Do anybody know how can we measure the effectiveness of a RCM Task selected to address failure modes with safety or environment consequences. John Moubray's RCM2 book says that a RCM task for failure modes with safety or environment consequences is worth doing if it can reduce the probability of the failure to a tolerably low level. But once a RCM team have decided a specific task, how we can establish if it was really worth doing, I mean how we can measure the reduction of probability of the failure to follow John Moubray's premise for evaluation?

Thanks in advance

I think the failure rate can be measured and indicator of the task or strategy effectiveness

"Doing the right task"

I think that I need to explain a little bit more my question. According to John Moubray's RCM2 book (page 102), the RCM team must decide the best task to address an specific failure mode. In the case of a failure mode with safety or enviromental consequences the best choice is a task which can reduce the probability of ocurrence for that failure mode to a tolerably low level. But my question is, in the moment of decision, how you can measure such probability reduction to asure that your choice was the best option. I mean, if you are deciding between thermography or ultrasonic inspection for instance. How you can measure the possible reduction of probability of failure to decide one or another option?
Thanks anyway for your previous answer.

You can measure or estimate the reduction in probability. The probability is reduced by the factor "number of times that the device fails and causes a problem (so the failure finding task did not find it first)" divided by "number of times that the device fails". It will take a long time to get data to measure these amounts, so you'll probably have to estimate it. Another way to estimate it is, 100% minus the chance that your failure finding task will find the failure before it causes a problem. You should know this from the RCM analysis.

Thanks BurraFoods, but I'm still doubtful. The point is, when you are in the meetings of the RCM group and the group is analysing a specific failure mode that could affect safety and the group choice was thermography inspection for instance, if we want to follow the rule, we need to measure in some way the reduction in probability of failure caused by our decision, to establish if the task could reduce it to a tolerably low level. But how we can measure it if you are not performing the task yet, you only have decided the task to start later, so you don't have any data to make some statistic. I have read the John Moubray's RCM2 book and he provides some examples but I think these examples are quite theoretical where you have numbers but he really doesn't tell us how you could get that numbers. In conclusion, I haven't found a method to estimate this reduction in probability of failure when you decide a task to address a failure mode with safety consequences in the moment of the RCM meeting, to know if the task is "worth doing" as stated in the Moubray's book. Any idea?

Les,

You are quite right when you say,

quote:

in the moment of decision, how you can measure such probability reduction to asure that your choice was the best option. I mean, if you are deciding between thermography or ultrasonic inspection for instance. How you can measure the possible reduction of probability of failure to decide one or another option?

The RCM team members have knowledge and experience; they should be able to judge what is most likely to solve the problem or reduce its consequences. If there is a safety consequence, there should not be an OR decision; should we use thermography OR ultrasonics, for example. Any combination of tasks that reduce the residual risk to a tolerable level is what is required. If either task alone will do it, then cost and do-ability considerations come into play. For non-safety consequences, cost-effectiveness is what determines the selection.
The estimation of risk reduction is a matter of judgement. We have trained and experienced operators, maintainers and specialist engineers in the RCM team for exactly this reason, to make such judgements. They could of course be wrong in some cases, and we may on occasion suffer the consequences. In which case we have to go back and revise the RCM study, perhaps after an RCA shows us what exactly went wrong.
I am of the view that RCM analysis must be done using a team, never with a single RCM expert. I know that others in this forum support the single RCM analyst approach, mainly for 'pragmatic' cost and resource availability reasons. I do not support their view. A team, knowledgeable in the culture of the company, the operating context and with a deep knowledge of the equipment is always superior to the RCM expert, however good he or she may be.

Vee,

Rightly or wrongly I think this was aimed at myself. Maybe not but I would like to comment on it anyway.

quote:

I am of the view that RCM analysis must be done using a team, never with a single RCM expert. I know that others in this forum support the single RCM analyst approach, mainly for 'pragmatic' cost and resource availability reasons

If it wasn't aimed at me, then let me clarify my position on this. I am not a supporter of sole analysts per se, I am a supporter of teams and of a combined team / sole analyst approach.

Let me explain further. When we need a team? If you think about it, and recall your own experiences in performing RCM, what happens normally?

There is a rigorous discussion surrounding functions, of which all are involved in. hould this be done with a facilitated meeting format. Definitely! All have an input into both the "wants and the cans" associated with the asset or asset system. (To paraphrase Moubray)

If they have been done correctly then the functional failures should fall out of them, so why do I need to get a group of people around a table to discuss this? NO need for a team here!!

The we get to failure modes. Do we need a team for this? If you recall you will have two or more days go by when the maintenance guys are very keenly discussing the situation, and the operational guys are less than fully utilised! (Wasted some might say!)

Why can't this be done by targetted interviews?

Effects - Operations, with some targetted interviews of maintainers where needed. But mainly the operations staff become highly animated during this part of the discussion and the maintainers are not so engaged. (Wasted time again)

Strategy creation - Targetted interviews of the maintainers, or of operators, or through short duration facilitated team meetings.

