Skip to main content

Full text of "DTIC ADA1036239: Estimating Aircraft Depot Maintenance Costs"

See other formats


ADA103623 



Estimating Aircraft Depot 
Maintenance Costs 


Kenneth E. fy/^arks, Ronald W. Hess 



J 

t 

This Document Contains 










The resfearnh described in this report was sponsored by t^'e Office of the 
Assistant Secretary of Defense/Program Analysis and Evaluation under Con¬ 
tract No. MDA903-77-C-0107. 


Library r' Congress Cataloging in Publication Data 

Marks, Kenneth E., 1945- 

Estimating aircraft depot maintenance costs. 

"R-2751-PA&E." 

1. United States, Air Force—Management, 

2. Airplanes, Military—Maintenance and repair-- 
Estimates. I. Hess, Ronald Wayne 
II. United States. Office of the Assistant 
Secretary of Defense (Program Analysis and Evalua¬ 
tion) in. Rand Corporation. IV, Title. 
UG1245.M57 558.4*185 81-11872 

ISBN 0-8550-0555-6 AACR2 


The Rand Publications Series: The Report is the principal publication doc¬ 
umenting and transmitting Rand’s major research findings and final research 
resiilts. The Rand Note reports other outputs of sponsored research for 
general distribution. Publications of The Rand Corporation do not neces¬ 
sarily reflect the opinions or policies of the sponsors of Rand research. 


Published by The Rand Corporation 








eCURi'tY CLASSIFICATION OF THIS PAGE (Wh^n Dmf Enffd) 


1 REPORT DOCUMENTATION PAGE 

READ INSTRUCTIONS 

BEFORE COMPLETING FORM 

: RE'^CRT NUMBER 

R-2731-PA&E 

Z. GOVT ACCESSION NO. 

3. REClPlE^’S CATALOG NUMBER 

4. title fand Sub(//le) 

Estimating Aircraft Depot Maintenance Costs 

S. TYPE OF REPORT 8 PERIOD COVERED 

Interim 

6. PERFORMING ORG. REPORT NUMBER 

7 AUTHORf*; 

Kenneth E. Marks, Ronald W. Hess 

B. CONTRACT OR GRANT NUMBERfaJ 

MDA903-77-C-0107 

9 PERFORMING ORGANIZATION NAME AND ADDRESS 

The Rand Corporation 

1700 Main Street 

Santa Monica, California 90401 

10. PROGRAM ELEMENT, PROJECT, TASK 
AREA 8 WORK UNIT NUMBERS 

1 1 CONTROLLING OFFICE NAME AND ADDRESS 

Director, Program Analysis- & Evaluation 

Office, Secretary of Defense 

Washington, D.C. 20301 

12. REPORT DATE 

July 1981 

*3. NUMBER OF PAGES 

195 

14 MONITORING AGENCY NAME 8 AOORESSff/d//rar*n< ttom Controlling Otliro) 

IS. SECURITY CLASS, (ol Ihio ttpottj 

UNCLASSIFIED 

IS*. DECLASSIFICATION/DOWNGRAOING 
SCHEDULE 

IS. DISTRIBUTION STATEMENT (ol thit Roport) 

Approved for Public Release; Distribution Unlimited 


t7. DISTRIBUTION STATEMENT (oi th^ sbttrmct tntwrmd in Block 20, H dilicront from Roporl) 

No Restrictions 


st?"^ 


^ 98 '' ‘'.'® 



8 supplementary notes 




19. KEY WORDS (Cont/nuo on reverse aid* it noc9*9Ty *nd identity by block numbmr) 

Shop Maintenance Logistics Planning 

Cost Estimates 

Aircraft 

Parametric Equations 


20 ABSTRACT (Contlnu* on rover** aid* it nocoooory ond Idontlfy by block numbor) 

See Reverse Side 


DD ,:°r73 1473 


UNCLASSIFIED 


SECURITY CLASSIFICATION OF THIS PAGE fWJitn D,f Bnffi) 






UNCLASSIFIED _ 

Y CL»SSI^ ICATION OF This PACEfHTiwi Dmia Enlatmd) 


Describes a series of parametric equations for use 
in estimating the depot maintenance cost of new 
Air Force aircraft, particularly for the five major 
maintenance categories: airframe rework, engine 
overhaul, airframe component repair, engine component 
and accessory repair, and avionics component repair. 
The equations are intended to provide cost estimates 
for Defense Systems Acquisitions Review Council 
Milestone II, at which point some design details of 
major aircraft subsystems (airframes, engine avionics) 
are available. The report presents a single set of 
equations that are the most representative and appli¬ 
cable to the widest range of estimating situations, 
but presents alternative equations and supporting data 
and analyses“use by the interested reader. (See 
also R-2552-pi&E.) (WH) 


SECURITY CLASSlFICATtON OF THIS PAGEnr?i*n Dmtm 













6^L 


// / / / R-2731 


■P^ 




t 

^imating^rcraftDepot ^ 
fiiiaintenance Costs. 




■]- . - f - 

Kenneth E/Marks/Ronald w/Hess 

. 


(/! 


Z-i 


July/^981 / 


C4S/ V / 


A report prepared for the 

Office of the Assistant Secretary of Defense/ 

Program Analysis and Evaluation 


/ 


(S 


I f 


/ ' 


f *>' 


_V'v' 


Ramd 

SANTA MONICA. CA •'OAOS 


DlSTRlBUnCM a'A'ltMCTTA 

Approved for public relaoM; 
IKstiibuticm Unlimited 


Li.' '/ 







iii 


PREFACE 

This report presents the results of Rand research on parametric 
methods for estimating the annual depot maintenance cost of new Air 
Force aircraft. The research focused on methods suitable for use at or 
near Defense Systems Acquisition Review Council (DSARC) Milestone II— 
that is, prior to the initiation of Full Scale Development. These 
methods make use of information about the design and maintenance 
characteristics of new aircraft to provide estimates suitable for life 
cycle cost analysis and planning studies. The methods are not 
intended for detailed programming and budgeting of depot maintenance 
activities. 

The work documented here was sponsored by the Office of the 
Assistant Secretary of Defense, Program Analysis and Evaluation 
(PA&E). The results should be of interest to cost analysts in both 
the OSD and the Air Force system acquisition and logistics communities. 

A related cost often associated with depot-level aircraft 
maintenance requirements is that of recoverable spares investment. 
Recent results on this subject are published in K. J. Hoffmayer, 

F. W. Finnegan, Jr., and W. H. Rogers, Estimating USAF Aircraft 
Recoverable Spares Investment , R-2552-PA&E, The Rand Corporation, 

August 1980. 

Accession For 

NT IS GP.A&I 

DTic t;.b □ 

U-jsnncanced □ 

Jus t i f i cat i on__ 

By- 

_pi»tributlon/ 
i Availability Codes 
I and/or 

:Dist ' Special 








V 

SUMMARY 


This report describes a series of parametric equations for use iq 
estimating the depot maintenance cost of new Air Force aircraft. The 
equations are intended to provide cost estimates for Defense Systems 
Acquisition Review Council Milestone II. At that point in time, some 
design details of major aircraft subsystems (airframe, engine, avionics) 
are available, but existing estimating equations reportedly do not make 
use of this information. This work specifically sought to include 
subsystem-level, as well as system-level, parameters in the equations 
developed. 

Experienced logisticians at three Air Force depot maintenance 
facilities were consulted to identify factors that affect the depot 
maintenance cost incurred by aircraft weapon systems. Their views were 
combined with knowledge accumulated at Rand during previous studies to 
form the basis for selection of potential explanatory variables for use 
in the parametric estimating equations. 

Data on the depot maintenance cost of most major USAF aircraft and 
aircraft engines for fiscal years 1975 through 1977 were obtained from 
the Air Force Logistics Command. The primary data source was the 
Weapon System Cost Retrieval System. Data were also obtained from the 
standard AFLC supply and maintenance cost reporting systems (D041 and 
H036B). The data were analyzed in conjunction with data on potential 
explanatory variables at both the system and subsystem levels. The 
analyses centered on the development of useful estimating relationships 
for the five major categories of depot maintenance: airframe rework, 
engine overhaul, airframe component repair, engine component and 
accessory repair, and avionics component repair. 

Equations were developed that relate airfram^ rework cost to 
flying hours, aircraft empty weight, depot production quantity, 
programmed depot maintenance policy, airframe manufacturing cost, 
aircraft age, and the percent of work performed organically. Similar 
equations for the depot-level repair of airframe components incorporate 






















empty weight, airframe manufacturing cost, sortie rate, cr»d a variable 
that denotes whether or not the aircraft engine has an afterburner. 

Equations developed for maintenance work on whole engines and 
engine components and accessories make use of turbine inlet 
temperature, pressure, specific fuel consumption, engine weight, 
thrust, removal rate, selling price, and variables that distinguish the 
aircraft mission, single versus multiple engine applications, operation 
by active versus reserve/guard units, and organic versus contract 
maintenance. 

Avionics component repair costs are estimated by equations based 
on the avionics suite weight, the number of suite black boxes, the 
number of suite functions, the mean time between suite demands, the 
suite procurement cost, sortie rate, and mission and all-weather 
capability designators. 

Several estimating equations are potentially useful for each of the 
depot maintenance categories. A single set of equations is presented as 
being, in our judgment, the most representative and applicable to the 
widest range of estimating situations. The alternative equations and 
supporting data and analyses are presented in the report, however, for 
use by the interested reader. 

Several issues that are beyond the scope of this study should be 
addressed in future research. Chief among these are the effects of 
recent changes in aircraft and engine design practices on depot 
maintenance costs. Our data did not, for example, permit an analysis of 
the implications of engine modularity (as in the FlOO engine) or of the 
increased use of composite materials (as in the F-15 and F-16). 
Similarly, we were unable to examine the effect of aircraft age on the 
cost of airframe rework and engine overhaul. Also, a detailed analysis 
of the basic H036 data could evaluate some data that were not included 
in the WSCRS data files. For example, data identifying individual 
facilities could be very useful in studies of maintenance concepts, 
indirect costs, or the relationships between the composition (and cost) 
of the labor force and the nature of the work performed. 














S 5 


Vll 


ACKNOWLEDGMENTS 


This study -ould not have been conducted without the cooperation of 
a large n';:.ier of Air Force Logistics Command personnel. Personnel in 
the Direc 1 - cates of Maintenance, Materiel Management, and Plans and 
Programs at the Ogden, San Antonio, and Warner Robins Air Logistics 
Centers provided advice and insights that were important in the 
development of hypotheses. All of the cost data used were provided by 
the headquarters. Spe;.ial thanks are due to Roger Steinlage, Robert 
Bou'.ais (AFL( 'ACMCC), and Captain John Wallace (formerly of AFLC/ACMCC) 
for suoplying data ft -m the Weapon System Cost Retrieval System and 
’■irovj life -information on the system periodically during the study. 


-9 

I 


4 

A 


I 

4 


> 

I 


I 






CONTENTS 


PREFACE. iii 

SUMMARY. V 

ACKNOWLEDGMENTS. vii 

LIST OF TABLES. xi 

LIST OF FIGURES. xiii 

GLOSSARY. xvii 

Section 

I. INTRODUCTION AND OVERVIEW. 1 

Approach au<l Principal Results. 2 

Other Results... 6 

Report Organization. 7 

II. CATEGORIES OF DEPOT MAINTENANCE ACTIVITY. 8 

Airframe Rework. 9 

Engine Overhaul... 11 

Component Repair... 12 

Support Equipment Maintenance. 14 

Common Considerations Affecting Depot Maintenance.. 16 

III. DATA BASE AND ANALYTICAL APPROACH. 18 

Cost Data. 18 

Explanatory Variable Data. 21 

Common Aspects of Analytical Approach. 38 

IV. DEVELOPMENT OF ESTIMATING EQUATIONS. 42 

Airframe Rework Analysis. 42 

Engine Overhaul and Repair Analysis . 50 

Airframe Component Repair Analysis. 62 

Engine Component and Accessory Repair Analysis. 70 

Avionics Component Repair Cost. 74 

Total Cost Equations. 80 

V. SUMMARY OF RESULTS AND CONCLUSIONS. 84 

Principal Findings. 84 

Application of Estimating Relationships. 89 

Possible Improvements. 91 

Appendix 

A. DEFINITIONS OF TERMS AND VARIABLES. 95 

B. DATA PROCESSING. 108 

C. DEPOT MAINTENANCE COST DATA. Ill 

D. EXPLANATORY VARIABLE DATA. 121 

E. DATA PLOTS. 127 

F. NOTE ON AlxaRAME REWORK COST. 181 













































-•*« 


XI 


TABLES 


1. Representative Set of Cost Estimating Relationships. 5 

2. Typical Annual Depot Maintenance Costs Per Aircraft. 8 

3. Potential Explanatory Variables for Airframe Rework. 23 

4. Potential Explanatory Variables for Engine Depot Over¬ 

haul and Repair Cost Elements...... 30 

5. Potential Explanatory Variables for Avionics Component 

Repair... 34 

6. Airframe Rework Costs: Averages for 1975-1977. 43 

7. Airframe Rework Cost Per Aircraft Estimating Relation¬ 

ships: Total Sample. 47 

8. Airframe Rework Cost Per Aircraft Estimating Relation¬ 

ships: Most Representative Series. 48 

9. Summary Engine Data by TMS: Averages for 1975-1977. 51 

10. Engine ATBO Estimating Relationships. 58 

11. Engine Cost Per Overhaul Estimating Relationships. 59 

12. Engine Annual Cost to Repair Estimating Relationships... 61 

13. Airframe Component Costs: Averages for 1975-1977. 63 

14. Airframe Component Repair Cost Estimating Relationships: 

Total Sample. 66 

15. Airframe Component Repair Cost Estimating Relationships: 

Most Representative Series. 67 

16. Airframe Component Repair Cost Estimating Relationships: 

Mission Subsamples. 68 

17. Engine Component and Accessory Repair Cost: Averages 

for 1975-1977. 71 

18. Engine Annual Component and Accessory Repair Cost 

Estimating Relationships. 73 

19. Avionics Component Repair Cost: Averages for 1975-1977. 75 

20. Avionics Component Repair Cost Estimating Relationships. 78 

21. Average Depot Maintenance Cost: 1975-1977. 81 

22. Total System Depot-Level Cost Estimating Relationships.. 81 

23. Summary of Results for Airframe Variables. 86 

24. Summary of Results for Engine Variables. 87 

25. Summary of Results for Avionics Variables. 88 

26. Alternative Total Cost Estimates Per Aircraft. 90 

A.l Definition of Airframe Components. 97 

A.2 Definition of Engine Components and Accessories. 97 

A.3 Definition of Avionics Components. 98 

A.4 Definition of Armament Components. 98 

A.5 Definition of Support Equipment. 99 

C.l Airframe Rework Total Cost Data. 112 

C.2 Elements of Airframe Rework Cost Data. 113 

C.3 Elements of Airframe Component Repair Cost Data. 114 

C.4 Armament Component Repair Costs. 115 

C.5 Elements of Engine Overhaul Cost Data. 116 

C.6 Elements of Engine Repair Cost Data. 117 

C.7 Elements of Engine Component and Accessory Repair Cost 

Data. 118 








































XI1 

C.8 Elements of Avionics Component Repair Cost Data: Annual 


Ay^erages for 1975-1977. 119 

C. 9 Depot Maintenance Cost Comparison with OSCAR Data: 

Annual Cost Per Aircraft. 120 

D. l Airframe Explanatory Variable Values. 122 

D.2 Engine Explanatory Variable Values. 125 

D.3 Avionics Explanatory Variable Values. 126 

F.l Total Airframe Rework Cost Equations: Full Sample. 182 

F.2 Total Airframe Rework Cost Equations: Most Repre¬ 
sentative Series. 184 

F.3 Total Airframe Rework Cost Equations for Fighter/Attack 

Aircraft. 188 

F.4 Airframe Rework-Cost-Per-Visit Equations: Total Sample. 189 

F.5 Airframe Rework-Cost-Per-Visit Equations: Most Repre¬ 
sentative Series. 190 

F.6 Airframe Rework-Cost-Per-Visit Equations For Fighter/ 

Attack Aircraft.... 193 

F.7 Airframe Rework-Cost-Per-Visit Equations For Bomber/ 

Cargo Aircraft. 194 

F.8 Production Quantity Equations..... 195 
















XllX 


FIGURES 


1. ATBO and Cost to Overhaul as a Function of Engine Age.... 27 

2. Variation of Annual Airframe Rework Cost Per Aircraft 

With Empty Weight. 45 

3. Variation of ATBO With Engine Pressure Term. 54 

4. Variation of Overhaul Cost With Engine Pressure Term. 55 

5. Variation of Annual Repair Cost With Engine Pressure Term 56 

6. Variation of Annual Airframe Component Repair Cost With 

Empty Weight. 64 

7. Variation of Annual Engine Component and Accessory Repair 

Cost With Engine Pressure Term. 72 

8. Variation of Annual Avionics Component Repair Cost With 

Suite Procure...ent Cost. 77 

3. Variation of Total Cost Per Aircraft With Inventory Size. 83 
E.l Variation of Total Airframe Rework Cost V/itfa Inventory 

Size.. 128 

E.2 Variation of Total Airframe Rework Cost With Production 

Quantity. 129 

E.3 Variation of Total Airframe Rework Cost With Fleet 

Flying Hours and PDM Policy. 130 

E.4 Variation of Production Quantity With Age. 131 

£.5 Variation of Production Quantity With Percent Organic 

Maintenance. 132 

E.6 Variation of Annual Airframe Rework Cost Per Aircraft 

With Empty Weight and PDM Policy. 133 

E.7 Variation of Annual Airframe Rework Cost Per Aircraft 

With Airframe Manufacturing Cost. 134 

E.8 Variation of Annual Airframe Rework Cost Per Aircraft 

With Production Quantity. 135 

E.9 Variation of Airframe Rework Cost Per Visit With Age. 136 

E.IO Variation of Airframe Rework Cost Per Visit With 

Airframe Manufacturing Cost. 137 

E.ll Variation of Airframe Rework Cost Per Visit With 

Percent Organic Maintenance. 138 

E.12 Variation of Airframe Rework Cost Per Visit With 

Production Quantity. 139 

E.13 Variation of ATBO With Turbine Inlet Temperature. 140 

E.14 Variation of ATBO With Engine Base-Level Removal Rate.... 141 

E.15 Variation of ATBO With Selling Price. 142 

E.16 Variation of ATBO With Specific Fuel Consumption. 143 

E.17 Variation of ATBO With Engine Weight. 144 

E.18 Variation of ATBO With Model Qualification Date. 145 

E.19 Variation of Overhaul Cost With Turbine Inlet 

Temperature. 146 

E.20 Variation of Overhaul Cost With Specific Fuel Consumption 147 

E.21 Variation of Overhaul Cost With Selling Price. 148 

E.22 Variation of Overhaul Cost With Engine Weight. 149 

E.23 Variation of Overhaul Cost With Military Thrust. 150 

E.24 Variation of Overhaul Cost With Model Qualification Date. 151 






































xiv 

E.^5 Variation of Annual Cost to Repair With Turbine Inlet 

Temperature. ^^2 

E.26 Variation of Annual Cost to Repair With Specific Fuel 

Consumption. 

E.27 Variation of Annual Cost to Repair With Selling Price.... 154 

E.28 Variation of Annual Cost to Repair With Engine Weight.... 155 

E.29 Variation in Annual Cost to Repair With Maximum Thrust... 156 

E.30 Variation in Annual Cost to Repair With Military Thrust.. 157 

E.31 Variation of Annual Cost to Repair With Model Qualifi¬ 
cation Date..... 158 

E.32 Variation of Airframe Component Repair Cost With 

Airframe Manufacturing Cost. 159 

E.33 Variation of Airframe Component Repair Cost With 

Airframe Manufacturing Cost. 160 

E.34 Variation of Airframe Component Repair Cost With 

Empty Weight and PDM Policy. 161 

E.35 Variation of Airframe Component Repair Cost With 

Empty Weight and Afterburner. 162 

E.36 Variation of Airframe Component Repair Cost With 

Sortie Rate. 163 

E.37 Variation of Annual Engine Component and Accessory 

Repair Cost With Turbine Inlet Temperature. 164 

E.38 Variation of Annual Engine Component and Accessory 

Repair Cost With Specific Fuel Consumption. 165 

E.39 Variation in Annual Engine Component and Accessory 

Repair Cost With Selling Price. 166 

E.40 Variation in Annual Engine Component and Accessory 

Repair Cost With Engine Weight. 167 

E.41 Variation in Annual Engine Component and Accessory 

Repair Cost With Maximum Thrust. 168 

E.42 Variation of Annual Engine Component and Accessory 

Repair Cost With Military Thrust. 169 

E.43 Variation in Annual Engine Component and Accessory 

Repair Cost With Model Qualification Date. . 170 

E.44 Variation of Annual Avionics Component Repair Cost 

With Suite Black Box Count. 171 

E.45 Variation of Annual Avionics Component Repair Cost 

With Suite Weight. 172 

E.46 Variation of Annual Avionics Component Repair Cost 

With Suite Functions. 173 

E.47 Variation of Annual Avionics Component Repair Cost 

With MTBD (Mean Time Between Demands). 174 

E.48 Variation of Annual Avionics Component Repair Cost 

With Annual Sortie Rate. 175 

E.49 Variation of Annual Avionics Component Repair Cost 
With Percent of Avionics Suite Item Cost Which 

Is Peculiar to MDS. 176 

E.50 Variation of Annual Avionics Component Repair Cost 

With All-Weather Dummy Variable. 177 

E.51 Variation of Annual Avionics Component Repair Cost 

With Aircraft First Flight Date. 178 



























XV 


F.l Variation of Total Airframe Rework Cost With Fleet 

Flying Hours... 181 

F.2 Variation of Total Airframe Rework Cost With Empty 

Weight. 183 

F.3 Total Airframe Rework Cost for Four Classes of Aircraft.. 186 

F.4 Variation of Airframe Rework Cost Per Aircraft With Empty 

Weight. 191 

F.5 Production Quantity Reflects Inventory Size. 196 











xvii 


GLOSSARY 

The values of all cost elements listed in this glossary are 
expressed in fiscal year 1978 dollars. 

ACFFD Aircraft first flight date 

ACI Analytical condition inspection 

AFCCST Annual airframe component repair cost per '■’ircraft 

AFMFGC Airframe manufacturing cost; cumulative average cost 
of first 100 units, including manufacturing labor and 
materials (millions of FY 1978 dollars) 

AFRWKC Annual airframe rework cost per aircraft 
AGE Aircraft average age, as measured and reported by 

AF/PAXRB (years) 

ALC Air Logistics Center; any one of five Air Force owned 

and operated depot maintenance facilities 
ANNCTR Annual cost to repair per engine 

ATBO Average time between overhaul (hours) 

ATE Automatic test equipment 

AVCST Annual avionics repair cost per aircraft 

AVGCOH Average cost per overhaul 

AWf Avionics suite weight (lbs) 

AWXDV All-weather capability Gi'Vjny variable (no = 1; yes = 2) 

BITE Built-in test equipment 

BLBOX Number of black boxes in suite (#) 

CER Cost estimating relationship 

CIE Controlled interval inspection 

CODMC Contracted-out depot maintenance cost 

DCLC Direct civilian labor cost 

DoD Department of Defense 

D041 The Recoverable Consumption Item Requirements Computation 

Systern 

DMC Depot maintenance cost 





xviii 


DMIF Depot Maintenance Industrial Fund 

DMLC Direct military labor cost 

DSARC Defense Systems Acquisition Review Council 

ENGACC Annual engine component and accessory repair cost per 
engine 

EW Aircraft empty weight (lbs) 

FH Fleet flying hours; the number of flying hours accumulated 

during a year by aircraft of a particular MDS 
FHRATE Same as FH 

FUNC Number of electronics functions performed by aircraft 

avionics suite (#) 

GAC General and administrative cost 

H036B Depot Maintenance Industrial Fund (DMIF) Cost Accounting and 

Production Report (H036B) 

INV Inventory size, the number of possessed aircraft 

LRU Line replaceable unit 

MAINTPCT Organic maintenance percentage; the percentage of cost 
that is associated with organic maintenance rather than 
conttactor maintenance 

MAXLDF Maximum load factor; the aircraft design load factor (g's) 
MAXTH Maximum thrust (lbs) 

MD Aircraft mission/design 

MDS Aircraft mission/design/series 

MILTH Military thrust (lbs) 

MISSDES Mission designator (1 = bomber/cargo; 2 = fighter/attack) 
MISSDV Mission dummy variable (1 = noncombat aircraft; 

2 = combat aircraft) 

MISTR Management of items subject to repair 

MTBD Mean time between OFM demands (hours) 

MTBO Mean time between overhauls (in engine flying hours) 

NENG Number of installed engines per airframe 

NRTS Not reparable this station 

NSN National Stock Number 

O&A Over and above; i.e., work over and above the usual 

requirement 


1 






XIX 


I 


ODC Other direct cost 

ODMC Other direct material cost 

OFM Organizational and field maintenance; the levels of 

maintenance below the depot level 

OIC Other indirect cost 

OSCAR Operating and Support Cost Analysis Report 

PDM Programmed Depot Maintenance (1 = no PDM; 2 = has PDM) 

PQ Production quantity 

PRSTERM Engine pressure term (psf) 

REMRATE Base-level engine removal rate (// per 1000 engine hours) 

RSVPCT Percentage of engine operating hours flown by Guard/ 

Reserve personnel 
SE Support equipment 

SEE Standard estimate of error 

SELLPR Engine selling price (unit 1000 in 1978 dollars) 

SFC Specific fuel consumption (Ibs/hr/lb) 

SINGDES Single engine designator (multiple = 1; single = 2) 
SORTENG Annual engine sortie rate (sorties/year) 

SORTRATE Annual aircraft sortie rate (sorties/year) 

SRU Shop replaceable unit 

SUITE 1 Procurement cost of avionics suite at unit 100 (1978 
dollars) 

SUITE 2 Sum of D041 item (NSN) procurement costs for all items 
in avionics suite 

TCSTPAC Total cost per aircraft; the annual cost per aircraft 
for total depot maintenance cost, including all 
categories of depot-level activity 
TEMP Turbine inlet temperature (degrees Rankine) 

TFDES Turbofan designator (1 = no; 2 = yes) 

TOTCST Total annual cost; the total annual cost per aircraft 
for a single cost category 
Technical repair center 


TRC 



TYPOTC Type maintenance designator (1 organic; 2 
’ United States Air Force 

Work performance category 
Weapon System Cost Retrieval System 
Engine dry weight (lbs) 


= contractor) 


WPC 

WSCRS 

WT 







1 


I. INTRODUCTION AKD OVERVIEW 

The Department of Defease (DoD) uses a basic three-level system of 
equipment maintenance: organizational, intermediate, and depot. In 
the Air Force, organizational and intermediate maintenance units are 
located at aircraft operating bases. Depot-level maintenance is 
typically the most complex work and is performed at a limited number 
of permanent facilities that are operated by, or under contract to, 
the Air Force Logistics Command. The depot facilities support the 
organizational and field maintenance units through overhaul, repair, 
and modification of aircraft, engines, support equipment, and their 
components, 

Depot-level maintenance accounts for a significant part of the 
support cost of military aircraft. Reliable, accurate estimates of 
depot maintenance cost (DMC) are needed during the acquisition of new 
aircraft if life cycle cost is to be a criterion in acquisition 
decisions. Methods of estimation currently in use reflect the amount 
of design data available at different points in time. Early in 
aircraft development, DMC typically must be estimated as a total cost 
based on system-level parameters such as aircraft weight and speed. 

When a new aircraft is almost ready to enter production, detailed 
bottom-up estimates for major elements of DMC can be developed based 
on detailed knowledge of the design. A problem arises during an 
intermediate period in aircraft development. There is a point at 
which some subsystem-level information is known, and this information 
could serve as the basis for an estimate of DMC, but no suitable 
estimating method is available to make use of this information. 

This point often occurs at or near the Defense Systems Acquisition 
Review Council (DSARC) Milestone II. 

An appropriate estimating method would include separate cost 
estimates for each of the major categories of depot maintenance 
activity--airframe rework, engine overhaul, component repair--with 
sensitivity to parameters specifically related to the subsystems 











2 


involved. Airframe rework activity is primarily inspection and repair 
of airframe structural components tc correct the effects of corrosion 
and structural fatigue. An estimating method useful during the middle 
of the aircraft development process would account for the influence of 
both system- and subsystem-level design features that affect the need 
for airframe rework. Similarly, the usefulness of a method for 
estimating engine overhaul costs is related in part to its sensitivity 
to features of the engine that affect the frequency and scope of 
overhauls. Separate estimating techniques for maintenance of various 
categories of aircraft components should likewise be sensitive to the 
specific parameters that affect the individual cost categories. 

An acceptable alternative approach would use parameters that are 
related to the individual categories, but use them in a single equation 
that estimates total DMC. Although this approach .vould not provide 
visibility of the relative contributions of different types of 
activity, it could offer useful sensitivity to subsystem characteristics. 

APPROACH AND PRINCIPAL RESULTS 

We have developed parametric estimating equations that provide 
improved sensitivity at a point in time near DSARC II. Statistic? ly 
derived equations were developed for airframe rework, engine overhaul, 
depot repair of three types of components (airframe, engine, and 
avionics) and for total DMC. (Data were collected for depot-level 
maintenance of support equipment, but the costs of this action are 
relatively small and were not addressed in the development of 
estimating equations.) 

