Ad 0856021

UNCLASSIFIED
AD NUMBER
AD856021
NEW LIMITATION CHANGE

TO
Approved for public release, distribution
unlimited
FROM
Distribution authorized to U.S. Gov't.
agencies and their contractors;
Administrative/Operational Use; FEB 1969.
Other requests shall be referred to Rome
Air Developmental Center, Griffiss AFB,
NY.
AUTHORITY
USAF ltr, 27 Aug 1973
THIS PAGE IS UNCLASSIFIED

0 RADC-TR-68-403
Final Technical Report

C.) February 1969
O0
RELIABILITY PREDICTION - MECHANICAL

STRESS/STRENGTH INTERFERENCE (NONFERROUS)
Charles Lipson
Narendra J. Sheth
Ralph L. Disney
Mehmet Altun
University of Michigan
This document is subject to special export

controls and each transmittal to foreign
governments, foreign nationals or repre-
sentatives thereto may be made only with
prior approval of RADC (EMERR), GAFB,
NY 13440.
DDC
This Document Contains IliK
Missing Page/s That Are AUG 6 1969
Unavailable In The
Original Document L"LiLBU
Rome Air Development Center

Air Force Systems Command
Griffiss Air Force Base, New York
When US Govemment drawings, specifications, or other data are used for any purpose other
than a definitely related government procurement operation, the government thereby incurs
no responsibiiity nor any obligation whatsoever; and the fact that the government may have
formulated, furnished, or in any way supplied the said drawings, specifications, or other
data is not to be regarded, by implication or otherwise, as in any manner licensing the
holder or any other person or corporation, or conveying any rights or permission to manu-
facture, use, or sell any patented invention that may in any way be related thereto.
CyST' WHITE f]
SECTION
99C BUFF
SECTION
' ..~~~............................. i,
S............................
Do anoor SPE•I
AVAIL, sAL
'AVA........ !.06
Do not etewn this copy. Retain or destroy.ý

St
RELIABILITY PREDICTION - MECHANICAL

STRESS/STRENGTH INTERFERENCE (NONFERROUS)
Charles Lipson
Narendra J. Sheth
Ralph L. Disney
Mehmet Altun
University of Michigan
This document is subject to special export

controls and each transmittal to foreign
governments, foreign nationals, or repre-
sentatives thereto may be made only with
prior approval of RADC (EMERR), GAFB,
NY 13440.
.7
• i
FOREWORD U
This report is the result of a study performed by the University of

Michigan, Department of Mechanical Engineering, Ann Arbor, Michigan, from
March 1967 to August 1968. The research team was headed by Professor
Charles Lipson, actively assisted by Mr. Narendra J. Sheth, Associate
Research Engineer. Mr. Sheth was instrumental in developing the tech-
niques and methods required for the project and headed the group respons-
ible for the data collection and the development of the computer
approach used in the data analysis. Mr. Mehmet Altun contributed
significantly toward the development of the computer approach and the
data analysis. Professor Ralph L. Disney of the Department of Industrial
Engineering developed and evaluated the analytical expressions of the
interference.
The following contributed to the various phases of the project: Messrs.

Raymond B. Pittman, John P. Godwin and Edward J. Gainer.
The study was performed for the Rome Air Development Center under
contract F30602-67-C-2014, Project 5519, Task 551902. The RADC project
engineer was Mr. Donald W. Fulton (EMERR).
The distribution of this document is limited under the U. S. Mutual a

Security Act of 1949.
This report has been reviewed and is approved.
DONAL W.-FULTON
Reliability Engineering Section
Re•liability Branch
Approve* e;I .era ~LL

oPrVft rig 41
Chief, A4 onced Studies Group
it /V
ABSTRACT
This study addressed itself to the development of a Stress/Strength

Interference Theory in form of a practical engineering tool, to be used for
designing and quantitatively predicting the reliability of mechanical parts
and components subjected to mechanical loading.
In our past investigation ("Reliability Prediction-Mechanical

Stress/Strength Interference Theory" by Charles Lipson, Narendra J. Sheth and
Ralph L. Disney; Final Report RADC-TR-66-710, March 1967, referred to in the
present report as Reference 1) the Interference Theory was developed for the
ferrous materials. In the present study, the work was extended to non-ferrous
materials.
Early practices in stress-strength relationship dealt almost entire-

ly along the lines of factors of safety. Utilization of such factors is
justified when they are based on considerable experience with parts not too
different from the one under consideration. ,rowever, when substantial changes
in the geometry, the processing, or the function of the part are contemplated,
a major error may result if the old factyrs of safety are projected to the new
set of conditions.
In the present investigation, an approach was used which attempted

to recognize the above limitations. Instead of an indiscriminate grouping of
all the variables affecting stress and strength (generally fatigue strength)
into one index (factor of safety) these variables were individually recognized.
The principal variable is the scatter in the stresses imposed on the part and
in the strength of the material resisting these stresses.
The prevailing practice is to use the mean values of the calculated

stress and strength, ignoring the natural scatter that each may possess. How-
ever, the variability in these two factors results in a statistical distribu-
tion of stress and strength. When these two distributions interfere, that is
when stress becomes higher than strength, failure results. Means of expres-
sing these distributions, in a practical engineering sense, and mans of cal-
culating the resulting interferences, represent the heart of the present stud)
In our past investigation(1) on ferrous materials, the problem of

determining the scatter in the fatigue strength andsubsequently, its distri-
bution function was resolved by a graphical method. Among the several dis-
tribution functions that were considered, only the Weibull distribution was
used because of its wide use and the difficulty of handling other distribu-
tions by a graphical method.
In the present investigation on the non-ferrous materials, a com-

puter approach was developed which made the determination of the distribution
functions of fatigue strength considerably less difficult (Section 6). Con-
sequently, several distributions were tried: Weibull, Largest Extreme Value,
Smallest Extreme Value, Logistic and Normal. The reasons for the selection
of these particular distributions are given in Section 6.3. It was later
found that the Logistic distribution had similar characteristics to the Normal
iii
distribution and therefore it was not followed any further.
Mechanical Properties Data Center in Traverse City, Michigan, was

found to be a very useful source of information for the fatigue strength data
for various non-ferrous materials under various conditions such as the type of
loadingsurface finish, stress concentration, heat treatment, temperature,
environment, etc. Considerable fatigue data thus obtained were then organized
and systematized according to materials and conditions.
A computer program was developed so that when these data (raw S-N
data) were fed into the program, the computer printed out the degree of fit
and the parameters for all the distribution functions tried for any given set
of fatigue strength data.
In the first part of this program, the scatter in life was converted
into the scatter in strength. In the second part, the distribution functions
(Weibull, Normal, Logistic, LEV and SEV) were fitted to these strength data.
The best fitting distributions (those with the highest degree of fit) and
their parameters were then chosen, and these are presented in a tabular form
in Appendix 1 for various non-ferrous materials and under various conditions.
In order to present these data in a pictorialand thus more easily

digestible form, the representative parameters were then plotted on log-log
scale against life for various materials and conditions (Section 6.7). In
order to put this information on the same consistent basis, all the data were
expressed in terms of the same distribution (Normal). The plots were made by
expressing the mean (P) and standard deviation (a) as a function of life
(Section 6.7). The reasons for the selection of the Normal distribution for
[
these graphs, the description of the computer approach, and the statistical
tools used in developing this approach are given in Section 6.
The problem of stress distrilltion (Section 7)was investigated in

more detail than in the previous study, ,,but no additional information was
located. When one speaks of "stress distribution" he usually refers to a
spectrum of loading or stresses to which a part is subjected. Indeed, most
of the available data on the subject is expressed in this manner. In an en-
gineering sense, this kind of a distribution means the number of times that a
liven part is subjected to a given load or stress. In the Interference Theory
however, this is not what is wanted. For consistency with the strength dis-
tribution, the number of parts subjected to a given stress is required instead.
In the present investigation, the required stress distribution was

obtained in the saw manner as is our past study. This was dona by convert-
Ing the stress spectrum, which generally has some mean stress, into a spec-
trum with zero man, with the aid of the Goodman diagram. This facilitated
the conversion of the resultant spectrum into an equivalent stress based on
sero mean stress. The conversion was accomplished by means of Miner's rule.
The required stress distribution was then expressed in term of the equivalent
stress. This distribution was then compared with the strength distribution to
determine the probability of interference.
i-
iva
I!
Most investigators in the past have assumed both the stress and
strength distributions to be normal. In those cases when they were not normal,
a Monte-Carlo technique was employed. This involved a sophisticated means of
randomly selecting a sample from one distribution and comparing it with a
sample from a different distribution.
In view of the serious limitations of the Monte-Carlo ;.echniquea

method of Integrals was developed (Appendix 3) and used in the present study.
This method involved developing an integral resulting from the interference
of two distributions and evaluating this complex integral. Numerical analysis
was carried out (Appendix 3) using an IBM 360 computer to solve these inte-
grals.
In our past investigation on ferrous materials(l), these integrals

and hence the interference values were evaluated and tabulated for the fol-
lowing combinations:
Stress Distribution Strenath Distribution

Normal Normal
Weibull -Weibull
Normal Weibull
In the present investigation on non-ferrous materials, data analysis

(Section 6) revealed many cases where fatigue strength followed distributions
other than Weibull or Normal. Hence, it was necessary to develop tables of
interference values for the following additional combinations:
Stress Distribution Streuxth Distribution

Normal Weibull (Extended-see
Section 4.2)
Normal Largest Extreme Value
Normal Smallest Extreme Value
These interference values (tabulated in Appendix 2) were found to be

highly non-linear. Therefore, for the purpose of Interpolating these values
in a given table or between the tables, it is recommended that a higher order
interpolation than linear be used. A suggested method of interpolation,
with some solved examples illustrating this method, is gIven in Section A-2.1.
In order to *how the application of The Interference Theory Tech-

nique developed in the present study two examples were solved, as described
in Section 9.
Av
V I
TABLE OF CONTENTS
Section Page No.

1. Introduction 1
2. Interference Theory 5
2.1 Two Normal Distributions 5
2.2 Consistency of Two Distributions 8
2.3 Non-Normal Distributions 10
2.4 The Integral Method 10
3. Objectives of the Present Study 11
4. Study Approach 12
4.1 Literature Search 12
4.2 Theoretical Analysis 12
4.3 Empirical Data 13
4.4 Factors Affecting the Statistical Distribution of Fatigue
Strength 14
4.5 Computer Analysis of Strength Data 16
5. Analytical Expressions for Interference 17 *

5.1 Introduction 17
5.1.1 Interference Probabilities 17
5.1.2 Calculation of Probabilities of Failure Is
5.1.3 Interference Tables, Appendix 2 22
5.2 Use of Interference Tables in Appendix 2 23
5.2.1 Parameters for the Weibull Distributed Strength,
Normal Distributed Stress 23
5.2.2 Parameters for the Smallest Extreme Value Distrib-
uted Strength, Normal Distributed Stress 23
5.2.3 Parameters for the Largest Extreme Value Distributed
Strength, Normal Distributed Stress 24
5.2.4 Use of the Tables, Explanation of Hissing Values
and Interpolation 24
6. Statistical Distribution of Fatigue Strength by Computer Approach 26

6.1 Least Squares Method of Fitting a Line to S-N Type Data 26
6.1.1 Method of Least Squares 26
6.1.2 Application of Least Squares Method to S-N Data 28
6.1.3 Correlation Coefficient 29
6.2 Conversion of Life Data to Strength Data 31
6.2.1 Computation of Parallel S-N Lines 31
6.2.2 Computation of the Scatter of Strength at a Given
life 33
6.3 Distribution Functions 33
6.3.1 Welbull Distribution 34
6.3.2 Normal Distribution 34
6.3.3 Extreme Value Distributions 35
viK /.
S6.4 The Linearizing Transformation Equations of the Dis-
tribution Functions 38 -
6.4.1 Weibull Distribution 40
6.5 Fitting the Linearized Distribution to the Strength Data 45
6.6 Determination of the Best Fitting Distribution and Its
Parameters 51
6.7 Graphs of the Fatigue Strength Distribution Parameters 53
6.7.1 Expression of Data in Terms of Normal Parameters 53
6.7.2 Explanation of the Graphs 57
6.7.3 Use of Graphs 60
6.7.4 Graphs 61
7. The Statistical Distribution of Stress 118

7.1 Stress Spectrum vs. Stress Distribution 118
7.2 Conversion of a Stress Spectrum to an Equivalent Stress
(Sequ) 118
7.2.1 Miner's Rule 120
7.2.2 Corten-Dolan's Rule 124
8. The Interference of a Stress Distribution with a Strength

Distribution 128
9. Application of Interference Theory to Design Problems 133

9.1 Example Problem No. 1 (Effect of Stress Concentration on
Reliability) 133
9.2 Example Problem No. 2 (Effect of Temperature on Reliability)143
Conclusions and Recommendations 147
References 149
bibliography 151
Appendix 1 Tables of the Distribution Parameters of Fatigue

Strength 155
A-lI1 Index to Distribution Parameters for Various Non-
Ferrous Materials 156
A-1.2 Distribution Functions and Their Parameters Used In
Tables 161
A-l.3 List of Abbreviations and Symbols Used in Tablas 165
A-l.4 Composition of the Non-Ferrous Materials 166
A-1.5 Tables 173
vii
•!• .' ].•. I, II~A_ .. . _____IL

A-1.5.1 Aluminum Alloys 174
A-1.5.2 Cobalt Alloys 236
A-1.5.3 Copper Alloys 243
A-1.5.4 Magnesium Alloys 250
A-1.5.5 Nickel Alloys 257
A-1.5.6 Titanium Alloys 283
A-1.5.7 Other Non-Ferrous Alloys 346
Appendix 2 Tables of Interference Values 351

A-2.1 A Method for Interpolating the Interference Values 352
A-2.1.1 Higher Order Interpolation Procedure 353
A-2.1.2 Example Problem for One-Way Interpolation 355
A-2.1.3 Example Problem for Two-Way and Three-Way
Interpolation 358
A-2.2 Stress Distribution - Normal, Strength Distribution -
Weibull 368
A-2.2.1 Stress Standard Deviation = 0 369
A-2.2.2 Stress Standard Deviation # 0 373
Normal 40C
A-2.3.1 Stress Standard Deviation = 0 401
Largest Extreme Value 405
A-2.4.1 Stress Standard Deviation 0 406
A-2.5 Stress Distribution - Normal, Strength Distribution - 9,.
Smallest Extreme Value 416
A-2.5.1 Stress Standard Deviation -.0 417
A-2.5.2 Stress Standard Deviation A 0 420
Appendix 3 Evaluation of the Integrals by Numerical Analysis 427

A-3.1 The Problem 428
A-3.2 Simplification of the Equations 429
A-3.3 Numerical Methods Considered for the Problem 430
A-3.4 Simplifying Approximations 434
A-3.5 Limits of Integration 435
A-3.5.1 Limits for Normal - SmAllest Extrere Value
Case 435
A-3.5.2 Limits for Normal - Largest Extreme Value
Case 435
A-3.6 Determination of Simpson's Rule Step-Size 436
A-3.7 Error AnalysiR 438
A-3.8 Summary 439
Appendix 4 Flow Chart of the Computer Program Used for Determin-

ation of the Fatigue Strength Distribution 440
A-4.1 Explanation of the Symbols Used in the Flow Chart 441
A-4.2 Explanation of the Flow. Chart 443
vii
-! I
A-4.2.1 Part One - Conversion of Life Data to Strength

Data and Non-Linear Mode Variables to
Linear Mode Variables 443
A-4.2.2 Part Two - Determination of the Best Fitting
Distribution 444
A-4.2.3 Part Three - External function Fit 445
A-4.3 The Flow Chart 446
A-4.3.1 Part One 447
A-4.3.2 Part Two 448
A-4.3.3 External Function Fit 450
I.[
__ t
,,c.... , ,.. . ,j... t•'• . .

