Modeling The Life Cycle Cost of Jet Engine Maintenance

Reglerentwurf zur Synchronisation einer hybriden
Aktuatorkonfiguration
Diplomarbeit
Modeling the Life Cycle Cost

of Jet Engine Maintenance
Students Research Project
by
Ralf Seemann
[email protected]
Hamburg
October 2010
Abstract
Cost incurred by aircraft maintenance, repair and overhaul (MRO) make up a considerable
proportion of the total life cycle cost of an aircraft. For an evaluation of the economic efficiency
of commercial aircraft, it is therefore crucial to estimate its MRO cost. The largest share of these
cost are incurred by the maintenance of the aircraft engines. Engine maintenance is performed
on-condition in dedicated workshops mainly independent from the regular maintenance check
events of the remainder aircraft. For a consideration of aircraft life cycle cost, it is hence also
necessary to predict the intervals of these engine shop visits.
The present study discusses an approach for estimating the life cycle cost of aircraft engine
maintenance. It provides an extensive review on basic concepts of jet engine MRO as well as on the
primary factors that affect the engine maintenance cost and intervals. Based on these resources,
a database was assembled from historic maintenance data provided by the aviation magazine
“Aircraft Commerce”. Through linear regression analysis of the database, cost estimating
relationships (CERs) describing the correspondence between maintenance cost/intervals and
basic engine specifications were derived. These CERs are complemented by a series of adjustment
factors that were developed in order to reflect additional influential effects, such as operational
severity or engine age. The resulting model demonstrates that reasonable figures for the engine
shop visit intervals and cost can be estimated by considering the engine take-off thrust, engine
dry weight, engine maturity, average flight length, applied derate and environmental conditions as
primary influence factors. Since the assembled database contains only maintenance information
of the currently mature engine generation, the validity of the developed model is limited to the
current engine generation. However, it is assumed that the basic maintenance characteristics
remain unchanged with the next engine generation. Plausibility tests, which compare the model
results with estimates for the maintenance cost of the next engine generation, indicate that the
more advanced engines can be represented by the developed model through the use of technology
factors.
The developed model is intended to complement the aircraft life cycle cost simulation tool
(LCC-tool), which is being developed at the Institute of Air Transportation Systems at Hamburg
University of Technology. The LCC-tool uses Matlab as programming environment and enables
the evaluation of technologies under the incorporation of expertise in form of technology factors.
Therefore, the resulting model was implemented into the structure of the existing Matlab
programme sequence.
I
Contents
List of Abbreviations V
List of Symbols VII
1 Introduction 1
1.1 Thesis Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Thesis Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Literature Review 3
2.1 Aircraft Gas Turbine Engines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1.1 Basic Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1.2 The Turbofan Engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Aircraft Engine Maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2.1 On-Wing Engine Maintenance . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2.2 Engine Overhaul - Shop Visit . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2.3 Engine Time On-Wing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2.4 Engine Maintenance Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.3 Modeling of Engine Maintenance Cost . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3.1 Reflection of EMC in DOC methods . . . . . . . . . . . . . . . . . . . . . 27
2.3.2 Parametric Cost Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3 Development of Cost Estimating Relationships 32

3.1 Database Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.1.1 Establishing the Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.1.2 Review of Data Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.1.3 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.1.4 Data Normalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.1.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.2 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.2.1 Candidate Relationship Screening . . . . . . . . . . . . . . . . . . . . . . 40
3.2.2 Regression Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.3 Results of the Parametric Cost Modeling . . . . . . . . . . . . . . . . . . . . . . . 43
3.3.1 Shop Visit Interval CERs . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.3.2 Shop Visit Cost CERs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
II
Contents III
4 Modeling of Engine Maintenance Cost 46

4.1 Model Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.1.1 CER-Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.1.2 Effect-Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.1.3 Spare Engine Charges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.2 Example Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.2.1 Base Costs and Intervals from CERs . . . . . . . . . . . . . . . . . . . . . 52
4.2.2 Adjustment of Intervals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.2.3 Final Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.3 Model Plausibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.3.1 Model Results vs. Original Database . . . . . . . . . . . . . . . . . . . . . 55
4.3.2 Model Results vs. Additional Data Sources . . . . . . . . . . . . . . . . . 58
4.3.3 Summary of the Plausibility Tests . . . . . . . . . . . . . . . . . . . . . . 60
4.4 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.4.1 Sensitivity of Model Output . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.4.2 Sensitivity of Life Cycle SVC . . . . . . . . . . . . . . . . . . . . . . . . . 63
5 Implementation into existing LCC-Tool 68

5.1 Function Definition and Input Modification . . . . . . . . . . . . . . . . . . . . . 68
5.2 Estimating the Shop Visits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
5.3 Processing the Predefined Shop Visits . . . . . . . . . . . . . . . . . . . . . . . . 72
5.4 Consideration of Spare Engine Costs . . . . . . . . . . . . . . . . . . . . . . . . . 72
5.5 Estimation of Required Shop Visit Number . . . . . . . . . . . . . . . . . . . . . 72
5.6 Definition of Last Shop Visits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5.7 Output Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
6 Summary and Conclusion 75
Bibliography 77
List of Figures i
List of Tables ii
Appendix ii
A Maintenance Costs iii

A.1 Engine MRO Cost Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
A.2 Shop Visit Cost Driver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
B Database iv
B.1 Aircraft Commerce Shop Visit Reserves & Intervals Example Table . . . . . . . . iv
Contents IV
B.2 Classification of Aircraft Engines . . . . . . . . . . . . . . . . . . . . . . . . . . . iv

B.3 Core Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
C Regression Analysis vi
C.1 First Interval SH Engines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi
C.2 Mature Interval SH Engines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
C.3 First Interval MLH Engines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
C.4 Mature Interval MLH Engines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
C.5 First Shop Visit Restoration Costs . . . . . . . . . . . . . . . . . . . . . . . . . . x
C.6 Mature Shop Visit Restoration Costs . . . . . . . . . . . . . . . . . . . . . . . . . xi
C.7 LLP Cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
D Model Parameters xiii

D.1 Averaged Short-Haul-Engine Severity Curve . . . . . . . . . . . . . . . . . . . . . xiii
D.2 Averaged Medium-Long-Haul-Engine Severity Curve . . . . . . . . . . . . . . . . xiv
D.3 Time & Material Factor Curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
E Model Analysis xvi

List of Abbreviations
AC . . . . . . . . . . . . . . . . . Aircraft
ACA . . . . . . . . . . . . . . . Aircraft Commerce articles
AD . . . . . . . . . . . . . . . . . Airworthiness directive
BLS . . . . . . . . . . . . . . . . Bureau of Labor Statistics
BPR . . . . . . . . . . . . . . . . Bypass ratio
CER . . . . . . . . . . . . . . . . Cost estimating relationship
DB . . . . . . . . . . . . . . . . . Data base
DMC . . . . . . . . . . . . . . . Direct maintenance cost
DOC . . . . . . . . . . . . . . . Direct operating costs
ECI . . . . . . . . . . . . . . . . Employment cost index
ECM . . . . . . . . . . . . . . . Engine condition monitoring
EF . . . . . . . . . . . . . . . . . Environment factor
EFC . . . . . . . . . . . . . . . . Engine flight cycle
EFH . . . . . . . . . . . . . . . . Engine flight hour
EGT . . . . . . . . . . . . . . . Exhaust gas temperature
EGTM . . . . . . . . . . . . . Exhaust gas temperature margin
EMC . . . . . . . . . . . . . . . Engine maintenance costs
EPR . . . . . . . . . . . . . . . . Engine pressure ratio
FADEC . . . . . . . . . . . . . Full authority digital engine control
FOD . . . . . . . . . . . . . . . . Foreign object damage
FR . . . . . . . . . . . . . . . . . First-run
HPC . . . . . . . . . . . . . . . . High pressure compressor
HPT . . . . . . . . . . . . . . . . High pressure turbine
IMC . . . . . . . . . . . . . . . . Indirect operating costs
IOC . . . . . . . . . . . . . . . . Indirect operating costs
IPC . . . . . . . . . . . . . . . . Intermediate pressure compressor
LCFH . . . . . . . . . . . . . . Life cycle flight hours
LLP . . . . . . . . . . . . . . . . Life limited part
LM . . . . . . . . . . . . . . . . . Line maintenance
LPC . . . . . . . . . . . . . . . . Low pressure compressor
LPT . . . . . . . . . . . . . . . . Low pressure turbine
MIF . . . . . . . . . . . . . . . . Maintenance inflation factor
MLH . . . . . . . . . . . . . . . Medium-long-haul
MR . . . . . . . . . . . . . . . . . Mature-run
V
Contents VI
MRO . . . . . . . . . . . . . . . Maintenance, Repair and Overhaul

OAT . . . . . . . . . . . . . . . . Outside air temperature
OEM . . . . . . . . . . . . . . . Original equipment manufacturer
PM . . . . . . . . . . . . . . . . . Parts manufacturer approval
PPI . . . . . . . . . . . . . . . . . Producer price index
SEC . . . . . . . . . . . . . . . . Spare engine costs
SF . . . . . . . . . . . . . . . . . . Severity factor
SH . . . . . . . . . . . . . . . . . Short-haul
SLOATL . . . . . . . . . . . Sea level outside air temperature limit
SV . . . . . . . . . . . . . . . . . Shop visit
SVC . . . . . . . . . . . . . . . . Shop visit costs
SVR . . . . . . . . . . . . . . . . Shop visit rate
SVRC . . . . . . . . . . . . . . Shop visit restoration costs
TE . . . . . . . . . . . . . . . . . Temperate environment
TIT . . . . . . . . . . . . . . . . Turbine inlet temperature
TMF . . . . . . . . . . . . . . . Time & material factor
TOC . . . . . . . . . . . . . . . Total operating costs
TOW . . . . . . . . . . . . . . . Time On-Wing
TOW . . . . . . . . . . . . . . . Time on-wing
TSF . . . . . . . . . . . . . . . . Three-spool factor
TSFC . . . . . . . . . . . . . . Thrust specific fuel consumption
TWR . . . . . . . . . . . . . . . Thrust-weight ratio
USD . . . . . . . . . . . . . . . . US Dollar
WPG . . . . . . . . . . . . . . . Workscope planning guide
List of Symbols
Symbol Unit Description
BaseInterval EF H Base interval between shop visits determined

through CER
U SD
BaseSV RC EF H EF H Base SV restoration costs per EFH determined
through CER
U SD
BaseSV R EF H Base shop visit rate determined through
BaseInterval
EF H LC EF H Accumulated engine flight hours throughout the

life cycle
Interval EF H Interval length between shop visits

U SD
LLP Cost EF C Costs per EFC incurred by LLP replacement
U SD
LLP CostEF H EF H Costs per EFH incurred by LLP replacement
SEC U SD Costs incurred by providing spare engines
SV C U SD Total costs for a shop visit

U SD
SV C EF H EF H Total SVC per EFH
SV C LC U SD Accumulated engine shop visit costs throughout

the life cycle
SV s
SV R 1000 EF H Number of shop visits per 1000 EFH
SV RC U SD Total costs for the engine’s performance restora-

tion
U SD
SV RC EF H EF H SV restoration costs per EFH
U til ann EF H Annual utilization of an engine
Y ears LC years Number of years of the engine life cycle
VII
1 Introduction
The global market for passenger and freight air transportation has tremendously grown over
the past decades and it is expected to keep expanding at a high pace. At the same time, the
airlines see themselves in a more competitive market environment, especially with the emerging
number of low-cost-carriers that has marked a turning point in the market structure. In order
to stay competitive, airlines need to constantly seek cost saving potentials. This ambition is
closely linked to evaluating new technologies and their possible contribution to reducing the
long term costs for owning and operating the entire aircraft system throughout its life cycle. A
considerable share of these life cycle costs (LCC) are expenditures for maintenance, repair and
overhaul (MRO) of the individual aircraft systems. The biggest proportion of the aircraft MRO
costs is incurred by the engine (fig. 1.1).
Components MRO
21%
Engine MRO
35%
Line Maintenance
21%
Modifications
9%
Airframe Heavy
MRO
14%
Figure 1.1: Aircraft MRO cost overview [Jet08]
Most aircraft engines used in today’s air transportation industry are gas turbine engines.
The mechanical complexity of these engines results in considerable labour costs required for
MRO related tasks such as disassembly, inspection, reassembly and test. In addition, the engine
design requires highly tensile and thermo resistant materials, which results in high material costs
for repair and replacement of worn parts. Therefore, engine MRO is considered as cost driver
and it is in the interest of aircraft operators to estimate the life cycle costs caused by engine
maintenance, when making decisions regarding their engine fleet.
1
1.1 Thesis Objectives 2
1.1 Thesis Objectives

The objective of this thesis is to develop a model that is capable of predicting the MRO costs
of commercial jet engines, using the method of parametric cost modeling based on available
historic data. The focus of the model is supposed to lie on the engine MRO that is performed
in regular intervals off-wing in dedicated engine workshops. Therefore, it should also enable
the estimation of engine maintenance intervals. This work is intended to complement that of
Schilling [Sch09], which established a method for considering the various maintenance events of
an aircraft life cycle as part of a LCC simulation tool realized in Matlab. The newly developed
model is supposed to elaborate the existing consideration of engine maintenance as part of
the LCC maintenance module. The LCC-tool enables the evaluation of technologies under the
incorporation of expertise in form of technology factors. Therefore, the model is not aimed to
forecast accurate figures of engine MRO costs and intervals. Rather, it is intended to qualitatively
reflect the general influence factors of engine maintenance and estimate reasonable cost and
interval figures accordingly.
In support of the development of the engine MRO model, this thesis is meant to serve as a review
on basic concepts and relationships in jet engine maintenance. The aim of this review is a better
understanding of the decisive characteristics that affect the maintenance costs of aircraft engines.
1.2 Thesis Structure

To meet the specific objectives, this thesis is structured into three major sections.
First Section The first section (chapter 2) includes a brief introduction to gas turbine engines in
general followed by an extensive literature review on engine maintenance and parametric cost
modeling.
Second Section The second part (chapters 3 and 4) describes the process of developing cost
estimating relationships using a methodical approach introduced in the first section. Therefore,
an adequate database is assembled based on available data sources. The resulting database is then
statistically analyzed in order to establish valid cost estimating relationships. Building on these
relationships and based on the conclusions of the literature review, the considerations that led to
the final engine maintenance model are subsequently described including a demonstration of an
example application. Finally the model is checked for its plausibility using additionally available
independent data and its sensitivity on changes of different input parameters is illustrated.
Third Section The third section (chapter 5) revolves around implementing the developed model
into the existing LCC simulation tool. The model is implemented in an independent sub module
that is integrated in the existing programme sequence. However, the implementation required a
few minor adjustments to the input file as well as to the tool structure itself. All changes are
respectively documented.
2 Literature Review
This chapter is an summary of the reviewed literature and provides a theoretical background for
the present study. It is structured in three sections. First an overview of the working principle
and the composition of aircraft gas turbine engines is given, followed by an analysis of the various
aspects in engine maintenance. The last section reviews relevant concepts for the modeling of
engine maintenance costs.
2.1 Aircraft Gas Turbine Engines

The first successful application of gas turbines engines for powering an aircraft in 1939 was
preceded by a long development time. Simple gas turbines have been used as windmills since
ancient times. However, it was not before the industrial revolution in the 19th century that first
attempts were made to use gas turbines for propulsion. As a result of this early gas turbine
engines were developed, shortly after the first successful engine driven flight in the beginning of
the 20th century. Since then, the development grew rapidly. Bill Gunston [Gun95] describes in
detail the evolution of the turbine aero engine from the early prototypes over the first applications
in second world war to modern civil and military aircraft engines. Today, there are several
different kinds of gas turbine engines, all sharing the same basic engine core principle [Tew07]. A
brief description of this engine core and its basic principles will be addressed in the next sections
along with a closer look at the turbofan engine, which is by far the most common engine used in
today’s civil aviation. Since the cost estimation model as result of this work is based on cost
information from turbofan maintenance, it is important to have a rough picture of their layout
and composition.
2.1.1 Basic Principles

The gas turbine engine is an internal combustion engine based on the following process: a
continuous flow of air is sucked in an inlet, densified by a compressor, heated up by burning fuel
in the combustion chamber and eventually leaves through a turbine. The compressor and the
turbine are placed on one shaft and is sometimes referred to as a spool. Therefore the compressor
is powered by using a part of the kinetic energy from the hot compressed air, that escapes
through the turbine. Gas turbine engines also handle the working fluid in a smooth continuous
flow and each part of the working cycle takes place simultaneously in a different part of the
engine, unlike in piston engines. This basic configuration consisting of a compressor, a combus-
tion chamber as well as a turbine is the engine core, which all gas turbine engines have in common.
3
2.1 Aircraft Gas Turbine Engines 4
fuel
2 3
1 4 The Brayto
air combustion exhaust q
pressure
chamber
2 3
com
ex
pa
pres
ns
ion
sio
n
W 1 4
volume
compressor turbine
Figure 2.1: Engine core of gas turbine engines
The working cycle of such gas turbines is called the Brayton cycle. The cycle efficiency depends
EGT [C°]
on the achieved temperature ratio T3 /T1 as well as on the given pressure ratio p2 /p1 . A closer
Engine DMC [$/EFH]

fuel Redline EGT
High Cost
look on the thermal and cycle efficiencies of gas turbine engines is given in [Cum97].
due to low
EGT Margin
3 The Brayton Cycle
utilization
4 Take-Off EGT
mbustion exhaust q q
pressure
pressure
hamber
2 3 2 3 Target
TOW
com
ex
pa
pres
ns
Corner Point SLOATL Engine T

ion
OAT [C°]
sio
n
W 1 4 1 4
volume temperature
Redline EGT The engine becomes less
EGT [C°]
turbine Figure 2.2: Working cycle of a gas turbine engine efficient, due to wear of
compressor/turbine blades
EGT Margin
There are two ways of generating thrust from this working
Deteriorated Engine principle. On the of
The loss one hand, has
efficiency the
hot high pressure airflow leaving the turbine can be accelerated to high speedto by
be compensated by
a nozzle behind
an increased fuel burn
the turbine. This is called jet propulsion. The other approach is to utilize the kinetic energy of
Engine DMC [$/EFH]
the air flow

Redline EGT mainly for providing shaft power to drive a propeller. In such aThe
New Engine setincrease
up, there isburn
in fuel still
High Cost results in a higher EGT
an exhaust air flow that
due tocan contribute to the over all thrust. However, the
low main share of the
Increasing cost
utilization
propulsion is generated by a turbine driven propeller or fan.to
due Whereas
extendedjet propulsion is based on
OAT [C°]
Take-Off EGTa relatively small mass of air at high speed, the
accelerating workscopes
propeller accelerates a large mass of
air at much lower speed [Bur97]. Several gas turbine engine types have developed out of this
EGTM Erosion [C°]
working principle. These types are briefly outlines

Target in the following:
TOW EGT Limit
Turbojet The turbojet is the earliest and simplest type of all gas turbine engines [Gun95]. It
SLOATL OAT [C°] of an air-intake an engine core asEngine
consists Time
described On-Wing
above and a nozzle to accelerate the exhaust
air flow. The turbojet is a pure jet propulsion engine, thus the turbine is supposed to extract
Restoration
just enough energy from the gas flow to drive the compressor,
Installation in Shop Visitso that as much energy as possible
Loss
Redline EGT is left in the The
flowengine
to form the propulsive
becomes less jet. The turbojet engine provides a great amount of
efficient, due to wear of
Engine Flight Cycles Engine Time
eteriorated Engine The loss of efficiency has

to be compensated by
thrust at high speed and high altitude, but has the disadvantage of low thrust at low forward
speeds (i.e. take-off).
Turboprop Turboprop engines consist of the engine core like the turbojet but with the addition
of a propeller output reduction gear and a second propeller shaft, which makes this engine type
more complicated and heavier than other gas turbine engines. Unlike turbojets it is the aim to
extract all the energy from the gas flow in the turbine and convert it into shaft power for driving
a propeller. Hence the turboprop generates only a small amount of jet propulsion. Turboprop
engines are characterized by a high propulsive efficiency at low airspeeds. The engine is therefore
able to develop very high thrust at take-off. However, together with the propulsive efficiency the
thrust falls rapidly at speeds above 800 km/h.
Turbofan Like the other gas turbine engines, the heart of the turbofan engine is the core turbine.
In addition, it has a duct-enclosed fan which is usually mounted on the front of the engine. The
air entering the engine passes through the fan and splits into two separated air streams. The core
stream provides the working fluid for the combustion cycle, whereas the second stream bypasses
the engine core, hence its name, bypass airflow. In the following section, the turbofan engine is
discussed in some detail.
Propfan A recent development of gas turbine engines is a combination of the Turbofan and the
Turboprop. The propfan, also known as ultra-high-bypass- or open rotor jet engine, is featured
by an unducted propeller of radically different design to conventional propellers [OCE91]. There
are two types of propfans, one with the propeller module in the front of the engine and one at
the rear of the turbine module. This design is said to result in a very low fuel consumption at
high sub-sonic speeds.
2.1.2 The Turbofan Engine

As mentioned earlier, the turbofan is the main engine used in today’s commercial aircraft. The
development of these engines began early, almost simultaneously with the first turbojet engines in
the 1930’s [Gun95]. In order to improve the low efficiency of the turbojet at take-off and subsonic
speeds, it was proposed to use extra turbine shaft power to drive a bigger compressor (fan), that
could accelerate a way bigger airflow at lower speeds. This low speed airflow is bypassed around
the engine core and usually expands in a separate nozzle at the outlet. The result of this concept
is the turbofan or bypass engine. It is usually realized in a multiple-spool configuration. This
allows the big fan to rotate independently from the compressor stages of the engine core. The
turbofan can be seen as a compromise between turbojet and turboprop [FAA04]. The turbofan
is featured by an increased thrust at low forward speeds, similar to a turboprop. However, its
thrust is not penalized with increasing airspeed up to about Mach 1. The thrust specific fuel
consumption (TSFC) as well as the specific weight of the turbofan engine fall between turbojet
and turbofan. Noteworthy is also the considerable lower noise level of turbofans due to the
low exhaust velocity of the bypassed airflow. The sum of these features makes the turbofan
applicable for a wider range of aircraft in comparison to turbojet and turboprop engines. A
more detailed comparison of the three gas turbine engines and their characteristics along with
descriptive diagrams that show the correlation between net thrust as well as TSFC and the
airspeed at sea level and in high altitude is given in [Tre79].
bypass airflow
core airflow
source: GE
[modified]
Engine Core
Intake Fan LPC LPT Nozzle
[HPC - Combustor - HPT]
Figure 2.3: GEnx-2B - high bypass twin-spool turbofan
The amount of air that is bypassed in relation to the airflow going through the engine core
is called bypass-ratio (BPR). Turbofans can be distinguished in low and high-bypass engines.
The former have a BRP in the range of 0.2:1 to 1:1 and can be found in super-sonic combat
aircraft due to their fuel economy at high speeds [Hue03]. Engines with a BPR of 5:1 and
more are termed high bypass-ratio engines. Today they practically make up all engines in high
sub-sonic military and civil aircraft. Similar to a turboprop, most of the total thrust of high
bypass turbofan engines is produced by the bypass air accelerated in the fan stage, whereas the
engine core primarily acts as gas generator providing the power to drive the turbines. Generally
speaking, a higher BPR leads to a reduced TSFC. However with increasing BPR also the size
and weight of the engine rise. As a result of this, the BPR is somewhat limited by factors like
available ground clearance under the wing or tolerable stress in the wing structure.
The design of conventional turbofan engines can also be distinguished between two-spool and
three-spool configurations. In the more common two-spool turbofan, the low-pressure compressor
(LPC) stages and the fan stages are mounted on one shaft together with the low-pressure turbine
(LPT). The second shaft is hollow and contains the high-pressure compressor (HPC) as well as
the high-pressure turbine (HPT). In order to reach higher bypass-ratios in an effort to reduce
the fuel consumption, fan diameters of turbofan engines have been steadily increased over the
last decades. However, the tip speed of fan blades is limited to less than supersonic, due to
material constraints. Since the fan and LPC of two-spool turbofans are places on the same shaft,
the rotational speed of the LPC is limited to the tolerable revolutions per minute (RPMs) of
the fan. Therefore, the LPC and HPC need a relatively high number of stages to achieve high
compression ratios. The three-spool concept was developed to overcome this issue. The fan and
the first compressor module in the core engine are mounted on different shafts so that they can
turn at different RPMs. The result of this is an intermediate-pressure compressor (IPC) that
can turn faster then the LPC of a two-spool engine. Since the compressors turn at higher RPMs,
they also require less stages, which leads to a shorter and lighter design and generally to more
durability than two-spool engines of the same size. The configuration with the fan, IPC and
HPC each placed on different shafts and driven by a dedicated turbine also makes it possible
that each spool can turn at optimized velocity. The disadvantage of the three-spool design is the
complexity of its construction [Air08a].
LPC + Fan Turbines Turbines

IPC
1 Introduction
HPC HPC
Fan
two-spool three-spool
1.1 Engine Systems in General
Figure 2.4: Comparison: two- and three-spool configuration
A turbine engine consists of its main components, which change the state
of the gas flow in the sequence of the thermodynamic working cycle. The
2.1.2.1 Layout and Module
design ofCharacteristics
modern turbofan engines follows a modular concept. Thus a
typical twin-spool turbofan engine, like the V2500-A5 shown in Fig. 1.1,
The design of today’sis turbofan
composed of the following
engines main
follows modules: concept. This modular design essentially
a modular
reflects maintenance•aspects.
Fan module
Each of the modules has its own identity, service history and specific
• Low pressure compressor module
inspection schedules.• During a shop
Core engine or gasvisit, any of the individual modules can be removed from the
generator
engine as an entire •unit
Lowwithout
pressure turbine module
disassembling it into its piece parts. Figure 2.5 illustrates the
• Accessory gearbox module
modular structure of aThe
typical two-spool
core engine consiststurbofan
of the highengine
pressure(IAE V2500-A5).
compressor, A short description
the combus-
of each of this maintion section and
modules the high
is given pressure turbine. This modular design of the en-
below.
gine mainly reflects maintenance aspects. During engine disassembly each
Fan Module Low Pressure Core Engine Low Pressure
Compressor Module Module Turbine Module
Accessory Gearbox Module
Fig.Figure
1.1 The main
2.5:modules of a V2500-A5
The main modules of a V2500-A5 [Lin08]
Fan The fan is simply a specialized type of a compressor and usually contains one stage. The
fan draws air into the engine, compressing the bypass airflow to produce most of the engines
thrust and supplying air to the gas turbine core. The fan module consists of the fan disk with
fan blades mounted to the low-pressure shaft. Today’s fan blades and disks are made of titanium
alloys. However, more and more blades of newer generation models are also made of carbon fibre
reinforced plastics (CFRP) [Rol07].
Low-Pressure Compressor The main purpose of the LPC is to increase the pressure of the air
through the gas turbine core. In this example, the LPC module contains not only the low-
pressure compressor case and stages, but also the fan case. Large civil gas turbine engines that
are considered in this paper have axial-flow compressors. That means the air is compressed in a
direction parallel to the engine axis. An axial-flow compressor is made up of alternating stages
of rotating blades and static vanes. In order to achieve a high pressure rise, the compression is
spread over a number of stages. Today’s LPC blades and vanes are generally made of aluminum
alloys [Cum97].
Core Engine The core engine module consists of the inner casings, a high-pressure shaft, a
high-pressure compressor, the combustion system as well as the high-pressure turbine. The HPC
is used in conjunction with the LPC and also contains alternating stages of rotor blades and
stator vanes, which further compress the air before it is supplied to the combustor. It is especially
the later stages of a HPC that handle an airflow at considerable higher temperature and pressure,
which is why the blades and vanes are made of more temperature resisting titanium and nickel
alloys. In the combustion system, fuel is burnt with the air received from the compressor modules,
sending hot gas downstream to the HPT. It consists of a combustion chamber, a fuel injector, an
igniter and nozzle guide vanes. The following HPT is made up of one or more turbine rotors as
well as a set of stationary nozzle guide vanes. The HPT converts part of the energy stored within
the hot gas into kinetic energy to drive the HPC and the accessory gearbox. Both combustor
and HPT are exposed to the maximum temperatures that occur in the engine therefore, cooling
air and ceramic coated nickel alloys are used to increase component lives. Generally, in a running
engine, it is the core engine module that is subjected to the most compelling conditions in terms
of temperature, pressure and rotational velocity. Thus, it will be the module that suffers the
fastest deterioration of performance [Ack10].
Low-Pressure Turbine The LPT module is located in the rear of the engine downstream of the
HPT module. It is an assembly of disks with turbine blades that are attached to the low pressure
shaft, nozzle guide vanes and a rear frame. The LPT removes the remaining energy from the
combustion gases to power the LPC module.
Accessory Gearbox The accessory gearbox is attached to the bottom or side of the engine.
The aircraft engine not only provides thrust, but it also supplies power for engine and aircraft
accessories. This includes starters, fuel and oil pumps as well as hydraulic pumps and generators
for cabin power. The accessory gearbox is where all this mechanical-driven components are
mounted to the engine.
This compound of main engine modules can be referred to as basic engine. However, this basic
engine by itself is not operable and cannot serve all necessary functions. In addition to its main
components the engine needs various systems to become operable. These engine-related systems
include amongst others an air cooling and sealing system, a lubrication system, a fuel distribution
system, an exhaust and thrust reverser system as well as an air inlet and a nozzle [Lin08].
2.1.2.2 Engine Operating Parameters
Modern aircraft are equipped with a multitude of gauges to provide the flight crew with feedback
information about the engine condition. The main operating parameters contain the speeds of
the engine spools and the engine pressure ratio (EPR) for performance monitoring, as well as
the temperatures of the turbine gases for health monitoring. A brief description of these key
operating parameters is given below:
N1 and N2/N3 speeds In a jet engine, every main revolving section has a separate gauge to
monitor its RPMs. Depending on the engine type, the N1-gauge keeps track of the LPC and/or
fan speed. The core section is monitored by the N2-gauge, whereas a three-spool engine has
an additional N3-gauge. Due to the high revolving velocities, the RPMs of the engine spools
are displayed as percentage of the design RPM rather than actual RPM. The N1-speed is the
primary indication of thrust on most turbofans [FAA04].
Engine Pressure Ratio (EPR) The EPR is the total pressure ratio across the engine and is defined
as the ratio of the pressure at turbine exit (exhaust) to the pressure at the intake. On some
turbofans, it serves as primary thrust indication gauge.
Turbine Inlet Temperature (TIT) The TIT is the gas temperature from the combustor exit as it
enters the first HPT stage. As the highest temperature inside a gas turbine engine, the TIT is
one of the limiting factors for the power output of an engine. However, it is difficult to measure
therefore, the exhaust gas temperature (EGT) is usually the parameter measured.
Exhaust Gas Temperature (EGT) The EGT is the temperature of the exhaust gases as they enter
the tail pipe, after passing through the LPT. It is expressed in degrees centigrade and can be
seen as one of the most important health monitoring parameters. The engine gas temperatures
have to be closely monitored, as exceeding temperature limits may lead to serious heat damage
to the turbine components [FAA04]. In addition, the EGT is a measure of the engine’s efficiency
in producing its design level of thrust. A high EGT may indicate that the engine has suffered
significant hardware deterioration during service. Generally, the EGT reaches its maximum
during take-off or right after lift-off, as the engine operates here at its peak.
EGT Margin (EGTM) fuel of an engine’s performance degradation,