The data used in this study covered fiscal years 1975 through 1977 
and included the major Air Force combat, support, and training aircraft 
systems active during that time span. The study capitalized on the 
special opportunity offered by a data retrieval system developed by the 
Cost Analysis Group in the Office of the DCS/Comptroller at 
Headquarters, Air Force Logistics Command. The Weapon System Cost 
Retrieval System (WSCRS) extracts data from standard Air Force data 
systems and integrates them into a single data file. A major advantage 


i- 













of WSCRS over previous methods of integrating DMC data is the treatment 
given to the cost of maintenance for aircraft components. Earlier 
procedures usually allocated almost all component repair costs. Some 
even allocated costs associated with components used on only one 
mission/design/series (MDS) of aircraft. Whenever a component is 
identified by stock number in the raw data, WSCRS allocates its repair 
costs to the MDSs that use that component. Repair costs for a specific 
component used on only one MDS are thus assigned to that MDS. Repair 
costs not charged to specific components are aggregated by Federal 
Stock Class and allocated to all MDSs. Because of the reduced 
dependence on cost allocation, these data provide a more accurate data 
base than could be obtained earlier—accurate in that the cost 
associated with each weapon system is a more realistic measure of the 
maintenance resources needed to support the system. (The USAF data 
reporting system developed for the Visibility and Management of Support 
Costs program has adopted the essence of the WSCRS data processing 
procedure.) 

A major limitation of the WSCRS data is that they do not include 
maintenance costs for components two or move levels below an end item. 
Costs for depot repair of items removed from an airframe or engine are 
included whether the items are removed at the depot itself or at base 
level. Whenever repair of these items involves removal of failed lower- 
level components (which are usually repaired at the depot), the costs 
associated with these lower-level parts are excluded from the WSCRS 
data. This excludes as much as 95 percent of depot-level component 
repair cost for a weapon system. 

In order to include lower-indenture component costs, the WSCRS 
data were supplemented for this study with data taken directly from 
standard AFLC maintenance and supply data systems (H036B and D041). 

These additional data were used to identify maintenance costs charged 
to specific components and to identify the MDSs that use those 
components. The cost for each lower-indenture component was then 
allocated only to the MDSs that use it, as was done by WSCRS for the 
first-indenture items. 


4 


The explanatory variables used to derive estimating equations 
with the AFLC data were selected to represent the factors that we 
and experienced logisticians believed to have significant influence 
on depot maintenance costs. 

Our analysis produced a number of potentially useful equations for 
each of the major categories of maintenance activity; all of these 
are shown and discussed in Sec. IV. The equations that we believe are 
the most representative and applicable to the widest range of 
estimating situations are displayed in Table 1. The explanatory 
variables in these equations had values in the data basa that span 
these ranges: 


Characteristic 

Airframe empty weight (lbs) 

Engine pressure terra (psf) 

Engine dry weight (lbs) 

Avionics suite procurement cost (?) 


Data Base Range 

4067-320,085 

3400-65,840 

367-7475 

220,000-10,410,000 


Despite their probable range of applicability, the estimating 
equations of Table 1 are not universal; nor are they clearly 
superior--in a statistical sense—to several of the alternatives 
documented in Sec. IV. We believe it best to review the results of 
the study as a whole before selecting the preferred set of equations. 
Moreover, this report has been organized to retain the salient data and 
plots needed to make that selection or to develop cost estimates by 
analogy. For these purposes, the interested reader should refer both 
to Sec. IV and to Apps. C, D, and E. 

In using the equations of Table 1, or any of the other equations 
developed during this study, it is important to keep in mind the 
limitations imposed by the nature of the data. The airframe rework 
equation produces much larger costs for aircraft with a PDM program 
(Programmed Depot Maintenance) than for those without PDMs. It should 
therefore be used V7ith caution, since it does not address offsetting 
cost differences that may occur in other support costs, such as base- 
level maintenance. The accuracy of the equations for predictive 
















5 

Table 1 

REPRESENTATIVE SET OF COST ESTIMATING RELATIONSHIPS 


Category 

Equation 


R^ 

SEE 

F 

N 

Airframe 

0 144 1 22 

AFRWKC = 183 PDM'^ 

.84 

.62 

52 

23 

Rework 

(.018) (.000) 





Engine 

Overhaul/ 

AVGCOH = 0.598 PRSTERM*^^^ 

WT-390 

.82 

.43 

32 

17 

Repair 

(.000) 

(.008) 






ATBO = 957000 PRSTERM~'^®^ 

-1 23 

MISSDES 

.51 

,67 

7 

17 


(.007) 

(.011) 






ANNCTR = 2.72x10"^ PRSTERM 

1.49 ,^1.24 

.61 

1.77 

10 

16 


(.026) 

(.028) 





Airframe 

AFCCST = 0.788 


.78 

.54 

116 

34 

Components 

(.000) 






Engine _ 77g n 677 

Components/ ENGACC= 0,0265 PRSTERM^^ ' WT '^ 

.84 

.52 

34 

16 

Accessories 

(.001) 

(.001) 





Avionics 

AVCST= 0.00455 SUITE2^’®^® 

FHRATE^'^^® 

.86 

.46 

41 

16 

Components 

(.000) 

(.012) 






Notes: All cost variables are in 1978 dollars. Statistics in 
right-hand coliunns are coefficient of determination, standard error of 
estimate, F-statistic, and sample size. Numbers in () are significance 
levels for individual variables. 




..-../..r..- 



6 


purposes will be greatest for systems that have characteristics within 
the ranges of the data base parameters. An examination of the lists of 
equations in Sec. IV shows a pattern of nigh standard errors. That is, 
there is some substantial amount of variance in th-i data that is not 
accounted for by these equations. Nevertheless, based on past 
experience with similar equations derived from similar data bases, we 
believe that these equations are accurate enough to be useful at or 
before the DSARC II milestone in new weapon system development. 

It should be noted that our equations were derived from data for 
aircraft that may not adequately reflect the technical and design 
concepts that will characterize future aircraft. For example, the F-16 
and A-10 were excluded from the analysis (and the F-15 included to only 
a limited extent), because there was little or no depot maintenance 
experience on them in 1975 through 1977. This is important because at 
least some of the new concepts are intended to reduce maintenance 
costs. These concepts include modular engines, increased use of 
built-in test equipment, and airframes designed to be supported without 
a rework program. To the extent that these concepts are successful, 
our equations may overestimate the depot maintenance costs, of future 
aircraft. 


OTHER RESULTS 

Although the primary reason for conducting this study was to 
produce estimating relationships, the results have value in another 
respect as well. The study results as a whole (equations, data plots, 
and tabulated data) provide a new look at the nature of depot 
maintenance for aircraft. All of the equations that meet our screening 
criteria (discussed in Sec. Ill) are included in the later sections of 
Lliib report so that interested readers can study them all. A large 
number of data plots are included to convey further information about 
the nature of the data base. Most of these are in App. E, but some 
are part of the discussion of analytical results presented in Sec. IV. 

Presenting a large number of equations and supporting data is 
worthwhile in two respects. First, the information contained in the 















_.iC'^ 








7 

equations can enhance understanding of the factors that affect depot 
maintenance cost. Thus, the estimator will have a richer context in 
which to judge the applicability of specific estimating equations. 
Secondly, we are offering the user alternatives for each cost 
category that may be better suited in a particular case than any 
single equation that we might have selected if we chose to document 
just one. This is important since, in general, the study did not 
produce one equation for each cost category that is clearly preferred 
over all others. The user should review all of the results before 
selecting the equation or .quations to be used in a particular 
situation. 

REPORT ORGANIZATION 

Section II offers descriptions of the natures of the individual 
categories of depot maintenance work. Section III describes the data 
base, discusses the explanatory variables selected for quantitative 
analysis, and describes the statistical analysis methods. The 
estimating equations that met the selection criteria specified in 
Sec. Ill are presented in Sec. IV. Section V summarizes the main 
findings of the study and suggests some ways to improve upon these 
results in future research. Appendixes are included to provide 
information more detailed than that presented in the body of the 
report. Appendix A gives definitions of various terns and variables. 
Appendix B describes the data processing steps used to produce the 
data base used in the statistical analyses. Cost and explanatory 
variable data are tabulated in ..ops. C and D, respectively. Plots of 
the data are collected in App. E. Appendix F describes some 
alternative ways of addressing airframe rework costs. 









8 


II. CATEGORIES OF DEPOT MAINTENANCE ACTIVITY 


Depot maintenance is performed on four major categories of items 
associated with aircraft: airframes, engines, aircraft components, and 
support equipment. Support equipment maintenance costs were not 
analyzed during this study because they are very small relative to the 
other categories. Component repair may be divided into four 
subcategories on the basis of the types of components repaired. The 
four are airframe components, engine components and accessories, 
avionics components, and armament components. Table 2 shows, for a few 
typical systems, the relative magnitudes of the costs in the various 
categories. 


Table 2 

TYPICAL ANNUAL DEPOT MAINTENANCE COSTS PER AIRCRAFT 
(Averages for 1975-77; FY 1978 dollars) 


Engine Component Repair 

Overhaul 


MDS 

Aircraft 

Rework 

and 

Repair 

Airframe 

Engine 

Avionics 

Armament 

A-7D 

13,090 

81,944 

5,035 

24,783 

19,749 

0 

B-52H 

230,913 

38,415 

73,698 

47,104 

160,808 

4,040 

F-4D 

45,482 

17,958 

16,175 

18,060 

31,755 

0 

F-106A 

55,583 

37,211 

25,119 

38,486 

69,226 

504 

F-lllF 

2,775 

101,830 

29,998 

57,230 

117,030 

0 

T-37B 

1,648 

3,681 

1,547 

1,824 

4,595 

0 


Although Rand has worked with various aspects of depot maintenance 
in the past, the current insights of Air Force personnel actively 
involved in depot maintenance activities were felt to be an important 
source of information. Experts at three Air Logistics Centers were 
consulted about their views of the parameters that affect each category 




‘A>.-it»'.4.<»vwi!f*--‘^v*'-»’; v-^’v '*»»‘'#F^«r“>'j'**>r 


It 


of depot maintenance. Their inputs were combined with expertise 
available within Rand to develop the k'''>wledge that formed the basis 
for selection of the potential explanatory variables that were 
evaluated during the study. Those variables are described in the next 
section. The rest of this section summarizes our general understanding 
of the natures of the four major categories of activity at the time we 
were selecting variables for quantitative analysis. In some cases the 
statistical results are consistent with our expectations; in other 
cases they are not. These expectations are presented here and in Sec. 
Ill in order to describe a comprehensive view of depot maintenance 
activities. The most accurate view of depot maintenance is perhaps 
given by the combination of these expectations and the collection of 
quantitative results presented later in this report. Vflien the 
quantitative results do not agree with the expectations, either the 
expectations may be faulty, because of incomplete knowledge about the 
factors that drive depot costs, or the data base may be unable to 
capture the effects that do exist. 


AIRFRAME REWORK 

When an aircraft needs maintenance that is beyond the capability 
of the organizations located at the Air force's operating bases, the 
needed work is accomplished at a central maintenance depot—either 
an Air Logistics Center or a contractor facility. The term 
"airframe rework" is used to identify depot-level work associated 
with whole aircraft (rather than individual components), but 
excluding the engines. 

Installation of aircraft modification kits is one type of work 
that is included in airframe rework. An aircraft may visit the 
depot for a modification alone, or modification work may be done 
along with maintenance work. A given modification may or may not 
significantly change the performance characteristics or other 
features of the aircraft. That which does is of a different nature 
than recurring maintenance of a fixed system and is not included in 
..his study. 










■m mm i 




10 


The nature of airframe rework changes from time to time. In 
recent years, PDM has been a major element of airframe rework for 
many aircraft. PDM consists of a package of depot-lev^l maintenance 
tasks performed at specified calendar inteirvals. Other elements of 
airframe rework are the Analytical Condition Inspection (ACI) program 
and the Controlled Interval Extension (CIE) program. Some aircraft 
are exempt from force-wide scheduled depot maintenance. For these 
aircraft an ACI program may make up most or all of the airframe 
rework activity. 

A PDM package typically includes a core requirement of 
depot-level tasks plus work that is over and above the core 
requirement (O&A tasks), and work that could be accomplished by 
organizational or intermediate maintenance organizations but which 
can be performed economically by the depot once the aircraft is 
torn down for the PDM (economy tasks). The O&A work is the same 
tiqje of work as the core requirements, but is planned as an 
aggregate man-hour requirement rather than as specific tasks. This 
is a way of providing for an am< unt of work that is required but 
that can be predicted only in the aggregate, and not in detail. 

Economy tasks differ from core and O&A tasks in that they do not 
call for depot-level skills or equipment. The amount of field-level 
work done at depot facilities has changed from time to time, at 
least partly because cf explicit policy changes. 

Despite the uncertainties associated with O&A tasks and past 
and (likely) future changes in field-level work performed by depot 
activities, the bulk of the work in a PDM package is driven by 
defects in the basic airframe structure. These defects are caused 
mainly by corrosion and structural fatigue. Analyzing the sources 
of these conditions gives clues to basic parameters that influence 
the cost of airframe rework. 

Corrosion is related to the age of an aircraft and the environment 
within which it is operated: The more time an aircraft spends in a 
humid environment, the greater the corrosion problem is likely to be. 

















11 





•?,... t .. .J. ^ . -« -. ‘ ' - ' > ' . ^ ^ V 



Structural fatigue is related to the aircraft's mission—to how it 
is used. Thus, different types of aircraft that perform different 
missions might be expected to have different PDM requirements. 


ENGINE OVERHAUL 

Periodically during its life, a jet engine undergoes major depot 
overhaul to restore it to a "zero-time" status. "In this process, 
the engine is completely disassembled and the parts go off in various 
directions to be reworked, modified, or condemned and replaced by new 



parts. Then, as the 'engine nameplate' moves down the depot floor, 
similar parts come back together and are reassembled. By the time 
the 'nameplate' gets to the end of the line, the whole engine is 
reassembled and is considered to be a zero-time engine; that is, one 
capable of achieving the full maximum overhaul time allowed for that 
engine before its next trip to the depot. Most of the parts now 
making up the engine were probably not in the engine when it 
arrived."* 

The maximum number of flying hours which may be consumed before 
an engine must be returned to the depot for overhaul, regardless of 
how well it is working, is termed the maximum time between overhaul 
(MTBO). Few engines actually reach the MTBO, however. Base-level 
inspections often reveal signs of degeneration that are beyond 
base-repair capability because of a lack of either personnel skills 
or appropriate support equipment. Depending on the degree of 
degradation and the time remaining until MTBO, the engine may be 
repaired or may undergo a complete overhaul. The average number of 
flying hours consumed before an engine undergoes overhaul is termed 
the average time between overhaul (ATBO). 

The MTBO is initially determined based on contractor inputs 
and initial testing. As the ATBO experience improves, through 
enhanced base-level repair capability and component improvement 
modifications, the MTBO is usually increased. Increasing MTBO is 

*J. R. Nelson, Life Cycle Analysis of Aircraft Turbine Engines , 
The Rand Corporation, R-2103, November 1977, p. 33. 







12 


an Air Force policy decision based on actual experience with ATBO. 

At some point, however, the MTBO is usually determined to be long 
enough and is not increased further. This upper limit on MTBO 
presumably represents a balance between the perceived risk of a 
higher probability of in-flight failure and corrosive damage to 
parts and the cost of more frequent, but less expensive, depot 
visits. 

The reasons for an engine being returned to a depot facility for 
repair are considerably more diverse than the reasons for overhaul. 

They include such things as premature part failure (misestimation of 
part life), unknown source of performance degradation, lack of proper 
maintenance support equipment, aircraft accident, and foreign objact 
damage. Thus, the causes of engine depot repair are not always 
directly related to either engine or application characteristics. 

In general, however, it is not unreasonable to suggest that the same 
characteristics which influence overhaul cost will also influence 
repair cost. 

The two most direct causes of engine maintenance are thermal 
fatigue and cyclic fatigue. Thermal fatigue (e.g., warping and 
cracking of turbine vanes and blades) is caused by both operation 
at high temperature and changes in temperature. Cyclic fatigue 
(e.g., wearing of discs and bearings) is caused by changes in the 
rotational speed of the engine. Thus, the frequency and amount of 
time at maximum power as well as the total number of throttle 
excursions are felt to have a significant impact on engine part 
life. 

Other factors that may affect engine depot maintenance cost 
include the level of technology embodied in an engine's design, 
the number and size of engine parts and assemblies, and maintenance 
concepts and policies. 

COMPONENT REPAIR 

The depot repair of aircraft components is managed by the MISTR 
(Management of Items Subject To Repair) system. Items are submitted to 





i.3»M »fc;'««c^ -^XtJti^i-i 


13 

MISTR from the operating bases and from the depot. When a component 
fails during operations, the base-level maintenance force removes the 
failed item and substitutes a working item from stock. Certain items 
can be repaired only at the depot and are shipped there directly. 

Other items are coded for base-level repair but because of a lack of 
spare parts, maintenance skills, test equipment or the like, are 
sometimes shipped to the depot (coded Not Repairable This Station— 

NRTS—with an appropriate indicator of the reason why repair cannot 
be accomplished). The depot airframe rework and engine overhaul 
processes also submit components to the MISTR system. 

The total population of components repaired at the depot includes 
airframe, engine, avionics, and armament items and assemblies. Each 
ALC is designated as a Technical Repair Center (TRC) for specific types 
of components. For example, the majority of avionics components are 
sent to Warner Robins ALC while landing gears are repaired at Ogden 
ALC. Therefore, like components will normally be funneled to the same 
depot. 

Other things being equal, the depot maintenance cost of individual 
components of all types should increase with item demand rate. The 
maintenance cost for a collection of components should therefore be 
related to a total demand rate. In addition, the types of materials 
used ana the complexity of the manufacturing tasks involved in 
producing components, as reflected in component procurement cost, may 
also be related to the amount and cost of material and labor needed to 
perform depot maintenance. 

Since most items are processed through a component repair line 
in batches, the cost of repair should also be affected by considerations 
that determine whether or not the most economical lot size is used. 
Shortages, for example, may lead to repair in lot sizes smaller than 
the most economical. 

Airframe and Engine 

Many of the factors that influence the repair cost of airframe 
and engine components and engine accessories are the same as those that 
influence the cost of airframe rework and engine overhaul. 






14 

Avionics 

The avionics subsystem is defined to include those components 
providing aircraft display, communication, navigation, fire control, 
countermeasure, and reconnaissance functions. The depot-level repair 
cost of these components depends on the frequency with which they are 
returned to the depot for repair and the extent of the required 
repairs. Factors that are believed to have a strong influence on the 
frequency and cost of avionics repair include the complexity and 
performance of the components, the environment in which they must 
operate, and the diagnosis and repair concept. 

Armament 

Aircraft armament consists of guns, bomb racks, missile 
launchers, and other components related to weapon delivery. The 
total repair cost of armament components for a weapon system is 
expected to increase with the system's number of guns, number of 
munitions stores hard points, total munitions load, and number of 
types of munitions carried. Each of these parameters reflects a 
different aspect of the amount of armament hardware on the aircraft. 
Some combination of them should be related to the overall scope of 
the maintenance effort needed for these components. The amount of 
work done at the depot level is extremely small—small enough that 
it is insignificant compared with other cost categories. Appendix C 
shows the data for the few aircraft that had armament costs at the 
depot during FY 75-77. Because those costs were so small, we did not 
analyze armament or develop estimating methods for it. 

SUPPORT EQUIPMENT MAINTENANCE 

As with armament, we did not prepare estimating relationships 
for support equipment (SE) costs; we collected some SE infoimation 
during the early research stages of the study, however, and summarize 
it here for completeness. 










15 

Direct depot-level maintenance costs for SE are associated only 
with SE used at aircraft operating bases. SE used in depot-level 
maintenance of aircraft is maintained by the shops that use the 
equipment or by a Precision Measuring Equipment Laboratory supporting 
these shops. The associated cost is an indirect cost of the 
operation of aircraft maintenance shops. Base-level SE is similarly 
maintained by the using base maintenance organization to the exlent 
possible, but SE that requires repair work beyond the capability of 
the base is either sent to an Air Logistics Center or to a 
contractor. This results in a depot maintenance cost within the scope 
of this study. 

SE includes training aids and devices and maintenance equipment. 
Some maintenance equipment can be further identified as automatic 
test equipment (ATE). ATE is more complex (and likely to be more 
expensive to repair) than other maintenance equipment. The SE repair 
cost per aircraft can be considered to be the sum of three terms; 

(1) The annual cost for repair of training aids and devices, 

(2) The annual cost for repair of ATE, and 

(3) The annual cost for repair of maintenance equipment other 

than ATE. 



li 



I 


I 

a 


Each term includes the cost of overhaul of SE end items and repairs of 
SE components. SE costs were not analyzed in cais study; they remain 
an appropriate area of investigation for future research. 

SE depot I Intenance costs probably vary by mission. In 
particular, combat aircraft are likely to have a greater cost than 
noncombat aircraft, because they are likely to have more 
sophisticated equipment on board and to be supported by more 
sophisticated ground equipment. 

SE depot maintenance cost may increase with increases in aircraft 
fleet size and flying activity, as measured in flying hours or sorties. 
The number of SE maintenance tasks that are performed is likely to be 
driven by the usage of SE, which is influenced by both number of 
aircraft and the level of flying activity. 



















16 




Maintenance equipment maintenance costs should be greater for new 
aircraft than for old aircraft, because electronics and automation are 
used more extensively with newer aircraft. 

Maintenance equipment depot maintenance costs are expected to 
increase with increases in the per aircraft procurement cost, weight, 
and power consumption of an aircraft's avionics. Procurement cost, 
weight, and power are indirect measures of the amount of avionics on 
the aircraft. 

Maintenance equipment depot maintenance cost should decrease with 
the use of built-in test equipment (BITE) in onboard avionics systems. 
The extent of the use of BITE can be measured by the fraction of 
avionics systems for which BITE is used. 

The depot maintenance cost of maintenance equipment other than ATE 
should increase with the size of the aircraft supported, with size 
measured by aircraft empty weight or basic operating weight. Aircraft 
size drives the size and procurement cost of various types of work 
stands and ground handling equipment; and larger, more expensive 
equipment should be more expensive to maintain. 


COMMON CONSIDERATIONS AFFECTING DEPOT MAINTENANCE 


One important issue affects all categories of depot maintenance: 
The costs charged for a given depot-level task may depend upon where 
the task is accomplished. This effect can be felt in one of three 
ways. 

First, the direct cost to perform a stated task may differ between 
ALCs, because their direct labor rates differ. An ALC charges for 
direct labor at a rate derived from the average pay of the direct labor 
personnel in the production division perfor ing the work. Each ALC 
will therefore have its own direct labor rare, reflecting the skill 
levels and experience of its workers and the general level of wages in 
its geographic area. 

Second, hourly charges for indirect and overhead costs can 
also vary between ALCs, which may result in different total costs 
even when direct costs are equal. These differences would be due 




to differences in the staffing of indirect and overhead functions 
and to differences in the allocation of these costs between the 
ALCs and other organizations on the same bases. 

Third, total costs for similar work will differ between an ALC 
and a contractor, and between contractors. Contractors can change 
the sizes of their work forces and the mixes of .skills within them 
more quickly than can the ALCs. This allows contractors to more 
readily match their personnel to changes in the types or amounts 
of work that come to them. PDM costs, for example, may vary 
between locations because of differences in aircraft condition. 

F-4s operating in the Far East (and reworked there) are likely to 
have a greater corrosion problem than F-4s operating in the 
southwestern United States (reworked in this country). This could 
drive the man-hours needed to perform a PDM. It could also affect 
the types of workers that contractors would hire to perform the 
PDM, resulting in differences in labor co£cs par man-hour. 

Contractor charges should therefore be more closely matched to the 
nature and scope of the work. As a result, two contractors with 
different total workloads are likely to have different costs for 
parts of their work that are similar. A contractor and an ALC are 
likely to have different costs for similar work because their 
total workloads are dissimilar and because the labor forces 
available for these similar tasks will not be alike. 

These common considerations may have important influence on 
the magnitude of depot maintenance costs, but data limitations prevented 
their analysis during this study. These considerations should be kept 
in mind during any application of the study results. 








III. DATA BASE AND ANALYTICAL APPROACH 


This section describes general aspects of the quantitative 
analysis that produce'’ estimating relationships presented in Sec. 

IV: the cost data base, the candidate explanatory variables, and 
elements of the analytic approach that are common to all maintenance 
categories. The scope of the cost data base is described, along with 
brief descriptions of the cost data sources. Appendixes A, B, and C 
present the cost element definitions, data processing steps, and tables 
of the cost data. The discussion in Sec. II of the nature of depot 
maintenance led to consideration of specific potential explanatory 
variables. These variables are discussed here, aud sources of data for 
them are identified. These variables are defined in App. A; 
tabulated data are included in App. D. 

COST DATA 

Data for three fiscal years (1975, 1976, and 1977) were 
collected and analyzed for most of the aircraft and engines currently 
in the Air Force inventory. These were the only years for which 
WSCRS data were available. Data were organized in the working data 
base by category of depot activity: 




I 


I 


Airframe Rework 

Engine Overhaul 

Component Repair 

Airframe Component Repair 

Engine Accessory and Component Repair 

Avionics Component Repair 

Armament component repair costs exist in the raw data for only a few 
weapon systems. Where they do appear, they are very small. 
Consequently, they are not included in the working data base or the 
analytical work. 






19 


The total maintenance cost of interest for each category includes 
the costs for maintenance proper and for installation of Class IV 
modifications, where these costs are identified in Air Force data by 
Work Performance Category (WPG). Relevant WPG definitions are given in 
App. A. Class IV modifications are changes to the physical makeup 
of an aircraft that do not alter the mission, performance, or 
capability of the aircraft. Such modifi''ations can be expected as a 
routine part of the support of new weapon systems, so their cost is 
included. The cost of modifications that do change the mission, 
performance, or capability of an aircraft are specifically excluded 
because they are outside the scope of normal system acquisition 
decisions. 

The data collected by the Air Force for engine maintenance show 
no costs for Glass IV modifications, so the data base for this study 
necessarily includes only costs labeled as being for maintenance work 
per se. 

It should be noted that some raw records for ''977 do not contain 
a WPG code. This meant that there was no way to determine whether or 
not the costs in these records were associated with maintenance 
activities relevant to this study. With no better information than 
this, it was decided not to include these costs in this analysis. If 
it were known that all of the costs in such records for airframe work 
were relevant, the airframe rework cost.s of, for example, the A-7D, 
B-52G, G-5A, and G-130E, would be between one and six percent higher 
than the values used in this study. 

Total cost is composed of seven individual cost elements: 


o 

o 

o 

o 

o 

o 

o 




Direct Givilian Labor Gost (DGLG) 
Direct Military Labor Gost (DMLG) 
Other Direct Material Gost (ODMG) 
Other Direct Gost (ODG) 

General and Administrative Gost (GAG) 
Other Indirect Gost (OIG) 
Contracted-Out Depot Maintenance Cost 




(CODMC) 






These are defined in App. A. 

Excluded are the following costs that, for other purposes, might 
be considered elements of depot maintenance cost: 

o Cost of components and assemblies submitted to the MISTR 

line (Management of Items Subject to Repair) durin'^ overhaul 
or repair-* This cost is sometimes referred to as Direct 
Replacement cost of condemned reparables (which is considered 
a supply function). 

o Depreciation of capital equipment, 
o Material Cost at Standard Cost to Repair, 
o Other Work Performance Categories such as conversion, 
activation, inactivation, reclamation, and storage, 
o Transportation to and from the depot, 
o Pipeline components. 

Three sources of information were used in the development of the 
cost data included in the working data base. The primary source of 
cost data was WSCRS. All of the cost information for airframe rework 
and engine overhaui/repair was taken from WSCRS. WSCRS also provided 
some component repair costs, specifically, costs of repairing line 
replaceable units (LRUs) and costs reported against a class of 
components rather than a specific component. The term LRU denotes a 
component that is removed from an aircraft or engine as a single 
unit. An LRU may contain removable elements t^hat are termed shop 
replaceable units (SRUs), SRU costs were obtained from the Depot 
Maintenance Industrial Fund (DMIF) Cost Accounting and Production 
Report (H036B). In order to link SRUs with the appropriate aircraft, 
application data were obtained from the Recoverable Consumption Item 
Requirements Computation System (DOAl). 

*These MISTR-related test and repair costs are considered in the 
component rework section of this depot cost model. 







I 



I 








21 


All costs were converted to fiscal year 1978 dollars, using the 
indices given below, and averaged over the three-year period: 


Cost Element 

1975 

1976 

19'’7 

DCLC 

1,265 

1.174 

1.076 

DMLC 

1.187 

1.128 

1.067 

ODMC 

1.220 

1.135 

1.068 

ODC 

1.246 

1.159 

1.071 

GAC 

1.246 

1.159 

1.071 

QIC 

1,265 

1.174 

1.076 

CODMC 

1.246 

1.159 

1.071 


Average costs were computed for the three-year period to minimize the 
problems associated with random year-to-year fluctuations in the 
magnitude of the maintenance work for any given system or category of 
activity. 


EXPLANATORY VARIABLE DATA 

The material presented in Sec. II was the basis for development 
of sets of explanatory variables for the various categories of 
maintenance activity. The variables and the sources of relevant data 
are shown in Tables 3 through 5. Appendix A contains definitions of 
all variables. The data are tabulated in App. D. Before a variable 
was accepted for use in this study, it had to satisfy three criteria: 

o Be logically related to cost (i.e., the variable must be felt 
to have a logical impact on the frequency or magnitude of cost) 
o Be readily available at DSARC II 
o Possess historical record 


The first point was satisfied through the development of the background 
material presented in Sec. II. The information conta-ned therein 






22 


about factors related to cost points to potentially useful variables. 
Our goal was to develop at least one quantitative variable for each 
factor—one variable which meets the other two criteria. Data 
availability at DSARC II is required because that is the point at 
which the equations are expected to receive the most use. A historical 
record was obviously a necessity if data were to be collected to 
support a quantitative analysis. 

Airframe Rework 

The main approach to airframe rework estimates the annual cost per 
aircraft. If a cost analyst can estimate a cost per aircraft, then he 
needs to know only the inventory size to get the total cost for a 
weapon system. Alternative approaches, considered in App. F, are to 
estimate (1) the average annual total cost for a fleet of aircraft, 
and (2) the product of average cost per rework and average number of 
reworks per year. 

