magnifying glass, etc…. And vice grips, I just want break down and beet my head against the wall. Without the proper tools, your technicians are forced to improvise, to figured out how to just Get-R-Done. Because no one bothered to Read the Fricking Instructions or management was too cost conscious (read as cheap) to buy the correct tools, you are setting yourself up for installation errors. And where there’s a set-up, there is an occurrence. I had to tap the bearing in place with a hammer because I had no bearing heater. I had no hydraulic press. Oops, I slipped and smacked the cage, put a little dent in her, no one saw, it’s O.K.; maybe. The instruction said to use a spanner on the lock nut, I didn’t have one so I used a hammer and punch. Oops, the locknut wasn’t tight enough, and released the preload. Excessive axial play………it’s O.K.; maybe. Compared to the cost of ‘doing it over again’, tools are pretty inexpensive and the skills to use them are easy to acquire. Last note here about the installation instructions, I guarantee the Fricking Instructions for bearing installation won’t be in the equipment O&M, they’ll be in the bearing manufacturer’s documentation. The O&M will contain equipment assembly instructions but when it comes to installing the bearing, reference the bearing manufacturers documentation.
So basically, Managing Installation means the job is planned, a temporary shop area is established for field installations, gaskets and seals are staged at the job site, the installation area is clean, the correct tools are available, the technicians have the skill set to use the tools and the new bearings are installed in accordance with the bearing manufacturer and the equipment manufacture’s Fricking Instructions. We get a precision installation. And last but not least, we lubricate the newly installed bearing in accordance with our documented lubrication program.
Managing Load……….
Have you ever seen this equation; L10 = (C/P)3 ? Does everyone understand the concept that L10 represents? Just as a review; L10, L50, LXX…….. is the uncertainty of bearing fatigue life. Basically L represents a bearing’s fatigue life in revolutions, minutes, hours, years, etc… and the 10 represents the statistical point where 10% of a group of like bearings will fail prior to reaching that bearing’s designed fatigue life. Same with 50; it represents the statistical point where 50% of a group of like bearings………….. Anecdotally L10 means that if I buy 100 of the same bearings and install them error free and with adequate quantity and quality of lubricant, statistically 10 bearings will fail prior to reaching their designed fatigue life. So if L10 is a representation of acceptable bearing life (we are accepting the 10% failure rate) then what do C and P represent?
C and P are load values, units of pound-feet or newtons, units of force. C is the bearing manufacturers advertised dynamic load rating, their catalog value. P is the actual load on the bearing as installed. Let’s look at P a bit closer. This value is the summation of all forces applied to the bearing. So what is the make up those applied forces. As a maintenance guy my list includes;
Maintenance Concepts
Typical Forces Applied to Bearings (P)
| |
Alignment Forces
|
Magnetic Forces
|
Balance Forces
|
Hydraulic Force
|
Gravitational Force (Weight)
|
Structural Force
|
Let’s look at our bearing life equation again; L10 = (C/P)3. For roller bearings, the calculation is L10 = (C/P)10/3.
As a maintenance guy, as the owner of a pre-designed and applied piece of equipment, I really don’t have any say in what bearing is installed so I do not have influence over the C factor, the catalog rating. However I have a tremendous influence over P, the applied force. So what happens to L10 if we reduce the total force applied to a given bearing, we lower the value of P? When I lower the value of P in our bearing life equation I am increasing, I am improving my L10. My bearings will last longer. The bearing still has a finite life, however my replacement frequency has decreased, the life cycle cost is reduced, profitability has improved, velvet robes will part, money will fall from the heavens and life will be good; lower P to increase L10. WOW!!! It really is that simple.
To Manage Load is to reduce the applied force (the P) on a bearing. How do I do this? I align my couplings; I balance my rotating elements; I ensure I have equal voltage applied across phases; I operate the equipment within its designed operating envelope, I make sure nuts and bolts are tight, I ensure resilient mounts are functional, I do everything I can to lower the value of P. I manage my equipment condition. I Manage Load.
