Hi gents

Why we do maintenance?


Is it to prevent failures or failures consequences !!! ?

It is, of course, about managing consequences.

However, there is a bit more to it obviously. Moubrays last article, published shortly after his passing, was called Times Arrow and was aimed at creating a link to a scientific basis for maintenace based on the second law of thermodynamics.

Was a pretty good piece I thought, I am not sure where you can find a copy of it but I imagine Terry has it somewhere on his site.

In the context of business, we do maintenance so the business can make money.

In the context of most everything - we also do maintenance to fulfill the mission.

Terry O

D - I do not have Time's Arrow posted anywhere but I did a fair bit of research on it once he wrote it. It lead down some very interesting roads - and there were predecessors that lead JM to that piece.

I will see if I can dig up some of the documents and post or link to the source as I would enjoy a discussion on them.

Man - you have a great memory.

Terry O

terry,

I would be interested in the precedents you refer to.

I know that the term "times arrow" is not a Moubray-ism at all. It is a standard term for referring to the second law of thermodynamics.

I also know that if anybody thought about it, this relationship is obvious, however John was the only one that I am aware of to illustrate a direct connection to it.

I may be wrong in this, I often am, but thats how I saw it.

This may be of interest to anybody looking into this issue at all:

http://www.secondlaw.com/two.html

I think the purpose of maintenance is to safeguard technical integrity while the purpose of design is to provide technical integrity.

Why?
To keep the business assets running at the best operational level and at the best economical solution we can achieve.
(I said economical, not necessarily lower cost).

At our plant the focus is still in Preventive Maintenance (PM). We define PM as inspections to detect conditions (failures) that might cause breakdowns, production stoppages, or detrimental loss of function combined with specific actions to eliminate, control, or reverse such conditions in the early stages.

PM consists of two basic activities: (1) periodic inspection, and (2) planned restoration of deterioration based on the results of inspections.

In the practice we do more than just inspect the equipment and report the failures. If the problem can be fixed at the moment and return the equipment to service we do it instead of just report it and open a corrective order and wait until the next oportunity that operations allows us to touch the equipment again.

The statistics (% of PM orders or hours vs % of Corrective orders or hours) may get somewhat gray, but if I can get the downtime hours lower in this manner I do not care if my KPIs look different than others in the industry.

RCM2,

In my book, details below, I explain why we do maintenance, what we should actually do, and when we should do it.

In brief, it is about managing business risks economically. These risks cover those relating to Technical Integrity (TI), as Josh points out and those pertaining to Profitability. TI in turn affects Safety, Environment and Public Perceptions (and Relations). Profitability depends on Availability, Quality and Productivity. Maintenance is about all these, and it achieves these results through the economic management of reliability.

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

I believe this is the article you are looking for: Time's Arrow

If my Thermodynamics professor back at the 80's could explain "thermo" like John Moubray !!!

Thanks all for ur feed back ,

What is your opinion with below difintion ?

We do maintiance in order to

- To preserve the functional requirements.
- To prevent premature failures as will.
- To mitigate the consequences of failure.

Waiting for your comments .

kind regards,

"We do maintenance because we believe that hardware reliability degrades with age, but we can do something to restore or maintain the original reliability that pays for itself.”

NAVSEA RCM Level 2 Training

Dear MotorDOC

said " We do maintenance because we believe that hardware reliability degrades with age"

What about those eq which they fail randomly, so their failures are not age related. You will not adopt any maintenance strategy on them!!

regards,

RCM2 (What a name..!)

To get back to the initial point. WHy do we do maintenance, to prevent failures or to manage consequences, it is of course to manage consequences.

We only do maintenance, regardles of whether it is (preventive restoration, preventive replacement, predictive, or detective) to eliminate or reduce the consequences of failures that we find to be unacceptable.

The unacceptable-ness (Is that a word?) of the consequences will be determined by the level of risk of failures with safety or environmental consequences in the first instance.

And then cost effectiveness in the second instance.

