Are there any simple methods for determining plant equipment criticality?

Some systems go by the experience of stakeholders, others use Reliability Software, others use an RCM type process. What do you use?

Details please.

Thanks.

Terry O

We rate all pieces of equipment on a scale of 1 to 5 for Criticality, with 5 being the most Critical and 1 having no Criticality. The scale is as follows -
5 Critical, No Redundancy
4 Critical with Redundancy
3 Not Critical, No Redundancy
2 Not Critical with Redundancy
1 No Criticality rating.

Therefore a piece of equipment that is critical to the operation and has no onsite replacement would be given a criticality of 5, one that is critical to the operation and has a replacement onsite would be given a criticality of 4, etc.

Alan Daykin
Maintenance Planner
Wastech Services

We use RCM-Turbo for individual rating/ranking and to develop our maintenance strategy for the piece of equipment.

Now, with many thousands of pieces of equipment, this is impractical on a mill-wide basis. We must reduce the number of RCM's we perform, since it is not practical to use the RCM process on every piece of equipment. Now we must (very generally) rank the equipment, based on a much quicker process to get the ball rolling.

RCM-Turbo rates each piece of equipment in groups of 4. 1-4 is Critical, 5-8 is medium criticality, 9-12 is low criticality and 13-16 is not critical.

Currently as our program develops, we are leaning more on the experience of the stakeholders. So, it will take some time to shift the thought paradigm.

A simple method or quicker method would be using the stakeholders for their own area, not for mill-wide criticality. The finishing area of Production would not be good at rating the raw-input side of the process, but they are great for their own area. Combining each area's expertise is the job of the Reliability Engineer with their assistance to gain a full understanding of the equipment's criticality. All documented, of course.

This would then be excellent as a starting point. From here, verify the stakeholder's experience by performing RCM analysis on the highest critical equipment. Work your way down to the level of criticality your business demands, and move on from there.

Very simple, but anyone can do that. You also don't need software. It is what some companies have been doing for years, before they every heard of Reliability. They knew what to do, when to do it, why they do it, how to do it, and where. Of course in most places it wasn't a documented process like the Reliability Based Maintenance philosophy, but it worked for some.

Hello,

this approach of setting the criticality in such a way (1-5, 1-9, ...) is a first approach.
But it does not take care of all the inflencing factors:
- production criticality
- environmental criticality
- save & healthy

You should consider all these points, rate them (based on a predefined scoring) and define the calculation of the (final) criticality. This allows you to calculate in a very simple the criticality of new equipment or redefine the criticality based on changed requirements.

A simple way would be:
- score each of the topics above from 0 (not critical) to 3
- add the scores of each topic (this give you a number from 0..9
- define the criticality from 0 (not critical) to 9

Normally the most discussed issue is the definition of the calculation.

A simple and quick one may be something like this:

Criticality A - Failure of equipment compromises HSE (Most/HSE critical)

Criticality B - Failure of equipment do not compromise HSE but affects production (Production critical)

Criticality C - Failure of equipment has no effects on HSE and/or production but cost of maintenence is above USD10K (Maintenance critical)

Criticality D - Failure of equipment has no effects on HSE and/or production but cost of maintenence is below USD10K (mostly run to failure)

If want to be more complicated, each of criticality above can further have 2 subdivisions such as N to indicate no redundancy and Y to indicate having redundancy.

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

We used this when I was still employed and implementing TPM's Planned Maintenance 4 Phases To Zero Unplanned Breakdown. Here's the summary

Rank A : Worst Machine - Highest Priority
Rank B : 2nd Worst Machine - 2nd Priority
Rank C : Good Machine - Least Priority

The step involved is to conduct a machine inventory in your plant for operations and facilities and then to conduct a machine ranking for each equipment, Once you have completed your machine ranking for all equipment's then you have a clear data on how many machines categorized as Rank A, Rank B and Rank C.

Our Planned Maintenance Efforts was geared on focusing and improving all rank A machines.

More of this in my materials on TPM-Planned Maintenance 4 Phases TO Zero Unplanned Breakdown
on the sales portion, we have great success on this, since we have a dramatic reduction of unplanned failures from 888 as of Jan 2001 and down to only 14 as of Sept 2001 on all 472 equipments we categorized as Rank A

Kindly refer to the attachment, I hope this helps you

Machine_Ranking_Matrix.xls (30 KB, 131 downloads)

Terry,

If we treat a Production Plant as a Black Box, where we feed in the raw materials and other inputs and get as an output Dollar Notes, then we can define criticality in a simple way.

