Thanks Oz.

I was just interested in other's perspectives. I do understand that criticality is not the only means of determining the maintenance strategy. What else I would consider in determining the maintenance strategy is (not a complete list): spare parts availability, lead time, cost, environmental concerns, safety, fire, whether a piece is in a remote location and not normally accessed by personnel, and of course political means. If it is not politically correct to have this piece of equipment running, it definitely could mean that the criticality would go up. But, we don't have politics in our places of work do we?

When you mentioned physics I was intrigued to see what you meant. No problem for the wait though. Don't want this to get in the way for you.

Thanks!

Happy New Year Daryl,

Let me answer your question like this: Look at all the replies I have received. Determining criticality seems to be a very well established activity in reliability. There are a number of methods I can choose from to determine criticality listed above.

Are you saying there is no need for criticality - or that we are simply misdirecting focus when we look for criticality from where it should be placed?

Your clarification is most appreciated.

For those lurking on this post - please jump on board in the spirit of participation - let us know how you deal with criticality at your plant.

Terry O

quote:

Originally posted by Josh:
Vee, any article or sample report on the computer simulation for asset criticality that you mentioned above to get the general idea better?

At the link below, there is no sample report but there is a description of the methodology for quantitative risk ranking of components of the space shuttle
http://foia.msfc.nasa.gov/docs/SAFIE.PDF

We have a computer model of our plant which does a similar analysis to what is shown at that link although not as sophisticated. We also have an expert human panel. The final ranking is a combination of the computer and expert panel rankings.

Happy New Year to you Terry,

I don't think I said there was no need for it, merely that it was misunderstood.

After goading me into writing this at 4am I hope you take the time to read it!!

My question was "why do you want to know criticality". The reason for this is that it is often misapplied, often used to create a very cumbersome approach to a simple issue, and sometimes I have even seen it used to determine frequencies of tasks without refering to the underlying mechanisms of failure that are driving the maintenance activity.

So, why do you want to use criticality? Is it:

• To determine which assets to attack, in which sequence, with some maintenance activity
  
• To provide a "live" indication of which assets are at risk at a certain time
  
• To integrate into your task selection activities to help you to derive the correct tasks for the failure mode
  

I use criticality methods at times, driven by fuzzy logic generally although I do use even more qualitative methods, but always it is horses for courses.

First approach;one of the things that consultants offering criticality analyses do not tell you is that it can only be done at the failure mode level.

This means that if you are using it to determine which assets to analyses, and in which order, then you will have to go through a project of defining all of the functions, functional failures, failure modes and effects of an asset /asset system prior to being able to say whether or not you would continue to analyse the equipment or asset system.

This is 3/4 of an RCM analysis that is required to tell you whether or not you need to do an RCM analysis. Seems a little overdoen doesn't it? Well, not always. There are applications for this which I can go into at other times.

If you wish to find a rapid, prioritization approach, to asset selection then one of the pair-wise type of prioritization methods such as Analytical Hierarchical Process may be more suitable. It is definitely faster, but again there are pros and cons of this approach also.

Second approach;I use this often in my projects after an RCM implementation has been done as a means of management and others to have a common reference point for discussing what are the hot issues of the day and what should be done about them. With todays technology it is not a difficult thing to place information from operational and maitnenance systems on a matrix, against already determined consequences, to give an indication of how assets are performing at present. Again not for everybody, there are ways of getting around this differently, but it does provide a quick method of communication.

Third approach; at its heart all maintenance selection activities do refer to criticality in some form or other. Criticality, as it is most commonly defined, is a reference to probability of failure and the consequences of failure, so it is a proxy for risk.

So, is it necesary to use criticality, in the form of a criticality matrix, to determine maitnenance tasks? Not at all, RCM doesn't for example. And when people try to integrate criticality into RCM it makes it a more cumbersome process I have found.

RBI, ont he other hand, does,, and for what RBI does that seems to be a good place for it also.

However, one of the things that really grates me, and something that has not stood up to scrutiny so far, is when the frequency of maitnenance tasks is changed merely because of where it sits on the criticality matrix.

Taks selection is a complex activity and one that is not as straight forward as "high criticality = more intensive maitnenance". This is a dumbing down of maintenance engineering I would suggest.

For example, and this is a pretty complex area to try to explain in a posting board, a high criticality may be driven by the consequences of the failure mode. Say, the failure mode of a failed bearing of a pump motor. Just because it is high doesnt mean that the frequency of the task should be higher (more frequent) than a similar failure mode at a lower level of criticality.