Because the aircraft in the data base vary greatly in age, we 
considered the possibility of basing the prediction of airframe rework 
costs on a model that would capture the various effects on cost that 
change over the life of a weapon system. This proved not to be 
feasible, because the time-histories needed to understand and quantify 
these effects are not available in any readily accessible form. 
Corrosion, for example, is thought by some experts to cause significant 
costs at periodic intervals. A fleet of aircraft that receives 
extensive corrosion repair will not need such work again for some time, 
until the effects that cause corrosion to occur have had some time to 
work. Th^n, when repair is necessary, it will probably be needed for 
all aircraft in the fleet at roughly the same time; and the cycle 
repeats. Quantifying such effects would require consistent data over 
several years. Such data are not readily available for any sizable 
number of MDSs. 

Although no sophisticated representation of age is possible, age 
is included in the list of explanatory variables dealt with in the 










Variable 


Source 


SIZE 

Empty weight 
Maximum takeoff weight 

TECHNICAL/PERFORMANCE 
Maximum speed 
Typical speed 
Typical altitude 

Dynamic pressure at maximum speed 
Dynamic pressure at typical speed 
and altitude 
Maximum load factor 
Airframe manufacturing cost 
Afterburner designator 
Fighter/attack designator 
Boraber/cargo designator 
Trainer designator 

UTILIZATION 
Fleet flying hours 
Inventory 
Age 

Sorties 

Percent of fleet operated by reserves 
Percent of fleet operating in 
humid climate 


SAC Charts#! 
SAC Charts 


SAC Charts 
SAC Charts 
SAC Charts 
Computed#2 

Computed 
SAC Charts 
Rand Data#3 
SAC Charts 
Assigned 
Assigned 
Assigned 


WSCRS 

WSCRS 

Hq USAF/PAXRB 
Hq USAF/PAXRB 
Air Force Planning Data 


See Note#4 


POLICY 

Organic maintenance percent 
PDM policy 
Production quantity 


Computed from Cost Data 

T.O. 00-25-4#5 

WSCRS 


Notes: 

#1 USAF Standard Aircraft/Missile Characteristics , Air Force 
Guide Number Two, various dates. 

#2 Computed from appropriate speed and atmospheric density. 

#3 Rand data collected for previous research on airframe 
development and production costs. 

#4 Derived frcm aircraft operating locations specified in Air 
Force planning documents and standard climate categories. 

#5 Depot Maintenance of Aerospace Vehicles and Training 
Equipment , Air Force Technical Order TO 00-25-4, various dates. 







24 







analysis. This allowed for the possibility of long-term effects that 
might be significant at a gross level even though they could not be 
modeled as detailed processes. 

Corrosion is related to the age of an aircraft and the environment 
within which it is operated: The more time an aircraft spends in a 
hujiiid environment, the greater the corrosion problem is likely to be. 

An older aircraft is therefore likely to incur more cost associated 
with corrosion treatment than a newer aircraft. 

Structural fatigue is related to the aircraft's mission—to how it 
is used. Thus, different types of aircraft that perform different 
missions might be expected to have different PDM requirements. At a 
gross level one can distinguish three major mission categories: bomber 
and cargo, fighter, and trainer aircraft. Bombers and cargo aircraft 
tend to carry heavy loads while flying straight and level for long 
periods of time. Fighters carry relatively light loads for shorter 
periods of time, but must endure the stresses of combat maneuvering. 
Trainers fly short sorties with many landings and are flown by 
inexperienced pilots. (Similarly, some logisticians believe that 
pilots in the Air Force Reserve and the Air National Guard may impose 
different stresses on an aircraft than active pilots who may fly that 
specific aircraft type more often.) 

Within a given type, different usage may be associated with 
differences in size, flight conditions, and levels of activity. 

Airframe weight and aircraft empty weight are measures of the size 
of the aircraft. Airframe weight is the more direct measure of the 
amount of structural material in the aircraft; but data on empty weight 
are more easily obtained, and empty weight is highly correlated with 
airframe weight. Maximum takeoff weight is a measure of the total mass 
of the vehicle, including fuel and payload. 

The altitude and speed that a specific aircraft uses on a typical 
mission may result in stresses of a different magnitude from those 
encountered by a similar aircraft under different flight conditions. 
Maximum altitude and maximum speed relate to the greatest magnitude of 
stress to be expected. Important features of tighter design are the 











25 

maximum load factor for which tho vehicle is designed and whether or 
not an afterburner is used. 

Also, for any aircraft, the number of landings per unit time is 
likely to be related to rework requirements for landing gear and 
related structural elements. Similarly, the numbers of flying hours 
and sorties per unit time are measures of the amount of use an aircraft 
receives. 

For any type of aircraft, two aspects of the airframe design are 
relevant. The type of material used should affect the cost of material 
used during rework and the number of man-hours needed to perform the 
work. ilso, it is possible that different design practices result in 
structures with different degrees of resistance to corrosion or 
fatigue. It is probably not possible to specifically identify these 
practices; but it contractors are consistent in their choice of design 
approaches, it may be that all aircraft designed by any one company 
have somewhat similar rework requirements. 

In addition to aircraft characteristics, maintenance policies 
significantly affect costs incurred for airframe rework. A major 
policy is whether or not to have PDMs. A number of USAF aircraft, 
including the newest (the F-15 and F-16) do not have PDMs. They visit 
a depot only for modification, for an ACI, or because of unusual 
damage beyond the capability of field maintenance units. 

The interval between PDMs on a specific airframe is, along with 
the scope of the PDM package, a major determinant of weapon system 
airframe rework cost. The maximum interval for a new aircraft is 
decided upon on the basis of the best available engineering 
information. The recommendations of the contractor building the 
aircraft receive considerable weight. The value of this initial 
interval is likely to be related to the same things that influence the 
scope of the PDM package, as described above. Typically, as experience 
with a weapon system increases, the maximum interval is extended. The 
maximum value permitted at any point in the aircraft's operating life 
is therefore a function of the initial value and the system's age. 

An aircraft that undergoes a PDM can be reworked by a crew of 
workers dedicated to a particular airframe in a given PDM dock or by 





26 


workers dispatched from pools of specialists. Warner Robins ALC uses 
dock crews; San Antonio ALC uses specialist pools. F-4 aircraft are 
reworked at five facilities—Ogden ALC and four contractor facilities. 
Depending upon which site it visits, a particular F-4 may be reworked 
either by a dock crew or by specialists. This distinction could affect 
both the man-hours needed for a PDM and the average cost of a man-hour. 


Engine Overhaul and Repair 

The two primary components in determining engine lifetime overhaul 
cost are the average time between overhaul (ATBO), which reflects 
frequency, and the cost per overhaul, which reflects the scope of the 
overhaul work. Discussions with ALC personnel suggest that these 
factors vary with engine age in the manner illustrated in Fig. 1. 

Based on this view of engine maintenance, a parametric model for 
an engine lifetime overhaul cost might then take the following form: 


(i) ATBO(i) = f(TECH, APPLIC, AGE(i)) 

(ii) OHAGE(j) = fCFLYPRG, ATBOPRG) 

(iii) NLOH = f(FLYPRG, ATBOPRG) 


(iv) COH(j) = f(TECH, APPLIC, OHAGE(j)) 
NLOH 

(v) LIFOHC = COK(j) 

j=l 


where 


AGE(i) = engine age in year i 
APPLIC = engine application characteristics 
(aircraft characteristics) 

ATBO(i) = ATBO in year i 

ATBOPRG = ATBO program (projected ATBO, by year, 
over engine life) 

COH(j) = cost of jth overhaul 

FLYPRG = engine flying program (projected flying hours, 
by year, over engine life) 

LIFOHC = engine lifetime overhaul cost 
NLOH = number of lifetime overhauls 
OHAGE(j) = engine age at time of jth overhaul 


TECH = engine technical characteristics 






















Engine depot repair cost presents a slightly different problem 
from overhaul cost. Whereas overhauls tend to be somewhat standard 
for a given engine model, repairs can be quite diverse in both type 
and frequency. Thus, engine depot repair cost would appear to be most 
logically estimated on the basis of an average annual cost per 
installed engine. Additionally, the average cost to repair is 
believed to vary in a manner similar to engine overhaul cost (see 
Fig. 1(b)). Based on these observations, a parametric model for an 
engine's lifetime depot repair cost might take the following form: 


(vi) REPFRC(i) = f(TECH, APPLIC, FLYPRG, AGE(i)) 

(vii) AVGCTR(i) = f(TECH, APPLIC, AGE(i)) 

(viii) ENGDRC(i) = REPFRC x AVGCTR(i) 

n 

(ix) LIFDRC = Z ENGDRC(i) 
i=l 

where AGE(i) = engine age in year i 

APPLIC = engine application characteristics (average 
characteristics) 

AVGCTR(i) = average cost per repair in year i 
ENGDRC(i) = annual depot repair cost per installed engine 
FLYPRG = engine flying program (projected flying hours, 
by year, over engine life) 

LIFDRC = engine lifetime depot repair cost 

n = number of years in engine life cycle 
REP. RC(i) = fraction of installed engines returned to 
depot for repair in year i 
TECH = engine technical characteristics 

While the preceding formulation is conceptually valid, it has two 
difficulties which preclude its testing at this time. First, cost data 
are available for only three years (1975, 1976, and 1977). Given an 
engine life of 15 years or more, such limited longitudinal data cannot 
be viewed with any degree of confidence. Second, the shape of the 
overhaul cost and repair cost curves (see Fig. 1(b)) represents a 
degree of sophistication considerably beyond the norm that now exists 







29 

at DSARC II. Consequently, the following simplified model will be 
tested instead. It assumes a "mature" engine;* that is, one which is 
past all the problems associated with the introduction of a new engine 
into the fleet. 


Overhaul Cost 


(x) 

ATBO = f(TECH, APPLIC) 

(xi) 

NLOH = (n X ANNFHR/ATBO) - 1 

(xii) 

AVGCOH = f(TECH, APPLIC) 

(xiii) 

LIFOHC = NLOH x AVGCOH 


Depot Repair Cost 

(xiv) 

ANNCTR = f(TECH, APPLIC) 

(xv) 

LIFDRC = n X ANNCTR 


where ANNCTR = annual cost to repair per installed engine 

ANNFHR = annual flying hours 
APPLIC = engine application characteristics 
ATBO = average time between overhaul 
AVGCOH = average cost to overhaul 
LIFDRC = engine lifetime depot repair cost 
LIFOHC = engine lifetime overhaul cost 

n = number of years in engine life cycle 
NLOH = number of lifetime overhauls 
TECH = engine technical characteristics 

Table 4 shows specific explanatory variables used in our 
quantitative analysis to relate ATBO, AVGCOH, and ANNCTR to technical, 
size, application, and other explanatory variables. 

The ATBO, the average cost to overhaul, and the annual cost to 
repair should be related to engine technical characteristics such as 
turbine inlet temperature, the thrust-to-weight ratio, the total 

*A mature engine is defined as an engine which has been in the 
fleet at least 5 years. 









30 





Table 4 


POTENTIAL EXPLANATORY VARIABLES FOR ENGINE 
DEPOT OVERHAUL AND REPAIR COST ELEMENTS 


_ Explanatory Variables 

TECHNICAL/PERFORMANCE 
Turbine inlet temperature 
(degrees Rankine) 
Thrust-to-weight ratio 
Pressure term (psf) 

Specific fuel consumption 
(psf) 

Maximum Mach number 
Removal rate (usage 
removals per 1000 hours) 
Selling price at 1000th 
unit ($ 1978) 

SIZE 

Weight (lbs) 

Maximum thrust (lbs) 
Military thrust (lbs) 

APPLICATION 
Annual engine sorties 
Mission designator (bomber- 
ca rgo/fighter-attack) 
Fighter/attack designator 
(air-to-air/air-to-ground) 
Single engine designator 
(multi/single) 
Reserve/Guard fraction 


Source 


_ Cost Element _ 

Average Cost Annual 
Sam- Time to Cost 
pies Between Over- to 
#1 Overhaul haul Repair 


Gray Book#2 1 

Table entries 1 

N-1242,Tbl 1W3 1 

Gray Book 1 

Gray Book 1 

AFLC Form 992 1 

N-1242,Tbl 49 1 

Gray Book 1 

Gray Book 1 

Gray Book 1 

HQ USAF/PAXRB 1 

Assigned 1 

R-2249,Tbl Al#4 3 

WSCRS 1 

AF Ping Data 1 


MISCELLANEOUS 

Turbofan designator (yes/no) Nomenclature 1 XX 

Manufacturer designator 

(GE/P&W) Nomenclature 2 XXX 

Type maintenance indicator 

(organic/contract) _ WSCRS _1_X_ X 

#1 Indicates extent of variable applicability in terms of sample: 

(1) Basic sample (all turbojet and turbofan engines in data base; turboprop 
and reciprocating engines are excluded). (2) Pratt & Whitney and General 

Electric engines only. (3) Engines on fighter/attack aircraft only. 

//2 Gray Book is USAF Propulsion Characteristics Summary , Air Force 
Guidebook Number Three. 

#3 Future V/STOL Airplanes: Guidelines and Techniques for 
Acquisition Program Analysis and Evaluation , J. R. Nelson, J. R. Gebman, 

J. L. Birkler, R. W. Hess, P. Konoske-Day, W. H. Erase, The Rand Cor¬ 
poration, N-1242-PA&E, October 1979. 

#4 Measuring Technological Change in Jet Fighter Aircraft , W. L. 
Stanley, M. D. Miller, The Rand Corporation, R-2249-AF, September 1979. 


f fart 











31 


pressure acting on critical engine components, the engine's specific 
fuel consumption, and the maximum Mach number. Generally speaking, 
the higher the values associated with these variables, the higher the 
level of technology which is embodied in the engine and the greater 
the degree of part complexity (in terms of configuration and material 
composition). In turn, this increased part complexity usually leads 
to a greater incidence of part failure as well as an increased cost 
to overhaul/repair. 

Other technical characteristics which may affect engine depot 
overhaul and repair costs are the removal rate and the selling price. 
Intuitively, higher removal rates should be associated with shorter 
ATBOs and higher annual repair costs. Similarly, more expensive engines 
tend to be more technologically advanced than less expensive engines, 
and therefore less reliable and more costly to overhaul/repair.* 

The number and size of engine parts can be expected to influence 
maintenance costs. Depot maintenance costs for turbofan engines should 
be higher than those for turbojet engines because of the additional 
number of parts associated with the fan section. Similarly, larger 
engines have larger parts and subassemblies which may cause greater 
handling difficulties and a more extensive inspection effort. Engine 
weight and thrust are assumed to be indicators of size. 

The application variables—sortie rate, mission designator, 
fighter/attack designator, single-engine designator, and the fraction 
of engines operated by Guard and Reserve units—should affect the 
ATBO and the annual cost to repair. Takeoff and landing cause full- 
tb-ottle excursions which, as stated earlier, contribute to cyclic 
failure. Thus, higher sortie rates should be associated with 
shorter overhaul intervals and higher repair costs. 

*It is of course possible that, other things being equal, higher 
selling prices reflect measures undertaken to improve reliability and 
maintainability. Generally speaking, however, we do not feel this to 
be a significant factor, particularly with respect to our data base, 
which consists largely of engines developed prior to the current 
emphasis on reliability and maintainability issues. 








32 


Excluding takeoff and landing, the engine power level profile 
(engine power level versus misrion time) for a bomber/cargo aircraft 
v.'ill be much more constant than for a fighter/attack aircraft. Thus, 
the fighter/attack aircraft are going through many more throttle 
excursions than bomber/cargo aircraft, thereby resulting in higher 
levels of thermal and cyclic fatigue. A further refinement of mission 
effects, applicable only to fighter and attack aircraft, suggests that 
engines on aircraft with an air-to-ground mission will have shorter 
overhaul intervals and hi^^her overhaul and repair costs than engines on 
aircraft with an air-to-air mission because of the higher stresses 
placed on engines operating at low altitude. 

Because an engine failure can be catastrophic on a single-engine 
aircraft, the engine of such an aircraft may be subjected to more 
frequent and thorough inspections and to more conservative mairter'.nce 
policies, and cost more to overhaul, than a similar engine 0.-3 a 
raultiengine aircraft. 

Engines on aircraft operated by Guard and Reserve units may have 
higher depot maintenance costs than engines on aircraft operated by 
active units. Factors that could cause this include the typically 
greater age of aircraft operated by the reserves. Another possible 
explanation is that some Guard and Reserve pilots fly a specific 
aircraft type less frequently than active duty pilots and therefore may 
make more throttle adjustments. 

Depot maintenance costs may also vary with the manufacturer of an 
engine. Manufacturers may incorporate unique and consistent design and 
manufacturing techniq.es and procedures in their jet engines that result 
in consistent depot maintenance cost differences. 

Finally, the cost to perforin a given overhaul/repair action may 
vary with the organization (depot or contractor) performing the work. 
Depots are bound by federal government regulations and policies and 
this may affect maintenance costs. 

One general area which, with the exception of the performing 
organiz.ition designator, is prominent by its absence from our analysis 
is maintenance policy, which includes such things as; 







33 


o Inspection interval or technique 
o Health monitoring program 

o Quantity and sophistication of base and depot 
support equipment 
o Engine modularity 

Such factors were omitted from the analysis for two reasons. First, 
consistent and sound performance measures could not be developed. 
Second, even if consistent measures could have been developed, many 
of the variables would lack sufficient data for a parametric analysis 
because of their relative newness (e.g., engine modularity and health 
monitoring). 

Component Repair 

Component repair costs will be estimated as annual costs per 
possessed aircraft. The repair costs for airframe components and 
for engine components and accessories are expected to be driven 
by the same set of factors that influence airframe rework and 
engine overhaul and repair costs. Thus, the variables listed in 
Tables 3 and 4 were used in the analysis of these categories of 
component repair as well <is in the analysis of costs for whole 
airframes and engines. 

Avionics depot repair cost will be estimated as an annual cost per 
possessed aircraft utilizing technical and application characteristics 
associated with the aircraft’s avionics suite. It should be noted that 
identifying "the" avionics suite for a mature MDS is a formidable task. 
A suite changes continuously over time, but not uniformly for all 
aircraft in the series. Thus, the determinati'n of values for avionics 
suite characteristics is subject to some uncertainty. 

Table 5 group*; specific explanatory variables investigated in uur 
analysis according to the aspect described: size, complexity, and 
application. 

Weight is a measure of size, and other things being equal, the 
greater the size, the greater the repair cost. Given the 


. 4 




















34 


Table 5 

POTENTIAL EXPLANATORY VARIABLES FOR AVIONICS COMPONENT REPAIR 


Variable 

Source 

SIZE (weight) 

Unpublished Rand data 

PERFORMANCE/COMPLEXITY 

Capability (aircraft first flight date) 

SAC Charts#! 

Number of "black boxes"#2 

SAC Charts 

Number of functions 

SAC Charts 

Suite procurement cost ($) 

Published#3,#4 and 

Mean time between OFM demands 
(flying hours) 

unpublished Rand data 

R-2552-PA&E#4 

Combat designator (combat/noncombat) 

Assigned 

All-weather capability (yes/no) 

Unpublished Rand data 

Mission group designator (bomber, cargo, 
fighter/attack, reconnaissance, trainer) 

Assigned 

APPLICATION 

Annual flying hours per aircraft 

WSCRS 

Annual sorties per aircraft 

HQ USAF/PAXRB 

Percentage of unique items (%) 

R-2S52-PA&E 


#1 USAF Standard Aircraft/Missile Characteristics , Air Force 
Guide Number Two, 

in "Black boxes" refers to individual pieces of avionics 
equipment, which are generally designated by AN (Army-Navy designation) 
number. 

#3 An Estimating Relationship for Fighter/Interceptor Avionic 


System Procurement Cost , C. Teng, The Rand Corporation, RM-4851-PR, 
February 1966. 

#4 Estimating USAF Aircraft Recoverable Spares Investment , 

K. J. Hoffmayer, F. W. Finnegan, Jr., and W. H. Rogers, The Rand 
Corporation, R-2552-PA&E, August 1980. 


conglomeration of integrated circuits, array antennae, discrete 
devices, magnetic amplifiers, etc., which exist for current inventory 
aircraft, the credibility of weight as a measure of avionics repair 
cost is clearly questionable. However, it is doubtful that a size 
measure exists that does not have this or a similar problem. 

We were not able to determine a fully satisfactory capability 
measure which applies to the suite as a whole, so aircraft first flight 
















35 



date is taken as a proxy. This assumes that capability i.s increasing 
uniformly over time. Another indicator of capability may be the number 
of individual black boxes in the suite. A higher number of black boxes 
is also associated with a higher part count and a greater degree of 
system integration than a lower number of black boxes. In turn, part 
count and the degree of system integration are felt to be significant 
influences on repair cost. The number of functions a suite performs is 
a measure of capability which differs from the number of black boxes in 
that the number of functions reduces the impact of redundant black 
boxes. Functions which will be counted are as follows: 

Communication/Identification 
Navigation 

Bomb Navigation/Fire Control 
Penetration Aids/ECM 
deconnai.ssance 

Cohtrols/Displays/Instrumentation 

Suite procurement cost reflects the types of materials used and 
the complexity of manufacturing tasks involved in producing the suite 
components. 

Avionics depot repair workload is influenced by the suite 
Organization and Field Maintenance (OFM) demand rate. As the suite OFM 
demand rate increases, the depot's share of that workload should also 
increase. 

Several aircraft characteristics may affect the cost of avionics 
depot repair. Intuitively, aircraft intended for combat should have 
more complex avionics and consequently should be more expensive to 
repair. Because an all-weather capability implies a more complex 
navigation function, aircraft with such a capability should be more 
expensive to repair. Avionics components on lower-performance 
aircraft (e.g., bombers and transports) are subject to lower levels 
of vibration and acoustic noise, are not packed as densely, and 
operate in a more benign temperature environment than do avionics 




36 


components on higher-performance aircraft (fighters and attack 
aircraft). The mission t’^Tpe also captures to some extent the average 
sortie length and the total hours flown. Higher sortie rates are 
generally associated with higher repair costs since as the number of 
sorties increases, the number of times the components are switched on 
and off increases, which in turn leads to a greater incidence of 
failures. Similarly, a higher number of annual flying hours should 
lead to more failures per year. Suite uepot repair cost should also 
be affected by the degree of component commonality among aircraft. 
Greater degrees of commonality should result in greater levels of 
repair-line standardization an^ therefore lower repair cost. 

There are several otner factors which we believe could influence 
avionics depot •epair cost but which were not tested, primarily 
because of definitional problems: Unambiguous definitions applicable 
at the suite level could not be developed. For example, conventional 
wisdom suggests that as an avionics system matures, it.s failure rate 
and repair cost should decrease. Because an aircraft's avionics 
suite changes constantly, however, it is extrenely difficult to 
determine a single value for suite age. .An aircraft's avionics depot 
repair cost should also be influenced by whether or not the suite 
represents a revolutionary or evolutionary technology change. 
Revolutionary change may occur in components (e.g., the change from 
solid state devices to integrated circuits), in the degree of system 
complexity (i.e., the component count), in system philosophy (e.g., 
functional integration vs. functional self-sufficiency), and in 
diagnosis and repair philosophy (e.g., inclusion of self-test 
functions), './hile revolutionary change may be beneficial in the long 
run, in be short run it is usually associated with more unreliable 
operation. Additionally, since a sizable portion of maintenance 
action time is normally attributable to diagnosis, the "ease" of 
diagnosis should also affect repair cost. 

There are two final concepts which will not only not be 
investigated because of definitional ambigi'.ity but for which the 
direction of change in cost can not be postulated with any certainty. 






37 


The first is the level of technology—discrete device or integrated 
circuit. Integrated circuits are probably more reliable than discrete 
devices but may be more expensive to repair. The second is the degree 
of functional integration. Suites with greater degrees of functional 
integration are generally regarded as more difficult to diagnose and 
therefore more expensive to repair. On the other hand, suites with 
greater degrees of self-sufficiency should also be more expensive to 
repair because of the additional components. 


General Variables 

Consideration was given to identifying variables related to 
policies and procedures involving different labor and overhead 
rates at facilities performing similar work. 

The direct labor rate charged by an ALC for a given category 
of work is related to the total amount of work in that category 
that the ALC performs and to the mix of skills possessed by the 
organizational unit doing the work. 

Overhead rates charged by an ALC vary with the total number 
of ALC personnel, the number of personnel performing operations 
overhead and GSeA tasks, and the total number of personnel on the 
base at which the ALC is located. The total "Other Indirect” cost 
is also known as "Operations Overhead," which in total varies with 
direct workload. The rate varies with component class and type of 
maintenance activity. The G&A total is essentially fixed. 

Total costs for similar work packages differ significantly 
among contract maintenance facilities and between contract and 
organic facilities. 

Different labor rates apply to different component classes 
and dilierent categories of maintenance activity, aue to different 
mixes of skills. 

Unfortunately, the WSCRS data files from which o’lr working 
data base was derived do not identify the organization performing 
the reported work. As a result, variablec related to organizational 
entities could not be defined. 











CO?lMof< ASPECr -. Oi- ANALYTICAL APPROACH 

Estimating equations were developed in this study for each of 
the following categories of depot maiutenance activity: (1) airframe 
rework, (2) engine overhaul and repair, (3) airframe component 
repair, (4) engine component and accessory repair, and (5) avionics 
component repair. The data base was divided into separate files for 
this purpose. Some aspects of the analysis are common to all 
categories and are discussed below. Section V presents the results 
and aspects of the analysis that were peculiar to each category. 

Multiple regression analysis was the technique used to examine the 
relationships between cost and potential explanatory variables. Only 
one equation form was used—logarithmic-linear: 


ln(Y) = a + b In(Xj) + c InCx^) + ..., 

where Y is the dependent variable, x^, X 2 , etc., are independent 
variables, and a, b, c, etc. are coefficients to be derived by 
regression analysis. The logaritlunic form was selected because it has 
the advantage that the assumption of normal distribution of error about 
the linearized equation leads to an estimating equation with constant 
percentage error. The alternative equation forms (linear and 
exponential) lead to constant absolute dollar errors. Since m.''>iy 
variables in the data base span large ranges of values, constant 
percentage errors were considered more appropriate. The analysis 
showed that a few variables might be handled better by some other 
transformation, such as a logit transformation, but this was left as a 
subject for future investigation. 

Potential explanatory variables for each depot maintenance 
activity were grouped into three major categories: size, technical/ 
performance, and application/utilization. (Airframe activities 





possessed a third category: policy.) Ideally, an estimating 
relationship would incorporate at least one variable from each 
category. Practically, however, it proved difficult to find such 
estimating relationships. Furthermore, equations incorporating only an 
application (or policy) variable would not be particularly useful since 
the hardware itself would not be defined. Consequently, acceptable 
equations incorporating size and/or technical/performance variables 
were determined first, and then application (or policy) variables were 
added where they were significant. In almost all cases, the number of 
possible variable combinations was small enough that all possible 
regressions could be run and examined to see the effects of each 
variable. 

The estimating relationships were evaluated on the basis of 
statLa^ical quality and intuitive reasonableness. Variable significance 
was utilized as an initial screening device to reduce the number of 
estimating relationships requiring closer scrutiny. Normally, 
only those equations for which all variables were significant at the 
5 percent level (in a one-sided t-test) were documented in this report. 
Occasionally this criterion was relaxed in order to provide a useful 
comparison with an equation that meets the criterion. 

Other statistical measures used in the analysis include the 
coefficient of determination, the standard error of estimate, and the 
F-statistic. The coefficient of determination was used to indicate the 
degree of association between the independent and dependent variables 
in the equation. The standard error was used to indicate the degree of 
variation of the data about the regression line. It is given in 
logarithmic form in this report but may be converted to a percentage of 
the predicted value by performing these calculations: 


+SEE 

e 

-SEE 

e 




J. 











40 


For example, a standard error of 0.30 yields standard error percentages 
of +35 and -26 percent. The F-statistic was used to determine 
whether or not the explanatory variables in an estimating relationship 
are collectively related to the cost variable. Those equations for 
which the probability of the null hypothesis being true (i.e., the set 
of independent variables being unrelated to the dependent variable) is 
greater than 0.05 are identified when the equations are presented. 

Collinearity in two-variable estimating relationships was avoided 
by not testing explanatory variable combinations whose correlation 
coefficient was 0.7 or greater. Collinearity in estimating 
relationships incorporating more than two explanatory variables was 
avoided by rejecting any result for which one explanatory variable's 
correlation with the other equation variables was 0.7 or greater. 

A few equations that did not meet this criterion were derived in the 
course of the analysis. A review of these gives the impression that 
a thorough analysis using a higher critical value, such as 0.8 or 0.9, 
would not be likely to produce equations more useful than those 
arrived at with the 0.7 criterion. 

Plots of equation residuals* were given cursory examinations in 
order to identify obvious patterns and to identify additional 
explanatory variables which might help to explain part of the remaining 
variance. Observations which were believed to be outliers were 
eliminated prior to statistical analysis. 

Finally, the estimating relationships were reviewed for 
reasonableness. All estimating relationships for which the sign of the 
variable coefficient is not consistent with a priori notions, or for 
which the magnitude of a coefficient produces results which do not seem 
credible, have been identified in the presentation of results. 

The acceptable estimating equations are presented in tabular form 
for each cost category. The equations are presented in their 

The most frequently used plots were residuals vs. predictions and 
residuals vs, time (aircraft first flight date or engine MQT). 










41 

* 

exponential form, although the regression analyses were performed 
using the log-linear form discussed above. Statistics presented with 
the equations include: the coefficient of determination (R square), 
the standard error of the estimate (SEE), the F-statistic (F), and the 
sample size (N). The significance level for each variable in an 
equation is shown directly below the mnemonic for the variable. 
Additionally, a comment column provides space for information regarding 
other aspects of the estimating relationships such as the 
reasonableness (sign and magnitude) of the variable coefficients. 

In developing a recommended set of depot maintenance cost estimating 
relationships, we initially tried to select relationships which satisfied 
the following conditions: 

o Each variable is significant at the 5 percent level. 