In order to measure the level
the so called EGT margin has been introduced. The EGT margin of an engine is the difference
2 3
between the maximum tolerable1EGT (Redline EGT) and the peak
4 EGT during take-off. This The B
redline EGT is the absolute
air temperature limit, which cannot be exceeded
combustion without damaging the
exhaust q
pressure
engine [Bra04]. Therefore, the EGT margin ischamber
a measure for how well below this limit the engine
2 3
operates in times of maximum power output at take-off. As the EGT of an engine increases over
com
ex
pa
time, due to hardware deterioration, the EGT margin decreases. Theoretically, an engine can
pres
ns
ion
sio
remain on wing until its EGT margin has become zero. It is normally at its highest level when
n
the engine is new or has just been refurbished. W 1 4
The EGT margin is furthermore highly influenced by the present outside air temperature (OAT). volume
For a given thrust setting, the EGT rises at a constant rate
compressor as the OAT increases. Figure 2.6
turbine
shows the relationship between take-off EGT and OAT.
EGT [C°]
Engine DMC [$/EFH]

Redline EGT
High Cost
EGT Margin due to low
utilization
Take-Off EGT
Corner Point SLOATL OAT [C°] En
Figure 2.6: Correlation between Take-Off EGT and OAT [Air06b]
Redlineschedule
The pictured curve is a result of the power management EGT of the digitalThe
engine controller
engine becomes less
EGT [C°]
efficient,
(FADEC). It is programmed to provide constant maximum thrust with increasing OAT. due
Asto the
wear of
OAT rises, the air density EGT Margin Therefore, the throttle has to be increased in order to
decreases.
Deteriorated
maintain constant thrust, which results in an Engine
increase in EGT. However, constant maximum
The loss of efficiency has
thrust is only maintained up to a certain OAT (corner point). The FADEC is then programmed
to keep the EGT constant for OATs higher then the corner point temperature. This power
management setting is called flat rating and makes
Newsure that the engine operates
Engine with enough
The increase in fuel burn
EGT margin also at high OATs. The constant EGT is maintained by reducing results in a higher
the engine thrustEGT
as the OAT rises beyond the corner point [Air06b]. Without flat rating, the EGT would continue
OAT
to rise with increasing OAT as the dashed line in fig. 2.6 [C°] The OAT at which the EGT
indicates.
would reach the redline EGT, if maximum take-off thrust was maintained is termed sea level
outside air temperature limit (SLOATL). The actual highest permitted thrust setting for a given
EGTM Erosion [C°]
OAT can be determined by calculating the SLOATL.

EGT Limit
Since the EGT margin is the main indicator for an engine’s health status, it is normally expressed
independently from the OAT. That means the EGT margin is given as the difference between
Restoration
Installation in Shop Visit
Loss
2.2 Aircraft Engine Maintenance 11
redline EGT and the actual EGT at maximum thrust at the corner point OAT (fig.2.6).
2.2 Aircraft Engine Maintenance

This section is supposed to give an overview over the different aspects and in engine maintenance.
The aircraft engine as a major airplane component, in terms of investment, operating cost as
well as its complexity, follows its own maintenance schedule mainly independently from the
regular maintenance check events of the remainder aircraft. For a more detailed discussion of the
maintenance programme of an aircraft, it is referred to [Sch09]. Engine maintenance is a broad
field that includes not only engineering but also various complex planning and management
problems. As such, it is often also referred to as engine management.
Modern engine maintenance is based on the so called on-condition method. After this method,
engines removals and overhauls only take place when the engine condition demands it [Rup00].
Whereas in the past, engines were removed and maintained after a fixed time interval (hard-time
interval), which had the disadvantage of engines being removed, even in case of a safe operating
engine. Similar to the maintenance of the remainder aircraft, engine maintenance can be divided
into on-wing and off-wing maintenance or overhaul, in the following referred to as shop visit
(SV). The next two subsections will discuss both maintenance components. Subsequently, engine
time on-wing and maintenance cost are discussed separately, due to their importance for the
proposed model.
2.2.1 On-Wing Engine Maintenance

On-wing engine maintenance, also known as engine line maintenance, includes all maintenance
and inspection activities that can be done without engine removal and disassembly on the flight
line. As such it is generally included into the line maintenance schedule of the aircraft operation.
As as result of the on-condition maintenance concept, a great share of on-wing maintenance
activities involves Engine Condition Monitoring. The aim is to monitor and analyze the main
operating parameters as well as the internal physical condition of the engine, in order to identify
potential problems before they become serious and to provide data that can be used to determine
the most economic times for engine shop visits. A second group of tasks can be summarized
as On-Wing Repair and Replacement. In the recent past, more and more actions have been
developed to access the site of engine damage directly on-wing and without complete disassembly.
As a result of this, more engine problems can be fixed on-wing, which significantly extends the
time on-wing (TOW) of the engine [Bur10]. The following is a more detailed discussion on both
kinds of on-wing maintenance.
2.2.1.1 Engine Condition Monitoring (ECM)
Today’s ECM systems evolved as a result of aviation authorities requiring flight crews to monitor
basic engine performance parameters from the flight deck instruments. The recorded data was
then used by the engineering departments of the airlines to determine the maintenance programme
for the engine. Therefore, ECM data was historically recorded manually and only during take-off
and once in the cruise. On modern aircraft, ECM information is gathered automatically in
higher number and quality and can even be recorded and transmitted to a ground station in
real-time [Air05a]. The engine performance parameters that are measured can be divided into
two categories. The first consists of parameters that are not heavily influenced by flight conditions
and engine thrust, like engine vibrations as well as oil temperature and pressure. The second type
of parameters comprise those that are affected by flight conditions and thrust. These parameters
include the gas path temperatures like the EGT, EPR, fuel flow as well as the N1 and N2 speeds.
In order to also provide data for indications of the present flight conditions, parameters like
altitude Mach number and air temperature are measured and recorded as well [Air05a]. The
key objective of ECM is to plot the performance trend data, so that it can be compared to a
model of how the engine is expected to behave under the experienced flight conditions. Shifts in
performance indicate hardware deterioration or operational problems. Combinations of specific
parameter changes are known to be indications for specific deviations in the engine. The data
can be further interpreted to find out which part of the engine is inducing the problems. This
analysis of the recorded data is undertaken by specialized ECM software usually provided by
the original equipment manufactures (OEMs). It is expected that future ECM systems will
capture more accurate data and have more elaborated data interpretation capabilities than the
current generation [Air06c]. In addition to the recording and analysis of engine performance
data, ECM also includes monitoring the physical condition of internal engine parts with the help
of inspection borescopes. An inspection borescope is an optical diagnosis tool comprising of a
long flexible tube and an optical lense, that gives an magnified and illuminated view of hardly
accessible areas inside the engine. It allows to inspect internal engine parts for defects such as
cracks, stress fractures and corrosion.
To sum up, ECM allows the concept of on-condition maintenance of aircraft engines. It helps
to manage the timing of both scheduled and unscheduled shop visits and it prevents excessive
hardware deterioration and it provides initial alerts that allow engines to be fixed on-wing
[Air05a].
2.2.1.2 On-Wing Repair and Replacement
Aircraft engines usually have a design life that exceeds the achieved actual shop visit intervals by
far. This is due to part failures and unexpected damages. For instance, the fan and LPT modules
are often the first areas to suffer environmental damage due to their exposure to birds and
debris [Bur10]. The ECM systems described above, are able to detect such problems and provide
information that help the maintenance engineers to decide if an on-wing repair or replacement
should be conducted. Together with ECM systems, the on-wing repair capabilities are getting
more and more sophisticated. Today, on-wing maintenance includes repairs that historically have
been high-cost shop repairs [Bur10]. However, as a result of progressive hardware deterioration,
an engine overhaul is eventually unavoidable. On-wing repair though, contributes to extend
the engine’s time on-wing as close to its design life as possible, despite unexpected failures or
damages. Also falling into this category is the replacement of line replaceable units (LRUs).
These are parts that are designed to be quickly replaced on the flight line. They are usually
sealed units like, sensors, pumps, filters or tanks and can be replaced independently from their
surroundings.
On-wing repair and replacement not only has the benefit that it saves the time and money
for engine removal and complete disassembly, but also that there is no need for a spare engine
in order to keep the aircraft in service. In addition, it can be included into the aircraft’s line
maintenance schedule. GE’s On Wing Support for instance, performs flight line repairs like
borescope blending of compressor blades, fan module and gearbox workscopes as well as top case
compressor repairs [GE 08]. Another technique that falls into this category is the so called engine
water wash. It can be done without requiring additional ground time and involves spraying about
a hundred liters of water repeatedly into the front of the turning but not burning engine, where
the engine cleans itself. This procedure reduces fuel consumption by improving the EGT margin
and therefore extends the on-wing intervals [KLM07].
2.2.2 Engine Overhaul - Shop Visit

Despite better on-wing maintenance technology, eventually every engine has to be removed
from wing and disassembled in order to get more extensive maintenance. Airlines usually have
access to a pool of spare engines. Therefore, the removed engine is immediately replaced by a
spare engine for the time the engine is being refurbished. Depending on the engine type, the
replacement can be performed by 3-4 mechanics in a full 8-hour shift. Thus, the engine removal
can be conducted on the flight line. The overhaul is performed in a dedicated engine workshop,
hence its name engine shop visit (ESV). This subsection discusses the main causes that force an
engine to be removed and overhauled followed by a brief description of the shop visit process
and shop visit management considerations.
2.2.2.1 Main Causes of Engine Removals
As previously mentioned, a modern aircraft engine’s condition is constantly monitored. This

allows one to predict the time when the engine has deteriorated to a level where an engine
overhaul becomes necessary. That means an ESV is generally a scheduled event that repeats in
regular predictable intervals. However, especially in situations where an unexpected part failure
or damage cant be fixed by on-wing maintenance efforts, an engine has to be removed and sent
to the workshop prematurely. The primary engine removal causes can be categorized in four
groups: EGT margin degradation, expiry of Life Limited Parts (LLPs), hardware deterioration
and other unscheduled removal causes. [Air00]. The causes of engine removals depend heavily on
the type of aircraft operation. Engines operating on short-haul routes show a higher percentage
of removals caused by EGT margin degradation and LLP expiry, while medium- and long-haul
operating engines tend to have a higher share of removals due to hardware deterioration and
EGTM degradation. The distribution of the engine removals on the removal causes depending
on the aircraft operation and the engine age status is illustrated in figure 2.7. In the following
fuel
2
2.2 Aircraft Engine Maintenance 3 14
1 4 The Brayton Cycle
air combustion exhaust
parapraphs the primary removal causes are briefly discussed. q
pressure
pressure
chamber
2 3
com
ex
pa
pres
ns
ion
sio
n
W 1 4 1
volume
compressor turbine
Figure 2.7: Removal causes depending on aircraft operation [Ack10]

EGT [C°]
EGT Margin Degradation The degradation of an engine’s EGTM is one of the dominating
Engine DMC [$/EFH]

Redline EGT
High
removal driver. EGT margin degradation is a result of the gradual wear Costcompressor and
of the
EGT Margin due to low Inc
turbine blades. This leads to a rising clearance between the blade tips and the surrounding
utilization due
shrouds and thus, to an increasing leakage of the Take-Off
working EGT
fluid in the compressor and turbine w
stages [Air08c]. Such leakage causes a decrease in the overall engine efficiency and performance.
Therefore, a deteriorated engine has to burn more fuel than a new engine in order to achieve the
Target
same required thrust level. As the engine wear continues, the EGT will ultimately rise until theTOW
EGT margin is so little that severe damage to the turbine components cannot be excluded. In
this case, the engine needs
Corner to get aSLOATL
Point performanceOAT
restoration
[C°] in order to remain operable. Figure
Engine Time On
2.8 illustrates the effects of engine wear on the EGT margin.

EGT [C°]

EGT Margin
Deteriorated Engine The loss of efficiency has
New Engine The increase in fuel burn

results in a higher EGT
OAT [C°]
Figure 2.8: Effects of engine wear on the EGT Margin [Ack10]

EGTM Erosion [C°]
The rate of EGT margin erosion normally depends on the thrust rating as well as on how the
EGT Limit on the factors that influence the EGT margin $/ESV
engine is operated. A more detailed discussion
erosion rate is given in the following section 2.2.3. Generally, the rates of EGT margin degradation
are highest during the first 1000 - 2000 engine flight cycles (EFC) after installation. This is
called Installation Loss. The erosion rate stabilizes thereafter and reaches a fairly constant
level until the the engine is removedRestoration
[Air05b]. In the following engine shop visit, the EGT
Installation in Shop Visit $/EFH
margin is restored.
LossHowever, it is normally not possible to fully restore the initial EGT margin
Engine Flight Cycles Engine Time On-Wi

EGT [C°]
2.2 Aircraft Engine Maintenance 15 blades
compressor/turbine
EGT Margin
of a new engine. Overhauled engines typically achieve 70%-80% of the initial to EGT margin
be compensated by
an increased
level. This contributes to the fact that an engine’s first TOW is usually the longest achievedfuel burn
during its life cycle. Figure 2.9 shows the qualitative relationship between EGT margin erosion
New Engine
and accumulated EFC’s. Since an engine’s EGTM is constantly monitored, it isThe increase in fuel burn
fairly easy to
determine its actual stabilized EGTM degradation rates. Therefore, it is possible to predict the
point where the EGT margin becomes zero. In other words,
OATEGT margin degradation generally
[C°]
leads to a scheduled ESV.
EGTM Erosion [C°]
EGT Limit
Restoration
Installation in Shop Visit
Loss
Engine Flight Cycles En

Figure 2.9: Trend of EGT margin erosion rates over accumulated EFC
Life Limited Part Expiry Life limited parts (LLPs) are defined as engine rotor and major static
structural parts whose failure could result in hazardous engine effects. Such effects include for
instance uncontrolled fire or complete inability to shut down the engine. The Advisory Circular
(AC) 33.70 [FAA09], issued by the FAA, regulates the standards for the design and testing of
engine LLPs. Life limited rotational parts include disks, spools, spacers and shafts, whereas
static structural parts generally include high-pressure cases and non-redundant engine mount
components. For each LLP, an operating limitation or life limit must be established in order to
ensure that no hazardous effect occurs. The life limit specifies the maximum number of finite
flights or engine flight cycles (EFC) a LLP is allowed to be in service. The life limit for rotating
parts is for example equal to the minimum number of EFC, that is required to trigger a crack of
about 7 mm in length by 3.5 mm in depth [FAA09]. The definition of maximum permissible life
times for certain engine parts is an exception from the general on-condition maintenance concept
for aircraft engines. The life limit of LLPs is defined by the engine manufactures and typically
ranges between 15,000-30,000 EFC. However, some individual parts can have restricted lives,
due to technical issues and imposed airworthiness directives (AD) [Air04a]. Once an engine has
accumulated as many EFC as the shortest life limit of all equipped LLPs, it has to be removed
and sent to a workshop in order to replace the used up LLPs. Hence, LLP replacement can be
scheduled in coordination with the expected point of full EGT margin erosion. LLP replacement
is a major cost driver in engine maintenance and as such it is subjected to several cost saving
measures.
Hardware Deterioration The third main removal cause is the deterioration of the engine hardware.
All engine components are exposed to different kinds of deterioration mechanisms. These include
amongst others, low and high cycle fatigue, thermo-mechanical fatigue as well as corrosion
[MM10]. These mechanisms lead to a degradation of the part lives or in worst case to a part
failure as well as to a loss of engine performance. In contradiction to LLPs, the remainder engine
parts are replaced on an on-condition basis. A safely operating engine therefore relies on a well
functioning ECM system to detect problems related to hardware deterioration. The engine’s
core module, being exposed to the highest temperatures and revolving velocities within the
engine, suffers in particular from the mentioned deterioration mechanisms. These engine failures
are practically not predictable but modern ECM systems are capable of detecting them soon
after they arise and therefore allow to prevent more severe damage. However, if such an engine
hardware deterioration problem cannot be fixed on-wing, it forces the engine to get a premature
unscheduled shop visit.
Other removal causes The last group includes unscheduled removal causes from foreign object
damage (FOD), engine system failures and engine vibration. FOD is engine damage resulting
from ingestion of foreign objects. Foreign objects include birds, ice or ash as well as a runway
debris [Ack10]. Especially the ingestion of larger objects like birds can lead to significant damage
of the fan- and the LPC blades. However, such an incident usually does not affect the safe
outcome of a flight and may not even be noticed by the flight crew [Tur04]. But the ingestion of
foreign objects poses a risk of latent effects, like minor cracks that can propagate by progressive
engine wear. Hence, it is important that the ECM is able to detect the occurrence of FOD. Also
falling into this category is engine system failures. Especially lubrication system problems, such
as leaks or oil pump malfunctions can result in severe engine damage if they are not fixed.
2.2.2.2 Engine Shop Visit Process
Engine shop visits can generally be categorized by the extend of the conducted workscopes and
the number of modules on which work is performed. The level of the shop visit has a strong
impact on the following removal interval. An extensive shop visit results in a significantly longer
time on-wing before the next shop visit becomes due. As mentioned in 2.1.2.1, the individual
engine modules are considered independently during a shop visit. This is necessary because
they have different rates of deterioration. Therefore each module normally requires different
workscopes at each shop visit [Air09]. The workscopes are typically performed in dedicated shop
departments, that are specialized on a certain module. It is also not unusual that engine shops
outsource the overhaul of individual modules or parts to shops with more capabilities in this
particular field. Shop visits that include work on all modules can last up to 50-90 days [ACT08].
A shop visit process following the general incoming inspection and the definition of the objected
workscopes is illustrated in figure 2.10. A detailed analysis of the different shop visit stages for
each of the main engine modules is given in [Air09].
Engine Engine
Disassembly Testing & Release
• Disassembly into
• Testing of the engine
engine modules
on-wing or in test cell
• Full disassembly of
• Updating of engine
modules into piece
documents
parts if necessary
GEnx-2B
Parts Repair & Certification &

Inspection Replacement Reassembly
• Cleaning of parts • Certification of
• Part repair in special
• non-destructive tests refurbished parts
workshops
for detecting cracks • Balancing of rotables
• Replacement of non-
• Dimensional checks • Reassembly of
repairable parts
on blades and vanes module units
Figure 2.10: Engine shop visit process
fuel Incoming
2.2.2.3 Shop Visit Management
Inspection
3 The primary objective of shop visit management
• The Brayton is to Cycle
minimize the long-term engine direct
4 Inspection of engine
exterior and interior
maintenance cost (DMC), expressed in cost per flight hour (USD/EFH) [Ack10]. A more detailed
combustion exhaust • Analysis of engines
breakdown of the composition of theqDMC service
of an engine follows underq2.2.4. A dominant factor
history
pressure
pressure
chamber
2 3 • Workscope definiton 2 3
that influences the engine maintenance cost is the time on-wing between shop visits. An increasing
com
Folie 4
ex
TOW will result in increased engine deterioration [Air07a]. Hence, the engine modulesVortrag
require
pa
> Autor > Dokumentname > Datum

pres
ns
ion
more extensive maintenance at each shop visit. This leads to rising shop visit cost as the TOW
sio
n
increases. W
Generally, the raised shop1visit costs 4are compensated
1 by the extended
4 removal interval,
due to the increased TOW. As a resultvolume
of this, the overall maintenance
temperature costs per flight hour
decrease. However, the increase in engine deterioration accelerates after a certain TOW, with
turbine
the result that the required shop visit workscopes extend so severely that the shop visit cost
are raised to a level, where the overall USD/EFH begin to increase again [Air07a]. Figure 2.11
illustrates this relationship between the engine’s TOW and the DMC per flight hour.
Engine DMC [$/EFH]
Redline EGT
High Cost
due to low Increasing cost
utilization due to extended
Take-Off EGT workscopes
Target
TOW
nt SLOATL OAT [C°] Engine Time On-Wing
Figure 2.11: Influences of the TOW on the DMC of an engine [Eng10]


The aim is to find a balance between shop visit cost and time on-wing, so that the lowest
cost per EFH are achieved. This is a challenging task, which requires thoughtful shop visit
management. Four key considerations concerning the shop visit management are subsequently
discussed:
Workscope Planning The level of workscope to be performed on a module or an individual

item is proposed by the workscope planning guide (WPG) , issued by the engine manufacturer
[Air09]. However, it also heavily depends on the result of the module inspection, the removal
cause, the time on-wing (TOW) since the last shop visit, as well as the extend of previous
shop visit workscopes [Air07a]. According to the set operational and economic engine build
targets, the workscope plan of a shop visit can individually deviate from the shop visit manual
[Jet08]. That means the performed workscopes can be adjusted to meet certain target on-wing
times or target shop visit costs. In general, there are three levels of workscopes: Minimum
Level, Performance Level and Full Overhaul. A full overhaul on a module involves a complete
disassembly to piece-part level and an inspection as well as repair or replacement of all parts.
Lighter workscopes, like performance restorations, usually require just partial disassembly and
repair works only on certain items [Ack10].
LLP Management The management of the LLPs is essential in minimising maintenance cost,
particularly for engines used on short- and medium-haul operations [Air04b]. In long-haul engines,
LLPs account for a smaller share of total maintenance cost. This is because LLPs have fixed
lives defined in engine flight cycles (EFC). Therefore they can last for many years in long-haul
operating engines, due to the low number of flight cycles (FC) these engines accumulate per year.
Short-haul engines on the other hand, accumulate a considerably more FC each year. Hence,
their LLPs have to be replaced every few years [Air04a]. As stated in 2.2.2.1 LLP life expiry is a
main removal cause that forces an engine into a shop visit. The task of LLP management is to
coordinate the remaining lives of the equipped LLPs with other criteria for shop visit timing.
Ideally, the LLP replacement would coincide with the optimal TOW as illustrated in Fig. 2.11
as well as with the date of full EGT deterioration. However, these events may occur at a time
when some LLPs still have a few thousand EFC left. If not replaced, this remaining life, also
called “stub life”, would limit the subsequent removal interval. In order to prevent an early next
shop visit, these LLPs have to be replaced and scrapped without utilizing all their available life.
However, wasting remaining LLP life raises the average cost per flight hour [Air04b]. It becomes
apparent that a compromise between utilization of the LLP lives and optimal time on-wing has to
be found in order to achieve the lowest long-term maintenance cost. In short-haul engines, where
LLP expiry is the main removal cause and their frequent replacement is a major cost driver, it
is common that the shop visit workscopes are tailored around the LLP lives, so that the shop
visits coincide with LLP replacement [Air99]. Another key consideration is that LLPs should
ideally be replaced during a heavier shop visit, when the engine has gone through a high level of
disassembly [Air04a]. This is because replacing LLPs also requires a high degree of disassembly
and reassembly. Man-hours for assembly works account for a large percentage of shop visit
cost. LLP replacement during a light shop visit would increase the necessary workscope and
therefore the cost. The lowest maintenance cost per EFH is accomplished when a heavy shop
visit coincides with full LLP utilization [Air04b].
PMA Parts PMA stands for Parts Manufacturer Approval and is a combined design and
production approval for modification and replacement parts. A manufacturer who holds the
PMA is allowed to produce and sell FAA approved parts that are eligible for installation on type
certificated aircraft [FAA10]. Aerospace original equipment manufacturers (OEMs) usually have
a strong monopoly position on replacement parts [Ben08]. This results in generally high prices
for original parts in the aftermarket. The OEMs legitimate the high margins on their parts with
the high level of investment in research and development, required to market a modern aircraft
engine. Parts from approved manufacturers are considerably cheaper and the savings potential
can range between 45%-75% compared to OEM pricing [Air06b]. With the OEMs continuing
to impose yearly price increases of five to six percent for spare parts [Ben08], the PMA market
is expected to carry on growing. However, the engine manufacturers are striving to strengthen
their position, by restricting technical support and warranty coverage when PMA parts are used
in their engines. Yet, the legal status of PMA parts is clear. According to FAA regulations, PMA
parts can be seen as equivalent to parts from the original equipment manufacturers [Hol08].
Parts Repair The increasing development of repairs for engine parts has considerably contributed
to the reduction of the overall engine maintenance cost. The two key objectives of parts repair are
to maintain the on-wing life and at the same time, reduce the shop visit costs [Air07a]. Repair
costs have to be considered against the costs of used parts on the surplus market and new parts
either from PMA or from the original parts manufacturer. Most repairs have been developed for
airfoils such as blades and vanes, since they have the largest economic impact. But repairs are
also developed for other parts that prove to be a major maintenance cost driver as the engine
ages. In general, it takes some time and investment for repairs to be developed for new engine
models. Hence, repairs are rather available for mature engines [Air07a]. Another consideration
is that repaired parts tend to lead to shorter subsequent on-wing life in comparison to engines
where parts were replaced. Generally, the reduction in costs makes up for the shorter following
shop visit interval and so the overall cost per EFH are lower when utilizing repairs. Repairing
parts costs five to 10 times less than replacing them [Air99]. However, the quality of repairs and
their effect on the remainder engine has to be considered, since poor quality repairs may result in
unscheduled removals. On most repairable parts, two repairs can be performed before they are
scrapped at the third removal. In doing so, the costs for the second repair will be higher than
the first. This also has to be considered when planning the parts repair strategy for a shop visit.
2.2.3 Engine Time On-Wing

The achieved time on-wing is an important parameter in engine maintenance. Until the engine
reaches its target TOW, a longer TOW generally leads to reduced over all shop visit DMC per
EFH (see figure 2.11). However, engines are often forced into shop visits before reaching this
target time on-wing. This happens not only because of unexpected engine damages but also due
to accelerated performance deterioration. The previous section demonstrated how shop visit
management influences the engine TOW. This section discusses the influence factors on the
hardware & EGTM deterioration rates. These factors resulting from the engine built and the
operation conditions also heavily influence the engine TOW. They are summarized as follows:
• Engine Thrust Rating
• Operational Severity
• Engine Age
2.2.3.1 Engine Thrust Rating
Normally, there are several thrust ratings for a given engine model. The CFM56-7B for instance,
comes in six different engine variants, all rated at different thrusts. The basic engine build is the
same, the rating is the result of the power control setting of an engine. The engine variants with
a higher thrust1 level generate higher gas path temperatures [Air05b]. This results in a lower
EGT margin and normally also in a more severe EGT and hardware deterioration, due to the
increased thermal stress. Low initial EGT margin and high EGT margin erosion rates will lead
quickly to complete EGT margin degradation. That means high rated engines are more likely to
be forced to get a shop visit because of full EGT margin deterioration than lower rated ones.
Table 2.1 sums up the different engine variants and their EGT related parameters.
Engine Variant -7B18 -7B20 -7B22 -7B24 -7B26 -7B27
Thrust [lbs] 19,500 20,600 22,700 24,200 26,300 27,300
Initial EGT mar- 125-130 125-130 105 100 60 55

gin [degC]
EGTM erosion 2.5-4.0 2.5-4.0 2.5-4.0 4.0-6.0 4.0-6.0 4.0-6.0
[degC/1000EFC]
Table 2.1: Initial EGTM and mature EGT erosion rates for CFM56-7B variants [Air08c]
As a result of the more severe hardware deterioration, higher rated engines generally also tend
to achieve shorter times on-wing than engines variants with low thrust ratings.
2.2.3.2 Operational Severity
Furthermore an engine’s time on-wing is heavily influenced by its operating conditions. More
demanding conditions will result in greater stress on the engine and therefore increase the wear
of the engine hardware. The major parameters of operating severity include:
1 thrust as engine specification generally means the maximum take-off thrust

