|
By Rolly Angeles
Infant Mortality Failures can best be explained from page 143 on the book of John Moubray on Reliability-Centred Maintenance, he quotes : It is also borne out by machine operators who says that every time the maintenance works on it over the weekend, it takes up to Wednesday to get it going again. Have you been a victim of this phenomena ? Operators often say sarcastically that if you have not performed your regular overhauling and replacement on this piece of equipment, I bet it will be running without any problems. But isn’t it the main purpose of Scheduled Maintenance is to ensure that the equipment performs as it is intended, but the opposite seems to happen whenever maintenance do something on the equipment. Is it profound ? NO! You are just a victim of what you called Infant Mortality Failures. In almost all cases, Infant Mortality Failures are caused by human errors and intrusive or forced maintenance. In fact many things can go wrong when we try to dismantle an equipment for overhauling purposes. Human errors such as slip and lapses can occur on the part of the maintenance performing the overhauls.
Infant Mortality Failures are failures which occur at the beginning of life, others refer them as commissioning failures, start-up failures or debugging failures. Many factors affect Infant Mortality Failures which includes poor equipment design, poor quality manufactured, incorrect installation, incorrect commissioning, incorrect operation, unnecessary maintenance, slip and lapses, human errors or simply bad workmanship.
The case of Infant Mortality Failure which is Pattern F of the Six Failure Pattern starts off with a high incidence of early failures which eventually drops to a constant or very slow increasing conditional probability of failure and ending in a no wear out zone.
|
|
Studies done by Nowlan and Heap on civil aircraft showed results that 68% of items that failed conformed to pattern F which is the case of Infant Mortality Failures. An example of some of the benefits achieved from this learning on infant mortality failures by the civil aircraft industry was based on a dramatic reduction in their scheduled overhauls in their DC-8 aircraft from 339 items for overhaul to only 7 items in their DC-10. One of the items no longer subject to overhaul was their turbine engines. Likewise on the initial program developed for Boeing 747, it took United Airlines 66,000 man hours on major structural inspections to reach an inspection interval of 20,000 hrs compared to their traditional 4 million man hours inspection on smaller and less complex aircraft such as DC-8. Truly, it had been a remarkable achievement and feat which saved them millions of dollars on maintenance without compromising its safety. But perhaps the greatest benefit on learning from Infant Mortality Failures is that it save lives. Below is the actual record for worldwide commercial fleet from 1959 to 1995 indicating the rate of accidents per million take-offs. |
|
Performing too much overhaul on our equipment to comply with PM specs and activities is not a guarantee that equipment will be reliable in the future. It is similar to the logic that a fat child is not exactly a healthy child, healthy and fat are not synonymous, in fact there are cases when a child is more prone to sickness when he/she is fat. Or eating plenty of food each meal is not an indication of a healthy person, eating the right balance food each meal can make us a healthy person. Similar thing holds through with maintenance, the more activities we perform on our equipment the more likely it will fail. I have seen a lot of cases that when an equipment is subject for overhaul and returned to operations, the operators have a hard time of running the equipment, as well as a lot of cases where maintenance tend to forgot some pieces that should be in the equipment most specially if the part is just a small item. Most people are shocked when I say during my training that the More PM you perform on your equipment, the more problems you will encounter, the less PM then the less problems you will encounter. I never said No PM means No Problem. PM will always have a place in the maintenance function. |
|
If Infant Mortality Failure exists, the question raised, is can we eliminate them ? Many factors contribute to infant mortality failure and it is impossible to eliminate them, the best we can do is to reduce its likelihood. Let us look at the following perspective :
Design Stage
No equipment is perfect by design, there is always a design weakness attributed within it. There are parts that fails prematurely during commissioning and even in actual operation. We need to identify these parts and conduct a physical investigation and Root Cause Analysis as to why these parts fail prematurely. Only when we understand fully the cause of failure can we recommend a redesign in the part such as changing its shape or strength of material.
Commissioning Stage
Problems occur during commissioning of equipment in our plant. Either some human error was involved during set-up or start up. Infant Mortality Failures are the reason why vendors and manufacturers provide some form of warranty on their equipment most specially during commissioning and start up activities of the equipment.
Scheduled Overhauls and Replacement
When equipment is operational, infant mortality failures occurs more frequently when maintenance intervenes the equipment to perform their time-based scheduled overhaul and replacement. There is always the assumption that when maintenance dismantle the equipment, they can put it back confidently together. It is not the intent of this newsletter to discredit Preventive Maintenance, but rather to educate each and everyone that there are cases when performing overhauls on equipment can introduce infant mortality into otherwise stable system. There are many ways to check the condition of the equipment without dismantling it. These techniques are known as Condition-Based Maintenance. Example, instead of overhauling an engine, why not check its oil. Normally when the equipment is not in good shape and actually needs reconditioning and overhauling, it will be indicated in the oil in the form of particulates mostly metals. Metals present in oil can indicate which part is actually on the verge of wearing out so that maintenance can focus with great accuracy on which parts to replace.
With this in mind, we should rethink again our strategy when performing PM overhauls and replacement on a scheduled frequency because this is when the equipments are more vulnerable to infant mortality failures. Perhaps, we might not get a high score on our PM completion and for manufacturing, there might be possibilities that Quality Control will provide a non-conformance to the maintenance department because PM was ignored. If we intentionally missed PM due to high incidence of Infant Mortality Failures in our plant, I believe our reasons are valid and that sometimes it is best for our equipment not to be disturbed.