0 The equation as a whole is significant at the 5 percent level, 
o Individual elements of the equation are credible, 
o Residual plots are free of systematic patterns that indicate 
possible bias in the estimating relationship. 

Once these initial conditions were satisfied, the objective was 
minimization of the standard error of estimate. Tradition suggests 
that a "good" estimate will be within +20 percent of the actual cost. 

As will be seen, however, few of the estimating relationships 
documented herein come close to this objective. 

A 

If the log-linear form is used for a regressii nation, the 
expected cost is given by an equation of the form 

vr / a b c . z 
Y = (e x^ X 2 ...)e 

where z = v/2 and v is the actual variance of the error term in the 
log-linear equation. Although the actual variance is unknown, it can 
be approximated by the square of the standard error of estimate, SEE. 







42 


IV. DEVELOPMENT OF ESTIMATING EQUATIONS 

Separate analyses were conducted for data pertaining to the 
categories of airframe rework, engine overhaul and repair, airframe 
component repair, engine component and accessory repair, and avioni.cs 
repair. An alternative approach was also evaluated: estimating 
annual depot maintenance cost as a total that includes the costs of 
these separate categories without dealing with them individually. All 
the results are presented in this section. Some data plots are 
included here to provide an understanding of the scope of the data 
base. Additional plots are assembled in App. E. 

AIRFRAME REWORK ANALYSIS 

Depot-level airframe rework cost was estimated on the basis of an 
annual cost per aircraft. Analysis of other forms of the dependent 
variable (total annual fleet cost and cost per visit/number of visits) 
is discussed in App. F. The most important descriptive data for 
these aircraft are shown in Table 6. Values for other candidate 
explanatory variables may be found in App. D. 

D ata Base 

Data for 35 different MDS aircraft are provided in Table 6. 
However, the A-lOA, though shown in the table, was not included in 
the analysis because it is so new that no significant depot costs were 
accumulated during the years covered by the data base. The range of 
size and technical characteristics covered by the remaining 34 
aircraft is shown below: 

Cha r acteristic Data Base Range 

Empty weight (lbs) 4067-320,085 

Maximum speed (knots) 325-1434 

Dynamic pressure at maximum speed (psf) 178-1566 

Maximum load factor (g’s) 2.0-8.7 




Table 6 


AIRFRAME REWORK COSTS; AVERAGES FOR 1975-1977 
(Costs in 1978 dollars) 


MDS 

Annual 

Fleet 

Cost 

($000) 

Annual 
Cost per 
Aircraft 
($) 

Cost per 
Visit 
($) 

Annual 

Depot 

Production 

Quantity 

PDM? 

Inven¬ 

tory 

Most 

Represen¬ 

tative 

Series? 

A-7D 

4,778 

13,090 

51,932 

92 

P 

365 

Y 

A-lOA 

3 

94 

— 

0 

N 

29 

Y 

A-37 

1,238 

10,952 

4,139 

299 

P 

113 

Y 

B-52D 

3,011 

33,828 

143,368 

21 

Y 

89 

N 

B-52G 

39,722 

245,195 

630,501 

63 

Y 

162 

Y 

B-52H 

20,551 

230,913 

587,178 

35 

Y 

89 

N 

C-5A 

26,469 

407,222 

71',391 

37 

P 

65 

Y 

C-130E 

10,634 

37,843 

98,461 

108 

Y 

281 

Y 

C-141A 

24,826 

100,105 

206,883 

120 

Y 

248 

Y 

F-4C 

15,254 

56,496 

98,413 

155 

Y 

270 

N 

F-4D 

20,194 

45,482 

87,419 

231 

Y 

444 

N 

F-4E 

28,506 

47,990 

98,980 

288 

Y 

594 

Y 

F-5B 

33 

3,667 

16,502 

2 

P 

9 

N 

F-5E 

915 

17,947 

17,602 

52 

P 

51 

Y 

F-15A 

713 

8,592 

4,542 

157 

N 

83 

Y 

F-IOIB 

337 

3,007 

56,127 

6 

N 

112 

Y 

F-105B 

580 

17,072 

5,635 

103 

P 

34 

N 

F-105D 

2,502 

25,275 

16,907 

148 

P 

99 

Y 

F-105F 

586 

30,830 

24,407 

24 

P 

19 

N 

F-105G 

2,121 

50,497 

151,490 

14 

P 

42 

N 

F-106A 

9,727 

55,583 

127,987 

76 

Y 

175 

Y 

F-106B 

2,161 

58,418 

39,299 

55 

Y 

37 

N 

F-lllA 

474 

5,094 

157,904 

3 

N 

93 

N 

F-illD 

766 

9,115 

85,077 

9 

N 

84 

Y 

F-lllE 

820 

10,380 

410,017 

2 

N 

79 

N 

F-lllF 

236 

2,775 

117,925 

2 

N 

35 

N 

T-33A 

709 

3,138 

8,059 

88 

P 

226 

Y 

T-37B 

1,045 

1,648 

8,707 

12 

N 

634 

Y 

T-38A 

2,606 

2,915 

6,260 

460 

N 

872 

Y 

T-39A 

796 

7,207 

98,200 

8 

P 

109 

Y 

FB-llU 

209 

3,161 

34,767 

6 

N 

66 

N 

KC-135A 

16,938 

25,938 

109,275 

155 

Y 

65? 

Y 

OV-lOA 

473 

5,439 

— 

0 

N 

?7 

. Y 

RF-4C 

15,601 

45,089 

73,243 

213 

Y 

346 

M 

TF-15A 

233 

10,572 

11,075 

21 

N 

22 

N 

NOTE: 

Y = yes 

; N = no; 

P = PDM program for part of 

data base 

time 


period. 










LSf£V..«3 









44 


Some models of aircraft are represented in the data base by a 
single MDS, others by four or five different MDSs. In order to 
evaluate the possible bias caused by this unequal weighting, certain 
parts of the analysis were repeated with a subsample composed of one 
series of each model. These "most representative series" aircraft are 
identified in Table 6. 

A plot of the annual rework cost per airframe as a function of 
empty weight is provided in Fig. 2. An examination of this plot 
yields the following observations: 

o Rework cost tends to increase as empty weight increases. 

0 Data tend to cluster by mission type. 

As a result of the latter observation, certain parts of the analysis 
were repeated with subsamples of fighter/attack and bomber/cargo 
aircraft. Such divisions have intuitive appeal. The fighters and 
attack aircraft tend to be small, fast, and maneuverable whereas the 
bombers and cargo aircraft tend to be large, slow, and not very 
maneuverable, 

Estimating Relationships 

Table 3 (Sec. Ill) lists at least two explanatory variables for 
each of the explanatory variable categories (size, technical/performance, 
utilization, and policy). Ideally, an estimating relationship would 
incorporate one variable from each of the four categories. Practically, 
however, it proved difficult to find such estimating relationships. 
Furthermore, equations incorporating only utilization or policy 
variables would not be particularly useful for predictive purposes 
since the airframe itself would not be defined. Consequently, 
acceptable equations incorporating airframe size and technical variables 
were determined first, and then utilization and policy variables were 
added where they were significant. 

Mnemonics used are as follows: 




























A1''MFGC = airframe manufacturing cost (cumulative average for 
100 airframes; millions of 1978 dollars) 

AGE = aircraft average age (years) 

AFRWKC = annual airframe rework cost per aircraft (1978 dollars) 

EW = aircraft empty weight (lbs) 

MAlNTPCr = percent of airframe rework activity performed organically 
rather than under contract 

RDM = PDM policy (1 = no PDM program, 2 = has a PDM program) 

PQ - production quantity (number of depot visits per year) 


Tot.iJ Sample. Estimating relationships incorporating variables 
significant at the 5 percent level are provided in Table 7. The 
equations are generally of poor statistical quality. Additionally, 
other reservations exist. The exponent of the PDM variable is 
relatively large. This suggests that the annual airframe rework cost 
for aircratt with PDM programs is approximately 10 times that of 
aircraft with no PDM programs. However, these equations say nothing 
about other costs that might be affected by such a decision. A PDM 
is only one p.irl of a scheduled maintenance program. Avoiding use of 
a PDM could require larger than normal costs for base-level scheduled 
inspections. Also, unscheduled maintenance requirements could be 
larger than otherwise would be expected. Such effects are beyond the 
scope of this study but must be addressed in any application of these 
e(piation.s. 

One should also note that the PQ exponent is counterintuitive: 

For every doubling of the production quantity, unit c s^s increase by 
approximately 35 percent. Finally, one should be a'-' of the dramatic 
changes in the empty weight and airframe manufacturing cost exponents 
when the FDM designator is added. 

Most Representative Series . Estimating relationships based on a 
sample consisting of only one observation per aircraft model are listed 
in Table 8. The statistical quality of the estimating relationships 
incorporating empty weight and airframe manufacturing cost improves 
markedly, while the quality of the two estimating relationships 
incorporating the PDM variable improves somewhat. On the other hand, 












47 


Table 7 

AIRFRAME REWORK COST PER AIRCRAFT ESTIMATING 
RELATIONSHIPS: TOTAL SAMPLE 



Statistics 



Equation 

r2 

SEE 

F 


Comments 

Size 






0.904 

AKRWKC = 2.75 EW 

(.000) 

0.46 

1.05 

28 

34 


Teohniaal, Performanae 






1.06 

AFRWKC =44.6 AFMFGC 

(.001) 

0.35 

1.12 

12 

25 


Size/Policy 






0.942 0.403 

AFRWKC = 0.355 EW PQ 

(.000) (.000) 

0.66 

0.86 

29 

33 

Sign of PQ 
exponent 

O.'H- 3.22 

AFRWKC = 183 EW PDM 

(.018) (.000) 

0.84 

0.62 

52 

23 

Exponent 

magnitude 

Teohniaalj PCX'formanae/Policy 






1.22 0.461 

AFRWKC =2.07 AFMFGC PQ 

(.000) (.000) 

0.63 

0.86 

19 

25 

Sign of PQ 
exponent 

0.602 3.43 

AFRWKC = 111 AFMFGC PDM 

(.014) (.000) 

0.85 

0.59 

41 

18 

Exponen' 

.magnitude 

Size/Technical, Performance: none 
Size/Utilization: none 

Technical, Performance, Utilization: 

none 





Size/Teahniaal, Performance/Utilization, Policy: 

none 



iium 











48 


Table 8 

AIRFRAME REWORK COST PER AIRCP.AFT ESTIMATING 
RELATIONSHIPS: MOST PJiPRESENTATIVE SERIES 



1.30 

.300 





AFRWKC =2.49 AFMFGC 

PQ 

0.72 

0.88 

13 

13 

(.000) 

(.070) 





.861 

2.96 





AFRVKC =27.0 AFMFGC 

PDM 

0.90 

0.56 

27 

9 

(.019) 

(.002) 






Siae/Teahniaal, Performance; none 
.jize/Vtilization: none 
Technical, Performance/Utilizatian; none 
Size/Technical, Performanae/Utilization, Policy; none 


PQ does not meet 
the 5% significance 
criterion; sign of 
PQ exponent 



49 



production quantity is no longer significant at the 5 percent level 
in the two equations reported, a not altogether distressing situation 
given the counterintuitive nature of its sign. 

Fighter/Attack Sample . No estimating relationships incorporating 
variables meeting our 5 percent significance level criterion could be 
identified for the fighter/attack sample. This result is not too 
surprising since this particular stratification eliminates much of the 
variation in the size and performance variables. 

Bomber/Cargo Sample . Only a single estimating relationship 
incorporating a variable meeting our 5 percent significance level 
criterion could be identified. The equation, based on airframe 
manufacturing cost, is as follows: 

AFRWKC =4.81 AFMFGC^'^^ 

(. 020 ) 

(R^ = 0.60, SEE = 0.77, F = 8, N = 7) 


Summary . The analysis of annual airframe rework cost per aircraft 
can be summarized as follows: 


o Surprisingly few estimating relationships were identified in 
which all equation variables met our 5 percent significance 
level screening criterion. 

o Of those estimating relationships which did meet our initial 
screening criterion, most were of dubious statistical quality 


The selection of a recommended estimating relationship would seem to 
focus on the following equations: 








50 


Total Sample 


'\UL ^ 22 

AFRWKC = 183 EWPDM 
(.018) (.000) 


R" 


SEE 


F N 


0.84 0.62 52 23 


AFRWKC =111 AFMFGC'^®^ 

(.014) (.000) 


0.85 0.59 41 18 


Most Represeatative Series 

4Q1 2 72 

AFRWKC = 44.8 EW^ PDM ' 0.87 0.63 30 12 

(.027) (.003) 

AFRWKC =27.0 AFMFGC'®^^ 0.90 0.50 27 9 

(.019) (.002) 


All equatioQs include the highly relevant PDM variable. However, as 
mentioned previously, the equations say nothing about base-level 
costs that might be affected by a PDM/no-PDM decision. 


ENGINE OVERHAUL AND R.EPAIR ANALYSIS 

The estimation of engine lifetime overhaul cost requires the 
development of two estimating relationships: the average time between 
overhaul (ATBO) and the average cost to overhaul. Engine lifetime 
repair cost will be estimated on the basis of an annual cost to repair 
per installed engine. ATBO, average cost per overhaul, and average 
annual repair cost data to be used in the analyses are summarized in 
Table 9. Candidate explanatory variable values may be found in 
App. D. 


Data Base 


An examination of Table 9 indicates that the T76 and 10-360 C/D 
apparently incurred no overhaul or repair costs during the 1975-1977 











51 









Table 9 


SUMMARY ENGINE DATA BY TMS: AVERAGES FOR 1975-1977 
(Costs in $ 1078) 


Overhaul Data 


Repair Data 



Engine 

<u 

rH W 
f-1 <1» 

C 

U 

U) CD 

C C 

M U3 

Annual 

Flying Hov, ; 
per Engine 

Average Time 
Between Overnau 
(ATBO) 

Average Cost 
per 

Overhaul ($) 

Average 

Number of 

Annual 

Overhauls 

Average Cost 
per 

Repair ($) 

Average Number 
of Annual 
Repairs 

Average Annual 
Repair Cost pei 
Installed Engii 

J33-A-35 

207 

345 

3260 

2,373 

38 

8,850 

5 

207 

J57-P-13A/B 

87 

258 

1560 

3,863 

1 

— 

— 

— 

-19W/29VIA 

1018 

263 

2978 

36,283 

62 

3,010 

28 

83 

-21A/B 

356 

227 

431 

32,552 

51 

7 

51 

1056 

-23B 

65 

285 

609 

— 

— 

- 

— 

— 

-43WB 

1601 

407 

2904 

29,578 

156 

2,880 

106 

191 

-55/55A 

218 

300 

J246 

32,560 

26 

21,000 

5 

482 

-59M 

2613 

333 

2377 

35,220 

257 

25,400 

3 

29 

J60-P-3/3A 

261 

934 

2210 

8,885 

53 

29,700 

1 

114 

J65-W-5F 

77 

395 

792 

17,280 

42 

— 

— 

— 

J69-T-25 

1397 

437 

3032 

4,255 

260 

840 

1 

1 

J75-P-17 

199 

343 

918 

30,638 

49 

16,000 

22 

1771 

-19/19W 

19A 

226 

921 

30,998 

43 

8,640 

40 

1782 

J79-GE-15 

2112 

258 

1057 

38,883 

462 

3,550 

111 

187 

-17/17A 

1286 

252 

lv48 

31,423 

267 

5,180 

52 

209 

J85-GE-5H 

1831 

40C 

2207 

10,231 

175 

3,140 

1 

2 

-13 

23 

298 

1182 

8,499 

3 

6,920 

1 

301 

-17A 

230 

226 

1528 

— 

— 

— 

— 

— 

-21 

201 

194 

1/6 

— 

— 

n inn 

5 

741 

1F30-P-3 

313 

247 

530 

51,380 

128 

2,400 

270 

5265 

-7 

116 

313 

523 

42,602 

51 

14,600 

32 

4016 

-9 

147 

236 

552 

57,702 

22 

17,000 

20 

2307 

-100 

174 

256 

342 

64,122 

77 

11,200 

51 

3292 

TF33-P-3 

735 

428 

2715 

29,250 

53 

3,000 

44 

180 

-5 

100 

675 

3423 

28,558 

10 

70,600 

2 

1412 

-7/7A 

1095 

1068 

6962 

26.394 

119 

5,460 

40 

200 

-9 

103 

764 

5350 

26,885 

9 

28,000 

2 

543 

TF3A-GE-100 

108 

214 

192 

— 

— 

— 

— 

— 

TF39-GE-1/1A 

277 

631 

1602 

44,324 

46 

14,700 

38 

2018 

TF41-A-1/1A 

354 

299 

355 

88,287 

140 

14,268 

435 

17,532 

FlOO-PW-lOO 

338 

157 

180 

55,561 

3 

26,800 

19 

1506 

-23A 

338 

157 

155 

— 

— 

7,150 

19 

402 

-23B 

338 

157 

183 

57,347 

3 

32,700 

21 

2030 

-23C 

338 

157 

172 

16,734 

6 

6,980 

4 

83 

-23F 

338 

157 

272 

12,039 

4 

16,700 

4- 

99 

-23G 

338 

157 

169 

2,64S 

4 

3,460 

t 

61 

T56-A-7B‘^ 

1596 

574 

2661 

11,592 

287 

3,2x0 

53 

107 

-9Ba 

549 

413 

1814 

13,990 

98 

1,430 

68 

177 

-15 

542 

524 

2588 

14,622 

64 

2,860 

19 

100 

G56-A-7B^ 

1032 

670 

2636 

— 

— 

890 

1 

1 

-9B‘^ 

547 

419 

1628 

12,117 

123 

1,170 

44 

94 

-15‘' 

1276 

481 

1368 

8.938 

436 

1,280 

60 

60 

T76-GE-10A 

92 

339 

1363 

— 

— 

— 

— 

— 

-12A 

90 

347 

1651 

— 

— 

— 

— 



■'T56 gearbox. 

ROTE: - = Ho data reported in WSCRS. 







-stv t'5 A«wt 



52 




time period and were therefore eliminated from the sample. This left 
the T36 as the only turboprop in the sample. Therefore, the T56 was 
also eliminated from further analysis. Two engines (the FlOO and the 
TF34) were eliminated from the sample because they were phasing into 
the inventory during the 1975-1977 time period and therefore did not 
meet the mature engine criteria. Finally, several engines were 
eliminated from the sample to reduce the problem of engine series 
weighting (e.g., the J57 has seven series, the J60 has one). Thus, 
for those engines with multiple series, a particular series was 
retained only if it represented a significant difference in 
performance or application from other series of that engine model. 

The final sample consisted of the following 17 engines: 


J33-A-35 

J57-P-19W/29WA 

-21A/B 

-43WB 

-59W 

J60-P-3/3A 

J65-W-5F 

J69-T-25 

J75-P-17 


J79-GE-15 

J85-GE-5H 

TF30-P-3 

-100 

TF33-P-3 

-7/7A 

TF39'GE-1A 

TF41-A-1/1A 


These engines cover a fairly wide range of technical and size 
characteristics as shown below: 


Characteristic 


Data Base Range 


Turbine inlet temperature (°R) 19C''-2810 

Pressure term (psf) 3400-65,840 

Specific fuel consumption (Ibs/hr/lb) 0.315-1.140 

Weight (lbs) 367-7475 

Military thrust (lbs) 1,025-40,805 






53 


Plots of ATBO, overhaul cost, and annual repair cost as a 
function of the engine pressure term* are provided in Figs. 3, 4, 
and 5, respectively. An examination of these plots yields the 
following observations; 

o For a given mission type, the overhaul interval 

generally decreases as the pressure term increases. 

o The average cost per overhaul increases fairly 
uniformly as the pressure term increases. 

c The annual cost to repair generally increases as 

the pressure term increases, but appears unaffected 
by the aircraft mission type. 

E stimating Relationships 

Table 4 (Sec. IV) lists several variables for each of the 
explanatory variable categories (technical/performance, size, and 
application). Ideally, an estimating relationship would incorporate 
one variable from each of the three categories. Practically, 
however, it proved difficult to find such estimating relationships. 
Furthermore, equations incorporating only an aprlication variable 
would not be particularly useful for predictive urposes since the 
engine itself would not be defined. Consequently, acceptable 
equations incorporating engine performance and size variables were 
determined first, and then application variables were added where 
they were significant. 

*The engine pressure term was selected as the plot parameter 
because it was one of the more successful explanatory variables 
throughout the engine analysis (including accessory and component 
repair). Additional plots for these cost categories utilizing other 
potential explanatory variables may be found in App. E. 

















• Fighter/attack 
■ Bomber/cargo 









j_I_I_I_!_i_ 

9.0 95 10.0 105 

Natural logarithm of engine pressure term 


Fig. 3—Variation of ATBO with engine pressure term 










Natural log of average cost per overhaul 



















• Fighter/attack 
■ Bomber/cargo 






■ ■ 




J_I_I_I_L 

85 9.0 95 10.0 105 

Natural logarithm of engine pressure term 


Fig. 5—Variation of annual repair cost with engine pressure term 







Mnemonics used are as follows: 


ANNCTR = annual cost to repair per engine ($) 

ATBO = average time between overhaul (hours) 

AVGCOH = average cost per overhaul ($) 

MAXTH = maximum thrust (lbs) 

MILTH = military thrust (lbs) 

MISSDES = mission designator (1 = bomber/cargo; 

2 = fighter/attack) 

PRSTERM = engine pressure term (psf) 

REMRATE = base-level engine removal rate (# per 1000 engine hours) 
RSVPCT = percentage of engine operating hours flown by 
Guard and Reserve Personnel 

SELLPR = engine selling price (unit 1000 in 1978 dollars) 

SFC = specific fuel consumption (Ibs/hr/lb) 

SINGDES = single engine designator (multiple = 1, single = 2) 

TEMP = turbine inlet temperature (°R) 

TYPMTC = type maintenance designator (1 = organic; 

2 = contractor) 

WT = engine dry weight (lbs) 


ATBO . Estimating relationships incorporating variables 
significant at the 5 percent level are provided in Table 10. The most 
notable feature of these equations is their generally poor 
statistical quality. Additionally, the magnitude of the turbine 
inlet temperature exponent is quite large in every case. Because of 
the poor statistical quality of these equations, the engine 
base-level removal rate was separated from the rest of the 
performance variables and several new combinations were tested. As 
the statistics indicate, these latter estimating relationships appear 
to be the best of a poor group. 

Average Cost per Overhaul . Costs per overhaul range from under 
$5,000 to over $85,000. Estimating relationships incorporating 
variables significant at the 5 percent level are provided in Table 11. 
Again, the magnitude of the turbine inlet temperature exponent seems 
unusually large for predictive purposes. However, the equations 
incorporating the engine pressure term and either weight or military 
thrust would appear to be acceptable estimating relationships. One of 
the more interesting aspects of these equations is the magnitude 







58 


Table 10 

ENGINE ATBO ESTIMATING RELATIONSHIPS 



Statistics 



Equation 

R® 

.EE 

F 

N 

Comnents 

Performnae 






ATBO - 6.69 X 10^^ 

(.026) 

.23 

.82 

4 

17 

Exponent magnitude; 

F value 

ATBO = 654000 PRSTERM""^®^ 

(.015) 

.28 

.79 

6 

17 


ATBO - 3520 REMRATE"‘^°® 

(.004) 

.39 

.73 

9 

17 


- 473 

ATBO = 936000 SELLPR 

(.025) 

.23 

.81 

5 

17 


Size 






None tested since no a priori rationale could be established. 


Pevfopmnae/Appliaation 






ATBO = (2.99 X 10^®) TEMP“^‘®® MISSDES"^'^® 
(.019) (.017) 

.45 

.71 

6 

17 

Exponent magnitude 

ATBO - (2.24 X 10^^) TEMP'^'^® SINGDES'^'^^ 
(.017) (.044) 

.38 

.76 

4 

17 

Exponent magnitude: 

F value 

ATBO = (2.60 X 10^®) TEMP'^'®* RSVPCt"’^^^ 
(.005) (.008) 

.50 

.68 

7 

17 

Exponent magnitude 

ATBO = 957000 PRSTERM*'®®^ MISSDES"^'^® 

(.007) (.Oil) 

.51 

.67 

7 

17 


ATBO = 167000 PRSTERM"'®®^ SINGDES'^'®® 

(.007) (.017) 

.48 

.69 

7 

17 


- 728 - 129 

ATBO = 1570000 PRSTERH RSVPCT 

(.002) (.004) 

.57 

.63 

9 

17 


ATBO - (5.67 X 10®) SELLPR"*®^^ MISSDES~^’®° 
(.003) (.003) 

.63 

.57 

9 

17 


ATBO = (1.^6 X 10®) SELLPR"'^^® SINGDES”^'^^ 
(.015) (.040) 

.39 

.75 

4 

17 

F value 

ATBO = 790000 SELLPR"'®^® RSVPCT"^'®^ 

(.016) (.027) 

.42 

.73 

5 

17 


Performance/Re liabi Vi ty 






ATBO “ (6.70 X lO"^®) temp”®’®® REMRATE”’^®® 
(.007) (.001) 

.61 

.60 

11 

17 

Exponent magnitude 

ATBO = (1.97 X 10®) PRSTERM"’®^°REMRATE"’^®® 
(.001 (.000) 

.12 

.51 

18 

17 


7 - 529 - 762 

ATBO “ (3.16 X 10 ) SELLPR ’ REMRATE ’ 

(.002) (.000) 

.68 

.54 

15 

17 













Table 11 


ENGINE COST PER OVERHAUL ESTIMATING RELATIONSHIPS I 

' ^ 

i 

—»» - ■■ - ■- -- ■■ ” ■ ■■ - ' ■■ ■■' " - < 

i 

Statistics I 


Equation 


r,® SEE 

F 

N 

Conme^ts ; 

Pei'fcrrrance 





7 

AVGCOH = (2.20 X TEMP^' 

(.003) 

.40 .76 

10 

17 

Exponent magnitude > 

AVGCOH - 1.2A PRESTERM^'®^ 

(.000) 

.73 .52 

40 

17 


AVGCOH » 18A00 SFC'^‘®° 
(.003) 


.41 .76 

10 

17 

Exponent magnitude 

AVGCOH - .166 SELLPR'^^^ 
(.000) 


.79 .46 

55 

17 


Sise 






AVGCOH = 68.1 

(.000) 


.53 .68 

17 

17 


AVGCOH = 11.6 MAXTH'®®® 
(.000) 


.66 .57 

29 

17 

' 

AVGCOH -12.1 MILTH'®^® 
(.000) 


.61 .62 

23 

17 


Performnae/Size 






AVGCOH - (6.34 X lo"^^) TEMP®'^^ MT’®''® 

(.012) (.002) 

.68 .58 

15 

17 

Exponent magnitude 

AVGCOH = .598 PRSTERM'^®® 
(.000) 

wr-®®o 

(.008) 

.82 .43 

32 

17 


735 

AVGCOH » .538 PRSTERM 

(.001) 

(.017) 

.80 .45 

29 

17 

r (PRSTERM, MILTH) - .67 1 


Perfomar e/AppHcation 


None 


Performance/Other 











60 


of the type-of-maintenance designator, which suggests that contract 
maintenance is 30 to 50 percent as costly as organic maintenance. 

Such a result strains credibility and clearly warrants analysis that 
was beyond the scope of this study before the type-of-maintenance 
variable is used in cost estimating. Perhaps the observation that 
most of the contract maintenance engines are on noncombat and reserve/ 
guard aircraft provides a partial explanation. 

Annual Cost to Repair . The annual cost to repair has an unusual 
distribution. Of the 16 observations, 10 are less than $210 per year; 

3 between $1000 and $2000 per year; 2 between $3000 and $5000 per year; 
and 1 over $17,000 per year.* Estimating relationships incorporating 
variables significant at the 5 percent level are provided in Table 12. 
The most notable feature of these estimating relationships is clearly 
the exponent magnitude. Only two equations (VTT and PRSTERM/WT) possess 
variables with exponent magnitudes of less than 2 and in only one case 
(PRSTERM/WT) are the exponents less than 1.5. 

Summary . The estimating relationships listed in Tables 10, 11, 
and 12 for the three elements of engine overhaul and repair (ATBO, cost 
per overhaul, and annual repair cost) have two common features: 
relatively poor statistical quality and large exponents. However, the 
large exponents would be a serious problem only when extrapolations 
beyond the range of the data base are made. Or. the positive side, 
many variables were found to be significant. Those displaying a degree 
of consistency across the three cost elements are as follows: 


Technical/Performance 

Size 

Application 

TEMP 

WT 

MISSDES 

PRSTERM 

MAXTH 

SINGDES 

SELLPR 

MILTH 

RSVPCT 


*The J65 had no repair costs in 1975-77 and therefore is not 
included in the analysis. 






