Would 99.9% reliability rating be viewed as desirable? If 99.9% were achieved, would 95% also be acceptable, if you knew 99.9% were achievable.

Management should review the tasks and procedures performed if 99.9% reliability were achieved. The review would be to evaluate and adjust tasks and manpower utilization that provided the 99.9% rating. But by limiting those tasks and manpower utilization, would a lesser reliability rating of 95% be good enough for the business?

Personally, tasks and manpower utilization should be routinely evaluated and depending on results, adjusted. But if 99.9% reliability were achieved, should we accept something less?

Any thoughts?

This message has been edited. Last edited by: Don Jones,

Don,
Everything comes at a cost. If the business needs 99.9%. e.g., if you are operating an airline, then yes, we should.
This cost rises exponentially. If it costs $X to achive 90% reliability, it can cost $10X to get to 99% and $100X to reach 99.9%.
If it was your money, wjat would you do?

Vee,
I understand the money side of the equation. Money is always considered but the process that would be affected has to be considered too. As I said the evaluation of tasks and manpower would be adjusted but services that became proficient at maintaining the process and the potential cost of a failure should be considered. I guess fear of Murphy’s Law sways my thoughts when viewing reliability concerns.

I suppose a person must trust management decisions in this area. It just seems that when these services are adjusted and failures occur (+- 95% reliability) the reliability team feels the pressure from those decisions and adjustments.

Concern for a way of thinking that is historic comes to my mind? (Everything is operating so good what do the reliability guys have to do or are they doing anything. Do we need them anymore?) How do you defend against that way of thinking when 99.9% was achieved?

Thanks,
Don

If you achieve 99.9% then you are probably doing on over-kill. Or, have achieved and need to cut cost by extending monitoring intervals.

Slow turning stuff typically can go to quarterly monitoring and those that do really slow should be doing quarterly or cut to quarterly. It's hard to extend anything beyond that.

1200 RPM fans are a canidate for quarterly monitoring. If your program is refined and MTBF has been optimized, then your PM program is doing its job and there's not as much need for monthly monitoring. And consideration can be for upgrading to full-time on those machine you always wanted full-time monitoring on - now you can justify that by a manpower cut. Not what you wanted to hear?

Reliability equal Availability? No! I think so. Why Don used Reliability indices for ....?

Vee, All the time of cost rise up when increase reliability?

Panuphan,
The question of 99% reliability success came about from a conversation. The idea of achieving a high percent of equipment reliability could mean reduction in staff, maintenance or reliability services. I have heard about this being done at other companies before, after all it is a business.

If maintenance and reliability activities are achieving 99% reliability just how much do you cut and choosing what to cut to reach 95%. Management wants to reach a balance of cost to activity but I think that can be an expensive decision and should be done very carefully as I'm sure it would.

This message has been edited. Last edited by: Don Jones,

The problem is the lag. It is not a condition where you cut your PdM manpower or reduce PM tasks and the reliability drops. Your costs will drop right away when you do this (Short term gain)and reliability by stay at 99% for a while. Maybe even a year or two. By the time you see a drop in reliability it may be a more significant drop from what you wanted.

To get it back may take a lot more effort and money than what you saved those few years.

quote:

Originally posted by Ron Stiemsma:
The problem is the lag. It is not a condition where you cut your PdM manpower or reduce PM tasks and the reliability drops. Your costs will drop right away when you do this (Short term gain)and reliability by stay at 99% for a while. Maybe even a year or two. By the time you see a drop in reliability it may be a more significant drop from what you wanted.

To get it back may take a lot more effort and money than what you saved those few years.

Well said Ron. The ebb and flow of mature PdM Programs. Pay me now or pay me later.

Panuphan,

quote:

Vee, All the time of cost rise up when increase reliability?

On the contrary. Increasing reliability does not mean more work. High reliability means few failures, so less breakdown work, and lower downtime.
The nature of the Availability-maintenance relationships can be summeriszd as follows:
1. At Availability below 85% and upto about 95%, improvements require more PM/PdM, more coists
2. Above 95%, the cost rises are steep and geometrical. Also, there is a practical limit to Availability. Doing more PM means more downtime too. Not to forget that doing maintenance adds 33% more maintenance as well.

Availability which is the only parameter that allows us to work safely and make $$$ increases directly with Reliability and Maintainability. We use the term Reliability as if we mean availability (this is onlky true for non-repairable items like light bulbs).

