by Rolly Angeles, Reliability Consultant

 

I have asked some of my friends from reliability forums if all failures can really be eliminated, some say yes and others simply say that we can only eliminate a fraction of them.  Hence, I tried to rephrase my question by asking which of these following statements would you think would be more appropriate, relevant and meaningful :

 

1st : We can eliminate failures by analyzing them through

        a thorough RCA and RCFA methodologies.

 

2nd : We cannot eliminate the likelihood of a failure but

         rather we can only prevent or predict the failure from

         occurring on its own.

 

3rd : Failures cannot be eliminated and that the best that we

        maintenance and reliability people can do is to reduce

        the frequency of failures.

 

4rth : Failures cannot be eliminated, the best that we can do

         is to delay or prolong the process of failure occurring.

 

First and foremost, All failures are not created equal (each failure will have their own consequence), and that the word failure itself is very broad and diversified and can have some different meaning or interpretation , sometimes confusion arise as when do we call it a failure, before answering my question, I would like to explain failure in the way I have understand it.

 

PATTERNS OF FAILURE (HOW FAILURE OCCURS)

 

Infant Mortality Failures - These are failures that occur at the beginning of its life.  Others call them start-up failures, commissioning failures which are likely to occur after a major overhaul or Preventive Maintenance had been initiated

 

Random Failures - These are failures that can occur at any given period, and this is where our routine Preventive Maintenance will be at its weakest point, recommended task that can be used for Random failures will be Predictive Maintenance if the failure provides a warning or potential failure and that it is on the verge of process of occurring, modification or simply run it until it fails if the consequences

when the risk and consequences will be low and tolerable.

Wear Out Failures or Age-Related Failures where parts will eventually survive to the age that it is suppose to fail

 

John Moubray author of RCM further indicates in his book  that these three failures mostly occur in combination by studying parts behavior and led to what we call the 6 failure pattern, But eventually those 3 patterns of failure (bathtub curve) can be seen as existing in TPM books, Statistical  books and Weibull Distribution.

 

CLASSIFICATION OF FAILURES

 

Hidden Failures, failures that will not become evident to the operator or maintenance when it occurs on its own such as failure of a standby pump with the duty pump running

 

Evident Failures, are failures that will become evident to the operator and maintenance when it occurs on its own

 

DIFFERENT TYPES OF FAILURES

 

Function Loss Failure - or failure of the primary function, wherein when a failure occurs the equipment will totally stop and will definitely affect operations.  RCM term this as failure of the Primary Function.

 

Function Loss Reduction - or failure of secondary functions wherein when a failure occurs the equipment can still be capable of running. Every equipment have their own primary and secondary functions that we must be aware off. One thing to note is that there are cases that the consequences of failure of secondary functions are far much worst than the failure of the primary function

 

From TPM Books, references and resources breakdowns and failures can either be Planned or Unplanned and what maintenance can prevent or eliminate or zero out is the unplanned breakdown and not to mention cases of chronic and sporadic failures and breakdowns.

 

Planning for an activity to be performed on our equipment, such as routine Preventive Maintenance, parts replacement will constitute the Planned Breakdown, and also Predictive Maintenance group spotting that a potential failure which is now in progress and the next step is scheduling the equipment for some intervention also constitute a Planned Breakdown . While Unplanned Breakdown simply means that when failure occurs we are unaware and are always caught by surprise on its occurrence.

 

Hence, going back and answering  these statements :

 

1st : We can definitely eliminate failures by analyzing them through a thorough RCFA and RCA

 

Wrong !!! Failures cannot be totally eliminated by analyzing them through RCFA and RCA. We must consider that there are a variety of causes to understand and analyze and every single failure have their own unique causes and when we  treat a single cause then there is a likelihood of that same part to fail again in the future due to a different cause, remember when we speak about Root Cause Failure Analysis and Root Cause Analysis, we are dealing with evidences as to what really have caused the part to fail and not all the probable causes that might have cause the part to fail.

As my good and dear friend Bob Nelms from Failsafe-Network once told me that you cannot move an ocean with a fork (not even a spoon) simply concurs that it is possible to remove or eliminate the causes but not the failure itself.

 

A bearing can fail for a multiple variety of reasons and taking care only on a single cause (perhaps the analysis shows that bearing failure was due to lubrication failure) will not eliminate its recurrence since it can fail in the future due to some other reasons such as brinelling, pitting fatigue, spa-lling, misalignment, careless handling, over lube, under lube, and many more.

 

2nd : We cannot eliminate the likelihood of a failure, but rather we can only prevent or predict the failure

from occurring.

 

Yes for some failures, it is possible to prevent them and this is true for age related failures or parts and components which inhibits some sort of wear out mode and which constitute around 15 to 20% of overall equipment failures and likewise we can predict failures for those who shows signs of potential failures.

But this is not the goal of maintenance, if a bearing fails prematurely, yes we can predict that it is on the verge of failing, but the life of the bearing had not been maximized at all, perhaps we were able to predict it from failing and we have advise operations about it but we should not stop from here, data from Predictive Maintenance can be most useful in analyzing why component fails, Predictive Maintenance is a one step forward towards achieving a Proactive Maintenance and not to mention that not all failures can be predicted and prevented.

3rd : Failures cannot be eliminated and the best that

Maintenance can do is to reduce them.

 

Wrong !!! Some think that when we have experience zero unplanned breakdown or have reduced failure tremendously as per seen on our breakdown indices for the past couple of years we think that we have eliminated the failure completely, but come to think of it technically have we have really eliminated the likelihood of a failure  or rather we have just delayed the failure process to occur ? Just try to think of buying a second hand car and driving it and giving it the best maintenance ever, eventually there are still many parts that are subject to  wearing out and when parts wear out then they have actually and eventually failed isn’t it ?

4rth : Failures cannot be eliminated and that the best that maintenance can do is to prolong or delay the

process of failure itself from occurring.

Correct !!! I think that we must not be misled that maintenance can eliminate failures, Failures cannot be eliminated. they will happen and they will occur, the best that we can do is to delay its process, control the timing of failure or eventually prolong the occurrence of failure, yet in the end failure and breakdowns will occur on our equipment and we must be ready for it.

CONCLUSION:

Failures cannot be eliminated and failures vary in consequences, the best the we maintenance and reliability people can do is to delay the process of failure and control the timing of failure but eventually and inevitably expect failure to occur in the future.

If you are experiencing a no failure situation in your equipment, it is just temporary, because of your good system of maintenance, the truth is you are just delaying the process of failures.

I've been in the field of TPM for so many years and its goal is to zero out unplanned breakdowns is idealistic and in its technical sense next to impossible if we really understand how part really behaves, when parts wear out, in its technical sense it failed. Hence, we are not really eliminating failure itself but just doing the best we can to prolong the life of the part or component.