EVALUATION
1. This study ms concerned with the development of a practical

engineering tool to be used in predicting the reliability of mechanical
parts fabricated of nonferrous metals and subjected to dynanic loading.
The tool vas to be compatible with the Stress-Strength Interference
Theory as reported in RADC-TR-66-710 and be usable by design engineers
with limited statistical background.
2. The study has provided a prediction technique which is almost "cook-

book" in nature and requires a minimum amount of computation except as
required in the interpolation of the tabulated interference values. The
tabular values are highly nonlinear; thus, linear interpolation cannot
be used. Procedures for higher order interpolation are given and
illustrated. Within the limitation that the material must be a nonferrous
metal, the reliability of any part subjected to fatigue loading can be
predicted. Eminles of such parts would include springs, fasteners, shafts,
beams, torsion bars, vehicle frames and suspensions, forgings and castings.
The accuracy of the technique depends on the knowledge the user has of the
stress distribution, i.e., measured or estimated. "
3. The results of this study and of the earlier study of ferrous metals
will be combined and presented in the MADC Nonelectronic Reliability
Notebook.
Relability Engineering Section

Reliability Biaeh
_______________________________1.
"""
SECTION 1 INTRODUCTION
A complete cycle of reliability is made up of four stages; 1. Speci-

fication of Reliability; 2. Prediction of Reliability; 3. Verification of
Reliability; 4. Preservation of Reliability. The heart of the second stage,
namely, the Prediction of Reliability, is the Interference Theory.
The basic idea behind reliability is that a given part has certain
physical properties which, if exceeded, will result in failure. The factor
which may cause these properties to be exceeded is the " tress imposed by the
operating conditions. Thus, in prediction of reliability it is not the stress
alone or the strength alone that is the determining factor but the combined
effect of the two.
Early practices in stress-strength relationshis dealt almost entire-

ly along the lines of factors of safety. Once a part was designed and the
ratio of strength to stress was in the range of approximately 5 to 10, it was
considered to be safe for service. In certain industries and in certain
applications, factors of safety as high as 20:1 were employed.
The definition of the factors of safety varied from user to user,

depending on the sophistication and the complexity of the problem.
Some of these definitions, found in literature, are listed below:
Factor of Safety = Ultimate Strength

Nominal Stress
Factor of Safety - Yield Strength

Nominal Stress
Ultimate Strength
Factor of Safety - Actual Working Stress
Factor of Safety - Yield Strength

Actual Working Stress
Factor of Safety -Maximum Safe Load
Normal Load
Factor of Safety = Computed Strength

Computed Load
Margin of Safety Strength-Stress

Stress
Design Factor Strength

Design Stress
Factor of Utilization - Stress

Strength
1
Magnitude of Variable
Producing Failure
Functional Reserve Factor Magnitude of Variable
at Operating Conditions
where the variable could be force,

power, torque, material, surface
finish, fillet radius, etc.
Apparently the factor of safety was meant to account for all the
variables which were known to affect the stress and strength of the member.
The utilization of a factor of safety of this kind has justification only when
its value is based on considerable experience, with parts not too different
from the one under consideration. However, when substantial changes in the
geometry, the processing, or the function of the part are contemplated, a
major error may result if the old factor of safety is projected to the new
set of conditions.
This can be illustrated by the problem of automotive axle shafts
which were failing in service in large numbers. These shafts had Leen
fabricated from a steel with a tensile strength of 240,000 psi, and yet, the
operating stresses as measured in actual service were f-und to be only 13,000
psi. This produced an apparent factor of safety of 240,000/13,000 - 18.5.
This is obviously a fictitious value, since the shafts were failing in ser-
vice, and the true factor of safety was less than one. The explanation lies
in -the fact that axle strength to be compared with the 13,000 psi operating
stress should not have been the ultimate strength of the material (240,000
psi) but the fatigue strength corresponding to the surface finish of the
shafts, the mode of loading to which the shaft was subjected, etc. When the
ultimate strength was reduced by these derating factors, the resultant value
was found to be 12,000 psi. This strength, when compared with the 13,000
psi stress, produced the realistic factor of safety of 0.9.
Examples such as this lead to the next phase in the relationship

between stress and strength, namely, to the concepts of a significant stress
and a significant strength. By significant stress is meant the actual stress
Imposed on the part, and it may include the effect of stress raisers, magnifi-
cation due to impact loading, residual stresses, etc. By significant strength
(generally, fatigue strength) is meant the actual strength of the part in its
fabricated form under actual operating conditions. A rational approach to
significant strength still employs ultimate strength as the basis. However,
instead of an indiscriminate grouping of all the factors affecting the ulti-
mate strength into one index, it attempts to evaluate quantitatively the
effect of each individual factor pertaining to the part and the conditions
under consideration. The result is a value which is strictly applicable to
the part under consideration and to the set of loading conditions to which the
part is subjected in service. The principal factors affecting strength and
which must be considered in determining the significant strength are: life
expectancy, type of loading (axial, bending, torsional, or a combination),
sie effect, surface finish, surface treatment, notch effects, mode of loading
(static, completely reversed dynamic, or a combination).
2
These concepts of significant stress and significant strength repre-
sent a major step toward a more realistic prediction of mliability and, as
such, they have been included in the present investigation. By themselves,
however, they are not sufficient. This is because the prevailing practice is
to use the mean values of the calculated strength and stress, ignoring the
natural scatter that stresses and strengths may have.
The variability in these two factors results in the existence of a

statistical distribution function of stress and strength (See Figure 2.1) and
is the heart of the Interference Theory. Thus, for proper prediction of relia-
bility, an estimate must be made of both the mean value and the dispersion
characteristics of both the strength and stress.
The strength of the part, as all prcperties of nor-homogeneous

materials, varies from specimen to specimen, in view of the variation in
hardness, surface finish, degree of stress concentration, etc. In our past
investigation ("Reliability Prediction - Mechanical Stress/Strength Interfer-
ence" by Charles Lipson, Narendra J. Sheth, and Ralph L. Disney, Final Report
RADC-TR-66-710, March 1967, referred to in the present report as Reference 1)
the problem of determining the variation in strength and subsequently its
distribution function was resolved by a graphical method. Among the several
possible distribution functions considered, only the Weibull distribution was
used because of its wide use and the difficulty of handling other distribu-
tions by a graphical method. In the present investigation a computer approach
was developed by which the determination of the distribution functions of
strength became considerably less difficult. (This is discussed in Section 6.)
Consequently, several distribution functions were tried. These were: Weibull,
Largest Extreme Value, Smallest Extreme Value, Logistic and Normal. The rea-
sons for selection of these particular distributions are given in Section 6.3.
It was later found that the Logistic Distribution had somewhat the same char-
acteristics as the Normal Distribution and, therefore it was not followed any
further. -The computer printed the degree of fit and the parameters for all
of the above distributions for a given set of strength data. The best fitting
distributions and their parameters were then chosen, and they are presented in
a tabular form in Appendix 1 for various materials and conditions studied.
The representative ones were then plotted in order to determine tbe effect of
various conditions on the strength of various materials. In order to put the
information on the same consistent basis, all the data were expressed in terms
of the same distribution (Normal) and the plots were made by expressing the
means (')and standard deviations (a) as a function of life (See Section 6).
The reasons for the selection of the Normal Distribution for these graphs are
given in Section 6.
As is the case of strength, the oisrating stresses vary too. These

stresses vary from time to time in a particular part, from part to part in a
particular design, and from environment to environment. Therefore, both the
mean value and the dispersion characteristics of stress and strength must be
determined.
Once the parameters of the strength and stress distributions are

found, percent interference and thus probability of failure can be determined
3
II
from the interference area (shaded area in Figure 2.1). Means of computing
these interferences represent one of the principal objectives of the present
investigation.
I :
L__
T
SECTION 2 INTERFERENCE THEORY
Suppose there are two barrels containing slips of paper, each having a
number printed on it. The numbers in barrel Y are distributed according to
distribution Y, as in Figure 2.1, and the numbers in barrel X are distributed
according to distribution X. If, at random, slips of paper from each barrel
are selected and paired, they may be classified into successes and failures.
A success is constituted by a strength value exceeding a stress value, as for
example, when xi > Yl. Failure will occur if x2 < Y2 as shown. It will be
noted that, although the shaded area is a measure of interference, it is not
interference itself: a pair of points x3 and Y3, although in the shaded area,
will not produce failure. By continued pairing of stresses and strengths at
random, pairs will be found where the stress will exceed the strength. By
continued experimentation a good estimate of the probability of interference
can be found.
2.1 TWO NORMAL DISTRIBUTIONS

From an exhaustive survey of literature made during the past investiga-
tion(1)l it was found that most studies have assumed both the stress and the
strength distributions to be normal. This is a natural assumption to make in
order to solve a practical problem, as no work was found dealing with an ana-
lytical expression for the interference of two non-normal distributions.
When the stress and strength distributions are assumed to be normal,

2
the probability of interference can be'determined from the equation( ):
Zp4
x0 y
X (2,1)
where y = mean stress
ox W mean strength
ox - strength variance
2
o2 stress variance
y
z - standardized normal variate determined from
standard tables (See Table 2.1)
Tus, if the average stress is 30 ksi with a standard deviation of 3 ki i

and the average strength is 50 kui with standard deviation of 10 kal, z - 1.91
and from Table 2.1 c, which represents interference, in found to be .0281.
Thus, the percent interference (the probability of failure) is 2.81%.
* Numbers in parenthesis designate References listed on Page 149.
\\ 5
¥ Applied x Strength
Stress
,U
U'
S..
u Measure of
S~Interference--
X'2 •Y 1 2
Yl xJ
3 xl
Stress (Strengthl)
Figure 2.1 Interference of Stress and Strength Distributions
. iii
NORMAL DISTRIBUTION
Tabulation of the values of a versus K. for the

Standardized Normal Curve. 2
0 _ z__Z
CY~zKc =1 2
= Area under the Standardized Normal Curve

from z = Kc to z nc
SZ 0
K f .00 .01 .02 .C3 .04 .0.5 .06 .07 .08 .09
0.0 WW.5000
.4960 .4920 .4880 .4840 .4801 .4761 .4721 .4681 .4641
0.1 .4802 .462 .4522 .4483 .4443 .4404 .4364 .4325 .4286 .4247
0.2 420 .4188 .4129 .4090 .4052 .4013 .3974 .39M8 .3897 .3809
0.3 .3821 .3783 .3745 .3707 .3669 .3632 .3594 .3557 .3520 .3483
0.4 .3448 .3409 .3372 .3336 .3300 .3264 .3228 .3192 .3158 .3121
0.5 .3085 .3050 .3015 .2961 .2948 .2912 .277" .2843 .2810 .2778
0.8 2743 .2709 .2678 .2643 .2611 .2578 .2564 .2514 .2483 .2451
0.7 .2420 .2389 .2358 .2327 .2296 .2286 .2238 .2206 .2177 .2149
0.11 .2119 .2090 .2061 .2033 .2005 .1977 .1949 .1922 .1894 .1867
0.9 .1841 .1814 .1788 .1782 .1736 .1711 .1685 .1660 .1635 .1011
1.0 .1587 .1562 .1539 .1515 .1492 .1469 .1446 .1423 .1401 .1379
1.1 .1357 .1335 .1314 .1292 .1271 .1251 .1230 .1210 .1190 .1170
1.2 .1151 .1131 .1112 .1093 .1075 .1056 .1038 .1020 .1003 .0965
1.3 .0968 .C951 .0934 .0918 .0001 .0885 .0969 .0853 .0638 .0823
1.4 .0808 .0793 .0778 .07e4 .0749 .0735 .0721 .0708 .0694 .0581
1.5 .0M8 .0655 .0643 .0630 .0618 .0608 .0504 .0582 .0571 .0559
1.6 .0548 .0537 .0528 .0516 .0505 .0495 .0485 .0475 .0465 .0455
1.7 .0446 .0436 .0427 .0418 .0409 .0401 .0392 .0384 .0375 .0367
1.8 .0359 .0351 .0344 .0336 .0329 .0322 .0314 .0307 .0301 .02
1.9 .027 .0281 .0274 .0268 .0262 .0258 .0250 .0244 .0239 .0233
2.0 0228 .0222 .0217 .0212 .0207 .0202 .0197 .0192 .0188 0183
2.1 .0179 .0174 .0170 .0166 .0182 .0158 .0154 .0150 .0146 .0143
2.2 .0139 .0136 .0132 .0129 .0125 .0122 .0119 .0116 .0113 .0110
2.3 .0107 .0104 .0102 . 00P.00964 .00939 .00914 .0089 .0066 .0 642
2.4 .00820 .00798 .00776 .00755 .00734 .00714 .00695 .86768 .00857 .00639
2.5 .WG21 .0080 .001587 .00570 .00554 .00539 .00523 .00508 .00494 .00480
2.6 .00466 .00'53 .00440 .00427 .00415 .00402 .00391 .00379 38 .003S57
2.7 .00347 .003381 .00326 .00317 W3 07O .00289 .00M .00272 .00M
2.8 .0025M .00248 .00240 .00233 .00226 .002219 .00212 .00205 .001 .00193
2.1 M.R17 .001811 .0075 .00I89
1 00114 .001591 00154 .00W14 .00144 .00139
Table 2.1 Normal Distribution(')

In practical applications of the Interference Theory the following pro-
blem arises: both distributions under consideration extend to plus and minus
infinity. It is apparent, therefore, that any two distributions will overlap
and cause interference. This, of course, is erroneous because some distribu-
tions, such as strength, must have a finite lower bound of zero. In many
situations the physical set-up and tbe sample size adjust for this lower bound.
For example, suppose a part is designed so that the mean of the strength dis-
tribution is placed 6 a away from the mean of the stress distribution, both
distributions having standard deviation equal to a. From Equation 2.1 it is
found:
6a = 4.24
and the probability of interference comes out to be .00001. This means that
only one part will fail in 100,000 parts produced. If actually only 50,000
parts are made, the physical problem has effectively truncated the distribu-
tions. However, the probability of failure of one part remains .00001. This
means that,due to sample size, the extreme portions of these distributions are
no longer important since the sample size is such that not even one failure
can be expected.
For the ease of application of the Interference Theory for the case of
two Normal distributions, a graph was constructed as shown in Figure 2.2. The
example solved previously through Equation (2.1) now yields:
50ks$-3Oksi a 6.67 and,

amin 3ksi
max 10ksi
- = 3.3
0
ain 3 ksi
and the percent interference (the probability of failure) comes out approx-
imately 3% as before.
2.2 CONSISTENCY OF TWO DISTRIBUTIONS
In the application of the Interference Theory, the following important

point must be considered: the distribution of stress and the distribution of
strength must be consistent with each other. In a fatigue test or in actual
application in service, a single part has a single fatigue life for a given
loading condition. Subsequent testing of additional parts under the same load
will show a scatter in life leading to a life distribution. Through more ex-
tensive tasting a strength distribution for a given life can be obtained, such
as the distribution in Figure 2.1 (the method of obtaining a strength distri-
bution from a life distribution is described -In Seetion 6).
It follows then that, for a consistent development, the stress distribu-

tion must be of the same nature as the strength distribution. That is, it
should represent the plot of the frequency of occurrence of an applied stress
versus- the applied stress. This Is _no the same as the stress distribution con-
ventionally derived from the spectrum of loading acting on the part. The con-
- ~~-
ST1
auzajIPOWj 90 AIT~fquqo.Ld
00 C) 4 r4
Ln C14 ,- - 1r C'J - 0i
- - 00
-4P
000,
0
r..
40 -4
-44
00
00
an.
cva.
43UOA*JA03u
0A JOi1..ut~l
version of one to the other must be accomplished through the use of the equi-
valent utress (Sequ) and Miner's or Corten-Dolan's Rules. This is the method
vised in the present investigation, as described in Section 7.
2.3 NON-NORMAL DISTRIBUTIONS

