Combining the Plant Wellness Methodology and Precision Maintenance into Your Company Asset Management Strategy and Practices

People often ask how do you get the equipment reliability performance achievable with precision maintenance results for the least effort. The best answer is to combine precision maintenance with the Plant Wellness Way asset management and maintenance management methodology.

The FAQ reply is long, since the Question is long. Within it are some of the most important answers for lasting high machinery reliability that you will ever read in your career.

 


 

Hello Mike,

One of my key responsibilities is to develop corporate equipment design, purchase, and installation standards. We are transitioning to a more proactive maintenance culture and are using life cycle cost methodology for new equipment purchases, refurbishments, and repairs.

While we recognize this will increase upfront costs, we also recognize this investment will pay for itself order of magnitudes over the useful life of the assets. I come from a reliability culture and came to this organization to help facilitate the transition. So you do not have to convince me of the merits of this approach, I’ve seen it work.

I am working on individual rotating asset standards in order of priority. Started with motors, then to centrifugal pumps because it is the most numerous driven asset, etc.

But I have recognized I need for a general precision design and assembly rotating asset standard as a catch all. (to cover assets missed and as a stopgap until I get more of the rotating asset classes finished.)

I am using the following ISO recommended overall vibration acceptance criteria for new equipment not covered under specific rotating asset classes. For the most part specific rotating classes have similar acceptance criteria.

Machine vibration limits to USA standards

I have a pretty good idea on what range of precision fits and base flatness I will need to obtain this but would like your thoughts. I also have some interesting questions concerning balancing standards.

We are transitioning from a culture where we rarely asked for these types of requirements from our supplier base. Some will be capable of performing to this level and some will not. I must have good justification for my requirements when I receive pushback from engineering and purchasing (which I will when the supplier base complains and people they are used to working with are no longer being considered.)

So your recommendation on the following categories would be greatly appreciated. Keep my overall vibration standards in mind as you answer these. They are not as stringent as some industries require.

Base flatness: I believe for general equipment I can get away with 0.005”/ft. I could specify a PIP style base or one machined to API standards (0.002”/ft) but I think this is overkill for my industry and the overall levels of vibration we are currently trying to achieve. Your thoughts?

Fits and tolerances: For bearing to shaft journal and bearing to housing bore fits, I understand the importance of precision tolerances. (whether it is the tight fit or loose fit side) for size, concentricity and cylindricity. I think I can specify IT6 and still get away with my overall vibration level. Do you think I have to go to IT5?

Balancing: Fans are already specifically covered by G2.5 Centrifuge rotors will be balanced to G1.0. The rotors mass on these two types of equipment are large in relation to the overall machine. They stay together as an assembly so in the example of the fan which is interference taper fit to the shaft it can be balanced as an assembly on the arbor and swapped out as a unit. If the shaft ever becomes damaged, we would send it to the fan manufacturer to remove, refurbish, or replace. We are capable of field balancing so the whole activity becomes a swap out.

Smaller rotors on general equipment of less significant mass where field disassembly routinely occurs are not as easy an answer. Shafts and hubs are an example. Most are single keyed and set screwed in my world. Their mass is not significant in the overall picture. You can specify an AMGA Class 9 hub and get a good overall balance as machined because of the concentricity tolerances. But if I understand manufacturing and disassembly processes correctly, as they happen in real life, the amount of allowable clearance for the single key and slip fit hub would negate anything I might require for balancing after the fact as we routinely assemble and disassemble these components in the field during replacement. The chance of offsetting the center of mass when we tighten the set screws is virtually assured. The only benefit I see for balancing is where drop in units such as the fan wheel assembly are replaced in the field.

I think I need to say something on the subject but am at a loss to how to word it. Maybe spec-out coupling to AMGA standards buy precision balanced shafts and other specific verbiage? This might be a tough one to put into exact words as it really depends on how the vibration is dampened, the machine is supported, and the size of the rotor. Know that for concrete/epoxy grout sub-base supported machinery, we will spec a 5X mass of the flooded equipment and use best practice grout installation. So in most cases we will have a good foundation to work from. And where equipment is attached to structural steel or the roof, we will design and purchase sturdy steel bases. (Wide flange beams, square tubing and gusseting, not channel) Your thoughts on how to word this?