On condition tasks are driven by physical science, in this case by the P-F Interval. (Which everybody has gone into many times here) So whether it is a high criticality or a low criticality then the frequency of this activity will stay the same.

What could be done is to use criticality to determine whether or not to do an on-condition task. But, hee again there are other considerations. For example, within an RCM analysis the approach to economic failures is to do an analysis on what is the most cost effective task to do taking into account total costs of failure and total costs of maintenance. This is not easily translated into a, say, 5 x 5 matrix for task selection.

In summar; criticality is a good thing generally, if it is properly applied, takes into account some of the other underlying aspects of task selection, and is either the easiest or the most adequate, means of deploying some form of maintenance improvement activity.

So, what do you want to use it for Terry??

Cheers,

Terry,

There is yet another measn of applying criticality that I neglected to put into the previous post. That of criticality of works in progress.

I use this religiously and would suggest it to anybody as a means of managing maitnenqance backlogs efficiently and effectively, reducing the level of emotion often driving the prioritization of works in progress.

Cheers,

James,

Thanks for your patience and I hope that my previous postings puts some of my comments into context in this discussion.

I do know a little bit about RCM turbo and was involved with it prior to its being sold to strategic when it was known as MPDS for BHP Engineering. (Along with just about every maitnenance engineer in AUstralia at the time it seems)

To try to answer your questions directly, and continuing the example of on condition maitnenance:

When looking at stock levels and trying to determine which stocks to hold there is a lot more to it than just criticality of course. For example, with the on condition task the driving issue is not whether it is critical or not (and lets assume that it is relatively important or you wouldn't be trying to prevent the unplanned failure) but the lead time between failure detection and failure or the P-F interval.

If the P-F Interval is enough for you to get the component to site, plan it and replace it, then there is no need to hold it in the stores at all.

But if you have a P-F of two weeks, and a lead time of four weeks, then the answer would be obvious.

Does this clear up my point of view here? Don't misquote me or misundertand me, I think criticality has its place and should be used at times. But it is by no means a panacea for all the plants ills.

Daryl,
You said,

quote:

First approach;one of the things that consultants offering criticality analyses do not tell you is that it can only be done at the failure mode level.

I am afraid I have to disagree with you. I know we have been over this before, so I will not repeat my earlier comments.
Criticality can be applied at system, sub-system, equipment or at failure mode level. (In the offshore Oil & Gas business, (up to 13) Safety Ctitical SYSTEMS are recognized - and accepted by the HSE in the UK. It is NOT necessary to do e.g. an RCM, RBI, or FMECA to identify system criticality.
System criticality helps identify focus, i.e., where should we put out effort and resources. In some or even many cases, critical systems may be obvious. Further, system criticality studies identify the need for redundant equipment. Often we can get rid of very low criticality equipment with little or no consequence to the Plant as a whole.
In most cases criticality is top-down approach, to identify the system, then the sub-system and then the equipment that has the potential to cause business losses (in HSE, $$$ or reputation terms). But we still have to do a sanity check to ensure nothing is missed out, e.g., Relief valves on low criticality system; this does not need RCM type analysis, only a quick P&ID review.
As you point out, physical failure mode level criticality helps identify the maintenance task and its frequency. Invariably, this is the only level where we can identify degradation mechanisms.
A simple RBD or FTA analysis at a high level can identify system level criticality, with perhaps an Excel spreadsheet to help. Working at a more detailed level invariably needs mathematical modeling tools. if these are to be realistic, they need Monte Carlo or Markov type simulators.
I am sure you know all this, but some other readers may be less well informed.
In sum, I agree entirely with terry's asking the question.

Vee,

What are the up to 13 safety critical systems which are accepted by the HSE UK? What is proper tittle of the HSE UK that you referred to above? Is it the UK Occupational Safety and Health Act?

I agree that we can assign criticality from systems up to parts but not at platform level for oil & gas industry.

There are some PM works which are assigned the platform tags eg SDS/ESD instrument input checks, vibration survey, valve greasing, lubrication, etc which are across numberous systems.

We assigned all PM works the same work priority three (3) and prioritize them based on asset criticality.

Because the platform is not assigned any criticality, those PM works with the platform tags as their locations will not have any criticality value. How to solve this problem?

We use the asset criticality to indicate safety critical PM works because safety critical systems and equipment will have a higher criticality which ideally need 100% PM compliance.

