2 Fundamental& of Malntenance Chapter

used to implement the maintenance tasks. We t�en add�es� the

ongoing process of adjusting that program during the lifetime of the
equipment. In Chap. 3, we establish �he goals and obje_ctives for _an
airline maintenance program that uiill serve the real-life operatw7:.
Chapter 4 discusses the extensive �e�tificati�n requiremen_ts leuied
on the aviation industry from the original design of the uehicle to the
1
establishment of commercial operators and the people who run them.
The documentation for the aircroft, its operation, and its maintenance,
Why We Have to Do Maintenance
is discussed in Chap. 5 and includes the documents produced by the
equipment manufacturers, by the regulatory authorities, and by the
airline itself.
Chapter 6 will identify those activities requ�r�d by the J:AA to
accomplish maintenance as well as tho�e additwn_al requiremenis
deemed necessary by operators to coordinate and implement an_
effective maintenance and engineering program .. Chapter..7 d�fin!s a
maintenance and engineering (M&E) organization. fo: � typi�al
midsized airline. Variations for larger and smaller airlines wil� also Thermodynamics Revlslted
be discussed. Part J, then, can serve as background to the �emainder Nearly all engineering students have to take a course in thermodynamics in
of the book and can, if desired,. be used _as _the b�is for a first or their undergraduate years. To sorne students, aerodynamicists and power
introductory course in the subject of aoiation maintenance plant engineers for example, thermodynamics was a major requirement for
management. graduation. Others, such as electrical engineers for instance, take the course
as a necessary requirement for graduation. Of course, thermodynamics and
numerous other courses are "required" for all engineers because these courses
apply to the various theories of science and engineering that must be under-
stood to effectively apply the "college learning'' to the real world. After all, that
is what engineering is all about-bridging the gap between theory and real-
ity.
There is one concept in thermodynamics that is often a puzzler to students.
That concept is labeled entropy. The academic experts in the thermodynamics
field got together one day (as one thermo professor explained) to create a clas-
sical thermodynamic equation describing all the energy of a system-any
system. When they finished, they had an equation of more than severa! terms;
and all but one of these terms were easily explainable. They identified the terms
for heat energy, potential energy, kinetic energy, etc., but one term remained.·
They were puzzled about the meaning of this term. They knew they had done
the work correctly; the term had to represent energy. So, after considerable
pondering by these experts, the mysterious term was dubbed "unavailable
energy''-energy that is unavailable for use. This explanation satisfied the basic
law of thermodynamics that energy can neither be created nor destroyed; it can
I only be transformed. And it helped to validate their equation.
Let us shed a little more light on this. Energy is applied to create a system
by manipulating, processing, and organizing various elements of the universe.
More energy is applied to make the system do its prescribed job. And whenever
the system is operated, the sum total of its output energy is less than the total