So far the discussion has been limited to the cases when both the stress
and strength distributions can be assumed to be normal. In cases when either

one or both are not normal the problem is much more involved. For example,
the intersection of a Normal and a Weibull distribution produces a distribu-
tion of an unknown origin.
In the past, problems such as this were solved largely through "brute
force," by a method commonly referred to as the Monte-Carlo Technique. Essen-
tially, the Monte-Carlo Technique consists of a sophisticated means of random-
ly selecting a sample from one distribution and comparing it with a random
sample taken from a different distribution. This is accomplished with the aid
of Tables of Random Numbers. The resultant paired data are plotted as a Cumu-
lative Distribution .Function on Normal, Weibull, etc. probability papers and
the percent interference is read from the graph.
2.4 THE INTEGRAL METHOD
In the present investigation a Method of Integrals was used in prefer-

ence to the Monte-Carlo Technique. This method involves determining, as an
integral, the expression for the interference of the two distributions Under
consideration and establishing percent interference from this integral.
The advantages of the Integral Method are:
1. For some distributions the integrals have been already tabulated

and percent interference can be read directly from the table.
2. In those cases where the integrals have not been already tabulated,
they can be evaluated by Numerical Analysis as done in the present
investigation.
3. The major shortcoming of the Monte-Carlo Technique is that it

requires a very large sample size for any accuracy. This short-
coming is avoided when the Method of Integrals is used.
4. One of the objectives of the present study is to develop and

evtluate an analytl.cal expression for interference of any two
distributions. Such an expression is possible when the Method of
Integrals is used but nQt when the Monte-Carlo Technique is employed
10
SECTION 3 OBJECTIVES OF THE PRESENT STUDY
The objectives of the study described in this report were:
1. To refine and to reduce to practice the Stress/Strength Inter-

ference Theory technique for designLng and predicting the quanti-
tative reliability of mechanical parts and components fabricated
from various non-ferrous materials under mechanical loading.
Maximum use was to be made of empiricalpractical engineering
values as well as a sound theory.
2. To study the effect of such factors as type of loading, surface

finish, temperature, heat treatment, stress concentration, surface
treatment, manufacturing processes, environment, etc. on the
statistical distribution of fatigue strength.
3. To determine, from the existing available empirical data, the

distribution of the fatigue strength under the effect of each
of the above factors.
4. To develop a computer approach [as contrasted with the graphical

approach of our past investigation(l)] for determining the sta-
tistical distribution function of fatigue strength at a given life.
This method has the advantage of high accuracy and time saving.
5. To develop the means of synthesizing the strength distribution

function when such function is non-time variant, i.e. infinite life
design, and when such function is time variant. i.e. finite life
design.
6. To develop an analytical expression of the distribution of inter-

ference for the general case where the two interfering distribu-
tions are different. In the present investigation the distributkos
studied were:
Stress Strenath
Normal Weibull
1A
t1
SECTION 4 STUDY APPROACH
4.1 LITERATURE SEARCH
In the past investigation on ferrous materials(l), an exhaustive liter-

ature search was made to determine the State of the Art in the field of Relia-
bility Prediction - Mechanical Stress/Strength Interference. Specific topics
covered were: Interference Theory, Mathematical Tools as related to the Inter-
ference Theory, Monte-Carlo Technique, Reliability Prediction, Fatigue of
Metals, Cumulative Damage, etc. As these topics were common to the past and
the present investigations, it was not necessary to repeat them in the current
study. Therefore, in the present project, no literature search was made.
4.2 THEORETICAL ANALYSIS
One of the objectives of this investigation was to develop and evaluate

an analytical expression for the interference of two distributions. When the
two distributions are Normal the interference can be simply expressed by a z-
distribution, as described in Section 2. From an extensive survey of litera-
ture made in our past investigation, no work was found dealing with the analy-
tical expression for interference when the two interfering distributions are
not Normal. The purpose of this phase of the investigation, then, was to
develop such expressions.
In the past investigation-the analytical expressions and, hence, the

tables of interference values, were developed by means of the Method of Inte-
grals as discussed in Section 2.4 of Reference I. These included:
Stress Distribution Strength Distribution
Normal Normal
Weibull Weibull
Normal Weibull
In the present investigation, data analysis (Section 6) revealed cases
where fatigue strength data followed distributions other than Weibull or Nor-
mal. Hence it was necessary to develop tables of interference values for
cases wherestreus distribution was Normal and strength distribution was Lar-
gSeat Extreme Value and, also, stress Normal and strength Smallest Extreme
Value. This was done by evaluating the complex integrals expressing the dis-
tribution of interference. Numerical analysis was carried out (Appendix 3)
using an IBM 360 computer to solve these integrals. Tables were then prepared
for the interference values as a function of the distribution parameters
(Appendix 2).
In addition, in the present investigation it was necessary to extend

the scope of study for the case when the stress distribution is Normal and the
strangth distribution is Weibull. In the past investigation on ferrous mate-
rialsl(), the study was limited to the cases where the stress-strength inter-
ference parameter. C
(C * strenlth index)
a stress index '
12
K.
was between 10 and 100. This was because the strength data on ferrous mate-
rials fell into this range of C values. In the present investigation on the
strength data of non-ferrous materials, C values fell between 1.0 and*q .0 in
many cases. Hence, it was necessary to extend the interference tables to in-
clude values of C between 1.0 and 10.0. Thus, tables of interference values
were prepared to include the combinations:

Normal Weibull (Extended)
These tables are given in Appendix 2.
4.3 EMPIRICAL DATA

The tables of interference values are the heart of the Interferen e
Theory as applied to engineering practice. Once the stress distribution and
strength distribution parameters are known, the percent interference, anq thus
the probability of failure,can be read directly from these tables (for t e pro.-
cedure see Section 9 and for the tables see Appendix 2).
Over 250 articles from literature and other sources were examined1 in
the past investigation, and practically no data were found concerning thi sta-
tistical distribution of stress. In the present investigation some furtjher
work was done in this area, but no additional information was found. TUe only
data located referred to spectrum of loads or stresses, which, as pointpd out
in Section 7, does not represent the stress distribution required for &e
Inturference Theory. Interference Tables, therefore, were constructed/so that,
for given dispersion characteristics of stresses, a percent interference can be
found. The range of these characteristics chosen here and corresponding to
engineering practice are (if the stress distribution is Normal): /
/
.01 .10
To determine the distribution of strength it was necessary to collect a

great deal of data in order to arrive at a meaningful distribution. To sys-
tematize the effort of collecting data, a format was prepared which included
the factors which are known to affect the final distribution bf strength. An
attempt was made to collect data in different areas in order to determine the
effect of such factors as type of loading, size, processes," surface conditions,
heat treatment, surface environment, temperature, surface treatnent, stress
concentration, etc.
The Mechanical Properties Data Center in Traverse City, Michigan, was
found to be a very useful source of information for the fatigue strength data.
They have been very cooperative in providing the necessary information. An
exhaustive collection of data on the fatigue strength of non-ferrous materials
was made from the Mechanical Properties Data Center. These data were then
13
systematized, evaluated in terms of thT distribution parameters (Section 6),
tabulated (Appendix 1), and the representative ones plotted (Section 6).
4.4 FACTORS AFFECTING THE STATISTICAL DISTRIBUTION OF FATIGUE STRENGTH
Since fatigue strength represents the major interest in the engineering

applications of the Interference Theory, this problem was studied in some de-
tail. The statistica.l distribution of the fatigue strength of a mechanical
component is a function of a number of factors, such as type of loading, sur-
face finish, stress concentration, heat treatment, temperature, environment,
size of specimen, frequency of loading, surface treatment, manufacturing pro-
cesses, grain size, impurities, and time. Each shows variability which is
characterized by some form of a distribution. The effects of these factors on
the statistical distribution of strength were studied in the present investi-
gation.
Fatigue strength can be defined as the maximum stress that can be sus-
tained for a specified number of cycles without failure, the stress being com-
pletely reversed within each cycle. In the case of steels, a component is
said to have finite fatigue strength if it fails between 10 3 and 106 or 107
cycles due to a given magnitude of cyclic load. For non-ferrous materials this
is generally extended to 5 x 108 cycles. In the present investigation the
data were extrapolated to 108 cycles.
Type of Loading: The three major types of fluctuating load encountered

in designing parts are axial, bending, and torsion. Experimentally determined
values-of the ratio of average fatigue strength for axial loading as compared
to bending load were reported in literature as ranging generally from 0.75 to
1.0.(49) Although a jreat deal of work has been done to obtain precise values
for this ratio, no detailed study has ever been made as to the statistical as-
pects of these strengths. Investigations have been conducted to find statis-
tical distributions (Normal, Exponential, Weibull, etc.) of fatigue strength
tested under a Aivej type of loading, such as bending. No work was done to
determine the effect on the distribution if the loads were other than bending.
In the present investigation, an attempt was made to study the effect of dif-
ferent loads on the statistical distribution of the fatigue strength. The
statistical parameters of the distribution for various materials under differ-
ant loads were determined, tabulated according to materials (Appendix 1), and
plotted in Section 6.
Surface Finish: The surface finish of a part does affect its endur-
ance strength. Hence, the condition of finish should be taken into account
when the design is based on fatigue. Surfaces which have an effect on the
significant strength can be classified into five broad categories: polished,
ground, machined, hot-rolled, and as-forged. The worse the surface condi-
tion the lower will be the mean fatigue strength but the higher will be the
scatter. As a result, the degree .of interference is likely to be pronoun-
cedly affected by the type of surface finish imparted to the member. Dif-
ferent surface effects were studied in the present investigation and the
fatigue strensth parameters were Labulated (Appendix 1) and plotted in
Section 6.
14
Stress Concentration: A notch or a stress raiser iD a part subjected
to fatigue loading can be regarded as a factor causing a local increase in
stress or a reduction in strength. For example, a notch with a stress concen-
tration factor of 2 can be thoughtof as doubling the stress or as halving the
strength. In the present investigation this factor was taken as a strength
reduction factor.
If all parts were made of materials which are completely homogeneous

and have perfectly polished surface finishes, the effest of a notch would be
to increase the stress by the factor Kt. Since actual materials are not per-
fectly homogeneous and actual surfaces are seldom perfectly polished, there
exist internal and surface stress raisers. For this reason, the addition of a
notch to a part, already havixw stress concentration due to geometry, generally
produces a smaller effect than would be predicted from the theoretical stress
concentration factor, Kt. The extent to which a notch reduces the endurance
limit of a part is referred to as the fatigue stress concentration factor, or
the fatigue strength reduction factor, and is designated by the symbol Kf.
This is defined as:(6)
Kf - endurance limit of specimen without the notch

endurance limit of specimen with the notch
In this study, an attempt w;,z made to determine the effect of stress

concentration on the statistical scatter of the fatigue strength. More speci-
fically, the objective was to find out whether this factor chauges the mean
strength only, whether it has an effect or. the standard deviation, or whether
it completely changes the nature of the distribution itself. The data were
collected for various materials, various stress concentrations, and at differ-
ent temperatures. Changes in the parameters of the statistical discribution
of the strength due to the effect of stress concentration for different mate-
rials at various testing temperatures are tabulated in Appendix 1 and plotted
in Section 6.
Heat Treatment: Different heat treatments such as annealing, solution

heat treating, normalizing, quenching, tempering, aging, etc., can be imparted
to materials to change their mechanical properties. Heat treatment may change
the average fatigue strength but also the statistical scatter. Pertinent
parameters are tabulated in Appendix 1 and plotted in Section 6.
Temperature: In a similar manner the effect of temperature on the sta-

tistical scatter of fatigue strength is shown in Appendix 1 and Section 6.
Environment: The effect of environment on the fatigue strength was

confined in this investigation to the corrosive environment since most of the
empirical strength data was in this area. It is known that a corrosive envi-
ronment has a significant effect on the fatigue strength, particularly at a
higher life. The statistical distribution parameters for various non-ferrous
materials under different corrosive environments were determined, tabulated
according to materials (Appendix 1), and plotted (Section 6).
15
Other Factors: In the present study it was observed that size and fre-
quency of loading do not have any consistent effect on the distribution of the
fatigue strength, whereas surface treatment, manufacturing processes, and im-
purities do exhibit consistent effects. The reason is that the latter may
change the microscopic structure which, in turn, is related to strength. The
statistical distribution parameters for various non-ferrous materials under
the above conditions were determined, tabulated (Appendix 1) and plotted
(Section 6).
4.5 COMPUTER ANALYSIS OF STRENGTH DATA
Data collected during this phase of the investigation were organized

and systematized according to materials and conditions. A computer approach
was developed to analyze these data (Section 6). In the past investigation
using the graphical approach, only the Weibull distribution was used in the
data analysis for the reasons discussed in Reference 1. In the present inves-
tigation, with the aid of the computer, other distributions besides Weibull,
such as Largest Extreme Value, Smallest Extreme Value, and Normal, were tried.
The reasons for the selection of these particular distributions are given in
Section 6.3. The best fittirg distribution and its parameters were determined
for various materials and conditions and these are tabulated according to
materials in Appendix 1.
16
I
SECTION 5 ANALYTICAL EXPRESSIONS FOR INTERFERENCE
5.1 INTRODUCTION
5.1.1 Interference Probabilities
In interference theory one supposes that the strength of a manufactured

part is not known with certainty prior to performing some test on it and that
the stress induced by a load is not known with certainty prior to actually
loading the part. Thus, for example, one does not know with certainty that
the strength of a part is exactly 50 ksi. He may know that the part cannot
have a strength greater than 58 ksi or less than 40 ksi. Or he may know that
the average strength that has been obtained in previous tests on these parts
is 49 ksi. He may have some measure of how dispersed the strength measures
are around this average strength. The point, of course, is that this type of
knowledge is quite different from knowing precisely what the strength is prior
to testing. For a multiplicity of reasons, strengths of seemingly identical
parts are not exactly the same, and precisely what strength a part will have
cannot be known until some type of strength test is performed. In the theory
of probability one says that the strength of a part is a random variable. Cer-
tainly the same type of reasoning applies to the stress. Thus for a mathema-
tical theory of interference one starts with the idea that strength is a ran-
dom variable, say X and stress 'is a random variable, Y.
In describing the properties of random variables, since their values

are not known exactly, one supposes that, associated with every set of values
that the random variable can take, there is a real number called the probabil-
ity that the random variable takes values in the set. These probabilities
are non-negative real numbers,, they are all less than 1 and in the sense given
below they "sum" to 1.
If x is any real number then there is a probability that the random

variable takes some value less than or equal to x. Symbolically,
Pr(X I x)
is a number such that 0 • Pr(X S x) S 1. Surely (i.e. with probability 1)

X s so that
Pr(X -
and
Pr(-- ' X) - 0.
Clearly Pr(X S x) depends on the real number x. Consequently one

defines a crobability distribution funcrton, F(x), by the relation
(x) - Pr(x S x).
17
2.'.
One sees immediately that
0 A F(x) < 1,
F(-w) -1
F 0
and that F(x) is a non-decreasing function.
In most engineering applications F(x) has a derivative for every value

of x, and one defines the probability density function, f(x), by
f(x) . dF(x)
dx
One takes f(x)dx - Pr(x < X I x + dx).
(i.e. the probability density function multiplied by dx is the probability

that X takes values in the neighborhood of x.) Since
f(x)dx = dF(x)
one has
f f(x)dx = dF(x) F(x)

