The Effects of Maintenance on Reliability
By Bill Brinkley AP / IA
Manager of Reliability and Development
USAirways Express / Piedmont Airlines
Publisher's Note: I will caution you that a) Bill
Brinkley comes from the commercial Aviation industry where
reliability is not optional and b) he pulls no punches when
speaking to industrial maintenance professionals. I an
interested in hearing what you can take away from this paper
that was presented at IMC-2007. You can
email me your comments or you can
post them here.
In an airline environment, maintenance is king. An aircraft
receives about 17 man-hours of maintenance for every flight
hour. That may seem excessive - unless you are the one riding in
that aircraft. So – the question is, do aircraft really break
all that often and do they need that much maintenance? Are they
that unreliable? The answer, of course, is no.
Aircraft are designed to be reliable, so why perform all that
maintenance? Over ninety percent of the maintenance performed on
an aircraft is preventive or servicing in nature. Preventive
maintenance is done to maximize availability of the aircraft for
operational service and minimize the number of failures
occurring at inconvenient times or places.
The remainder of the maintenance is either routine scheduled
maintenance, random failed component replacement, or various
types of damage.
Does all this maintenance have any impact on overall
reliability? Study and analysis on the relationships between
reliability / safety and scheduled maintenance activity has been
accomplished by major aircraft and component manufacturers.
Their conclusion was that the reliability of only a small number
of components is directly dependant on scheduled maintenance.
Most components do not benefit from it.
Components can generally be divided into 4 groups:
1. Components whose reliability is independent of age
2. Components subject to wear-out beyond equipment life
3. Components whose inherent reliability levels are directly
dependent upon scheduled maintenance activity
4. Components that are susceptible to age related deterioration
Each group has a unique reliability / age relationship
Group 1 - Components who’s Reliability is Independent of Age
This group includes components whose failure rate is
exponential, such as electronic equipment or other equipment
with no moving parts. They may or may not exhibit infant
mortality, and regularly scheduled maintenance has no effect on
Regularly scheduled maintenance on these components is only
effective as a "failure finding" task. This group represents
about 90% of aircraft equipment.
Components in Group One tend to have either a fairly high infant
mortality rate or none at all, but once that initial burn-in
period is over, component reliability stabilizes. If it is going
to fail, it will fail early on after installation. Once it gets
past this infant mortality period, it can pretty much be counted
on to last.
Predictive, proactive, or preventive maintenance on these types
of components will have little or no effect on their
An example on your car would be the radio. If it gets past the
first few weeks trouble free, it will likely not fail for the
life of the vehicle. There is no preventive maintenance task you
could perform that would affect the reliability. It will work
until it doesn’t work anymore. Then you change it.
Group 2 - Components Subject to Wear-out Beyond Equipment Life
This group includes Hydro / Electro Mechanical Equipment that
has adequate design margins to push the wear-out region beyond
the life of the equipment it is installed on.
They behave a lot like Group 1 Components, and their reliability
is generally unaffected by scheduled maintenance. Failures, when
they occur, tend to be abrupt and random, and provide little or
no pre-failure indication.
Occasionally Group 2 components experience premature failure or
infant mortality, but in general, failures of components in this
group are discovered through effective reliability monitoring
rather than through the maintenance program.
Components in this group have an inherently stable design that
under normal circumstances will have a life that significantly
exceeds that of the equipment it is installed on. Some might
require an occasional lubrication or cleaning task, but most do
not benefit from ongoing maintenance. Not only do they not
benefit from it, many do not even have provisions or
requirements to perform any maintenance. With these components,
there is generally little you can do to extend component life,
and since the component is already designed to exceed the life
of the equipment it is installed on, there is little need to
An example on your car might be the windshield. Barring any
accidental damage, it should last well beyond the life of the
vehicle it is installed on. When you tow that old clunker to the
scrap yard, chances are the windshield will still be okay
provided no one has broken it along the way. There is nothing
you can do from a maintenance standpoint to extend the life of
the windshield, and even if there were, you would have to
carefully consider whether it was worth the time and effort to
do it. If there were something you could do that would extend
the life of the windshield by fifty percent, would there be any
benefit in doing it? Probably not, since it will still be fully
functional when the car is scrapped anyway.
