WHAT ARE ALL THE METHODS/RESOURCES TO SPECIFY THE PM FREQUENCY FOR AN EQUIPMENT(SPECIALLY FOR ROTARY EQUIPMENTS)?

It has been said here, based on failure development periods.

Please do not shout!!! (write in capitals)

Could you be more specific? I consider ferris wheels also rotating equipment

http://en.wikipedia.org/wiki/Ferris_wheel

But to threat it as my ventilator, which my wife periodically cleans (PM) would not be acceptable.

thanks josh
According to me these data are only one source to do this,
But what about setting the PM frequencies where previous data are not available (e.g setting frequencies for the first time in a new plant)?

Hi svanels
Actually both the equipments you mentioned are rotary equipments,
Now the question is ,
Is there any differance between the Resources witch can be used to specify the freq. for those two equipments or they may have their own resources.
Foe example;
One source can be FAILURE DEVELOPEMENT period or it may be the MTBF which in both cases should be driven from previous data, But if there is (for example)any published standard,then the resousces MAY be differant.
That means there may be differant standards for each equipment.

Check OEM manual. Else, Start once a month and increase or decrease from here, depending on how fast your equipment deteriorate.

This message has been edited. Last edited by: Josh,

Method / Resources:
* Failure Analysis results
* Experience with similar equipment
* Manufacturers recomendations at manuals, technical documents, website.
* Handbooks on Plant Maintenance, Maintenance Engineering, or Mechanical Engineering

I do not think the equipment type makes a difference on method to determine PM Frecuency.

This message has been edited. Last edited by: Eugene,

quote:

Originally posted by Eugene:
* Handbooks on Plant Maintenance, Maintenance Engineering, or Mechanical Engineering

eugene
could you be more specific about "handbooks on plant maintenance"?
Which type of handbooks?
can you just give an example?

Since it is a new plant/equipment, I would recommended performing an RCM to identify the PM as well as using the OEM for reference. Than, determine the criticality of the equipment based engineering specs and process flow and business impact. Place the equipment into three categories: Critical, Moderate, Low. The Critical Equipment should have monthly PMs, Moderate Monthly to Quarterly, and the Low Quarterly to Semi-Annual or even longer. After a year, re-evaluate the program by looking MTBR/MTBF and so on as others have described.

First of all no handbook will tell you how to maintain a equipment model 123XYZ fabricated by ACME Co.; but if the 123XYZ unit has bearings, motors, variable frequency drives, couplings, a lubrication system, a good handbook will tell you how these systems work and how to maintain them. For specific details on a 123XYZ you must consult the equipment manufacturer (manuals, engineers, etc.).

Well, I got the following handbooks for reference:
* Standard Handbook of Plant Engineering, by Robert C. Rosaler, Mc Graw Hill Co.
* Maintenance Engineering Handbook, by Lindley R. Higgins, Mc Graw Hill Co.
* Mark's Standard Handbook for Mechanical Engineers Mc Graw Hill Co. ( http://www.mcgraw-hill.co.uk/cgi-bin/browse_tpr1.pl?subject=sctech )

Check other editorial companies like:
* John Wiley & Sons http://www.wiley.com/WileyCDA/
* Industrial Press Inc. http://www.industrialpress.com/en/default.aspx
* Prentice-Hall http://www.prenticehall.com/

I will throw some writers in the pot:

Heinz Bloch, Joel Lewitt, Terry Wireman, Keith Mobley, Vee Narayan

Editors
Plant Engineering
Industrial Press Inc.

In determing PM frequencies we need to understand more about the failure or potential failure of the equipment. Time based PMs only account for a small percentage of PM activities. Most PMs we need to perform are based on the health of the equipment. I would start by understanding the PF Curve and PF Interval. I wrote an article recently for Plant Services Magazine where I discuss the subject in more detail. Go to: http://www.plantservices.com/articles/2006/136.html I hope this article helps.

The next question would be: How to develop the Pf curve for critical equipment 010X001 if I have its maintenance history records available.

Also, somewhere in the tread the topic of new equipment showed up. If new, no historic data, question, how much data must be available to develop the PF curve? One year is enough?

Eugene,

Eugene,

If you have equipment history it should help you define what can be done to prevent or predict failure of the equipment (visual inspection, Vibration analysis, etc.). Understand what you are looking for is known or likely failure modes (equipment failures). On one critical piece of equipment this could be 1 or it could be 100. Along with equipment history use operator and maintenance personnel’s knowledge of how the equipment fails in order to identify the PF Interval (the PF Interval is the segment of the curve between the "P" Potential Failure Point, this is the point where equipment degradation is detectable, and “F” is the point where Functional Failure has occurred). You can then hopefully the point on the PF Interval where you need to schedule an outage and do something to the equipment (restore, change out, etc). Identifying the point on the curve where you have time to schedule the outage before the equipment functionally failed is the toughest part.

If you do not have equipment history on equipment then the process is the same using the knowledge of the operators and maintainers along with manufacturer’s specs, and any other information which help you begin this process. What is great wants this process is established is that if a failure does occur you should learn more about the PF interval for that specific failure mode.

I attached 4 PowerPoint slides (be sure and read the notes which are attached to the slides) that shows the PF Curve with the PF Interval. I would recommend everyone takes a course in RCM or read a book with information on FMEA and it will change your view of failure. I have an article coming out in August with Plant Services Magazine talking about FMEA and how to use it to design your current PM Program.

I hope this helps.