3
 Why We Have to Do Malntenance 5
4 Fundamental& of Malntenance
a number of other constraints which we could refer to as "man-made entropy."
energy input. While sorne of this can be attributed to heat loss through frict�on A design engineer may be limited from making the perfect design by the tech-
and other similar, traceable actions, there is still an imbalance o� energy. Deñning nology or the state of the art within any facet of the design effort. He or she may
entropy as the "unavailable energy'' �f.ª system re�tifies .that 1m�alance. be limited by ability or technique; or, more often than not, the designer may be
The late Dr. Isaac Asimov, hiophysicist and prolific wnter of science �act and limited by economics; that is, there just is not enough money to build that nearly
science ñctíon,' had the unique ability to explain the most difficul: science to perfect system that is on the drawing board or in the designer's mind. Although
the layman in simple, understandable terms. Dr. Asimov sa!s tha: if you wa�t the designer is limited by many factors, in the tradition of good engineering prac-
to understand the concept of entropy in practica! terms, think of it a� the dif- tice, the designer is obliged to build the best system possible within the con-
ference between the theoretically perfect system you have on the drawmg bo�rd straints given.
and the actual, physical system you have in hand. In other wor�s, we can dasign Another common situation in design occurs when the designer has produced
rea� world.
perfect systems on paper but we cannot build perfect syste?1s m the what he or she believes is the optimum system when the boss, who is respon-
The difference between that which we design and that which we can build con- sible for budget asks, "How much will it cost to build this?'' The designer has
stitutes the natural entropy of the system. meticulously calculated that these widgets can be mass produced for $1200
each. "Great," says the boss. "Now redesign it so we can build it for under a thou-
sand dollars." That means redesign usually with reduced tolerances, cheaper
A Saw Blade Has Wldth materials, and, unfortunately, more entropy. More entropy sometimes translates
This concept of entropy, or unavailable energy, can be illustrated by a simple into more maintenance required. The design engineer's primary concern, then,
example. Mathematically, it is possible to take a half of a numbe: repeatedly for- is to minimize (not elimina te) the entropy of the system he or she is designing
ever. That is, half of one is 1/2; half of that is 1/4, half of that is 1/8, and s� on while staying within the required constraints.
to infinity. Although the resulting number is smaller and smaller each time
you divide, you can continue the process as long as you can stand to do so, and The Role of the Mechanlc
you will never reach the end.
Now, take a piece of wood about 2 feet long (a 2 x 4 will do) anda cros�cut saw. The mechanic (technician, repairer, or maintainer), on the other hand, has a dif-
Cut the board in half (on the short dimension). Then take ?ne ofthe preces and ferent problem. Let us, once again, refer to the field of thermodynamics. One
cut that in half. You can continue this until you reach a point where you can �o important point to understand is that entropy not only exists in every system,
longer hold the board to saw it. But, even if you could find sorne way to hold it but that the entropy of a system is always increasing. That means that the
while you sawed, you would soon reach a point where th� piece you h�ve left to designed-in level ofperfection (imperfection?) will not be permanent. Some com-
cut is thinner than the saw blade itself. When (if) you saw rt one more time, there ponents or systems will deteriorate from use and some will deteriorate from lack
will be nothing left at all-nothing but the pile of sawdust on :he floor. The of use (time or environment related). Misuse by an operator or user may also
number of cuts made will be far less than the infinite number of times that you cause sorne premature deterioration or degradation of the system or even out-
divided the number by two in theory. . right damage. This deterioration or degradation of the system represents an
The fact that the saw blade has width and that the act of sawmg creates .ª increase in the total entropy of the system. Therefore, while the engineer's job
kerf in the wood wider than the saw blade itself, constitutes the entropy �f this is to minimize the entropy of a system during design, the mechanic's job is to
system. And no matter how thin you make the saw blade, the fact that it has combat the natural, continua! increase in the entropy of the system during its
width will limit the number of cuts that can be made. Even a laser bea� has operational lifetime.
width. This is a rather simple example, but you can see that the r�al �orld is �ot To summarize, then, it is the engineer's responsibility to design the system
the same as the theoretical one that scientists and sorne engineers live in, Nothing with as high a degree of perfection (low entropy) as possible within reasonable
is perfect. limits. The mechanic's responsibility, on the other hand, is to combat the con-
tinua! increase in entropy during the operational lifetime of the equipment.
i
The Role of the Engineer
Two Types of Malntenance
The design of systems or components is not only limited by t�e. imperf�ct�ons of
the physical world (i.e., the "natural entropy" ofthe system), rt is also limited by Figure 1-1 is a graph showing the level of perfection of a typical system. One
hundred percent perfection is at the very top of the y-axis. The x-axis depicts
time. Tbere are no numbers on the scales on either axis since actual values have
*Dr. Asimov wrote over 400 books during bis lifetime.
 Why We Have to Do Malntenance 7
6 Fundamentals of Malntenance

100%
100%
ATIAINABLE LEVEL OF \ DESIGNED - IN LEVELj
ENTROPY OF PERFECTION
ENTROPY PERFECTION
J ------ ------ ---
---------------------------
a b

SCHEDULED
z � � MAINTENANCE e
o ¡:::
o ... ai
..J

� �o:$
w
u. NATURAL DECAY OF SYSTEM
a: w
w ..J
UN-SCHEDULED
a:
w (lncreasing Entropy) o. a: MAINTENANCE
o.