If you are going to do RCM in the 21st century, and there are other more advanced technologies starting to appear, then why would you try to force people down a path that became redundant a decade ago?

The resources are not physically available to engage in the luxury of fully team facilitated analyses. Pragmatism is what is required, not strict adherence to an implementation method that was well sold to the community, (And John Moubray had a lot to do with that) but is now out of date.

Best regards,

Vee,

Your quotation:

quote:

The estimation of risk reduction is a matter of judgement. We have trained and experienced operators, maintainers and specialist engineers in the RCM team for exactly this reason, to make such judgements.

I don't think this is always the case. That is, there are means to predict the level of risk reduction quite accutately using some of the technologies available in todays market place.

Best regards,

Thanks to all gentlemen, I was absent from my job a couple of days so I could'nt read the posts until today. After read the two positions I think that both have reason in part. Some cases don't require an exact calculation of probability to determine the effectiveness and the good judgement of O&M people provide the answer. But for some other instances you could apply quantitave methods as Daryl say "to predict the level of risk reduction". I'm not a statistical expert so I want to know about the basics of such methods and where I could get some information of them (I mean links). I apologize for my ignorance of this quantitave methods and I want some advise in this theme. I have heard about a technic called "Monte Carlo Simulation" to make probability estimations but I want to know more about of this concern.
Do you know something about?


Best Regards
Lester Pino

Daryl,
Please read my post in relation to the question posed by Les; it should make more sense when read in context. I was attempting to answer him, and was making a specific point about teams.
You are not alone in supporting the solo RCM-analyst approach, and there are others in this forum itself, let alone the wider world who strongly agree with you. So my comment was fairly general. You are of course entitled to whatever approach you think is best. Allow me to express my view that I do not accept this position.

quote:

If you are going to do RCM in the 21st century, and there are other more advanced technologies starting to appear, then why would you try to force people down a path that became redundant a decade ago?

Some time ago you commented on my use of the words right and wrong, calling them emotive. They were not, and certainly less emotive than your reference to the word emotive itself. I chose to remain silent then, as I did not want a non-productive debate. Now you are pushing those of us who prefer to use the team approach into the last century. You imply that your method is superior and technologically advanced, while the rest of us are in the backwoods.
I am a great believer in teams. This is the result of doing many RCAs, RCMs and similar work, While single interviews have their place in data gathering, I am of the view that teams are far superior and faster in doing RCM or RCA. They act as great memory joggers, allowing hitch-hiking of ideas sometimes bringing in lateral thinking. I believe they produce better quality of analysis that solo-flyers, but these views are conditioned by my experiences. Your experience is obviously different, and I accept your right to have a different view. But please dont state emphatically that those who dont agree with you are in the last century (perhaps that is not emotive in your view) however much you believe its validity.

Vee,

I have not explained myself fully, so I will try to do so here.

I don’t agree with the sole analyst approach and don't advocate it. Period!

I do advocate a mixture of team facilitation plus targeted interviews, with the analyst carrying out certain functions. (Which is different entirely than a sole analyst approach)

I wrote an article on it that was published pretty widely a while ago.

With regard to a redundant approach, I stand by that, but I think there is a need to explain it a little.

In the 90's when RCM started to become truly popularized, there were many more people around. Personally, I was able to get hold of resources to perform RCM analysis relatively easily. As the decade went on this became increasingly more difficult and by the commencement of the 2000’s it had become almost physically impossible.

Today there remains a need to get RCM and reliability methods out into industry, and they shouldn’t be restrained by an implementation method that came out of a period when there were more resources available for companies to utilize.

So, forcing companies down the route of fully team facilitated resource intensive approaches, where their people are under-utilized and spend more time away from the front line, is a redundant approach.

It is redundant not because it no longer works, it clearly does and we both agree with this, but because it is no longer possible on a large scale. (And becoming even more difficult every year)

This isn't my opinion, this is a fact. I am sure that you come accross this almost daily in your own consulting activities.

So, I stand by my point, fully team facilitated implementations of any reliability initiative are a redundant approach and have been so for almost a decade.

Just to clarify a little further, this does not mean to infer anything onto yourself.

Vee, the use of the terms “right” or “wrong” are obviously emotive terms, I don’t see anyway that they couldn’t be. (And It was only a couple of days ago I’m sure of that)

Best regards,

This message has been edited. Last edited by: <Ozgipsy>,

Les,

TO get back to your original question:

quote:

Do anybody know how can we measure the effectiveness of a RCM Task selected to address failure modes with safety or environment consequences.

In my book, The Maintenance Scorecard I wrote about something called the RCM Scorecard.

This was the first time the RCM scorecard had been published and it focussed almost entirely on the effctiveness of RCM strategies to "do what it says on the tin".

Namely, increase productivity and profitability of the asset base for a given level of risk.

There are proactive methods about for predicting, quantifiably, the effectiveness of certain strategies in detecting or otherwise avoiding the consequences of failure.

Moreover, as discussed int he NAVAIR guide to RCM, based on SAE JA1011 and a range of other documents, there is a need to understand the accuracy of the task you have put into place.