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

Criticality analysis of running unit is not that critical than to do at design stage. Today's designs are very complex hardly any teams engaged to work on RCM at design phase of any project.
The significant number of criticality analysis done on very complicated applications have not resulted in any dominant findings that have suggested any great achievements to the legacy models that are followed from ages. By this experience it seems that RCM is at the beginning of its civilization and needs time to reach to its perfection. Don't you believe so?

quote:

Originally posted by Vee:
Daryl,
You said,

quote:

First approach;one of the things that consultants offering criticality analyses do not tell you is that it can only be done at the failure mode level.

I am afraid I have to disagree with you. I know we have been over this before, so I will not repeat my earlier comments.
Criticality can be applied at system, sub-system, equipment or at failure mode level. (In the offshore Oil & Gas business, (up to 13) Safety Ctitical SYSTEMS are recognized - and accepted by the HSE in the UK. It is NOT necessary to do e.g. an RCM, RBI, or FMECA to identify system criticality.
System criticality helps identify focus, i.e., where should we put out effort and resources. In some or even many cases, critical systems may be obvious. Further, system criticality studies identify the need for redundant equipment. Often we can get rid of very low criticality equipment with little or no consequence to the Plant as a whole.
In most cases criticality is top-down approach, to identify the system, then the sub-system and then the equipment that has the potential to cause business losses (in HSE, $$$ or reputation terms). But we still have to do a sanity check to ensure nothing is missed out, e.g., Relief valves on low criticality system; this does not need RCM type analysis, only a quick P&ID review.
As you point out, physical failure mode level criticality helps identify the maintenance task and its frequency. Invariably, this is the only level where we can identify degradation mechanisms.
A simple RBD or FTA analysis at a high level can identify system level criticality, with perhaps an Excel spreadsheet to help. Working at a more detailed level invariably needs mathematical modeling tools. if these are to be realistic, they need Monte Carlo or Markov type simulators.
I am sure you know all this, but some other readers may be less well informed.
In sum, I agree entirely with terry's asking the question.

Azy,
With regard to your post,

quote:

Criticality analysis of running unit is not that critical than to do at design stage. Today's designs are very complex hardly any teams engaged to work on RCM at design phase of any project.

This has not been my experience. I was personally involved in mathematical modeling of several projects, and in doing RCM work at the design stage. In fact, a few projects had all the maintenance tasks, procedures, drawings, spares, tools lists etc. in the CMMS before the Plant was commissioned. This is the best way to proceed, in my view, if we are to be top performers.
In several projects, significant capital cost savings were made by doing criticality analysis at an early stage of design.

Vee,

For some reason you are not understanding what I am saying here, so let me clarify again. Please read through what I am writing here and internalise it prior to returning to your corner and stating the same things again.

As a side note I was actually against criticality as a means of determining frequency without reference to the underlying failure mode. (See examples above)

Also, like you I also have done this once or twice. In fact I just finished a pretty large project in Asia focussing on this as one of the elements.

Terry, what do you want to apply criticality for? There was never any inference that he shouldn't have posed the original question, merely a repeated request for clarification.

The reason why I have asked this several times is because, as I have pointed out, there are many different ways of applying either criticality or other prioritization process.

First; if we are talking about criticality then lets clear the air and agree exactly what we are talking about.

Criticality as it is commonly used, is a means of representing probability of failure and consequence of that failure, agreed?

For me this is represented either by some for of quantified number against the failure mode, or one of the many criticality matrices around the place, or even by some form of none numerical grade. But they all represent probability and consequence.

And for us to truly understand the consequences and probability of failure then we need to understand not only what an asset does, but also how it can fail, agreed again?

This is a fundamental point to agree, because if we are not doing it this way then what are we doing? We are prioritizing - establishing a relative importance based on criteria other than those that make up criticality. (Probability and consequence)

So regardless of whether you want to apply this at a system, sub-system, platform or remote utility level, there is always a need to base it around the failure modes, how things fail, what happens when they fail etcetera. So criticality has to be bottom up. (If we are talking about consequence and probability of failure)

Second; I know from past experience that creating the difference between criticality and prioritization seems to fire you up. Regardless of this it is a valid point and one that is supported by systems, methods and processes in use within asset management as well as further-a-field.

I put it to you that anything coming from the top down, for sequencing of effort, and for determining where best to apply effort is not criticality as I have defined it earlier, but prioritization. (Thus removing the confusion from the over used term "criticality" and aligning us with other methods that are widely used.)
So, if we are coming in from the top down we cannot, by definition of coming in from the top down, have a definite idea of the probability of failure. We may understand how frequently an asset type or system breaks down based on general information, but this is a far cry from what we are speaking about as you will relate to.