61 










Table 12 

ENGINE ANNUAL COST TO REPAIR ESTIMATING RELATIONSHIPS 


Statistics 


Lqustion 



R^ 

SEE 

F 

N 

Concents 









ANNLTR ' (2.77 X 

(.004) 


.42 

2.09 

10 

16 

Exponent tsagiilcude 

ASSCTR - (3.72 x lo'^) 

2.31 

PRSTERM^ 

(.002) 


.48 

1.97 

13 

16 

Exponent magnitude 

-4 30 

ANNLTR • 995 SFC 

(.016) 



.29 

2.31 

6 

16 

Exponent magnitude 

ANKCtR - (2.41 X lO'^) 

SELUPR^’®® 

(.001) 


.54 

1.85 

17 

16 

Exp tgnitude 









ANNCTR - (3.28 X lO'*) 

(.002) 


.48 

1.98 

13 

16 

Exponent magnitude 

AtiNCTR - (2.52 x lO'^) 

KAXTH^'^^ 

(.000) 


.59 

1.76 

20 

16 

Exponent magnitude 

ANNCTR - (5.16 X lO"^) 

2 13 
MILTH^'^-* 

(.001) 


.53 

l.SS 

16 

16 

Exponent magnitude 

Perj'ctv72nG<i/6t,ze 








ANKCTR • (3.44 x 

Te^IO.8 „1.47 
(.015) (.007) 

.64 

1.70 

12 

16 

Exponent magnitude 

ANNCTR • (2.72 x 10‘*) 

1 49 

prst:rm 

(.026) 

(.028) 

.61 

1.77 

10 

16 

Exponent magnitude 

Perfoivwios/AppUcation 

TEMP^®-® SISCDES®'®^ 
(.001) (.007) 






A.NNCTR - (1.47 X 10'^^) 

.64 

1.71 

11 

16 

Exponent magnitude 

ANNCTR - (2.62 X lO'**) 

temp^^'^ rsvpct'^®^ 
(.001) (.020) 

.58 

1.84 

9 

16 

Exponent magnitude 

ANNCTR - (2.26 X 10*^') 

9 90 -4 11 

TEKP^ TYPHrr 

(.028) (,010) 

.62 

1.74 

11 

16 

Exponent magnitude 

ANNCTR - (1.74 X 10"’) 

PRSTERM^’^^ 

(.000) 

SINCDES^*®^ 

l.OOO) 

.80 

1.27 

26 

16 

Exponent taagnitude 

ANNCTR - (3 65 X lO'®) 

PRSTERM^'^^ 

(.000) 

RSVPCT*^ 

(.001) 

.66 

1.65 

13 

16 

Exponent magnitude 

ANNCTR - 29.8 SFc"^'®^ 
(.001) 

MISSOES*’^^ 

(.012) 


.52 

.96 

7 

16 

Exponent magnitude 

ANNCTR - 107 SFC“®'®® 
(.001)' 

4 92 

SINGOES 

(.013) 


.60 

1.79 

10 

16 

Exponent magnitude 

ANNCTR ■ 649 

(.003) 

.337 

RSVPCT 

(.015) 


.51 

1.99 

7 

16 

Exponent magnitude 

ANNCTR • (5.02 x lo"^^) 

2 19 
SEUPR^**’ 
(.000) 

SINCOES*"'® 

(.001) 

.78 

1.33 

23 

16 

Exponent magnitude 

Size' ^ ^'lioation 

ANNCTR - (2.48 x lO”^) 

HISSDES’*^® 

(.000) (.00?: 

.67 

1.62 

13 

15 

Exponent magnitude 

ANNCTR - .00321 

(.008) 

TFDES^*'*' 

(.044) 


.59 

1.83 

9 

16 

Exponent magnitude 

ANNCTR • (1.78 X 10'®) 

HAXTH^'^® 

(.000/ 

2 94 

MISSOES*’’ 

(.014) 

.72 

1.52 

17 

16 

Exponent magnitude 

ANNCTR - (2.44 x lo"®) 

MAXTH^’®^ 

(.000) 

3.04 

SINCDES^ 

(.017) 

.71 

1.52 

16 

16 

Exponent magnitude 

ANNCTR • (6 77 X 10‘“) 

(.000) 

MISSDES*'**’ 

(.000) 

81 

1.24 

27 

It 

Exponent iMgnltiKje 

ANNCTR - (3.74 x lo'®) 

(.000) 

SINCDEs’’*"’ 

(.017) 

.67 

1.63 

13 

16 

Exponent k%arnitude 





n 



1 

i 

I 

I 

I 
















WUJlIUUi '■'fl'* 

■*J 




62 


One possible "set" of estimating relationships which tends to m inimize 
potential problems consists of the following equations: 

7qo '^QO 

AVGCOH = 0,598 PRSTERM WT 


ATBO = 957000 PRSTERM"'^®^ MISSDES"^'^^ 


-R 1 49 1 24 

ANNCTR = 2.72x10 PRSTERM^ WT 



AIRFRAME COMPONENT REPAIR ANALYSIS 

The category "airframe components" includes structural 
components, landing gear, utilities, and a variety of other 
miscellaneous systems. These subcategories are defined in more 
detail in App. A. 

Airframe component repair cost will be estimated on the basis of 
an annual cost per aircraft. Cost data used in the analysis are 
summarized in Table 13. Potential explanatory variables are listed in 
Table 3 (Sec. III). Values for the candidate variables may be found 
in App. D. 

Data Base 

The sample used in the airframe component repair analysis is the 
same sample used in the airframe rework analysis. The A-lOA has again 
been omitted because of the lack of cost data in the 1975-1977 time 
period. Despite the deletion, the annual airframe component repair 
cost per aircraft still varies by two orders of magnitude (from $1500 
for the OV-lOA to $150,000 for the C-5A). 

A plot of the annual airframe component repair cost as a function 
of the aircraft empty weight is provided in Fig. 6. An examination of 
tue plot yields the following observations: 

o Component repair cost tends to increase as empty weight increases. 

o The data points tend to cluster by mission type. 


i 










63 




Table 13 

AIRFRAME COMPONENT COSTS: 
AVERAGES FOR 1975-1977 

(Costs in 1978 dollars) 


MDS 

Cost per 
Aircraft 
($) 

Most 

Representative 

Series? 

PDM? 

A-7D 

5,035 

Y 

P 

A-lOA 

389 

Y 

N 

A-37 

2,097 

Y 

P 

B-52D 

62,521 

N 

Y 

B-52G 

70,120 

Y 

Y 

B-52H 

73,698 

N 

Y 

C-5A 

153,221 

Y 

P 

C-130E 

34,165 

Y 

Y 

C-14U 

61,791 

Y 

Y 

F~4C 

16,022 

N 

Y 

F-4D 

16,175 

N 

Y 

F-4E 

14,602 

Y 

Y 

F-5B 

19,954 

N 

P 

F-5E 

3,359 

Y 

P 

F-15A 

5,115 

Y 

N 

F-IOIB 

10,436 

Y 

N 

F-105B 

14,723 

N 

P 

F-105D 

12,239 

Y 

P 

F-105F 

19,835 

N 

P 

F-105G 

14,895 

N 

P 

F-106A 

25,119 

Y 

Y 

F-106B 

40,649 

N 

Y 

F-lllA 

24,729 

N 

N 

F-lllD 

28,177 

Y 

N 

F-lllE 

30,635 

N 


F-lllF 

29,998 

N 

N 

T-33A 

3,376 

Y 

P 

T-37B 

1,547 

Y 

N 

T-38A 

2,205 

Y 

N 

T-39A 

7,351 

Y 

P 

FB-llU 

29,518 

N 

M 

KC-135A 

12,665 

Y 

Y 

OV-lOA 

1,539 

Y 

N 

RF-4C 

18,A32 

N 

Y 

TF-15A 

12,195 

N 

N 


NOTE: Y = yes; N = no; P = PDM for 
part of data base time period. 












65 






Thus, as was the case with airframe rework, certain parts of the 
analysis were repeated with subsamples of fighter/attack and 
bomber/cargo aircraft. Additionally, as was also the case with airframe 
rework, certain parts of the analysis were repeated with a subsample 
of one series of each a" 'craft model in order to evaluate the possible 
bias caused by the unequal series weighting. These "most representative 
series" aircraft are identified in Table 13. 

Estimating Relationships 

The structure of the airframe component repair analysis is 
identical to that used for the airframe rework anaJvis' Acceptab ]j 
equations incorporating airframe size and technica. \a’'iables were 
determined first, and then utilization and policy variables were added 
where they were significant. 

Mnemonics used are: 

ABDES = afterburner designator (1 = no afterburner, 

2 = atterburner) 

AFMFGC = airframe manufacturing cost (cumu.l.': ive average for 100 
airframes; 1978 dollars) 

AFCCST = annual airframe component repair cost per aircraft (1978 
dollars) 

EW = empty weight (lbs) 

SORPAC = average number of annual sorties per possessed aircraft 

T otal Sampl e. Estimating relationships incorporating variables 
significant at the 5 percent level are provided in Table 14. The 
equations are relatively good, although the standard error of estimate 
is somewhat higher than desirable. 

Most Representative "p i.es . Estimating relationships incorporating 
variables significant at , 5 percent level are provided in Table 15. 

Contrasting these equations to those developed for the total sample, 
we find there is very little difference. 

Mission Sam p les . Estimating relationships incorporating variables 
significant at the 5 percent level are provided in Table 16. 

Contrasting these equations with those developed for the total sample. 







. - »'*^^ .-s- - * . s 

66 

Table 14 

AIRFRAME COMPONENT REPAIR COST ESTIMATING 
RELATIONSHIPS: TOTAL SAMPLE 




Statistics _ 

_Ration R^ SEE F N 

Size 

AFCCST = 0.788 .78 .54 116 34 

(. 000 ) 

Teahniaal, Performanae 

AFCCST = 19.0 AFHFGC^*®^ .78 .44 81 25 

(. 000 ) 

Size/Tedhniaal, Perfomkmoe 
None 

Size/Uti lizatim 

AFCCST = 0.394 ABDES®*^^^ .82 .50 72 34 

•.000) (.008) 

Teahniaal, Pevfci'mnoe/Vti tizaticn 

AFCCST = 0.808 AFMFGC^’^® SOEPAC®*^®^ .81 .41 48 25 

(.000) (.024) 

Size/Policy 
None 


TeahniaaZ, Performanae/PoZicy 

















68 


Table 16 

AIRFRAME COMPONENT REPAIR COST ESTIMATING 
RELATIONSHIPS: MISSION SUBSAMPLES 




Statistics 


Equation 


SEE F 

N 

Fighter/Attaok Subsample 

Size 




AFCCST = 

0.894 

(.000) 

0.48 0.58 17 

20 

Teehniaal, 

Performance 



AFCCST = 

144 AFMFGC®*’^^ 
(.003) 

0.43 0.39 11 

16 


No other fighter/attack estimating relationships 
uncovered. 


Bomber/Cgpgo Subsar/jple 


Si:.-- 

0 974 

AFCCST = 0.603 EW”*’' 
(.011) 

0 61 

0.51 

9 

8 

Technical, Performance 





1 03 

AFCCST = 23.1 AFMFGC 

(.007) 

0.73 

0.44 

14 

7 

.0 other bomber/cargo estimating relationships 
uncovered. 


we find that based on the standard error of estimate ther^ is little 
difference. Furthermore, the empty weight exponent remains remarkably 
stable regardless of the sample selected. On the other hand, the 
airframe manufacturing cost exponent fluctuates considerably for the 
fighter/attack sample. 

Excursion . The PDM designator, when utilized as a dummy variable, 
did not meet our 5 percent significance level criterion. An excursion 
was made in which the total sample was split according to PDM policy 
(see Table 13). Those aircraft which switched policy during the 




P - 








69 

1975-1977 time period were excluded. The results of this analysis 
are as follows: 


Aircraft Without PPM 

1 10 

AFCCST =3.06 AFMFGC 

(. 000 ) 


AFCCSr = 0.019 

(. 000 ) 


R^ SEE _F_ CO MMENT 

0.99 .10 499 8 

0.97 .22 308 11 Exponent 

magnitude 


Aircraft With PPM 

A OOA 

AFCCST = 81.2 AFMFGC”* 0.40 .55 7 12 

(.014) 

AFCCST = 65.3 EW 0.51 .49 10 12 

(.004) 

As indicated, the estimating relationships for aircraft without 
a PDM program are quite good. However, one should be aware that the 
"without PDM" sample does not include any large bomber or cargo aircraft 
Consequently, the "without PDM" estimating relationships should probably 
not be used for large bombers or cargo aircraft. This is particularly 
true for tf _■ equation containing empty weight because of the large 
exponent (1.32). The equations for aircraft with a PDM program are 
not nearly as attractive, but do not compare very unfavorably with 
estimating relationships derived for samples not differentiated by 
PDM policy. 

S ummar y. Empty weight and airframe manufacturing cost both seem 
to do a reasonable job of explaining annual airframe component repair 
cost. Additior.<.ily, although sample stratification by mission type 
does not provide any significant benefit, the distinction between 




70 


aircraft with a PDM program and those without would seem to be 
important. For general usage, one of the following estimating 
relationships is suggested: 

AfCCST = 0.788 

AFCCST =19.0 AFMFGC^'®^ 

ENGINE compoct:i;t and accessory repair analysis 

Engine component and accessory repair cost will be estimated on 
the basis of an annual cost per installed engine. Cost data to be 
used in the analyses are summarized in Table 17. Candidate 
explanatory variable values (the same as those used in the engine 
overhaul and repair analysis) may be found in App. D. 

Data Base 

The data base is initially limited to the same 17 engines used 
in the engine overhaul and repair analysis: 


J33-A-35 

J57-P-19W/29WA 

-21A/B 

-43WB 

-59W 

J60-P-3/3A 

J65-W-5F 

J69-T-25 

J75-P-17 


J79-GE-15 

J85-GE-5H 

TF30-P-3 

-100 

TF33-P-3 

-7/7A 

TF39-GE-1A 

TF41-A-1/1A 


Additionally, an examination of Table 17 indicates that the J75 
is at least an order of magnitude less expensive than any other 
engine on the list above. Further examination of the costs in a more 
disaggregated form strongly suggests an error in the raw 
data. The J65 is therefore deleted from the sample. 




















Tab:e 17 

ENGINE COMPONENT AND ACCESSORY REPAIR COST: 
AVERAGES FOR 1975-1977 

(Costs in 1978 do’^ars) 


Engine 

Installed 

Engine 

Annual 
Engine 
Component 
Repair Cost 
($/engine) 

Engine 

Installed 

Engine 

Annual 
Engine 
Component 
Repair Cost 
($/engine) 

J-33-A-35 

207 

1,374 

TF33-P-3 

735 

5.888 

J-57-P-13A/3 

87 

8,168 

-5 

100 

4 

-19W/29W.V 

1018 

11,066 

-7/7A 

1095 

13,604 

-2U/B 

356 

12,291 

-9 

103 

10 

-23B 

65 

7,795 

TF34-GE-100 

108 

3,684 

-43WB 

1601 

8,273 

TF39-GE-1/1A 

277 

43,774 

-55/55A 

218 

24.551 

TF41-A-1/U 

354 

24,783 

-59W 

2613 

5,738 

FlOO-PW-lOO 

338 

7,926 

J60-P-3/3A 

261 

3.325 

-23A 

338 

126 

J65-W-5F 

77 

98 

-23B 

338 

41 

J69-T-25 

1397 

912 

-23C 

338 

10 

J75-P-17 

199 

38,486 

-23F 

338 

3,923 

-19/19W 

194 

25,730 

-23G 

338 

12 

J79-GE-15 

2112 

9,030 

T56-A-7B 

1596 

5,747 

-17/17A 

1286 

7,598 

'"B 

549 

not incl. 

J85-GE-5H 

1831 

1,550 

*5 

542 

not incl. 

-13 

23 

7,209 

G56-A-7B 

1032 

59 

-17A 

280 

1,314 


547 

not incl. 

-21 

201 

402 

-15 

1276 

not incl. 

TF30-P-3 

313 

27,198 

T76-GE-10A 

92 

1,035 

-7 

116 

27,745 

-12A 

90 

1,011 

-9 

147 

15,653 




-IOC 

174 

28,615 

, J 





Figi re 7 is a plot of annual engix.e component an« accessory repair 
cost as a function of the engine pressure term. Appendix E contains 
additional plots using other potential explanatory variables. 

Estimating Relationships 

Estimating relationships incorporating variables significant at 
the 5 percent level are provided in Table 18. Mnemonics used are as 
follows: 














Natural logarithm of annual engine component and accessory repair cost 












73 


Table 18 


ENGINE ANNUAL COMPONENT AND ACCESSORY REPAIR 
COST ESTIMATING RELATIONSHIPS 




Statistics 



Equation 

r2 

SEE 

F 

N 

Coismencs 

Pei‘foi‘i!uncc 






ENGACC = (1.67 x lO"^^) TEMp’"^® 

(.001) 

.68 

.90 

13 

16 

Exponent magnitude 

ENGACC = .086 PRSTERM^'^^ 

(.000) 

.65 

.73 

26 

16 

Exponent magnitude 

ENGACC - 5920 SFC“^'^^ 

(.001) 

.50 

.88 

14 

16 

Exponent magnitude 

ENGACC - .00367 EELLPR^'^^ 

(.000) 

.79 

.57 

52 

16 

Exponent magnitude 

Size 






ENGACC - .605 

(.000^ 

• &6 

.73 

27 

16 


ENGACC • .196 MAXTH^'^* 

(.000) 

.81 

.54 ' 

61 

16 

Exponent magnitude 

ENGACC - .299 

(.000) 

.76 

.64 

40 

16 

Exponent magnitude 

Bei'fornanae/Size 






ENGACC - (6.58 X 10“^*) TEMP*’^* 

(.002) (.000) 

.83 

.54 

31 

16 

Exponent magnitude 

ENGACC = (6.69 X lo"^°) TEMP^'®^ MAXTH'®®^ 

(.036) (.000) 

.86 

.69 

39 

16 

Exponent magnitude 

ENGACC =■ (1.30 X lO"^^) TEMP^'®^ MILTH’®^® 

(.036) (.000) 

.80 

.58 

26 

16 

Exponent magnitude 

778 677 

ENGACC - .0265 PRSTEBM WT' 

(.001) (.001) 

.86 

.52 

34 

16 


ENGACC = 70.6 PRSTERM'®®® 

(.008) (.001) 

.86 

.52 

33 

16 

r (PRSTERM, MILTH) •= .67 

Performanoe/Application 

ENGACC = (7.36 X lO"^*^) TEMP^’®® SORTEHG 

.58 

.86 

9 

16 

Exponent magnitude; sign of 
SORTENG 

ENGACC - .0311 PRSTERM^'®^ SINGDES^'^® 

(.000) (.034) 

.73 

.67 

18 

16 

Exponent ma.gnitude 

-2 92 .103 

ENGACC - 6090 SFC RSVPCT’^"-’ 

(.000) (.067) 

.60 

.82 

10 

16 

Exponent magni'udc 

ENGACC - .00238 SELLPR*"^® SINGDES’^®^ 

(.000) (.050) 

• 

'3 

32 

16 

Exponent magn .tude 

Size/Application 






903 .971 

ENGACC - 6.68 TFDES' 

(.000) (.050) 

.73 

.68 

17 

16 


ENGACC ’ .0186 MILTH^’®® MISSDES^’^® 

.83 

.53 

33 

16 

Exponent magnitude 


(.000) (.009) 









t-i'^lfi^s^'^j^' V .WA^-yykfjrv- ^ *><> 


74 


ENGACC = 


MAXTK 

MILTH 

PRSTERM 

RSVPCT 


SELLPR 

SINGDES 

SORTENG 

TEMP 

TFDES 

WT 


annual engine component and accessory repair cost per 
engine ($) 

maximum thrust (lbs) 
military thrust (lbs) 2 

engine pressure term (Ibs/ft'^) 

percentage of engine operating hours flown by Guard/ 
Reserve personnel 

engine selling price (unit 1000 in 1978 dollars) 
single engine designator (multiple = 1; single = 2) 
annual engine sortie rate (sorties/year) 
turbine inlet temperature (°R) 
turbofan designator (1 = no; 2 = yes) 
engine weight (lbs) 


As was the case with the engiro o’^erhaul and repair estimating 
relationships, t< - -'omponent and accessory repair CERs also exhibit 
some fairly large -■ ^onents. However, the size and performance CERs 
are, as a group, the best of the engine depot-level estimating 
relationships documented in this report. As to the choice of which 
component repair equation to actually use, consistency with the 
overhaul and repair equations would suggest the following estimating 
relationship; 


ENGACC = .0265 PRSTERM*^''® 


AVIONICS COMPONENT REPAIR COST 


Avionics component repair cost will be estimated on the basis of 
an annual cost per aircraft. Cost data to be used in the analysis 
are summarized in Table 19. Candidate explanatory variable values 
may be found in App. D. 


Data Base 


The A-lOA, F-15A, and TF-i5A were excluded from the analysis 
because they were phasing into the inventory during the 1975-1977 
time pe;ici and consequently are of dubious value to a study oriented 
to the osts of mature systems. The 32 remaining avionics suites 
were included in the analysis according to the availability of input 
data. Suite characteristics proved difficult to obta^-u because of 




Table 19 


AVIONICS COMPONENT REPAIR COST: 
AVERAGES FOR 1975-1977 


(Costs in 1978 dollars) 





I 


Annual 






MDS 

Inventory 

Annual Fleet 
Flying Hours 

Avionics Repair 
Cost per Aircraft 

A-7D 

365 

94,556 

19,749 

A-lOA 

29 

13,270 

5,454 

A-37 

113 

28,537 

4,218 

B-52D 

f9 

31,752 

99,897 

B-52G 

162 

69,240 

116,793 

B~52H 

89 

38,182 

160,808 

C-5A 

65 

44,430 

289,866 

C-130E 

281 

170,188 

40,859 

C-141A 

248 

277,727 

83,207 

5-4C 

270 

62,261 

38,288 

F-4D 

444 

106,309 

31,755 

F-4E 

594 

152,329 

29,306 

F-5B 

9 

3,260 

28,895 

F-5E 

51 

12,121 

16.717 

F-15A 

83 

18,544 

8,289 

F-IOIB 

112 

26,779 

21,360 

F-105B 

34 

8,015 

31,065 

F-105D 

99 

21,645 

32,777 

F-105F 

19 

3,921 

50,660 

F-105G 

42 

8,818 

42,923 

F-106A 

175 

53,969 

69,226 

F-106B 

37 

11,532 

129,032 

F-lllA 

93 

17,602 

97,902 

F-lllD ' 

8« 

16,837 

176,154 

F-lllE 

79 

20,010 

125,397 

F-lllF 

85 

21,609 

117,030 

T-33A 

226 

64,234 

6,262 

T-37B 

634 

291,079 

4,595 

'f-38A 

872 

350,926 

7,946 

T-3yA 

109 

101,996 

24,436 

FB-lllA 

66 

17,520 

136,303 

KC-135A 

653 

212,491 

23,585 

OV-lOA 

87 

29,205 

13,010 

RF-4C 

346 

91,975 

51,338 

TF-15A 

22 

5,436 

15,711 



























76 


the number of black boxes and contractors involved with each aircraft 
and because suites change constantly, not only between aircraft of a 
given series but also on a given aircraft (tail number). Additionally, 
since most avionics contracts are firm fixed price, contractors are not 
required to divulge costs. However, the data base we have been able to 
put together covers a wide range of characteristics: 


Characteristics 



Range 

Number 
of Points 

Suite weight (lbs) 



230-6200 

16 

Number of black boxes 



5-34 

31 

Number of functions 



2-6 

31 

Suite procurement cost 

#1 

($)b 

26,000-3,705,000 

29 

Suite procurement cost 

#2 

($)^ 

220,000-10,410,000 

16 

Mean time between OFM 

demands (FH) 

0.75-13.72 

16 


^Procurement cost of avionics suite at unit 100 in 1978 dollars 
(based on contract data). 

^Sum of D041 item (NSN) procurement costs for all items in 
avionics suite (average of 1975-1977 entries). 


Figure 8 is a plot of the annual avionics component repair cost 
as a function of the suite procurement cost. Appendix E presents 
additional plots using other potential explanatory variables. 


Estimating Relationships 

Estimating relationships incorporating variables significant at 
the 5 percent level are provided in Table 20. Mnemonics used are as 
follows: 


AVCST = annual avionics repair cost per aircraft ($) 

AVWT = avionics suite weight tlbs) 

AWXDV = all-weather capability dummy variable (no = 1; yes = 2) 
BLBOX = number of black boxes in suite (#) 

FHRATE = annual flying hour rate (hours) 

FUNG = number of electronics functions performed by aircraft 
avionics suite (//) 

MISSDV = mission dummy variable (1 = noncombat aircraft; 

2 = combat aircraft) 















Natural logarithm of annual avionics component repair cost 






A 

• Fighter/attack 





■ Bomber 





A Cargo 




• 

— o Reconnaissance 




■ 

Q Trainer 


• 


■ 

• ■ 

— 



A 

• ■ 

— 


• 

• 

• 

• 



• 


• • 

A • 


a 

• 


• 

• 



o 




-Q 

□ 

f 1 

□ 

1 


1 1 

1 




















78 


Table 20 

AVIONICS COMPONENT REPAIR COST ESTIMATING PELATIONSHIPS 


Statistics 


Hquation 


SHE F N Comments 


1.09 


AVCSr =22.2 AVWT 

(. 000 ) 


.65 .69 26 16 


Cortiptexity 


AVCST = 

1.49 

918 BLBOX ^ 

(.000) 

.50 

.78 

29 

31 • 

Exponent magnitude 

AVCST = 

4080 FUNC^'^^ 

(.000) 

.49 

.79 

28 

31 

Exponent ir.agnltude 

AVCST » 

34.5 SUITEl'^^^ 

(.000) 

.66 

.68 

51 

29 


AVCST - 

.415 SUIT22'®^^ 

(.000) 

.79 

.54 

54 

16 


AVCST = 

136000 MTBD'^'°^ 

(.000) 

.56 

.79 

18 

16 


AVCST = 

14000 AWXDV^'*^ 

.53 

.75 

34 

32 

Exponent magnitude 


(. 000 ) 


Size/Pevfsnmnae, Complexity 
.742 

' BLB 

(.006) (.030) 


AVCST = 17.9 BLBOX'^®^ 


.74 .62 18 16 r (AVWT, BLBOX) » .66 


Size/App lication 


AVCST = 65.5 AVHT^'*’^ MISSDV'^'^® 


.75 .61 19 16 Exponent sign 


(.000) (.023) 

Performnce, Complexity/Misaion Descriptors 


1 50 959 

AVCST - 527 BLBOX HISSDV^ ^ 


(. 000 ) 


.858 


(.018) 
1.65 


AVCST = 2040 BLBOX AWXDV 
(.005) (.002) 


AVCST = .074 SUITE2'^^® MISSDV"*^^^ 
(.000) (.035) 


Performance, Complexity/Application 


AVCST - 34200 BLBOX^'^^ SORTRAT'E"^'^® 


AVCST = .485 SUITEl 

(. 000 ) 


(. 001 ) 

.593 


(.036) 
.654 


raRATE 

(.013) 


AVCST = .00455 S0ITE2'®^® TWRATE'*^® 


(. 000 ) 
,-1.2 
(. 000 ) 


AVCST - 281 MTBD"^'^^ FHRATE 


(. 012 ) 
1 . 0 / 
(. 000 ) 


57 

.73 

19 

31 

Exponent magnitude 

63 

.68 

24 

31 

Exponent magnitude 

.84 

.49 

34 

16 

Exponent sign on MISSDV 

,56 

.75 

17 

31 

Exponent sign 

,72 

.63 

33 

29 


.86 

.46 

41 

16 


.72 

.65 

17 

16 

Exponent magnitude 


Si „ 


















79 


MTBD = mean time between OFM demands (hours) 

SORTRATE = annual sortie rate (sorties) 

SUITE 1 = procurement cost of avionics suite at unit 100 (1978 

dollars) 

SUITE 2 = sum of D041 item (NSN) procurement costs for all items 

in avionics suite ($) 

Several of the estimating relationships have a familiar problem— 
exponent magnitude. Others have exponents with counterintuitive 
signs. For example, the sign of the mission dummy variable in the 
AVWT/MISSDV and SUITE2/f'IISSDV estimating relationships is negative, 
indicating that the cost of avionics component repair for combat 
aircraft is less than for noncombat aircraft. Equally interesting is 
the fact that the mission duimny variable has a positive exponent in the 
BLBOX/MISSDV estimating relationship. This flip-flopping of the 
exponent sign suggests a relatively unstable variable. 

On the positive side, the CERs utilizing suite procurement cost 
appear quite acceptable. The SUITE2 variable actually does better 
from a statistical standpoint than does the SUITEl variable. 
Intuitively, one might expect this, since SUITE2 is based on the 
curr'=*nt cost of suite items (including costs of items incorporated as 
the result of modifications) while SUITEl is based on quantity- 
normalized historical data. Table D.3 indicates significant 
differences between the two procurement cost values, but the 
correlation between the two is relatively high (r = .90). If one 
should want to use an estimating relationship utilizing suite 
procurement cost, then one of the following is recommended: 

©1 1 

AVCST = .415 SUITE2 

AVCST = .00455 SUITE2'®^® FHRATE'^^® 

The SUITE2 value for proposed aircraft can be approximated by 
the estimated avionics suite cost at the projected production. 
Attempting to estimate mature aircraft SUITE2 costs would be an 
exercise in futility, because of modifications unknown during the early 
stages of design. 






80 


TOTAL COST EQUATIONS i 

The equations developed for the different categories of depot | 

maintenance activity have one common feature: They have poorer 
statistics (e.g., higher standard errors of estimate) than we would 
like. An alternative approach was therefore considered: the use of 
subsystem parameters to estimate the total cost of all types of depot 
maintenance activity. Although this does not give insight into the 
relative costs of the individual categories, sensitivity to subsystem 
characteristics can be retained. 

The cost and explanatory variable data used for the individual 
categories gave full coverage of the data needed for 19 weapon systems. 

The total cost for each system was evaluated and analyzed as a cost per 
possessed aircraft. These values are presented in Table 21.* 

The objective of the analysis was to develop equations that 
include variables describing the airframe, engine, and avionics 
subsystems as well as the aircraft utilization. Table 22 presents the 
best results obtained from this data. Mnemonics used are: 


Variable 

Subsystem 

Definition 

ACFFD 

Avionics 

Aircraft first flight date (months since 
January 1943) 

EW 

Airframe 

Empty weight (lbs) 

INV 

System 

Inventory size, the number of possessed 
aircraft 

MILTH 

Engine 

Military thrust (lbs) 

NENG 

System 

Number of installed engines per airframe 

PRSTERM 

Engine 

Engine pressure term (psf) 

SELLPR 

Engine 

Selling price for 1000th engine (1978 dollars) 

SFC 

Engine 

Specific fuel consumption (Ibs/hr/lb) 

SUITE1 

Avionics 

Procurement cost of avionics suite at unit 

100 (1978 dollars) 

TCSTPAC 

System 

Annual depot maintenance cost per aircraft 
(1978 dollars) 

WT 

Engine 

Engine dry weight (lbs) 


*Appendix C presents a comparison of these costs with 
corresponding costs taken from output for 1977 from the Air Force's 
Operating and Support Cost Analysis Report (OSCAR). 