-00 -10
the probabilities given by f(x)dx "*sum" to 1.
In the mathematical theory of interference one assumes that the proba-

bility density function for the random variables X(strength) and Y (stress)
are known. Sections 6 and 7 following show how these functions caai be found
from engineering data. Thus the "givens" of the mathematical theory of inter-
ference use the random variables X and Y, the set of values that they each can
take (usually the non-negative real line), and the probability density or dis-
tribution function F(x) [or f(x)] and G(y) [or g(y)].
The problem to which interference theory addresses itself is that of

finding the probability of failure. Failure is said to occur whenever the
stress exceeds strength. Thus from the known probabilities for the X and Y
random variables one wishes to find
Pr(Y b X)
which is the probability that stress exceeds strength or the probability of
failure.
5.1.2 Calculation of Probabilities of Failure
There are two useful ways of determining the probability of failure

from the known properties of X and Y.
18
(a) Since one wishes to find Pr(Y • X) it is convenient in some cases
(e.g. when stress and strength are normally distributed) to define a new ran-
dom valbable, Z, by the relation
Then if one can find the probability density function of Z, h(z), the proba-
bility of failure will be simply the probability that z S 0. In terms of h(z)
this is found by
Pr(failure)- h(z)dz.
The problem in general is then to find h(z) from the known probability density
functions f(x), g(y). A complete discussion of this method and its applica-
tions is found in SectioA A-3.2 of Reference [1].
In the important special case in which both stress and strength are
normally distributed random variables it is well known that Z is also normally
distributed with parameters.
"UZ Ux - Uy
and a2 Ma2 +a2
Consequently the probability of failure can be found directly from tables of

the normal curve areas. One wants the area from - to 0 from these tables.
A complete discussion of how to do this is given in Section 2.
(b) For most applications method (a) above is unnecessarily complex

because one must first find the entire density function of the random variable
Z before finding the probability of failure. Since the random variable Z is
of no practical value for Z < 0, the approach in part (a) is unduly long. The
methods described in this section are more direct and, from our experience,
more useful in general.
One can derive the probability of failure as follows. Suppose we super-

impose the stress and strength density function on the same graph as shown in
Figure 5.1.
Although Y is a random variable, let us fix attention on a particular,

small interval that T can take values in. Let us fix y Y y + dy. Then
let us find the probability that the random variable X takes values less than
this fixed Y. One can show that this probability is
Pr(X yY
y <yY y + dy) -u f(x)dx
The left hand side of this expression is called a conditional urobabilitv. It

is the probability that the random variable X takes values lose than the real
number y when it is known ("given that") the random variable Y is "nearly" y.
19
i ,;,
Stress Density
Function
Y Strength Density
Function
yx
Interference-I
• •.1
x or y
Figure 5.1 Interference of Stress and Strength Distributions
20 "
By the definition of f(x), this probability is obviously the same as the right
hand side of the expression. If now we multiply this conditional probability
by . \
"Pr(y < Y < y + dy) g(y)dy,
we obtain the oint probability that X • y and y. < Y - y + dy which symboli-

cally is Pr(X S y I y < Y S y +
dy) Pr(y < Y S y + dy)
"aPr(X - y; y < Y S y + dy) .
This probability is given by the integral
Zo f(x)dx g(y)dy
From the joint probability one obtains the probability of failure by:
Pr(failure) -o Jo f(x) g(y) dx dy
Thus this double integral gives the probability of failure directly.
If one recalls the relation between f(x) and F(x), it is clear that the
double integral is easily reduced to the single integral
Pr(failure) * F(y) g(y) dv.
If F(x) is easily obtainable then this expression is easier to work with than
the double integral. For example, one knows that for a strength with a Weibull
probability density function the probability distribution function F(x) is
Riven by
-Xx ý
F(x) 1- e Q•-X
In these cases the probability of failure io then given by •

3(0 ol
Pr(fallure) - (1- 0 ) g(y) d
I S(y) dy
The latter expression follows because
S(y)dy 1
21
Ii
if the random variable Y takes only positive values which is the usual case in
interference theory.
5.1.3 Interference Tables, ApDendix 2
In Section 5.1.2 it was shown that the probability of failure could be

expressed as an integral involving the known probability density or distribu-
tion functions. In certain cases this integral can be evaluated in closed
form (e.g. when f(x) and g(y) are both exponential functions). In some cases
this integral can be evaluated in terms of other well known and tabulated
functions (e.g. when f(x) and g(y) are both gamma functions or when f(x) and
g(y) are both normal functions). In general it is not to be expected that the
integral for the probability of failure can be evaluated in closed form or in
a form involving other well known functions. (e.g. when f(x) and g(y) are
both Weibull functions or when f(x) is a Weibull function and g(y) is a normal
function.) In those cases one must resort to numerical evaluation of the
integral.
Since the integrals giving the probability of failure cannot be ex-

pressed in terms of well known functions, in general, we have evaluated the
integral numerically. Tables of the probability of failure are given in Sec-
tion A-2 of [l]. A full discussion of the numerical methods used and the
errors of approximation appropriate to the tables are given in Section A-4 of
Additionally, one is interested in the case in which the strength dis-

tribution is of the form
1 - e when 1 > 0,
x 1x+=
called the Smallest Extreme Value probability distribution function. One is

also interested in the case in which strength is distributed as
-8(x-M)
e when 8 > 0,
which is called the Largest Extreme Value probability distribution function.
Section A-2.2 givesthe probability of failure for the case in which

strength is Weibull distributed aid stress is normally distributed,
Section A-2.5 gives the probability of failure for the case in which
strength is distributed with the Smallest Extreme Value distribution function
and stress is normally distributed.
22
22 "
___________________
Section A-2.4 gives the probability of failure for the case in which
strength is distributed with the Largest Extreme Value distribution function
and stress is normally distributed.
5.2 USE OF INTERFERENCE TABLES IN APPENDIX 2
5.2.1 Parameters for the Weibull Distributed Strength, Normal Distributed
The form for the integral involved in finding the probability of failure
when the strength is Weibtll distributed and the stress is normally distribu-
ted is given in Section A-3.3.5 of [1]. Tables of these probabilities are
given in Section A-2.1 of [1]. A discussion of the numerical analysis, error
and accuracy of the tables is given in Section A-4.1.7 of [11.
The form of the distribution of the strength has been given in Section
5.2.1 of [i].
In the past investigation, extensive tables of interference values

(Secti.on A-2.1.2 of [1]) were prepared for the parameter C (as defined below)
equal to 10 or above. In the present investigation on non-ferrous materials,
the values of parameter C much lower than 10 were needed. Hence, the Tables
A-2.1.2 of (1] were extended to cover the range of 1 1 C < 10. Therefore, the
Tables of this report list probabilities of failure for
B(x) - 1, 1.2, 1.3, ... 3.2
C - 1., 2., ... 10.
A - 0, .2, .4, ... 2.8 and -. 2 to -10.0
where
b - the slope of the strength distribution. In the
x table3 this is called B(x).
(OxXo )/0 - the ratio of the difference of the characteristic

strength and the truncation parameter (or lower
bound of strength) to the standard deviation of the
stress. In the tables this is called C for typo-
graphical simplicity.
-)/a - the difference between the strength truncation para-
meter and the mean stress divided by the standard
deviation of the stress. In the tables this is
called A for simplicity.
The values in the body of the table are the probabilities of failure
for the parameters given at the heading of the tables. From the Z1scussion
given in Section A-4.1.7 of (1] these probabilities are correct to 65 x 10-5.
5.2.2 Parameters for the Smallest Extreme Value Distributed Strenath,

Normal Distributed Stress.
In this came it has been found that the integral giving the probability
of failure can be expressed in terms of two parameters which are denoted by
23
L
I. Ii
a and Y. In the tables these parameters are given by
y - B(U-M) - the difference in the location parameter (p) for

the normally distributed stress and the mode (M) for the
Smallest Extreme Value distributed strength. This differ-
ence is multiplied by 8, the slope parameter of the extreme
value distribution.
a = 8- the product of the standard deviation of the Normal

distribution for the stress variable and the slope para-
meter for the Smallest Extreme Value distributed strength.
The tables give the probability of failure for
a: .001, .005, .025, .05, .075, and .1 to 3.0 in

steps of 0.1 and 3.5 to 10 in steps of 0.5.
y: 0 to -40 in steps of 1.0
The values in the tables give the probability of failure corresponding

to the values of a,y given at the column and row heals. These probabilities
are correct to within *2 x 10-4 according to the discussion given in Section
A-3.
5.2.3 Parameters for the Largest Extreme Value Distributed Strength,
Normal Distributed Stress
In this case, the integral expressing the probability of failure is

almost identical to that in the previous case (Section 5.2.2). The defini-
tions of the parameters a and y are identical. The tables give probabilities
of failure for the ranges of a and y listed in Section 5.2.2. These proba-
bilities are also correct to *2 x 10-4 as discussed in Section A-3.
5.2.4 Use of the Tables, Explanation of Missing Values and Interpolation
Numerical examples of the use of the tables are given in Section 9. In

general the user will enter the table with known parameters, for example, b
ex, x 0 vand the appropriate parameters for the stress distribution, and wishx
to fi'd the probability of failure. This is a direct table look-up. In some
design problems the user will have a given probability of failure to achieve
and will know the general shape of the distribution of stress and strength
appropriate to the material that he Is using. The table will then give him
the relative parameters (there may be many of these) to design for. It would
be expected that a cost analysis would give the acceptable parameter ialues
for each distribution. As long as the relative values are as given in the
table, the probability of failure will be the same no matter what the values
for each distribution are.
In the Weibull Strength-Normal Stress tables, the missing values within
the ranges of the parameters used are nemrly zero and hence have not been tabu-
lated. When A - (xo-ia)/o ) 3.5, the probability of failure is less than 1 x
10-4 since the area under the normal curve from 3.5 to * is less than 3 x 10-4.
24,
___________ -i
For the Normal-Extreme Value cases, the probabilities of failure are
zero (to 4-places) whenever Y 4-39 and a<l10.
When 0 = 8O = 0 (or a = 0), then the stress distribution for either

the Normal-Smallest Extreme Value or Normal-Largest Extreme Value case is a
straight-line distribution, and the probabilities of failure can be computed
as in Appendix 2. These formulas are repeated here since they give very good
approximations for the probabilities even when a $ 0 but is small (say a4.001).
For a = 0:
- ~ Y
Pr GailureJ = 1 - e-e
(Normal-Smallest)
Y wherey = 0 (-M)
..
Pr failure - e-e
(Normal-Largest)
It can be seen that the tables are non-linear for almost all values
of the parameters. This can cause inaccuracies when the tables are inter-
polated. The absolute value of the interpolation error depends on which
tables are interpolated. For precise values the user should use a higher
order interpolation formula (as given in Section A-2.1) rather than linear
interpolation. We have not explored the relative errors of interpolation
closely. In those cases checked, the relative errors are small.
25
I.'
SECTION 6 STATISTICAL DISTRIBUTION OF FATIGUE STRENGTH
BY COMPUTER APPROACH
Most fatigue testing involves subjecting specimens or parts to a fluctuatinE

stress to failure and repeating this process at various stress levels. The
data thus obtained, known as life data, are used to construct the conventional
S-N diagram. In this case, the scatter obtained is the scatter iii life at a
given stress. In the present investigation the attention was focused on the
nature of the scatter in fatigue strength at a given life. In the past inves-
tigation(l), where the study was made mainly on the ferrous materials, the
problem of determining the scatter in the fatigue strength and, subsequently,
its distribution function was resolved by a graphical method discussed in
Reference 1. Among the several possible distribution functions that were con-
sidered, only the Weibull distribution was used because of its wide use and
the difficulty of handling other distributions by a graphical method. In the
present investigation on the non-ferrous materials a computer approach was
developed by which the determination of the distribution functions of strength
became considerably less difficult. Consequently, it was possible to try out
such other distributionsas Largest Extreme Value, Smallest Extreme Value,
Logistic, and Normal besides the Weibull which was tried in the past investi-
gation. The rearons for selection of these particular distrtbutions are given
in Section 6.3. The computer program was so developed that when the raw data
(conventional S-N type data) were fed into this program, the computer printed
the degree of fit and the parameters for all the above distributions for a
given set of strength data. In the first part of this program the scatter in
the life was converted into the scatter in the fatigue strength. In the se-
cond part, the distribution functions (Weibull, Normal, Logistic, Largest Ex-
treme Value, and Smallest Extreme Value) were fitted to these strengths data
simultaneously, and the one with the highest degree of fit was the best fittng
distribution. The detailed discussion on this computer approach and the L.e-
cessary statistical tools that were used in developing the approach are given
in the following sections. The flow chart of the computer program uaed to
determine the fatigue strength distribution is Riven in Anpend$,x 4.
6.1 LEAST SQUARES METHOD OF FITTING A LINE TO S-N TYPE DATA
6.1.1 Method of Least Squares
The principle underlying the "best fit" of a line by the least squares
method is that the sum of the squares of the deviations in the y-direction of
the data points (ZI, Yi) from the most probable line is minimum (See Figure
6.1). The equation of this line is:
Y- a+ OX (6.1)
where a is intercept on Y-auis
and 0 is the slope of the line.
The parameters a and B are found from the following relationships(2):
26
_ ____ ____
Least Squares Line
'3(Id' Minimum)
- deviation
(xi,!i)I
I
IQ
IS
Figure 6.1 Least Squares Line
I•
SI•••
n n n
nl l Y-Xj Xi Yj (6.2)
n n
I Yi-B xj
i n i(6.3)
where n - the total number of points
Xi - abscissas of points
Yi - ordinates of points
6.1.2 Application of Least Squares Method to S-N Data
When stress-life (S-N) data are plotted on a log-log paper or log (S)
and log (N) are plotted on a Cartesian paper, an approximately straight line
results. Hence, the S-N data are related by the equation:
in (S) - c + B in (N)
When in (S) and in (N) are substituted for Y and X in the Equations(6.2)
and (6.3), they reduce to:
n (lu Ni)(ln Sin N in S(
(ln Ni)2) n in Ni 2)
n n
[ (ln S) - inN14
1
a- i (6.5)
n
where a and B are the intercept and the slope of the S-N line (See
Figure 6.2)
n - the total number of test points
Ni - the ith life, corresponding to the ith stres, Si
28
inS 0
ln[SiS
in[S1 ] -- - - -- O C
th
iI point
I
I
ln[Ni] inN
Figure 6.2 inS - in N Diagram
6.1.3 Correlation Coefficient
Correlation Coefficient (R) is an index of the goodness of fit of the

least squares line to a set of data. In the previous section it was pointed
out that a straight line can be fitted to a set of data with known coordinates
for example, in (S) and in (N). This section deals with a computation of cor-
relation coefficient (R) which is a measure of the goodness of fit. The cor-
relation coefficient is defined as:
R - (0)2 x-• (6.6)