Overhung Load: Cantilever loads, like drive belts, is another one of those areas where left unspecified you rarely get in trouble. But it is that 1 out of 100 example where the OEM in a competitive environment does not use a conservative enough safety factor for design, coupled with someone in maintenance using their finger as the calibrated tension gauge that gets us in trouble.

I have seen statements in technical articles concerning overhung load which read something like this.

“Calculate the actual overhung load at worst case design conditions and make sure it is within the equipment manufacturer’s allowable load.” OK great, so vendor XYZ in a competitive environment comes along and his equipment is marginally designed but he does not want to go to the next biggest shaft, widget, etc. I.E. it meets the literal wording of the statement but not the intent of conservative design. Then along comes a ‘know it all’ mechanic who doesn’t believe in tension gauges for tightening belts. He knows Operations will not shut the equipment down after belt break-in no matter the quality of his work, so he does his own belt tightening strategy with finger pressure instead of measuring deflection as required by the OEM.

I guess what I am asking if it was you, what safety factor would you put in place to not over design equipment but reduce failures in the real world?

Here is an analogy from our fan standard concerning stress associated with centrifugal force.

‘Stress due to centrifugal force shall not exceed 75% of the yield strength of the material on any portion of the fan rotor at maximum design operating temperature.’

So our overhung load statement might read:

Manufacturer shall calculate overhung load under worst case design operating conditions. Calculated overhung load shall not exceed XX% of equipment manufacturer’s allowable load. What should ‘XX’ be?

Is this requirement a good addition to the standard, or is it not worth considering?

I run into too many people who know individual reliability rules of thumbs but do not know their significance in overall equipment reliability. This is especially frustrating in the area of life cycle cost analysis. I have put together real life case studies where we have had successes, but I find them difficult to complete as maintenance data is hard to come by. (Especially from equipment suppliers)

Any thoughts on any of the above topics would be greatly appreciated.


 

Dear Reader, to provide context to my comments, the original Question is repeated below with answers and suggestions placed at the applicable points in the text.

 

Dear Friend,

Please read my imbedded comments (coloured blue) throughout your email below.

The guiding philosophy I espouse in the Plant and Equipment Wellness (PEW) methodology for getting high reliability is to act on changes in machine health rather than reacting to machine condition acceptance limits. Machine condition acceptance limits are set at the worst value that the equipment is allowed to exhibit. By the time a condition limit is reached the affected parts in a machine are already on the path to breakdown. Once limits are breached all that is left for you to do is stop the machine and rectify the problem. If instead you observed changes in condition health you recognise problems very early in the degradation cycle. The change in health tells you that something is affecting the equipment. You can then act in time to address the issue well before it becomes a repair and/or you can manage the degradation more carefully.

Monitoring machine health changes is much like using blood pressure to observe the human circulatory system condition. High blood pressure values are a worry since we know that they indicate greater likelihood of early death from heart failure. If a doctor only reported blood pressure readings once the value was above a too high limit, you would already be in too great-a-danger of early death by the time they told you that you had high blood pressure. But doctors do not do that. They observe blood pressure readings looking to see if circulatory system health has changed away from normal values. They then give you early warning of changes in your health because they know that a bad change is a precursor to future problems. This is also the best philosophy to use for monitoring machinery condition.

In some of my imbedded comments I present specifications using the Accuracy Controlled Enterprise (ACE) 3T format that indicates Good / Better / Best performance parameters (you could call the scale of results Bronze / Silver / Gold if you prefer that analogy). In an ACE the actual result must, as a minimum, be within the ‘good’ tolerance. When a result is at ‘best’ level you know that it is a far better performance than ‘good’. The scale gives your people and management a gauge to understand the impact of quality on reliability. The diagrams below show how the 3Ts produce work quality that is in statistical control.

The 3Ts provide a way to get work quality into statistical control
The 3Ts set Target, Tolerance, and Proof Test work quality specification

Usually the ‘good’ value is what the Original Equipment Manufacturer (OEM) specifies. The ‘best’ is set at world best practice. (I find that most OEM recommended values are not good enough to create highly reliable machines, instead they make us do maintenance on their machine.) The ‘better’ value is a stretch target between ‘good’ and ‘best’. Often ‘better’ is the half way point, but it does not need to be.