• Average Flight Time
• Take-Off Derate
• Outside Air Temperature
• Environment
Average Flight Time The measurement of time length an engine is operating on-wing can be
quantified in both engine flight hours (EFH) and engine flight cycles (EFC) . A flight hour
represents one hour of flight, whereas a flight cycle represents one sequence of take-off, cruise and
landing. Both measurements can be frequently found in the literature. However, the number of
accumulated EFC is generally the more appropriate measure for engines operating on short-haul
routes, while the time on-wing of engines that are operated on medium and long flight times
should be considered in terms of EFH [Air08b]. During one flight cycle, it is the take-off and
climb phase, where the engine is exposed to the greatest thermal stress and engine wear. The
engine hardware deteriorates considerably less during the following cruise and landing. Therefore,
the number of accumulated EFH is not a representative time measure for an engine operated on
short cycle lengths. Rather it is the number EFC in service, that is an indicator for the engine
wear. On the other hand, for medium and long-haul operating engines the accumulated EFH has
made its way as a common time on-wing measure in terms of maintenance. Figure 2.12 compares
the flight profiles of a short-haul and a medium-haul aircraft. The flight profile of an aircraft can
be expressed by the flight hour to flight cycle ratio (FH:FC), also known as the flight leg length
or flight time. The average FC:FH is an important parameter for the operational severity of an
engine. In general, an engine that operates on a short average flight time will suffer a more rapid
performance deterioration and therefore has shorter shop visit intervals and higher DMC per
flight hour. Conversely, as the the FC:FH increases the engine is exposed to less wear and can
remain longer on wing with reduced USD/EFH. The mean time on-wing between shop visits
is often represented by the shop visit rate (SVR) . This characteristic is analogue to the direct
maintenance cost per flight hour (USD/EFH) in terms of operational severity. It is defined as
Short-Haul Operation FH:FC = 1.0

De
b
Cruise
Clim
sc
SVR & DMC [$/EFH]

ent
ff &
&L
e-o
and
Tak
1 flight hour 1 flight hour 1 flight hour

i ng
Medium-Haul Operation FH:FC = 3.0

De
limb
Cruise
sc
&C
ent
&
ff
e-o
Lan
Tak
3 flight hours
din
short
g
cycle time c
Figure 2.12: Two example flight profiles [Ack10]
EGT [C°]
1.6
rity Factor
Redline EGT
EGT Margin 1.4
Increasing Dera
Take-Off EGT 1.2
the number of shop visits per 1000 EFH for a given engine [SV/1000 EFH]. A mean TOW of
17000 EFH is for instance equal to a SVR of 1000/17000 = 0.0588. Equation (2.1) describes the
conversion from shop visit intervals into SVRs.
1000
SV R = (2.1)
ShopV isitInterval
More severe operation conditions will lead to an increase of both, SVR and DMC. The qualitative
relationship between EFH:EFC and the SVR as well as the DMC is illustrated in figure 2.13.
eration FH:FC = 1.0
SVR & DMC [$/EFH]
flight hour 1 flight hour
Operation FH:FC = 3.0

De
Cruise
scent
&
Lan
ight hours
din
short long
g
EFH:EFC [h]
cycle time cycle time
Figure 2.13: Shop visit rate and DMC in relation to the flight hour flight cycle ratio
1.6
Take-Off Derate
EGT Another issue that influences the operational severity is the manual reduction
Severity Factor
Redline
of the maximum thrust
1.4 at take-off. This is usually referred to as take-off derate. Derating the
Increasing Derate = Lower Severity
engine during take-off is in the discretion of the pilot. It ranges between 0-20%, and typically
Take-Off EGT 1.2
falls between 10-15% [Air04c]. Derating is used when the take-off weight is below the maximum
Base Point
take-off weight of the1.0
aircraft, a long runway is available or the ambient temperatures during
Derate
5% and thus a reduced
take-off are relatively low [Air06c]. The result is a lower EGT at take-off
0.8 10%
rate of engine deterioration and prolonged time on-wing. In general,15%
engines that operate on
short average flight times
1.0 benefit1.5
more from 2.0 take-off 2.5
derate than
3.0 those operated on long-haul
nt SLOATL OAT [C°] EFH:EFC [h]
routes. Also, it is generally that the first 5% of derate have a bigger impact in terms of reducing
the operational severity than following derate steps of 5% [Air06d].
Since both the average

Redline EGT flightbecomes
The engine time and the level of derate heavily affect the engine’s deterioration,
less
efficient,
engine manufacturers due severity
develop to wear ofcurves to illustrate the combined influence of both param-
eters on the severity of an engine’s operation [Air06d]. These curves are collected from statistical
distributions andThe
Deteriorated Engine allow
lossthe operatorshas
of efficiency to perform benchmarking and sensitivity studies in order
to be
to achieve the lowest compensated
cost by Severity curves are often developed separately for each
of ownership.
engine variant of an engine model [Ack10]. Figure 2.14 illustrates an example severity curve.
OAT [C°]
$/ESV
De
SVR &
Cruise
sc
ent
& Lan
light hours
din
short long
g
2.2 Aircraft Engine Maintenance EFH:EFC [h] 23
cycle time cycle time
1.6
Severity Factor
Redline EGT
1.4
Increasing Derate = Lower Severity
Take-Off EGT 1.2
Base Point
1.0 Derate
5%
0.8 10%
15%
1.0 1.5 2.0 2.5 3.0
int SLOATL OAT [C°] EFH:EFC [h]
Figure 2.14: Example severity curve
Redline EGTEach of the multiple

The engine becomes
curves in theless
graph represents one derate level as stated at the right.
These multiple severity curves areblades
compressor/turbine often also expressed in form of a matrix. The output of a
severity curve is a certain severity factor (SF) , that is used to adjust the SVR or the maintenance
Deteriorated Engine
cost according to The
the loss of efficiency
operational has The following calculation based on the severity curve
severity.
illustrated above will demonstrate
an increased this concept.
fuel burn
An example engine is operated at an average flight time of EFC:EFH = 2 and an average

derate of 5%. Under thisinbase
results conditions
a higher EGT the SVR is known to be 0.05 [SVs/1000 EFH] and
the maintenance cost per flight hour are 100 [USD/EFH]. To estimate how the SVR and the
[C°]
OAT maintenance cost change when the same engine is operated at a flight time of EFC:EFH = 3.0
and a derate of 10%, the operator can determine the severity factor for these two parameters
and subsequently adjust the base cost and SVR.
$/ESV
Base engine severity: EF H : EF C = 2.0 [h], Derate = 5%
SV s U SD
Base data → SV R = 0.05 and Cost = 100
1000 EF H EF H
Adjustment to deviating severity: EF H : EF C = 3.0 [h], Derate = 10%

Restoration
in Shop Visit 1. Determination of severity factor via severity curve →
$/EFH SF = 0.8
2. Multiplication of SF with base SVR and Cost
Engine Flight Cycles EngineSV
results in → SV R = 0.8 · 0.05 = 0.04
s On-Wing
Time and Cost = 0.8 · 100 = 80
U SD
1000 EF H EF H
Therefore the operator can expect 20% reduced SVR and cost, if the engine is operated under
less severe conditions as proposed in the example.
Outside Air Temperature As demonstrated in figure 2.6, the EGT during take-off is directly
influenced by the ambient air temperatures. In order to prevent the engine from operating at
EGTs that could result in severe damage, the digital engine control keeps the EGT and EGT
margin constant at all OATs above the corner point temperature, by reducing the engine’s thrust.
However, at OATs below the corner point, the thrust is kept constant and the available EGT
margin increases as the OAT decreases [Air07b]. In other words, low ambient air temperatures
result in low gas path temperatures, which reduces the thermal stress on the engine’s hardware
and thus prolongs engine time on-wing.
Environment Also contributing to the severity of an engines operation are environmental

conditions. Particulate matter that results from air pollution, such as dust, sand, industry
emissions or volcanic ash can erode compressor and turbine blades as well as block cooling
holes. Salty environments in coastal areas will accelerate corrosion and oxidation of the engine
components [Jet08]. These environmental conditions can have a severe impact on the engine’s
hardware deterioration and thus, on the time on-wing.
2.2.3.3 Engine Age
A general observation from analyzing engine maintenance data is that older engines remain
on-wing shorter and cost more to maintain than newer engines. In terms of maintenance, engines
can therefore be distinguished in first-run and mature-run phases. There is no clear definition
when an engine’s mature phase starts. Maturity may begin as early as after the first shop visit,
depending on the engine model. In general, first-run engines will achieve considerably longer
times on-wing than subsequent runs, as a result of increasing rates of hardware deterioration as
the engine ages. However, once the engine reaches maturity, the shop visits intervals and cost
stabilize to a relatively steady state [Ack10]. The engine phase also has a significant influence on
the cause of engine removals as previously seen in figure 2.7.
2.2.4 Engine Maintenance Costs

Prior to this section, several references to the influence factors on engine maintenance costs
(EMC) have been made. This subsection shall further discuss the EMC as part of the total
operating costs (TOC) of an aircraft including a cost breakdown structure as it is considered in
the scope of this paper. Aircraft TOC are generally divided into direct operating costs (DOC) and
indirect operating costs (IOC) [Ros90]. DOC comprise of all costs, that can be clearly allocated
to the aircraft operation including fuel costs, crew costs and maintenance, while IOC consist of all
general costs indirectly related to the aircraft operation, like costs for planning and organization
as well as marketing and ticket sales. A more detailed cost breakdown structure of aircraft TOC
can be found in [Pet08]. Engine maintenance costs are generally considered to be part of the
direct operating costs. Modeling the DOC of an aircraft has historically been an important tool
to evaluate the economics of an aircraft design. Thus, there are various different DOC methods
that include EMC as part of the total maintenance costs. Many DOC methods are based on
the ATA 1967 DOC method [Air67]. An example is the DOC method after Roskam [Ros90],
where the EMC per flight hour comprise three components: labour, material and maintenance
burden costs. The division of maintenance cost in labour- and material costs can be found in
most cost estimation publications reviewed in the scope of this project. The maintenance burden
are allowances to reflect general charges for facilities, spare engines and engine leasing, training
of the staff or general engineering and administrative services related to engine maintenance.
Therefore, some DOC methods consider these charges as indirect operating costs and do not
include maintenance burden, as in [Sch98] or [AEG00]. In this context, engine maintenance costs
are often also divided into direct maintenance cost (DMC) and indirect maintenance costs (IMC).
The focus lies in general on the estimation of the DMC since they are the more meaningful
benchmark for comparing two engine designs and they are directly influenced by the aircraft
operation [CFM09]. Section 2.3.1 analyzes the reflection of EMC within common DOC models
in more detail.
2.2.4.1 Cost Breakdown Structure of Engine Maintenance
The cost breakdown is the foundation of a cost estimating model. The objective is to provide a
reliable structure that includes all elements the cost estimate will cover [NAS08]. The following
structure is based on the DMC breakdown in Ackert [Ack10], expanded to also include indirect
maintenance costs.
Engine Maintenance Costs
Engine DMC Engine IMC
Line DMC Shop Visit DMC Maintenance Burden

• Providing Facilities:
• Labour:
offices, workshops
• Line Labour: assembly/disassembly
Line inspection cleaning, inspection,
• Staff member training
Troubleshooting
• Material: replacement
LRU replacement • Administration:
of parts and material
planning, engineering
• Line Material: • Repair of parts
consumables
• Fees, testing, logistics Spare Engines
• access to spare engines,
leased or owned
Input of Shop Visit DMC

Shop Visit DMC [$/EFH] = Shop Visit Cost (SVC)/TOW
SVC [$] = Restoration Cost + LLP Cost
Figure 2.15: Engine maintenance cost breakdown structure
While engine line maintenance incurs costs continuously as it is performed in relatively short
intervals on the flight line, shop maintenance takes place after comparable long intervals thus
causeing costs only during shop visits. Due to the different character of line maintenance and
shop visit, they are reflected separately within the engine DMC. The engine line maintenance
is included in the aircraft maintenance planning document (MPD). Thus, it has already been
implemented into the existing LCC-tool. The IMC include maintenance burden also known as
2.3 Modeling of Engine Maintenance Cost 26
overhead costs and charges for spare engines.
Figure 2.15 also illustrates an alternative way of accounting for the shop visit DMC. In contrast
to the traditional separation into material and labour costs, it is also common to divide the
shop visit costs (SVC) into restoration costs and LLP costs [Ack10][Bec09]. Restoration costs
are charges for labour and material related to restoring the engine’s performance, while LLP
costs reflect expenditures for the LLP replacement. The shop visit DMC per flight hour are
then calculated by dividing the SVC by the mean time on-wing. This cost breakdown correlates
with the concept of adjusting maintenance cost via severity factors. It splits up the shop visit
cost in one component that depends on the severity and one that is mainly independent from
the operational conditions. Only the restoration cost are escalated according to the operational
severity. The LLPs are replaced after a hard time independently from the severity of the flight
conditions.
2.2.4.2 Common forms of engine maintenance contracts
Whereas many airlines historically performed the overhaul of their engines in their own workshops,
today the engine overhauls of most airlines are contracted to external MRO service providers.
The commissioning of engine overhauls to external shops is based on contracts that contain
all services to be performed. There are generally three basic types of payment methods that
are typically arranged in the contracts [Rup00]. In the so called Time and Material contracts,
the MRO provider issues for each engine a detailed invoice with the required man hours and
materials. The customer pays the bill according to the arranged labour rates and material costs.
In return, the contracted shop guarantees a certain minimum subsequent time on-wing of the
engine. In contrast to this are Fly-By-Hour arrangements, where the airline pays a fixed amount
of money per engine flight hour to the MRO provider. The contracted shop has to finance all
coming shop visits from the received advance payment. This results in a good predictability
of the maintenance cost for the airlines. The third payment method is based on Fixed Prices
for certain shop services to be performed. The arranged workscopes vary from full overhauls to
limited overhauls on certain modules. There are also diverse hybrid forms possible. The airlines
generally try to arrange a customized contract that fits the requirements of the airline and results
in low cost and reliable cost forecasting.
2.3 Modeling of Engine Maintenance Cost

An overview of concepts for modeling the maintenance costs of aircraft engines is given in
the following. Starting with a summary of the reflection of EMC in common DOC methods.
Subsequently the basics of cost estimation with focus on parametric cost modeling are also
reviewed as background for the development for cost estimating relationships (CERs).
2.3.1 Reflection of EMC in DOC methods

The analysis of the consideration of the EMC within different DOC methods is a first reference
on how engine maintenance costs maybe modeled. Therefore, table 2.2 summarizes the reflection
of engine maintenance costs in different DOC models available. For a more explicit comparison
of several DOC methods in general it is referred to Mildt [Car00].
DOC
Cost Breakdown Labour C lab Material C mat Input IMC
Method
C lab = h TT O : take-off thrust,

5.43 · 10−5 · EP ·
h C mat = labRmhr : labour rate,
TT O
Roskam C EM C = C lab + 0.718 + 0.0317 · 1000
·
i MTBR, EP: engine 3
[Ros90] C mat + C burden price, Esppf: spare
· K 1
i Esppf − 0.47
1100 part price factor,
M T BR
+ 0.10 · labRmhr Hem
Khem :MTBR factor
h
0.05·TT O h
C lab = 0.645 · · 0.05·TT O

104 C mat = 25 · ·
AAE 104 TT O : take-off thrust,
C EM C = C lab +
0.434
i i labRmhr : labour rate, 3
[AAE04] C mat + C burden 0.566 + FT
· FT · 0.38 FT: flight time
0.62 + F T · FT
labRmhr
C mat =
h TT O : take-off thrust,
TO T labRmhr : labour rate,
AEG C lab = 3.63 + 0.91 · 1000
7
C EM C = C lab +C mat K2 : material coeffi-
[AEG00] 3.26 · 10−5 · TT O · labRmhr i cient, ECM: economic
· K2 + 5.07 · ECM efficient
C LM,lab =
C LM,M at =
TT O : take-off thrust,

a + FbT + F T c · 2
a + b · TT O + c · TT O ·
labRmhr : labour rate,
p,day
C EM C = Emat a,b,c,d:regression co-
LH C LM,lab + C LM,mat + labRmhr · Elab efficients, Elab,mat : 7
[TB07] C SV,lab + C SV,mat +
C SV,mat = a · fetops ·
cost escalation factor,
C LLP + C N AC C SV,lab = a · fetops · fe tops: cost factor,

b c d
O · BP R · (year − 1970) · SF: severity factor,
b c d
TT
TT O · BP R · (year − 1970) · FT:flight time
Emat · SF
labRmhr · Elab · SF
Table 2.2: Comparison of EMC consideration in different DOC methods
The structure of the DOC methods is quite similar. All consist of cost components for labour
and material adjusted by several regression coefficients and cost factors. However, the LH method
sticks clearly out, due to a higher level of cost breakdown and more complex estimation functions.
It is the only method that considers the costs for line maintenance (LM) and shop visit (SV)
maintenance separately. In addition, there are explicit cost estimation functions for the LLPs
(C LLP ) and the engine nacelle (C N AC ). All models include the take-off thrust as major input
parameter, while the LH method also considers severity factors for engine de-rate and average
flight time and even the Bypass-Ratio (BPR) is reflected. The factor (year − 1970)d models the
influence of the age of the engine design.
2.3.2 Parametric Cost Estimation

Parametric cost estimation is based on historical data and mathematical expressions, that relate
cost as a dependent variable to selected cost-driving independent variables. The result are so
called cost estimating relationships (CERs), which are defined as:
Cost Estimating Relationships (CERs) are mathematical expressions relating cost as

the dependent variable to one or more independent cost driving variables. [Bru96]
methodologies) for the data available to develop the most accurate cost estimate possible.
Based upon the phase that the project/system is entering and the data available to conduct the
estimate, follow the quick reference chart shown in Table 1-3 to select the cost estimating
methodology (or methodologies).
Table 1-3. Cost Estimating Methodology Selection Chart
A typical CER for instance isPre-Phase A

estimating thePhase A Phase B
manufacturing cost of aPhase C/D
product Phase E
by using the
Parametric 4 4 2 2 0
product weight [Lon00]. The implicit assumption is that the future cost of a product are affected
by the same forces that affected the
Analogy 4cost in the 2
past. This approach
2 is generally
2 applied, 0
when
only a few key parameters are known.
2
Therefore,
2
it suits the
4
present problem
4
of estimating
4
Engineering Build Up
the maintenance cost of aircraft engines with limited access to detailed primary data sources.
4 2 0
However, Legend: Primary
the key parameters Applicable
of an engine, such Notthrust,
as weight and Applicable
as well as to a certain
extend the overall maintenance costs are openly available and can be utilized in a parametric
Parametric Cost Estimating
cost analysis. A major advantage is that once the CERs are established, the cost estimates can
Estimates created using a parametric approach are based on historical data and mathematical
be conducted quickly and easily replicated [NAS08], which is necessary for an implementation
expressions relating cost as the dependent variable to selected, independent, cost-driving
into a variable software tool. The problem is that the collection of the necessary data and the
variables through regression analysis. Generally, an estimator selects parametric cost estimating
subsequent
whendetermination
only a few key of CERs
pieces is aare
of data complex
known, and
suchtime consuming
as weight process.
and volume. The scope of
The implicit
this study, however,of is
assumption limited on
parametric costengine maintenance,
estimating which
is that the same makes
forces it possible
that affected to the
cost in apply
past the
will
affect
parametric cost in theapproach.
estimation future. ForThe
example,
NASA NASA
cost cost estimates
estimating are frequently
handbook of space
[NAS08] systems
provides or
further
software. The data that relates to estimates of these are weight characteristics and design
resources on the applicability of different cost estimating methods including a summary of their
complexity respectively. The major advantage of using a parametric methodology is that the
pros andestimate
cons. Figure 2.16 be
can usually shows the methodological
conducted procedure
quickly and be easily of parametric
replicated. Figure 1-12cost estimating.
shows the steps
This procedure is the
associated withfoundation
parametric of the
cost data analysis in chapter 3. Therefore, the key stages are
estimating.
briefly discussed in the subsequent paragraphs.
Define
Estimating Collect
“Hypothesis” “Relationship” Evaluate &
Data Normalize Analyze Data
Data for Candidate
Relationships
Perform
Statistical
(Regression) Test
Analysis Relationships Select Cost
Estimating
Relationship
Figure 1-12. Figure

Parametric2.16: Parametric
Cost Estimating Cost
Process Estimating process steps [NAS08]
Steps
Volume 1 Page 1-27

2.3.2.1 Estimating Hypothesis Definition
The objective of defining an estimating hypothesis is to identify potential cost driving variables
and to propose logical cost relationships. This demands a good understanding of the technical
character and the requirements of the examined project. The result is a hypothesis of a forecasting
model necessary to develop CERs [Bru96].
2.3.2.2 Data Collection and Evaluation
The assembly of a database is essential when deriving cost estimating relationships. A lack of
valid CERs is often the result of an inappropriate database [Bru96]. The first step in building
up a good database is the data collection. There are generally two types of data, cost and
non-cost data. Non-cost data includes technical information coming from drawings, specifications,
certification documents or direct measurement as well as schedule and programmatic information
that can be obtained from operations departments [BJ82]. Cost data comprises labour hours
or direct cost information extracted from accounting systems or through interviews. The data
typically comes from many different sources. It is important that the estimator judges the quality
of the data origin and identifies the best source [Gal08]. Data can be obtained from internal
sources, such as accounting or workshop databases, programme recaps or engineering notes,
as well as from external sources like professional articles or public record informations. The
disadvantage of external sources is that the user has no knowledge of the procedures used to
collect and process the data. It is further distinguished between primary data that is directly
obtained from the original source, and secondary data which is derived and possibly “sanitized”
from primary data. Hence, primary data is generally considered best in quality and reliability
[Gal08].
2.3.2.3 Data Normalization
When establishing a database, it is often discovered that the collected raw data turns out to
be irregular and inconsistent or partly in the wrong format for analytical purposes. Therefore,
adjustments to the raw data have to be made to ensure a comparable and consistent database
[BJ82]. For instance the normalization of raw data adjusts inconsistencies in currencies, mea-
surement units and the scope of the data. Historical data should furthermore be adjusted for
anomalies, improvement in technology as well as inflation. Any kind of adjustment or judgments
used in processing historical data should be fully documented. The data collection, evaluation
and normalization is a fundamental step in generating a parametric cost estimating model. Thus,
a considerable amount of time is devoted to assembling a database [Bru96].
2.3.2.4 Data Analysis
The first stage of the data analysis is screening the database for candidate relationships between
the dependent and independent variables. This process is built on the hypothesis. However, there
may be additional relationships that were not foreseen during establishing hypothesis. Once the
candidate relationships have been established, one can perform a regression analysis to model
the CERs. The objective of regression analysis is to determine the parameters for the function
that fits the set of data best. The data is fit using techniques like:
• Linear Regression: unknown parameters are estimated from the data using linear functions
• Nonlinear Regression: applied for data that is not essentially linear
For CERs, the dependent variable is always the cost to be estimated and the independent variable
will be the cost driver [NAS08]. The dependent variable responds to changes of the cost driver
according to the determined regression function. An example would be the hypothesis that
the cost of a product development maybe driven by the weight of the final product. With this
assumption one could plot the historical data on cost over weight, with the possible result of the
chart in figure 2.17.
cost
r(x) = ax + b
data
linear regression
weight
Figure 2.17: Example data points for cost-weight dependency
In this case, a linear regression has already been performed with the aim to fit a straight line
to the data points. The result is an equation that describes the line, expressed by r(x) = ax + b.
In this CER, x represents the weight and r(x) equals the estimated costs. Often, there are
more than one independent variable, that have an effect on the cost. Multivariate regression is
capable of observing and analyzing the effect of multiple independent variables on the dependent
variable, through the addition of possible explanatory coefficients. Usually a computer software
is used to assist in determining the regression coefficients. For a closer look on the mathematical
background of the different regression methods, in context with parametric cost modeling it is
referred to the parametric cost estimating handbook from the US department of defense [Bru96].
2.3.2.5 Testing the Relationship Results
After the determination of a CER through regression analysis, it is crucial to evaluate and test
the regression results. Therefore, it is necessary to have a look on more than just one criteria
[Chu08]. Only the consideration of a multitude of factors will give the whole picture of the quality
of the CER. Table 2.3 summarizes the key criteria that should be evaluated when reviewing the
quality of a CER. As for the regression analysis, a computer software is widely used to conduct a
quick and reliable determination of the statistical criteria. For further information on the theory
of probability and statistics, it is referred to the statistics ebook from the UCLA [UCL].
Symbol Description Reference Evaluation/Test

Check data range, number of are there enough data points for a represen-
X,Y data observation data points and especially out- tative CER? can the outliers be explained or
liers corrected?
ranges from 0-1, while 1 as the maximum rep-
coefficient of measure for the strength of the
R2 resents the best overall fit of the model to the
determination relationship
data
measure for the contribution similar to R2 , however it only increases if an
R2 adjusted for
R2 adj of additional explanatory coef- added explanatory coefficient considerably im-
degrees of freedom
ficients proves the model
root mean square measures the accuracy of the check if the actual mean deviation of the data
RM SE
error relationship points to the model is acceptable
the lower the p-value, the more significant is
measure for the significance of the hypothesis, a hypothesis is generally re-
p probability value
the hypothesis jected when the p-value is higher than the sig-
nificance level α, which is often equal to 0.05
measure for the validity of generally, a t-ratio above 2 leads to the accep-
t t-ratio adding a particular cost driver tance of the hypothesis, that a cost term adds
variable predictive value to the CER
Table 2.3: Criteria for the evaluation of regression results

3 Development of Cost Estimating Relationships
The purpose of this chapter is to document the process of developing the cost estimating
relationships needed for the engine maintenance model. The objective is to generate cost-to-non-
cost CERs for the two dependent variables:
• Shop Visit Cost per Flight Hour

• Shop Visit Interval
The development procedure follows the methodological approach described in the previous chapter
under 2.3.2. This approach is not a fixed single sequence of steps to be conducted. Rather,
it is as an iteration loop as illustrated in figure 2.16 on page 28. The outcome of each step is
evaluated to determine if the next step can follow or if one has to go back a few stages to start
all over again.
3.1 Database Assembly

As mentioned before, the database (DB) assembly is crucial for the success in developing CERs.
Therefore, most time was spent on collecting and processing the data. The following describes
the proposal of the hypothesis as a starting point for the data collection. Subsequently, different
data sources are reviewed and the process of data normalization is reported.
3.1.1 Establishing the Hypothesis

As a result of the literature review, the technical background regarding aircraft engine maintenance
has been worked out. From this information, the cost estimating hypothesis can be derived. Both
cost and interval length are heavily influenced by operating parameters, such as average flight
time, derate, OAT and environmental conditions. However these parameters are no independent
variables that can be directly allocated to the engine built. Since these operational parameters
still have a great impact on the dependent variables, they are subjected to a normalization of
the collected data. Another influence factor is the thrust rating of an engine. The thrust rating
can be represented by its thrust-weight ratio (TWR) . The weight1 reflects the constant built
of an engine model. The more thrust2 is generated from the hardware, the higher the TWR,
which leads to higher shop visit rate and also higher cost within one engine model range. The
assumption is made that the TWR can also serve as variable to compare the thrust ratings of two
1 within the framework of this study, the “weight” as engine specification always means the dry weight
2 “thrust” as engine specification always refers to the take-off thrust
32
3.1 Database Assembly 33
different engine models. Therefore, the TWR is introduced as first independent variable. From
the DOC methods, it can be derived that the engine take-off thrust TT O is also an important
cost driver. In addition, it is assumed that the engine weight as general measure for the engine
size is directly related particularly to the engine maintenance costs. Table 3.1 sums up these
three independent variables and the corresponding proposed relationships.
Summary of CER Hypothesis
TWR higher TWR ⇒ higher SVR and cost
Take-off Thrust higher thrust ⇒ higher cost
Dry Weight higher weight ⇒ higher cost
Table 3.1: Summary of cost estimating relationship hypothesis
3.1.2 Review of Data Sources