2
S
y
where B - the slope of the fitted line.
n n
I i ( Xi) 2 Sample variance of X
Sx(n-) (abscissas) (6.7)
n IY ! I 2 Sample variance of Y
Sn (n-2) (ordinates)
If the correlation coefficient (R) is 1.0, every data point falls on

the least square line and the data has the best fit. On the other hand, if
the correlation coefficient (R) is equal to zero, all the data are around a
circle and the data has very poor fit. (See PFiure 6.3)
29
y
Least Square Line
(a) best fit

when R 1.0
(b) very good fit

when R 1.0
(c) very poor fit

when R * 0
x
Figure 6.3 Goodness of Pit for Various Correlation Coefficients (R)
30
• ii
A
Five distributions (Weibull, Normal, Logistics, Largest Extreme Value,

and Smallest Extreme Value) were fitted to each set of the fatigue strength
data, and the values of the correlation coefficient (R) for each fitted distri-
bution were computed from Equation (6.6). The distribution having the largest
value of R was taken as the best fitting distribution.
6.2 CONVERSION OF LIFE DATA TO STRENGTH DATA
In the past investigation(1) on ferrous materials the conversion of life

data to strength data was accomplished graphically where the data were plotted
on the conventional S-N diagram and the least squares S-N curve was then fit-
ted to the test points. Passing through each point an S-N curve parallel to
the least squares curve was drawn. These made a.set of parallel S-N lines.
(See Figure 6.4). A vertical line was drawn at a given life, say N1 , inter-
secting the family of S-0 curves. The points of intersection $l, S2, S3 ...
represent the scatter of the fatigue strength at the life N1 (See Figure 6.4).
Here, it was assumed that to each specimen of the population can be attributed
an individual S-N curve, and that there exists for any population of specimens
(at fixed test conditions) a family of non-intersecting S-N curves which can
be determined with any desired accuracy, each curve corresponding to a given
probability.
In the present investigation on non-ferrous materials, this basic as-

sumption was held still true, and for the purpose of developing the computer
approach, it was necessary to develop an analytical approach for the above
graphical method. This is given in the following sections.
6.2.1 Computation of Parallel S-N Lines
The general form of the equation of the parallel S-N curves is:
In S - a + B In N (6.9)
where B is the slope of the lines as computed from Equation (6.4)
Sis the intercept of the line passing through a given point

(Ni, Sj)
and £ is the subscript which takes values from 1 to n where n is
the number of parallel lines
In order to determine the equation of each line, it was necessary to

compute the corresponding values of ai from the following equation:
a- (In Sd)- 0 (In Ni) . (6.10)
By substituting the values of S-M! 4ta, (el, NJ), (S2 1 N2 ), (03, N3 )..., In
Equation (6.10), a,, ce, o3... can be computed. Hence, using Equation (6.9),
a set of equations for each parallel line was obtained.
31
Least Squares S-N Curve
(u
Q
0
00
44
O4
Aj
IAlii
N1 Life, cycles (log scale)
Figure 6.4 S-N Diagram for converting Life Data

to Strength Data
32
InS - a, + lnn
inS:-x2 + 0inN
inS - a3 + a inN (6.11)
6.2.2 Computation of the Scatter of Strength at a Given Life
A life, say NI, was chosen at which the scatter of strength was desired.
By substituting in N1 in Equation (6.11), a set of values of in Sl, in S2,
in S3 ... , were obtained. By taking antilog of these values, the scatter of
fatigue strengths Si, S2 , S3 ... for the life N1 was obtained. This was re-
peated for several values of life Nj, and the resulting scatters of facigue
strengths, Sij, were stored by computer for further analysis.
Median ranks(2) were then assigned to the scatter of the fatigue
strength data, Sijp arranged in increasing order. The analytical expression
for the median ranks that was used for the computer programming was(7):
*-3
Median Rank -
n+ .4
where j - order number and
n total number of date p-ints.
6.3 DISTRIBUTION FUNCTIONS
The most common probability distributions that were found from litera-
turc. for expreising ttie fatigue strengths are: Weibull, Largest Extreme Value,
Smallest Extreme Value, Normal. and Logistic. These cover a cosiderable
variety of distribution patterns. For example, Weibull distribution generates
a wide variery of distributions by choosing different values-of its parameters,
Xo, 9, and b, (this is disctissed in the folloeinS section). Extreme .alue
distribution (Largest and Smallest) takes care of some of the unexplainable
phenomenon observed in connection with the fiacturing of materials under the
applied dynamic load where the puzzling question Is: Why is the strength of a
specimen considerably smaller or largai- than its expected value? The reason
for this difference in the values lies in the fact that there may uxist flaws
in, thu specimen that will weaken It. Normal distribution was used for the
obvious reason that it can be justly called the most common and well known
diseti-butlon. Log.stic distribution, which has somewhat the same charaeter-
istics as the Normal distribution, was tried, but later it was found from
th.- coiAputer results that Ncrmal fits the data just as well as Logistic.
Hence, in those cases where Logistic fitted the data best, the Normal was
used, instead, to express the fatigue strength. Besides, Normal distribution
is widely used and is easier to understand for engineers with little statis-
tical background. It is for these reasons that the discussion on the Log-
istic distribution does not appear in the following section. Other distrib-
utions are discussed 'elow.
33
6.3.1 Weibull Distribution
The Weibull equation is a three parameter function. The general expres-

sion for the Weibull density function is:
0!
\ b
.x -X
,~ b xjXOjb-1l 1 X0
px) = e , (6.12)
O-X
0 dOA
and the general expression for the cumulative distribution function is:
_x-X2 b
P(x) 1 -e , X0 'x d0 (6.13)
whcre, as used in this study,
x is the independent variable (fatigue strength)
X is the lower bound of fatigue strength

0
9 is the characteristic fatigue strength, where 63.2% of the

population have strengths less than or equal to this value(2)
b is the Weibull slope
6.3.2 Normal Distribution
Normal distribution is a tw'o parameter function. The general expression

for Normal density function is:
(xu2
p(x) -
1 e
2a2
2 a x -d - (6.14)
where x is the independent variable (fatigue strength)
the pcpulation rvean Sis
a is the population standard deviation

By p simple transformation, this can be reduced to:
i2
P(y) e dz (6.15)
34
where z = standardized normal variate
a
and P(y) = probability of z A y.
6.3.3 Extreme Value Distributions
Gumble(8) designed what is known as the Extreme Value Distribution

which is generally used to estimate the tail-ends of the frequency distribu-
tions. This distribution dctermines the probability of occurrence of an event
Largest values or smallest values of an event can be estimated by Largest Ex-
treme Value distribution or Smallest Extreme Value distribution respectively.
The general expression for the density function is:
Largest Extreme Value (L.E.V.):
(x-M) (x-M) (6.16)
a-e (x-M) - ) +6 (6.17)

p(x) - 0 e- 0 (x-M) - e
+0 (x-M)
Smallest Extreme Value (S.E.V.):
84+(x-M) e+- (x-M)
(x-M) . Pa + --1- x (6.18)
The general expression for the cumulative distribution function is:
Largest Extreme Value (L.E.V.):
P W •-e -0(-)(6.19)
Smallest Extreme Value (S.E.V.):
P - 1 - e-a+ XM (6.20)
where =intensity function (slope)
M -Extreme Value (mode)
8 and M are also called the extremal parameters. Comparison of characteris-

tics of Extreme Value distribution with Normal distribution in given in Table
6.1 which is self-explanatory.
35
Characteristic Largest Value Smallest Value Normal Varia
A. Distribution
1. Cumulative probability func- -y y 1 2

tion for reduced variate, 0 e 1~-eyv-
2. Reduced variate y- a(x-u) y= c(x-u) z-.(x-u)
3. Original variate, x, in
terms of reduced variate x'u+y/a x-u+y/--
B. Reduced Variate y
4. Moment generating function 1i2

G(t) - -- - - - - -- -- r (l-t) r (l+t) e'T
5. Mode ------------------- 0 =0
22
6. Mean -y-0.577 -0.57722 2-0
7. Median Y ---------------- -lglg2-036651 -0.36651 i-O
8. Standard deviation a -------- l/,-1.28255 1.28255 1
9. Skewness 8---------------- 1.29857 -1.29857 0
10. Kurtosis 2----------- 27/5 27/5 3
C. Deviates yo Corresponding to Given Probability Points 0
11. 0 U .95 --------- 2.97020 1.09718 1.64485
12. 0 n .99------------------ 4.60015 1.52718 2.32635
Table 6.1 Comparison of Characteristics of Extreme Value Distribution

8
and Normal Distribution( )
36
.____________ ________________________
Characteristic Largest Value Sma3lest Value Normal Variatel
D. Area PKa Included Within Mode - Ka
-Kr 0 +Kr y
13. K-I; PKa=4(a)-0(-o) ------------ 0.73064 0.73064 0.68269
14. K-2; PKa=0( 2 a)-0(- 2 o) -------- 0.92597 0.92597 0.95450
15. K=3; PKo-0( 3o)-0(- 3a) --------- 0.97890 0.97890 0.99730
16. Variate y for which Py-0.68269 1.14071 1.14071 1.00000

17. Variate for which Py-0.95450- 3.06685 3.06685 2.00000
E. Original Variate, x
18. Mode R------------------------ u -u
19. Mean R--------------------- uy/ -u-Y/c x+aX=
20. Median I------------------- u+0.36651/ a -u-0.36651/ac

1 1w 1 1 1.
21. Standard deviation ax-- -.-.. a-.---
Table 6.1 (Continued) Comparison of Characteristics Extreme

Value Distribution and Normal Distribution(j
37
%V
6.4 THE LINEARIZING TRANSFORMATION EQUATIONS OF THE DISTRIBUTION FUNCTIONS
When x, the independent variable (in this case the fatigue strength),
and P, the dependent variable (in this case the cumulative probability of the
fatigue strength), are plotted on a Cartesian graph paper, the resultant plot
is non-linear (sigmoidal or S-shaped plot) as shown in Figure 6.5.
Such a curve is difficult to work with. To fit this curve to a set of

data it is necessary to use the curvilinear regression analysis. This analy-
sis is quite involved as compared to the linear regression analysis which can
be used to fit only the straight lines. Hence, if the non-linear function can
somehow be transformed into a linear function, the linear regression analysis
can be employed. This can be done by transforming the dependent (P) and inde-
pendent (x) variables of the function to a new set of dependent [T(P)] and
independent [G(x)] variables. This process is called the Linearizing Trans-
formation. The functional relationships between the variables (independent
and dependent) of the sigmoidal form of the distribution (Figure 6.5) and
variables of the linear (Figure 6.6) form of the distribution are called the
linearizing transformation equations. These are given as:
X - G(x) - independent variable plotted on abscissa

Y - T(P) - dependent variable plotted on ordinate
The linearizing transformation is used in many engineering applications. For

example, in strength-life (S-N) diagrams with the ordinate and the abscissa
having logarithm scale;
Y ln S
and X - ln N are the linearizing transformation equations where
ln S and ln N are called the linear mode variables and S and N are the origi-
nal variables. Or, in case of Weibull distribution, nln x and ln (x-Xo)
are the linear mode variables where as P(x) and x are the original (or non-
linear mode) variables. In the analysis of fatigue strength data, lineariza-
tion transformation can be stated as:
Y - T(P) (6.21)
X = G(x) (6.22)
where T and G are the linearizing transformation equations relating original

variables P and x to linear mode variable Y and X respectively. The functions
G and T should be such `that the inverse transform operation on them gives the
original var•4bles. 'These inverse transformation equations are:
P T (Y)
x G" (X).
The inverse transformation equations are important for finding the parameters
of original distribution functions. This will be discussed in the section
that follows.
I0
4.1
J'A "~4
Ln8 800n
0~~~~ ~ ~ 0 0 0 c*
9TVOS~~$. OTV IVUTUN-
839
fA &
C4 W
a 0 0 0
C)0
0 m Iccovi
C;0
STODSAWOUI
POTVA 9020
39N
r-.~ ~~~ A0..0.0...0 ...

Either or both the abscissa scale and the ordinate scale may be trans-
formed for linearizing the frequency distribution functions. In the following
sectiofi the linearizing transformation equations are derived.
The general expression for the Weibull cumulative distribution function,

Equation 6.13, is: xXo b
I- b--
p-l e 0 iX0
with original variables
P - probability of failure
x - strength
or, (x-Xo)b
1 0
i-p
lnln (,'i) - b in (x-Xo) - b in (Q-Xo) . (6.23)
This has a form of Y - OX + a where:

Y- nln (6.24)
i-p
X - in (x-Xo) (6.25)
8 - b (6.26)
cc= -b In (9-Xo) (6.27)
The Equations (6.24) and (6.25) respectively for ordinate and abscissa scales,
are the linearizing transformation equations for Weibull distribution. Hence,
if the new scales (ordinate and abscissa) are constructed corresponding to
these equations, the points following Weibull distribution should plot as a
straight line on these new scales.
In the case of Weibull distribution, the density function (Equation

6.12) can be integrated in the closed form to give the cumulative distribution
function (Equation 6.13),
x-Xo b
40 •[
whereas in the case of Normal distribution, it is not possible to integrate
the density function in the closed form to give Normal cumulative distribution
function. Hence, this is expressed as:
+• i
1 2o2
2a
P(x) - J• - e dx
Whereas, in the case of Weibull distribution, the Equation 6.13 can be trans-
formed into a straight line function (See Equation 6.23), there is no simple
transformation for the case of Normal distribution that can be used to relate
the original variable P(x) to a linear mode variablewhich is lnln --
I- in
the case of Weibull. Hence, the following method was used to linearize the
Normal distribution.
The general equation of .he linear form (straight line) of a distribu-
tion is:
Y - a + OX (6.28)
where X is some function of strength x [for example, X - ln (x-Xo) for Weibull]

and Y is some function of probability P(for example, Y - lnln ?L3 Since
Y is not a simple function of probability (P), one needs to find a new variable
which should satisfy the Equation (6.28), and it should be related to P in
such a manner that it can be used to find necessary values of this new variable
for any given corresponding value of P.
Such a variable can be the standardized normal variate, z. The stan-

dardized normal variate, z;
1. is related to P values and could be found directly from the Tables
of Normal distribution, such as given in Section 2, Table 2.1.
2. satisfien the Equation (6.28). This is shown below:
z is expressed as:
Z X--•
*
oro
This has the same form as Equation (6.28), where
Y z, - , J - a, and x -strength.
Hence, the linearizing transformation equations for Normal distribution are:
Y - z, the Standardized Normal Variate (6.29)
X x (6.30)
41
If the transformation just described does linearize the Normal distri-
bution function, it will plot as a straight line on a Cartesian paper. A
graphical example demonstrating the linearization of Normal distribution is
shown:
The Normal distribution function with a mean of P - 28 and a - 4.25 was

plotted on a Normal probability paper (See Figure 6.7) from which the follow-
ing values of x and its corresponding values of P were selected:
X P
15 .0015
20 .035
25 .25
35 .95
For the corresponding values of P, the values of z were found from the
Tables of Normal distribution:(9)
x P T{P)- z X-
15 .0015 2.97
20 .035 1.82
25 .025 0.67
35 .95 -1.65
The values of x on the abscissa and z on the ordinate were plotted on a Carte-
sian coordinate paper, and the resultant plot was found to be a straight line
(See Figure 6.8). Hence, this proved that this method did linearize the Nor-
mal distribution and the linearizing transformation equations are EquationR
(6.29) and (6.30);
y- Z
and X - x .
The general expression for the cumulative distribution function of the

Largest Extreme Value distribution is, from Equation 6.20):
-e (x-H)
By taking the ln twice, this equation reduces to a form of a straight line:
-lnln 1 Ox O (6.31)
or Y Ox + a
42
I44
TT
-44-~~
4fL1 d 4~
t 4
454
InA
4,
0
IN .
44
where Y -lnln(•) (6.32)
P
X =x (6.33)
a= -S (6.34)
Equations (6.32) and (6.33) are linearizing transformation equations

for Largest Extreme Value distributions. For Smallest Extreme Value distribu-
tions, the linearizing transformation Equations (6.35) and (6.36) were found
in a similar manner. They are:
Y - lnln (6.35)
X = x (6.36)
a = -8M (6.37)
6.5 FITTING THE LINEARIZED DISTRIBUTION TO THE STRENGTH DATA
In the previous sections, the parameters, the linearizing transforma-

tion equations, and the linear mode variables of the distribution functions
used In this study were presented. The best fitting distribution is the one
which has the highest correlation coefficient, R. In order to determine the
@
values of R for each distribution, they should be fitted to Probability-
Strength (P,x) data by fitting a least squares line (Y - a + SX) to the linear
mode variables of these data. This was done for all the distributions in the
following manner:
The original variables of the data are Probability and Strength (P and
7). From the linearizing transformation Equations (6.24) and (6.25);
1|
Y Inln 1-P
and X ln (x-Xo) , the linear mode variables are lnln
and 1n (x-Xo), where x is the original variable (strength), P is the cumulative
probability from the median rank table, Xo is the lower bound of strength, and
X and Y are the linear mode variables.
nlys-ubstitution of X and Y in Equations (6.2) and (6.3), the values of
the parameters a and B for the least squares line (Y - a + SX) were computed:
1 1 n n
n (Inln ) in(xiX 0 ) _ In (xN-X) (-XYnl
n i- [ln(x 1 -X 0 )1 2 -.- 22 (6.38)
!145 L.
n n
dnln aiI n (xi-X)
and iii i- n 1 (6.39)
In order to compute a and 8 it was first necessary to determine Xo from

the Probability-Strength (P, x) data. Once the X was determined, the values
of a and $ were computed, and the correlation coefficient R was found by sub-
stituting the values of a and 8 in Equation 6.6. In the past investigation(l),
X was determined by a graphical method; in the present investigation this was
done by a computer approach, as follows:
1. A value for X was assumed which should be somewhere between zero

and(1).the lowest strength value of the test data. Let this Xo be
X 0
2. Once the value of X0 is known or assumed, the original variables

of the data points can be transformed to the linear mode variables
by using the transformation equations
1
Y - lnln 1-P
X - ln (x-Xo).
3. By substitution of each value of X and Y corresponding to each

value of the original data points (P,x) in Equations (6.2) and (6.3),
the parameters a and 8 of the least squares line Y a + MX were
then computed.
4. Using the values of a and a, the value of correlation coefficient R

for this fit was computed from Equation (6.6). Let this correla-
tion coefficient be R(M).
5. Another Xo was then assumed close to (greater or less than) Xo(l).