The aim of having the Good/Better/Best scale is to provide a new solution to fix most of industry’s reliability problems. I believe having only a single point accept/reject standard, as in an Inspection and Test Plan, leads to the creation of poor quality and its eventual acceptance.

A single point pass/fail value indicates the worst acceptable quality result. But you actually want the quality of performance and results that produce high machinery reliability. Using only a pass/fail point test actually causes poor performance because all that is necessary is to ‘just get pass’. The test results will bunch-up just on the pass side of the acceptance point. By setting only a worst case acceptance criteria you create a wasteful Taguchi Loss Function situation.

Instead, an ACE 3T scale lets people see the improving quality values that actually deliver higher reliability—you create an optimal Taguchi Loss Function with ‘best’ positioning the optimal zone. ACE 3T quality standards leaves no doubt in peoples’ minds as to what quality really delivers world class results. I am not saying that people need to reach world class standards, but they do need to know the difference between a ‘good enough’ result and a world class outcome if you want to challenge them to perform better.

I noticed that you did not mention lubricant chemistry and solids contamination standards for your machinery—have they already been decided? My 3T Good/Better/Best scale for oils using the ISO 4406 count is 14,_,_/ 13,_,_/12,_,_. These values are the worst solids contamination condition that the oil is allowed to have.

I can also advise that the contents in our online store machinery maintenance materials, purchased at all-machinery-maintenance-training-materials, may prove to be a useful reference for you.

It is dangerous for me to talk generalities when it comes to your machinery. Each machine’s function and service must be well understood if you are setting health condition parameters for that machine. I hope that I do not lead you astray or give you inappropriate values in my following comments. I must include this disclaimer, as really I am offering advice in complete ignorance of the real situation. In the end you will need to make your own decisions on the machine reliability standards you set.

If you want to use our expertise in future, or assist you in your change efforts, we would be happy to submit our consulting rates at your request. Please contact us and we will send you what information you need.


 

One of my key responsibilities is to develop corporate equipment design, purchase, and installation standards. We are transitioning to a more proactive maintenance culture and are using life cycle cost methodology for new equipment purchases, refurbishments, and repairs.

While we recognize this will increase upfront costs, we also recognize this investment will pay for itself order of magnitudes over the useful life of the assets. I come from a reliability culture and came to this organization to help facilitate the transition. So you do not have to convince me of the merits of this approach, I’ve seen it work.

I am working on individual rotating asset standards in order of priority. Started with motors, then to centrifugal pumps because it is the most numerous driven asset, etc.

But I have recognized I need for a general precision design and assembly rotating asset standard as a catch all. (to cover assets missed and as a stopgap until I get more of the rotating asset classes finished.)

I am using the following ISO recommended overall vibration acceptance criteria for new equipment not covered under specific rotating asset classes. For the most part specific rotating classes have similar acceptance criteria.

Here is a colored chart that shows ISO 10816-3 general machinery allowable vibration by measuring non-rotating parts information in a different way. Be aware that these vibration values are overall values and do not explain why the vibration is happening. You will also need to analyse and interpret the harmonics of both the velocity and acceleration signatures to find the cause of the vibration.

Machine vibration limits to USA standards

In an ACE 3T standard format for a new or rebuilt machine I would use Good/Better/Best as… Boundary of Zone A (top of green area) / two third of distance from zero to Boundary of Zone A / one third of distance from zero to Boundary of Zone A. Some people would justify extending the tolerance to the top boundary of Zone B (yellow area). But that is not precision thinking—that is breakdown mentality and it breeds a reactive culture. Your production plant would suffer accordingly if you only act once a health condition upper limit is breached. To me it is better to use Zone A as indicator of health and use Zone B as grace time in which to get the machine fixed before it breaks.

The values in the table above are RMS and not PEAK. RMS is square root of PEAK. What scale you use depends on the wave form provided by the vibration analyser equipment.

I have a pretty good idea on what range of precision fits and base flatness I will need to obtain this but would like your thoughts. I also have some interesting questions concerning balancing standards.