Random failures are failures that occur at any given period. This means that the probability that an item will fail in any one period is the same as it is in any other. This means that the conditional probability of failure is constant.
To further illustrate my point on random failures, imagine that you are driving your car and a small piece of rock hit your windshield which cause a very small fracture. My question is when do you think it will happen again in the future ? That is a difficult question in which the answer will always be a form of guess. To know the occurrence of random failures is like catching a lighting with a Polaroid camera on that precise second it shows up. |
|
In the given example above , it is a lists of failure distribution of the same bearing that failed during operations. Every time this same bearing failed, maintenance will record the date when it failed. We are speaking of the same bearing and that our history showed that 5 bearings failed on the first year, 15 on the second year and so on. No bearings achieved a life of more than 9 years. With this distribution, my question is when is the best period and time to replace this bearing ? If you answer on the first year, then our maintenance effort would be very costly since most of the bearings will last for more than 1 year. This means that we are replacing good parts or parts which are not yet on the verge of failing. If you answer that the best time to replace the bearing will be on the 8th year, then we have a case of reactive maintenance or run to fail situation in our operations. Perhaps this would be a good idea if the consequences of failure can be limited to the cost of repair or a redundancy is in place. But if this is not the case then a run to fail scenario would not possibly be the best strategy. My point is no amount of scheduled maintenance can fully address a case of random failure. So if this is the case your experiencing with some of your parts in your equipment, the following are some of the options to address random failures :
Use of Predictive Maintenance Strategy
Predictive Maintenance is valid if the part whose failure is random posses a potential failure. A potential failure is actually a symptom or indication that a part is on the verge of failing, although it is not yet in its failed state or it have not reached its functional failure. Sound, noise, excessive heat, vibration are all signs of Potential Failure that can be detected with an ailing bearing. Predictive Maintenance instruments can aid us in determining the potential failure of certain parts of our equipment such as this bearing on our case. Once a potential failure is spotted, then we need to determine the P-F Curve and perform some maintenance before it actually reached its functional failure stage.
Analyze Using Root Cause Failure Analysis
Understanding the causes of failures through Root Cause Failure Analysis. Failures happen for a reason and there is always a cause and effect on it. Most people make the mistake on jumping to redesign or modification without understanding the cause behind the failure itself and only to realize that a new born problem will emerge as a result of this redesign and modification. Performing a thorough investigation will provide us some shed of light in understanding why the failure occurred in the first place as based from the evidence. Hence, when performing a Root Cause Failure Analysis, the depth of probe should always end up on the Latent Cause of Failure.
Run To Fail Strategy
Run to Fail is feasible and applicable to use if the consequence of failure will be limited to the direct cause of failure and would not compromise safety, environment and operations. This is also applicable if the system have some form of back-up or redundancy making the failure tolerable and safe as in the case of the Airline Industry. If something fails when the plane is airborne, I cannot provide you a rope to check the condition of the engine or perform some sort of Predictive Maintenance. They allow failure to happen simply for the reason that they have some form of redundancy making failure tolerable and safe.
Hence, to conclude this August 2008 Edition of our Reliability Newsletter, for failures which are random in nature, never, ever use Preventive Maintenance or time-based strategy as this is where this strategy will be useless and weakest. Applicability of using Preventive Maintenance overhauls and replacement is applicable to parts which conform to a wear out pattern. Just remember, the more automated and complicated your equipment are, the more cases of Infant and random failures we will encounter and what is important is understanding which tasks is the most feasible to use. |
|
In the real world of maintenance, parts with failure patterns on Infant Mortality and Random Failures are much more frequent than age-related or wear out failures. And that these two failure patterns require a different strategy than Preventive Maintenance |
|
You are receiving this newsletter and email because at some point in time you opted to be included in our Monthly Reliability Newsletter mailing list from our site.
Our newsletter will be sent out once or twice a month and provide you with quality issues and resources on our most common link which is all about reliability and maintenance, as well as regular updates on our articles. I would like to personally invite you to regularly visit our website and check out updates on our articles and training courses. |
RELIABILITY NEWSLETTER |
|
Issue No. 16 - August 2008 |
|
Since establishing this website last May of 2007, I have made many efforts to improve this site and provide some useful insights about our common link which is all about improving our equipment reliability and the way we do maintenance.
Finally should you be interested to contribute to our articles section or share any feedback, I encourage you to email me at rollyangeles@rsareliability.com. If for any reason you wish to unsubscribe from our newsletter, kindly send my a blank email with unsubscribe as the heading and we shall remove you from our mailing lists. Once again welcome to our August 2008 edition of our Monthly Reliability Newsletter for this year 2008 and I hope that you enjoy reading.
My Warm Regards,
Website : www.rsareliaiblity.com Email : rollyangeles@rsareliability.com |
|
RELIABILITY & MAINTENANCE CONSULTANCY FIRM
|
|
Where the learning just never stops . . .
|
|
RSA
|
|
For our complete archive of previous reliability newsletter |
|
RSA Reliability and Maintenance Consultancy Firm values the privacy of your email. Should you wish to unsubscribe or do not want to receive any more email or messages from us kindly send me a blank email with unsubscribe on the subject and we will remove you from our mailing lists. Click here to Unsubscribe ! |