POINT AT WHICH
SCHEDULED
MAINTENANCE IS DONE

TIME
TIME
Figure 1·2 Restoration of system perfection·
Figure 1·1 The difference between theory and practice.

overhaul of parts or subsystems. Since these breakdowns occur at various,

no meaning in this theoretical discussion. The left end of the curve shows the unpredictable intervals, the maintenance actions employed to correct the prob-
level of perfection attained by the designers of our real world system. �ote that lem are referred to as unscheduled maintenance.
the curve begins to turn downward with time. This is a repres�nta�1on of the
natural increase in entropy of the system-the natural deter1ora�1on _of the Rellabillty
system-over time. When the system deteri_orates_to som_e lo�er (arb1tr_arily set)
level of perfection, we perform some correctíve action: adjusting, t�eakii:ig, se�- The level of perfection we have been talking about can also be referred to as the
reliability of the system. The designed-in level of perfection is known as the
icing, or some other form of maintenance to restore. the �y�tem to its d�s1_gned-m
level of perfection. That is, we reduce the entropy to its original l�vel. This is c�ll�d inherent reliability of that system. This is as good as the system gets during real
world operation. No amount of maintenance can be performed to increase the
preventive maintenance and is usually performed at regular mtervals. This �s
done to prevent deterioration of the system to an unusable level and to ke�p it system reliability any higher than this inherent level. However, it is desirable
in operational condition. It is sometimes re!erred to as sche1uled mainte- for the operator to maintain this level of reliability (or this level of perfection)
nance. This schedule could be daily, every flight, every 200 flight hours, or at all times. We will discuss reliability and maintenance in more detall in Chap.
19· But there is one more important point to cover here-redesign ofthe equip-
every 100 cycles (a cycle is a takeoff anda landing).
Figure 1-2 shows the system restored to its normal level (curv�s a _and b). ment.
There are times, of course, when the system deteriora tes rather rapidly m serv-
ice to a low level of perfection (curve c). At other times the system breaks down Redeslgn
completely (curve d), In these cases, the maintenance acti�ns nec�ssary to
restore tbe system are more definitive, often requiring extens1�e testmg, trou- Figure 1-3 shows the original curve of our theoretical system, curve A. The
bleshooting, adjusting, and, very often, the replacement, restoration, or complete dashed line shows the system's original leve! of perfection. Our system, however,
e Fundamentals of Malntenance Why We Have to DoM alntenance 9

IMPROVED RELIABILITY
(Reduced Entropy)
100% ( DUE TO REDESIGN
(3 cases; see text)
lncrease in. Perfection is
Loganthmic