An acceptable fit and tolerance is not enough to confirm a shaft or housing is suitable for use. You must also confirm its shape meets FORM requirements—circularity, cylindricity, straightness, squareness, flatness, and whatever other FORM parameter applies for the part. The bearing manufacturer specifies the worst shaft and housing machined tolerance and FORM allowed. The required surface finish is also specified by the bearing manufacturer.

We are transitioning from a culture where we rarely asked for these types of requirements from our supplier base. Some will be capable of performing to this level and some will not. I must have good justification for my requirements when I receive pushback from engineering and purchasing (which I will when the supplier base complains and people they are used to working with are no longer being considered.)

Your managers will go where the money is. They have been trained by our society to be profit-driven maniacs. Show them a dollar more than what they can get now and they will follow you anywhere.

One thing that you can do to get management support for the more excellent machinery standards that you want to introduce is to work out the business-wide cost of failure for each of your ‘bad actor’ assets. If when a ‘bad actor’ machine fails the business loses $10,000s and even $100,000s of operating profit you have a great business case to justify using higher quality standards that reduce the odds of failure. If by moving to more stringent reliability standards you reduce failures to the point that a ‘bad actor’ is no longer in that category, you can easily quantify the financial value of adopting the standards by the additional profit that they will bring to the business. Money talks, and all you need to do is show people just how much more money your higher standards will make for the production group and for the company.

In the PEW methodology we always start its introduction by finding out how much money its use will generate for a company. We take each asset’s failure history and annualise its total business-wide losses from past failures. We then determine the future business profits from the reduction in breakdowns and forced production stoppages once PEW is adopted. Usually you can find $10,000,000s of new profits for a company over the coming years. It makes for a very powerful business case to justify rapidly moving the company to operational excellence performance (Further down the page information is noted on the risk matrix approach used in PEW to identify potential savings).

You only need to start by telling your fabricators and suppliers two things: First, what results you really want and the tolerance that you will accept (the Target and Tolerance from the 3Ts); second that you will favor those Suppliers who deliver the best quality for the same price. Over time your Suppliers will react accordingly and you will get better quality for the least price.

You do not ask for a more demanding grade—that will cost you money. You advise people that you will accept what the manufacturer of the item concerned requires as the minimum grade in order to get full service life from the item (that becomes the ‘Good’ point in the Good/Better/Best 3T scale). But you prefer a higher grade if it can be provided for free (i.e. the ‘Better’ or ‘Best’ from the 3T scale) and you will reward the order to the Supplier who delivers the higher grade for the same price. In time, human ingenuity, entrepreneurialism and the capitalist system will do the rest.

You want a higher grade where possible even if you do not need it. We make the presumption that the nearer to perfection a thing is the less trouble it will cause (i.e. the Taguchi Loss Function perspective), because variations that produce defects which later lead to failure become less and less as we approach perfection. By setting a scale from ‘good’ to ‘best’ you create the opportunity for world class performance to be achieved (for free) whereas using only a pass/fail point tells nothing about the quality needed to produce world class reliability results.

What you do by following the above approach is you set a challenge for people to beat. People love a challenge against which to test their capability. Try it, over time you will see a positive difference in their quality performance and price. You may first want to ask your suppliers how they are going to go about meeting the standard and price combination and see what they say.

So your recommendation on the following categories would be greatly appreciated. Keep my overall vibration standards in mind as you answer these. They are not as stringent as some industries require.

Base flatness: I believe for general equipment I can get away with 0.005”/ft. I could specify a PIP style base or one machined to API standards (0.002”/ft) but I think this is overkill for my industry and the overall levels of vibration we are currently trying to achieve. Your thoughts?

A three sigma outlier value of 0.005in/ft is a severe softfoot problem for a machine running on roller bearings. Depending on bearing type the clearance in a roller bearing can range from initial preload to a few micron gap. If the machine is then bolted down on a base that is 0.005in/ft un-flat (that is 375micron/m un-flatness) it will equate to dozens of microns more preload within the bearing from machine distortion due to softfoot. Even API baseplate standard is not as good as you need for outstanding reliability. The really silly thing is that it is as easy to machine a base flat to high precision as it is to machine it un-flat.