With the hypothesis established, available data sources are reviewed to find the necessary data
for the proposed relationships. The review of the data sources is separated into technical- and
cost data sources.
3.1.2.1 Technical Data
Standard technical specifications of aircraft engines, like take-off thrust and dry weight, are
generally no sensitive data. Hence, they can be be obtained directly from the engine OEMs
(website or specification sheets). This is a primary source and it can be considered as very reliable.
It is also possible to find engine specifications in public databases. This has the advantage that
the data for a wide range of engine models and variants is concentrated in one single source. Two
such single sources are the Database Handbook for Turbofan and Turbojet Engines from Élodie
Roux [Rou07] and the Jet Engine Specification Database from Nathan Meier [Mei05]. Even
though these databases are strictly speaking no primary sources, it can be assumed that the
data is still reliable, since it is simply a summary of the primary source without any deviation.
This assumption was also confirmed by a few random comparison checks. Both sources provide a
huge database for a wide range of engine models and variants and their specifications, including
specifications beyond the engine dry weight and take-off thrust.
3.1.2.2 Cost and Interval Data
In contrast to technical engine data, cost and removal interval information are highly sensitive
and well protected by the MRO providers and airlines. Hence, it was not possible to make
primary sources accessible. However, there is a range of secondary data sources, that were
available in the framework of this study. These sources are briefly described in the following
paragraphs:
Form 41 Databases The US department of transportation1 maintains databases for aircraft

traffic, capacity data and other operational data for air carriers operating to and from the United
States. The database includes monthly data of engine maintenance containing labour, repair and
material costs. However, all costs are given only for the different aircraft models. There is no
indication of the engine model version. Therefore, this database is not adequate for collecting
data for specific engine models and variants [Bec09].
MRO Prospector Aviation Week is a weekly magazine reporting on the aerospace industry. Part
of their portfolio is the MRO Prospector2 , a online database for fleet data and contract details.
In contrast to the Form 41 database, it provides comprehensive tables with cost data for a wide
range of specific engine models. The problem is that, there is no indication of the engine variant.
However, the thrust rating of one engine model can vary considerably. Since thrust and TWR
are key independent variables, it is necessary to have more detailed maintenance cost information
on engine variant level. Furthermore, there are no informations given about the operational
conditions the data is based.
Aircraft Commerce Archive Aircraft Commerce3 is an aviation magazine published every two
months. It provides intelligence on fleet planning, maintenance costs and aircraft leasing for
the commercial aircraft industry. In regular intervals, it publishes detailed operator & owner
guides dedicated to specific engine models. These articles give comprehensive information about
the engine’s shop visit planning, removal causes, hardware degradation, LLP management and
the influence of the operational severity. They also summarize data about shop visit intervals
and costs in clearly represented tables. In appendix B.1 an example table from the magazine is
illustrated. Usually, there are distinct information about each variant within an engine model
range. The maintenance costs are generally expressed in estimated reserves per EFH or EFC, in
which the reserves are distinguished between restoration reserves and LLP reserves. It is also
indicated how the reserves change as the engine ages, by showing distinct reserves for first, second
and third or mature shop visits. The articles are fairly consistent in their structure throughout
the years. The magazine maintains an online archive with articles of the past ten years. This
enables to collect and summarize the data.
The aircraft commerce archive is clearly the best available source for building up the database,
since it provides information for specific engine variants including indications about the operating
conditions. The disadvantage is that all the articles have to be collected and particularly read
in order to get all necessary information. Another issue is the fact that it is mainly unknown
how exactly the data was collected and the reserves estimates were established. On request,
the editorial office of the magazine stated, that the data is gained directly from maintenance
facilities. Also, it was possible to clarify further questions about the data collection through
1 http://www.bts.gov/data_and_statistics/
2 http://mrop.aviationweek.com/
3 http://www.aircraft-commerce.com/
direct correspondence with the editorial staff of Aircraft Commerce. Therefore, it is considered
an adequate secondary source for the database assembly.
AeroStrategy AeroStrategy1 is an aerospace consulting firm providing strategic consulting

services to aerospace clients. Amongst others it provides market estimates on engine SV intervals
and SV restoration costs. Costs for LLPs are unfortunately not included. The respective data
tables are not to be disclosed in public. Therefore, they are not displayed in the framework of
this study. The AeroStrategy data distinguishes between first and mature removals and it also
shows distinct estimates on certain engine variants depending on the aircraft they are equipped
to. Even though it is not clearly stated what flight conditions the estimates are based on, the
information on which aircraft the engines are applied to enables assumptions on the average
flight time the estimation relates to. In sum, the AeroStrategy data tables are less appropriate
for the database assembly. However, they form an adequate independent data source for testing
the plausibility of the targeted model.
3.1.3 Data Collection

The collection of the raw data was done in one single excel table. The cost and interval data
from the Aircraft Commerce Articles (ACA) was collected first, since it is only available for
certain engine models and variants. This data was separated in first-run and mature-run data.
Thus, it was assumed that each engine reaches maturity after the first shop visit. This is a valid
assumption, since the available data indicated that the cost and intervals reached a fairly steady
level already after the first removal. For each engine it was noted what year and month the
respective articles were published. This is necessary for a subsequent normalization of inflation.
In addition, the average flight time, on which the cost and interval estimation of the engines is
based on, was collected. This made it possible to collect the interval data as both, EFC and EFH
with the average flight time as conversion factor. Direct information about the environmental
conditions and the derate could not be extracted. However, from studying the articles, the
assumption was made that the average standard derate for estimating the cost and interval data
equals 10 %. Since the articles also provide detailed information about the LLP management
and the EGT degradation rates of the engines it was considered to collect this data as well. The
idea was that especially the LLP cost, LLP lives and number of LLPs in an engine may have a
detectable relationship with the dependent variables. However, it turned out that the articles do
not report this information consistently. As a result of this, the focus laid on the collection of
the cost and interval data. The LLP reserves were generally collected in USD/EFC, whereas
the restoration reserves were collected in USD/EFH. After the collection of the data from the
ACA, the rows were filled up with the respective engine specifications. The collection of the
technical data was extended to additionally include available engine specifications like pressure
ratio, BPR, fan diameter, engine length and the number of stages in each turbine and compressor.
1 http://www.aerostrategy.com/
The objective was to determine if one of these additional independent variables can significantly
contribute to the CER development. Table 3.2 displays the basic structure of the colltected raw
data table.
Engine Specifiations First Removal Mature Removal
Thrust Weight EFC:EFH Interval Interval Restoration LLP Interval Interval Restoration LLP
...
[lbf] [lb] [h] [EFC] [EFH] Reserves Reserves [EFC] [EFH] Reserves Reserves
. . . . . . . . . . . .
. . . . . . . . . . . .
Table 3.2: Structure of the collected Data
From the previous chapter, it is known that engines designed for short-haul (SH) aircraft have
different maintenance characteristics than medium-long-haul (MLH) operating engines. Because
of this, the DB has been arranged in a way that both engine types are listed in separate groups.
This enables both the combined and separate analysis of the two engine types. Appendix B.2
explains what considerations led to the classification of the database engines into SH and MLH.
3.1.4 Data Normalization

General data inconsistencies caused by the varying presentation of the information in the ACA
were adjusted simultaneously with the collection of the data. However, adjustments of the data
for instance due, to inflation and operational severity have to be done subsequently to the data
collection.
3.1.4.1 Inflation
The reviewed ACA were published over a time span of eight years, which makes fluctuating labour
rates and material prices an issue when comparing cost data. Thus, the cost data is normalized
by adjusting it through inflation factors. In general, material and repair & replacement costs
tend to exhibit a higher price fluctuation. This is mainly because of the increasing application
of more advanced and expensive materials and the generally greater imbalances in supply and
demand for these materials. To account for this, two separate economic indices correlating to
both labour and material are utilized to determine the overall inflation factor for each engine.
These indices are the Employment Cost Index (ECI) for aircraft manufacturing wages & salaries
and the Producer Price Index (PPI) for industrial commodities. The proportion of total engine
maintenance costs is in the order of 35% labour and 65% material [Ack10]. The escalation
year and month were set to May 2010. The economic indices for the escalation month and the
respective base month of the engine data were obtained from the website of the US Bureau of
Labor Statistics1 (BLS) . The method used to calcualte the maintenance inflation factor (MIF)
is expressed in the following formula:
1 http://data.bls.gov/cgi-bin/srgate
ECIesc P P Iesc
M IF = 0.35 · + 0.65 ·
ECIbase P P Ibase
ECIesc = mean ECI of the 3 months prior to escalation month

ECIbase = mean ECI of the 3 months prior to base month of collected engine cost data
P P Iesc = mean PPI of the 3 months prior to escalation month
P P Ibase = mean PPI of the 3 months prior to base month of collected engine cost data
To balance short term fluctuation, each index is averaged over the three months prior to the
actual month. For each engine of the DB, the corresponding MIF was calculated and the LLP
cost and restoration costs were adjusted accordingly.
3.1.4.2 Flight Time
As discussed in 2.2.3.2, both the shop visit DMC and the SVR/mean TOW are influenced by
the average flight time. This effect has to be normalized, if possible. The database contains the
average flight time for all cost and interval estimates. The objective is to normalize this data
to a standard flight time level. Theoretically, this is possible if for each engine of the DB the
corresponding severity curve was available (see figure 2.14). In this case, one could predefine a
standard flight time and calculate for all data points the severity factor that would adjust the cost
and interval data to the level of the standard flight time. The problem is that, each engine model
and even each engine variant has a distinct severity curve. These curves are sensitive information
that could not be obtained from the engine manufacturers. However, it was possible to get
an example curve for a short-haul operating engine (A320) as well as for a medium-long-haul
aircraft engine (Boeing 777). Together with the scattered information on severity factors from
the ACA, it was succeeded in assembling averaged severity curves for both SH engines and for
MLH engines, based on the two example curves. The assumption is made that the entire range
of distinct SH severity curves can be adequately approximated by one averaged severity curve. It
is assumed that the same applies for the group of MLH aircraft engines. This is an assumption
made by the author of this thesis. It results from general observations made while studying
the aircraft commerce archive. The two determined average severity curves are subsequently
displayed in shape of a table for a derate of 10%.
EFH:EFC 0.5 1.0 1.5 1.9 2.5 3.0 4.0 5.0 6.0
Severity Factor 2.40 1.75 1.30 1.00 0.86 0.78 0.706 0.66 0.63
Table 3.3: Determined average SH severity curve for a derate of 10%
EFH:EFC 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0
Severity Factor 2.20 1.70 1.40 1.23 1.08 1.00 0.93 0.88 0.86 0.84 0.82 0.80
Table 3.4: Determined average MLH severity curve for a derate of 10%
Since the LLP replacement is mainly independent from the operational severity and the LLP
reserves were collected in USD/EFC, they were excluded from the flight time normalization.
Therefore, only the restoration cost reserves and removal intervals of the database were adjusted
with these severity factors. Both engine categories have been escalated to the base flight time of
the respective average severity curve (Severity Factor = 1.0). That means the SH engines were
normalized to a flight time of 1.9, whereas the MLH engines were normalized to a flight time of
6.0.
3.1.4.3 LLP Reserves
The deviation of the LLP reserves between the first shop visit and subsequent ones is minor. In
addition, there is no trend detectable. Therefore, the LLP reserves of all available data have
been averaged for each engine. This average LLP reserves serve as basis for the cost analysis.
That means the LLP reserves are not divided into first-run and mature-run data.
3.1.4.4 Remaining Anomalies
According to the previous chapter, there are several other effects which have a significant impact
on the maintenance cost and the time on-wing. Ideally, these effects would be also normalized to
a standard level. The following summarizes these effects and discusses how they were considered
while assembling the database.
Derate The ACA do not clearly state on what derate the estimates are based on. However,
there are often information about the average derate the engines are operated with. Most engines
operate on average with a derate of 10%. Thus, it is assumed that all estimations were based on
this derate. Therefore an adjustment of the data according to the derate level does not apply.
Environment and OAT Since the information about the environment and the OAT are very
scarce, it was not possible to utilize this as foundation for a data normalization. In general, the
estimates reflect a worldwide average and thus it is assumed that all values were established
based on a temperate environment (TE) .
Engine Age The effects of the engine age are already included in the database. The aircraft
commerce guides publish estimated cost reserves and intervals for first and mature shop visits.
This data is represented separately in the database. Therefore it is possible to do a distinct
analysis of first SVs and mature SVs. As mentioned before, this is based on the assumption that
the engines reach maturity in terms of maintenance after the first shop visit. This may vary
especially for engines that have only a few shop visits during their life cycle. However, for these
engines, it is also applicable that the first interval is at most times considerably longer than
subsequent ones.
Improving Technology and Learning Curve When building up a database from historic data, it is
also an issue to consider effects from improved technology and developing know-how. Over the
past two decades, the trend in engine maintenance went to longer intervals and thus reduced
maintenance costs per EFH. This was a result of improvements throughout the maintenance
process such as better materials, ECM, on-wing repairs and also design for durability. These
effects were clearly visible in the reviewed data sources. However, it was not possible to collect
enough data from older engine models for establishing reliable escalation factors. Therefore, the
database only contains maintenance data for newer generation engine designs based on similar
technology1 . However, neither does the DB contain engines from the latest generation, since
most of these engines have not even been through their first shop visit and thus there is no
historic maintenance data available.
For normalizing learning curve effects, it applies the same as for the technology improvement. For
a few engines, there is data accessible indicating that an early engine built has raised SVRs and
costs compared to later revisions of the same engine model. This is due to improvements in both
design and maintenance as the engine model is in service. However, since available information
was limited, a normalization could not be performed. In order to make sure that these two
effects do not influence the data analysis, data of older engine generations and data from newly
marketed engines was marked and excluded from the general data analysis. However, the existing
data can be utilized as input for a determination of technology factors for the maintenance model.
Number of Spools In section 2.1.2, it was discussed that the spool configuration of an engine
influences the achievable shop visit intervals. The database contains only six engines with a
three-spool configuration, all from the same OEM. As with the learning curve and the improving
technology, it was not possible to collect enough information to adjust the intervals to the level of
two spool engines. However, the accessible data confirmed this effect. Therefore, the three-shaft
engines were also marked and excluded in the interval analysis.
3.1.5 Summary
A lot of findings and decisions made while assembling the database were the result of first data
analysis procedures, which were necessary to evaluate the adequacy of the DB. This process is
not reported into detail, in order to keep this documentation clearly presented. The result of the
database assembly is the foundation for the subsequent extensive data analysis. However, one
has to keep in mind that the collected data is based on only one single secondary external source.
As a result of the inconsistency of the presentation of the shop visit data in the ACA and the fact
that the exact manner the data was collected is unknown, a lot of assumptions had to be made
to fit the data to the defined database structure. The DB is divided column by column in engine
specifications, first-run, mature-run and normalized intervals and cost reserves. The rows are
furthermore split after the engine type. Short-haul and medium-long-haul operating engines are
represented separately. The same applies for three-spool engines and older generation engines.
Noteworthy is also that some engine variants appear repeatedly with varying average flight times.
1 most engines of the DB entered the market between 1990-2000

3.2 Data Analysis 40
In appendix B.3, the assembled database is displayed with reduced number of columns. Since
several influential effects have been normalized during the DB assembly, it is important to report
the conditions the final DB is based on. Table 3.5 sums up these base conditions for short-haul
and medium-long-haul engines.
Parameter SH MLH
EFH:EFC 1.9 6.0
Derate 10% 10%
No.Spools 2 2
Environm. TE TE
Table 3.5: Summary of DB base conditions
3.2 Data Analysis

The data analysis is divided into determination of candidate relationships and regression analysis
of the found relationships. Both stages were aided by the extensive use of statistical computer
software1 . The dependent variables shop visit interval and shop visit cost per EFH were
Shop Visit Interval Shop Visit Cost
1 2 3 4 5
First Mature
First Interval Mature Interval Restoration Restoration LLP Reserves
Reserves Reserves
Table 3.6: Summary of the preliminary dependent variables to be analyzed
split according to the structure of the database. The SV intervals consist of intervals for first-
and mature removals, while the SV costs consist of restoration reserves for first- and mature
removals and average LLP reserves for all removals. A further differentiation according to the
engine type is also evaluated. As seen in table 3.6, the minimum number of CERs is therefore
five.
3.2.1 Candidate Relationship Screening

The first step is to determine candidate relationships. This includes the evaluation of a possible
further separation of the dependent variables. The procedure was to do a database screening for
all five dependent variables (see table 3.6). The applied statistic software provides a screening
function, that assists in finding independent variables that significantly contribute to modeling
the analyzed dependent variable. This enables an interpretation of the proposed CERs from the
hypothesis simultaneously to an analysis of the additional engine specifications. Figure illustrates
the results of such a data screening.
1 JMP 8.0 from SAS

Short_cost_Analysis- Screening of 3rd cost adj 2 Pag
3.2 Data Analysisfor 3rd cost adj
Screening 41
Contrasts
Lenth Individual Simultaneous
Term Contrast t-Ratio p-Value p-Value
thrust 78,6617 43,43 <,0001* <,0001*
HPC Stages -7,4830 -4,13 0,0043* 0,0506
BPR -9,4701 -5,23 0,0015* 0,0192*
OPR -13,2651 -7,32 0,0004* 0,0035*
length 5,5572 3,07 0,0169* 0,1589
weight -7,9594 -4,39 0,0032* 0,0386*
LPC Stages -6,2270 -3,44 0,0110* 0,1016
Airflow -1,0819 -0,60 0,5760 1,0000
SFC -0,6221 -0,34 0,7512 1,0000
thrust*thrust 4,2789 * 2,36 0,0341* 0,3368
thrust*HPC Stages -0,1146 * -0,06 0,9522 1,0000
HPC Stages*HPC Stages 4,0165 * 2,22 0,0428* 0,4114
thrust*BPR -1,2031 * -0,66 0,5332 1,0000
HPC Stages*BPR -0,8808 * -0,49 0,6505 1,0000
BPR*BPR 0,4734 * 0,26 0,8099 1,0000
thrust*OPR 1,0578 * 0,58 0,5839 1,0000
HPC Stages*OPR -1,2074 * -0,67 0,5315 1,0000
BPR*OPR -0,8278 * -0,46 0,6706 1,0000
OPR*OPR -3,4446 * -1,90 0,0719 0,6043
Half Normal Plot
Figure 3.1: JMP screening function
90
80 thrust
For each of the independent variables, the respective t-ratio’s and p-values are displayed.
70
This is a first hint of what independent variable could contribute to the model of the analyzed
Absolute Contrast
60
dependent variable. In conjunction with the screening, there is the possibility to conduct a
50
quick regression
40
analysis with the highlighted variables as input. This first regression analysis
enables an
30evaluation of the capability of certain variable combinations to model the dependent
variable. 20
The result of this initial screening was, that all additional engine specifications do not
OPR
provide any
10 valuable contribution
LPCHPC
length
toBPRmodeling the CERs. Another important result was, that
weight
Stages
Stages
thrust*thrust
HPC
OPR*OPRStages*HPC Stages
0 analysis turned out to be more complicated. It was not possible to find acceptable
the interval
-10results for an interval analysis of the entire engine range of the database. Therefore,
regression
0,0 0,5 1,0 1,5 2,0 2,5
the separation of the database in SHQuantile
Half Normal and MLH engines was utilized. The shop visit intervals were
further
Lenth divided in first and mature-run intervals for SH engines and MLH engines. While the
PSE=1,81115
Asterisked terms were forced orthogonal. Analysis is order dependent.
engines with three spool configuration were excluded from the interval analysis, they turned out
P-Values derived from a simulation of 10000 Lenth t ratios.
to beMake
eligible
Modelfor the cost analysis. Older generation engines or engine data from newly marketed
Run Model
engines have been excluded entirely from the data analysis. The initial screening confirmed the
assumption that this data could not be modeled adequately together with the remainder data
points. As visible on figure 3.1, the software also automatically evaluates the significance of
variable combinations. This is an useful function, since it is hard to predict how such variable
combination could contribute to the model.
Shop Visit Interval Shop Visit Cost
1 2 3 4 5 6 7
Medium-Long- Short-Haul Medium-Long- First Mature
Short-Haul
Haul First Mature Haul Mature Restoration Restoration LLP Reserves
First Interval
Interval Interval Interval Reserves Reserves
Table 3.7: Summary of the final dependent variables to be analyzed
Table 3.7 summarizes the seven dependent variables for which the respective cost estimating
relationships are developed in the following section.
3.2 Data Analysis 42
3.2.2 Regression Analysis

With all CERs established, the next step was to find the best fit of valid combinations of
independent variables to the analyzed dependent variable. This stage was also conducted through
applying the JMP 8.0 software. The software does the statistical data analysis independently.
That means it determines for a predefined set of independent variables the prediction function
that results in the best least square fit to the analyzed dependent variable. Promising sets of
independent variables can be derived from the data screening. The best combination of variables
was then simply determined by evaluating and comparing the regression results of different
variable sets. An example output of the regression results established by the used computer
software is shown in figure 3.2. From the initial data screening, the possible independent variables
have been narrowed down
LLP- Fit Least to the main engine specifications: take-off thrust and dry weight as
Squares
well as the ratioResponse
of both. This simplified the process of finding the combinations that result in
LLP Reserves
the best overall fit.
Whole Model
Actual by Predicted Plot
900
800
700
LLP Reserves
600
Actual
500
400
300
200
100
0
0 100 300 500 700 900
LLP Reserves Predicted
P<.0001 RSq=0,95 RMSE=44,13
Summary of Fit
RSquare 0,952884
RSquare Adj 0,950528
Root Mean Square Error 44,13013
Mean of Response 254,4868
Observations (or Sum Wgts) 64
Analysis of Variance
Sum of
Source DF Squares Mean Square F Ratio
Model 3 2363135,5 787712 404,4799
Error 60 116848,1 1947 Prob > F
C. Total 63 2479983,6 <,0001*
Parameter Estimates
Term Estimate Std Error t Ratio Prob>|t|
Intercept -115,3133 13,6202 -8,47 <,0001*
weight 0,0194512 0,007095 2,74 0,0080*
thrust 0,0031206 0,001069 2,92 0,0049*
(weight-8608,78)*(weight-8608,78) 2,6924e-6 3,188e-7 8,44 <,0001*
Residual by Predicted Plot
Figure 3.2: JMP regression results example output
250
200
150
LLP Reserves
Residual
100
50
0
-50
-100
-150
0 100 300 500 700 900
3.3 Results of the Parametric Cost Modeling 43
3.3 Results of the Parametric Cost Modeling

This sections sums up the results in shape of the determined prediction functions separated in
interval and cost CERs. The developed CERs have at most three different components as input.
This also includes combinations of the three remaining independent variables: thrust, weight
and T W R. The units of measurement of these independent variables were adopted from the
database. Their definition is summarized in table 3.8. A regression summary as illustrated in
figure 3.2 for each of the seven CERs is displayed in appendix C.
Input Symbol Unit
thrust lbf pound-force
weight lb pound
lbf
TWR lb
pound-force per pound
Table 3.8: Defined standard units for the input parameters of the CERs
3.3.1 Shop Visit Interval CERs

As previously established, the SV interval as a major dependent variable was divided in four
sub-variables, in order to account for occurring anomalies in the database. The corresponding
CERs are subsequently summarized. From the hypothesis, it was expected that the interval
length mainly depends on the TWR of an engine. This was not confirmed by all four developed
CERs. The first intervals of both, SH and MLH engines are only marginally under the direct
influence of the TWR. However, these two CERs also heavily depend on the input of both, the
engine thrust and the engine weight, so that the hypothesis is still represented. The development
of the interval CERs was based on intervals given in EFH, thus the output of the following CERs
is also defined in EFH as standard interval measurement unit.
3.3.1.1 First Interval for SH-Engines
The first removal intervals for short-haul operating engines can be expressed as a function of
T W R, weight and weight2 :
Interval F R,SH = 68466 − 8267.81904 · T W R − 1.00444 · weight

+ (weight − 5407) · [(weight − 5407) · 0.00012125] (3.1)
3.3.1.2 Mature Interval for SH-Engines
For mature removal intervals the regression analysis showed that the CER is just a function of
the thrust-weight ratio (T W R):
Interval M R,SH = 40684 − 5022.8116 · T W R (3.2)

3.3.1.3 First Interval for MLH-Engines
The determined CER for the first removal interval of medium-long-haul operating engines includes
the dependent variables weight, thrust and weight2 :
Interval F R,M LH = 22539 + 1.4329 · weight − 0.3147 · thrust

+ (thrust − 76305) · [(thrust − 76305) · 0.0000034421] (3.3)
3.3.1.4 Mature Interval for MLH-Engines
Mature intervals for medium-long-haul operating engines are, similar to equation (3.1) reflected
as a function of T W R, weight and weight2 :
Interval M R,M LH = 34415 − 2759.25322 · T W R − 0.36625 · weight

+ (weight − 12072) · [(weight − 12072) · 0.000101795] (3.4)
3.3.2 Shop Visit Cost CERs

As a result of the data source structure, the shop visit costs as second major dependent variable,
have been divided into restoration cost and LLP cost. Since the LLP reserves in the database
are represented in USD/EFC, the LLP CER generates LLP costs given in USD/EFC. The LLP
costs turned out to be fairly stable as the engine ages. Therefore, only the restoration costs have
been furthermore divided in first-run and mature-run costs. The restoration cost CERs were
developed based on shop visit restoration costs given in USD/EFH. Hence, the output of the
restoration costs CERs is also defined in USD/EFH. The developed CERs largely reflect the
predictions resulting from the hypothesis. The three corresponding CERs are presented in the
following.
3.3.2.1 First-Run Restoration Costs
The restoration cost per EFH for first shop visits have been modeled as a function of the thrust.
This CER is valid for the entire range of engines:
SV RC F R, EF H = 7 + 0.002361887 · thrust (3.5)
3.3.2.2 Mature-Run Restoration Costs
The mature shop visit restoration cost were also modeled with the thrust as only input component:
SV RC M R, EF H = 46 + 0.002886118 · thrust (3.6)

3.3.2.3 LLP Costs
Eventually the CER estimating the LLP costs of all engines of the database is a function of
weight, thrust and weight2 :
LLP Cost = − 115 + 0.0194512 · weight + 0.0031206 · thrust

+ (weight − 8608.78125) · [(weight − 8608.78125) · 2.69234 · 10−6 ] (3.7)
4 Modeling of Engine Maintenance Cost
As mentioned earlier, the foundation of a cost estimating model is the cost breakdown, that
includes all relevant costs. The cost breakdown structure from the perspective of this study has
been established in figure 2.15 on page 25. Since the engine line maintenance is already included
in the existing LLC-Tool, the focus here lies on modeling the shop visit costs. Maintenance
costs are normally divided in direct and indirect costs. Therefore, the SV costs can be further
differentiated in SV DMC and SV IMC. Since the collected data reflects calculated prices from
MRO providers, it is assumed that charges for maintenance burden are already included in the
collected cost data. However, the costs for maintaining a spare engine pool or leasing spare
engines for the duration of a shop visit are generally not included. Therefore, spare engine
costs as part of the indirect maintenance costs are considered separately. The objective is to
develop a qualitative maintenance cost model that focuses on estimating the shop visit DMC and
intervals. Charges for spare engines are optionally added based on average leasing rates. The
assembled database and the resulting cost estimating relationships developed in the previous
chapter, predetermine parts of the DMC model structure. However, since the CERs are based on
a normalized database, they do not reflect the impact of major influence factors such as, flight
time, derate, number of spools or environment. These effects have to be modeled subsequent
to the CERs. This chapter summarizes the considerations that led to the final cost estimating
model structure.
4.1 Model Structure

The first step in establishing the model structure is to determine what the input and what
the output parameters are. The objective of the engine maintenance model is to estimate SV
intervals and SV costs. The developed CERs distinguish between first-run and mature-run shop
visits. Thus, there are four output parameters: SV interval and SV costs for each engine phase.
The input parameters depend first of all on the necessary input for the developed CERs.
These parameters are, the engine thrust and the engine weight. Since the interval CERs are
further divided in short-haul and medium/long-haul engines, an additional input parameter that
determines what CER is applied, has to be introduced. This additional parameter was termed
engine application and is considered as an engine specification, since it is a static parameter
linked to the engine variant. All of these input parameters can be derived from the developed
CERs. However, as mentioned above, there are important effects that are not modeled in the
CERs. Therefore, there are more necessary input parameters. These include the number of
engine spools as well as operational factors like flight time, derate and information about the
46
4.1 Model Structure 47
severity of the environment.
Engine Specifications
SV Interval
Weight
FR Removal
Thrust Interval
No. Spools MR Removal

Interval
Application
AC Engine
Maintenance
Cost Model
Engine Utilization
SV Costs
EFH:EFC FR Shop Visit
Cost
Derate
MR Shop Visit
Environment Cost
*FR: First-Run
*MR: Mature-Run
Figure 4.1: Black box of maintenance model
Figure 4.1 illustrates the maintenance model as black box with a summary of all input and
output parameters. With the model input and output established, the inner structure of the black
box can be generated. In order to match the output parameters, the model contains in general
two separate lines: a cost-line and an interval-line. The results of the CERs are normalized
values for costs and intervals, based on the engine’s weight and thrust. Thus, they do not model
any operational severity effects. These effects are modeled in conjunction with the normalized
values from the CERs. Therefore, the inner structure of the model has been splitFolieinto
3 two serial
Vortrag > Autor > Dokumentname > Datum
modules. The first module reflects the developed CERs, while the second represents all additional
effects influencing the shop visit costs and intervals. The two modules are thus termed as follows:
• CER-Module
• Effect-Module
The CER-Module determines normalized base values for the shop visit costs and intervals. These
base values are then adjusted in the Effect-Module with a series of adjustment factors. The
adjustment factors are determined in correspondence to the respective input parameters. The
entire model structure is illustrated in figure 4.2. The two modules are described in some more
detail in the following subsections.
CER-
CER-Module
Base Shop Visit Cost Base Shop Visit Interval
Restoration Cost Short-

Short-Medium Haul Medium-
Medium-Long Haul
LLP Cost Func. FR Cost Func. FR Interval Func. FR Interval Func.