B A C D

DESIGNED • IN LEVEL
OF RELIABILITY

POINT AT WHICH
SCHEDULED Figure 1-4 Períection vs. cost.
MAINTENANCE IS DONE

Conversely, if the decay rate is the same as befare, as shown in (C) of Fig. 1-3,
or less steep, as shown in (D), then the maintenance interval would be increased
TIME and the overall amount of preventive maintenance might be reduced. The ques-
Figure 1-3 Effecte oí redeeign on eyetem reliability. tion to be considered, then, is this: Does the reduction of maintenance justify
the cost of the redesign? This question, of course, is a matter for the designers
has now been redesigned to a higher level of perfection; that �s, a hígher l�vel to ponder, not the maintenance people.
of reliability with a corresponding decrease in total entropy. During this redesign, One of the major factora in redesign is cost. Figure 1-4 shows the graphs of
new components, new materials, or new techniques may have b�en .used to two familiar and opposing relationships. The upper curve is logarithmic. It rep-
reduce the natural entropy of the system. In sorne cases, a reduct1on m man- resenta the increasing perfection attained with more sophisticated design efforts.
made entropy may result because the designer applied tighter tolerances, The closer we get to perfection (top of the illustration) the harder it is to make
attained improved design skills, or changed the design philosophy. a substantial increase. (We will never get to 100 percent.) The lower curve
Although the designers have reduced the entropy of the �yst�m, t�e system depicts the cost of those ongoing efforts to improve the system. This, unfortu-
will still deteriorate. It is quite possible that the rate of detenorat1on will change nately, is an exponential curve. The more we try to approach perfection, the more
from the original design depending upon numerous fac�rs; thl;1s, the slope of it is going to cost us. It is obvious, then, that the designers are limited in their
the curve may increase, decrease, or stay the same. Wh1ch�ver is the case, the goal of perfection, not just by entropy but also by costa. The combination of
maintenance requirements of the system could b� affected_ m sorne w�y. . these two limitations is basically responsible for our profession of maintenance.
Ifthe decayis steeper, as in (B) in Fig. 1-3, the pomt at wi:uch preventive mam-
tenance needs to be performed might occur sooner, and �he mterval �etween sub- Fallure Rete Patterns
.' sequent actions would be shorter. The result is that mamtenance will be ne�ded
more often. In this case, the inherent reliability is increased, but m?re mamte- Maintenance, of course, is not as simple as one might conclude from the above
nance is required to maintain that level ofreliability (level ofperfecti�n). Unl�ss discussion of entropy. There is one important fact that must be brought out: not
the performance characteristics ofthe system have been impr.ove�, this redesign ali systems or components fail at the same rate nor do they ali exhibit the same
pattern of wear out and failure. And, as you might expect, the nature of the main-
may not be acceptable. A decision must be made to determn.1e if the �erfor�-
ance improvement justifies more maintenance and thus an mcrease m mam- tenance performed on these components and systems is related to those failure
rates and failure patterns.
tenance costs.
 Why We Have to Do Malntenance 11
10 Fundamental& of Malntenance