In an ACE 3T standard format I would use Good/Better/Best as… API flatness / One IT # Lower / One IT # Lower.

Fits and tolerances: For bearing to shaft journal and bearing to housing bore fits, I understand the importance of precision tolerances. (whether it is the tight fit or loose fit side) for size, concentricity and cylindricity. I think I can specify IT6 and still get away with my overall vibration level. Do you think I have to go to IT5?

The tolerance and FORM specifications are not up to you to chose. The roller bearing manufacturer tells you in their catalogue what FORM grade you need to get the full bearing design life. You can chose a less demanding FORM, but you must then expect less than full design life from your roller bearings.

In an ACE 3T standard format I would use Good/Better/Best as… Brg Manuf maximum FORM / One IT # Lower / One IT # Lower.

Balancing: Fans are already specifically covered by G2.5 Centrifuge rotors will be balanced to G1.0. The rotors mass on these two types of equipment are large in relation to the overall machine. They stay together as an assembly so in the example of the fan which is interference taper fit to the shaft it can be balanced as an assembly on the arbor and swapped out as a unit. If the shaft ever becomes damaged, we would send it to the fan manufacturer to remove, refurbish, or replace. We are capable of field balancing so the whole activity becomes a swap out.

I have no doubt that the balancing shop will give you a certificate that proves the balance grade is what you asked for. As you state below, the problem is loss of balance grade when rebuilding the machine. The problem is that the G1 balance is lost during the equipment rebuild.

To maximize the chance of keeping G1 grade during rebuild I use drive components that remove as much human error as possible (e.g. keyed hubs with a close sliding fit on a cold shaft—this also makes the technician remove burrs from the shaft, else the hub does not slip easily along the shaft.). I also ensure the technician is well trained and practiced in precision assemble methods. I write rebuild procedures that specify exactly how to reassemble the machine so that the G1 balance grade is maintained in the operating equipment. Here again I specify FORM parameters for all the rotating components such as shafts, hubs, bearings, etc—straightness, concentricity, circularity, etc. I would expect the technician to take and record accurate micrometer readings of dimensions and I would expect accurate records of total indicator run-out during FORM measurements.

All component measurements would need to meet FORM limits else the item would be replaced or rectified. Depending on the rotational speed and the distance from the center of rotation I would get the mechanic to weigh the bolts to check they are all sufficiently close in weight. Similarly the key length and size would be specified and proven accurate as part of the assembly procedure. I would do the same for washers and any other item that may prevent G1 grade balance.

None of the above assembly practices are difficult or surprising to do for a competent precision maintainer—it is simply applying the precision practices of an expert.

These days I ask the technicians to digitally photograph each installed part and I save the photos as evidence of machine condition at time of rebuild.

Smaller rotors on general equipment of less significant mass where field disassembly routinely occurs are not as easy an answer. Shafts and hubs are an example. Most are single keyed and set screwed in my world. Their mass is not significant in the overall picture. You can specify an AMGA Class 9 hub and get a good overall balance as machined because of the concentricity tolerances. But if I understand manufacturing and disassembly processes correctly, as they happen in real life, the amount of allowable clearance for the single key and slip fit hub would negate anything I might require for balancing after the fact as we routinely assemble and disassemble these components in the field during replacement. The chance of offsetting the center of mass when we tighten the set screws is virtually assured. The only benefit I see for balancing is where drop in units such as the fan wheel assembly are replaced in the field.

I think I need to say something on the subject but am at a loss to how to word it. Maybe spec-out coupling to AMGA standards buy precision balanced shafts and other specific verbiage? This might be a tough one to put into exact words as it really depends on how the vibration is dampened, the machine is supported, and the size of the rotor. Know that for concrete/epoxy grout sub-base supported machinery, we will spec a 5X mass of the flooded equipment and use best practice grout installation. So in most cases we will have a good foundation to work from. And where equipment is attached to structural steel or the roof, we will design and purchase sturdy steel bases. (Wide flange beams, square tubing and gusseting, not channel) Your thoughts on how to word this?