SMHE MLHE
MR Cost Func.
MR Interval Func. MR Interval Func.
SMHE MLHE
FR Rest.Cost [$/EFH] FR Base Interval [EFH]

LLP Cost [$/EFC]
MR Rest.Cost [$/EFH] MR Base Interval [EFH]
Effect-
Effect-Module
Time&Material Factor Three-Spool Factor
EFC:EFH Ratio
Severity Factor
Environment Factor
FR Cost [$/EFH] FR Interval [EFH]
MR Cost [$/EFH] MR Interval [EFH]
Figure 4.2: Inner structure of the cost estimating model
4.1.1 CER-Module
The CER-Module basically comprises of nothing more than the seven CERs developed in the
previous chapter (see blue frames in figure 4.2). With the input of the engine weight and thrust
plus the information if it is a SH or MLH engine, the CER-Module generates five outputs:
Folie 2
• LLP Costs [USD/EFC]
• FR Restoration Costs [USD/EFH]
• MR Restoration Costs [USD/EFH]
• FR Base Interval [EFH]

• MR Base Interval [EFH]
These base outputs are valid only for the normalized conditions on which the CER development
was based (see table 3.5). The adjustment to the operational severity is performed in the following
effect-module. For this adjustment, the determined shop visit intervals as output of the interval
CERs have to be converted into a shop visit rate (SVR). In this instance, the SV interval and
the SVC are represented in analogue measurements (both relating to [1/EFH]). This enables the
adjustment of both values with the same factors.
4.1.2 Effect-Module
The Effect-Module, generates the factors necessary to adjust the base costs and intervals from the
CERs according to the input of the operational severity and the number of engine spools. There
are five factors (red frames in figure 4.2), which are subsequently discussed. The output of the
effect-module are adjusted SV intervals and SVC per EFH divided in first-run and mature-run
shop visits. Since the effect-module merges the LLP costs and the restoration costs from the
CER-module, it generates four outputs:
• FR Shop Visit Costs [USD/EFH]

• MR Shop Visit Costs [USD/EFH]
• FR Shop Visit Interval [EFH]

• MR Shop Visit Interval [EFH]
With this output, the absolute shop visit costs can be calculated through multiplying the SVC
per EFH with the respective SV interval.
4.1.2.1 Severity Factor
The concept of severity factors extracted from severity curves has been introduced in 2.2.3.
Severity factors adjusts restoration costs and shop visit intervals corresponding to the average
flight time and derate under which the engine was operated. It was not possible to obtain the
severity curve of each engine in the DB. Therefore, two average severity curves that approximate
the severity curves of a range of engines, have been developed (see 3.1.4.2). These average
curves were already applied to normalize the flight time of the DB engines (see tables 3.3 and
3.4). This normalization did not include an adjustment of the derate, since the DB entries were
assumed to have a constant derate. However, the average severity curves were developed to also
include multiple curves for each of the common derate levels. The two developed severity curves
are fully illustrated in appendix D. These average curves are now the basis for modeling the
effects of flight time and derate on the restoration costs and shop visit intervals as part of the
effect-module. With the flight time and the derate as input, the severity curve simply gives out
the corresponding severity factor which is then multiplied with the restoration costs and the
interval (as seen in the example calculation on page 23).
4.1.2.2 Time & Material Factor
The time & material factor (TMF) has been introduced to account for the effect, that the
absolute shop visit restoration costs (SVRC) generally increase with increasing TOW. When the
severity factor is applied alone, the absolute SVRC remain constant regardless of the flight time
or derate. This is because the SF adjusts both the interval and the restoration costs per EFH
simultaneously. However, the increased TOW due to raised derate and flight time should result
in increasing SVRC. The time & material factor models this effect. Therefore, one could expect
that the TMF can be expressed similar to the severity factor via multiple curves, only inverted
so that the factor increases with decreasing flight time and derate. Due to lack of data it was not
achieved to develop such multiple curves. However, it was possible to obtain a single example
curve from a contact person in the engine maintenance industry. This curve does not reflect
the impact of the derate. Since the influence of the derate on the TOW is generally less severe
and the accessible data is limited, the contribution of different derate levels was neglected. The
available single curve was considered as basis for developing T&M curves valid for all derates.
As with the severity curves, two curves have been developed. One for all SH engines and one for
all MLH engines. The two curves are subsequently illustrated in shape of a table. A graphical
illustration of the time & material curves can be found in appendix D.
EFH:EFC 0.5 1.0 1.5 1.9 2.5 3.0 4.0 5.0 6.0
T&M Factor 0.90 0.95 0.98 1.00 1.02 1.03 1.04 1.05 1.06
Table 4.1: SH Engine Time & Material factor with respect to the flight time
EFH:EFC 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0
T&M Factor 0.85 0.91 0.94 0.96 0.98 1.00 1.03 1.05 1.07 1.09 1.10 1.11
Table 4.2: MLH Engine Time & Material factor with respect to the flight time
4.1.2.3 Three-Spool Factor
The three-spool factor (TSF) models the extended TOW of engines with a three-spool config-
uration compared to the more common two-spool engines. In general, there was no detailed
additional information on the impact of the three-spool configuration on the achievable SV
intervals accessible. However, since the DB indicates that three-spool engines achieve significant
longer SV intervals, the available data from the assembled DB was enabled to determine a simple
constant factor that models this effect. This factor was determined through averaging the offset
of the original three-spool data points over the generated intervals from the two-spool CERs
with the respective three-spool engine specifications as input. However, it has to be noted that
all three-spool engines of the DB are MLH engines. It is assumed that SH engine are influenced
in a similar manner. The result of the analysis was the following offset factor:
• TSF = 0.7 for three-spool engines

• TSF = 1.0 for two-spool engines
In case the input indicates that the proposed engine is a three-spool engine, the SVRs generated
from the interval CERs are simply multiplied with the TSF = 0.7 to account for the expected
longer TOW of a three-spool configuration. The TSF simply equals 1.0, in case of a standard
two-spool engine. For a qualitative consideration, this simple approach is sufficient to model the
influence of the number of spools.
4.1.2.4 Environment Factor
The environment factor (EF) reflects the impact of the present environmental conditions including
the outside air temperature on engine maintenance. Studying the ACA indicated that the
environment influences the SV intervals and costs considerably. This was also confirmed through
the correspondence with different professionals in the aircraft engine maintenance field. However,
it was difficult to locate clear data on this topic. As guideline for modeling the environmental
impact served a paper from Ackert [Ack10]. Ackert defines three gradual levels of environmental
severity and relates each level to a certain escalation factor. These environment levels and their
correlating EFs are listed in table 4.3. The respective EF is then multiplied with the overall SVR
and SVRC in order to adjust the intervals and costs to the present environmental severity. Table
4.3 also indicates typical regions for each environment level.
Environment EF Typical Regions
Temperate 1.0 North America, Europe, Australia
Hot/Dry 1.1 Middle East, North Africa
Erosive 1.2 Coastal China, SE Asia, India
Table 4.3: Environment factors for different environmental conditions
4.1.2.5 EFC:EFH Ratio
Strictly speaking the EFC:EFH ratio is not a factor that is intended to model a certain influential
effect on engine maintenance. The EFC:EFH ratio, as the reciprocal value of the flight time
(EFH:EFC), translates costs represented in USD/EFC into USD/EFH. The effect module sums
up the LLP costs and the restoration costs in order to obtain one measure for the overall shop
visit costs per EFH. However, the LLP costs are generally given in USD/EFC, whereas the
intervals and the restoration costs are based on EFH. Therefore, the LLP costs have to be
converted into USD/EFH in order to enable the summation of LLP costs and restoration costs.
4.1.3 Spare Engine Charges

Aircraft engines that are removed and sent to the workshop are normally replaced with spare
engines, in order to keep the aircraft in service while its original engines are overhauled. Costs
incurred by either owning or leasing spare engines are generally considered as a cost driver in
engine maintenance. Therefore, the targeted model is supposed to enable the consideration
of spare engine charges. For a qualitative estimation model, it is not practical to model the
expenditures for spare engines into detail. That would require comprehensive information on
the airline’s engine fleet situation and access to spare engine pools as well as on current leasing
market developments. However, in correspondence with one of the product managers of Lufthansa
Technik1 , it was established that a reasonable estimation of spare engine costs can be achieved
through current leasing rates. Leasing rates can vary between 2000-5000 USD/day. This deviation
is not only related to the engine type, but it also heavily relates to the current supply and
demand situation for the respective engine. Therefore, there were no CERs developed that reflect
spare engine leasing rates. In the framework of this study, the spare engine costs are simply
1 http://www.lufthansa-technik.com
4.2 Example Application 52
estimated with a predefined leasing rate and the information of the duration of the shop visit.
Current engine leasing rates are not sensitive information and can be obtained on request from
engine lessors. In case there are no leasing rates accessible it is proposed to assume an average
leasing rate of 3500 USD/day.
A reasonable average shop visit duration is 80 days. Since the developed SV DMC model considers
the workload of each shop visit as equal, this average of 80 days is established as standard shop
visit duration. Therefore, the total spare engine costs (SEC) for an average shop visit yield to:
SEC = LeasingRate · SV duration ≈ 3500 · 80 = 280,000 U SD (4.1)
If more detailed information is available, the two parameters of this simple approach can be
adjusted at all times. The proposed values here give an idea of the dimension of the costs and
will serve as default values of the model.
4.2 Example Application

This section illustrates the functionality of the cost estimating model through an example
calculation that includes all equations necessary to generate the output of the model. The
example is based on the following input parameters:
Engine Specifications
Engine Utilization
Parameter Input
Parameter Input
Thrust [lbf] 79900
EFH:EFC [h] 8.0
Weight [lb] 14545
Derate 10%
No. Spools 2
Environment Temperate
Application MLH
Table 4.5: Input engine utilization

Table 4.4: Input engine specifications
This input relates to a Pratt & Whitney 4077 operated on a long-haul route typical for a
wide-body aircraft like the Boeing 777-200. With this input, the CER-module first determines the
base intervals and costs, which are subsequently adjusted through a series of adjustment factors.
Since the input parameters are already given in the required measurement unit, a conversion of
the units does not apply.
4.2.1 Base Costs and Intervals from CERs

The PW 4077 is classified as MLH engine. Therefore, the equations (3.3) and (3.4) are applied
to determine the base interval lengths:
BaseInterval F R = 22539 + 1.4329 · 14545 − 0.3147 · 79900

+(79900 − 76305) · [(79900 − 76305) · 0.0000034421] (4.2)
= 18500 EF H
79900
BaseInterval M R = 34415 − 2759.25322 · − 0.36625 · 14545
14545
+(14545 − 12072) · [(14545 − 12072) · 0.000101795] (4.3)
= 14700 EF H
The restoration cost for FR and MR shop visits are calculated using the eqs. (3.5) and (3.6):
U SD
BaseSV RC F R, EF H = 7 + 0.002361887 · 79900 = 194 (4.4)
EF H
U SD
BaseSV RC M R, EF H = 46 + 0.002886118 · 79900 = 275 (4.5)
EF H
The LLP cost are eventually determined through equation (3.7):
LLP Cost = −115 + 0.0194512 · 14545 + 0.0031206 · 79900

+(14545 − 8608.78125) · [(14545 − 8608.78125) · 2.69234 · 10−6 ] (4.6)
U SD
= 509
EF C
All intermediate results generated through the CERs are summarized in table 4.6.
U SD U SD U SD
Interval F R [EF H] Interval M R [EF H] SV RC F R, EF H EF H
SV RC M R, EF H EF H
LLP Cost EF C
18500 14700 194 275 509
Table 4.6: Summary of CER results for the example input parameters
4.2.2 Adjustment of Intervals

The adjustment factors of the effect-module are defined so that they relate to the elapsed EFH.
Therefore, the determined intervals of the CERs have to be translated into SVRs. This enables
an analogue application of the adjustment factors for both the SV intervals and SV restoration
costs. With equation (2.1) the two determined intervals from (4.2) and (4.3) are converted to:
1000 1000 SV s
BaseSV R F R = = = 0.054 (4.7)
BaseInterval F R 18500 1000 EF H
1000 1000 SV s
BaseSV R M R = = = 0.068 (4.8)
BaseInterval M R 14700 1000 EF H
In order to obtain the final SVR, the base SVRs are multiplied with the three-spool factor (TSF),
the severity factor (SF) and the environment factor (EF) as illustrated in figure 4.2.
SV R = BaseSV R · T SF · SF · EF (4.9)
The input indicates that both the TSF and EF equal 1.0, since the example engine has two spools
and is operated in a temperate environment. The SF is obtained from the average severity curve
for MLH engines (appendix D.2). For a flight time of 8.0 hours and a derate of 10% the severity
factors yields to SF = 0.88. Therefore, the actual first-run and mature-run SVRs result in:
SV R F R = 0.054 · 1.0 · 0.88 · 1.0 = 0.048 =

b 20800 EF H (4.10)
SV R M R = 0.068 · 1.0 · 0.88 · 1.0 = 0.060 =

b 16700 EF H (4.11)
These final SVRs can be converted back into an interval expressed in EFH, as performed above.
4.2.2.1 Adjustment of Costs
The total shop visit costs per EFH consist of LLP costs per EFH (LLP Cost EF H ) and the
adjusted restoration costs (SV RC EF H ).
SV C EF H = LLP Cost EF H + SV RC EF H (4.12)
The LLP costs are not adjusted by any effect factors. However, since the LLP costs are given in
USD/EFC, they have to be converted into USD/EFH:
EF C 1 U SD
LLP Cost EF H = LLP Cost · = 509 · = 64 (4.13)
EF H 8.0 EF H
The base restoration costs from eqs. (4.5) and (4.5) have to be multiplied with the time &
material factor, the severity factor and the environment factor in order to get the final restoration
costs.
SV RC = BaseSV RC EF H · T M F · SF · EF (4.14)
The TMF is determined with the respective time & material curve for MLH engines (appendix
D.3). With a flight time of 8.0, it yields to TMF = 1.05. The severity factor and environment
factor are known from before. Thus, the FR and MR restoration costs result in:
U SD
SV RC F R, EF H = 194 · 1.05 · 0.88 · 1.0 = 179 (4.15)
EF H
U SD
SV RC M R, EF H = 275 · 1.05 · 0.88 · 1.0 = 254 (4.16)
EF H
With the results from the eqs. (4.15) and (4.15), the total shop visit cost per EFH are then
calculated through eq. (4.12):
4.3 Model Plausibility 55
U SD
SV C F R, EF H = LLP Cost EF H + SV RC M R, EF H = 64 + 179 = 243 (4.17)
EF H
U SD
SV C M R, EF H = LLP Cost EF H + SV RC M R, EF H = 64 + 254 = 318 (4.18)
EF H
4.2.3 Final Results

Eventually the total shop visit costs can be calculated through the multiplication of the SVC per
EFH and the respective shop visit interval or time on wing.
SV C = SV C EF H · Interval (4.19)
SV C F R = SV C F R, EF H · Interval F R = 243 · 20800 = 5.1 mil U SD (4.20)
SV C M R = SV C M R, EF H · Interval M R = 318 · 16700 = 5.3 mil U SD (4.21)
The final output of the cost estimation are first-run and mature-run shop visit intervals and shop
visit costs. The results for the proposed example are summarized in table 4.7.
Interval F R [EF H] Interval M R [EF H] SV C F R [U SD million] SV C F R [U SD million]
20800 16700 5.1 5.3
Table 4.7: Final output for the example input parameters
4.3 Model Plausibility

The plausibility of the model was continuously monitored while developing the CERs and creating
the model structure. This intermediate plausibility tests significantly contributed to the decisions
made throughout the development process. This section illustrates the examination of the
credibility of the final model. In general, it is important to avoid using the same data that was
applied to develop the model for subsequent plausibility tests. It can be expected that the model
reflects the collected data of the database. However, since the database has been normalized and
the final model structure includes not only the developed CERs but also a series of adjustment
factors, it is first analyzed how well the final model reflects the original data points, prior to the
flight time normalization. Subsequently, the model results are compared to additional available
data sources.
4.3.1 Model Results vs. Original Database

As indicated, it is established that the developed CERs reflect the normalized data points quite
well (see regression results in appendix C). However, the objective of the following analysis
is to illustrate how the combination of the different CERs and the subsequent effect-module
reflect the original data before the flight time normalization. Therefore, each of the primary
output parameters of the model is plotted over the respective actual original data points from
the database. Ideally, the resulting points would lead to a graph that equals the standard linear
curve f (x) = x. In this case every predicted value would be equal to the respective actual value.
This ideal linear curve is plotted as a blue dotted line. However, it can be expected that the
plotted data points do not lie perfectly on this line. Furthermore it is possible that the ideal
curve does not even represent the trend line of the data points. Therefore, a linear regression
line that fits the data points is developed and additionally plotted as a red continuous line.
Coinciding red and blue dotted lines indicate that the data points can be fitted by the ideal curve
through linear regression. Clearly crossing lines would indicate opposing trends and thus refer to
a bad reflection of the actual data through the model. In the following tables, the resulting plots
illustrating the comparison of the model output with the original database are displayed. Each
engine application is considered separately. In addition, each plot displays the root mean square
error (RMSE) between data points and ideal curve.
4
x 10 FR Interval SH-Engines 4
x 10 MR Interval SH-Engines
3.5 2.5
3 RMSE: 3088 EFH RMSE: 2094.6 EFH

Actual Interval [EFH] 2
Actual Interval [EFH]
2.5
2 1.5
1.5
1
Data Points Data Points
Regression of Data Regression of Data
Ideal Curve Ideal Curve
0.5 0.5
0.5 1 1.5 2 2.5 3 3.5 0.5 1 1.5 2 2.5
Predicted Interval [EFH] x 10
4 Predicted Interval [EFH] x 10
4
4
x 10 FR Interval MLH-Engines 4
x 10 MR Interval MLH-Engines
3 2.5
RMSE: 1588 EFH RMSE: 1042.1 EFH

2.5
2
1.5
1.5
1
1
2 Shaft Data 2 Shaft Data
0.5 0.5
0.5 1 1.5 2 2.5 3 0.5 1 1.5 2 2.5
4 Predicted Interval [EFH] x 10
4
Intervals The plots show that the model generally represents the original database intervals
well. In the MLH plots, the three-spool engines have been highlighted. It can be seen that
the three-spool data points follow the trend of the remaining data quite well. This is achieved
through the three-spool adjustment factor. In addition, the RMSE values of this analysis relate to
the RMSEs of the regression analysis that was performed to develop the interval CERs (appendix
C).
FR SVC per EFH SH-Engines MR SVC per EFH SH-Engines

800 800
U SD U SD
700 RMSE: 30.7 EF H 700 RMSE: 35.3 EF H
Actual Cost [USD/EFH]

600 600
500 500
400 400
300 300
200 200
100 100
100 200 300 400 500 600 700 800 100 200 300 400 500 600 700 800
Predicted Cost [USD/EFH] Predicted Cost [USD/EFH]
FR SVC per EFH MLH-Engines MR SVC per EFH MLH-Engines

900 900
800 800
U SD U SD
RMSE: 32.4 EF H RMSE: 52.7 EF H
700 700
600 600
500 500
400 400
300 300

200 3 Shaft Data 200 3 Shaft Data
100 100
100 200 300 400 500 600 700 800 900 100 200 300 400 500 600 700 800 900
Predicted Cost [USD/EFH] Predicted Cost [USD/EFH]
Shop Visit Costs per EFH These plots generally indicate a weaker reflection of the original data
through the developed model. This probably results from the fact that the SVC per EFH are
made up of two independently modeled cost components (SV RC and LLP Cost). Especially the
SH engine comparison reveals a clearly visible deviation between model and database. The data
points here are clustered around a low cost level, while only a few data points reach higher cost
dimensions. The explanation for this is that the database of SH engines mainly consists of small
engines with a thrust level of about 20,000-30,000 lbf. The few data points that stick out are
made up of the CF6-80C2A series engines. These are the only bigger size short-haul engines of
the database. Therefore, the credibility of the model in this region is somewhat limited. It seems
the model tends to predict generally higher costs for such big SH engines. For MLH engines, the
picture is slightly more favorable. The data points are not as clustered around a certain cost
level. The two data points that stick out stand for the PW4074/77 operated on a short haul
route. The model is capable of reflecting this engine sufficiently however, one has to be careful
again, since there are only two data points that confirm the displayed trend in the higher cost
level of short-haul operated MLH engines.
4.3.2 Model Results vs. Additional Data Sources

It is crucial to also compare the developed model with additional available data. However,
adequate data sources are limited as it was pointed out in 3.1.2. Solely, the AeroStrategy
database proved to be appropriate for a plausibility test. However, since this database does
not provide all information that are compulsory for running the model, a few assumptions
had to be made. The average derate is again assumed to be 10%, while the environment is
considered as temperate. In addition it was necessary to define a legitimate flight time for each
data point of the AeroStrategy table. As mentioned earlier, the AeroStrategy table indicates
for most engines the respective aircraft, on which the estimation is based. Together with the
ACA aircraft operator & owner guides it was determined what the global average flight time
for these aircraft is. This average flight times were assigned to each data point and served as
input for the model. Subsequently, the AeroStrategy database was further extended to also
include the respective weight, thrust, number of spools and the engine application of each data
point. Hence, all necessary input parameters have been defined to perform a comparison of the
model with the addtional data. The plausibility based on the AeroStrategy database is tested
for six output parameters. On the cost side, the SV restoration costs per EFH and the total SV
restoration costs are compared separately for first and mature shop visits. Since the database
does not provide any cost figures for LLP charges, this cost component could not be checked
for its credibility. In addition, the first and mature removal intervals are compared between
model and AeroStrategy estimates. The presentation of the plausibility analysis correlates to the
previous plots. As before, the displayed RMSE values refer to the error between the data points
and the dotted blue line (ideal curve).
x 10 FR Interval - Model vs. AeroStrategy x 10 MR Interval - Model vs. AeroStrategy

4 4
3 2.2
2
RMSE: 1745.2 EFH
2.5 RMSE: 3169.6 EFH 1.8
1.6
2
1.4
1.2
1.5
1
0.8
1
Data Points 0.6
Data Points
0.5 0.4
0.5 1 1.5 2 2.5 3 0.5 1 1.5 2
4 Predicted Interval [EFH] 4
x 10
Intervals At first glance it becomes apparent, that the predicted intervals generally tend to be
longer than the intervals of the AeroStrategy data tables. However, the model clearly reflects the
prevailing trend of the reference data table. The resulting RMSE values are in the same scale
like the RMSEs of the other interval analysis plots, which indicates that model relates similarly
to both the Aircraft Commerce and AeroStrategy database.
FR SVRC per EFH - Model vs. AeroStrategy MR SVRC per EFH - Model vs. AeroStrategy
400
250
U SD U SD
RMSE: 32.4 EF H 350 RMSE: 41.2 EF H
Actual SVRC [USD/EFH]
Actual SVRC [USD/EFH]

200 300
250
150
200
100
150
50 Ideal Curve 100 Ideal Curve
50 100 150 200 250 100 150 200 250 300 350 400
Predicted SVRC [USD/EFH] Predicted SVRC [USD/EFH]
SV Restoration Costs per EFH Comparing the SVRC per EFH of the developed model with
AeroStrategy also reveals a satisfying picture. The model represents the trend of reference data
points very well and the displayed RMSEs are in an acceptable scale.
FR SVRC - Model vs. AeroStrategy MR SVRC - Model vs. AeroStrategy

6 6
5 RMSE: 0.6 mil USD 5 RMSE: 0.6 mil USD

Actual SVRC [USD million]
Actual SVRC [USD million]
4 4
3 3
2 2
1 1
0 0
0 1 2 3 4 5 6 0 1 2 3 4 5 6
Predicted SVRC [USD million] Predicted SVRC [USD million]
Total SV Restoration Costs The plots that are relating to the comparison of the total SVRC
generally show a bigger deviation between model and AeroStrategy. This is somewhat expected,
since now the combined results of intervals and costs per EFH in shape of the total restoration
costs are compared. It becomes apparent that the model generally predicts higher costs than
estimated by AeroStrategy for engines that require high investment for an overhaul. These
estimations mainly belong to the newest generation engines of widebody aircraft that were not
part of the assembled database. Hence, it can be assumed that new generation engines generally
achieve longer intervals with reduced cost per EFH. This is a trend that has been confirmed
through comparing the older engine generation of the database with the current generation
4.4 Sensitivity Analysis 60
engines that formed the foundation of the database. It is likely that this trend continues now,
leading to even longer intervals and lower shop visit costs per EFH for the newest generation
engines. However, the overall deviation is considered as acceptable and these results are still
regarded as confirmation of the developed model.
4.3.3 Summary of the Plausibility Tests

Considering that the objective was to develop a qualitative model that is capable of predicting
realistic SV intervals and costs with respect to basic relationships concerning engine maintenance,
the model relates to the expectations. Undeniably, there are considerable deviations between
model and available databases. However, the model does not attempt to give an exact forecast
of shop visits costs. It was proven that the model qualitatively reflects the general correlations
that define shop visit intervals and costs. In addition, the predicted absolute values lie in the
expected dimension. The present variations are considered admissible, granting that the exact
forecast of SV intervals and costs is to hard to achieve.
4.4 Sensitivity Analysis

This section presents the results of a basic sensitivity analysis on the developed model. Excluding
the engine application, the model relates to six input parameters (see fig. 4.1).
Parameter SH base values MLH base values

Thrust [lbf] 27000 78000
Weight [lbs] 5139 14545
EFH:EFC [h] 1.9 6.0
Derate [%] 10 10
EF 1.1 1.1
TSF 1.0 1.0
Table 4.8: Input parameters and their base values for the sensitivity analysis
Based on these parameters the analysis was split into two parts. First, it was examined how the
isolated output of the model reacts to changes on one single input parameter when the remaining
parameters are kept constant. The six analyzed parameters and the values they are held constant
to are summarized in tab. 4.8. The base SH engine values relate to the IAE V2500-A5 (A320
family) and the MLH values are derived from the PW 4077 (Boeing 777). These are two very
common engines in their respective field of application. The environment was assumed to be
hot and dry (EF = 1.1) and the two engines have two spools (TSF = 1.0). The impact of the
environment and the number of spools was only modeled in rough discrete steps, while the model
enables continuous changes of the remainder input parameters. As a result of this, the analysis
of the four continuous parameters - thrust, weight, flight time and derate - was grouped together,
while the impact of the number of spools and the environment is illustrated separately. This
ensures a consistent presentation of the results. The second part is a sensitivity analysis on the
impact of the six input parameters on the entire SV life cycle cost of an aircraft engine.
4.4.1 Sensitivity of Model Output