The United Airlines study showed that only about 11 percent of the items
TABLE 1-1 Fallure Rate Patterns
included in the experiment (those shown in curves A, B, and C of Table 1-1)
A. lnfant mortality; constant or slightly rising would benefit from setting operating limits or from applying a repeated check
\, ) failure rate; definite wear-out period (4%) of wear conditions. The other 89 percent would not. Thus, time of failure or
B. No infant mortality; slightly rising failure deterioration beyond useful levels could be predicted on only 11 percent of the
rate; definite wear-out period (2%) iteras (curves A, B, and C of Table 1-1). The other 89 percent (depicted by
- C. No infant mortality; slightly rising rate; no curves D, E, and F of Table 1-1) would require sorne other approach. The
- definite wear-out period (6%) implication of this variation is that the components with definite life limits
D. Increasing failure rate at outset; constant or and/or wear-out periods will benefit from scheduled maintenance. They will
slightly rising rate; no definite wear-out period (7%) not all come due for maintenance or replacement at the same time, however,
E. No infant mortality; constant failure rate but they can be scheduled; and the required maintenance activity can be
throughout life; no definite wear-out period (14%) spread out over the available time, thus avoiding peaks and valleys in the
F. Infant mortality; constant failure rate throughout workload. The other 89 percent, unfortunately, will have to be operated to
life; no definite wear-out period (68%) failure before replacement or repair is done. This, being unpredictable, would
result in the need for maintenance at odd times and at various intervals; i.e.,
SoURCE: F. Stanley Nowlan and Howard F. Heap: Reliability-Centered Maintenance; U.S. unscheduled maintenance.
Department of Commerce, National Technical Information Service, Washington, DC, 1978.
These characteristics of failure make it necessary to approach maintenance
in a systematic manner, to reduce peak periods of unscheduled maintenance.
UnitedAirlines did sorne studies on lifetime failure rates and found six basic The industry has taken this into consideration and has employed severa} tech-
patterns.* These are shown in Table 1-1. The vertical axes show failure r�tes niques in the design and manufacturing of aircraft and systems to accommo-
and the horizontal axes indicate time. No values are shown on the scales smce date the problem. These are discussed in the next section.
these are not really important to the discussion.
Curve A shows what is commonly referred to as the ''bathtub" curve, for obvi-
ous reasons. This failure rate pattern exhibits a high rate of failure during the Other Malntenance Conslderatlons
early portion of the component's life, known as "infant mortality." This is one of
the "bugaboos" of engineering. Sorne components exhibit early failures for severa! The aviation industry has developed three management techniques for address-
reasons: poor design, improper parts, incorrect usage. Once the bugs are worked ing the in-service interruptions created by the items that must be operated to
out and the equipment settles into its "pattern," the failure rate levels off or rises failure before maintenance can be done. These are equipment redundancy, line
only slightly over time. That is, until the later stages of the component's life. The replaceable units, and mínimum aircraft dispatch requirements.
rapid rise shown in curve Anear the end of its life is an indication of "wear out." The concept ofredundancy of certain components or systems is quite common
The physical limit of the component's materials has been reached. in engineering design of systems where a high reliability is desirable. In the case
Curve B exhibits no infant mortality but shows a level, or slightly rising fail- of redundant units-usually called primary and backup units-if one unit fails,
ure rate characteristic throughout the component's life until a definite wear-out the other is available to take over the function. For example, in aviation most
period is exhibited toward the end. commercial jets have two high-frequency (HF) radios. Only one is needed for
Curve C depicts components with a slightly increasing failure rate with no communications but the second one is there for backup in case the first one fails.
infant mortality and no discernible wear-out period but, at sorne point, it A unique feature of redundant units also affects the maintenance require-
becomes unusable. ments. If both primary and backup units are instrumented such that the flight
Curve D shows a low failure rate when new (or just out of the shop), which crew is aware of any malfunction, no prior maintenance check is required to
rises to sorne steady level and holds throughout most of the component's life. indicate that incapability. On the other hand, if neither system is so instru-
)Curve E is an ideal component: no infant mortality and no wear-out period, mented, maintenance personnel would need to perform sorne check on both pri-
just steady (or slightly rising) failure rate throughout its life. mary and backup systems (at the transit or other check) to determine
Curve F shows components with an infant mortality followed by a level or serviceability.
slightly rising failure rate and no wear-out period. Very often, however, one system (usually the backup) is instrumented to show
serviceability to the crew. If a maintenance check is performed on the other (i.e.,
the primary) the crew can be assured that it is serviceable. In the case of fail-
·Nowlan, F. Stanley and Howard F. Heap, Reliability-Centered Maintenance. National Technical ure, then, they already have a positive indication, through the instrumentation,
Information Service, Washington, DC, 1978.
 Why We Have to Do Malntenance 13
12 Fundamentals ofMalntenance