Increasing reliability can be done in two ways; eliminate/reduce failures using RCA, and doing the right maintenance work at the right time to the right quality, using RCM, RBI, IPF to help. Doing RCA or RCM costs money and resources. As we get better and better, the law of diminishing returns applies.

Vee - where did you get these figures?

quote:

The nature of the Availability-maintenance relationships can be summeriszd as follows:
1. At Availability below 85% and upto about 95%, improvements require more PM/PdM, more coists
2. Above 95%, the cost rises are steep and geometrical. Also, there is a practical limit to Availability. Doing more PM means more downtime too. Not to forget that doing maintenance adds 33% more maintenance as well.

My analysis of equipment and failure rates tell me that all equipment has inherent reliability / availability which is determined by the design and the operating context. There is a level of maintenance that can be done to achieve the inherent levels. Doing more does not increase performance from that point. Now you may reduce the dowtime of equipment by keeping more stores - adding expense in this area increases avaialability for random and sudden failures.
However, in general a discussion on where this so called optima or point of least costs is irrational from where I see things. Maintenance is essentially a fixed cost - not a variable. Doing more of it does not increase reliability. Doing less can cost more overall and result in more breakdown...But once you achieve the inherent levels than that is it.
If you get your mathematics hat on, you can prove the least cost interval of condition based maintennce is always the PF Interval(reliable inspection method and immediate repair)itself in the case where the cost of inspection is greater than the cost of failure, there is no point at any interval that changes that result. Fixed time maintence does vary however we rarely have the numbers to determine the curve....
Regards
Steve

Steve,
I think we are in violent agreement.
I did not talk about inherent reliability to minimise confusion, but I agree with you.
I am not a believer in the maths you quote, i.e of cost optimisation. o me the only thing that matters is risk management. When we do that properly, costs fall, as a consequence,

Vee,
I presume you own a car. It has a maintenance program specified by the manufacturer. Now it is not too bad. It is mostly condition based. If you analysed the program for your own operating context - the way you drive and the conditions, then you might reduce or increase the level of maintenance on your car - you should use the concepts of RCM to do this. You arrive at a maintenance program developed for your own operating context. Most of the program will be run to fail, next will be condition based and some will be fixed time replacement. The latter being the least.
Lets say you do this analysis and you change a few things from the recommended schedules based on your PF intervals. You probably don’t have the data to change the fixed time replacement items.
Once you have done that evaluation, you would not expect to reduce the failure rate by doing more inspections. And if you changed out the parts more frequently ,then in some cases (head gasket, valves etc) you would make things worse. Now if you live in an area where corrosion is a big problem, I would agree that the more you wash your car, the longer it will last - and thus my statement that fixed time maintenance does vary. But you would not increase the reliability of your windscreens or eliminate punctures by doing more inspections. Nor would changing them out on a fixed cycle have any impact. If you inspected the wear on your tyres every day compared to monthly – would your tyre last longer.
I know there are a lot of text books out there that contain curves of about costs and how reliability increases the more you spend on maintenance and there is some “sweet spot” - similar to the argument you are proposing - but I think these were created before Nowlan and Heap and the people had the old traditional thinking… at least this is the only reason I can see for the logic. Most other things have minima like stores and inventory order quantities. However maintenance is not a continuous function in that respect. Mathematically, the economics of condition monitoring change after PF is reached.
As you can see, these curves don’t make sense to me.
Also – still interested in your numbers?

quote:

The nature of the Availability-maintenance relationships can be summarized as follows:
1. At Availability below 85% and up to about 95%, improvements require more PM/PdM, more costs
2. Above 95%, the cost rises are steep and geometrical. Also, there is a practical limit to Availability. Doing more PM means more downtime too. Not to forget that doing maintenance adds 33% more maintenance as well.

Rgds
Steve

Another way of improving reliability of an item would be improving the design.

Although most of my career in thermal power generation was in "maintenance" areas, the work included investigation of plant performance shortcomings and design of a solution. Today I guess this work would be called "proactive" or "reliability" engineering. Not all work called "maintenance" work was the janitorial aspects!

Merry christmas and Happy New Year 2009 to everybody!!!

I think % reliability success has no quantum measurement figure. Everybody has different interpretation on the reliability success figure.

Equipment will fail eventually no matter how good the PdM maintenance apply due to obsolete parts for replacement.