Let us say that within the Black Box there are 100 equipment items, some connected in series, some in parrallel as installed spares, some with recycle lopps etc., and we number them from 1 to 100. Let us say the output, which may be cartons of cola, boxes of soap, gallons of gasoline, strips of medicine tablets or similar, can simply be equated to a daily value of $100k.

Let us now see what happens if we are able to raise the output of each equipment item in turn, one at a time, by 1%. What, if anything happens to the output? as an example. let us say that, for

a. Item 1 output becomes $103k
b. Items 3-20, no change in output
c. Items 21, 22, 23, output in each case is $ 101.5k
d. Items 24-50, no change in output
e. Item 51, output is $ 105k
f. Item 52-75, no change in output
g. Item 76, output is $102k
h. Item 77-99, no change in output
i. Item 100, output is $102.5k

From this we see that items 24-50, 52-75, 77-99 are not critical. In order of criticality are
items 51, 1, 100, 76 and items 21-23.

We can check this be seeing what happens when we REDUCE output of each of the items in turn, by 1%. The results should be the same.

What about safety and environmental consequences? In principle, large safety & environmental consequences affect production, some more drastically than others. For example, in an offshore installation, if the fire pumps dont work, we have to stop production and de-man the Platform. If there is cross contamination of medicine batches due to equipment failure, the line will have to be stopped. A fataliity or explosion will invariably stop production, even if the Plant is able to continue running. So, for the purpose of this exercise, all failures can be treated as $ losses.

Now to address the 1% changes I asked you to make earlier. This can come from 4 sources;
1. Nameplate capacity, i.e. what is the item capable of producing?
2. Its reliability, i.e how often does it fail (measured by e.g. MTBF, Weibull parameters etc)
3. Its maintainability, i.e how long does it take to fix the failure (measured e.g., by MTTR)
4. Planned downtime for scheduled shutdowns, setup time etc.

Obviously, it takes some effort and knowledge to arrive at the sensitivity figures i qouted earlier. We can do this by using mathematical modeling using Reliability Block Diagrams, Fault Tree analysis etc.

As general points,
- redundancy reduces criticality
- buffer storage reduces criticality
- low MTTRs reduce criticality
- high MTBFs reduce criticality

As far as shutdowns are concerned, keep them short and infrequent, else they become very critical.
I dont know if the above is confusing, but it usually takes a day in a class room to explain this clearly.

Terry,

Why would you want to develop criticality in the first place?

Depending on what it is that you are trying to do there are a range of approaches. Some of these are real criticality approaches and others are more along the lines of prioritization approaches.

(Some foolishly try to use criticality to determine maintenancefrequencies, flying in the face of physics generally)

Criticality is one of the most misunderstood and misused areas of reliability, and I think this thread is a good indication of why.

So... why do you want to generate a criticality matrix in the first place?

Ozgipsy wrote:

quote:

Depending on what it is that you are trying to do there are a range of approaches. Some of these are real criticality approaches and others are more along the lines of prioritization approaches.

(Some foolishly try to use criticality to determine maintenancefrequencies, flying in the face of physics generally)

So, with your perspective, let me probe your understanding a little bit. What do you mean by "real criticality approaches" if physics is used as you mentioned. I'm just trying to understand your philosophy or reason behind your approach.

I do believe from my understanding (emphasize my understanding - it doesn't mean it is the only way!) that criticality is essential to produce maintenance frequencies and an overall equipment strategy. Otherwise, what else would you use? This is where I need to understand your perspective to have the correct view. Please help me!

I also agree with you regarding prioritization being that is what most people think criticality is from outside of an RCM understanding or from a production standpoint. Criticality and prioritization are radically different.

Also, you mentioned:
Ozgipsy wrote:

quote:

Criticality is one of the most misunderstood and misused areas of reliability, and I think this thread is a good indication of why.

What is the reason that this thread demonstrates the misunderstandings of criticality? Again, I am really curious as to why you mention this. I may agree with you, but I am probing your thoughts.

I look forward to your reply.

Hello James F.,

you are roight, that criticality is one criterion to determine maintenance strategies. But it is not the only criterion for it.

And the criticality gives you some more handles: It serves as a base for the priority (which is not the criticality) when a fault occurs.
If you calculate the criticality also based on production information, you even get more:
By the (dynamic) calculation of the criticality with regard to the actual production it helps in identifying the important equipment.

Vee, clarifications:

Can the 1 % increase in eqpt output be done in real plant operations? I heard about plant optimization in this manner which will show catalyst limitations but not quite eqpt criticality. Is this approach widely practiced elsewhere?

I can determine redundancy & buffer storage easily but MTTR & MTBF are not readily obtainable even with CMMS if data analysis is not done. Can we use simple averages of times to repair and time between repairs as indicators?