I am a great believer in highly detailed procedures that lead a Mechanic through every step and explains with diagrams exactly what activities to do in a step, and exactly what work quality to deliver with each activity. I incorporate ACE 3Ts into each activity needed to complete a task. I put in as much visual management content as possible (so I use fewer words but include more annotated diagrams and pictures). Once you write a ACE procedure on exactly how to do the job right in the right way you give people the information and answers to deliver excellent performance.

Overhung Load: Cantilever loads, like drive belts, is another one of those areas where left unspecified you rarely get in trouble. But it is that 1 out of 100 example where the OEM in a competitive environment does not use a conservative enough safety factor for design, coupled with someone in maintenance using their finger as the calibrated tension gauge that gets us in trouble.

The failure events with long odds of happening carry huge business implications and lasting operational risks that few people appreciate. Whether you accept that the loss of a piece of plant has negligible likelihood of occurrence, and thus you will do nothing to prevent the incident, really depends on how much money your company can afford to lose should the outlier event happen. If you cannot afford to have a disaster then you must either prevent its occurrence with your business practices and procedures, or take out full insurance cover to pay for the consequences of the disaster. To me, spending 10% more capital buying better equipment which greatly reduces the odds of operational failure is easy to justify when compared to buying lower cost equipment that heightens the chance of disaster.

For example, I carry fire insurance for my house. The odds of having a fire in this house are incredibly small. If I managed my risk by saying that because the odds of a fire here are (say) one in a ten thousand houses per year, and because I will live in this house for only 20 years, and a fire will be an incredibly rare event to happen during my stay, I thus do not need fire insurance; I could save $400 a year on insurance. But I must also think about the consequence of the event. Remember that the odds are a mathematical construct by insurance companies—it is actually a lie because the ratio is a point value at the average of a distribution of many events. It is not the real truth of my situation. As my house gets older its construction and its electrical systems are aging and their reliability must be falling. My wife and I are aging. We will fatigue quicker with age and we will make more mistakes due to age. That 1 in 10,000 houses/year event could happen tomorrow if various scenarios unfolded. I cannot afford the $500,000 it would take to rebuild my house. And so I pay for fire insurance on which I never expect to collect just because I cannot afford to lose my house. The odds of having a fire happening are totally unimportant, it is the consequence of the event that I must protect against.

Many managers in business will do little to prevent highly unlikely disasters because they use the remoteness of the chance of the event as reason to do nothing. What they do not then see is that over time such an approach to making choices by people throughout the company creates thousands and even tens of thousands of unmitigated outlier risk situations in their business (Not only are your Managers and Supervisors making decisions, your Mechanics and Operators regularly make three sigma risk decisions too). With thousands and even tens of thousands of undermanaged negligible risk scenarios now in the business it naturally follows that some must occasionally produce unexpected failure events. Their business now starts to suffer random failure events continually and work becomes a never ending ‘fire fighting’ exercise for everyone in the company. All this comes from thinking that negligible failures can be neglected. PEW addresses this problem throught its ACE 3T error proof procedures.

In the case of your three sigma outlier overhung equipment failures my suggestion is to protect against them with better equipment and better operating and maintenance practices if your operation cannot afford to have the machines fail. The lasting consequence will then be fewer production equipment breakdowns and loss outages forevermore.

In the PEW methodology we identify three sigma outlier failure events and make sure that the management in a company really understands the true size of the neglected risks that their business is carrying.

I have seen statements in technical articles concerning overhung load which read something like this.

“Calculate the actual overhung load at worst case design conditions and make sure it is within the equipment manufacturer’s allowable load.” OK great, so vendor XYZ in a competitive environment comes along and his equipment is marginally designed but he does not want to go to the next biggest shaft, widget, etc. I.E. it meets the literal wording of the statement but not the intent of conservative design. Then along comes a ‘know it all’ mechanic who doesn’t believe in tension gauges for tightening belts. He knows Operations will not shut the equipment down after belt break-in no matter the quality of his work, so he does his own belt tightening strategy with finger pressure instead of measuring deflection as required by the OEM.