Sweet....

If 99.9% reliability is a goal... then why wouldn't 100% reliability be your goal ??? As stated in some posts, it is all about how you measure it as well. Are you equating that to 99.9% productivity or equipment uptime (run without any failures)??? Someone used a car as an example, most all will need some types of "repairs" during their lifetime. Do these repairs count against the cars "reliability". Not really, unless it brakes down on you on the freeway and or these brake downs are frequent enough. Rutine Maintenance, like checking/changing the oil and tires etc. and periodic inspections will keep the car running good for a longer period of time. The point is to address issues before your car (or equipment)brakes down, and to schedule to have it serviced when it needs it. That being said it is not prudent to take your car to the dealer every week to have it inspected. In order to achive 100% reliability in a car (for example) you would almost have to have another one in stand by ready for you to jump into. Now ask yourself, can you or your company afford that? In most cases NO, it is not financally practical. So 95-99% is a reasonable goal. Keep critical and hard to find parts on hand, perform rutine inspections, have a staff of highly qualified technicians, react quickly to concerns and make repairs correctly.

Steve,
Sorry about the late response. I have had limited access to the internet as I am away from Scotland.
My numbers come from some 35 years in Industry and 6 years as a Consultant. They may differ from yours, but hey, that is what I have learnt. I happen to differ with you on topics such as Weibull or Maths in Reliability etc., but I think we can disagree as we have different experiences.
I am a firm believer that our job is to manage risk. In my view, any optimisation is aimed at reaching the desired level of risk, never at a requird level of cost. It so happens that managing risk properly does reduce costs and improve uptime, so we will be winning anyway. But going after costs as a prime objective can lead to high risk scenarios, e.g. Longford or Bhopal.
Your primer on RCM (car maintenance) is instructive, and supports this view. So I am not sure why you think I (specifically as your note is addressed to me), need such instruction. Nevertheless, your point is well made, and I believe we are on he same page.

I wish you and all the other readers a very happy New Year.

Hi Vee,
First - the car example was presented for the interest of the forum members, but it is also used to try to explain the proposition that achieving high levels of reliability dont mean adding cost which is what your experience is and the statement about geometrical cost increases to acheive above 95% availability.
There is a good case study from Dupont (bit old now) but as they increased reliability / availability costs reduced considerably. They did not increase geometrically. If the organisation is out of control then spending money may help reliability, but otherwise, I dont think that after a certain spend, the additional money helps at all.
This might be my last post for the year - it is new years eve here in Oz.
Talk again next year.
Rgds
Steve

99.9% Reliability? Two words, GET REAL. Unless the operation is small and equipment/parts are cheap no CEO/Manager in his right mind is going to shoot for 99.9% (IMO) If a CEO/Manager sets a goal this high... I'm at a total loss for explanation. My department goal for downtime was at 2.0% back in 1991 and for 16 months it was between 1.3 and 2.2% but, But the equipment was New (a cutting, packaging operation for PS Plates). This was a 24/7 operation with TWO maintenance Techs (one Mechanical and one Electrical but both were duel skilled somewhat). After several equipment upgrades and additions there's 7 Technicians and 2 Engineers that maintain the same equipment with 12% Downtime. (OUCH)

Unless you're dealing with Human Lives forget 99.9% the cost to maintain would be astronomical!!!!!

(Management wants to reach a balance of cost to activity but I think that can be an expensive decision and should be done very carefully as I'm sure it would.)

It was never said that it did not deal with human life and for conversation sake what if the cost of some equipment having unscheduled downtime (failure during critical times) would be astronomical and disrupt years of progress. Then 98% may not be good enough when the cost to have 99.9% reliability is 1/8th the cost of that unscheduled downtime.

All maintenance is reviewed and cuts are made, but sometimes the cuts are more costly than keeping up the reliability.

Don,
Availability equates to potential safety and profitability. Reliability is one parameter affecting Availability, but so is Maintainability.
People often say they want 99% Reliability when they mean 99% Availability.
An item in an As Good As New (AGAN) condition is (at best) 100% reliable at the instant we install it. One millisecond of use later it is no longer 100%. The figure keeps dropping, the droop depending on time in service and failure rate. It will only get back to 100% when we restore to AGAN through a maintenance effort.
100% Reliability mean NO failures EVER; welcome to Utopia!