At all times we do maintenance to manage consequences. Not just for the heck of it.

RCM2,

Randomness of failures is a matter of timing, not of cause.

The cause of failures is degradation, caused by e.g., fouling, fatigue, creep, wear, corrosion, poor quality work, or stressing beyond the design intent.

ALL failures have one or more causes. The initiation of failure may be at an unpredictable point in time. That does not make the failure cause unpredictable.

The P-F curve tells us about the rate of physical degradation. At the 'P' point, failure has ALREADY started. The so called 'Predictive Maintenance' predicts the time-to-functional failure, or the 'F' point, once the P-F curve has been initiated. But the 'P' point itself remains unpredictable.

The probability density curve tells us about the statistical probability of failure at any point in time. It tells us NOTHING about the physical degradation, but a LOT about the correct timing of maintenance.

Regards.

RCM2,

We do maintenance to manage risk.

This can be done either by managing the probability of failure or its consequence.

Using a process such as RCM, we tend to focus on conswequences of failure. If we used a different process such as FTA or HAZOP, we tend to focus on reducing the probability. Even when we do a process such as RCM, sometimes we can only work on the probability, e.g. when the task is a resdesign.

Eventually, what matters is that the risk is at a tolerably low level.

RCM2,

At the end of the day we are talking about managing the consequences of failure, and the probability of realising those consequences.

This is what I stated earlier.

When we are dealing with operational consequences it becomes a matter of cost effectiveness. Either the cost of failure + consequences, and the liklihood that it will occur, requires you to predict, prevent or eliminate it, or it does not.

Either way it is about the lowest cost over a reasonable length of time. Again, a decision based on consequences of failures, not of failures per se.

Thats quite a few different ways of stating the same thing now I would suggest.

thanks all for ur useful feed back


Let me through this Ex , In our PM program we change the oil for X pump in time based period , the reason We do that is to prevent the bearing failures although we have a spare pumps so there will be no failure consequences ,While in other single pumps we do predictive maintenance to give us the status of the bearing , because as much as we do the predictive maintenance we will not prevent the failures so the purpose of this exercise is to prevent failure consequences not the failure it self

So, what I can see here , that the answer for the purpose of maintenance is to prevent failure consequences is applicable for predictive maintenance only.

What do u think

quote:

Originally posted by Vee:
RCM2,

Randomness of failures is a matter of timing, not of cause.

The cause of failures is degradation, caused by e.g., fouling, fatigue, creep, wear, corrosion, poor quality work, or stressing beyond the design intent.

ALL failures have one or more causes. The initiation of failure may be at an unpredictable point in time. That does not make the failure cause unpredictable.

The P-F curve tells us about the rate of physical degradation. At the 'P' point, failure has ALREADY started. The so called 'Predictive Maintenance' predicts the time-to-functional failure, or the 'F' point, once the P-F curve has been initiated. But the 'P' point itself remains unpredictable.

The probability density curve tells us about the statistical probability of failure at any point in time. It tells us NOTHING about the physical degradation, but a LOT about the correct timing of maintenance.

Regards.

Dear Vee what u mentioned is correct , but what about electronics or some of instrumentation which failed randomly which I think u can not apply P-V curve on them & this why we apply detective maintenance on them as u mentioned in ur articles which posted in the Web

regards,

RCM2,

Detective strategies apply ONLY to hidden function failures.

Failures of electronic instruments or panels may be evident or hidden, depending on configuration and availability of indicating lights or other devices. Failures are often due to terminations, dust or moisture tracking, supply voltage spikes and sometimes due to ageing of individual elements such as capacitors or transistors on boards. Because of the complexity of a board, failures will be exponentially distributed for the board as a whole. Usually there is no notice of impending failure, so designers try to put in some indicator lights or other devices to tell you when a board has failed. If not such failures are hidden, and in theory at least a test is appropriate. In practice, boards are replaced after they have failed, and preventive strategies are not always economic.