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

Josh,
First the 1% change in equipment productivity is a method to check how sensitive the final plant output is this small change. The more the plant output changes wrt any one equipment item, the more critical that item becomes. Can you follow this logic?
Second, these sensitivity studies are done in a mathematical model, because it is not practical to tinker with one item at a time in a real plant. However the model, if accurate, should mimic reality.
Third, in reality you can improve equipment performance by one or all ofthese ways:
- improve its reliability: use whatever measure you can to monitor it. e.g MTBF, failures/month etc.
- improve maintainability; the faster you rectify a fault and get the equipme nt back on line, the better. Measure it with any means at your disposal e.g., MTTR
- increase its capacity by redesign, e.g., larger impellor for pump, higher speed for packing machine etc.
- reduce setup time (for manufacturing processes using e.g machine tools, tabletting machines etc)
- increase interval between shutdowns/overhauls
- reduce shutdown durations

In a process using catalysts, there is a clear degradation pattern, with age-related loss of performance. This is a natural process which we cant do much about. We can however compensate for such losses by working on the other factors listed above. Catalyst technology is changing quite rapidly, and you can now get better longer-lasting catalysts for many applications. You may have to pay more for the catalyst, but work out the life cycle (or at least shutdown cycle) costs first. Nothing stops you from working to improve reliability, maintability etc. except resources. If you model your plant and do sensitivity studies, you will know which equipment or sub-system is critical and you can focus your attention on those few items.
As to MTBF, MTTR calcs., you have to read my next book "100 Years of Maintenance" to be published later this year. I cant explain it easily on forum pages, sorry.
-

Ok Vee. Since you mentioned computer simulation, I understand it can be done. I guess the process licensor has this kind of computer model.

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

Josh,
Nope, that is very unlikely. The model needs your inputs on
- mode of operation
- loading density
- reliability info
- maintainability info
- spares availability and logistics
- actual capacity, not nominal nameplate capacity
- consequences of failure
- etc.
These are only known to you, the operator, not the designer

I am afraid the buck stops with you. You can get Consultants to guide you. The software packages run reliability models, not process models. MAROS, MIRIAM, WITNESS, SPAR, SPARC etc. are popular packages. They cost quite a lot to licence and building/running the model can be quite excpensive too. It is not something you should embark upon casually.

In typical petrochem plant : say 10% equipment will be critical (fire water pumps, ESD systems, most of pressure relief valves, single line equipment etc.. All spared operating pums as "C", spare pumps are either B or C and remaining stuff as D. Also definition of criticality varies from organisation to organisation based on risk bearing capacity, local environment safety and Health rules and reputation of the company..

It sounds like there are a lot of schemes out there.

One important aspect that I would highlight for a large facility - once the work is done to classify components, it should be captured in a widely-available (and living) database.

The reason is that almost everyone in the plant needs this info for one reason or another. Vib guy needs to help select his monitoring frequency. Work scheduler needs it to prioritize his work. Selection of PM and overhaul frequencies etc. For everyone individually to try to learn the systems and develope the ranking would be grossly inefficient.

At our plant, the risk-ranking process it was a very involved process involving computer simulations combined with expert panel input. I was not heavily involved in that process, but I have the results available to help me every day in deciding what levels of resources are appropriate for various pieces of requipment.

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

Electricpete, I think it's a simple/quick approach vs a complicated/vigorous one.

It's like for developing maintenance strategy, do we want a simple & quick ones based on generic & vendor strategies or wait for a complicated & rigorous RCM, RBI, IPF etc to complete?

Of course, it can evolve from the simple to the complicated one.

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

Josh - I guess there are a variety of circumstances. In my circumstance, determining equipment criticality is not a quick simple process.

I work at a nuke plant. We have a multitude of interrelated systems (more complex than Navy plants). In the old days (before we had a formal ranking available), when addressing an equipment issue, I would have to spend some time figuring out how the component fits into the system and what are the consequences of possible failures. While that is very valuable stuff to know, it is time-consuming and not productive for everyone (me) to be an expert on everything (effect of component failures on systems). Having the the risk ranking done by others (expert panels) and available to me in our plant database has been invaluable for me.

Josh,

quote:

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

I dont think you will find these in the public domain, but try looking at the websites of these software products.
There is a chapter on this subject in my forthcoming book "100 Years of Maintenance"

Dear James,

My apologies for my silence. I am right under the hammer at present unfortunately, but I do want to continue this discussion and thanks for engaging in it.

I will try to sort out a few clarifications over the next few days. But I was hoping for an answer to the question I posed to Terry...

Why do you want to use criticality Terry?

Cheers,