that the backup system is available and useable. The purpose for this arrange- T�is hst provides �nf�rmation on dispatch of the airplane in the event that cer-
ment is to strike a middle ground between how much instrumentation is used ta�n panels are rmssmg or when other configuration differences not affecting
and how much maintenance is required to ensure system serviceability. In sorne satety are noted.
cases the backup system is automatically switched into service when the pri- �though fail�es on these complex aircraft can occur at random, and can come
mary system fails. Flight crew needs during the flight are primary concerns in r.t mopportune tim_es, these three management actions-redundancy of design
making such decisions. me replaceable uruts, and minim� �spatch requirements-can help to smooth
Another common concept used in aviation is the line replaceable unit (LRU). out the workload and reduce service mterruptions.
An LRU is a component or system that has been designed in such a manner that
the parts that most commonly fail can be quickly removed and replaced on the Establishing a Maintenance Program
vehicle. This allows the vehicle to be returned to scheduled service without
undue delay for maintenance. The failed part, then, can either be discarded or Althou�h ��ere has been a considerable amount of improvement in the qualit
and rehabihty of compon�nts an� systems, as well as in materials and proc;
repaired in the shop as necessary without further delaying the flight.
The third concept for minimizing delays for maintenance in aviation is known �ures, o�er. the 10�-year life of avíatíon, we still have not reached total perfec-
as the mínimum equipment list (MEL). This list allows a vehicle to be dis- tíon. �v1ation e.qmpment, no matter how good or how reliable still needs
certain items inoperative provided that the loss offunc- attention from time to time. '
patched into service with
tion does not affect the safety and operation of the flight. These items are Scheduled m_ainten�nce and servicing are needed to ensure the designed-in
carefully determined by the manufacturer and sanctioned by the regulatory level of perfection (rehability). But, because the real world is as it is sorne of
authority during the early stages of vehicle design and test. The manufacturer tt�s� compone�ts a�d systems will, sooner or later, deteriorate beyond a toler-
issues a master mínimum equipment list (MMEL) which includes all equip- a e evel or will fail com�letely. In other instances, users, operators, or even
ment and accessories available for the aircraft model. The airline then tailors m�mtenance people who mterface with these components and systems can
the document to its own configuration to produce the MEL (more on this in Chap. or eve� abuse the equipment to the extent of damage or deterioration
tthat
at will require the need for sorne sort of maintenance action.
5). Many of these MEL items are associated with redundant systems. The con-
We have seen _that com�onents and systems fail in different ways and at fil.
cept of the MEL allows deferral of maintenance without upsetting the mission
ferent rates. This results m a requirement for unscheduled maintenance th t
requirements. The maintenance, however, must be performed within certain pre·
scribed periods, commonly 1, 3, 10, or 30 days, depending on the operational is s?mewhat erratic and uncertain. There are often waves ofwork and no-wo:k
penods that ne�d to be managed to smooth out the workload and stabilize the
requirements for the system.
The items are ídentified in the MMEL by flight crew personnel during the manpower requírements,
latter stages of new aircraft development. Thus, flight personnel determine . Those components exhibiting life limita or measurable wear-out characteris-
what systems they can safely fly the mission without orina degraded condi- tics can �e part of a systematic, scheduled maintenance program. Design redun-
tion. These flight crew personnel also determine how long (1, 3, 10, or 30 days) dancy,. hne replaceable units, and mínimum dispatch requirements have been
estabhshed as management efforts to smooth out maintenance workload. But
they can tolerate this condition. Although this is determined in general terms
there are numerous .�o�ponents and systems on an aircraft that do not lend
prior to delivering the airplane, the flight crew on board makes the final deci-
sion based on actual conditions at the time of dispatch. The pilot in command themselves. to s1;1ch adjustment for convenience." Occasionally, inspections
(PIC) can, based on existing circumstances, decide not to dispatch until repairs and/o� �odifications of equipment are dictated-within specified time limits-
are made or can elect to defer maintenance per the airline's MEL. Maintenance by av1a!1on regulators as �ella� by manufacturera. It is necessary, then, that
must abide by that decision. the mamtenan�e and engmeenng organization of an airline be prepared to
Associated with the MEL is a dispatch deviation guide (DDG) that contains address the mamtenance of aircraft and aircraft systems with a well-thought-
instructions for the line maintenance crew when the deviation requires sorne out and well-executed program. �e remainder of this textbook will address the
many-faceted pr�cess known as aireraft maintenance and engineering.
mJJ.intenance action that is not necessarily obvious to the mechanic. A dispatch
deviation guide is published by the airplane manufacturer to instruct the �e program discussed h�rein h�s ?een created over the years by concentrated
mechanic on these deviations. The DDG contains information such as tying up and mtegrated efforts by pilota, airlines, maintenance people manufacturera
cables and capping connectors from removed units, opening and placarding cir- co�ponent an� sy�tem ª';1P�liers, regulatory authorities, and �rofessional and
cuit breakers to prevent inadvertent power-up of certain equipment during busmess o�garuzations within the aviation industry. Not every airline will need
flight, and any other maintenance action that needs to be taken for precau- to �e .º�g�zed and operated in the same manner or style, but the programa and
actívítíes discussed in this text will apply to all operators.
tionary reasons. Similar to the MEL is a configuration deviation list (CDL).