Let this be Xo(2). The steps 1 through 4 were repeated. Let the
correlation coefficient thus computed be R(2).
6. The next value of Xo-X (3) was then computed in the following man-
ner: (See Figure (69
AR - (2)-R(l)
Whenever the ratio R =X(2)-XR(l) is positive, the value of
Xo(3) was assumed to be larger than Xo(2); and if the ratio is ne-
gative, the value of Xo(3) was assumed to be smaller than Xo(2).
The larger the standard deviation of -the strength data, the larger
will be the difference [Xo(3) - Xo(2)] that can be assumed.
7. After Xo(3) was found, the steps 2,3, end 4 were repeated; the cor-
responding value of correlation coefficient R(3) was computed.
S. This was repeated twelve tiaes, and twelve Xo'e and their corres-
ponding values of R were computed. (Note: The number of repeti-
tions, twelve, was selected arbitrarily).
46
_ ______.-- - - . . .. . •_ __ _ _ _ _ .. . _
0 A
4--r
00
".44
Out of these 12 lower bound X0 1 s, the one which gave the maximum correlation
coefficient (R) (the best fit) was taken as the "true" lower bound of strength
X0 . Once the true value of Xo was determined, this and the values of x and P
were substituted in Equations (6.38) and (6.39) to compute a and 0; ahd, sub-
sequently, the correlation coefficient R was determined from Equation (6.6).
To fit a Normal distribution to a set of data means fitting a straight

line (least squares line Y = a + $X) such that:
Y = z, the standardized Normal variate

and X = x, the fatigue strength.
The values of z correspond to the values of median ranks, and they were found
from the Normal Distribution Table.(9)
For example, if the total number of test points is three, their core-
sponding values of median ranks are (2): .2Q63, .5000 and .7937 as read from
the median ranks table. These values represent the areas under the Normal
distribution curve (See Figure 610). Hence, the values of z corresponding to
the values of the median ranks, .2063, .5000, and .7937 are -. 8194, 0, and
+.8194 as read from the Normal Distribution Table. The values of z or median
rank z values were found in the similar manner for various values of median
ranks. The table of median rank z values similar to the table of medipn ranks
was constructed, see Table 6.1. These tables of median ranks and median rank
z values, which are in matrix form, were stored into the computer for the pur-
pose of the data analysis. The values of X - x (fatigue strength) and Y - z
(the corresponding value of the standardized Normal variate) were substituted
in Equations 6.2 and 6.3, and * and B were computed to give the equation of
the least square line, Y - o + BX, fitted through these points.
The next step was then to determine the correlation coefficient R.

This was done using the Equation 6.6.
Since the Normal distribution is a little more involved than the other
distributions (Weibull and Extreme ValueX an example is given to illustrate
the procedure of fitting the Normal distribution to data.
Suppose the fatigue strength distribution data at 10 5 life cycles are:

27.0, 24.0, 32.5, and 29.0 ksi. (Although the number of test points were at
least 10 or more in our study, four test points are, given to make the correla-
tion simple)
The data were then arranged in an increasing order of value and the
appropriate median ranka z valueas for sample size n - 4 were read from Table
6.1 as follows.
48
_____________________- ______________- _____________
Area =.2063 Area -. 7937
(Median Rank) (Median Rank)
f (z)
-~-.8194 0 +.8194
z
Figure 6.10 Illustration for the Median Rank z Values
MEDIAN RANKS z VALUES
Order Sample Sizet n

Number,1 2 3 4 54
1 0.0000 -0.5450 -0.8194 -0.9982 -1.1290
2 +0.5450 0.0000 -0.2887 -0.4826

3 +0.8194 40.2887 0.0000
4 +0.9982 0.48i26
5 +1.1290
40
Table 6.2 Median Ranks a Values
* 49
I A ks i Median Ranks z Values
-t
24.0 -0.0982
27.0 -0.2887
29.0 +0.2887
32.5 +0.9982
By substitution of X - x and Y = z values in Equations (6.2) and (6.3),

they reduced to: n n n
nX zixi- I xi I zi
i=l i=l i-i
,n n xjj 2
in n X
and a i
n
The values of a and 0 were then computed to determine the equation of the
least squares line, Y a a + OX. Knowing the value of 0, the correlation coef-
ficient R was determined from Equation 6.6:
R. (0)2 x . (0)2 Sx~s
2 S2
Sy Sz
1
n
n Xj
2 ýj 2Samp
2~x e Variance
n I - 2 Sample Variance
and $y2 i-l of x
""and- n(n-l) (ordinate)
6.5.3 Extreme Value DistrLbutions
The linearizing transformation equations for Largest Extreme Value dis-

tribution (Equations 6.35 and 6.36) are:
Y -- ln3.
and X x
By substitution of X and T in Equations (6.2) and (6.3), the values of

the pawaoters a and B for the least square line (Y - a + OX) were computed:
50
_ __ -J-.-.---.-.-.---.
n (-nni) n n -
Sn -a
(-l nl
n
(x) -d (xi) - (-lnln
Tv (6.40)
n xi
n n
(-1njn
ln -_) - Z (xi)
and a= (6.41)
n
Knowing the value of 0, the correlation coefficient R was determined

from Equation (6.6).,
(0) 2s2
R-=
Sy 52 ofx(.2
n- x Sample Variance
2 = of x (6.42)
where x (n-1) (abscissa)
anV 2 i-i .- 1 (-ni ~JSample Variance

:1 and= ... . . Sn 2
n(n-l)
-
of y
(ordinate) (6.43)
where y - -lnln
In a similar manner, Smallest Extreme ValuA distribution was fitted to

the data such that:
Y a lnln 1-r
1-F
and X - x.
The values of a and B were computed using Equations (6.2) and (6.3) and the
correlation coefficient R using Equation (6.6).
6.6 DETERMINATION OF T!4E BEST FITTING DISTRIBUTION AND ITS PARAMETERS
In the previous sections, the statistical distributions, their par&-

meters, and their linearizing transformation equations were discussed. The
distributions were then fitted to the data by fitting a least squares line
(Y a a + OX) to the linear mode variables of these data by computing the
values of a and 0 from Equations (6.2) and (6.3). Knowing a and B, the
51
the correlation coefficients R were determined from Equation (6.6). The best
fit distribution was the one which had the largest value of correlation coef-
ficient R. This distribution could be taken as the "true" representation of
the scatter of the strength data. The linear mode of the distributions [for
example, ln(x-Xo) and (lnln I2j) for Weibull] are not used in practice, and
therefore, the original parameters (b, Xo, and Ofor Weibull) of the distri-
bution should be computed. These parameters were determined from the slope 8
and the interception a of the fitted least squares line. This could be done
by using the inverse-transformation equation as mentioned in Section 6.4. The
specific inverse-transformation equations to compute the parameters of the
four distributions used in this study are given in the following sections.
As mentioned before, Weibull distribution has three parameters. They

are:
Xo - lower bound of strength
b - the slope
9 - characteristic strength
The method of determining the value of Xo was explained in Section 6.5.1.

b is also the slope of the least squares line whose parameters are a and 8.
Therefore, b 8. 90 was determined from Equation 6.27:
a -bln (0- Xo)

in (0- Xo) -
Therefore, eb + Xo (6.44)
and b - (6.45)
6.6.2 Normal Di'tribution,
The Normal distribution has only two parameters. They are: f

u*mean
a - standard deviation
The linear form of the distribution is expressed as
and the equation of the least squares line is :

Y-0X+o
Bance, the parmeters of this line are:
52
a a
aand V were computed from rearranging these equations:

1
a = (6.46)
U -aa -- a (6.47)
8
Both of these distributions have two parameters, and they are referred
to as extremal parameters. However, in this study, these parameters are re-
ferred to as:
b - slope
M - Mode
The slope b is the slope of the least squares line (as it was the case
for Weibull distribution). Hence, b - 0.
The mode H was determined from Equation 6.34:

M -.
B (6.48)
and b - 8 (6.49)
These equations would apply to determine the parameters (H,O) for both
Largest and Smallest Extreme Value distributions even though the numerical
values of a and B would be differetin each case.
Using the statistical and analytical tools discussed here, a computer

program was developed whereby the computer prints out the fatigue strength dis-
tribution functions, their parameters, and the correeponding degree of fit
(correlation coefficient R) for several preassigned lives.
Data on various non-ferrous materials under various conditions were

collected and analysed by computer-aided procedure as described above. The
results are tabulated in Appendix 1 in terms of the best fitting distribu-
tion function and its parameters for several lives. The most representative
parameters were then plotted, as shown in Figures 6.14 to 6.64. The discussion
as related to these plots is given in Section 6.7.
6.7 GRAPHS OF THE FATIGUE STRENGTH DISTRIBUTION PARAMETERS
6.7.1 Egression of Data in Terms of Normal Parameters

fatigueA
By aid of the computer approach discussed in Section 6, the
strength data were analysed where different distributions (Weibull, Normal,
Largest Extreme Value, and Smallest Extreme Value) were fitted to a set of
data. The computer printed out the parameters and the degree of fit (correla-
tion coefficient R) for each of these distributions for a given set of data.
53
4 i
This was done for various materials under various conditions. A representa-
tive computer output sheet for a material and conditions shown below is given
in Table 6.2.
Material: 7075 Aluminum (Bar), Su=86 ksi
Life: 10 5 cycles
Type of Loading: Rotary Beam Bending
Theoretical Stress Concentration Factor: Kt . 2.0
Other Conditions:, See Page 233

and Code Number (240)
The distribution with maximum value of correlation coefficient R was

considered to be the best fitting distribution. For example, for the data
given in Table 6.2, the maximum value of R is 0.948724, and this corresponds
to Largest Extreme Value distribution. Hence, Largest Extreme Value distribu-
tion is the best fitting distribution of the fatigue strength of the above
material urder the given conditions. Its parameters are:
BETA (0) - .301634

and MODE (M) = 35.155
The fatigue strength data of various non-ferrous materials under various con-
ditions (effects) were analyzed in a similar manner. The results were tabu-
lated in terms of only the best fitting distribution (this could be Largest
Extreme Value, Smallest Extreme Value, Weibull, or Normal) and its parameters
according to the materials in Appendix 1. The tabulated values were rounded
to four digits. It was felt that the representative parameters should be plot-
ted to study the effects of such factors as stress concentration, surface fin-
ish, type of loading, etc. on the distribution of the fatigue strength. This
may lead to confusion as illustrated by the following example. Say, for a
given material and conditions, the Weibull parameters are Xo, 0, and b with
Kt - 2.4. However, for the same material and conditions, if the design should
be changed so that Kt becomes 1.5, the distribution function may change from
Weibull to Smallest Extreme Value (with parameters $ and M); and if Kt should
become 1.0, it may change to Nor-..ial (with parameters p and a).
When this information is plotted to determine the effect of stress con-

centration on the distribution parameters, a plot such as in Figure 6.11 re-
sults. This obviously cannot make muea sense to the reader, as it is very
difficult to compare cne distribution's parameters (say X b, and 9) for one
value of effect (say Kt - 2.4) with another distribution's parameters (say
and M) for another value of'the effect (say Kt = 1.5). This problem can be re-
solved by plotting all thr; data in terms of the same distribution. This means
if, for example, Normal distribution is chosen, then the parameters of the
other distributions slould be expressed in terms of Normal parameters. This
can be done by using, for all the test data, the values of the Normal para-
meters d1rectly from the computer sheet (a sample is shown in Table 6.2). Al-
54
I
WEIB.ULL SMALLEST EXTREME VALUE
THETA (0) - 37.8210 BETA (8) - .301694
Xd - 30.5332 MODE (M) - 38.701
b - 1.9047
R - 0.889976 R - 0.863512
LARGEST EXTRE1E VALUE NORMAL
BETA (0) - .301634 MEAN (P) - 36.9284
MODE (M) w 35.155 STD. DEVIATION (a) - 4.2207
R - 0.948724 R - 0.882308
Table 6.2 A Computer Output Data Illustrating the Distribucion Parmeaters and
the Corresponding Correlation Coefficients for a Given Set of Data
55
1 -41
oW
0 ,0
4U)
Ca)
too
CJU
445
- ~ .. .- U) 44
-V 0
C) 0 0 ' a C
oXe ID- 'g CU
4.5-. 46
I
though in some cases the best fitting distribution might be other than the
Normal distribution, it would not loose the significance by more than 5 to !0%.
For example, from Table 6.2, if Normal distribution with R - 0.882308 (that is,
88.2%) is used instead of Largest Extreme Value (the best fitting distribution
for this particular case) with R = 94.9%, the level of significance of the fit
is decreased by only 6.7% (94.9 - 88.2 - 6.7%). Although this approach results
in somewhat lower accuracy, it has the advantage of greater clarity.
Out of the four distributions considered in the present investigation,