81 

Table 21 

AVERAGE DEPOT MAINTENANCE COST; 
1975-1977 

(In $ thousand 1978) 



Annual Cost 

MDS 

per Aircraft 

A-7D. 

.. 145 

B-52D . 

. 

B-52G . 

. 

B-52H . 

. 551 

C-5A. 

. nhQ 

C-141A . 


F-4C. 

. ^Lf^ 

F-4D. 

. 129 

F-106A . 

. ?m 

F-106B . 


F-lllA . 

. 229 

F-lllE . 

. 280 

F-lllF . 

. 

T-33A . 

. 15 

T-37B . 

. 11 

T-38A . 

. 20 

T-39A. 

. 53 

KC-135 . 


RF-4C . 



Table 22 

TOTAL SYSTEM DEPOT-LEVEL COST ESTIMATING RELATIONSHIPS 


Equation 

1. TCSTPAC =7.49 PRSTEEM®’ 


2. TCSTPAC =4.66 INV 

3. TCSTPAC = 8.51 INV 


5. TCSTPAC =5.57 INV' 


Statistics 
SEE F N 

.93 .37 62 19 


(.008) 

(.000) 

(.001) 






,-0.332 

—0.573 . 

0.467 







tW SELLER ■ 


.93 

.36 

64 

19 

(.005) 

(.000) 

(.001) 






,-0.408 

_..0.698 . . 

0.862 







EW ACFFD 


.93 

.37 

62 

19 

(.001) 

(.000) 

(.001) 






,V-0.434 3^^-1.45 

SUITE1°'^^® 


.90 

.47 

38 

17 

(.008) 

(.006) 

(.001) 






,-0.402 

ACFFD^'®^^ 

WT°'®^® NENG°’^^^ 

.96 

.28 

85 

19 

(.000) 

(.000) 

(.000) 

(.000) 





-0.354 

ACFFD°'®^^ 


NENG®’^^^ 

.97 

.23 

130 

19 

(.000) 

(.000) 

(.000) 

(.000) 




























82 


As shown in Table 22, no estimating relationships incorporating 
one airframe, one engine, one avionics, and one utilization variable 
could be found. The parameters associated with each subsystem tend to 
be correxated with those of the other systems. The best equations, as 
shown in the table, include parameters representing, at most, two of 
the three subsystems. The statistics of these equations are better 
than those of the equations for the individual categories. The one 
variable included in all these equations is inventory size, INV. Since 
its exponent is negative, we have evidence of a fixed cost being 
associated with depot maintenapce. Figure 9 illustrates this effect. 

The last two equations in the table have the best statistics of 
any generated during this study; 

TCSTPAC = 5.57 ACFFD'^^^ WT'®^^ NENG'^^^ 

- 691 QS"! 522 

TCSTPAC = 2.51 INV ACFFD MILTH NENG 



• Fighter/attack 
■ Bomber/cargo 
▲ Other 


J_I_I I_I_ ti 

4.0 4.5 Si) 5.5 6i) 6J 

Natural iogarithin of inventory size 


Fig. 9—Variation of total cost per aircraft with inventory size 











84 


V. SUMMARY OF RESULTS AND CONCLUSIONS 

The previous section presented a large number of worthwhile 
estimating relationships. We now consider the overall implications 
of the study results. A summary of the significant variables for 
airframe, engine, and avionics maintenance is followed by an example 
of some applications of some subsystem- and system-level estimating 
relationships. This section closes with suggestions for possible 
extensions of this research during future studies. 

PRINCIPAL FINDINGS 

Several equations, instead of only one, were generated for each 
cost category. Thus, rather than having only a single preferred 
equation in each category, one has an opportunity to consider a 
number of equations and select the one most likely to capture the 
effects that are critical to a particular situation. There is also a 
potential for deriving an improved understanding of the nature of 
depot maintenance through an examination of the full set of 
equations. As a start, the following paragraphs summarize the 
results for each maintenance category. 

Airframes 

The variables found to be significantly* related to at least one 
of the relevant measures of cost (at the 5 percent level in at least 
one estimating relationship) include fleet flying hours, inventory, 
age, maximum load factor, empty weight, maximum takeoff weight, 
airframe manufacturing cost, the afterburner designator, sortie rate, 
and PDM policy. In some cases separate equations were developed for 
fighter/attack and boraber/cargo aircraft, r'^^lecting the influence of 
mission type. Factors found to be not significantly related to 
airframe rework cost are the operating climate, and speed and altitude 

'"'Subject to the satisfaction of other statistical criteria (e.g., 
exponent size, collinearity) and the ground rule that utilization and 
policy variables would be useful only as a supplement to size and 
performance variables. 




1 i 








85 


measures. A few interesting parameters were not tested because data 
were not available in sufficient quantity. These are landing rate, 
material composition, contractor identity, PDM interval, and the use of 
dock crews or specialists to perform PDMs. 

Results for airframe component repair are similar. The two most 
useful variables are airframe manufacturing cost and aircraft empty 
weight. They are about equally useful, which is not surprising since 
they are highly correlated. 

Table 23 summarizes these results. 

Engines 

At least one significant variable was found in each of the three 
variable classes (i.e., technical, size, or application variables) for 
each of the four depot-level engine cost categories: average time 
between overhaul (ATBO), average cost per overhaul, annual cost to 
repair, and annual component and accessory repair cost. Table 2A 
summarizes those explanatory variables which were found statistically 
significant* at the 5 percent level in one or more estimating 
relationships. 

As indicated, several variables show up consistently in each cost 
category. Unfortunately, most of the estimating relationships in which 
these variables appear are of relati^'^ely poor statistical quality and 
have unusually large exponents, thereby creating serious reservations 
about their utility. 

Avionics 

For annual avionics component repair cost, 14 explanatory 
variables were grouped according to whether they describe the size, 
performance/complexity, or application aspects of avionics suites. At 
least one significant variable was found in each of these three cla.sses. 
Table 25 summarizes those explanatory variables which were found 

*Subject to the satisfaction of other statistical criteria (e.g., 
exponent size, collinearity) and the ground rule that application 
variables would be useful only as supplements to size and performance 
variables. 







86 


■vSS* 




Table 23 

SUMMARY OF RESULTS FOR AIRFRAME VARIABLES 





Results 






Rework® 



Explanatory 

Total 

Cost per 

Cosi, per 

Production 

Component 

Variable 

Cost 

Aircraft 

Visit 

Quantity 

Repair 

SIZE 






Empty weight 
Maximum takeoff 


X 

X 


X 

weight 

X 

X 

X 


X 

Tr..HNICAL/ 

MKFGRMANCE 






^peed 

Altitude 






Dynamic pressure 
Load factor 

X 

X 

X 


X 

Airframe cost 

X 

X 

X 


X 

Afterburner 

X 


X 



Mission 

X 

X 

X 


X 

UTILIZATION 






Flying hours 

X 



X 

X 

Inventory 

X 



X 

X 

Age 

X 


X 


X 

Sorties 

Reserve 

X 



X 

X 

percent 

Climate 



X 

X 

X 

POLICY 






Organic mainte- 






nance pr-fcenc 



X 

X 

X 

PDM policy 

X 

X 



X 


See App. F for results for Total Cost, Cost per Visit, and 
Production Quantity. 


X = Significant at 5 percent level in one or more relationships. 








87 


Table 24 


SUMMARY OF RESULTS FOR ENGINE VARIABLES 






Annual 




Annual 

Component 



Cost per 

Cost per 

and Accessory 

Explanatory Variable 

ATBO 

Overhaul 

Repair 

Repair Cost 

TECHNICAL/PERFORMANCE 





Turbine inlet temp. 

X 

X 

X 

X 

Thrust-to-weight ratio 
Pressure term (psf) 

X 

X 

X 

X 

Specific fuel cousiamption 
Maximum mach number 


X 

X 

X 

Removal rate 

X 

NT 



Selling price 

X 

X 

X 

X 

SIZE 





Weight 

NT 

X 

X 

X 

Maximum thrust 

NT 

X 

X 

X 

Military thrust 

NT 

X 

X 

X 

APPLICATION 





Mission designator 

X 

NT 

X 

X 

Fighter/attack designator 


NT 



Single engine designator 

X 

NT 

X 

X 

Reserve/Guard fraction 

X 

NT 

X 

X 

MISCELLANEOUS 





Turbofan designator 
Manufacturer designator 
Type maintenance 

NT 



X 

indicator 

NT 

X 

X 



Notes; 

NT = Not tested for cost category because a priori rationale could 


not be established. 

X = Significant at 5 percent level in one or more relationships. 



















89 


APPLICATION OF ESTIMATING RELATIONSHIPS 

The equations in Sec. IV provide a number of useful approaches 
to estimating the total depot maintenance cost of a new aircraft. 

The following set of equations, first presented in Sec. I, should be 
applicable to a wide range of situations: 


Airframe Rework 

Engine Overhaul/ 
Repair 


Airframe Components 

Engine Components/ 
Accessories 

Avionics Components 


AFRWKC = 183.2 

n 793 0 390 

AVGCOH = 0.598 PRSTERM^ 

-ft 1 49 1 24 

ANNCTR = 2.72x10 PRSTERM WT ’ 

AFCCST = 0.7877 

ENGACC = 0.0265 PRSTERM°*^^^ 

AVCST = 0.00455 SUITE2°‘®^® FHKATE^-^^° 


The mnemonics used in these equation.' are as follows: 


AFCCST = annual airframe component repair cost per aircraft 
(1978 dollars) 

AFRWKC = annual airframe rework cost per aircraft (1978 dollars) 
ANNCTR = annual engine cost to repair (1978 dollars) 

AVCST = annual avionics repair cost per aircraft (1978 dollars) 
AVGCOH = average engine overhaul cost (1978 dollars) 

ENGACC = annual engine accessory and component cost per aircraft 
(1978 dollars) 

EW = aircraft empty weight (lbs) 

FHRATE - annual MBS flying hour rate 
PDM = PDM policy designator 
PRSTERM = engine pressure term (psf) 

SUITE2 = avionics suite procurement cost 
WT = engine weight (lbs) 


Table 26 shows the results obtained when these subsystem equa.ions 
are applied to five recent aircraft in the data base. The A-7D, 

B-52H, C“141A, and F-lllF are used here because they are the newest 






90 


attack, bomber, cargo, and fighter aircraft that are present in the 
data base in considerable numbers. The F-4D is included because it is 
somewhat more typical of fighters in general than is the considerably 
heavier F-lllF. Also shown are the results of applying to these same 
aircraft the total cost per aircraft equations listed in Table 22. At 
the bottom of Table 26 is shown a mean absolute relative deviation 
computed for each equati a from the five results shown in the table. 
This provides a measure of how well each equation would predict the 
depot maintenance costs of these recent aircraft, the aircraft most 
likely (of a?-.y in our data base) to be similar to the aircraft with 
which cost estimators will be dealing in the future. 

Each of the two estimating approaches is valid and useful. Some 
of the total cost equations have lower deviations than the results 
derived from this set of subsystem equations. On the other hand, the 

Table 26 

ALTERNATIVE TOTAL COST ESTIMATES PER AIRCRAFT 
(In $ thousand 1978) 




Subsystem 

Equation 


Total 

. Cost 

Equat 

j a: 

ion 


IIDS 

Data 

Base 

1 

2 

3 

4 

C, 

6 

A-7D 

145 

109 

108 

97 

108 

182 

116 

150 

B-52U 

551 

621 

589 

479 

643 

705 

560 

649 

C-141A 

317 

352 

398 

295 

317 

267 

255 

328 

F-4D 

129 

167 

105 

125 

115 

118 

157 

139 

F-lilF 

309 

328 

455 

379 

393 

417 

413 

388 

Mean absolute 
relative deviation 

0.25 

0.16 

0.16 

0.23 

0.19 

0.12 


a 


See Table 22. 







91 



subsystem equations provide more information about the makeup of the 
total cost and about subsystem-level cost drivers. There is no one 
approach, one estimating equation, or set of equations, that is best 
for all uses. The approach preferred in a given case will depend upon 
the objectives of the cost estimator in that particular case. 

POSSIBLE IMPROVEMENTS 

Several issues that arose during this study could not be dealt 
with completely. Some of these are important enough that they should 
be part of studies of depot maintenance costs that might be undertaken 
in the future. Some involve data limitations; others relate to the 
analysis itself. 

Data Issues 

Data limitations in this study prevented a full analysis of the 
effects of system age on the costs of airframe rework and engine 
overhaul. Such an analysis would have to be based on data covering 
at least several (and perhaps many) years of the operating life of a 
significant number of systems. There seems to be no standard source 
of data for either airframes or engines that could provide all of the 
data needed. The G098 system maintained at the San Antonio Air 
Logistics Center may have useful data for some airframes, starting in 
1971; but a recent status report indicates that all of the data files 
in G098 are not complete for all years. A good analysis of the 
effects of age might hinge on a data base that could be assembled 
from bits and pieces of data collected from a number of standard and 
nonstandard data sources. 

Use of WSCRS data has led to a data base that could not be 
quickly regenerated in the future if it were desirable to repeat the 
analysis with data from a later time period. The WSCRS programs have 
been modified in a number of ways since they were used to develop the 
raw data for this research. Some of the modifications were needed to 
accommodate changes in data elements that occurred as part of the 
changes in 1977 to a unified depot accounting system. The cost 












elements now used in reporting depot maintenance data are more 
detailed than those available to Rand for the 1975-1977 time period. 
This provides new opportunities to learn more about the nature of 
depot costs, but it will be necessary in future work to use somewhat 
different processing steps. A good time to update this analysis might 
be after WSCRS has become an official Air Force data system. Changes 
in format and processing steps may occur less frequently in an 
official data processing system than in a set of programs with no 
official status. 

Additional research could be done with the data collected for 
this project. The basic H036 data are especidly valuable because 
they contain some information that was not included in the WSCRS 
files. For example, data identifying individual facilities could be 
very useful in studies of maintenance concepts or investigations of 
indirect costs or the relationships between the composition (and 
cost) of the labor force and the nature of the work perfonned. 

The use of dock crews or functional specialists for PDMs is a 
policy that might be examined using H036 data. PDM costs could be 
accumulated by facility. If it were known which facilities use dock 
crews and which use specialists, the costs of PDMs for the two 
policies could be compared. 

Identification of facilities would also make it possible to 
identify the relative cost of organic and contract maintenance more 
accurately than was possible in this research, which examined this 
issue only in a limited way because of various limitations. A more 
thorough analysis would need both cost and production quantity data 
by type of facility. These data are available from H036. 

Having cost data by facility for similar types of work would be 
necessary in a study of the factors that determine the values of 
direct and indirect cost rates. Do skill levels or experience levels 
vary enough from one facility to another to result in differences in 
average direct labor rates? Do staffing differences between 
facilities result in different rates for operations overhead or 
general and administrative costs? Answering these and similar 









93 






questions would depend partly on having cost and quantity data by 
facility. 

Analysis Issues 

Analysis of the cost of avionics component repair would probably 
be more enlightening if conducted for individual subsystems or 
functions rather than for an entire avionics suite. The results of 
the analysis might not be useful at DSARC II, because data on 
subsystems are not available at that point. The improved 
understanding of avionics repair cost would likely be worth the 
effort, though, if only through its indirect influence on future 
decisions. A better understanding of what drives avionics repair 
cost would both help in decisionmaking and point out appropriate 
research topics for future needs. 

A major consideration in the application of our results is the 
significant technology changes taking place today in aircraft design. 
This issue could not be addressed fully within this study, but it 
should be considered by potential users of these equations. New 
technology shows up in new materials anu design practices that are 
introduced either to meet high performance requirements or to reduce 
maintenance demands. Airframes are being designed with increased 
reliance on composite materials, which have repair requirements 
considerably different from < lose of conventional materials. Some 
airframes are being designed with the expectation ^hat they will not be 
subject to a regular program of airframe rework. The trend in engine 
design is toward modularity, which allows individual modules to be sent 
to the depot instead of an entire engine. Avionics systems are making 
increased use of digital computers and built-in test equipment. Many 
people expect the ne. result of changes such as these to be a sizable 
reduction in the amount of depot maintenance required for future weapon 
systems. To the extent that such reductions are realized, the depot 
maintenance costs of future airc.-aft may be lower than our equations 
will predict. 










95 


Appendix A 

DEFINITIONS OF TERMS AND VARIABLES 

COMPONENT REPAIR CATEGORIES 

Five categories of components were identified for the purpose 
of processing cost data by category: airframe components, engine 

accessories and components, avionics components, armament components, ; 

and support equipment components. Data for each category were | 

identified through the use of a Federal Stock Class (FSC) or Group 

1 

and either a Work Breakdown Structure Code or Group Code. The 
contents of the categories are listed in Tables A.l through A.5. 

Work Breakdown Structure Codes and Group Codes are listed elsewhere 
in this appendix. FSCs are defined in Cataloging Handbook H2-1, 

F ederal Supply Classification , Part 1, Groups and Classes . 

COST ELEMENTS 

The cost of depot maintenance is the sura of the following 

individual cost elements. i 

( 

f 

o "Direct civilian labor cost" is the cost of civilian labor 

hours that are associated with a specific maintenance i 

objective. Included are civilian pay, the cost of leave 
time (annual leave, sick leave, etc.), and government 
contributions to employee benefit programs, 
o "Direct military labor cost" is the cost of military labor 
hours that are associated with a specific maintenance 
objective. The hourly rate is derived from annual composite 
rates furnished in DoD 7220.9-H, Accounting Guidance 
Handbook . These rates include basic pay, allowances, and 

certain government expenses such as Social Security taxes and i 

I 

reenlistment bonuses. j 


























"Other direct material cost" is the cost of material that is 
specifically required to carry out an authorized maintenance 
task and that loses its identity as a result of the 
maintenance task, either by becoming part of the item being 
repaired or by being consumed. This is the cost of direct 
material "other" than repairable components that are 
exchanged for serviceable items during the maintenance task. 
The cost of repairing these exchangeable components is 
potentially distributed throughout all seven cost elements. 
There are two contributions to "other direct costs": 
purchased services and travel. Wl*en a depot maintenance 
activity contracts out work incidental to maintenance that 
it is performing, the work done under that contract is 
considered a purchased service. The organization performing 
the purchased service may be either a government activity or 
a commercial firm. Travel and per diem expenses are direct 
costs when incurred in connection with work that v/ill be 
charged as direct labor. 

There are also two elements of indirect costs—costs not 
charged direct to job orders. "General and administrative 
costs" are the expenses of organizational units that do not 
perform direct maintenance tasks. The term "Other Direct 
Costs" applies to the overhead expenses of direct maintenance 
units. 

When a complete maintenance task is carried out for the Air 
Force by a commercial firm or by another military service, 
the cost of the task falls into the cost element labeled 
"contracted out depot maintenance cost." 










































98 





/wen 


r‘Ki55:S»^f J£.AWKi' ’H^lK.'VtZ 


Table A.3 

DEFINITION OF AVIONICS COMPONENTS 


FSC 

WBS 

Code 

Group Code 

FSC 

WBS Code 

Group Code 

I2xx 

not 

xx6 

not 

SU 

6605 

not xx6 

not SU 

5805 

tl 

tt 

It 

tt 

6610 

It It 

II It 

5810 

(t 

tt 

tt 

tl 

6615 

II tl 

tl tt 

5811 

It 

tt 

tl 

It 

6620 

II tt 

tl tl 

5815 

t» 

It 

tt 

tt 

6645 

It It 

tl tl 

5821 

tt 

It 

tt 

It 

6650 

It tl 

tl It 

5826 

tt 

tt 

It 

tt 

6660 

It It 

tl tl 

5831 

It 

It 

tt 

tt 

6680 

II It 

tt tl 

5835 

tt 

It 

tt 

tl 

6685 

It tl 

It tt 

5841 

tt 

tl 

tt 

tt 

6695 

II tt 

tl tt 

5850 

tt 

tt 

tt 

tt 

67xx 

It tl 

It tt 

5855 

tt 

It 

It 

tt 

59xx 

xx4 

AV 

5860 

tl 

•t 

tl 

tt 

61xx 

tt 

tt 

5865 

tt 

tt 

ft 

tt 

62xx 

tl 

tl 

5895 

tt 

tt 

It 

It 





Note: X = any character 


Table A.4 


DEFINITION OF ARMAMENT COMPONENTS 



FSC 


WBS Code 

Group Code 

lOxx 



not xx6 

not SU 

llxx 

(not 

1190) 

It tt 

11 It 

13xx 

(not 

1377,1398) 

It It 

tt tl 

14xx 

(not 

1450) 

tt tl 

It tl 

30xx 



xx5 

AR 

43xx 



It 

It 

48xx 



tl 

tl 

59xx 



tl 

It 

6lxx 



tt 

It 

62xx 



tl 

tl 


Note: X = any character 





















Table A.5 


DEFINITION OF SUPPORT EQUIPMENT 


FSC 

WBS Code 

Group Code 

FSC 

WBS Code 

Group Code 

1190 

any 

any 

44xx 

any 

any 

1390 

tf 

tt 

49xx 

tl 

tt 

1450 

»l 

It 

51xx 

If 

tt 

17xx 

tl 

tt 

52xx 

ft 

tt 

25>.x 

tt 

tt 

6625 

II 

t; 

26xx (not 2620) 

tf 

ft 

6630 

tt 

tt 

2805 

tt 

tt 

6635 

tt 

II 

2815 

tt 

tt 

6636 

II 

tl 

2835 

tt 

tt 

6640 

tl 

It 

2850 

tt 

tf 

6665 

tl 

It 

2895 

tt 

tt 

6670 

tt 

If 

2910 

tt 

tf 

69xx 

tt 

tl 

2920 

tt 

tt 

70xx 

tt 

tt 

2930 

tt 

tl 

74xx 

If 

tl 

2940 

tt 

tt 

81xx 

tt 

tt 

2990 

39xx 

tt 

ft 

ft 

tt 

Any 

xx6 

su 


Note: X = any character 


EXPLANATORY VARIABLES 


The following variables are the potential explanatory variables 
for which data w^- re collected and analyzed during this study. 


Airframe Rework and Airframe Component Repair 

Fleet Flying Hours: the number of flying hours accumulated during a 
year by aircraft of a particular MDS. 

Inventory: average number of aircraft of a partic’ilar MDS possessed 
by operating units of the Air Force. 

Sorties: the number of sorties flown in a year by aircraft of a 
particular MDS. 


Flying Hours per Aircraft: average annual flying hours per possessed 
aircraft (fleet flying hours divided by inventory). 






















100 


Sorties per Aircraft: average number of annual sorties per possessed 
aircraft (sorties divided by inventory). 

Empty Weight: the weight of an aircraft when crew, fuel, oil, 

armament, cargo, bombs, and di.sposable or special equipment 
are excluded (pounds). 

Maximum Takeoff Weight: maximum gross weight at takeoff for normal 
operating conditions (pounds). 

Maximum Speed: the highest speed obtainable in level flight at 
conditions most favorable to speed (knots). 

Typical Speed: this is the speed most characteristic of an aircraft's 
basic mission, e.g., average cruise speed for bombers and 
transports and combat speed for fighter and attack aircraft 
(knots). 

Typical Altitude: the altitude most characteristic of an aircraft's 
basic mission (feet). 

Dynamic Pressure at Maximum Speed: dynamic pressure evaluated at the 
aircraft's maximum speed and for the standard atmospheric 
density corresponding to the altitude for maximum speed. 

Dynamic Pressure at Typical Speed: dynamic pressure evaluated at the 
aircraft's typical speed and at the standard atmospheric 
density corresponding to its typical altitude (psf). 

Maximum Load Factor: the design load factor (g's). 

Airframe Manufacturing Cost: cumulative average cost of first 100 

units, including manufacturing labor and materials (millions 
of FY 1978 dollars). 

Reserve Percentage: the percentage of the inventory operated by units 
of the Air Force Reserve or Air National Guard. 

Climate Percentage: the percentage of the inventory operating from 
bases in humid climates. 

Organic Maintenance Percentage: the percentage of a cost that is 

associated with organic maintenance rather than contractor 
maintenance. 

Afterburner Designator: = 1 if aircraft does not have an afterburner, 
= 2 if aircraft does. 








101 


Contractor Designator: dummy variable to identify manufacturer of 
the aircraft: 


1 = Boeing 

2 = Cessna 

3 = Fairchild 

4 = General Dynamics 

5 = Lockheed 


6 = LTV 

7 = McDonnell Douglas 

8 = North American 

9 = Northrop 
10 = Republic 


Fighter/Attack Designator: ~ 1 for fighter and attack aircraft, 

= 0 for all others. 

Bomber/Cargo Designator: = 1 for bomber and cargo aircraft, 

= 0 for all others. 

Trainer Designator: = 1 for trainer aircraft, = 0 for all others. 

Complete PDM Designator: = 1 if aircraft has a PDM p?.-cgram, 

= 0 if it does not. 

No PDM Designator: = 1 if aircraft has no PDM program, = 0 if it does. 

PDM: as a variable, = 2 if aircraft has a PDM program, = 1 if it does 
not, = 0 if aircraft in sample is not a clear case of PDM 
program or no PDM program. 

Representative Series Select Code: = 1 for the MDS most representative 
of an M/D, = 0 for any other MDS. 

Age: aircraft average age, as measured and reported by AF/PAXRB (years). 


Engine Overhaul, Repair, and Component Repair Analyses 

Overhaul Depot: depot responsible for engine overhaul (1 = Oklahoma 
City; 2 = Sau Antonio). 

Source: AFLC Form 992 

Model Qualification Date: date engine passed nonrated 150-hour 

model qualification date (in months since January 1943). 
Source: Gray Book 

Turbine Inlet Temperature: maximum turbine inlet temperature (degrees 
Rankine). 

Source: Gray Book 

Thrust-to-Weight Ratio: ratio of maximum engine thrust to engine 
weight. 

Source: Current table entries 








102 


Pressure Term: product of the maximum dynamic pressure for the flight 
envelope and the engine's maximum design pressure ratio 
(Ib/sq ft). 

Source: N-1242-PA&E, Table 11. Based on discussion with 
J. R. Nelson, values for engine series were determined by 
adjusting model pressure term by ratio of series pressur •. 
ratio to model pressure ratio. 

Specific Fuel Consumption: specific fuel consumption for military 
power at sea level (Ibs/hr/lb). 

Source: Gray Book 

Maximum Mach #: maximum airplane Mach number at which the engine 
can operate. 

Source: Gray Book 

Removal Rate: rate of unscheduled engine removals requiring base 
maintenance and depot overhaul plus engines removed for 
periodic inspection per 1000 fleet engine operating hours. 

It does not include maximum-time engines removed for overhaul 
or engines removed for non-usage reasons (aircraft accident, 
modification, removal to facilitate other aircraft maintenance, 
removal for experimental purposes, etc.). 

Source: AFLC Form 992 

Unit 1000 Selling Price: selling price for 1000th unit in 1978 dollars. 
Source: N-1242-PA&E, Table 49 

Weight: dry weight of turbine engine (lbs). 

Source: Gray Book 

Maximum Thrust: maximum thrust that engine can generate at sea level 
(with afterburner if engine has one) (lbs). 

Military Thrust: maximum thrust that engine can generate at sea level 
at military power throttle position (lbs). 

Source: Gray Book 

Annual Sortie Rate per Engine: total annual MDS sorties divided by 

product of average possessed aircraft and munber of engines 
per aircraft. 

Source: HQ USAF/PAXRB 

Mission Designator: dummy variable distinguishing engines with 

boraber/cargo applications (=1) from those with fighter/ 
attack applications (=2). 

Source: WSCRS data 

Fighter/Attack Designator: dummy variable distinguishing engines on 

fighter/attack aircraft with air-to-air roles (=1) from those 
on fighter/attack aircraft with air-to-ground roles (=2). 
Source: R-2249-AF, Table A.l 















103 


Single Engine Designator: dummy variable distinguishing engines on 
aircraft with multiple engines (=1) from those on aircraft 
with only a single engine (=2). 

Source: WSCRS data 


Reserve/Guard Percentage: percentage of engine operating hours flown 
by reserve/guard personnel. 

Source: Based on data in PA 76-1 and 78-1 (Aerospace 
Vehicles and Flying Hours) 


Turbofan Designator: dummy variable distinguishing turbojets (=1) 
from turbofans (=2). 

Source: Engine nomenclature 


Manufacturer Designator: dummy variable distinguishing General 
Electric (=1) from Pratt & Whitney (=2) engines. 
Source: Engine nomenclature 


Type Mainteriduce Designator: dummy variable distinguishing organization 
performing overhaul (1 = depot, 2 = contractor). 

Source: WSCRS data 


Maximum Time Between Overhaul: the maximum time (in operating hours) 

an engine may be retained in service without a major overhaul. 
Source: AFLC Form 992 


Average Time Between Overhaul: the average age (in operating hours) 
of all premature and maximum-time engine removals requiring 
major overhaul. 

Source: AFLC Form 992 


Installed Engines: serviceable engines physically mounted on an 
aircraft at end of fiscal year. 

Source: AFLC Form 992 


Annual Flight Hours per Engine: operating hours flown per engine per 
year. 

Source: AFLC Form 992 


Avionics Component Analysis 


Suite Weight: weight of electronics group equipment (excluded 
installation weight) (lbs). 

Source: Rand data 


Aircraft First Flight Date: first flight date of aircraft series (in 
months since January 1943). 

Source: Green Book 


jtwi n "iTiT 

. —■-n.m ^- 


igiaTwri.yibi)irtiLjMwi»i 

.A- 



5 i\ 












Number of Black Boxes: number of electronic components, or units, 

usually identifiable by AN designation, which are normally 
considered part of aircraft's avionics subsystem. 

Source: Green Book 

Number of Functions: number of electronics functions performed by 
aircraft's avionics subsystem. Functions counted are: 


1. 

controls/displays/instrumentation 

2. 

communication/identification 

3. 

navigation 

4. 

bomb navigation/fire control 

5. 

reconnaissance 

6. 