PFCurveFailureCurves.ppt (877 Kb, 135 downloads) PF Curve Slides

Eugene,

Sorry I did not answer your second question. In order to know the PF Interval for new equipment I would use manufacturer information and ask other companies who use the same equipment to provide you failure information and begin the process of learning from each failure that occurs. I would recommend RCFA to assist.
One experience I had (I wish I could forget it)
I had a hydraulic pump once that was hurting me and my company. It took me three months of failures on 7 pumps (7 production lines in three countries) in order to collect enough information to define the PF Interval but when I found it what a change we made to the failures of the equipment. Instead of looking at the hydraulic pump destroying itself in order to tell use when it failed we began to monitor it's health through monitoring flow, pressure, and taking oil samples and found where on the PF Curve these pumps would functional fail (6.8 GPM). We save millions of dollars quickly. That three months was a nightmare because even the pump manufacturer and production equipment manufacturer could not provide me with good information. This equipment was new technology so they were learning from the end user. I gave them none of the information or detection methods we used because they wanted to give it to our competition.

Just one of my experiences. I am sure you have had similar.

Ricky

Ricky,

I thoroughly enjoyed reading your slides and notes. Let me remark though that a question still remains answered: How do you determine the time intervals to perform the inspections in order to detect the P moment?

Thanks,

Rui

quote:

This equipment was new technology so they were learning from the end user. I gave them none of the information or detection methods we used because they wanted to give it to our competition.

Ricky, please forgive me for highlighting this sentence of yours, but I think you just raised a very important issue which, I think, deserves to be discussed in more depth.

I used to be a manufacturer of equipment myself a few years ago and I used to try hard to learn from my client’s experience, in order to continuously improve my products and serve them better in the future. Fortunately I never experienced information being denied what so ever. Today, I am on the other side of the “line” and I find it to be very useful to collaborate with manufacturers as I usually get the very last and up to date recommendations in return. Technology progresses so rapidly that we hardly buy a piece of equipment nowadays of which there is enough knowledge and experience both from the manufacturer and from ourselves with similar equipments, that enable us to put an effective and definitive maintenance program in place since the first moment.

Maybe I am being naïve…but I still think that it can be more advantageous to report experience (which is often done during periodic and informal visits paid by the manufacturer) than not to. I would like to hear our peer’s comments on this. Or perhaps it would be better to start a new thread on this issue…

Thanks,

Rui

Rui Tudo Bem!

I am not sure if I am following this thread correctly but please take this as input to the discussion.

As Ricky has spoken about, the entire point of on-condition maintenance as defined withint he SAE standard on RCM is to detect the "P" moment as you put it.

That is, the moment where the potential failure is detectable via the technology that you have chosen to utilise.

SO far so good, everybody has already spoken about this muich...

Once the P-F interval has been identified.
(And identifying this is a whole other discussion. It has been my experience that manufacturers often do not have this information for the assets in the operating context that you use them in)

Then the frequency of the task selected to detect this potential failure has to be less than the P-F Interval.

There are differing schools of thought on this. I have had some discussions with military types who believe that a task frequency of 1/3 of the P-F Interval is best. This is because the P-F is generally an estimation in most circumstances, and also to take into account the fact that the task may not be 100% accurate in detecting the potential failure.

One of the effects of this is increased routine maintenance activity. There are 3 checks for every P-F interval. So if the interval is two weeks then you are checking the item three times in this period.

Another view is that 50% is best. I subscribe to this generally unless it is a failure mode with safety or environmental consequences.

Either way, the interval is driven by the P-F interval.

Over time you will find that the interval was conservative, as it is generally based on engineering judgement rather than "crashing a few more assets". And you will be able to adjust the frequency to suit the more accurate P-F estimates.

But this is only part of the original question...

Statements like "based on failure development periods" are representative of the style of thinking that used to dominate the field about twenty years ago. Today, this is not quite the case. (Fortunately)

Modern RCM, as defined within teh RCM standard SAE JA1011, has led us to the point where we recognise there are 4 different types of routine maintenance, 3 different types of corrective maintenance, as well as reactive maintenance.

Within the routine maintenance areas there are:

• Predictive Maintenance
  
• Preventive Restoration
  
• Preventive Replacement
  
• Detective Maintenance
  

Of these only the second and third are recognised within RCM as "PM's" or preventive maintenance routines. (Unless PM means planned maintenance of course)

Each of these maintenance types use different means to determine the optimal frequency for a routine task. This could be driven by the P-F interval, by the "life" of the asset as driven by this failure mode, or by the tolerable probability of failure as it applies to a particular situation.

I hope this contributes to the discussion.

quote:

There are differing schools of thought on this. I have had some discussions with military types who believe that a task frequency of 1/3 of the P-F Interval is best. This is because the P-F is generally an estimation in most circumstances, and also to take into account the fact that the task may not be 100% accurate in detecting the potential failure.

One of the effects of this is increased routine maintenance activity. There are 3 checks for every P-F interval. So if the interval is two weeks then you are checking the item three times in this period.

Another view is that 50% is best. I subscribe to this generally unless it is a failure mode with safety or environmental consequences.

Hi Daryl,
I have had this debate with no satisfactory explanations as to why condition based task intervals should be "factored". Doing the on condition tasks three times during the PF interval means potentially you increase the cost of your condition based maintenance by nearly three times... Why would you want to do this if you are taking a conservative estimate of PF in the first place (which is what happens when there is uncertainty)??? Sure, if the inspection task is not robust, then the mathematics will allow you to factor this in, but if you are deploying CBM tasks that are not robust, then my approach would be to find out why they are not and fix this or look for other means of prediction. I think these rules came from one of those old wives that wrote the old wives tales... or as someone once so delicately put it at one of our workshops.... " around here - someone writes something on the dunny wall and before you know it it is policy".