Normal distribution, with parameters p and a, was selected for making these
plots. The reasoni for the selection of the Normal distribution are:
1. Normal distribution with its parameters pi(mean or average value)

and a (standard deviation or measure of scatter) is easier to understand for
engineers with little statistical background.
2. The plot of these two parameters can be represented by a single

line for each effect (See Figur 6.12) as compared to, say, Weibul1, where
three lines (one for each o' tiL_ three parameters Xo, b, and 9) are needed to
represent each effect.
3. Largest 'r Smallest Extreme Value distribution parameters (0 and M)

are not as easy to understand as In the case of Normal parameters.
A
It should be noted that: 1) the plots are meant only to represent the effect
of diffe:ent factors on the fatigue strength. For the purpose of solving the
actual dsign problam, it is recommended that the actual distribution and its
parameters be used from the tables in Appendix 1; 2) only the representative
values from the entire analysis are plotted, whereas the tables in Appendix 1
give all the values.
6.7.2 Explanation of the Graphs
Mean strength v was plotted against life on a log-log paper, one line
representing only one effect (see Figure 6.12). To the right of the graph a
specific value of a/v accompanies each line, this value being independent of
life. Therefore, once u is determined from a graph for a given life and given
effect, a can be calculated from the a/pvalue.
The meaning of a/h is discussed below:
For any given set of conditions at any given life the values of v and o were
generated by the computer. Consider the illustrative example in Section 6.7.1
graphically shown by the line corresponding to Kt * 2.0 in Figure 6.12,
This is repeated in Figure 6.13. At 105 cycles, mean strength is pie

Locate at 105 cycles (ii + a ),and through this point draw a line parallel to
the original line. (Thl asshuption for drawing parallel lines is justified in
Section 6.2, specifically in Figure 6.4.) At 10 cycles in example quoted
above, a - 4.2207 and P - 36.9284, and, therefore, o/ . .1145.This
in Figure 6.12 corresponds to Ul/V1 at 105 cycles. At an• '"°ther life such
as 100, locate P2 and (a 2 + 42). Ihe following demonstrates that
57
_ _ _ _ _ _ " •1
7075 ALUMINUM -(BAR)
U. 6 s ROTARY BEAM BENDING h
Kt l.
104 1
43?9J1 1S*j a 5 710
Life, Cycles
Figure 6.12 Plot of the Normal raramet~ers
58 -~
MINNOW
2.I.
hd II+
10 10
Lie cce
Fiur 6.13
I I9
U2/112- al/•1i .1145 and is thus independent of life. Since the lines are
parallel (Figure 6.13),
in Il- In (Ul-al) = in U2 - in (12-02)
or 1n -- ln
i-a01 "2-'2
or - 2-
1i101 12-02
or by simple algebra,
01 - -- - Constant
U~l U2
6.7.3 Use of Graphs
In this manner representative data were plotted for various materials

and under various conditions. These are shown in Figure 6.14 to Figure 6.64
The use of these graphs can be illustrated by a representative example shown
in Figure 6.12. For a life, say, 10 5 cycles and Kt - 2.0, the mean strength
1- 28.0 ksi and a/11- 0.1145. Therefore, 0 0.1145 x 28- 3.2 ksi.
60
6.7.4 Graphs
61.
INDEX TO GRAPHS OF NORMAL PARAMETERS
(Effect of Various Factors on Mean and Standard Deviation of Strength)
Matgrial Normal Parameters can be found:
Page No. Figure No.
ALUMINUM ALLOYS
7.5 Zn - 2.5 Mg Aluminum

Effect of Environment 67 6.14
Effect of Heat Treatment 67 6.14
M-257
Effect of Stress Concentration 68 6.15
Effect of Test Temperature 68 6.15
M-276
2014 Aluminum
Effect of Frequency (T-4) 70 6.17
Effect of Frequency (T-6) 70 6.17
Effect of Grain Direction 71 6.18
Effect of Manufacturing Processes 72 6.19
2014 Alclad Aluminum
Effect of Frequency 73 6.20
2024 Aluminum (Sheets)
Effect of Specimen Configuration 74 6.21
Effect of Frequency 74 6.21
Effect of Notch Radius 75 6.22
2024 Aluminum (Bars)
Effect of Stress Concentration (T-4) 76 6.23
Effect of Stress Concentration (T-3) 76 6.23
2024 Aluminum (Extrusions)
62
62
II It '
I
Page No. Figure No.
2025 Aluminum
Effect of Surface Finish 78 6.25

Effect of Surface Treatment 79 6.26
2219 Alumninum
2618 Aluminum
7001 Aluminum
(longitudinal)
Effect of Stress Concentration 82
86 6.29
Effect of Stress Concentration
(transverse) 82 6.29
7075 Aluminum (Sheets)

7075 Aluminum (Bar)

7075 Aluminum (Extrusions and Forgings) 84
C.OBALT ALLOYS
Stellite-31
NAGNEZIUN ALLOYS
ZK60A Magnesium
NICKEL ALLOYS
F$-27 Nickel
63
Page No. Figure No.
Hastelloy - R235
Udimet 500
Effect of Coating 90 6.37
Effect of Test Temierature 90 6.37
6 Mo. Waspalloy
TITANIUM ALLOYS
Ti-A55
Ti-75A
Ti-0.2 02
Ti-.2 C
Effect of-Stress Concentration 95 6.42
Ti-.07 N2 -. 2 02-.2 C
K-151 A Cermet
K-162 B Cermet
K-183 A Cermet
Effeqt of Stress Concentration 99 6.46
Ti-4 A1-3 Mo-IV
Ti-5 AI-2.5 Sn-.07 N2
Ti-5 A1-2.5 Sn-.2 02
64
I.I
Page No. Figure No.

Ti-5 A1-2.5 Sn-.2 C
Ti-5 A1-2.5 Sn-.07 N2 -. 2 C-.2 02
"Ti-6 Al
Effect of Shot Peening 105 6.52
Ti-6 Al-4 V
Ti-6 AI-4 V-.07 N2
Ti-6 AI-4 V-0.2 02
Ti-6 AI-4 V-0.2 C
Ti-6 AI-4 V-.07 N2 -. 2 02-.2 C
Ti-3 Mn-0.2 C2
Effect of Stress Concentration ill 6.58
Ti-3 Mn-.07 N2
Ti-3 Mn-.2 C
Ti-3 Mn Complex
Ti-8 Mn
Effect 6f Heat Treatment 115 6.62
Ti-13 V-l Cr-3 Al
65
Page No. Figure No.
Ti-16 V-2.5 Sn
SU40 kal ROTARYUMNU
BEAM BENDING /
T2(See Pa 7
orroAio Fatigue (9)
IIIi
IIf
ido q, ''1
-
* tal * vsw rsq
Lifue,C.ycles
Sit~w
40-6 kaeit~e arAR theMcd erofteaBuae ie
III I I I 1 1671
M-25 7
.110
M 1.1644
23 4 56790 2 32 4S 7 8936789107~
Life, Cycles
40
U) .1100
T-00(25 .0820
10~~~~~~~~ aa''*'. H I S19

RS~~O
Life, Cycles
(Numbers in parentheses are the code numbers of the tabulated values.)
Figure 6.15
68
M-276
as. Effect of Stress Concentration
Lie,
Cy.le
- ().0952
T * 00F
Ewa~*0 saso aa*sa.0

i parenthess
(Numbers are thelow eNNW bi tth au edvle.
.KtFiur 6.162)39
06
1 6 1 1
#910 4 5 s 91 6 3 3 4 is olo
S Lie
v Vle
2014 ALUMINUM
S 05ks ie Cce LATE BENDING C/
- - - Effect of Frequency -
30
1 .0~~0aa'S"~
9307
loz Life, Cycles
Effetgfurequ6.17
7013
2014 ALUMINUM
72 ks3LT ENDING 4M..
6.189 Ktur
7057
2014 ALUMINUM
S -72-78 ksi PAEBEDN
Z-.0683
Cod Drawn (61) .0578
0 *3 17St3 2 3 s 47 106~ 2 3 07
Life, Cycles
PLATsiE BENDING L
.1186
Hot olle (63
.1689
lo 113 a I *AaaO
to 7
Lifes Cycles
(Numbers in parentheses are the code numnbers of the tabulated values.)
Figuee 6.19
72
2014 ALCLAD ALUMINUM
Su 64 ksi PLATE BENING UU
Effect of Frequency
70
400f io cp (6)
-. ftft.0986
f3cp (7 .1123
Life, Cycle.
Su 67 ksiif'EBNDG
LIeC ifteI
(Nmbr infpaetee a00e tpe coevwbr o h tblae ale
r iue62
=73
I. .1010_____________________________________
2024 ALUMINUM (SHEETS)
S>EE
74kiAXA /
Efec of~lSpecmenConigua tion 436610
4043
AO a '''i 1191
CAA
uAXIAL cr/
,Ktv
40
.126
low %ft t 2.0531
7077
.0963
CAi
2024 ALUMINUM (BARS)-
(T-4)
Kt 101
20) ;Z.11025
65
.11
10 3 4 S,?S9]11 3 4ae
Life, Cycles
e6 ] 2 6ý11
E fec fiSures 6.23nraio

(T-3
2024 ALUmiNUM (EXTRUSIONS)
Reryt 1kiRTAYBA
ze6 ED l
Recr talized_
2 077
-. . .6 i07I-f--- .a 3 1 S 00
2025 ALUMINUM
S _ 56 a ROTARY BEAM BENDING ah
.01
ItI
10 ~~Life, Cycles 1
Us Effect ofiSurfae6.25 h
05151
2025 ALUMINUM
u ROTARY BEAM BENI ah
I* L
IL 07
(1ter
paenhees
i ret~e cdenuber o te tbuatd vlus0
Lifse,Cy.26
* 79
55-70 ka '1
2219 ALUMINUM
U
noS 59-67 ksiAXL
;z 20 7.0732
X .0866
01?,105
59ksi.
3 4 6
Life, Cycles
''1 6 11 101
*
ion
1. 15) .*7Kt
Iwo
10 7.. ss0
(Numersin
ar
th
coe
prenhess
nuber
of
he
abuatedvaleL.Lie Cycle
2618 ALUMINUM
ROTARY BEAM BENDING 0dj
to I i lII I I
so~Lie Cyclt-10e62
ar th cde
(Nuber inparnc~se umbrs ~ ce abuate aue.053
Igr 6.2
00 .4 (1
Kt
7001 ALUMINUM
su 100 ksi AXIAL
.144
I~~ #1S3' ' *'
Life: Cycles .i
EffectfijStres 6.2enraio
(Tasvre
C Jj ..--.-- "-. .. ..------------..

S,, 2-87ksiAXIAL c/
ISO
7
10 , 1.0 (2i0 10
S00
Me-anca Po0s (228 .0682
Life, Cycles
Efcof Surfae6.30 s
IGOi
7075 ALUMINUM(AR
uROTARY BEAM BENIN
10 ~~
'''10 1 4 1 3
l3q 6 a2*1 4 ''* 6 3 4 4s314

7so 7
10 1-110 Life, Cycles1
Fiur 63
.060
7075 ALUMINUM
EXTRUS IONS AND FORGI,'rGS
u 7-
87 ksi ROTARY BAMIBENDING a
11 1 1 1111 1
111 185
70
60 Firure66.32
M1
2
Extrwe .188
".to
toP
Y;
K _ f90 __ As61a90
STELLITE -31
S -123 ksi
u AIA
Efec of Stes Conbentaton
inprnhss r h o0'umeso
(I~~e0 taKlte value307)
3iur 6.33
=[Wo 7- -- k 4
ZK60A MAGNESIUM
104
--- K 3.4
U3
104 3 3 G790 3 4 s31s 9106 2 3 51907

Life, Cycle.
Figure 6.34
* 87
;N
FS-27 NICKEL
S 75ksiPLAT BENDING 0/1
070
w~ =T 2000OF
Life, Cycles
Figure 6.35
88
QRaw_
HASTELLOY -R235
u -64 ksi
S0 AXIAL 0rV
-T001200*F (390)Z
-;z.0624
:1049
.0518
00~
3 ~ Effect of Test Temperature
10 ~Life,
'
-~~~ Cce
Figure 6.36
89
UDIM4ET -500
SU 18 ksiPLATE BENDIN ah
II
.001
T 8011'
j411
j ~S U -156 ksl.AXA
6 Ho WAS PALLOY
M a
60
Kt2. 48)
IIH
'''''V 10 1 s.. 140 a 3 * 190

S 16
siLife, Cycles
U. AXIAL
150* T. - (4899)
!0323
EffectF
ofT-Tet7e30at
~io a a a0a 6
LieCyTe
0
paetee0t oenmer ftetbltdvle.
0i0re63
Lli
ElA
-~~L -I----..I
fS = 120 ksi
Ti-A5
ROTARY BEAM BENDING ah
M'
2t 3 4 07890 2
LifeCycle
ROTAR BEAM BEDI8
*.056
S 1~20 k4si l Life, Cycles
K2 1.1
-______~~~ -_ _- O -. .
'It T -75 ROTARY BEAM BENDING qV
to SII W uf
ace rud (518)
RolleadD
T__ 2Tf 3r 14S3.
6 r ac Rol an D. Ground .518
a i f
_Efet ins f ufae

Ti -0. 202
S
___ 82 k iR T R B M B NIi
11
IIIS1
Ti -0.2 C
S 85 ksi:
I lI I I I I I I II I I I l
,= I ,,... ,1.0z~ (534)'
o ..-- 0482
IW
Kt: 3.0 (535)rI1 I.03
00
10
'I,,.II ll ,
j~~jr
:l
Efec
li : i:L il
of Stress
,ioU
Concentration
a*s
isl 910 6? oo
10 i s 3 II 3i sl
Life, Cycles
(Numbers hth code numbers of the tabulated values.)
Figure 6.42
' i ' i ii i i i ii iii i
95
Ti -. 07 N2 -. 2 02 -. 2 C
S -121 ksia
u ROTA.Y BEAM BENDIN
'al
30
I I IlI ILA
Life, Cycles
Figure 6.43
96 .
K-151A CERMET
SU ~~ ~ ~ ie Cycle. PLT BNDNG a

in a
(Number E ftfee Thes
t ore coenmbersafturea U.
L 0
is97(51-
If____ I ______1 1
K-162B CERNET
S-U112 ksi PLATE BENDING
.1267
2000" (5 4.1401
10 -,Life,
Cycle.
(Numbers In parentheses are the code numbers of the tabulated values.)
Figure 6.45
98
K-183A CERMET
S = 72 ksi at 1800*F PLATE BENDIN 0r
w -
-Effect of Stress Concentration
.104
.059
.1227
tKa 2 .000
(5i48 -
Life, Cycles
.059
Ti -4 Al -3 Mo -1 V
30 A .1548
'28 1
'2.
0
10 3 4 56 19105 a3 s47 s9101 2 3 4 67810
Life, Cycles
1008
Ti- 5 Al -2.5 Sn -. 07 N2
SU ~ROTARY BEUM BENDING Ul
.0697
30
LLil
110 . 43 1 S4 7910
17090.)a 8 34 56 1910
,-.---
..-.- Life Cycles -.-------------
ji. 5 Al 2
2. Sn -. 2 02
Su 132 ksl ROTARY BEAM BENDING ch
13- -- Kt -1.0 (616)