The impact of the continuous parameters is presented in two tables. The tables consist of four
plots, each plot relating to one changing input parameter. Each table on the other hand relates
to one certain output parameter. The results here are presented only for the range of SH engines.
The MLH sensitivity analysis generally produced similar results. These results can be found in
appendix E.
4.4.1.1 Continuous Parameters
x 10
4 Thrust - Interval Sensitivity x 10
4 Weight - Interval Sensitivity
2.8 2.8
First-Run First-Run
2.6 2.6
Mature-Run Mature-Run
2.4 2.4
2.2 2.2
Interval [EFH]
2
Interval [EFH] 2
1.8 1.8
1.6 1.6
1.4 1.4
1.2 1.2
EFH:EFC = 1.9[h] EFH:EFC = 1.9[h]
1 Derate = 10% 1 Derate = 10%
Weight = 5139[lbs] Thrust = 27000[lbf ]
0.8 0.8
2.2 2.4 2.6 2.8 3 3.2 3.4 4000 4500 5000 5500 6000 6500
Thrust [lbf ] 4
x 10 Weight [lbs]
x 10
4 EFH:EFC - Interval Sensitivity x 10
4 Derate - Interval Sensitivity
2.8 2.4
First-Run First-Run
2.6
Mature-Run 2.2 Mature-Run
2.4
2.2 2
Interval [EFH]
Interval [EFH]
2
1.8
1.8
1.6 1.6
1.4
1.4
1.2
Derate = 10% EFH:EFC = 1.9[h]
1 Thrust = 27000[lbf ] Thrust = 27000[lbf ]
1.2
Weight = 5139[lbs] Weight = 5139[lbs]
0.8
1 1.5 2 2.5 3 3.5 4 0 5 10 15 20
EFH:EFC [h] Derate [%]
Table 4.9: SH Engines - Impact of the continuous parameters on the SV Interval
Intervals The results of the interval sensitivity reflect the theory very well. An increase in thrust
would lead to a rapid drop of the achievable SV intervals, while an increase of the weight has
an opposed effect (upper two diagramms in tab. 4.9). This represents the expectation that a
higher thrust rating generally leads to shorter removal intervals. The influence of the derate and
the flight time directly relates to the implemented severity curves (see D). In addition, mature
SV intervals are generally shorter than first SV intervals, which has also been predicted by the
outcome of the literature review.
Thrust - SVC per EFH Sensitivity Weight - SVC per EFH Sensitivity
220
200 First-Run
First-Run
Mature-Run 190 Mature-Run
200
180
SVC per EFH [$/EFH]

SVC per EFH [$/EFH]
180 170
160
160
150
140
140
130
120
120 EFH:EFC = 1.9[h] EFH:EFC = 1.9[h]
Derate = 10% 110 Derate = 10%
100 100
2.2 2.4 2.6 2.8 3 3.2 3.4 4000 4500 5000 5500 6000 6500
Thrust [lbf ] 4
x 10 Weight [lbs]
EFH:EFC - SVC per EFH Sensitivity Derate - SVC per EFH Sensitivity
350 210
First-Run First-Run
200
Mature-Run Mature-Run
300 SVC per EFH [$/EFH] 190
SVC per EFH [$/EFH]
180
250
170
160
200
150
140
150
130
100 Derate = 10% 120

EFH:EFC = 1.9[h]
Thrust = 27000[lbf ] 110 Thrust = 27000[lbf ]
Weight = 5139[lbs] Weight = 5139[lbs]
50 100
1 1.5 2 2.5 3 3.5 4 0 5 10 15 20
Table 4.10: SH Engines - Impact of the continuous parameters on the SV costs per EFH
Shop Visit Costs per EFH The sensitivity of the shop visit costs per EFH is generally expected
as well. The flight time and the derate affect the SVC per EFH as defined in the severity curves.
An increase of the thrust yields to a considerable linear increase of the costs per EFH, while the
weight affects the costs only to a minor extent. The little effect the weight has on the SVC per
EFH results mainly from the fact that the CER for the restoration costs per EFH only depend
on the thrust (see eqs. (3.5),(3.6)).
4.4.1.2 Discrete Parameters
The remaining input parameters are implemented as discrete variables. Therefore, their impact
is illustrated in bar plots. Apart from this, the presentation is analogue as seen before. Each
couple of plots relates to one output parameter and each single plot shows the influence of one
changing input parameter, while the remaining parameters are held constant according to tab.
4.8. The following plots illustrate the results for MLH engines using the example of the PW4077.
The effect of the EF and the TSF on the SH engine intervals and costs are practically identical.
Thus, they are not illustrated explicitly. The results largely reflect the expectations and directly
relate to the implementation of the environment and three-spool factors.
x 10
4 Environment - Interval Sensitivity x 10
4 No. Spools - Interval Sensitivity
2.2 3
First-Run First-Run
2 Mature-Run Mature-Run
1.8 2.5
1.6
2
Interval [EFH]
Interval [EFH]
1.4
1.2
1.5
1
0.8
1
0.6
0.4 0.5
0.2
0 0
temperate hot&dry erosive 3 Spools 2 Spools
Environment No. Spools
Environment - SVC per EFH Sensitivity No. Spools - SVC per EFH Sensitivity
550 450
First-Run First-Run
500 Mature-Run 400 Mature-Run
450
350
SVC per EFH [$/EFH]
SVC per EFH [$/EFH]
400
300
350
300 250
250 200
200
150
150
100
100
50 50
0 0
Table 4.11: MLH Engines - Impact of the discrete parameters on the direct model output
4.4.2 Sensitivity of Life Cycle SVC

The developed model is dedicated to serve as a module for a more complex aircraft life cycle
cost tool. Therefore, it is now analyzed on how changes to the input parameters affect the
accumulated shop visit costs of the entire life cycle of an engine. The life cycle shop visit costs
SV C LC based on the developed model can be estimated as follows:
SV C LC = Interval F R · SV C F R, EF H + [EF H LC − Interval F R ] · SV C M R, EF H (4.22)
The remaining EFH after the first removal are multiplied with the mature SVC per EFH to
account for all mature SVs. The accumulated engine flight hours of the life cycle (EF H LC ) are
calculated with the number of years in service Y ears LC and the annual utilization U til ann .
EF H LC = Y ears LC · U til ann (4.23)
The annual utilization of the two example engines has been determined from the aircraft owner
& operator guides of the aircraft commerce archive. According to [Air08d] a Boeing 777 on a
long-haul route of 6.0 EFH:EFC equipped with a PW4077, typically achieves an annual utilization
of 4500 EFH. An IAE V2500-A5 mounted to an A320 that is flying on 1.9 EFH:EFC short-haul
route is likely to achieve around 2800 EFH [Air06a]. With an estimated life cycle duration of
Y ears LC = 20 years, the total number of EFH in service for the two example engines yields to:
EF H LC, SH = 20 · 2800 = 56,000 EF H (4.24)
EF H LC, M LH = 20 · 4600 = 90,000 EF H (4.25)
The sensitivity analysis of the total life cycle SVC is analogue to the previous analysis of the SV
intervals and SVC per EFH. The following tables show the effect of the four continuous and the
two discrete parameters. Subsequently it is illustrated through tornado charts what parameters
have the most significant impact. The applied base values relate to table 4.8.
4.4.2.1 Continuous Parameters
The sensitivity of the influence of the continuous parameters is illustrated for each engine
application separately.
x 10
6 Thrust - LCC Sensitivity - SH x 10
6 Weight - LCC Sensitivity - SH
12 10.1
11.5 10
11
9.9
Life Cycle SVC [USD]
10.5
9.8
10
9.7
9.5
9.6
9
9.5
8.5 EFH:EFC = 1.9[h] EFH:EFC = 1.9[h]
Derate = 10% 9.4 Derate = 10%
8
7.5 9.3
2.2 2.4 2.6 2.8 3 3.2 3.4 4000 4500 5000 5500 6000 6500
Thrust [lbf ] 4
x 10 Weight [lbs]
x 10
7 EFH:EFC - LCC Sensitivity - SH x 10
7 Derate - LCC Sensitivity - SH
1.8 1.15
1.6
1.1
1.4
1.05
1.2
1
1
0.95
0.8
Derate = 10%
Thrust = 27000[lbf ] 0.9 EFH:EFC = 1.9[h]
0.6
Weight = 5139[lbs]
0.4 0.85
1 1.5 2 2.5 3 3.5 4 0 5 10 15 20
Table 4.12: SH Engines - Impact of the continuous parameters on the life cycle SVC
Short-Haul Engines The results for the range of SH engines largely relate to the sensitivity
analysis of the SVC per EFH in tab. 4.10. However, the reduced life cycle SVC with increasing
engine weight seem to be slightly odd on the first glance. Generally, one would expect that
increasing weight leads to increased total SVC. However, the shown curve is a result of the mainly
constant SVC per EFH and significantly prolonged first intervals with increasing engine weight
(compare with tab. 4.10). From these four continuous input parameters, only the flight time and
the derate can actually adopt a wide range of values depending on the operation of the aircraft.
The engine thrust and weight are depending on the aspired performance level somewhat limited
by design constraints. When observing the plots, it becomes apparent that especially the average
flight time defines the resulting life cycle SVC. For the given example, it ranges from 6 mil USD
in case of EFH:EFC = 4.0 to about 17 mil USD for a short-haul operation with EFH:EFC = 1.0.
x 10
7 Thrust - LCC Sensitivity - MLH x 10
7 Weight - LCC Sensitivity - MLH
3.55 3.45
3.5
3.4
3.45

3.35
3.4
3.35 3.3
3.3
3.25
3.25
EFH:EFC = 6[h] 3.2 EFH:EFC = 6[h]
3.2 Derate = 10% Derate = 10%
3.15 3.15
7.4 7.6 7.8 8 8.2 8.4 8.6 1.2 1.25 1.3 1.35 1.4 1.45 1.5 1.55 1.6
Thrust [lbf ] 4
x 10 Weight [lbs] 4
x 10
x 10
7 EFH:EFC - LCC Sensitivity - MLH x 10
7 Derate - LCC Sensitivity - MLH
5 3.7
3.6
4.5
3.5
4
3.4
3.3
3.5
3.2
3
EFH:EFC = 6[h]
Derate = 10% 3.1 Thrust = 78000[lbf ]
Thrust = 78000[lbf ]
Weight = 14545[lbs]
Weight = 14545[lbs]
2.5 3
3 4 5 6 7 8 9 10 11 12 0 5 10 15 20
Table 4.13: MLH Engines - Impact of the continuous parameters on the life cycle SVC
Medium-Long-Haul Engines For a better understanding of the sensitivity of the life cycle SVC
for MLH engines, it is additionally referred to the respective tables in appendix E illustrating
the sensitivity of the isolated model output for MLH engines. In general, the results resemble
the previous SH engine plots. Solely, the impact of the weight is opposed. For MLH engines
the dependency of the life cycle SVC on the engine weight relates to the general expectation.
This results from the fact that unlike with the SH engines, the SVC per EFH of MLH engines
considerably increase with increasing engine weight. Again the life cycle SVC considerably range
with the average flight time.
4.4.2.2 Discrete Parameters
As with the analysis of the isolated model output, the influence of the discrete parameters is
illustrated only for MLH engines. The resulting plots for SH engines would show the exact same
tendencies just in a different scale.
x 10 Environment - LCC Sensitivity - MLH No. Spools - LCC Sensitivity - MLH

7 7
x 10
4 4
3.5 3.5
3 3

2.5 2.5
2 2
1.5 1.5
1 1
0.5 0.5
0 0
Table 4.14: MLH Engines - Impact of the discrete parameters on the life cycle SVC
4.4.2.3 Sensitivity Analysis via Tornado Charts
Tornado charts are often used to illustrate the sensitivity of a target parameter with respect to
changes on all input variables simultaneously. The typical tornado shape results from arranging
the input parameters in descending order according to the significance of their impact on the
output. In order to show what parameters influence the life cycle SVC most considerably, tornado
charts were generated for both SH and MLH engines. Therefore, it was determined how a
deviation of -10% to +10% around the base values from tab. 4.8 affect the life cycle SVC. The
three-spool factor was excluded from this consideration, since it only provides two discrete steps.
The three discrete steps of the environment factor on the other hand, happen to fit in the chosen
pattern. The EF for hot&dry environments serves as base value, while the two remaining steps
roughly relate to the ±10% deviation, that was applied for the continuous input parameters.
The results are illustrated in the following tornado charts.
Base Result: 12.657
Name Low High Delta +

Derate 12.882 12.522 0.360
Weight 12.891 12.518 0.373
EF 11.755 13.558 1.803
4.4 Sensitivity Analysis Thrust 11.527 13.799 2.272 67
EFH:EFC 13.897 11.336 2.561
- 10% + 10%
LCC - Sensitivity - SH Engines

EFH:EFC
Thrust
EF + 10%
Weight - 10%
Derate
11.0 11.5 12.0 12.5 13.0 13.5 14.0

Life Cycle Shop Visit Costs [USD million]
Base Result:
Figure 4.3: Tornado 30.670
chart on the sensitivity of the life cycle SVC of SH engines
Name Low High Delta
Derate 31.021 30.469 0.552
Short-Haul Engines For SH engines
Weight 29.870 the model 1.759
31.629 indicates that especially the average flight time, the
EFH:EFC 32.248 29.255 2.993
thrust and the environmentEF
factor
28.388
have a significant
32.952 4.564
impact on the accumulated SVC throughout
the life cycle, while the weight28.130
Thrust and the derate5.065
33.195 play a comparable minor role.
LCC - Sensitivity - MLH Engines

Thrust
EF
EFH:EFC + 10%
Weight - 10%
Derate
28,0 29,0 30,0 31,0 32,0 33,0 34,0

Life Cycle Shop Visit Costs [USD million]
Figure 4.4: Tornado chart on the sensitivity of the life cycle SVC of MLH engines
Medium-Long-Haul Engines The tornado chart for the MLH engines reveals a generally similar
picture. However, the effect of the average flight time is considerably less severe, while changes
on the engine weight gained in significance compared to SH engines. The explanation for this is
that the chosen base flight time of EFH:EFC = 6.0 is rather a long-hong-haul route. In long-haul
operations, slight changes on the average flight time generally have a less significant impact on
the engine maintenance (see severity curve in appendix D.2).
5 Implementation into existing LCC-Tool
For the implementation of the model, the engine maintenance was sourced out into a dedicated
function that is called in the main executive m-file lccmain.m of the LCC-tool. Up to this
point, the developed LCC-tool required a fixed predefined number of checks. These checks
included the engine shop visits, while it was assumed that engines generally have three shop
visits in their life cycle. As a result of this thesis, it became obvious that this approach not
necessarily reflects the reality. The number of shop visits can range significantly depending
on the achieved shop visit intervals and on the total flight hours of the proposed life cycle.
Therefore, the programme structure has to be modified to enable a flexible number of checks.
The existing structure dictates that the number of total maintenance checks has to be defined
prior to the utilization module lccmaintutil. Since the generated engine maintenance module
lccmaintengine determines the expected intervals and thus the required number of SVs, it has
to be executed before the call of lccmaintutil. The outsourcing into a dedicated function was
done to concentrate the contribution of this thesis to the existing LLC-tool in one central place.
The objective was to change the existing surrounding structure as little as possible. One main
requirement for the implementation is that the new engine maintenance module enables both the
estimation of shop visits according to the developed model and the predefinition of shop visits
information extracted from available sources. In the following sections, the implementation of
the key functionalities of the developed maintenance module is briefly described. However, it
should be noted that a complete understanding of the subsequent explanations requires basic
knowledge about the existing programme sequence. In case of unclarity, it is also referred to the
commented programme code.
5.1 Function Definition and Input Modification

The generated function as it is implemented in the global executive m-file is defined as follows:
[Maint] = lccmaintengine(Aircraft,...
Routes,...
Maint,...
General,...
File,...
CostTechFactor);
The various input parameter necessary for the this function are subsequently briefly described:
• Aircraft Holds aircraft specifications as they are defined in lcc_frame_in_xxx.xml

including key engine specifications
68
5.1 Function Definition and Input Modification 69
• Routes Contains all route informations, including flight time, derate and environment
• Maint Maint includes all maintenance data for the aircraft and engine. Thus, it contains all
engine specifications as well as ESV intervals and costs, as they are set in lcc_maintineng.
xml. It holds the key parameters that are applied to account for all maintenance events in
the following programme sequence.
• General Contains general information, like life cycle length and basic utilization parameters
such as, number of curfew hours or flight days per week.
• File Includes the information of the loaded xml files and controls what files are loaded.
• CostTechFactor Incldes all cost technology factors.
The existing global programme structure accesses the Maint struct for the following reflection
of the aircraft maintenance as part of the AC life cycle. Hence, the existing Maint struct,
which is also an input of the developed function, is modified and represents the only output.
The performed modifications result from the outcome of the implemented engine maintenance
model. The model implementation is based on the assumption that the utilization of the aircraft
is constant throughout the life cycle. That means all input parameter that are defined in the
routes branch of the lcc_frame_in_xxx.xml are assumed to be constant for the entire life
cycle.
The engine maintenance function provides different control settings, which allow the user to
influence the processing of the functions. These settings were implemented with simple true-
false queries. They are summarized as follows:
Variable Name true “1” false “0” struct xml file

engine maintenance model predefined SV intervals and
estimates SV intervals and costs from engineMroFile
Maint.ctrl.engMroType Maint lcc_frame_in_
costs according to input pa- are applied for life cycle
rameters modeling
engine specifications engine specifications
for maintenance estima- for maintenance estima-
File.input.engInputSource File lcc_frame_in_
tion are extracted from tion are extracted from
engineMroFile lcc_frame_in_xxx
spare engine costs are con-
spare engine costs are NOT
Maint.ctrl.spareEngineCost sidered and included in the Maint lcc_frame_in_
considered
maintenance costs
last shop visit is performed last shop visit is fully per-
Maint.ctrl.engLastSvType for a targeted remaining formed regardless of ex- Maint lcc_frame_in_
TOW until end of life cycle pected end of life cycle
Table 5.1: Summary of the main control settings of lccmaintengine
All these control setting variables have been added to the main xml input file lcc_frame_in_
xxx. In addition, several input parameters that are required for executing the developed engine
maintenance model were not yet defined and respectively had to be added to the input xml files.
This includes, the the derate, environment, engine application, number of spools as well as shop
visit duration, spare engine leasing rate an a few technology factors. Since the developed function
allows to extract the required engine specifications from two different input xml files, both the
5.2 Estimating the Shop Visits 70
engineMroFile and the lcc_frame_in_xxx have been modified. All additional modifications
on each xml file are summarized in the tables 5.2 and 5.3.
Added Parameters in lcc_frame_in_xxx.xml
Parameter Description xml branch Input into LCC-Tool
average derate flown, can Routes.deRate

deRate [%] Routes
range between [0.0-0.2]
accounts for the environment Routes.environment
environment Routes
condition, [1.0 1.1 1.2]
nSpools number of engine spools, [2 3] Aircraft Aircraft.engine.nSpools
accounts for the aircraft’s de- Aircraft.engine.range

range Aircraft
sign application [’SH’ ’MLH’]
leasingRate expected leasing rate for Aircraft.engine.leasingRate
Aircraft
[USD/day] spare engines
sVduration expected average shop visit Aircraft.engine.sVduration
Aircraft
[days] duration
Table 5.2: Summary of the lcc_frame_in_xxx.xml modification
Added Parameters in engineMroFile.xml
Parameter Description xml branch Input into LCC-Tool

accounts for the aircraft’s de-
range range Maint.engineMaint.range
sign application [’SH’ ’MLH’]
leasingRate expected leasing rate for
leasingRate Maint.engineMaint.leasingRate
[USD/day] spare engines
sVduration expected average shop visit
sVduration Maint.engineMaint.sVduration
[days] duration
Table 5.3: Summary of the engineMroFile modification
5.2 Estimating the Shop Visits

The programme code is generally separated into two main branches. Depending on the setting
of Maint.ctrl.engMroType either one of them is active. Most of the executed operations
are dedicated to their respective branch. Thus, they appear twice in slightly modified from
in the programme code, while only the operations of the active branch are executed. In
case Maint.ctrl.engMroType is set to ’0’, the function estimates the shop visit costs and
intervals according the developed engine maintenance model. Thus, this first branch requires the
implementation of the maintenance model structure as developed in 4.1. Therefore, each of the
seven CERs as presented in 3.3 was simply implemented as a dedicated function. The LLP cost
CER is for instance carried out by the follwoing function:
function [LLP_cost] = LLP_function(thrust,weight)

%% WHAT DOES IT DO?
%this function generates the Base LLP costs [dollar/EFC] of the engine
% INPUT: thrust and weight of analysed engine

% OUTPUT: LLP costs [dollar/EFC]
5.2 Estimating the Shop Visits 71
% Author: Ralf Seemann

% Date: 11.08.2010
%% PREDICTION FUNCTION
LLP_cost = −115.31326 ...
+0.0194512.*weight ...
+0.0031206.*thrust ...
+(weight−8608.78125).*((weight−8608.78125).*2.69234e−6);
Analogue to the developed CERs, the severity and time & material curves are also reflected by
dedicated functions. While the TMF curves can be each implemented by one single vector (see
tables 4.1 and 4.2), the severity curves have to be represented by matrices. According to the input
parameters the output value is then interpolated between the values defined in the curve vector
or multiple curve matrix (see appendix D). The interpolation between two values simultaneously
as required for the severity matrices was achieved by the matlab function interp2. An example
programme code for the implementation of the severity curves is given below:
function [severity_factor]= SM_severity_value(FcToFH,DeRate)

%% WHAT DOES IT DO?
%this function generates the severity matrix for short−haul engines

%and gives out the respective severity factor according to the input
% INPUT: FC:FH Ratio, De−Rate

% OUTPUT: severity_factor
% Author: Ralf Seemann
% Date: 11.08.2010
% Severity Matrix
%rows = [0.5 1 1.5 1.9 2.5 3 4 5 6 ]−−−> FC:FH
%columns = [0% 5% 10% 15% 20%] −−−> De−Rate
severity_matrix = [2.8 2.6 2.4 2.28 2.16
2.10 1.93 1.75 1.65 1.54
1.62 1.47 1.32 1.23 1.14
1.20 1.10 1.00 0.94 0.88
1.06 1.11 0.88 0.83 0.77
0.96 0.88 0.80 0.75 0.71
0.85 0.78 0.71 0.66 0.62
0.78 0.72 0.66 0.62 0.59
0.74 0.69 0.63 0.60 0.56];
% defining derate and FcToFH steps

derate_vec = [0 5 10 15 20];
FcToFH_vec = [0.5 1 1.5 1.9 2.5 3 4 5 6];
%% interpolation using interp2

severity_factor = interp2(derate_vec,FcToFH_vec,severity_matrix,DeRate,FcToFH);
5.3 Processing the Predefined Shop Visits 72
With the four effect curve functions the total number of generated functions sums up to eleven.
The remaining effects do not have to be reflected by dedicated functions, since they result directly
from the input. The model branch that is active when the SV intervals and costs are supposed
to be estimated, calls these functions and performs all necessary conversions as dictated by the
model structure seen in figure 4.2. The implementation of the developed model structure is also
represented by the example calculation in 4.2. The input parameters for the CER and effect
functions are extracted either from lcc_frame_in_xxx.xml or from engineMroFile, depending
on the setting of File.input.engInputSource.
5.3 Processing the Predefined Shop Visits

The second main branch of the programme code is active when Maint.ctrl.engMroType is set
to ’1’. In this case, the shop visits informations are extracted from the engineMroFile, where
they have been predefined. The following estimation of the required number of shop visits, the
consideration of spare engine costs as well as the definition of the last shop visits and the output
generation are performed in each branch separately. For estimating the number of required
shop visits, the average mature shop visit interval is calculated based on the defined shop visit
intervals extracted from engineMroFile. Therefore, the intervals of all shop visits following the
first one are averaged. This relates to the general assumption that engines reach maturity after
their first shop visit. The following programme parts are defined in each of the two branches.
5.4 Consideration of Spare Engine Costs

Depending on the setting of Maint.ctrl.spareEngineCost, this programm part either estimates
the costs for spare engines according to eq. (4.1) or it sets them to zero. The spare engine costs
are then added to the shop visits costs during the output generation.
5.5 Estimation of Required Shop Visit Number

The objective was to implement a flexible total number shop visits. Therefore, the number of
necessary shop visits throughout the life cycle has to be estimated. In order to achieve this, one
has to estimate the expected total flight hours during the life cycle. The exact total life cycle
FH are not known before the utilization function lccmaintutil has been executed. However,
the number of shop visits has to be fixed for running the utilization function. This problem was
solved by simplifying the calculation of the total life cycle FH (LCFH) . Therefore, the routine
for calculating the maximum possible FH during the LC was adopted from the lccmaintutil
function. The reduction of this maximum number of flight hours due to the various maintenance
events was neglected. This calculation is one of the few programme parts that is executed globally,
since it is applied for both programme lines.
5.6 Definition of Last Shop Visits 73
%% ESTIMATION OF YEARLY UTILIZATION PRIOR TO THE UTILIZATION MAINT MODULE

% this is necessary for the estimation of how many SV are necessary
opsHoursWeek_h = General.opsDaysWeek_d*(24−General.curfewHoursDay_h); % [h]
weeklyFC = (Routes.relativefrequency(:) .* opsHoursWeek_h) ./ Routes.cycleTime_h(:);
weeklyFH = weeklyFC.* fcToFh'; %fcToFh = flight_time
weeklyFC = sum(weeklyFC); %number of FCs per week
weeklyFH = sum(weeklyFH); %number of FH per week
%total flight hours of engine life with the factor 0.98 for reducing total
%flight hours due to maintenance events (reduced available FHs)
lcFH = 52*weeklyFH*yearsInService*0.98;
The following estimation of the shop visit number is dedicated to the respective branch. Depending
on the defined or estimated first and mature SV intervals, the number of required SVs is calculated
through a while loop. Noteworthy is that this loop only calculates the number of full shop visits.
This is explained by the following example. The first SV interval is assumed to be 15,000 EFH,
mature intervals equal 10,000 EFH and the LCFH equals 48,000 EFH. In this calculation, the
third shop visit would take place after:
3rdSVEF H = IntervalF R + IntervalM R + IntervalM R
= 15,000 + 10,000 + 10,000 = 35,000 EF H (5.1)
In this pattern the fourth SV would take place after an accumulated flight time of 45.000 EFH.
However, since the life cycle ends already after 48,000 EFH, it is not necessary to perform a
full fourth shop visit that would enable the engine to remain on-wing for another 10,000 EFH
resulting in accumulated 55,000 EFH. In this case, the loop gives out that the engine requires ’3’
full shop visits during the proposed life cycle. The last shop visit is considered separately in a
distinct programme part.
5.6 Definition of Last Shop Visits

The consideration of the last shop visit relates to the setting of Maint.ctrl.engLastSvType. If
it is set to ’1’ the programme determines the remaining TOW between last shop visit and end of
life cycle. In the example above, this relates to:
remainingT OW = LCF H − 4thSVEF H = 48,000 − 45,000 = 3,000 EF H
Thus, the last shop visit would have to restore the engine to a level that it can remain on-wing
for another 3.000 EFH. Therefore, the last shop visit is considered as targeted SV with reduced
workscopes. The incurred costs for this targeted SV are calculated with the mature shop visit
cost per EFH (see eq. (4.18)) multiplied with the remainingT OW . If it is assumed that a
mature SV in the previous example costs 3 mil USD, then the matre SVC per EFH yield to:
5.7 Output Generation 74
SV CM R 3,000,000 U SD
SV C M R, EF H = = = 300
IntervalM R 10,000 EF H
Hence, the cost for the last shop visit as indicated in the example result in:
SV C last = SV C M R, EF H · remainingT OW = 300 · 3,000 = 900,000 U SD
This calculation is performed in both branches with the respective interval and cost data resulting
either from the implemented engine maintenance model or from the predefined shop visit data.
In case Maint.ctrl.engLastSvType is set to ’0’, the programme handles the last shop visit like
all previous SVs as full shop visit regardless of the expected remaining TOW until the end of the
life cycle. This setting should be preferred when selling the engine on the surplus market after
the end of the aircraft’s life cycle is considered.
5.7 Output Generation