ECM 

Source: 

Green Book 


Suite Procurement Cost #1: procurement cost of avionics suite at unit 
100 ($78). 

Source: RM-4851-PR, other Rand data 

Suite Procurement Cost #2: sum of current D041 item procurement costs 
for all items which are assignable to the avionics subsystem. 
Source: Data base established for R-ii52-PA&E 

Mean Time Between OFM Demands: mean time between avionics subsystem 
demands placed on base-level organizational and field 
maintenance. 

Source: Data base established for R-2552-PA&E 

Annual Flying Hours per Aircraft: total annual MDS flying hours 
divided by average possessed aircraft. 

Source: WSCRS 

Annual Sorties per Aircraft: total annual MDS sorties divided by 
average possessed aircraft. 

Source: HQ USAF/PAXRB 

Peculiar Percent Based on Dollars: percentage of total avionics spares 
investment managed by flying hours material program which is 
unique to MDS (%). 

Source: R~2552-PA&E, Appendix B 

Peculiar Percent Based on Item Count: percentage of total number of 
recoverable avionics items managed by flying hours material 
program which is unique to MDS (%). 

Source: R-2552-PA&E, Appendix B 

Combat Dummy Variable: dummy variable distinguishing noncombat 
aircraft (=1) from combat aircraft (=2). 

Source: Nomenclature 


105 


All-Weather Dununy Variable; dummy variable distinguishing aircraft 
with all-weather capability (=2) from others (=1). 

Source: Rand data 

Mission Group Dummy Variables: duiuuy variables distinguishing aircraft 
mission types. 

Source: Nomenclature 


GROUP CODES 

Group Codes were used in the DMIF Cost Accounting/Production 
Report prior to FY 1977. Headquarters, Air Force Logistics Command 
provided the following code definitions: 


AF - Airframe Repair 

AA - Aircraft Accessory Repair 

EA - Engine Accessory Repair 

EO - Engine Overhaul 

AV - Avionics Repair 

AR - Armament Repair 

SU - Support Equipment Repair 


WO RK BREAKDOWN STRUCTURE CODE 

The work breakdown structure came into use v’ith the unified cost 
accounting system in FY 1977. The following codes relate to aircraft 
or general use:* 


All - Aircraft, Fighters, Airframe 

A12 - Aircraft, Fighters, Engine 

A13 - Aircraft, Fighters, Aircraft and Engine Accessories and 

Components 

Al4 - Aircraft, Fighters, Electronics and Communications Equipment 

A15 - Aircraft, Fighters, Armament 

Al6 - Aircraft, Fighters, Support Equipment 

A17 - Aircraft, Fighters, Other 

* Office of the Deputy Assistant Secretary of Defense (Management 
Systems), Department of Defense Depot Maintenance and Maintenance Support 
C ost Accounting and Production Reporting Handbook , DoD 7220.29-H, 

October 21, 1975, Appendix D. 








106 






A2** - Aircraft, Bombers* 

A3** - Aircraft, Transport 

A4** - Aircraft, Trainers 

A5** - Aircraft, Utility 

KXX -■ General Purpose Equipment 

Lll - All Items Not Identified to Another Category 

Work Performance Categories 

This study addressed costs charged against eight Work 
Performance Categories. As defined by DoD,** these are: 

Code A — Overhaul . The disassembly, test, and inspection of the 
operating components and the basic structure to determine and accomplish 
the necessary repair, rebuild, replacement and servicing required to 
obtain the desired performanc i. It is considered to be synonymous 
with the terms "rework" and "rebuild." 

Code B — Progressive Maintenance . A predetermined amount of work 
that represents a partial overhaul under a program that permits the 
complete overhaul to be accomplished during two or more time periods. 

It is considered synonymous with the texmis "cycle maintenance," 
"restricted availability,' "preventive servicing," or "recondition." 

Code C — Conversion . The alteration of the basic characteristics 
of an item to such an extent as to change the mission, performance, or 
capability. 

Code G — Analytical Rework . The disassembly, test, and inspection 
of end items, assemblies, or subassemblies to determine and accomplish 
the necessary rework, rebuilding, replacement, or modification required. 
It includes the technical analysis of the findings and determination 
of maintenance criteria. Includes prototype teardown, analysis, and 
rework of an item to determine job and material specifications on a 
future workload. 

Code H — Modification . The alteration or change of the physical 
makeup of a weapon/support system, subsystem, component, or part in 
accordance with approved technical direction. 

Code I — Repair . Action taken to restore to a serviceable 
condition an item rendered unserviceable by wear, failure, or damage. 

* XX = Any third character. 

DoD 7220.29-H, pp. E-1 and E-2. 







107 


Code J — Inspection and Test . The examination and testing 
required to determine the condition or proper functioning as related 
to the applicable specifications. 

Code K — Manufacture . The fabrication of an item by application 
of labor and/or machines to material. 


Other categories exist, but they identify work that either is maintenance 
support, as opposed to maintenance proper, or is most likely to be 
connected with systems that are not fully operational, e.g., storage 
or reclamation work. 











108 


Appendix B 
DATA PROCESSING 

AIRFRAME A ND ENGINE DATA 

The main effort in development of the airframe rework data base 
was the fairly straightforward aggregation of data contained in WSCRS 
files. WSCRS data records contained MDS, WAC, individual cost 
elements, production quantity, flying hours, and inventory-month data 
for a fiscal year. Processing consisted of the following steps for 
each MDS: 

1. Drop records for WACs other than those of interest. 

2. Sum each cost element across records to compute a subtotal 
by year, MDS, and WAC. 

3. Convert each cost element to FY 1978 dollars. 

4. S’un the individual elements to compute i total cost by year. 

5. Average the cost, production quantity, flying hour, and 
inventory-month data, obtaining a single value for each 
variable representing the three years of raw data. 

6. Divide the inventory-month variable by 12 to compute the 
average inventory size. 

7. Explanatory variable data were obtained from several 
sources and input manually. 

Cost data for engine overhaul and repair were also obtained 
directly from WSCRS, with similar processing. The main difference is 
that engine costs are associated with the engines themselves rather 
than with the systems in which the engines are used. 

COMPONENT REPAIR 

Development of the component repair data base was considerably 
more complex than the processing of the rework and overhaul data. 









109 


The main reason is the necessity of associating data that are 
recorded by component stock number with the appropriate weapon system 
(or weapon systems). WSCRS was the primary source of component 
repair cost data. WSCRS data files provided by the Air Force covered 
costs for components one indenture below an end item and costs 
associated with only a Federal Stock Class instead of a complete 
stock number. These WSCRS data had two features that were important 
to this study. One was that they identified first-indenture 
components with the appropriate MDSs and engines. The other was that 
the lower-indenture components that were not included account, at 
least for some aircraft, for a significant fraction of the cost of 
component repair. 

Because of this, tapes containing copies of the Air Force's 
depot maintenance cost accounting data were requested from AFLC. 

These tapes contained data on all depot maintenance costs for the 
years of interest to this study. From this data we developed the 
needed cost information for components below the first indenture. 
These costs were then included with the cost data taken from WSCRS to 
produce a complete data base. 

Three files with different formats were the basic sources of 
cost data: 

1. WSCRS file of first-indenture component costs. 

2. WSCRS file of costs identified by Federal Stock Class. 

3. Accounting system data (H036B). 

In addition, data specifying the indenture structure (rela :ing 
SRUs to LRUs) were obtained from supply system records (D041). The 
following steps were used to collect all of this information into a 
single complete data base: 

1. Use D041 data to define the complete list of stock numbers 
used on each MBS and engine, computing the number of each 
component used on each end item (quantity per application). 






110 


2. For each stock number applicable to end items used in the 
data base, develop from H036B cost records in the same 
format as the WSCRS records. Add to these records the 
quantity per application data. 

3. Merge the cost records from H036B with the WSCRS files. 

4. For the new records, compute and save an allocation factor 
based on total component operating hours and operating 
hours on each end item. 

5. Process data as described above for airframe and engine data. 

6. Separate the file into subfiles for airframe components, 
engine components and accessories, avionics components, 
armament components, and support equipment components. The 
differentiation is based on Federal Stock Class, work 
breakdown structure, and Group Code as described in App. A. 

7. Add to each subfile the explanatory variable data obtained 
from separate sources. 










Ill 


Appendix C 

DEPOT MAINTENANCE COST DATA 


Table Title 

C.l Airframe Rework Total Cost Data 

C.2 Elements of Airframe Rework Cost Data 

C.3 Elements of Airframe Component Repair Cost Data 

C.4 Armament Component Repair Costs 

C.5 Elements of Engine Overhaul Cost Data 

C.6 Elements of Engine Repair Cost Data 

C.7 Elements of Engine Component and Accessory Repair 

Cost Data 

C.8 Elements of Avionics Component Repair Cost Data 

C.9 Depot Maintenance Cost Comparison with OSCAR Data 


The cost data used in this study are presented in Tables C.l 
through C.8, which also show the magnitudes of the individual cost 
elements associated with each category of depot maintenance activity. 
When the costs of all activities are combined to give a total annual 
depot maintenance cost by weapon system, the results are the costs 
shown in the left-hand column of Table C.9. The other coliunn in this 
table is a set of corresponding costs taken from output for 1977 from 
the Air Force's Operating and Support Cost Analysis Report (OSCAR). 
Differences between the two sets of da; a are due to differences in 
the allocation of common component repair costs as well as to the use 
of a three-year data base in this study, as opposed to the OSCAR annual 
data base. 







112 


Table C.l 

AIRFRAME REWORK TOTAL COST DATA 


MDS 

Total Cost 

Cost per Aircraft 

Cost per 
Depot Visit 

A007D 

4777745 

13090 

51932 

AOlOA 

2717 

94 

- 

A037 

1237538 

10952 

4139 

B052D 

3010723 

37828 

143368 

B052G 

39721537 

245195 

630501 

B052H 

20551246 

230913 

587178 

COOSA 

26469470 

407222 

715391 

C130E 

10633790 

37843 

98461 

C141A 

24825940 

100105 

206883 

F004C 

15254006 

56496 

98413 

F004D 

20193799 

45482 

87419 

F004E 

28506099 

47990 

98980 

F005B 

33003 

3667 

16502 

F005E 

915307 

17947 

17602 

F015A 

713143 

8592 

4542 

FIOIB 

336762 

3007 

56127 

FIOSB 

580444 

17072 

5635 

F105D 

2502250 

25275 

16907 

F105F 

585772 

30830 

24407 

F105G 

2120858 

50497 

151490 

F106A 

9726981 

55583 

127987 

F106B 

2161466 

58418 

39299 

FlllA 

473711 

5094 

157904 

FlllD 

765696 

9115 

85077 

FlllE 

820033 

10380 

410017 

FlllF 

235850 

2775 

117925 

T033A 

/09208 

3138 

8059 

T037B 

1044856 

1648 

87071 

T038A 

2606442 

2915 

6260 

T039A 

795601 

7207 

98200 

FBI11A 

208600 

3161 

34767 

KC135 

16937623 

25938 

109275 

OVOlOA 

473211 

5439 

0 

RF004C 

15600781 

45089 

73243 

TF015A 

232582 

10572 

11075 


NOTE: All costs in FY 1978 dollars. 

















nsisA Bi 














i 
















. ..s-v £jJf 











All values arc averages for 1975-1977 time period and are on 



















































Q^'<foor«-.fOQO'Oovr««»co»‘^'^iAP*^osOmr«*.ocovocsc'j*^r>Or«4irjNOvocomOoOi-H 

*^^m^4CNC\OsoiriOcou^oo\r-(cosO'Or^vOfMC'ifOOmoNcovi5aNsfr>ooo«-<c'’>rH 

r>.•^J‘C>Joor^oooococ^lc^^^^cooo^^<^»fOOr^vOO^'NOC^»-^coO<Nl^>o^^a■fOi^lO■*o^^ 

0^‘A<^0^^0OO^O<*^00^H0^C0^DC0r^l-^C^IOC^0^0^^^'.0^A^^\0*^0^«4•^0f0f0»•^lA 
IH O\f-lNO00«^C0f0t’^CSlCsJfH C'JCOrOm<J‘vDC'JO\r«.CMi-H CsJfOC'JfHiT'fH 

i-Hr-4 rHrH r-i 


vO^*HOOvOOOvOOO*4‘r*»«4“0\ >AOOrJfOO'i-<-^r>^fn<T\oOh**COC'4fOr^OvOmoO>«TC>4 


^sOsOo0^fN.<ffooOs3-cocvjf^a\mOfH<}''^rx.i/^co*^'OcSfooomvr)iAfHr>..vO«s 

vovOsr\ONOo\<ro\-»f^tr>\ovors.cc‘c^eMrH'N3’a»voOf>*a4iAcorHooOmc'ji/^4ntHm 

ci«HvooO'^oo<^c^*^o\‘<j-ONOmo\focMONir>r^m<scoo\fO'^*^«-4T-4mOOcocovo 

0000'OOO^|«-4Ow^r^0Nf*^OCvlr*^CVJ\fi-^CS4CS*^«itf».TvD\Dr^irjC00Nr^v0iH«—CCMlA 
Q4nrN.ONfvii-<<fC'icorooNOvomooONi/')«^^oo»-4r««.oaHON3'f-(iHr^oONOcn'^*^ 
C»4»H'^OOONfOOO 'OCOO•^CvJCO^DCOOcgO^CO^Hr^rH^*.O^O^v^O^O^vOC^>3‘^-^rHCn 


00 00 «« 00 

^ w 


0 0*^1^ 
C^i 1-4 fH 


r^C^i-«3■O^•*.3■O^O^l-lCSsOC^^00^A1-^r>. 


o^oo*<rf-'0^tAooc^o«3‘04ntA*d’ooc>jiriONCOf-icoc>icMcgo\r-icocn<ro\tnr<-tcNin 

i/^r^o\r^<nooo4r^c^osNOO\omf*^ooi^f*^oo»-4r^»4“0»/^oovOvfiir»oo«3''AcM*-iONO> 

^©OONa\<<J‘CO'Or-««-f\OOOf'^00>r>COf-400eMrHOOOO'^r^ONr^'.a-iHvDCO*0“fMOOrxlA 
fs,o\f*i*^f*>.inoo CNicOf^ooc'40iHONcocNJcsi\^cooo<r**3‘r**oO'^ir»f^*<j-Cootaoo 
00 <Ns;fO<»^oo »A*a‘%coo H*d‘C'JOcOH^DcMirifOir>r>‘Ocna\fH 


0> <^lAOf00^tr^c0v^OO00v0^^r^^00^r-^0^^«»-J■»OOO•.yc^^^OC0C^O00^«•^-^lA 

*J U^'0^'OCOxOCOO<riOO’sI\D'3‘\DiHCOO\tH.I>I^r^iHOiH»^cn«^CSimtnrHCMC*JfH 

O 0\<^0•<^■<^•^0<OvOcOf^COC'400^0^^r-.tOc^^cn^-<r<^4C^JCOOcOl^^O'OsJ■OOl-^ 

U i) u 

e) ^ « I •<tOhs»-4f>»c<^f^u^ON\f»oomOcoro»sr‘<j‘Or>»ONsO'rHf'>^0\ONr^i-ir».ooooa'COino‘. 

^*rtO| CO ^ OOcntA^ ^-»-»t^C^CO^-^O^l^>^«‘'O^A•«3'^H^r>O^C^^CM^^‘r^vOr*H\Of^OvOr>.f^ 

«*ocj «4‘ 0 \ 0 »AO cors.f>.r^ fH ir*. cvjcocvj<rfOCMO\fOtHcvj«^r«scn\oON»ncNjoo 

OC .. 


cn f-H 1-4 CM CSJ CM 


<d «H 
k TJ c 

s §-a 

5 ^ 


00‘^'^00r^»^Or^C0'O?xtAOCMC0r^rHHO'vDf-4OvDC0'^v^<f00NvDC0CMCM^Dv3- 

HOO<'<Mf-4^000«-4COOOHCO^Dcn•^^s.•^C^vO^OCOCM^Olr^CMOOlAO^COrHr>«.CMt^tOO 

U-JO^s^-4^Ar^^-tCOCOC\00^^^0■<^•3“CO^A^^^y^>»f^^Or>COOCMO^O^<^|l-^CMCMO^O\r«• 

cn^'^corN.ocn ov<Jsoms5'r*«.cocoooiAoocMHr«*ooOcocovOMtr%vOONJA.-4co 
H f^Hooi-4cno»or*>^0'<r»AMp^Kfo<Min»-40 

A.I mm4 (\t e\l ef\ aM JMK _■ 


O .U t 
OJ 0) (0 
JZ u p 
U *H O 
O O I 


CM •Hr4corN.<rir>Oco\C<-<aNOCMfOtH 

H CO ^ 


»-ICMOf-4^Cr-1<MCMCMCMv01HCOOCM»HfOOO 
CO cn r-4 CM r-J 


U C3 

V4 U •H 4J 

01 0) (4 

j: wi 0 ) o 

4J •H 4J U 

o Q <g 


I COCMO•^CMrs,^000'^t*«COC'l^>OOO^^^^CO'^OOlO^^^-^I—^^Or>•CM^OvOCOCO^HlOO^CO 

»H®*~*OCN<rr>.NOoocM»HvOco»Hf-4r*-r^osf>»r^-.^00'iAfor>.co*^ir>cor^*-4'OcoO' 

^«»\OC^l^^^OOO^^O^*»tOO*N^OOO'OCMl^^CM^OO^CO'DvOt^^OOOO^O^-4vDCT'^ACO'00^^ 

*^»H*.3'tnir»r>»iA«'yoNcococo«^««^ir>*^rHr*ssOiAror^oocMr*sX^ 

0> OO'Of^'3’ «HsDfO«-4COT>COr^CMmvOOCM\£>CMir>vOCOiHr«« 

CMVO^CO mCMCO'^ tH 'T-HCO'^rcOfO rH CMsj- 


u & 

O (0 V4 0) 
0) .tj O 44 
n -H jO S 

•H H « o 
p :H hJ X 

X 


CM»*H<^r^»C^'HO>CMCOfHxOCM>»OcO^MrvO««3-CO'OrHCOr««.'0«^OOt-»rHcO»HON»rj-<T»H 
ir><Mr^rHr«N'.a'C0 cooocMir»r*«.covO»HC\o«-^’«.3'»;rcooNr«-iocMcOf-4*<r'^oosOf-4tr> 
VO r-sovo-^ intnO'CO rHCMlACMCOCMiriOOsrr^COiHiHOOCMCMOrHCO 

f-fCMCMf-lCM-4»HCM rHCMCMCMCM CMCMCv? 


>v 

u u 

O CO Wi 
0) 4J O 4J 
44 •H ^ (4 
•H tH (0 O 

P ••{ P u 


<^^r^O'0>Mru'Of*^c^cMC'r^tnr^«^OoOrH-TvOir»o>coc^»-ir*«.ONirjf>.c\cooo 
fs. 'OCM'^r- cocMvfiH cn ■Of^^^JO'Cl^»^Oco^o*^r<M1~•sD^-4cocoo^O^^ 

00 ON«-4-*3“r*' O'i'ooco -<rcMOcMrocMCMfOt-«sors.f^^cor)»n<y'Hoo 



C 



vO 

r-. 

xn 


OX CO ^ 00 CM 

>«• «n 

vO 

CO 

CO 

xO 

lO 

»n 

ox 

CM 

rH O 

CO CO 

O 





o 






u 

CO 









O CO 

00 

m 

o 

CO CO 

CM 

MT 

O 

CO 

MT 

r>> p 

CO XO 

CM xO 








u 

•H 

44 

(4 

CX 



*n to O 

•-I ox 

xO 

Cn. 

CM 

OX 

xO 

CM 

CM 

O 

MT 

lA 

ox mT 

\0 

ox 

00 WO xO 

-H 

CM 

r>. 

CM 

Ox 

CO CO 

V4 

•H 

.o 

3 

m 

CM 

00 

cn 

C>» 00 

\o 

r«-. 

CO r«. 

xO 

00 

CO 

OX 

CM 

1-4 

to 

o 

CM 

Ox 








lA 


o 

o 


•H 

?» 

(0 

0 

tn 




^ OX 



Mr 

«o 


p 


xO 

CO 

Ox 

CO 

to 

CO CM 

<t 

CO 




o 

o 




P 

•i4 

U 

p 

X 

i 







CO '3’ 

«o 









CO 

iH CM 

CM 

CM 

CM 




CM 

wo 


wo 





i 

! 00 

ox 

<»■ 

o 

00 00 

CM xn 

m 

XO xO 

Ox 

XO 

ox 

»A 

CM 

CM 

lA 

CM 

O 


OX \0 

O 

o 





xO 






q 



; o 






u*> 

CO iH 

o 

ox 

rv 

tH 

CO 

CO 

IM 

WO 

CM 

CM 

CM rH 


o 


C^ 








4J 

53 



' CM 




«-4 r*» 

•-4 CO 


O 

uo 

1-4 

CM 

vO 

CO 

rH 

OX 

CO 


CO 

CM fM, 

CO 

p 

P >0 

CO 

OX 

CM 

r*. 

00 

Ox 

wo 

CO 

0) 

a4 

o 

u 1 

1 


CM 

4 

»-4 OO 

r-4 vO 

Ox CM O* 

MT 

CO 

CO O MT 

xO 

CO xO 

Ox 


O CO 




o 







44 >H 

p 

(4 ! 

! CM 





CO 

m MT 

UO 

Ox ^5- 

1-H 

Cs. 

CO 

MT 

CM 

xO 

'O 

CM xC 



lA d 






CO 



•H 

;> 

<0 

H 1 

1 ^ 






o 

•O ox 

tM 



iH xO 

CO 

O 

CO 

lA 

I'— 

••S' MT 


vO 

XO lA 

C7X CO 

CO 

CM 

CM 

CO 



P 

•H 

p 

P I 

1 ** 

















0, 














U 


1 

1 





CO 

r** 

CO Mr 

o 






*-5 



CO 

P CM 

CM 

CM 

CM 


rH 


CM 

WO 


xO 



P!55!^^^‘^‘^^'^^OWPWir»OOOOOOOHrHfH»H 

«-M CO m aTv «n LA mmi ^ _ 1—1 _■_! _ 1 _ 1 


(0(:QQ(t4 0<P<P'Jtki ^in< < 

<»-«»r»v^ininNO\Or^fHi-4LH<fQ<<HcOOOm 
mOOOOOOOHrHrH*HCOrs.COONHrHM'*5 H 
rHf-1e^iHr-4fHrHrHf-lA4^rSCOCOfOC*^ I | | | | 


i P5L;^^‘^<<^'^^OWPWir»OOOOOOOHrHiH»HCOr^COONHrHM^ H 
ir^<HCp(n*n*^in*H-H«3’'4'^IAlC^r-(f-iA<«Hr-(rH«Hr>4r-44-4f-lfHr>COCnc<)| | | | \ 

ll}llll|l||||||i|II|:|i|llllllcorj^Ct4ts4 


ta-UAint^A 

















(In $ thousand 1978) 


MDS 

Current 
Data Base 

OSCAR 

Data for 1977 

A-7D 

145 

156 

B-52D 

320 

664 

B-52G 

535 

422 

B-52H 

551 

623 

C-5A 

1109 

965 

C-X4L4 

317 

300 

F-4C 

146 

165 

F-4D 

129 

166 

F-106A 

201 

266 

F-106B 

279 

260 

F-llU 

229 

414 

F-lllE 

280 

370 

F-lllF 

309 

269 

T-33A 

15 

15 

T-37B 

11 

7 

T-38A 

20 

24 

T-39A 

53 

57 

KC-135 

1.05 

92 

RF-4C 

153 

172 










Appendix D 


EXPLANATORY VARIABLE DATA 




Table D.l 







123 


'O 

01 

s 

ti 


a 

o 

o 

I 

I 


H 


e « 
ti ru 
60 


« u 

C -w 
O (Q 


M I 

a c 

O U f 

e u 

<9 <9 > 

Vo 

tM 3 iJ V 

Vo c 0) a 
•H 5 O h 
CE <J >. 


M 3 a 
g W >%'3 
CO M H V 
c a> 4) 
>« V4 c. 
Q Pk CO V} 


g Cm 

2 ” 

o 6 Q 

o V4 *H w 
•H 3 X 

g « eg *3 
<0 <0 S a* 


C V 


4 ) 


fO -H ft) 

a. w ft) 

>*•-« CM 

H < ^ 


OOC'JOCcC\«HONcOCM<»%iA'»OOC4t/>cn<^CMc3t^w»wwwwi'»i-ik-^*-iw-ww»-ii-i 
OcO OCT'OcONv03\f^cOOOCO»Ar>. r^OcOcO\OOOp vOCMCcipr-ipOOr^ 


ocoooor^iHCNjfnONDor^cn 

^ ^ ^ 

rH 


<n 

0\ m 

H <n 


\OcCcDOOr^C4<SC4 
iCOfncOC^CcO^^CMC^I I 
c^<^cc<nrv«-fcofiooo 

fH •tf W 


r^O\0\avO\<^(*^COOOCOCO 
I cOr<HHf-4fHOOO\OcO\Oc 
s3>rvfHr«»r%mirioooo 

^ iH fH rH 


O r- CM 

I CO >3 I CM I CM I 

fH tH CO CO 


OcoO'»'»oocnu^«>«nw%«ncoc^<noo»N.r>.r>»rN,oocoor>c*^f*>r**cooooo»Ac*> 

rvr«%vOCMcncMCM'N<Moooooor«‘f>‘f>-s£>cococooor«‘'3r««r^r>r>.h*»3r«*rvcoeMoocorN. 


cc>a\<-MiHmcM<noccMCM<M-^OiHr^r>>cooccMCM-^»9^9cac9cacocr^rx.cO'4’>^r-4fHr^ 

iAr>>incn'^'^>0'^(n'>7«4’«ncooovroocor«ror«>rNmr>(omiHf-iHor»rN.oc09v 

0>'4’cnCMCMCMCM<-4M4000Vv-|<MO%rHOCrN.inir><-<i-i-4<-<fHfH(nrMrMCMrHC4 CMO> 
CMCMrM ft ft ft ft w-t CMCMCMCM ^*0 


cno><Mcnr>«^iH^'.^00'3000-^r«»0'«HOciAW*ymOOcO'3cOC'»HCOCMfHOOO*'J 

•-4r>>>0'>^fnr»r>Or><cocOcooiO'^»9^cMCMaccMOcocinococ>H«H9vr-4oorM<Hr«>0'^ 

o>cocnin‘C-»cncM<o>T«^»nmr>»fno'>»''»»Amcococor«^p*»r^co»^cO'^c^>TH»ncn 


oooooooocnoooooooooooooooooooooooooo 
OOOOOCOOOr^OOOOOOOOOOOW^OOOOOOOOOOOCmO 
iHrM»HCMr*»a'OCMCO*-<f-l«-IOOOO»-iiHrHrHc3c3r-ii^iHrMOOOOCMOOOO 
cnt-trMocMioc ooorM oo otoocoeniAmoo 

ncncncMcMcn cr>ir» -^cM-^cncMcn 


Q< QOX’<M«£<^aWGQ|x3<pDnQt^U<(0<;Qt<tfa<CA<<<< <(J< 
r«>O(^c>icMCMtnow'9'>7«»(A4ntn<HminirtincOcOfHi-4f-4iHcnr<^00ON<Hmo<rm 
OiHc0(r)inmp<n*9OOOOOiHOOOOOOOiHt-(«-(»H(nmcncnf-icnr-(OfH 
0000000<HfHOOOOOO<H<H>H«HfHr-t«-l«HfHiHrMOOOO*HfHOOO 
<<<aa3aC.;UOS>«&up4ti«Pu&uU«Pu(hP«tt4U4tio(x*(>*(htt4HHHH{QU>(><U4 

C ^ O fiO H 












ENGINE EXPLANATORY VARIABLE VALUES 









li’it r .t^-r::^ ; ■■*''■• ■■-? 


125 


OJUL‘UOJU|t»j 


^uuu'. |•^.l(l JdJii]3VjnurK 
, jojvu^<lb 3 <i ucjoqjni 


! SI 


(oJUJiAV ^ 61 - 5 ^ 61 ) 
J()L pjl«‘f>*)/»A 4 »S 3 y 

..joivuJjbOd .>ujauj JI^UIS 

101VUoI&^ )(3C)1V/43)q){li 

.JojvuSjsoa UOJSSJK 


h"i 


(o^kjSAB 

3 ' N< 3 |]J 0 S lenuu-/ 


IblUia X 4 P 3 H 1 K 


i 


(S<ll) 

I5.fili«i «rumxwK 


v«qi) aM 3 i*'*M 


(«- $) 

ouOl lj«/l 


(oawjjAU U-Si 61 ) 
939)1 tCA 0 « 9 )i 


1139 ^ U'WTXBK 




2S® 88 


o -AO 


'>O*««««OCCts>< 50 &CCS OO 
m rs. ^ 9- 




O O O d O 
o o o o 


S o OOOAOO^OOOOOOOQOO qo 

O O0O^'^O0O<~«OA0>A^/^.Al*0 oo 

(> 4 ^^ ^ 4 O<^O'^'-* 0 'CD<^'"CDAr^f^-^A<A 


O O A o c 
O 0 < O O O' 
^ O' CO ^ ^ 


> 0 <s>r«'OsC ^O' 


OOOOO oo O^A>AOOOOOOOOOOOOO oo 

oSoOO oo 00*^AJOOONAf>4«/%OOA^OO $o 

C^O* <»*OA40*A«AOBO«0^*-COOAC'^<0«0 


.-I O ^ p% f» «4 . 