For example, will an on-condition maitnenance task be able to detect the warning signs of failure every time? Or every third time? Or depending on the skills of the operator? Or provided we get acces to the equipment... etcetera?

Sometimes, most often, the best methods you have available to you are probabilistic methods that gie you a degree (measurable) of confidence that what you are trying to achieve is correct.

The RCM Scorecard, however, was focussed on measuring the work order information, failure rates and types, potential failure rates, costs of maintenance versus cost of repair, and other indicators to make sure that the RCM strategies that you had put into place were achieving the desired result.

Best regards,

Les, if I may add to this discussion on RCM2 I am like you in we are currently setting up the programme in our factory.
We have established 2 teams - one for manufacturing side, one for packing.
We have so far completed 3 pilot projects and we have done this using the Moubray model (team in a room).
I guess we are fortunate in that we have a consultant getting us off the ground _ something I would highly recommend - why reinvent the wheel - and I have been more than happy with how he has adapted the process to our company.

But I am also training as an RCM facilitator and have some practical and written tests to come - goodluck to me!

Reading through your statements I suspect that you are concerned with making sure you have everything covered. In my short experience so far (I definetly don't have the experience of my learned collegues above) I think you need to be able to apply a bit of gut feel without really truely knowing that you have the asset covered 100%.
I am not sure even RCM will give you that but the way it needs to be viewed is as a start and a working document that will continually be added to through the course of the assets life.

Gee - I know for the assets we've piloted we've uncovered a whole heap of fM's that know one even would have known about if we had not done the homework - and a lot of it was done by debate between 4 guys in a room over 2 days.
Over 100 FMs (which is probably not as many as some of the big machines you can get) however we know that between us working it out (operator included) we have come up with probably 95% of them and have made the changes to PM's and operating procedures where necessary.

And - the operator now knows a whole heap more about the machine!!!!

I don't know about the risk factors and you may have a lot of safety issues involved which would make anyone a lot more cautious, I guess, in making sure they had at least those critical failure modes completely covered.
But I think making sure you at least have a system in place for reviewing your RCMs on a regular basis and updating the database when required makes it a lot more effective.

Hope this ramble helps Les.

Regards, Mike.

PS should add tradesmen know a whole lot more too - they learnt off the operator!!!-- and I learnt off all of them !!!!

Mike,
Thank you for sharing your experience. I am glad to see you are using the team approach with a good external facilitator to get your RCM effort going. I hope you are measuring progress.

Perhaps I'm being picky here, but the examples used were infrared thermography or ultransonics. Neither one of these technologies actually reduce the probability of failure given that nothing else changes on the machine. These technologies are used to "inspect" and determine that a failure might occur in the future. The machine, the maintenance it receives, and its operating context are what they are, and these are the conditions that determine when the machine will fail. I don't think you can increase the reliability of a particular asset or machine unless it's redesigned of the basic care of that machine changes. In this case it would then be proper to ask is it worthwhile to do the change given a decrease in the probability of failure or an extension in life.

Inspection technologies will not decrease the probability of failure but will provide information that we can use to mitigate that failure (schedule maintenance to rebuild, replace, etc). Given you have some past failure history and you can competently inspect the machine, you should have a high probability of predicting any future failure.

Am I off base here?

Hi guys, I'm here again. I was on vacation so I didn't read the posts until today. First of all I want to give thanks to all of you for trying to provide an answer to my question. I'm still doubtful about this subject, I mean if you can measure or not in some quantitative way the reduction on risk caused by a RCM task once you have decided to apply it to a failure mode with safety or environment consequences, to be something, if we could call, orthodox about the rule established by John Moubray in his book. Maybe the best option to determine the effectiveness is qualitative way based on the experienced judgment. Are you agree?

Les,
No I dont. You can measure trends in
- MTBF
- uptime
- safety/environmental incidents
- maintenance costs
Also the number of 'new' hidden failures you find in an RCM study gives you some idea of improvement in Technical Integrity.
This is not to say that qualitative indicators are not important, but that there are hard quantitative ones available.
By the way, this is applicable for measuring ANY reliability improvement effort, be it RCM, RCA or whatever else you choose to use. Of course you only get results by implementing the results of the analysis, not just by doing the analysis. This is where many companies fail.

Les - I don't think you'll find a way to measure the effectiveness of any given method of predictive maintenance "before the fact".
What Vee has oulined are measures we can use "after the fact" and that will show you have a better -more reliable - asset due to your efforts.
As you say it is really experience with those predictive technologies and knowing when and how to apply them that will give you that gut (qualitative) feel.

Mike.

Les - just as a continuation of what I said - in some instances it may be beneficial to apply more than one predictive technology for a failure mode - you may have read some of the other forums which deal with this in more detail - but for example a gearbox can be monitored in both vibration and oil sampling for various failure modes.
This has the benefit of fault confirmation within a system and gives you that extra assurance you would not get with just one technology.
Wrong forum for this but if you are interested go to posts about technologies and techniques for condition monitoring.

Mike.