Most people want to do good work. The trouble is that they were poorly taught at the beginning of their education. (You only have one chance to put the right skills into a person. Once they learn wrong knowledge it will be hard to break the bad habits that result.) They now need to be retrained and taught what is the best practice to use. Training alone will not change their bad practices. You must also include into the job a written instruction on the correct practices to use and audit the job to prove that the right way is now being done (you become the ‘policeman’ who makes sure people follow the new rules).

If you want to move your new company quickly to high reliability performance you will need to do a lot of re-education; firstly managers and supervisors, then the workforce. Later you follow-up with the policing audits.

These days it is easy and quick for a technician to correctly do a belt alignment with laser light and proper tension measurement. Knowing that the best chance to influence Maintenance Crew practices is if the Crew Supervisor fully supports and champions a change, I would work with the Mechanic’s Supervisor to develop a belt alignment and tensioning procedure, train the Mechanics to do it right, ensure the procedure is available with every Work Order in which belt tightening will be done, and lastly ask the Supervisor to do a monthly audit when such Work Orders are being done to confirm his people are following the right procedures. Most Mechanics would love to know exactly how to do world class work. It makes them proud of their skills.

I guess what I am asking if it was you, what safety factor would you put in place to not over design equipment but reduce failures in the real world.

Parts’ failure is due to the material and atomic structure failing from loads the structure cannot support. If you control the stress levels in the material structure, so that operating loads are well within the structural capability, then the component will not fail. You can manage with incredibly low safety factors if you control material of construction stress. In the aircraft industry that have factors of safety as low as 1.1 to 1. But they also have great quality control, licensed mechanics with great skills and great monitoring of the actual stress put on components.

In general industry the operating stresses and loads are not that well controlled, so in general industry I would be conservative. In general industry machine parts will have to be able to absorb more stress and overload than design values because the operating processes used offer poor control of the damage inflicted on parts’ materials of construction.

That is why your use of life cycle costing it so important to your operation’s future. You can model the conservative choice verses the borderline choice. If by being conservative you can save a lot of money (compared to being borderline and suffering subsequent breakdowns) then the money saved would more than pay for the stronger equipment. If the conservative option reduces breakdowns and stoppages, and thereby increases the company’s operating profit, the life cycle financial modeling will make the best choice clear to company management.

You can use borderline equipment if you also accept that they will require more care and more maintenance which requires more downtime and more resources that the company must pay forevermore until the item is decommissioned.

Here is an analogy from our fan standard concerning stress associated with centrifugal force.

Stress due to centrifugal force shall not exceed 75% of the yield strength of the material on any portion of the fan rotor at maximum design operating temperature.

So our overhung load statement might read.

Manufacturer shall calculate overhung load under worst case design operating conditions. Calculated overhung load shall not exceed XX% of equipment manufacturer’s allowable load. What should XX be?

Is this a good adder to the standard or is it not worth considering?

A maximum 75% yield stress in the shaft does not make sense to me. That level of stress on a ferrous shaft puts the operating stress well up the material fatigue life curve. Such an overhung load would bend the shaft at the nearest bearing causing excessive fatigue stress on the shaft surface as it rotates. It would also generate massive localized loads within the bearing rolling elements and raceways. In machines under continuous use at 75% of yield stress the presence of regular rolling bearing failures, or occasional snapped and/or bent shafts would not surprise me.

I would be concerned if a radial load from a drive belt bends a typical industrial machine shaft off its true center of rotation by more than one thousandth of an inch per foot (0.001in/ft) (Larger diameter shafts would not bend so easily as smaller diameter shafts). The 0.001in/ft parameter is noted in the book ‘Machine Design theory and practice’ by Aaron D, Deutschmen et al.

You should really do a stress calculation for the shaft and identify the maximum deflection that ensures the maximum shaft fatigue stress does not reduce the material fatigue life to less than the required number of rotations needed during the entire operating life of the machine. (That is how they do it in the aircraft industry.) That stress value will then be the ‘XX’ you want for the shaft. Such an approach incorporates the dynamic situation and is not just a static load analysis. For ferrous shafts the ‘XX’ for infinite fatigue life is about half the ultimate tensile stress.