Also, the further we get from the failure mode level my personal experience is that decisions start to be made on what the asset does and its operational importance, rather than the effects of how it can fail. (Another underlying element of criticality) So we are not only having problems with the accuracy of probability of failure information, but also with consequence information.

So, if we are not using information derived at the failure mode level, then the further away we get from that level, the less accurate the criticality analysis is. I suggest to you that it is not criticality; regardless of what you call it, it is not based on probability and consequence but on issues of economics and whatever are the corporate hot issues at the time.

This is prioritization where we are coming in from the top and selecting assets, defining systems or whatever based on the corporate drivers at the time. I have referred people previously to a document prepared by a company called Expert Choice. I have it on my blog here: http://physical-assets.blogspot.com/2006/04/prioritization-or-criticality.html

I am not associated with these guys and don't know any of them either, but I have used their systems previously. Pair wise systems such as these are gaining in prominence as a means of doing exactly what you are describing, but doing it rapidly, with some degree of sensitivity analysis, and tying it solidly into corporate objectives at the time.

This sort of prioritization analysis is about to be used to determine several billion pounds worth of spending in two of the UK's regulated industries, has been used already within the UK rail industry for prioritizing maintenance initiatives, and is regularly used for selecting systems for data collection by several petro-chem companies that I am familiar with. So it has a pretty strong history in my knowledge, outside of my own areas I believe it is also used widely within the nuclear industry.

So; my point was that often consultants don't let on that it needs to be done at the failure mode level. This stands as my personal observation. Hopefully now it is clearer where I am coming from and why criticality (probability x consequence) is something for the failure mode level only.

Importantly; I am not against criticality, it has its place and I have used it in safety sensitive industries (munitions), in cash strapped companies where billions are needed (electricity), and where there could be a need to defend a decision to proceed a certain way. (Rail)

Cheers,

Oz:

Thank you for your concerted effort in explaining your understanding of Criticality/Priority. As you could gather from my posts, I am not a consultant. I would love to be a consultant! Then I could concentrate on the knowledge of the craft as opposed to just applying it at a plant level for one industry. You look at it at many levels and many plants.

With all the questions and many answers here, we may all be on the same page, or aka. headed in the right direction, but how we get there may be different. None of that makes our/your perspectives wrong, or inappropriate. I have noticed that whatever people's experiences are that is where they tend to lean with their posts. I try to lean on the craft itself (Reliability/RCM/Root Cause, etc), and where it can be applied specifically at a plant maintenance, hourly craft level.

I must say that your posts=(book) is a very good explanation of the understanding of your craft. I will have to think about it for a while before I can truly understand what you've said and give an appropriate and meaningful reply.

Nontheless, I do appreciate you answering the questions that have been posted here. I know we all have experience that we can help others with. That's where I come in. I like to learn, and apply. If it cannot be applied here where I work, then that doesn't make it wrong, just not applicable here!

Thanks again.

James,

Thank you very much. Fortunately, like most others on here, I truly enjoy what we have all chosen as a career.

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

Azy,

With regard to what you posted previously:

quote:

Criticality analysis of running unit is not that critical than to do at design stage. Today's designs are very complex hardly any teams engaged to work on RCM at design phase of any project.

RCM and criticality are not necesarily the same thing at all. In the RCM standard (SAE JA1011) there is no mention of criticality as I recall it.

Vee

Did you see my clarifications about the 13 safety critical systems above?

Josh,
Here is a partial list:
- Structural integrity systems
- Ignition Control systems
- Process containment systems
- Fire, smoke, gas detection systems
- Fire protection systems
- Shutdown systems
- Blowdown and relief systems
- Emergency response systems
- Life saving systems
- Heating, Ventilation and Air Conditioning systems
- Communication systems
- Blow-out prevention systems

-

Josh,
HSE is Health & Safety Executive, the uK Regulator.

Is the 13th one being the potable water system for drinking etc? My suspect is because somebody somehow assigned criticality 1 for potable water pumps.

Is the list published in the UK Safety and Health Act or written in a circular or guideline? I would really like to get hold of this list. Do you have a copy or will it be included in your 100 years Maintenance book (he...he...he)?

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

No wonder at one time when the telecom system was down, all hot works were suspended until full recovery.

I didn't realize that HVAC system is considered safety critical.

The rest is quite obvious.

Josh,

HVAC systems can communicate smoke from one part of the living quarters to others, including the control room and radio room, magnifying consequences dramatically. That is why we have fire-dampers to isolate sections. These are critical items to maintain,. and must be tested periodically.