O O CO A 06 
00 O' O <» r*i 


OOOOO 
f* ^ 

00 0 < <^ •« C'O 


0*A 00'>r''>«AOA>A>«r«0'*c^*'*O^A AO 

r- <®•^^►.w^C0^^«^O«CAfSlr»o^O p»a 

S 04 <'A<940«'^flOO''OjOAA«l<COMOO'9 

✓i «» t* » 1 . 4 ^r%PN »»•» 


O A • 
«» 04 r« r 
A «» >9 « 


«} A O 

OO A A 


HOdVW^UiN! QSOailM^li 


C 4 a 


Pt A 

>4 '4 


«aMP4MA'4r'<04A«A^»944A*M 
^OflUNOOOOOCOAV. NAfs© 


O O 

o c 


(qW4q/<iO 

U0|3<jBf’«U03 ICl>i ii;?59is 


OiOx 9 jn 9991 X 


0 | 39)1 

3 nS|a.-n. 03 » 3 sn 4 Hi 


i I 


(«♦) 9in3C40<jMi 

301UI au^qjni 


a96l ^Jenucf 
aauiti bM 3 uoH) X(>v> 


^ 3 oJo(i oimsuodsay 


O A A A A 
*9 A C* A A 
A 00 A «> 00 


A O 

A 06 

d d 


AOOOOOOOOOOOOO'OOO 


O 4 © »• ' 
A© —d f 


A M © © © 


AOA4('4r4-^<4AA^©AC09>OOA 

A 9 'OA»coaow©©o>e©« 0 'eA 

d©dA©dooAM©A©©©©© 


A © 
A A 

d d 


S O' A A © 
4 A «9 A 

A A A o A 

© d d © d 


;SS 

s O <« 

> © d 


K A e 6 . 

3 > 


Hill 


gg gsggassgsssgsssss ss 

MM MAA^AAOO'AAAIstACOACBP* 9 •* 


04 O 00 A ^ ,6 0 


90©0'^AP44A9> 


gggSi 

^ A A A O 
C^ sO ^ CO A 


A © © A M 
P4 A A A A 


4AA>O400AAl<^<^AA^ 

0444O^aA^'4AA«9 


^ O' A M 


OOOOO 


® s 

o o 


__ _ ^AAnOP 

SSSmAAAAAAP 


■2S 


O O O A A 
<0 A •■• A A 
O A C6 O 


3 A O 4 A 


4 O A A A 




M O ec 


< O A • • 

O IM A A. • b O O' 

> A.A. o (b O OC A 
pMOa'A.S^CO'aA 

>aO<0'aa^ — .. 

lAOfeAOO:^ lAAAOOWWi-^ 
I aaHA|Jaa.a,AAAaa44Ai/ 
PI IA.IA.IOOI • I I 1 I I I I 
eOU<teaVhiWMH4aA(^Wk.fa.HU 


. U. < A A b. 


M A 
A "A 
A X 




• A O A 

r'Tii 

. © u ee 


4 

7 cl) 

o to 


« O < O A U V 

I A IPI A A 9 O 


;sc: = s 


ClkUUUUUSCO 


: 3 < a 2 
■* O' A A A 
< A A A 4 

!r Or (L a> a. 


8 < ~ 


I A A A A A 


OAO'AAO'OA 


. A A A © or X 


w a. e. e 

Hie 

A A A - 

X U. U. J 


» u -t 

4 O' -• 
b. tk J 


g 

I n 

27 

I < 

8 <0 


• A 'P' S 
I 9 U 

• 9 > V 

' !.> A *> 

b e S *0 


























127 


Appendix E 


DATA PLOTS 
















Appendix E 


DATA PLOTS 


Figure Dependent Variable 

E.l Total airfratne rework cost 

E.2 Total airframe rework cost 

E.3 Total airframe rework cost 

E.4 Production quantity 

E.5 Production quantity 

E.6 Airframe rework cost per acft 

E.7 Airframe rework cost per acft 

E.8 Airframe rework cost per acft 

E.9 Airframe rework cost per visit 

E.IO Airframe rework cost per visit 

E.ll Airframe rework cost per visit 

E.12 Airframe rework cost per visit 

E.13 Average time between overhauls 

E.14 Average time between overhauls 

E.15 Average time between overhauls 

E.16 Average time between overhauls 

E.17 Average time between overhauls 

E.18 Average time between overhauls 

E.19 Engine overhaul cost 

E.20 Engine overhaul cost 

E.21 Engine overhaul cost 

E.22 Engine overhaul cost 

E.23 Engine overhaul cost 

E.24 Engine overhaul cost 

E.25 Engine cost to repair 

E.26 Engine cost to repair 

E.27 Engine cost to repair 

E.28 Engine cost to repair 

E.29 Engine cost to repair 

E.30 Engine cost to repair 

E.31 Engine cost to repair 


Independent Variable 

Inventory (PDM Policy) 
Production quantity 
Fleet flying hours (PDM Policy) 
Age 

Percent organic maintenance 
Empty weight (PDM Policy) 
Airframe manufacturing cofit 
Production quantity 
Age 

Airframe manufacturing uost 

Percent organic m'' tenance 

Production quantity 

Turbine inlet temperature 

Engine removal rate 

Selling price 

Specific fuel consumption 

Engine weight 

Model qualification date 

Turbine inlet temperature 

Specific fuel consumption 

Selling price 

Engine weight 

Military thrust 

Model qualification date 

Turbine inlet temperature 

Specific fuel consumption 

Selling price 

Engine weight 

Maximum thrust 

Military thrust 

Model qualification date 







Dependent Variable 


129 


Independent Variable 


Figure 

E.32 

E.33 

E.34 

E.35 

E.36 

E.37 

E.38 

E.39 

E.40 

E.41 

E.42 


E.43 

E.44 
E.45 
E. 46 
E.47 
E.48 
E.49 
E.50 
E.51 


Airframe component repair cost 
Airframe component repair cost 

Airframe component repair cost 

Airframe component repair cost 

Airframe component repair cost 

Engine component and 
accessory repair cost 

Engii’n component and 
accessory repair cost 

Engine component and 
accessory repair cost 

Engine component and 
accessory repair cost 

Engine component and 
accessory repair cost 

Engine con^ionent and 
accessory repair cost 

Engine component and 
accessory repair cost 

Avionics component repair cost 

Avionics component repair cost 

Avionics component repair cost 

Avionics component repair cost 

Avionics component repair cost 

Avionics component repair cost 

Avionics component repair cost 

Avionics component repair cost 


Airframe manufacturing cost 

Airframe manufacturing cost 
(PBM Policy) 

Empty weight (PDM Policy) 
Empty weight (afterburner) 
Sortie rate 

Turbine inlet temperature 
Specific fuel consumption 
Selling price 
Engine- weight 
Maximum thrust 
Military thrust 

Model qualification date 

Black box count 

Suite weight 

Suite functions 

Mean time between demands 

Sortie rate 

Percent peculiar cost 

All-weather variable 

First flight date 














































































Natura,' logarithm of annual airframe rework cost per aircraft 


136 



Fig. E.7-Varialion of annual airframe rework cost per aircraft with 
airframe manufacturing cost 









• Fighter/attack 
■ Bomber/cargo 
A Other 

















138 


13.5 


• Fighter/attack 
■ Bomber/cargo 
A Other 













K 


I 


65 7.0 75 

Natural logarithm of airframe manufacturing cost 


Fig. E. 10—Variation of airframe rework cost per visit with 
airframe manufacturing cost 












Natural logarithm 



Fig. E.11—Variation of airframe rework cost per visit with 
percent organic maintenance 








MMiM 















airframe rework 


1351 - 


,3 .0 H 


12.5 I— 


I 12.0 


11.5 


11.0 


I 105 


2 10.0 

3 


9.5 


9.0 I— 


8.5 H 


• ■ 


• A 


• Fighter/attack 
■ Bomber/cargo 
A Other 


J_L 


B • • 




















■ 


Fighter/attack 

Bomber/cargo 



Fig. E'.13—Variation of ATBO with turbine inlet temperature 





143 


• Fighter/attack 
■ Bomber/cargo 



Natural logarithm of engine removal rate 


E.14—Variation of ATBO with engine base-level removal rate 
























145 


■ 


9 Fighter/attack 
■ Bomber/cargo 


• • 


9 


9 


9 



Fig. E.16—Variation of ATBO with specific fuel consumption 


• 9 






















9 

• Fighter/attack 
■ Bomber.'cargo 

















• Fighter/attack 
■ Bomber/cargo 


I_ L 


1 


7.60 7,65 7.70 7.75 7.80 7.85 

Natural logarithm of turbine inlet temperature 


7.90 


Fig. E.19—Variation of overhaul cost with turbine inlet temperature 













Natural logarithm of average cost per overhaul 



* Fighter/attack 
■ Bomber/cargo 



1 


I 


1 















































• Fighter/attack 
■ Bomber/cargo 




75 8.0 


85 9.0 95 10.0 105 


Natural logarithm of military thrust 


Fig. E.23—Variation of overhaul cost with military thrust 
























• Fighter/attack 
■ Bomber/cargo 







■ 

■ 






J_I_I_1_L 

100 200 150 250 300 

Model qualification date (months since January 1943) 


Fig. E.24—Variation of overhaul cost with model qualification date 











Fig. E.25—Variation of annual cost to repair with turbine inlet temperature 















• Fighter/attack 
s Bomber/cargo 


I I I I I I * 

-1.0 -.8 -.6 -.4 -.2 0 

Natural logarithm of specific fuel consumption 


Fig. E.26—Variation of annual cost to repair with specific fuel consumption 











J.30 


• Fighter/attack 
■ Bomber/cargo 






* 








■ 


■ 




J_I_I_I_L 


s 


s 


s 


s 











Natural logarithm of annual cost to repair 


9 


j_ • Fighter/attack 
I ■ Bomber/cargo 


8 


7 


6 


5 




4 


3 


2 


1 



-L 

6.0 


,.l.I_ \ _1_L 

65 7.0 75 8.0 85 

Natural looarithm of enoine weioht 













> Fighter/attack 
I Bomber/cargo 


• ■ 








I 


1 


1 


1 























Natural logarithr.i of airframe component repair cost per aircraft {jCn^CSTPAC) 


11.5 


11.0 I— 


1051— 


10.0 l- 


9.5 H 


9.0 h- 


8.5 V- • 


8.0 1 


• • 
• ■ 


1 




9 Fighter/attack 
K Bomber/cargo 
A uther 


1 


1 


1 


55 65 6.5 7.0 7 5 8.0 

Natui ’ lowrithm of airframe man .factoring cost Af MFGC) 

Fig. £.3;^—Va-iation of airframe component repair cost with 
airframe manufacturing cost 


















Natural logarithm of annual airframe component repair cost per aircraft 
























■ 




■ 



• 


■ ■ 




• 

■ 




• 




• 

• 




• 

• 

• 

• •• 





• 


■ 



• 




■ 



• Afterburner 





■ No afterburner 


— 

■ 




•l 1 

i L 


I i 1 

1 

9.0 9. 



1.0 11.5 12J) 

125 


Natural logarithm of empty weight 


Fig. E.35—Variation of airframe component repair cost with 
empty weight and afterburner 




m 















— 


■ 


■ 

■ 


■ 


— 

• 

•• • 

• 

• 

■ 

• 


-- 

■ 

•• • 

• 

• • • 

• • 

• 





A 

— 


• • 


• 

■ 

Fighter/attack 

Bomber/cargo 

A • 


A 

Other 

A 

• 


_L 

_L- 

1 ^ 1 ^1 



Natural logarithm of annual sortie rate 


Fig. E.36—Variation of airframe component repair cost with sortie rate 


















Natural loflarithm of annual angina ootnponant and accaaaory rapair coat 












167 



-1.0 JB -.6 -.4 -.2 0 


Natural logarithm of specific fuji consump ' i 


Fig. E.38—Variation of annual engine component and accessory repair cost 
with specific fuel consumption 













Natural logarithm of annual engine component 


















Natural logarithm of annual engine component and accessory repair cost 


105 


• Fighter/attack 
■ Bomber/cargo 



Fig. E.40—Variation in annual engine component and accessory repair cost 

with engine weight 








Natural logarithm of annual engine component and accessory repair cost 





Fig. E.41—Variation in anntial engine component and accesf -..y repair cost 

with maximum thrust 












Natural logarithm of military thrust 


Fig. E.42—Variation of annual engine component and accessory repair cost 

with military thrust 












40,000 


• Fighter/attack 
■ Bombar/cargo 


35,000 


S, 30,000 


>5,000 


> 0,000 


5,000 




0,000 




5,000 


■ 






J_ I T I _L 

100 150 200 250 300 


Model qualification date (tnondis since January 1943) 


Fig. E.43—Variation in annual engine component and accessory repair cost 

with model qualification date 











Natural logarithm of annual avionics component repair cost 



















Natural logarithm of suite weight 

Fig. E.45—Variation of annual avionics component repair cost with suite weight 



































Natural logarithm of annual avionics component 















Fig. E.48—^Variation of annual avionics component repair cost with 

annual sortie rate 














A 


• Fighter/attack 
■ Bomber 
A Cargo 

O Reconnaissance 
D Trainer 


■ 




■ 




A 




O 


A 




12 3 4 

Natural logarithm of percent of avionics suite item cost which is peculiar to MDS 


Fig. E.49—Variation of annual avionics component repair cost with 
percent of avionics suite item cost which is peculiar to MDS 


• • 















Natural logarithm annual avionics component repair cost 



Natural logarithm of all-weather dummy variable 


Fig. E.50—Variation of annual avionics component repaii cost with 
all-weather dummy variable 


i 














Annual avionics component 




• Fighter/attack 
■ Bombei 
^ Cargo 

O Reconnaissance 
□ Trainer 


150.000 


100,000 


50,000 



Aircraft first flight date (months since January 1943) 


Fig. E.51—Variation of annual avionics component repair cost with 
aircraft first flight date 













181 


Appendix F 

NOTE ON AIRFRAME REWORK COST 


This appendix considers two alternative representations of 
airframe rework. In the first, the total annual cost for a fleet of 
aircraft is estimated directly. This method naturally has an inventory 
size or activity variable as a major explanatory variable; but with a 
large enough sample size, additional useful variables can be introduced 
into the equation. The particular nature of airframe rework makes this 
approach more appealing than it is for the other categories of 
maintenance activity. Airframe rework is a combination of PDMs and 
several other, less extensive, maintenance tasks, some of which take 
place during PDM visits. Airframe rework cannot be divided into two 
distinct activities (as can engine overhaul and engine repair). On 
the other hand, it is the result of a limited number of depot visits 
each year rather than a large number of small tasks (as is the case 
for component repairs). Estimating total cost also allows for the 
possibility of accounting for economies of scale. 

The second alternative represents annual cost for a single weapon 
system as the product of (1) the average cost of a rework visit to the 
depot and (2) the average number of visits per year. For many systems, 
the prescribed interval between PDM visits was increasing during the 
years covered by our data. These increases are the result of 
management decisions based on knowledge gained during previous years of 
operation of the various weapon systems. The value of the prescribed 
interval may therefore be related to the age of the system. Since 
the actual average interval takes many months to transition from one 
prescribed value to another, there may be only a weak relationship 
between the production data in this study's data base and the intervals 
prescribed during the period covered by the data. 


1 










182 


TOTAL ANNUAL FLEET COST 

Based on the full sample of all r.ypes of aircraft, four variables 
were found to be related to total annual fleet cost: fleet flying 
hours (FH^, emr iy weight (EW), production quantity (PQ), and PDM 
designat (PDM). According to these results, fleet rework cost is 
driven i more by aircraft size and utilization rate, and by policy 
decisions, than by technical characteristics. The variation of total 
cost with flying hours and empty weight is shown in Figs. F.l and 
F.2. Other total cost plots are included in App. E. Tables F.l 
and F.2 list the equations developed using these variables. The 
eqi ations wi'^h the best statistics use combinations of two or three of 
the four variables. As much as 90 percent of the variance is accounted 
for ny v-^re equations. In these tables TOTCST is the total annual 
fl^-.^r > vframe rework cost in 1978 dollars. 

Jhe exponent of the PDM policy variable in Table F.l leads to a 
factor of 15 as the difference between the equation used for aircraft 
with a PDM program and for those without one. In other words, 
according to this result, not having a PDM program on a new aircraft 
would save about 93 percent of the airframe rework cost that would be 
incurred if a PDM were required. This equation, of course, says 
nothing about other costs that might be affected by such a decision. 

A PDM is only one part of a scheduled maintenance program. Avoiding 
use of a PDM could require larger than normal costs for base-level 
scheduled inspections. Also, unscheduled maintenance requirements 
could be larger than otherwise would be expected. Such effects are 
beyond the scope of this study but must be addressed i any application 
of the equations developed by the study. 

A close examination of Fig. F.l shows that, for high levels of 
flying activity, the total annual airframe rework cost data can be 
grouped into four classes. For a gxven flying-hour figure, the lowest 
rework cost is associated with trainers. The next class is made up of 
the light cargo aircraft: C-130E, C-141A, and KC-135. Even higher costs 
are associated with fighter/attack aircraft. The most costly aircraft, 
in terms of aircraft rework, are the heavy boi ^er and cargo aircraft; 





• Fighter/attack 
■ Bomber/cargo 












wsm 




Table F.l 


TOTAL AIRFRAME REWORK COST EQUATIONS: FULL SAMPLE 


Statistics 


Equation 


SSE F N 


Size 


TOTCST = 86A.0 

(.0009) 


Utilization 

TOTCST = 249.4 FH 


Pol Lay 

TOTCST = 388,100 PQ 

Size/Uti lizatim 


,0.9513 

(. 0000 ) 

,0.7424 


(. 0000 ) 


TOTCST = 0.0570 


(.0000) (.0000) 


Size/Polioy 
TOTCST = 14.52 

(. 0000 ) (. 0000 ) 

Utilization/Poliay 
TOTCST = 2492 

(. 0001 ) (. 0000 ) 

TOTCST = 911.6 PqO.4944 

(.0013) (.0016) 


0.27 1.56 12 34 


0.43 1.37 24 34 


0.43 1.38 23 33 


0.66 1.08 30 54 


0.71 0.99 38 33 


0.90 0.59 94 23 


0.58 l.:.l 21 33 


Size/Utilization/Poliay 

TOTCST = 0.1152 0.83 0.77 48 33 

(. 0000 ) (. 0000 ) (. 0000 ) 


» 'ki4i«E2k, 







Natural logarithm of total airframe rework cost (-^rC'TOTCST) 


























186 


Table F.2 


TOTAL AIRFRAME REWORK COST EQUATIONS: 
MOST REPRESENTATIVE SERIES 



Statistics 



Equation 

r2 

SEE 

F 

N 

Comments 

Size 






0 ‘.'514 

TOTCST = 368.8 EW 

(.0002) 

.54 

1.12 

20 

19 


Utilization 






0 7444 

TOTCST = 2273 FH *' 

(.0161) 

.24 

1.44 

5 

19 


Size/Uti tizazion 






TOTCST = 0.2909 

(.0019) 

.73 

.88 

22 

19 


Size/PoHoy 






TOTCST = 56.15 EW 

(.0001) (.0051) 

.69 

.94 

17 

18 


Uti lization/Po Hay 

TOTCST = 2629 

(.0474) (.0571) 

Size/Uti lizaiion/Poliay 

.33 

1.38 

4 

la 

Fails F-test, PQ 
coefficient does not 
meet 5% significance 
criterion 

TOTCST = 0.2063 FH°'^^°® 

(.0037) (.0000) 

PqO. 4143 g2 

(.0062) 

.45 

21 

18 























Table F.3 

TOTAL AIRFRAME REWORK COST EQUATIONS FOR 
FIGHTER/ATTACK AIRCRAFT 


Statistics 


Equation 


SEE 

F 

N 

Utilization 





1 124 

TOrCST = 4.354 FH 

f.OOOO) 

.69 

.99 

43 

21 

Uti lization-Tealmiaal/Performanee 





TOTCST = 0.003047 MAXDLF^"^^^ 

(.0000) (.0250) 

.74 

.93 

24 

20 

Uti lizaticn/Poliay 





TOTCST = 35.22 Pq0.4020 

(.0001) (.0023) 

.00 

.82 

34 

20 

TOTCST = 78.37 

(.0017) (.0001) 

.94 

.47 

76 

12 

TOTCST = 11,550 

(.0001) (.0002) 

.95 

.43 

91 

12 

Tedhniaa t/Pevformance-Poliaxj 





TOTCST = 8.457 MAXLDF^'^^^ PDM^''°^ 
(.0010) (.0000) 

.95 

.44 

86 

12 


Commeats 


MAXLDF is not sig¬ 
nificant when used 
alone. Ei^onent 
magnitude. 


MAXLDF is not sig¬ 
nificant when used 
alone. Exponent 
magnitude. 


airframe production cost. These variables are highly correlated, and 
using them together generally produces an equation with unacceptable 
multicollinearity statistics. The best statistics are for equations 
that use all of these five variables except airframe cost. 

Care should be used in applying the equations that include age as 
an explanatory variable. The age exponents are as large as 0.69, which 
would make a 20-year-old aircraft eight times as costly to maintain as 
a one-year-old aircraft. Examination of the data base reveals that the 
influence of age is derived largely from the F-15A and TF-15A, which 









Table F.4 

AIRFRAME REWORK-COST-PER-VISIT EQUATIONS: 
TOTAL SAMPLE 


Equation 


CSTPVST = 5.A65 EW 


jO.9462 

(. 0000 ) 


Teahnioa l/Perfomtanae 
CSTPVST = 136.0 APTITGC^ 


(.0003) 


Utilization 
CSTPVST = 43,010 age' 


CSTPVST = 43,010 AGE°'^°®^ 

(.0039) 

Policy 

CSTPVST = 9755 MAINTPCT^*^^^® 

(.0003) 

Size-Utilization 

CSTPVST = 5.850 AGE°*^^®^ 

(.0000) (.0024) 

Size-Po Hoy 

CSTPVST = 6.871 MAIHTPCT®'^°^^ 

(.0001) (.0050) 

CSTPVST = 14.31 Pq-0.2263 

(.0000) (.0250) 

Technioal/Performanae-Poliay 

CSTPVST = 88.43 MAINTPCT®'^®^® AFMFGC®*®°°^ 

(.0015) (.0010) 

Utiliza tion-Po Hay 

CSTPVST = 1196 MA1NTPCT°'^^^^ AGE°*^^^* 

(. 0001 ) (. 0002 ) 

Size-UtiHzation-PoHay 

CSTPVST = 13.18 AGE®’^^^® pQ-0.2016 

(.0000) (.0011) (.0184) 

CSTPVST = 6.059 MAINTPCT®-^^®® AGE®’^‘®® 

(.0001) (.0053) (.0003) 

CSTPVST > 11 15 AGE®’®®^^ MAINTPCT®’®®®® pQ-0.1521 

(.0001) (.0002) (.0134) (.0471) 


Statistics 


R SEE F N 


.48 1.04 29 33 


.41 .96 16 25 


.22 1.24 8 31 


.32 1.20 14 33 


.64 .86 25 31 


.58 .95 21 33 


.54 .99 18 33 


.61 .80 17 25 


.53 .98 16 31 


.69 .81 20 31 


.72 .78 23 31 


.75 .75 19 31 












190 










Table F.5 

AIRFRAME REWORK-COST-PER-VISIT EQUATIONS: 
MOST REPRESENTATIVE SERIES 


Equation 


Size 

CSIPVST - 9.542 

(.0004) 

Teahniaal/Pepfopmcaica 


CSTPVST - 248.3 AFMFcd 


,0.9736 

(. 0021 ) 


Utilizr^tion 
CbTPVST - 33,760 ACe' 


.0.6096 
(.0367) 

Policy 

CS7{VST - 13,690 MAINIPCT°‘*^^ 

(.0075) 

Size-Utilization 

CSTPVST - 4.222 AGE®’®*®° 

(.0002) (.0094) 

Size-Policy 

CSTPVST • 12.81 HAINTPCT®'®^®^ 

(.0022) (.0398) 

CSTPVST - 55.15 EH®’^®®^ pg-0.4816 
(.0001) (.0078) 

Technical/Pepformnce-Poliay 

CSTPVST - 145.9 MAINTPCT®'AFMFGC®’®^®® 
(.0273) (.0037) 

' 'A lization-Po ticy 

CSTPVST ■■ 916.7 MAINTPCT®'®®*® ACE®’^®^* 


(.0007) (.0021) 


Size-Utilization-Policy 


Statistics 


SEE F N 


Cotaieiits 


.52 1.11 17 18 


.54 0.90 13 13 


Not significant 

.20 1.45 4 17 (falls F-test) 


.32 1.32 7 18 


.69 0.94 16 17 

.61 1.03 12 18 

.68 0.94 16 18 

.69 0.78 11 13 

.62 1.03 12 17 


CSTPVST • 

4.692 

£.„0.6235 

MAINTPCT®’®®®® AGE®'®^®® 

.82 

0.74 

20 

17 



(.0012) 

(.0044) (.0007) 





CSTPVST - 

18.79 

E^®.87®7 

age®'®®®® Pq"®'®^®^ 

.78 

0.82 

15 

17 



(.0001) 

(.0144) (.0179) 





CSTPVST - 

16.80 

E„0.6561 

^ Ge ®'5773 maINTPCT®'®®®® Pq"®'®^®^ 

.89 

0.61 

24 

17 



(.0002) 

(.0007) (.0026) (.0094) 













Natural logarithm of airframe rework cost per aircraft K£n. CSTPAC) 








■ 

— 



■ 


— 



■ 


— 

.*: 

• • 




— 

• 

■ 

■ 



• 


■ 



• 





• 

• 

• 



A 



• Fighter/attack 


— 


• 

■ Bomber/cargo 

A Other 


-A- 

-J_L 

• 

_1_L. 

-JL-j_ 

-J_1_ 

J_ 


















192 




Statistics 


Equation 

SEE F N 

Connnents 


CSTPVST = 0.02166 EW 

(.0010) 

.40 

1.02 

13 

21 


Uti lization 






CSTPVST = 28,930 

(.0022) 

.39 

0.99 

11 

19 


Policy 






CSTPVST = 10,930 MAINTPCT^'^^^^ 

(.0043) 

.31 

1.09 

9 

21 


Size-Utilization 






CSTPVST = 0.3080 

(.0131) (.0014) 

.62 

0.77 

13 

19 


Size-Policy 






CSTPVST = 0.4302 MAINTPCT°' 

(.0148) (.0070) 

.53 

0.89 

10 

21 


Utilization-Policy 






CSTPVST = 2454 MAI.NTPCT°'AGE®'^^®^ 

(.0044) (.0013) 

.57 

0.82 

11 

19 


CSTPVST » 60,352 AGE®’®^^^ PQ“®‘^^®^ 

(.0059) (.0525) 

.44 

0.97 

6 

18 

PQ slightly exceeds 
the 5% significance 
criterion 

CSTPVST = 35.22 Pq-0.5980 

(.0001) (.0001) 

.63 

0.32 

15 

20 

FH is not signifi¬ 
cant by itself 

CSTPVST = 79.24 FH®’^^®^ AGE®*^^®® Pq-0.4654 
(.0015) (.0064) (.0007) 

.69 

0.73 

11 

19 

FH is not signifi¬ 
cant by itself 









193 


Table F.7 

AIRFRAME REWORK-COST-PER-VISIT EQUATIONS 
FOR BOMBER/CARGO AIRCRAFT 

Statistics 

2 

_ Equation _ R S EE F N 

Size 

CSTPVST = 0.009506 .64 .58 9 7 

(.0159) 

Techniaal/Performanae 

CSTPVST = 4.633 AFMFGC^'^^^ .67 .55 10 7 

(.0119) 


Within the Air Force should visit the depot less often than one 
maintained under contract. Production quantity for the sample data is 
plotted versus inventory size in Fig. F.5. 

In selecting an equation for use in estimating, one might first 
consider the problems associated with i>redicting the number of depot 
visits or production quantity. The results in Table F.8 fit the data 
rather poorly, so it might be best not to use them if an alternative 
can be found. If one has no other way of predicting the parameter, 
then it would perhaps be best to avoid the cost-per-visit equations 
and equations that use production quantity as an explanatory 
variable. 
















’•c'*w5T^'-*'f**^ •»> 


Table F.8 


PRODUCTION QUANTITY EQUATIONS 



Statistics 


Equation 

SEE F N 

Comments 


Full Sample 


Ut% Ixzatvan 


PQ = 28.38 + 0.3120 INV 
(<.0005) 

.44 

77 

24 

33 

PQ = 55.08 + .3542 INV - 
(<.0005) 

3.983 AGE .52 

(<.15) 

75 

15 

31 

Uti lizatian-Poliay 





PQ = 82.70 + 0.2843 INV 
(<.0005) 

- 0.7184 MAINTPCT .50 
(<.05) 

74 

15 

33 


AGE coefficient is not 
significant at 5% and 
is not significant 
when used alone 


Sample of Representative Series 


Utilization 

PQ = 50.15 + 0.2479 INV 

(<. 01 ) 

PQ = 106.4 + 0.3023 INV - 7.641 AGE 
(<.025) (<.05) 


.30 97 7 18 

.46 91 6 17 


Jti lization-Policy 

PQ = 87.08 + 0.2383 INV - 0.6207 MAINTPCT .35 97 4 18 

(<.025) (<.20) 


MAINTPCT coefficient i 
not significant at 5% 
and is not significant 
when used alone 









Natural logarithm of production quantity 


195 



Fig. F.5—Production quantity reflects inventory size 























^•*'}_ •Ai-- ^ 






ERRATA 


^ANJ> 

R-2731-PA&E 


SstlniatlnR Aircraft Depot Maintenance Costs , 
by Kenneth E. Marks. Ronald W. Hess. July 1981. 


The following correction should be made on pp. xvii, 46, 65, and 100: 


AFMFGC Airframe manufacturing cost; cumulative average cost of first 
100 units, including manufacturing labor and materials 
(tens of thousands of FY 1978 dollars) 


The following correction should be made to the heading appearing in 
Table D.l on p, 123: 

Airframe 

Manufacturing 

Cost 

(78 $ X 10^ ) 

Explanation: For equations using airframe manuiacturing cost as an 

independent variable, cumulative total cost of the first 
100 units, in millions of dollars, was inadvertently used 
in the statistical analysis. Thus, in order that the 
values of airframe manufacturing cost used in the statistical 
analysis correspond to the cumulative average definition used 
throughout the report, the units must be reduced from millions 
to tens of thousands.