Now all of this is well and good when we are talking fatigue life; balls and races. However, there is one component of the bearing that does not fit within this model. Can you guess what component that is? It is of course the bearing cage. Cage failures are a bit different in the fact that there is generally very little stress on the cage. Its job in Conrad-style bearing is just to hold the balls in position. Until that is, you don’t pay attention to lubrication (too much or too thick) or you don’t pay attention to ball velocity changes. And then there’s installation error; we have already covered that.
Lubrication first; Did you know that bearing shields, not seals, but shields, were added in the late 1950’s early 1960’s by Electrical Motor Manufacturers to prevent Over-Greasing motor bearings. It seems that ‘back in the day’, owners were so zealous in greasing motor bearings that the excess lubricant actually stressed the bearing cage causing the cage to yield. Balls roll out of position, jam between the races and cause early fatigue failure (a ball skid condition). So shields, not seals, were added. This prevented the grease’s soap, the thickener, from entering the bearing. Then how is the bearing lubricated? As the bearing and equipment heat up, oil is released from the soap and flows around the shields onto the balls and races. WOW!!! Oh yes, and shields are generally associated with grease lubrication, not oil lubrication. Oil lubrication has other idioms, e.g. do not overfill the sump. Back to the grease; there is still debate today as whether to consider shielded bearings a sealed, lubricated for life, bearing. I don’t buy it, I for one still add grease to shielded bearings, if there’s a zirk fitting and a purge path, it gets a shot of grease. I manage on the conservative side.
Ball Velocity Change; Roller chains, universal joints, toothed belts, are murder on bearing cages. The reason; these dr
ives incorporate a velocity change to function. In Cardan drives (universal joints) and Chordal drives (roller chain and sprockets, toothed belts) this velocity change is basically a perpetual speeding up and slowing down of the balls and cage within the bearing. If the velocity change is high enough, the inertia of the balls can exceed the yield strength of the cage, allowing the cage to deform, causing the balls to roll out of position, causing the balls jam between the races thus causing early fatigue failure (a ball skid condition). For universal joints, the fix is to have a second universal at the other end of the shaft, mounted at the same but negative angle and 90° out of phase with the other universal. On chain drives, keeping the number of sprocket teeth high enough to dampen out the velocity change is the correct answer.
ives incorporate a velocity change to function. In Cardan drives (universal joints) and Chordal drives (roller chain and sprockets, toothed belts) this velocity change is basically a perpetual speeding up and slowing down of the balls and cage within the bearing. If the velocity change is high enough, the inertia of the balls can exceed the yield strength of the cage, allowing the cage to deform, causing the balls to roll out of position, causing the balls jam between the races thus causing early fatigue failure (a ball skid condition). For universal joints, the fix is to have a second universal at the other end of the shaft, mounted at the same but negative angle and 90° out of phase with the other universal. On chain drives, keeping the number of sprocket teeth high enough to dampen out the velocity change is the correct answer.
Again, you need to Manage Bearing Load. Only this time you need to focus on cage loading from the balls themselves.
Managing Shaft Condition………….
Fact; every time I replace a bearing I degrade shaft condition. If I’m lucky enough to pull the bearing from the shaft with minimal effort, as a minimum, I’m pulling metal, damaging the bearing seating area. If I have to burn, grind, hammer or otherwise abuse the equipment to get the old bearing off, there is a very high probability that I am impacting the straightness and the metallurgy, e.g. the toughness of the shaft. Even if I just leave a small nick in the shaft, I have spot that can start crack propagation.
So what do you do? You go over the shaft with a fine tooth comb; we measure the shaft diametrically, we measure the shaft for straightness, we may check the hardness, we may want to MT or PT the shaft. Nicks and scratches are thoroughly inspected and documented. These are things you would define in a PLANNED Work Order.
Shaft dimension is by far the greatest concern when installing new bearings. Generally, if the shaft is rotating, bearings are an interference fit on the shaft, the bearing inner race I.D. is smaller than the shaft by a few thousandths of an inch, e.g. I need to expand the bearing to get it on the shaft. If the shaft diameter is too small, we get into fretting, bearing creep, etc…. if too large, the bearing will overheat due to the clearance reduction between the balls and races.