Group 3 - Components Whose Inherent Reliability Levels are
Dependent on Scheduled Maintenance Activity
These typically have heavily loaded dynamic interfaces or high
speed rotating parts.
Scheduled tasks generally involve lubrication and / or
analytical inspection. Components in this group clearly benefit
from routine maintenance.
Sticking with the car example, this group would include things
like oil and filter changes. You can operate your vehicle
without performing this task – for a while – but doing so will
accelerate the failure rate.
Performing the maintenance task, though, does not absolutely
eliminate the failure from occurring, but it does delay it and
may lessen the severity when it does occur.
Group 4 – Components Susceptible to Age Related Deterioration
Components in this group will wear and will eventually fail
regardless of the type or amount of maintenance performed.
You can sometimes delay the inevitable failure by employing
operational procedures, but maintenance procedures will have
little if any effect, and may yield no real value.
An example would be the brakes on your car. They will eventually
wear to the point of failure. Period. There is nothing you can
do from a maintenance standpoint that will prevent or even
significantly delay that failure from occurring. All you can do
is plan for the failure and be prepared for it when it occurs.
It is not a question of “if” they are going to fail, it is
merely a question of “when”.
There are a few things you can do from an operational standpoint
– things like not “riding” the brake pedal – that will extend
brake life a small amount. From a maintenance point of view, all
that can be done is to check the components on a routine basis
and change them when wear patterns indicate that they need to be
Reliability ≠ Scheduled Maintenance
The majority of the equipment in Groups 1 and 2 do not have a
direct relationship between scheduled maintenance activity and
reliability. If they fail, it will be a random, unpredictable
type of failure. Routine maintenance would be ineffective and
Sort of like rearranging the deck chairs on the Titanic.
Scheduled maintenance should be limited to those Group 3
components where inherent reliability levels are dependent on
scheduled maintenance activity and those Group 4 components
where they are susceptible to damage or other forms of age
related deterioration. Basically, if it ain’t broke, don’t fix
it. If performing routine maintenance on it won’t have any
effect on component reliability, why do it?
If you have an exceptional reliability program and can predict
with any degree of accuracy when a Group 3 or Group 4 component
will fail, changing that component as a preemptive measure might
be worthwhile, particularly if you want to be able to control
when and where that component gets replaced. In an airline
environment, there are many parts we change for that very
For example, if an aircraft tire passes inspection during a
routine maintenance check at the hangar, but the technician
knows that the tire will probably be out of limits before the
next time the aircraft comes in for maintenance, it is better to
change the tire early than to have to change it at a remote
airport a few days later and cause a flight delay. Sure, you
lose a little life on the existing tire, but you have to compare
that cost to the cost of a flight delay.
Sometimes maintenance can help improve reliability, but in many
cases, it can’t.
Doing more does not always bring about a commensurate increase
An additional factor that makes aviation reliability different
from other types of equipment reliability is the differences in
how reliability is calculated. For most reliability
calculations, the standard is developed based on MTBF – or Mean
Time Between Failures. For an airline, this is no doubt the
worst possible factor to use in a reliability calculation.
Simply, an airline has to be proactive – and running a component
to the point of failure is certainly not proactive – and it
isn’t safe. Equipment and component failures on an aircraft have
the potential to ruin your whole day – as well as those of the
Granted, failures do occur, but every possible precaution is
taken to limit it. Mean Time Between Removals – or MTBR – is not
necessarily a good metric of measurement, either, as most major
components are on a time or cycle controlled replacement
schedule anyway, so MTBR is known before the component is even
The factor that becomes important in the airline world is Mean
Time Between Unscheduled Removals – MTBUR. While MTBUR includes
MTBF, as a failure is an unscheduled removal – there are also
unscheduled removals that are not the result of a failure of the