I would want to measure shaft deflection under belt tension as well as measuring the sideways belt deflection required by the OEM to set the necessary tension. In an ACE 3T standard format for a drive belt on an overhung small diameter shaft under full tension load, I would use Good/Better/Best as… 0.002” per ft TIR / 0.001” per ft TIR / 0.0005” per foot TIR. The deflection movement would be seen with a dial gauge placed to touch the overhung shaft at the maximum load point.

First you would need to check if the shaft is already bent and then remove the effect of the bend from the TIR value (a shaft bent more than 0.001”/ft is a problem). The deflection also depends on shaft bearing type and the bearing operating internal clearance. Some of the shaft deflection may be due to clearance release when the equipment is cold (a radial clearance movement of 0.0005”, half a thousandth of an inch, between operating and ambient temperature would not surprise me—but that depends on the bearing and its duty). The effect of internal bearing clearance would also have to be removed from the belt tension deflection.

What you really want is a straight shaft. What you really want is a perfectly straight shaft. But no one understands what you really want because you do not tell them. You only tell them what you will accept as bad enough to be good enough to pass the single point test. Without a scale to gauge the result against what you really want, no one will bother doing better. A one point pass/fail test guarantees you will have occasional quality problems with your suppliers and with your mechanics’ workmanship. Whether that matters depends on how much of a disaster results from the problems an occasional bent shaft will cause your company.

I would specify a 3T ‘best’ shaft that has less than 0.0001in/ft deflection—almost perfectly straight. My Good/Better/Best specification would be 0.001in/ft—0.0005in/ft—0.0001in/ft. I do not ask for 0.0001in/ft accuracy—I will accept 0.001in/ft—but I really want 0.0001in/ft if I can get it for free. (See the comment above at the ‘You only need to start by telling your fabricators and suppliers two things:’ paragraph about how to approach suppliers with the 3T quality scale.)

I run into too many people who know individual reliability rules of thumbs but do not know their significance in overall equipment reliability. This is especially frustrating in the area of life cycle cost analysis. I have put together real life case studies where we have had successes, but I find them difficult to complete as maintenance data is hard to come by (especially from equipment suppliers).

In the PEW method we show each operating assets total annualized business losses from past failures on a risk matrix like the one below. If the direct maintenance costs of historic repairs are not already available, the maintenance planner calculates the direct maintenance cost in detail. In a process operation we factor-up the total business cost to be 10 times the direct maintenance costs—a $10,000 maintenance repair actually costs a company $100,000 in total losses. (I have found that for process plants 10:1 is a reasonable indicative ratio to use. Batch plants tend to be around 5:1.)

Everything in PEW is financially justified before making any changes in the workplace. Once you show management how much more money they can make from having high reliability you will see their eyes open wide. When you see that happen you know that you have got them! By doing the justification this way you bait the hook with a juicy prize.

Plant and Equipment Wellness first financially justifies reliability improvements before undertaking them

Any thoughts on any of the above topics would be greatly appreciated.

You are on the right track to higher reliability. You cannot improve a process without first establishing the performance range that the process must deliver. In the image below you can only create a capable and controlled process after the acceptable performance limits are set. So set your standards high so that you can build a high performing company. If you set average standards, or below average standards, then that is the sort of company that you will get—how can it be otherwise.

How to go about creating proceses that are in-control and capable

As you have pointed out, you will have trouble with people who think that what is now done is sufficient. In such situations I advise those people that if they are happy with the company/department performance and its results then do not change a thing. That is what you will have forevermore. But if they want a better future with better results, then those changes which will bring them the future that they want must be made first. Only after changing can a new, better future arise.

I believe at this stage that the best approach for you to use to start shifting your company reliability culture in the right direction is to specify 3T (Target-Tolerance-Test) quality criteria in the equipment standards that you write. Let the standards you propose become the least acceptable pass/reject quality (a ‘good enough’ performance), plus include 3T target standards which are world class quality (the quality your company needs in order to be the best). Then everyone will know how close their work performance is to world class results.

I hope the above thoughts and ideas help you.

 

My best regards to you,

Mike Sondalini
Managing Director
Lifetime Reliability Solutions HQ

 

PS. If you require advice on industrial asset management, industrial equipment maintenance strategy or plant and equipment maintenance and reliability, please feel free to contact me by email at info@lifetime-reliability.com