PM Optimization vs RCM

[Raju]

Dear All,
1.We need to establish a Reliability Center at a steel manufacturing plant. As the plant has been in operation for over eight years, we have decided to adopt PM optimization technique to refine the existing PM program. Are we barking up the right tree?
2.Could anyone suggest the course of action that we should adopt for establishing the Reliability Center? Is our decision to keep statistical methods out of it (due to paucity of time, and also since the plant has been in successful operation for a long period) correct?
Regards

[Rui Assis]

You may find it instructive to take a look at this site which contains quite a few posts on RCM versus PMO.

http://maintenanceforums.com/eve/ubb.x/a/frm/f/209103451

There is also a very interesting article on the same subject "Is Streamlined RCM Worth the Risk?" from John Moubray which I can forward to you if you send me an e-mail.

[tarik]

Very good question,
You are suggesting keeping statistics out of it, what would be the instrumental tool that you are going to adopt to improve the performance?

In addition to changes in management and practices, statistics can you give the type of visibility that are needed when complete physical understanding about the plant equipments is lacking, put the data you have collected into work. Data is another huge company asset, which, in many times, not used to its full potential.

ReliaSoft recently introduced a good tool for Reliability Centered Maintenance RCM++ http://www.reliasoft.com/rcm/index.htm which I advise you to consider (along with other ReliaSoft tools)

[Rui Assis]

Tarik raised a very important issue – there is no way how to improve operations reliability without the tool of statistics. Both PMO or RCM need quantitative methods and specifically statistics when it comes to the decision diagram (the “worth doing” sort of questions…) in order to decide what maintenance policy to adopt. RCM++ from Reliasoft can be of great help in guiding the way.

[Raju]

Hello Rui, Tarik
Thanx for your immediate response. But to be frank, you have only managed to scare me rather than start me off on my tracks! I am very new to reliability and, as my understanding goes, it is all about (at least, mostly about) improving the Planned Maintenance program. Even without too much of statistics (save the no. of times a particular failure mode has occurred, hours spent on PM / PdM et al), can't I analyse the PM tasks for their usefulness and/or appropriateness (or whatever the noun for 'appropriate' is!)? Just by studying the failure modes that they are supposed to be addressing, and then studying the failure data to establish whether those particular tasks need to be carried out at the existing frequency or not, whether a modification could be thought of, etc?
Please keep in mind that I am very new to this field, as I had mentioned earlier, and hence it would be great if you could provide some implementable methods.
Thanx
Regards
-Raju

[Raju]

"In addition to changes in management and practices, statistics can you give the type of visibility that are needed when complete physical understanding about the plant equipments is lacking, put the data you have collected into work."

As regards Tarik's opinion reproduced above, do remember that we have an enormous knowledge database built up over the past eight years - a sufficiently long period to have gathered a 'near-complete' physical understanding about the plant equipments! So the decisions on the "worth doing..." aspects could come along quite easily in this case.
Regards
-Raju

[Andrew]

Raju,

You are asking an often asked question in most every industry! I also definitely recommend the John Moubray paper Rui Assis recommends. Regardless, it's not as scary as it may sound at first. Your facility is only eight years old. I implemented RCM at a facility that was 28-years old with over 425,000 maintenance records. And, yes, we used statistics to analyze the failure data in those 425,000 records.

With regard to your original post, I offer only my opinions. On the point of your first question, my opinion is that you ARE barking up the wrong tree. PMO tends to be a corner-cutting technique which, I believe, is its attraction. For example, you stated, "Just by studying the failure modes that they are supposed to be addressing..." What about the failure modes that the current tasks aren't addressing?

We went through about eight iterations of 'PMO' before moving to RCM. Under RCM we considered all potential failure modes and their consequences. We found that, not only did PMO fail to implement applicable and effective tasks, it had actually deleted them! In the end, 15% of the components that RCM identified as causing immediate plant shutdown, PMO had decided to run to failure. After implementing the RCM program, unplanned shutdowns were reduced 75%.

Bottom line, when you consider the potential consequences of equipment failures in your plant, do you really want to cut corners?

In your last post you state, "...studying the failure data to establish whether those particular tasks need to be carried out at the existing frequency or not..." How, without statistical analysis of your failure data, do you plan to adjust current task frequency?