What about the housing fit on our rotating shaft piece of equipment? Well by golly, this would generally be a clearance fit, e.g. the bearing would slide into the housing with minimal effort. Too tight of a fit and the bearing could overheat due to the clearance reduction between the balls and races. Too loose of a fit and we get into fretting, spinning the race or inadequate rotor clearances.
The reference that will list the shaft diameters, the shaft tolerances will be the equipment O&M manual or some other document provided by the equipment manufacturer; sometimes a shop manual, sometimes a drawing, sometimes a national or international standard (ISO 286). Sine you are not the designer of the equipment, stick with the equipment manufacturer, the OEM for shaft tolerance data. If I am installing Mounted Bearings (pillow block bearings) I need to reference the bearing manufacturer’s instructions. If you gain nothing else from this article, please at least take this to heart;
- Manufactured Equipment Shaft Specifications (except Fans) use the Original Equipment Manufacturer’s (OEM) shaft tolerance data. Examples; pumps, motors, gearboxes, compressors, etc…
- Mounted Bearing Shaft Specifications use the Bearing Manufacturer’s shaft tolerance data. Examples; fans, conveyors, hoists, etc…
Save yourself heartache here, this is free advice, an Added Value just by reading this article. The mantra for success in Managing Shaft condition; PLANNED work, clean and inspected shaft (no cracks, no deformities, no damage), the shaft is diametrically correct and straight, the shaft material and treatment is correct, Read the Fricking Instructions for shaft condition and tolerance. Read the Fricking Instructions for bearing installation and equipment assembly.
Manage Indication………
Here we go again, talking about vibration. In a previous article we discussed High Frequency and Low Frequency vibration monitoring. So what do we use in this circumstance? If we want to measure true bearing condition we will use High Frequency Monitoring. High Frequency = Bearing Condition, High Frequency = Bearing Condition, High Frequency = Bearing Condition…………. Point taken.
The common technologies associated with High Frequency Monitoring are Shock Pulse and High Frequency Demodulation (enveloping). Both of these technologies are based on Pulse Theory, e.g. a pulse in the time domain gives us an infinite signal in the frequency domain.
And is there a use for Low Frequency Monitoring in this discussion? Yes there is, when we discuss Managing Load, because these are all of the typical things we monitor for using Low Frequency Monitoring. We use our 0 to 2000 Hz broadband (filter out) signal analysis to measure the P of our L10 C to P ratio. If things like alignment, balance, structure are managed, e.g. classical conditions we use vibration monitoring to identify, then simply verifying that we have a low overall Low Frequency vibration, tells me that I am managing my bearing loads, it’s my QC to ensure I did my job. Low Frequency = Equipment Condition (P), Low Frequency = Equipment Condition (P), Low Frequency = Equipment Condition (P)……………Point Taken.
And how often would I monitor for bearing condition? As a minimum, Monthly. And how often would I monitor equipment condition (P)? As a minimum, Semiannually. Why these periodicities? If I did my job and successfully manage my bearing load(s), my P, then there is no reason to perpetually monitor for equipment condition; e.g. weekly, daily, continuously. This practice is not cost effective; it's a wasted labor effort. However, since you are dealing with a consumable, where 1 out of 10 like components can randomly fail before their designed service life, we need to be more diligent, we need to monitor bearing condition more frequently; monthly is a good starting point and you can adjust from there.
Conclusions………………
So there you have it, Manage the things related to the cause of bearing failure, Manage the indications of bearing failure and you will get the best return on your investment. Enough said, end of discussion.
So the next time you hear of a lecture, a symposium, on-line training that spews bearing failure can be eliminated, ask yourself; Am I in the market for swampland? You can’t eliminate the inevitable, a consumable’s end of life. However, you can manage its health and welfare throughout its life and get the best return, the best longevity, the best serviceable life on your investment.
Maintenance, what a Concept!!!!!
MMJennings