In the existing programme structure, the shop visit intervals and costs are read from the
engineMroFile xml file and then written in Maint.engineMaint.shopVisit according to the
xml structure. This data is then applied to determine the utilization and to define the maintenance
event costs. The objective was to keep this structure. Therefore, the programmed engine
maintenance module simply overwrites the Maint.engineMaint.shopVisit entries according to
the outcome of the engine maintenance function. Hence, the output has the exact same structure
as it is defined in in the engineMroFile xml file, with the addition of a field for the spare engine
costs. This enables an uncomplicated implementation of the developed function into the global
programme sequence.
The following maintenance cost function lccmaintcost classifies all check expenses into four
categories. The cost elements from the modified Maint.engineMaint.shopVisit variable are
allocated as follows:
Allocation of SVC on CheckExpenses
manhours -
materialcost LLP costs
fixcost restoration costs + spare engine costs
laborcost -
Table 5.4: Reflection of the engine maintenance costs in the lccmaintcost function
This allocation enables a differentiated adjustment of the restoration cost and the LLP cost
through the existing technology factors for each CheckExpenses category.
6 Summary and Conclusion
The objectives of this work were to review the literature on maintenance of commercial jet engines
and based on that, to develop a qualitative model that estimates engine shop visit costs and
intervals depending on the major influence factors on engine maintenance. Furthermore, it was
intended to implement this model into the existing maintenance module of the LCC simulation
tool.
After building up a comprehensive review on the prevailing concepts of engine maintenance, cost
estimating relationships (CER) regarding the engine shop visit costs and intervals were developed
using the methodological approach described in the NASA Cost Estimating Handbook [NAS08].
Therefore, a database that contains numerous current engine model variants and their shop visit
intervals and costs was assembled from an extensive review of the operator & owner guides of
the Aircraft Commerce magazine archive. However, important effects like the environmental
conditions and the operational severity have been normalized for the database assembly, which
led to CERs that do not reflect the influence of these factors. Therefore, the developed CERs
were complemented by a subsequent effect-module that adjusts the results of the CERs according
to the severity of the engine’s operation. The final model relates to six different input parameters:
engine thrust, dry weight, number of spools, average flight length, applied derate and the present
environment. Since the literature review and the assembled database indicated that short-haul
operated engines generally exhibit different maintenance characteristics than engines that are
operated on medium-long-haul routes, the model was split into two separate paths, each dedicated
to one of these engine applications. In addition the model distinguishes between first-run and
mature-run shop visits to account for the generally longer intervals and lower maintenance cost
per EFH of new engines compared to engines that reached maturity.
The resulting model was then tested for its plausibility by comparing the model results with
available cost and interval estimations from AeroStrategy. The conclusion of these plausibility
tests were that the general trend of the developed model and the Aerostrategy estimations
coincide. However, the AeroStrategy estimates for new generation engines tend to lie below the
predicted values of the model. This was not unexpected, since the past has shown that newer
generations engines generally achieve longer intervals and lower maintenance costs per EFH than
the previous generation. Since the database assembly was limited to engines that have been in
operation for the last two decades, the developed model reflects the current generation engines
best. The problem is that there is no reliable data on the average intervals and costs for the
newest engine generation. However, with these information available one could determine a
technology factor that adjusts the model results and enables a better forecast also for the these
75
6 Summary and Conclusion 76
new engines. The basic engine maintenance characteristics are assumed to remain constant also
with newer generation engines. Therefore, the applied CERs and adjustment parameters could
be also replaced with newly developed relationships that are based on available data for the
newer generation engines, while the rest of the model structure could remain unchanged.
The developed model was subsequently implemented into the LCC simulation tool as an inde-
pendent module. Therefore, the engine maintenance was excluded from the existing maintenance
module. This ensures that the contribution of this thesis to the LCC-tool is clearly separated and
it enabled the consideration of a flexible number of shop visits, while the previous maintenance
module relied on a fixed number of shop visits. Therefore, the new engine maintenance module
estimates the anticipated number of shop visits of the life cycle depending on the utilization
input and the estimated shop visit intervals. Since the developed engine MRO model requires a
few new input parameters that have not been included in the original input files, the xml input
files have been modified accordingly. The existing global structure of the LLC simulation tool
remained unchanged. The new module keeps the functionality that the shop visit intervals and
costs can be predefined if known from reliable data sources. The problem is that the actual shop
visit costs and intervals heavily depend on the engine’s operational severity. That means, the
applied predefined shop visit estimations have to relate to the utilization defined in the input file.
This applies especially for the flight time as major influence factor. Alternatively, the available
maintenance data could also be adjusted with the average severity curves established in the
framework of this thesis.
Bibliography
[AAE04] AAE 451 - Aircraft Design: Method for Calculating Direct Operat-
ing Cost (2004), URL http://www.soton.ac.uk/~jps7/Aircraft%20Design%
20Resources/Cost%20data/aircraft%20operating%20cost%20equations.pdf
[Ack10] Ackert. S: Engine Maintenance Concepts for Financiers: Elements of Engine Shop
Maintenance Costs (2010)
[ACT08] ACT: Keep on running - engine maintenance and cost reduction. AIRCRAFT TECH-
NOLOGY (Jun/Jul2008), (94): S. 52–60
[AEG00] AEG Working Group: Aircraft Evaluation Guidelines: Ref:
V62/CMP/SSA/16700C, issue 2, rev. 2 Aufl. (August 2000)
[Air67] Air Transport Association of America: Standard Methode of Estimating
comparative direct Operating Cost of Turbine Powered Transport Airplanes (1967)
[Air99] Aircraft Commerce: Fine tuning engine maintenance costs. AIRCRAFT COM-
MERCE (Jul/Aug 1999), (6): S. 42–48
[Air00] Aircraft Commerce: Ageing CF6-50 delivers acceptable maintenance costs: Main-
tenance & Engineering. AIRCRAFT COMMERCE (Nov/Dec 2000), (14): S. 18–24
[Air04a] Aircraft Commerce: LLP management for widebody engines. AIRCRAFT COM-
MERCE (2004), (39): S. 29–36
[Air04b] Aircraft Commerce: LLP management for short-haul engines. AIRCRAFT COM-
MERCE (Apr/May 2004), (34): S. 36–44
[Air04c] Aircraft Commerce: CFM56-3B1/B2/C1 maintenance cost analysis. AIRCRAFT
COMMERCE (Feb/Mar 2004), (33): S. 27–34
[Air05a] Aircraft Commerce: Using ECM to reduce engine maintenance costs. AIRCRAFT
COMMERCE (Aug/Sep 2005), (42): S. 36–41
[Air05b] Aircraft Commerce: CFM56-7B maintenance costs & reserves analysis. AIRCRAFT
COMMERCE (Feb/Mar 2005), (39)
[Air06a] A320 Family Maintenance Analysis & Budget. AIRCRAFT COMMERCE (2006),
(44): S. 18–31
[Air06b] Aircraft Commerce: Owner’s & Operator’s Guide: CFM56-3 Series. AIRCRAFT
COMMERCE (Apr/May 2006), (45): S. 9–31
[Air06c] Aircraft Commerce: Systems for Engine health monitoring. AIRCRAFT COM-
MERCE (Feb/Mar 2006), (44): S. 53–57
77
Bibliography 78
[Air06d] Aircraft Commerce: CF6-80C2 maintenance analysis & budget. AIRCRAFT

COMMERCE (Oct/Nov 2006), (48): S. 18–30
[Air07a] Aircraft Commerce: The economics of repairing engine parts. AIRCRAFT COM-
MERCE (Aug/Sep 2007), (30): S. 33–37
[Air07b] Aircraft Commerce: CFM56-5A/5B series specifications. AIRCRAFT COM-
MERCE (Feb/Mar 2007), (50): S. 6–9
[Air08a] Aircraft Commerce: Big engine in-service performance & maintenance. AIRCRAFT
COMMERCE (Aug/Sep 2008), (59): S. 41–53
[Air08b] Aircraft Commerce: V2500 maintenance analysis & budget (Feb/Mar 2008), (56): S.
17–31
[Air08c] Aircraft Commerce: CFM56-7B maintenance analysis & budget. AIRCRAFT
COMMERCE (Jun/Jul 2008), (58): S. 18–28
[Air08d] Aircraft Commerce: 777 family maintenance analysis & budget. AIRCRAFT
COMMERCE (Oct/Nov 2008), (60): S. 12–23
[Air09] Aircraft Commerce: Analysing the engine shop visit process. AIRCRAFT COM-
MERCE (Feb/Mar 2009), (62): S. 45–52
[Bec09] Beck A.: Engine Maintenance Cost Monitoring (29.10.2009)
[Ben08] Bennett J.: PMA parts penetration and acceptability. THE ENGINE YEARBOOK
(2008): S. 74–79
[BJ82] Birkler J.L., Garfinkle J. B.: Development and Production Cost Estimating Rela-
tionships for Aircraft Turbine Engines, RAND (1982)
[Bra04] Braeunling, W.: Flugzeugtriebwerke, Springer, Berlin Heidelberg, 2 Aufl. (2004)
[Bru96] Brundick B.: Parametric Cost Estimating Handbook (1996)
[Bur97] Burton M.: Piston Engines Propellers Gas Turbine Engines: The commercial pilot’s
study manual series (1997)
[Bur10] Burchell B.: Wanted: More On-Wing Wizardry. AVIATION WEEK (Apr 2010)
[Car00] Carsten Mildt: Entwicklung einer Methode zur Abschätzung der Kosten
für die Instandhaltung der Flugzeugzellen und -systeme im kommerziellen
Flugverkehr, TU Berlin (2000), URL http://www.dfld.de/Downloads/20000725_
Wartungskosten.pdf
[CFM09] CFM: CFM56 Cost of Ownership (October 2009)
[Chu08] Churchwell P.: Fundamentals of CER Development: Defense Acquisition University
(June 2008)
[Cum97] Cumpsty N.A.: Jet Propulsion: A simple guide to the aerodynamic and thermodynamic
design and performance of jet engines, Cambridge University Press (1997)
Bibliography 79
[Eng10] Engine Yearbook: Labour qualitiy, not labour rates drives maintenance cost opti-
mization. THE ENGINE YEARBOOK (2010): S. 52–56
[FAA04] FAA - Federal Aviation Administration: Airplane Flying Handbook: FAA-H-
8083-3A (2004)
[FAA09] FAA - Federal Aviation Administration: GUIDANCE MATERIAL FOR AIR-
CRAFT ENGINE LIFE-LIMITED PARTS REQUIREMENTS (2009)
[FAA10] FAA - Federal Aviation Administration: Aircraft Certification: Parts Manufac-
turer Approval (PMA) (2010), URL http://www.faa.gov/aircraft/air_cert/
design_approvals/pma/
[Gal08] Galorath D.: Parametric Estimating Handbook: 4th Edition (2008)
[GE 08] GE Aviation: Engine Education: Vocabulary (2008), URL http://www.geae.com/
education/vocabulary.html
[Gun95] Gunston B.: The Development of Jet and Turbine Aero Engines, Patrick Stephens
Limited (1995)
[Hol08] Holland J.: End of the PMA legitimacy debate. AIRCRAFT TECHNOLOGY
(Oct/Nov 2008), (96): S. 64–71
[Hue03] Huenecke K.: Jet Engines: Fundamentals of Theory, Design and Operation, Airlife
Publishing Ltd (2003)
[Jet08] JetEngine Consulting: Our Consultancy Philosophy (2008), URL http://www.
jetengineconsulting.com/pdf/JEC_Consulting_Philosophy.pdf
[KLM07] KLM: Engine Water Wash at KLM Engineering & Maintenance (2007), URL
http://www.klm.com/travel/csr_en/images/Engine%20Water%20Wash%20EN_
tcm256-100165.pdf
[Lin08] Linke-Diesinger A.: Systems of Commercial Turbofan Engines: An Introduction to
Systems Functions, Springer Verlag, Berlin Heidelberg (2008)
[Lon00] Long J.: Parametric Cost Estimating in the New Millenium (2000)
[Mei05] Meier N.: Jet Engine Specification Database: Civil Turbojet/Turbofan Specifications
(2005), URL http://www.jet-engine.net/civtfspec.html
[MM10] Mueller M., Staudacher S.: Probalisitsche Modellierung des Einflusses variierender
Umwelt- und Betriebsbedingungen auf die Triebwerksinstandhaltung (2010)
[NAS08] NASA: Cost Estimating Handbook, http://www.ceh.nasa.gov (2008)
[OCE91] Otis C. E., Vosbury P. A.: Aircraft Gas Turbine Engines of the World: Dictonary of
the Gas Turbine, IAP, Inc. (1991)
[Pet08] Peter Nolte: Bewertung des Zusammenhangs zwischen globalen Szenarien und dem
ökonomisch getriebenen Flugzeugentwurf: Diplomarbeit (2008)
Bibliography 80
[Rol07] Rolls-Royce Ltd.: Gas turbine technology: Introduction to a jet engine (2007),
URL http://www.rolls-royce.com/Images/gasturbines_tcm92-4977.pdf
[Ros90] Roskam J.: Alirplane Design: Part VIII: Airplane Cost Estimation: Design, Develop-
ment, Manufacturing and Operating, Lawrence, Kansas (1990)
[Rou07] Roux E.: Turbofan and Turbojet Engines: Database Handbook, Elodie ROUX (2007)
[Rup00] Rupp O.: Instandhaltungskosten bei zivilen Strahltriebwerken (2000)
[Sch98] Scholz D.: DOCsys - A Method to Evaluate Aircraft Systems, Bonn (1998)
[Sch09] Schilling T.: Methoden zur Modellierung von Flugzeuginstandhaltung - Diplomarbeit
(2009), URL http://www.tu-harburg.de/ilt/
[TB07] TU-Berlin: DOC-Abschätzungsmethoden (2007), URL http://www.ilr.tu-berlin.
de/LB/
[Tew07] Tewari A.: Atmospheric and Space Flight Dynamics: Modeling and Simulation with
Matlab and Simulink, Birkhäuser, Boston Basel Berlin (2007)
[Tre79] Treager I.E.: Aircraft gas Turbine Engine Technology: Second Edition, McGraw-Hill
Book Company (1979)
[Tur04] Turso J.A.: A Foreign Object Damage Event Detector Data Fusion System for
Turbofan Engines: NASA/TM—2004-213192 (2004)
[UCL] UCLA: Probability and statistics, Dissertation, URL http://wiki.stat.ucla.edu/
socr/index.php/Probability_and_statistics_EBook#Two-Way_ANOVA
List of Figures
1.1 Aircraft MRO cost overview [Jet08] . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2.1 Engine core of gas turbine engines . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2.2 Working cycle of a gas turbine engine . . . . . . . . . . . . . . . . . . . . . . . . 4
2.3 GEnx-2B - high bypass twin-spool turbofan . . . . . . . . . . . . . . . . . . . . . 6
2.4 Comparison: two- and three-spool configuration . . . . . . . . . . . . . . . . . . . 7
2.5 The main modules of a V2500-A5 [Lin08] . . . . . . . . . . . . . . . . . . . . . . 7
2.6 Correlation between Take-Off EGT and OAT [Air06b] . . . . . . . . . . . . . . . 10
2.7 Removal causes depending on aircraft operation [Ack10] . . . . . . . . . . . . . . 14
2.8 Effects of engine wear on the EGT Margin [Ack10] . . . . . . . . . . . . . . . . . 14
2.9 Trend of EGT margin erosion rates over accumulated EFC . . . . . . . . . . . . 15
2.10 Engine shop visit process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.11 Influences of the TOW on the DMC of an engine [Eng10] . . . . . . . . . . . . . 17
2.12 Two example flight profiles [Ack10] . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.13 Shop visit rate and DMC in relation to the flight hour flight cycle ratio . . . . . 22
2.14 Example severity curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.15 Engine maintenance cost breakdown structure . . . . . . . . . . . . . . . . . . . 25
2.16 Parametric Cost Estimating process steps [NAS08] . . . . . . . . . . . . . . . . . 28
2.17 Example data points for cost-weight dependency . . . . . . . . . . . . . . . . . . 30
3.1 JMP screening function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

3.2 JMP regression results example output . . . . . . . . . . . . . . . . . . . . . . . . 42
4.1 Black box of maintenance model . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

4.2 Inner structure of the cost estimating model . . . . . . . . . . . . . . . . . . . . . 48
4.3 Tornado chart on the sensitivity of the life cycle SVC of SH engines . . . . . . . 67
4.4 Tornado chart on the sensitivity of the life cycle SVC of MLH engines . . . . . . 67
i
List of Tables
2.1 Initial EGTM and mature EGT erosion rates for CFM56-7B variants [Air08c] . . 20
2.2 Comparison of EMC consideration in different DOC methods . . . . . . . . . . . 27
2.3 Criteria for the evaluation of regression results . . . . . . . . . . . . . . . . . . . 31
3.1 Summary of cost estimating relationship hypothesis . . . . . . . . . . . . . . . . 33

3.2 Structure of the collected Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.3 Determined average SH severity curve for a derate of 10% . . . . . . . . . . . . . 37
3.4 Determined average MLH severity curve for a derate of 10% . . . . . . . . . . . . 37
3.5 Summary of DB base conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.6 Summary of the preliminary dependent variables to be analyzed . . . . . . . . . 40
3.7 Summary of the final dependent variables to be analyzed . . . . . . . . . . . . . 41
3.8 Defined standard units for the input parameters of the CERs . . . . . . . . . . . 43
4.1 SH Engine Time & Material factor with respect to the flight time . . . . . . . . . 50
4.2 MLH Engine Time & Material factor with respect to the flight time . . . . . . . 50
4.3 Environment factors for different environmental conditions . . . . . . . . . . . . . 51
4.4 Input engine specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.5 Input engine utilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.6 Summary of CER results for the example input parameters . . . . . . . . . . . . 53
4.7 Final output for the example input parameters . . . . . . . . . . . . . . . . . . . 55
4.8 Input parameters and their base values for the sensitivity analysis . . . . . . . . 60
4.9 SH Engines - Impact of the continuous parameters on the SV Interval . . . . . 61
4.10 SH Engines - Impact of the continuous parameters on the SV costs per EFH . 62
4.11 MLH Engines - Impact of the discrete parameters on the direct model output . 63
4.12 SH Engines - Impact of the continuous parameters on the life cycle SVC . . . . 64
4.13 MLH Engines - Impact of the continuous parameters on the life cycle SVC . . 65
4.14 MLH Engines - Impact of the discrete parameters on the life cycle SVC . . . . 66
5.1 Summary of the main control settings of lccmaintengine . . . . . . . . . . . . . 69

5.2 Summary of the lcc_frame_in_xxx.xml modification . . . . . . . . . . . . . . . 70
5.3 Summary of the engineMroFile modification . . . . . . . . . . . . . . . . . . . . 70
5.4 Reflection of the engine maintenance costs in the lccmaintcost function . . . . 74
ii
A-HW(QJLQH&RQVXOWLQJ
Maintenance Costs
Operational
A.1 CostCost Analysis
Engine MRO
Engine MRO Cost Analysis
HPT aifoils
Repairs Other 21%
3% Fees 28%
12% DAT*
Life Limited Parts
25% US$ 3.9 billion (LLP) Other airfoils
Parts Repair 4% 21%
FAn
5%
anfallenden Kosten können im wesentlichen in Stationary parts
drei Bereiche aufgeteilt werden: Kosten für die Combustor 15%
6% PMA Parts
uß, den die Montage/Demontage des Triebwerks, Reparatur-
3%
altungskosten kosten für Einzelteile
60% und Materialkosten für den Life Limited Parts
eines Trieb- Ersatz von nicht mehr reparierbaren
Material Bauteilen. Die
US$ 9.3 billion (LLP)
Material
Verteilung der Kosten auf bestimmte Bauteile und
Reinigungs- 19%
tfernung von Bereiche ist von großer Bedeutung, da sie Rück-
chaufeln und schlüsse auf potentiell erreichbare Kostenreduzie-
stellung der rungen für die einzelnen Bauteile - beispielsweise Used Material
ufbetrieb des durch die Entwicklung neuer Reparaturen - zuläßt. 12% New Material
Kohlepulver *) Disassembly, Assembly, Test 66%
itzt werden. BILD
Source:42007zeigt beispielhaft
AeroStrategy die Repair
Aeroengine Parts Verteilung der
& Material Forecast
nreinigungen Instandhaltungskosten bei einer Überholung auf
s Verdichters die
24 Bauteile,
November 2006 unterteilt nach Materialkosten (blau 4
er reduzierte gekennzeichnet) und Arbeitskosten (rot bzw. gelb
uch die Be- gekennzeichnet). Ausdrücklich sei hier darauf
hingewiesen, dass lebensdauerbegrenzte Teile
A.2
(LLP's)Shop VisitmitCost
hier nicht Driver sind.
berücksichtigt
others 10 % Disassembly
Combustor 2,5 % Assembly 15 %
Cases 5 %
Bearings 2,5 % Airfoils 14 % *LLP cost not included
Blue: material cost
Red: labour cost
Airfoils 30 % Accessories 5 %
vement of MTBSV [Rup00]
ntroduction of Stationary Parts 8 %
ar Coke Cleaning
an Jul Rotating Parts 4 %
Seals 2 % Combustor 2 %
ng und BILD 4. Kostentreiber bei der Überholung von

e auf die Triebwerken.
Auffallend ist der große Kostenanteil für Airfoils -

Durchführung insgesamt fast 50% der Überholungskosten.
eispielhaft in Primär entstehen diese Kosten durch den Ersatz
n Fall führte von nicht mehr reparablen Airfoils aus dem
9 eine interne Bereich der Hochdruckturbine. Dies sind sicherlich iii
ge Reinigung mit die am höchsten belasteten Bauteile im
ke Cleaning) gesamten Triebwerk, wobei erschwerend hinzu-
B Database
B.1 Aircraft Commerce Shop Visit Reserves & Intervals Example Table
44 I MAINTENANCE & ENGINEERING
POSSIBLE MANAGEMENT, SHOP VISIT PATTERN & LLP REPLACEMENT TIMING OF CFM56-7B SERIES ENGINES
Removal Interval Accumulated Workscope Cost-$ $/EFC LLP LLP cost LLP Total Total
EFC EFC content replacement $ $/EFC $/EFC $/EFH
-7B27
1st 10,000 10,000 Core 1,200,000 120 - 79 199 111
2nd 7,000-8,000 17,000-18,000 Core & LPT 1,500,000 200 Core & LPT 1,211,000 79 279 155
-7B26
1st 13,000 13,000 Core 1,250,000 96 Core 785,000 91 190 106
2nd 12,000 25,000 fan/LPC & LPT 1,700,000 142 Fan/LPC & LPT 731,000 91 233 130
-7B24
1st 16,000 16,000 Core & LPT 1,550,000 97 Core & LPT 1,211.000 87 184 102
2nd 14,000 30,000 Core & fan/LPC 1,700,000 121 Fan/LPC 305,000 71 192 107
-7B18/20/22
1st 17,000-18,000 17,000-18,000 Core & LPT 1,600,000 92 Core & LPT 1,211,000 78 170 95
2nd 12,000-13,000 28,000-30,000 Core & fan/LPC 1,700,000 136 Fan/LPC 305,000 73 209 116
B.2 Classification of Aircraft Engines

may be a complete overhaul of all 45,000-50,000EFH. This will be equal to replacing LLPs.
modules. 15-17 years of operation for most Crawford estimates that a heavier
Putting aircraft engines into categories according to their application is rather subjective. A
“The -7B18/20/22 will have a core airlines’ operations. The highest thrust shop visit that included work on the LPT
regional airline may classify a certain route as middle-haul, while an intercontinental
refurbishment without LLP replacement, rated -7B27 will have accumulated about module would operating
add up to another 500
plus fan and booster refurbishment with 20,000EFC and 38,000EFH at the second MH, $200,000 for materials and parts
airline could classify
LLP replacement,” explains the same route
Crawford. shopasvisit,
shortafterdistance. The
a total time of following
about 13- table
and reflects
another $100,000theforperspective
sub-contract
This would be at a total accumulated 15 years’ service. repairs. The same labour rate of $70 per
of the author of this thesis. In the This
time close to 30,000EFC.
framework of this thesis, the engines
indicates that the second shop
of the database have
MH would take the total cost for the
been Thedivided
medium-ratedin -7B24 will have (SH)
short-haul a visits
andwill not occur for about another
medium/long-haul shop visit to The
(MLH) engines. $1.5-1.6 million.
classification is
similar second shop visit. seven years for the oldest and highest Crawford estimates that a complete
notHigher
basedratedon-7B26
theengines
absolute range rated
that did of the aircraft,
engines. on which
It is therefore too earlythe
to engines
overhaulare mounted.
would Rather up
use about 3,000MH, it
not have work done on the LPT at the estimate the intervals to the third shop to $1.2 million in materials and parts,
corresponds to have
first shop visit would the all
rough
major average flight
visits time
and their these workscopes.
subsequent aircraft and respectively
and about $500,000 theforengines were
sub-contract
modules worked
designed for.onHowever, engines can by all means operate outsiderepairs.
during the second theirThis would range.
design take total cost for the
A good
shop visit, as well as replacement of LPT shop visit to about $1.7 million. Some
guideline
and fan/LPC forLLPs.classifying
This would bethe Shop
at a engines for visit inputs model is the engine
the developed common distinction
shops estimate that thebetween
MH inputs
total accumulated time of 25,000EFC. Inputs for shop visits in terms of man- for these heavier shop visits may be as
short-haul
“The highestaircraft (e.g.
rated -7B27s A320, Boeing
would 737, materials
hours (MH), Embraer E-Jets)
and parts and wide
and cost high asbody
4,000. aircraft (e.g. A330,
have another core refurbishment and of sub-contract repairs for the -7B, can be
A340, Boeing 777/747). From the perspective of the developed model, wide-body aircraft are
workscope on the LPT, with LLPs being estimated on the basis of inputs for the -3
replaced in all
powered bythese modules,”
typical MLH says engines,series.
short-haul aircraft respectively by LLPSH amortisation
engines.
Crawford. This raises the issue of the “Man-hour inputs are expected to be LLP amortisation has to consider
fan/LPC module. Total accumulatedShort-Haul
time similar to the -3 for similar shop visit Long-Haul
Medium-Haul probable intervals to the third shop visit.
on-wing at this stage will be 17,000- workscopes, perhaps slightly lower,” says This is because LLPs replaced at the first
< 3FH 3 − 6FH > 6FH
20,000EFC, and so there would be up to Beale. “Material costs for a -7B are shop visit will be replaced at the third
13,000EFC remaining until LLPs in the expected to be 20-35% more than a -3 shop visit. Their cost should thus be
fan/LPC needed replacing. The fan/LPC and the -7B’s cost of sub-contracted amortised over the combined second and
module could thus be worked on at the repairs will be 5-15% more than those third removal intervals. Although some
third shop visit. experienced by the -3.” modules have LLPs replaced together
Sekinger estimates a core during the first or second shop visit, they
refurbishment plus some work on the may then get LLPs replaced for a second
Engine management LPT will use 2,700 man-hours (MH). time at different shop visits. LLPs for the
These approximate on-wing intervals Charged at a labour rate of $70 per MH fan/LPC have a list price of $305,000, iv
and LLP lives strongly influence engine this will take total cost to about LLPs in the HP system a list price of
management and shop visit workscopes. $140,000. Cost of materials and parts $785,000 and LLPs in the LPT a list price
Total time to the second shop visit will be will be about $700,000-800,000, on the of $426,000.
B.3 Core Database v
B.3 Core Database