Using a PMO approach, we had adjusted PM task periodicities from 9-years to 39-years! There was no technical basis for such a huge change, in fact PMO had identified 41 failures of the component type in question. However, in all 41 instances, the failure mode wasn't the failure mode that the PM task was "supposed to be addressing." PMO didn't implement an appropriate PM task at the correct periodicity, it merely extended the periodicity of an inappropriate task.

RCM drives you to modifications whenever an applicable and effective PM task can't be found. PMO just stops and leaves you vulnerable.

RCM is a living process that provides an auditable trail between the PM task and the plant design "weakness" that is addresses. PMO tends to be a snapshot of current practices and lacks the auditable trail such that you need to do it over and over again as the "history" of the PM program is forgotten.

A final point about PMO is it fails to consider costs; cost of PM, cost of failure, etc. If you plan to optimize your maintenance policies, what are you going to optimize? You're either going to optimize availability or O&M costs. That's sort of the fundamental thrust of RCM and the genesis of the RCM methodology in the first place.

With regard to your "'near-complete' physical understanding about the plant equipments", I would beg to differ. Failure modes, or at least failure mode probabilities, change as equipment ages. With only eight years of operation, I suspect you have alot left to understand about your plant's equipment.

[Raju]

Hi Andrew,

Thanx for a very informative response! The issue in question is the set of failure modes that the existing PM program does not address. However, in the PM Optimisation technique, there exists sufficient scope to identify those failure modes as well. Discussions with the maintenance engineers & the operators, and also analysis of the P&IDs, which constitute two of the most important steps in the analysis, provide ample scope for identifying the failure modes not yet encountered or conceived of by the facility.

Also, it would be highly appreciated if excerpts of the PM Optimisation programs undertaken by your facility could be shared with us.

Regards
-Raju

[Alvin]

Hi Raju,

PM Optimisation is a revolutionary concept that can work wonders in my opinion. I have been practicing RCM in my manufacturing plant for the last six years, and despite that, I would advocate PM Optimisation as the primary reliability improvement program in any already up-and-running industrial scenario. The reason is the speed and accuracy possible with the technique. Moreover, it passes as the best possible 'living-program' as of now.

All the major airlines and air forces have a variant of PM Optimisation program built into their maintenance management systems - Srevicing Schedule Development. This program collects inputs from the user stations periodically on the existing servicing schedules, carries out brain-storming sessions with operators and manufacturers and reviews the PM programs for their fleet on an annual or a bi-annual basis. Needless to say, airlines are amongst the most reliable industries in the world!

People who oppose PMO tooth and nail have very poor understanding of the technique or are hard-core conservatives who are happy that they are experts in their present domains!

However, I fully endorse the view that no reliability program can be successful without statistical analysis, as brought out by many eminent writers in this thread.

Good Luck
Alvin

[Raju]

Hi Alvin

Thank you for that interesting message which has helped us to make our final decision - we ARE going to go ahead with the PM optimization program! http://63.227.80.170/discus/clipart/happy.gif

I would like to place on record my sincere thanks to Rui, Tarik, Andrew and yourself for a most educative thread. http://63.227.80.170/discus/clipart/happy.gif

Thanks All
-Raju

[Rui Assis]

I am currently involved with my first case in RCM – a machine that picks and places electronic components on PCB´s for Pioneer car radios. The project started in July and is estimated to end by September.

I had the same doubts at the beginning. I read a lot including John Moubray and other authors defending the OPM alternative and made up my mind towards RCM after reading John Moubray arguments considering the legal side of the picture in the article "Is Streamlined RCM Worth the Risk?". I bought the two recommended norms: SAE JA1011 issued 1999-08 and SAE JA1012 issued 2002-01. Most of the text is founded on Moubray’s book. I am doing everything accordingly and I feel quite confident on how to proceed.

Good luck

[suresh]

hello rui assis
i wanted to know which software you are using? is it reliasoft products only? what all sw one needs to carry out RCM on a machine? i think only rcm++? what do you suggest? i want to carry out rcm technique on a big machine that has many components, mostly mechanical. how should i proceed? i want concrete practical steps to guide me. not articles by moubray etc.
thanks
-suresh

[tarik]

Suresh,
The next issue of Reliability Edge http://www.reliasoft.com/newsletter/
will have a good example for RCM and RCM++. The title of the article will be "Comparing Maintenance Strategies Based on Cost and Availbility".

[Steve Turner]

Raju
Its been a year now. I am new to the forum and would be interested to learn about your experiences with PMO.
Regards
Steve