Specifications First Removal Mature Removal Normalized Data
removal Removal LLP SV removal removal LLP SV 1st 1st mature mature
T_TO Weight EFH: MIF Interval Interval Reserves Reserves Interval Interval Reserves Reserves Interval cost interval cost Mean
OEM Model [lbf] [lb] TWR EFC EFC EFH [$/EFC] [$EFH] EFC EFH [$/EFC] [$EFH] SF SVR Adj Adj SVR adj Adj LLP Cost
CFM CFM56-3-xxx 18500 4276 4,33 1,40 1,10 18000 25200 88,00 53,90 10000 14000 83,05 96,25 1,30 32760 37,69 18200 74,04 85,53
CFM CFM56-3B1 20000 4276 4,68 1,40 1,10 16000 22400 97,90 60,50 8000 11200 72,60 119,90 1,30 29120 42,31 14560 92,23 85,25
CFM CFM56-3B2 22000 4301 5,12 1,40 1,10 10000 14000 80,30 68,20 6500 9100 80,30 130,90 1,30 18200 47,69 11830 100,69 80,30
CFM CFM56-3C1 23500 4301 5,46 1,40 1,10 7500 10500 93,50 90,20 5000 7000 84,70 141,90 1,30 13650 63,08 9100 109,15 89,10
CFM CFM56-5A1 25000 4995 5,01 1,20 1,09 8000 9600 100,28 178,76 1,46 14016 122,44 100,28
CFM CFM56-5A5 23500 4975 4,72 1,20 1,09 8000 9600 100,28 178,76 1,46 14016 122,44 100,28
CFM CFM56-5B3 32000 5250 6,10 1,50 1,09 7000 10500 119,90 95,92 5000 7500 119,90 171,13 1,22 12810 72,13 9150 140,27 119,90
CFM CFM56-5B5 22000 5250 4,19 1,80 1,09 15000 27000 113,36 80,66 10000 18000 94,83 109,00 1,00 27000 74,00 18000 109,00 104,10
CFM CFM56-5B6 23500 5250 4,48 1,80 1,09 15000 27000 113,36 80,66 10000 18000 94,83 109,00 1,00 27000 74,00 18000 109,00 104,10
CFM CFM56-5B7 27000 5250 5,14 1,80 1,09 14000 25200 113,36 80,66 9000 16200 94,83 109,00 1,00 25200 74,00 16200 109,00 104,10
CFM CFM56-7B18 18500 5216 3,55 1,80 1,18 18000 32400 92,04 71,98 13000 23400 86,14 101,48 1,00 32400 61,00 23400 101,48 89,09
CFM CFM56-7B20 20600 5216 3,95 1,80 1,18 17500 31500 92,04 71,98 13000 23400 86,14 101,48 1,00 31500 61,00 23400 101,48 89,09
CFM CFM56-7B22 22000 5216 4,22 1,80 1,18 17000 30600 92,04 71,98 13000 23400 86,14 101,48 1,00 30600 61,00 23400 101,48 89,09
CFM CFM56-7B24 24000 5216 4,60 1,80 1,18 16000 28800 93,22 75,52 12000 21600 93,22 90,86 1,00 28800 64,00 21600 90,86 93,22
CFM CFM56-7B26 26400 5216 5,06 1,80 1,18 13000 23400 107,38 75,52 11000 19800 107,38 105,02 1,00 23400 64,00 19800 105,02 107,38
CFM CFM56-7B27 27300 5216 5,23 1,80 1,18 10000 18000 93,22 90,86 8000 14400 93,22 142,78 1,00 18000 77,00 14400 142,78 93,22
GE CF34-3B1 9220 1670 5,52 1,16 1,10 15000 17400 50,05 51,70 8000 9280 72,60 103,40 1,50 26100 31,33 13920 68,93 61,33
GE CF34-8E5 13800 2470 5,59 1,28 1,07 8984 11500 70,62 65,27 5859 7500 70,62 142,31 1,39 15984 43,88 10424 102,38 70,62
GE CF34-8E5A1 15000 2470 6,07 1,28 1,07 7422 9500 70,62 79,18 5078 6500 70,62 164,78 1,39 13205 53,24 9035 118,55 70,62
GE CF34-10E5 18285 3700 4,94 1,28 1,07 14062 17999 64,20 50,29 7812 9999 64,20 139,10 1,39 25019 33,81 13899 100,07 64,20
GE CF34-10E6 19000 3700 5,14 1,28 1,07 13281 17000 69,55 53,50 7031 9000 69,55 148,73 1,39 23630 35,97 12510 107,00 69,55
GE CF34-10E7 20300 3700 5,49 1,28 1,07 10156 13000 73,83 57,78 4688 6001 73,83 169,06 1,39 18070 38,85 8341 121,63 73,83
IAE V2522-A5 22000 5230 4,21 1,90 1,01 11500 21850 112,11 67,67 9500 18050 112,11 136,35 1,00 21850 67,00 18050 136,35 112,11
IAE V2524-A5 24000 5139 4,67 1,90 1,01 11000 20900 112,11 70,70 9000 17100 112,11 143,42 1,00 20900 70,00 17100 143,42 112,11
IAE V2527-A5 27000 5139 5,25 1,90 1,01 8750 16625 136,35 94,94 7750 14725 136,35 164,63 1,00 16625 94,00 14725 164,63 136,35
IAE V2530-A5 30000 5139 5,84 1,90 1,01 7400 14060 121,20 104,03 6000 11400 121,20 168,67 1,00 14060 103,00 11400 168,67 121,20
IAE V2533-A5 33000 5139 6,42 1,90 1,01 7000 13300 121,20 109,08 5750 10925 121,20 175,74 1,00 13300 108,00 10925 175,74 121,20
GE CF6-80C2A5 61300 9389 6,53 1,00 1,10 5000 5000 198,00 550,00 1,63 8125 338,46 198,00
GE CF6-80C2A5 61300 9389 6,53 2,00 1,10 4500 9000 220,00 278,30 0,98 8820 283,98 220,00
GE CF6-80C2A3 60200 9360 6,43 3,00 1,10 4000 12000 209,00 199,10 0,80 9624 248,25 209,00
GE CF6-80C2A5 61300 9389 6,53 4,00 1,10 3500 14000 222,20 174,90 0,71 9884 247,73 222,20
GE CF6-80C2A2 53500 9360 5,72 3,00 1,10 6000 18000 198,00 135,30 5000 15000 198,00 176,00 0,80 14436 153,37 12030 219,45 132,00
PW PW4074 77440 14545 5,32 1,50 1,00 10000 15000 514,00 267,00 7500 11250 514,00 441,00 1,22 18285 219,03 13714 361,77 342,67
PW PW4077 79960 14545 5,50 1,50 1,00 9333 14000 514,00 286,00 6800 10200 514,00 410,00 1,22 17065 234,62 12434 336,34 342,67
PW PW4158 58000 9213 6,30 1,50 1,05 3300 4950 242,55 420,00 1,22 6034 344,54 242,55
GE CF6-80C2B6 60800 9670 6,29 6,00 1,10 2500 15000 253,00 163,90 1,00 15000 163,90 253,00
GE CF6-80C2B1F 58090 9499 6,12 7,00 1,10 2400 16800 209,00 146,30 0,93 15540 158,16 209,00
GE CF6-80C2D1F 61960 9850 6,29 7,00 1,10 2200 15400 213,40 168,30 0,93 14245 181,95 213,40
PW PW4052 52200 9213 5,67 3,00 1,05 4250 12750 157,50 159,60 1,40 17850 114,00 157,50
PW PW4056 56750 9213 6,16 8,00 1,05 1900 15200 197,40 138,60 0,88 13376 157,50 197,40
PW PW4060 60000 9213 6,51 7,00 1,05 2100 14700 252,00 139,65 0,93 13598 150,97 252,00
PW PW4062 62000 9213 6,73 7,00 1,05 2100 14700 252,00 139,65 0,93 13598 150,97 252,00
PW PW4168 68600 12400 5,53 7,00 1,00 2857 19999 418,00 135,00 2150 15050 418,00 247,00 0,93 18499 145,95 13921 267,03 278,67
PW PW4090 91790 15584 5,89 7,00 1,00 2643 18501 661,00 243,00 2000 14000 661,00 366,00 0,93 17113 262,70 12950 395,68 440,67
PW PW4098 99040 16500 6,00 7,00 1,00 2643 18501 661,00 243,00 2000 14000 661,00 366,00 0,93 17113 262,70 12950 395,68 440,67
PW PW2037 36600 7185 5,09 3,00 1,00 6667 20001 200,00 100,00 5167 15501 200,00 181,00 1,30 26001 76,92 20151 139,23 133,33
PW PW2040 40100 7185 5,58 3,00 1,00 6667 20001 200,00 110,00 4667 14001 200,00 200,00 1,30 26001 84,62 18201 153,85 133,33
GE CF6-80E1A2 67500 11162 6,05 3,00 1,00 4500 13500 370,00 170,00 3400 10200 370,00 284,00 1,40 18900 121,43 14280 202,86 246,67
GE CF6-80E1A2 67500 11162 6,05 6,00 1,00 2833 16998 343,00 147,00 2167 13002 343,00 235,00 1,00 16998 147,00 13002 235,00 228,67
GE CF6-80E1A2 67500 11162 6,05 8,00 1,00 2250 18000 343,00 172,00 1812 14496 343,00 217,00 0,88 15840 195,45 12756 246,59 228,67
GE GE90-85B 84700 15596 5,43 6,00 1,00 3000 18000 700,00 211,00 2417 14502 700,00 297,00 1,00 18000 211,00 14502 297,00 466,67
GE GE90-110B 110000 18260 6,02 8,00 1,00 2450 19600 807,00 215,00 2100 16800 807,00 278,00 0,88 17248 244,32 14784 315,91 538,00
GE GE90-110B 110000 18260 6,02 10,00 1,00 2150 21500 807,00 200,00 1850 18500 807,00 240,00 0,84 18060 238,10 15540 285,71 538,00
PW PW4168 68600 12400 5,53 6,00 1,00 3000 18000 375,00 140,00 2333 13998 375,00 243,00 1,00 18000 140,00 13998 243,00 250,00
PW PW4168 68600 12400 5,53 3,00 1,00 4500 13500 400,00 178,00 3500 10500 400,00 292,00 1,40 18900 127,14 14700 208,57 266,67
PW PW4168 68600 12400 5,53 8,00 1,00 2750 22000 375,00 123,00 2100 16800 375,00 233,00 0,88 19360 139,77 14784 264,77 250,00
PW PW4077 79960 14545 5,50 8,00 1,00 2750 22000 389,00 173,00 2250 18000 389,00 244,00 0,88 19360 196,59 15840 277,27 259,33
PW PW4090 91790 15584 5,89 8,00 1,00 2500 20000 556,00 200,00 2125 17000 556,00 279,00 0,88 17600 227,27 14960 317,05 370,67
RR RB.211-535E4 40100 7264 5,52 3,00 1,00 6000 18000 185,00 190,00 1,40 25200 135,71 185,00
RR Trent 772-60 71100 10550 6,74 3,00 1,00 5333 15999 330,00 250,00 4000 12000 330,00 367,00 1,40 22399 178,57 16800 262,14 220,00
RR Trent 772-60 71100 10550 6,74 6,00 1,00 4000 24000 330,00 215,00 2667 16002 330,00 275,00 1,00 24000 215,00 16002 275,00 220,00
RR Trent 772-60 71100 10550 6,74 8,00 1,00 3750 30000 330,00 191,00 2500 20000 330,00 220,00 0,88 26400 217,05 17600 250,00 220,00
RR Trent 884-17 86910 13100 6,63 8,00 1,00 3000 24000 429,00 196,00 2375 19000 429,00 255,00 0,88 21120 222,73 16720 289,77 286,00
RR Trent 895-17 95000 13186 7,20 8,00 1,00 3000 24000 698,00 200,00 2375 19000 698,00 263,00 0,88 21120 227,27 16720 298,86 465,33
PW JT9D-7R4E 50000 8905 5,61 3,50 1,29 1857 6500 144,48 366,36 144,48
PW JT9D-7J 50000 8850 5,65 7,00 1,29 1350 9450 144,48 237,36 144,48
PW JT9D-7Q 53000 9295 5,70 5,00 1,29 1600 8000 144,48 335,40 144,48
PW JT9D-7Q 53000 9295 5,70 8,00 1,29 1150 9200 199,95 245,10 199,95
PW JT9D-7R4G2 54750 9135 5,99 6,00 1,29 1500 9000 201,24 279,93 201,24
PW JT9D-7R4G2 54750 9135 5,99 8,00 1,29 1250 10000 216,72 251,55 216,72
GE CF6-50C2 52500 8731 6,01 5,00 1,29 1500 7500 185,76 258,00 185,76
GE CF6-50E2 52500 8768 5,99 6,00 1,29 1500 9000 135,45 236,07 135,45
GE CF6-50C2 52500 8731 6,01 2,00 1,29 2500 5000 180,60 451,50 180,60
PW PW4158 58000 9213 6,30 3,00 1,29 3300 9900 189,63 238,65 189,63
PW PW4060 60000 9213 6,51 6,50 1,29 1850 12025 243,81 180,60 243,81
PW PW4062 62000 9213 6,73 7,00 1,29 1700 11900 261,87 180,60 261,87
PW PW4056 56750 9213 6,16 8,00 1,29 1750 14000 175,44 161,25 175,44
GE CF6-80C2A2 53500 9360 5,72 1,00 1,29 5000 5000 258,00 490,20 258,00
GE CF6-80C2A3 60200 9360 6,43 3,00 1,29 2500 7500 220,59 309,60 220,59
GE CF6-80C2B1F 58090 9499 6,12 7,50 1,29 1700 12750 223,17 168,99 223,17
GE CF6-80C2D1F 61960 9850 6,29 7,00 1,29 1500 10500 216,72 196,08 216,72
Short-Haul Three-Spool
Medium-Long-Haul Old and Newly marketed Engines

C Regression Analysis
SE_Data- Fit Least Squares
Response
C.1 First Interval SH1stEngines
Interval adj
Whole Model
35000
30000
1st Interval
adj Actual
25000
20000
15000
10000 20000 25000 30000

1st Interval adj Predicted
P<.0001 RSq=0,77 RMSE=3247,8
Summary of Fit
RSquare 0,769685
Sum of
Model 3 846039131 282013044 26,7350
Error 24 253162938 10548456 Prob > F
C. Total 27 1099202069 <,0001*
Parameter Estimates
Intercept 68466,325 5575,065 12,28 <,0001*
TWR -8267,819 942,4547 -8,77 <,0001*
weight -1,004437 0,435151 -2,31 0,0299*
(weight-5407)*(weight-5407) 0,0001212 5,961e-5 2,03 0,0531
Effect Tests
Sum of
Prediction Function
Source Nparm DF Squares F Ratio Prob > F
TWR
Interval f irst,SH = 1 1
68466.325284 −811801410
8267.81904 · T W R −<,0001*
76,9593 1.00444 · weight
weight 1 1 56202317 5,3280 0,0299*
weight*weight + (weight
1 −
1 5407) · [(weight4,1379
43648191 − 5407) ·0,0531
0.00012125]
6000
4000
2000
j Residual
st Interval
vi
0
-2000
C.2 Mature Interval SH Engines vii
SE_Data- Fit Least Squares
Response
C.2 Mature Interval SHmature Interval adj
Engines
Whole Model TWR
Actual by Predicted Plot Leverage Plot
25000 25000
Leverage Residuals
mature Interval adj
20000 20000
mature Interval
adj Actual
15000 15000
10000 10000
5000 5000
5000 10000 15000 20000 25000 3,5 4,0 4,5 5,
mature Interval adj Predicted TWR Leve
P<.0001 RSq=0,79 RMSE=2112,5
Summary of Fit
RSquare 0,790264
Sum of
Model 1 554897670 554897670 124,3409
Error 33 147269489 4462711,8 Prob > F
C. Total 34 702167159 <,0001*
Lack Of Fit
Sum of F Ratio
Source DF Squares Mean Square 11,6996
Lack Of Fit 31 146461838 4724575 Prob > F
Pure Error 2 807651 403825 0,0817
Total Error 33 147269489 Max RSq
0,9988
Parameter Estimates
Intercept 40684,376 2392,421 17,01 <,0001*
TWR -5022,812 450,443 -11,15 <,0001*
Effect Tests
Sum of
Prediction Function
TWR 1
Interval mature,SH 1
= − 5022.8116
554897670 124,3409
40684.37633 · TWR
<,0001*
5000
4000
3000
mature Interval
2000
adj Residual
1000
0
-1000
-2000
-3000
-4000
-5000
5000 10000 15000 20000 25000
mature Interval adj Predicted
C.3 First Interval MLH Engines viii
BE_Data_wo_3s- Fit Least Squares
Response
C.3 First Interval 1st interval
MLH Enginesadj
Whole Model
28000
26000
24000
1st interval
adj Actual
22000
20000
18000
16000
14000
14000 18000 22000 26000
1st interval adj Predicted
P<.0001 RSq=0,91 RMSE=969,3
Summary of Fit
RSquare 0,911592
RSquare Adj 0,88949
Sum of
Model 3 116253879 38751293 41,2449
Error 12 11274503 939541,89 Prob > F
C. Total 15 127528381 <,0001*
Lack Of Fit
Sum of F Ratio
Lack Of Fit 5 5046845 1009369 Prob > F
Pure Error 7 6227658 889665 0,4233
0,9512
Parameter Estimates
Intercept 22539,976 1210,754 18,62 <,0001*
thrust -0,314694 0,078371 -4,02 0,0017*
weight 1,4329074 0,496822 2,88 0,0137*
(thrust-76305)*(thrust-76305) 3,4421e-6 4,789e-7 7,19 <,0001*
Effect Tests
Sum of
Prediction Function
Interval f irst,M LH = 22539.9757 + 1.4329 · weight − 0.3147
thrust 1 1 15149153 16,1240 0,0017*· thrust
weight 1 1 7815403 8,3183 0,0137*
thrust*thrust 1 − 48533396
+1(thrust 76305) · [(thrust − 76305)
51,6564 · 0.0000034421]
<,0001*
C.4 Mature Interval MLH Engines ix
BE_Data_wo_3s- Fit Least Squares
Response
C.4 Mature Interval3rd
MLHinterval adj
Engines
Whole Model
21000
20000
19000
3rd interval
18000
adj Actual
17000
16000
15000
14000
13000
12000
12000 15000 17000 19000
3rd interval adj Predicted
P<.0001 RSq=0,80 RMSE=880,27
Summary of Fit
RSquare 0,798364
Sum of
Model 3 58293477 19431159 25,0764
Error 19 14722696 774878,73 Prob > F
C. Total 22 73016173 <,0001*
Lack Of Fit
Sum of F Ratio
Lack Of Fit 12 10460068 871672 Prob > F
Pure Error 7 4262628 608947 0,3263
0,9416
Parameter Estimates
Intercept 34415,709 3013,422 11,42 <,0001*
TWR -2759,253 485,7675 -5,68 <,0001*
weight -0,366246 0,06245 -5,86 <,0001*
(weight-12072)*(weight-12072) 0,0001018 1,816e-5 5,61 <,0001*
Effect Tests
Sum of
Prediction Function
Interval mature,M LH = 34415.70939 − 2759.25322 · T W R<,0001*
TWR 1 1 25001151 32,2646 − 0.36625 · weight
weight 1 1 26650979 34,3937 <,0001*
weight*weight 1 −24347698
+1(weight 12072) · [(weight
31,4213− 12072) · 0.000101795]
<,0001*
C.5 First Shop Visit Restoration Costs x
Base_Data- Fit Least Squares
C.5 First ShopResponse 1st cost adj

Visit Restoration Costs
Whole Model thrust
250 250
Leverage Residuals
200 200
1st cost adj

adj Actual
1st cost
150 150
100 100
50 50
50 100 150 200 250 0 20000 5000

1st cost adj Predicted thrust Leve
P<.0001 RSq=0,93 RMSE=20,538
Summary of Fit
RSquare 0,925128
Sum of
Model 1 244971,87 244972 580,7369
Error 47 19825,98 422 Prob > F
C. Total 48 264797,85 <,0001*
Lack Of Fit
Sum of F Ratio
Lack Of Fit 30 14101,817 470,061 Prob > F
Pure Error 17 5724,161 336,715 0,2371
Total Error 47 19825,979 Max RSq
0,9784
Parameter Estimates
Intercept 7,1451068 5,602079 1,28 0,2084
thrust 0,0023619 0,000098 24,10 <,0001*
Effect Tests
Sum of
Prediction Function Source Nparm DF Squares F Ratio Prob > F
thrustRCostf irst1 = 7.14511
1 + 0.002361887
244971,87 · thrust
580,7369 <,0001*
50
40
30
20
adj Residual
1st cost
10
0
-10
-20
-30
-40
-50
50 100 150 200 250
1st cost adj Predicted
C.6 Mature Shop Visit Restoration Costs xi
Base_Data- Fit Least Squares 2
Response
C.6 Mature Shop Visit mature cost adj
Restoration Costs
Whole Model thrust
400 400
350 350
Leverage Residuals
mature cost adj
mature cost 300 300
adj Actual 250 250
200 200
150 150
100 100
50 50
50 100 150 200 250 300 350 400 0 20000 50000
mature cost adj Predicted thrust Levera
P<.0001 RSq=0,81 RMSE=38,591
Summary of Fit
RSquare 0,807809
Sum of
Model 1 388089,29 388089 260,5955
Error 62 92332,90 1489 Prob > F
C. Total 63 480422,19 <,0001*
Lack Of Fit
Sum of F Ratio
Lack Of Fit 41 78144,940 1905,97 Prob > F
Pure Error 21 14187,961 675,62 0,0065*
0,9705
Parameter Estimates
Intercept 46,528678 10,12921 4,59 <,0001*
thrust 0,0028861 0,000179 16,14 <,0001*
Effect Tests
Sum of
Prediction FunctionSource Nparm DF Squares F Ratio Prob > F
thrustRCostmature
1 = 46.52868
1 260,5955 · thrust
+ 0.002886118
388089,29 <,0001*
100
adj Residual
mature cost
50
-50
-100
50 100 150 200 250 300 350 400
mature cost adj Predicted
C.7 LLP Cost xii
LLP- Fit Least Squares
C.7 LLPResponse
Cost LLP Reserves
Whole Model
900
800
700
LLP Reserves
600
Actual
500
400
300
200
100
0
0 100 300 500 700 900
LLP Reserves Predicted
P<.0001 RSq=0,95 RMSE=44,13
Summary of Fit
RSquare 0,952884
Sum of
Model 3 2363135,5 787712 404,4799
Error 60 116848,1 1947 Prob > F
C. Total 63 2479983,6 <,0001*
Lack Of Fit
Sum of F Ratio
Lack Of Fit 48 101360,93 2111,69 Prob > F
Pure Error 12 15487,16 1290,60 0,1780
0,9938
Parameter Estimates
Intercept -115,3133 13,6202 -8,47 <,0001*
weight 0,0194512 0,007095 2,74 0,0080*
thrust 0,0031206 0,001069 2,92 0,0049*
(weight-8608,78)*(weight-8608,78) 2,6924e-6 3,188e-7 8,44 <,0001*
Effect Tests
Sum of
Prediction Function
weight= − 115.313261 + 0.0194512
LLP Cost 1 · weight 7,5152
14635,69 + 0.0031206 · thrust
0,0080*
thrust 1 1 16609,48 8,5288 0,0049* −6
+ (weight −1 8608.78125)
weight*weight 1 · ((weight
138881,53 − 8608.78125)
71,3139 <,0001*· 2.69234 · 10 )
D Model Parameters
D.1 Averaged Short-Haul-Engine Severity Curve
Averaged SH-Engine Severity Curve

3
2.5
Derate: 0%
2 Derate: 5%
Severity Factor
Derate: 10%
1.5 Derate: 15%
Derate: 20%
0.5
0
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
Cycle Time [h]
Flight Time [h] 0% 5% 10% 15% 20%
0.5 2.800 2.600 2.400 2.280 2.160

1.0 2.100 1.925 1.750 1.645 1.540
1.5 1.600 1.450 1.300 1.210 1.120
1.9 1.240 1.120 1.000 0.940 0.880
2.5 1.000 0.910 0.860 0.792 0.744
3.0 0.920 0.840 0.780 0.738 0.696
4.0 0.826 0.766 0.706 0.670 0.634
5.0 0.770 0.715 0.660 0.627 0.594
6.0 0.740 0.685 0.630 0.597 0.564
xiii
D.2 Averaged Medium-Long-Haul-Engine Severity Curve xiv
D.2 Averaged Medium-Long-Haul-Engine Severity Curve
Averaged MLH-Engine Severity Curve

3
2.5
Derate: 0%
2 Derate: 5%
Severity Factor
Derate: 10%
1.5 Derate: 15%

Derate: 20%
0.5
0
1 2 3 4 5 6 7 8 9 10 11 12
Cycle Time [h]
Flight Time [h] 0% 5% 10% 15% 20%
1.0 2.800 2.500 2.200 2.020 1.900

2.0 2.000 1.850 1.700 1.610 1.520
3.0 1.600 1.500 1.400 1.340 1.280
4.0 1.405 1.315 1.225 1.171 1.117
5.0 1.260 1.180 1.100 1.052 1.004
6.0 1.140 1.070 1.000 0.958 0.916
7.0 1.045 0.985 0.925 0.889 0.853
8.0 0.990 0.935 0.880 0.847 0.814
9.0 0.970 0.915 0.860 0.827 0.794
10.0 0.940 0.890 0.840 0.810 0.780
11.0 0.920 0.870 0.820 0.790 0.760
12.0 0.890 0.845 0.800 0.773 0.746
D.3 Time & Material Factor Curves xv
D.3 Time & Material Factor Curves
SH Time&Material Factor Curve
1.15
1.1
1.05
T&M Factor
0.95
0.9
0.85
0.8
0.75
0.7
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
Cycle Time [h]
EFH:EFC 0.5 1.0 1.5 1.9 2.5 3.0 4.0 5.0 6.0
T&M Factor 0.90 0.95 0.98 1.00 1.02 1.03 1.04 1.05 1.06
MLH Time&Material Factor Curve
1.15
1.1
1.05
T&M Factor
0.95
0.9
0.85
0.8
0.75
0.7
1 2 3 4 5 6 7 8 9 10 11 12
Cycle Time [h]
EFH:EFC 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0
T&M Factor 0.85 0.91 0.94 0.96 0.98 1.00 1.03 1.05 1.07 1.09 1.10 1.11
E Model Analysis
Sensitivity Analysis for MLH engines
x 10
4 Thrust - Interval Sensitivity x 10
4 Weight - Interval Sensitivity
2 2.1
First-Run First-Run
2
1.9 Mature-Run Mature-Run
1.9
1.8
1.8
Interval [EFH]
Interval [EFH]
1.7
1.7
1.6
1.6
1.5 1.5
1.4
1.4
EFH:EFC = 6[h] EFH:EFC = 6[h]
Derate = 10% 1.3 Derate = 10%
1.3 Thrust = 78000[lbf ]
Weight = 14545[lbs]
1.2
7.4 7.6 7.8 8 8.2 8.4 8.6 1.2 1.25 1.3 1.35 1.4 1.45 1.5 1.55 1.6
Thrust [lbf ] 4
x 10 Weight [lbs] x 10
4
x 10
4 EFH:EFC - Interval Sensitivity x 10
4 Derate - Interval Sensitivity
2.4 2.1
First-Run First-Run
2
2.2 Mature-Run Mature-Run
1.9
2
Interval [EFH]
Interval [EFH]
1.8
1.8
1.7
1.6 1.6
1.4 1.5
Derate = 10% 1.4 EFH:EFC = 6[h]

1.2
Thrust = 78000[lbf ] Thrust = 78000[lbf ]
Weight = 14545[lbs] 1.3 Weight = 14545[lbs]
1
3 4 5 6 7 8 9 10 11 12 0 5 10 15 20
xvi
E Model Analysis xvii
Thrust - SVC per EFH — Sensitivity Weight - SVC per EFH Sensitivity
380
380 First-Run First-Run
Mature-Run 360 Mature-Run
360
SVC per EFH [$/EFH]
SVC per EFH [$/EFH]

340
340
320
320
300
300
280
280
EFH:EFC = 6[h] 260
EFH:EFC = 6[h]
Derate = 10% Derate = 10%
260 Thrust = 78000[lbf ]
Weight = 14545[lbs]
240
7.4 7.6 7.8 8 8.2 8.4 8.6 1.2 1.25 1.3 1.35 1.4 1.45 1.5 1.55 1.6
Thrust [lbf ] 4
x 10 Weight [lbs] 4
x 10
EFH:EFC - SVC per EFH Sensitivity Derate - SVC per EFH Sensitivity
550 400
First-Run First-Run
500 Mature-Run Mature-Run
SVC per EFH [$/EFH]
SVC per EFH [$/EFH]
450
350
400
350
300 300
250
Derate = 10% EFH:EFC = 6[h]
200 Thrust = 78000[lbf ] Thrust = 78000[lbf ]
Weight = 14545[lbs] 250 Weight = 14545[lbs]
3 4 5 6 7 8 9 10 11 12 0 5 10 15 20

Modeling The Life Cycle Cost of Jet Engine Maintenance

Uploaded by

Copyright:

Available Formats

Modeling The Life Cycle Cost of Jet Engine Maintenance

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Modeling The Life Cycle Cost of Jet Engine Maintenance

Uploaded by

Copyright:

Available Formats

Reglerentwurf zur Synchronisation einer hybriden

Modeling the Life Cycle Cost

Students Research Project

List of Symbols VII

3 Development of Cost Estimating Relationships 32

4 Modeling of Engine Maintenance Cost 46

5 Implementation into existing LCC-Tool 68

6 Summary and Conclusion 75

A Maintenance Costs iii

B.2 Classification of Aircraft Engines . . . . . . . . . . . . . . . . . . . . . . . . . . . iv

D Model Parameters xiii

E Model Analysis xvi

MRO . . . . . . . . . . . . . . . Maintenance, Repair and Overhaul

Symbol Unit Description

BaseInterval EF H Base interval between shop visits determined

EF H LC EF H Accumulated engine flight hours throughout the

Interval EF H Interval length between shop visits

SEC U SD Costs incurred by providing spare engines

SV C U SD Total costs for a shop visit

SV C LC U SD Accumulated engine shop visit costs throughout

SV RC U SD Total costs for the engine’s performance restora-

U til ann EF H Annual utilization of an engine

Y ears LC years Number of years of the engine life cycle

Figure 1.1: Aircraft MRO cost overview [Jet08]

1.1 Thesis Objectives

1.2 Thesis Structure

2.1 Aircraft Gas Turbine Engines

2.1.1 Basic Principles

Figure 2.1: Engine core of gas turbine engines

Engine DMC [$/EFH]

Corner Point SLOATL Engine T

the air flow

working principle. These types are briefly outlines

eteriorated Engine The loss of efficiency has

2.1.2 The Turbofan Engine

Figure 2.3: GEnx-2B - high bypass twin-spool turbofan

LPC + Fan Turbines Turbines

Accessory Gearbox Module

2.1.2.2 Engine Operating Parameters

EGT Margin (EGTM) fuel of an engine’s performance degradation,

Engine DMC [$/EFH]

Corner Point SLOATL OAT [C°] En

Figure 2.6: Correlation between Take-Off EGT and OAT [Air06b]

OAT can be determined by calculating the SLOATL.

2.2 Aircraft Engine Maintenance

2.2.1 On-Wing Engine Maintenance

2.2.1.1 Engine Condition Monitoring (ECM)

2.2.1.2 On-Wing Repair and Replacement

2.2.2 Engine Overhaul - Shop Visit

2.2.2.1 Main Causes of Engine Removals

As previously mentioned, a modern aircraft engine’s condition is constantly monitored. This

Figure 2.7: Removal causes depending on aircraft operation [Ack10]

Engine DMC [$/EFH]

Redline EGT The engine becomes less

efficient, due to wear of

New Engine The increase in fuel burn

Figure 2.8: Effects of engine wear on the EGT Margin [Ack10]

Engine Flight Cycles Engine Time On-Wi

Engine Flight Cycles En