-JI
N .1142
.0781
3.01(617
10 a 3 & IfalO
91. ''60i06 k4 1s
Life, Cycles
(Numbers in parentheses are the code numbers of thie tabulated values.)
Miure 6.49
102
JJ
Ti - 5 Al - 2.5 Sa - .2 C
S 1ROTARY BEAM BENDING
" "Kt" 1.0 (618
".0937
.0455
L~ ~ ~~ie Cycles.;.... ,,
(Nuubers in parenth•*Ge gre the co?,* numbers of the tabulated values.)
FiRure 6.50
103
Ti 5 Al -2.5 Sn -. 07 N2 -. 2 C -. 2 02
uROTARY BEAM BENING 0/
.0697
.1306
Life, Cycles
(Numbers:i: arnhse r thge d ubeso6h tabulated values.)
1.04
Ti -6 Al
Su3 s ROTARY BEAM BENDING 0/
iNo Peening (622)-
30
Efec o ShotPeng
2- 3 7 0 1 4 2 81 6 3 4 567 90
2 3' ~ Life, Cycles
0 0
910
Ti 6 Al 4 V
S u 150ksiROTARY BEAM BENDING 0
.0557
3 .0643
Kt (0636)
10 -~Life,
Cycle.
S_- 132-142 k8i AIAL o'
.1081
.1488
.0653
10~ F 6 1
Life, Cyclse
Figure 6.53
106
Ti -6 Al -4 V -. 07 N2
u ROTARY BEAM BENDIN
040
E4ffct of Stress Concentration
17 j6 701
a9153 4 la.13 6 s6 s 10
Life, Cycles
Figure 6.54
107
Ti -6 Al -4 V -0. 202
u ROTARY BEAM BENDIN
INS ---.0801
.1544
10 r~ ~~~I ieCce I If1
110
108.
Ti -6 Al -4 V -0. 2 C
__ ROTARY BEAM BENDIN 0
-. -K 1.0 (648) M 2
Kt 3. (6S4.9.). 3~
Fi~ure 6.56
109
Ti -6 Al -4 V -. 07 N2 -. 2 02 -. 2 C
__ ~ROTARY BEAM BENDIN
.0958
.0584
(Numers
pa Eftfeses of Sthescodenuers ft hetaulte vlus
Fiaure 6.57
110
Ti 3 Mn -0.2 02
-u 131 ksi ROTARY BEAM BENDING (/
30 .0926
-Kt - 3.0 (668)
Effect of Stress Concentration,
10 Life, Cycles
Fiaure 6.58
- ---
----
Ti -3 Mn- .07 N2
u ROTARY BEAM BENDING a
.0301
_-A I I A:.1003
33 4 7110 1 A 3 90 34 41 t6is9io7
10 Life, Cycle.
Fiture 6.59
112
Ti 3 Mn- .2 C
S 126 ksi
Su ROTARY BEAM BENDING
Kt: 1 .0 (671) "
'= • • '.0578
'K 7 3. (62
S
-Efcto Stress Concentration
Life, Cycles
(Nuber
inparnthsesare the code numbers of the tabulated values.)
Figure 6.60
S. . . . _-_"" -- a - 4
-11-
Ti -3 Mn COMPLEX
Su 126-149 ksi ROTARY BEAM BENDIN ah
.0.4
IfI
if.16141
Effect of Hteat Troatentaio
Life, Cycles
6.616 Ktur
1161
___________________________________
K t_________3 .0 1 1 1 1___________
1 1 f ill
Ti -8 Mn
SU 3 s ROTARY BEAM BENDIN
I I HT-.58 (689)-
.0788
HIT-59 (690)~
V41
M f
4
- - Effect of Heat Treatment -
ILife, Cycles
U ---
K .0,467
I
r ý 11
r711
115
ubV
Ti -13 V -11 Cr -3 Al
So
U 3-7 s ROTARY BEAM BENDIN 0
.0411 (705Y
-r4 .0846(706)
30 -HT-61 (705) _
.102
Effect of Setes Coeacentr-io
Life, Cycles
fit 1161 T 1
*Ti -16 V -2.5 Sn
ISO~.11 (59 760)76
C100
44 40 ~.1715(79
M.16742(7
Effect of STest CoTemptratioe
* Life, Cycles
(NmeS paenhsesaetecd
i80 itetbltdve. ubr
6.64 AXIALe
117 710
SECTION 7 THE STATISTICAL DISTRIBUTION OF STRESS
7.1. STRESS SPECTRUM VS STRESS DISTRIBUTION
The problem of stress distribution in the Interference Theory appears

to be much more involved than the problem of strength distribution. Consider,
for example, the problem of an aluminum cylinder head in a reciprocating
engine. Because of the variation in hardness, surface finish, etc., the
fatigue strength will vary from one cylinder head to another. This will re-
sult in a distribution curve in which the strength will be plotted on the
abscissa and the number of cylinder heads having a given strength (i.e., fre-
quency of occurrence) on the ordinate.
Consider now the stress distribution in the cylinder heads. The
stresses in the head result from the gas pressure loading. If the attention
is now focussed on a single cylinder head, then the variation in the gas pres-
sure will produce a distribution of stresses in this particular head. The re-
sultant curve will be a plot of the stresses in the head on the abscissa and
the number of times that this stress occurs in this particular head on the
ordinate [Figure 7.1 (a)].
This, however, is not what is wanted in the application of the Inter-

ference Theory, because this distribution of stresses cannot be matched with
the distribution of strength. In the strength distribution the ordinate gives
the number of cylinder heads having a given strength. Therefore, in the
stress distribution the ordinate must be the number of cylinder heads having
a given stress (and not the number of times a given stress occurs in a single
head). This can be obtained by considering the fact that different engines
will be subjected in service to different operating conditions, and therefore,
the distribution of gas pressure loading will vary from engine to engine. As
pointed out in Section 7.2, a spectrum of stresses must be converted to an
equivalent stress for the purpose of Interference Theory. Therefore, if a
spectrum of loading due to different service conditions varies from engine to
engine, in a population of cylinder heads, the equivalent stress will vary
from head to head. Thus the statistical stress distribution desired for the
Interference Theory may be obtained [Figure 7.1 (b)]. In this distribution
the equivalent stress will be plotted on the abscissa and the number of cylin-
der heads (frequency of occurrence) having that stress on the ordinate. This
distribution then can be compared with the strength distribution to obtain
the probability of Interference.
7.2 CONVERSION OF A STRESS SPECTRUM TO AN EQUIVALENT STRESS (Sequ)
By definition, the equivalent stress is a completely reversed stress of

constant amplitude which, when imposed on a part, should cause failure at the
same life as if the stress spectrum was imposed instead. Thus, the damage
accumulated at any given life, due to this equivalent stress, will be the same
as if due to the spectrum of stresses.
The first step towards converting the spectrum to a single stress (Sequ)
is to convert the operating stresses, which may have some mean stress associ-
ated with them, to zero mean stress, that is, to a completely reversed stress.
118
VII
- (a) Distribuion Not Wanted
*H ci)
w 0
-H (Vr
o w )-P
Stresses in a Cylinder Head
(b) Distribution Desired
$-4 .1
4
4-4
Stresses in Cylinder Heeds
Figure 7.1 Stress Distribution for the

Interference Theory
119
(Figure 7.2). This can be done by means of the modified Goodman diagram. Draw
the Goodman diagram as shown in Figure 7.3. From the spectrum of operating
stresses plot each stress cycle on this diagram as shown, for example, line AB.
Connect CA and CB and extend to the vertical line where the mean stress is
equal to zero. Hence, XY is the zero mean stress equivalent to AB. After re-
ducing all such stress cycles to zero mean stress, the stress spectrum will
have all the stress cycles completely reversed. The magnitude of XY will be
different for different stress cycles. Therefore, the original operating
stress spectrum [Figure 7.2 (a)], with various mean stress levels, is thus re-
duced to a stress spectrum with zero mean stress, that is, to a completely
reversed stress (Figure 7.4).
This spectrum can then be reduced to a single equivalent stress of con-

stant amplitude by means of Miner's or Corten-Dolan's Rule.
7.2.1 Miner's Rule
Miner's Rule(OQ) assumes that the total life of a component can be esti-
mated by simply adding the fraction of life consumed by each over-stress cycle.
Overstress can be defined as the stress above the endurance limit of the mate-
rial which, if applied, will damage the part. This rule is expressed as:
n n n n
1 +2 +3+ k
N1 N 2 N3 Nk
i1k ni
or -- 1 (7.1)
i-i Ni
where nl, n2, n3...fnk represent the number of cycles at specific overstress
levels; N, N2 , N3...NNk thenjife cycles to failure at these levels as read
from the S-N curve, and Z
E ni cycles. 7 is the fraction of the total life consumed by
The equivalent life (the total life) of a part (Nequ) under a spectrum
of stresses may be fount. by proportion:
i-k
N • n
1 ilk n
Suppose, for example, there are three stress levels, 90, 70, and 50 ksl,
in a given spectrum. With reference to the curve in Figure 7.5, 1/(6 x 104)
is the fraction of life consumed by each 90 kui stress cycle, 1/(5 x 105) by
each 70 ksi cycle, 1/(8 x 105) by each 55 ksi cycle, etc. Using Equation (7.2):
Nequ 1 1+ 1 +1
1 + 1 1.5 x 10 5 cycles
6xl04 5x105 8U0h
120
I.... _ .....
...
I.
(a) 'Spectrum of operating Loads or'Stresses
4j
Me~an
Load Nor
Stress
0-
V Time
(b) The Equivalent Stress
Sequ
Figure 7.2 Conversion of a Stress Spectrum to An Equivalent Stress
121
S
m
Mean Stress, kei
Vigure 71.3 A Modified Goodman Diagram
122
00A
0) t
02
AJJ
'4.0
4
** 1
loll esqui
123e
Thus, the life of the part under the above spectrum of stresses will be
equivalent to a life of 1.5 x 105 cycles. The completely reversed stress equi-
valent to this life is (from Figure 7.5) 75 ksi. Hence, the damage that the
part accumulates due to the above spectrum of varying stress amplitude will be
the same as if stress cycles of constant amplitude equal. to Sequ (in this case
75 ksi) were imposed for Nequ (1.5 x 105 cycles).. Thus, the spectrum of stres-
ses can be replaced by a single stress.
Miner's Rule, as stated in Equation (7.1), gives one (1.0) as the cri-
terion for failure. Miner's original tests showed that the value for the sum-
mation in Equation 7.1 actually varied between 0.61 nd 1.45. His more recent
dataOl gives a range of 0.7 to 2.2. Other sources(J2) quote a range as high
as 0.18 to 23.0. In view of all this scatter, it is generally agreed that the
value of one (1.0), originally proposed by Miner, is probably the best overall
estimate that can be made at this time.
7.2.2 Corten-Dolan's Rule
The application of this rule in converting the stress spectrum to a

single equivalent stress (Sequ) is identical to that of Miner's rule, except
that the S-N curve used to obtain the life values Nl, N2,...Nk is modified.
This modification is done, as shown in Figure 7.6, by changing the slope of
the S-N curve. A line is drawn with an inverse slope, d, passing through the
point Nl on the S-N curve, having the maximum stress amplitude (in this case,
SI) occurring in the stress spectrum. This new line is known as the Corten-
Dolan line.
From available data (13lh,) it appears that for specimens having no
stress concentration (K^ = 1), the value of Wd = 0.8 is a reaso:nable esti-
mate. A recent study by Harris and Lipson 5 indicates that when stress
concentrations are present the following relationship can be used
d/d' = (0.73 + 0.07 Kf) (7.3)
This can be graphically expressed as in Figure 7.7. It will be noted that if

Kf - 3.5, d/d' 6 1, and this becomes equivalent to the criterion obtained from
Hiner's Rule.
124
!'-• 2. . . ..- •i ". ..- .. .. .. . . .. . . .. . . .. . . .. . . .. . •• • .. .. . . . .. .

1.
I.
I..
'1 -.
0
0
'-4
7 _______ 'C
0
7- 0
K
In
In.-.
S
InK,
-- S
A
I S
I 3k
-g
e.i
10
I
I
* 0
-
U RU
- e44'
%r VY
S
1>'
4%
125
a
S
9su S-N Curve

U
0
n Cre-olan Line
sN
S
n3
3
AI ---------- •5
103 104 AAS 105 10-
Life, cycles
Figure 7.6 The Corten-Dolan Line vs. the S-N burve
126
•-- -- -r•--- -
S ( 1.0
f
S 1.95 S-N line

S.sU K .0for Kf = 1.0
*Af
b1 0I
3 41o
Life, cycles
5 16 *
Figure 7.7 Corten-Dolan's Lines for Various Stress
Concentration Factors, According to Harris and Lipson-(15
127 '
SECTION 8 THE INTERFERENCE OF A STRESS DISTRIBUTION
WITH A STRENGTH DISTRIBUTION
After the strength distribution and the stress distribution are deter-
mined (Sections 6 and 7 respectively), the two are compared and the percent
interference is determined as discussed in Section 2, Section 5, and in detail
in Section 9. For a given strength distribution the percent interference will
depend on the distribution of the equivalent stress Sequ. A search through
literature and other sources produced a considerable amount of data leading
to the distribution of strength but very little information on the distribu-
tion of stress.
In some engineering applications there is very little scatter in

stresses. This leads to a stress distribution with a standard deviation equal
to zero. This distribution can be represented by a straight line, aq in
Figure 8.1, and the interference can be determined as shown.
For a given Sequ, interference may increase or decrease if the life to

which the components are designed is changed. This is shown in Figure 8.2,
and in terms of S-N diagram in Figure 8.3. The shape of the distributior
curve in Figure 8.2 is different from those in Figure 8.3 because the former
are plotted on a linear scale and the latter on a log-log scale.
In those engineering applications where the scatter in sLresses is

appreciable, the above approach will obviously not apply. On the basiE of
past experience, in the present investigation the s:rcs, diLtribution (Sequ)
was assumed to be normal and the range of standard deviations to be not less
than .01i and not more than .10p where v is equal to Sequ. The resulting
interference is represented qualitatively in Figure 3.4.
Examples of design problems employing this method are giver fn

Section 9.
128
-~i
I
f(x) Strength
Stress
'1. '
0
Interference-
0 Sequ x
Stresk er 5trength, ksi
Figure 8.1 Interference with Standard Deviation

of Stress equal to Zero
129
Stress Fatigue Strength
at 103 Cycles
f (x)
10
Cycle
equ
fl(x)
10 5 Cycles
Sequ x
f(x)
Interfrence106 Cycles
Sequ
Intorference - 06Cce
f(x) .n
Figure 8.2 Interference of Sequ with Strength "

Pietributiom at Different Lives
130
i! --
0
P-4 S-ti curves with

.1%Probability
.~ 50 Z Probability
99.9% Probability
1WStrength Strength
Distribution Distribution
Interference 0
Life# cycles
Figure 8.3 S-tI Diagram Representing the Dependent*

of Interference on Life
131
Stress Strength
(a - .Oli.)
(d = .050)
I -
r Measure of
Interference
Stress (Strenqut)
Figure 8.4 Interference of Sequ with Strenpth Distribution

for Various Values of Standard Deviation
132
SECTION 9 APPLICATION OF INTERFERENCE THEORY
TO DESIGN PROBLEMS
Once the parameters of the strength distribution (X ,G,b: or v,a; or 8,

0O
M) and the stress distribution (G=Sequ and a-Ku, where K represents a frac-
tion of the average stress) are determined, as shown in Sections 6 and 7 res-
4 pectively, the percent interference and consequently the reliability (Relia-
bility= 1- Interference) can be computed with the aid of the tables in Appen-
dix 2. The specific steps to be taken in solving particular design problems
are illustrated by the following examples.
9.1 EXAMPLE PROBLEM NO. 1 (EFFECT OF STRESS CONCENTRATION ON RELIABILITY)
The present design for an aircraft part specifies the following:
Material: 2024 Aluminum S =70 ksi, S -50 ksi

u y
Design Life: 106 cycles
Type of Loading: Axial
Size: .125" sheet
Surface Finish: Electropolished
Stress Concentration Factor: Kt - 2.4 (Hole)

2 Operating Temperature: Room Temperature
An alternative design calls for reducing the stress concentration at

the critical section, at additional cost, by milling an edge notch instead of
drilling a hole. This alternative design reduces the theoretical stress con-
centration factor, Kt, to 1.5. Will the resultant reduction in stress concen-
tration affect the reliability of the part?
9.1.1 Strength Parameters ",
The first step is to determine the strength distribution in terms of its

parameters corresPonding to the design conditions. From the table on Page 194
(Code No. 100), at 106 cycles the strength distribution of the present design
is found to be:
Strength Distribution: Weibull
Strength Parameters: X - 13.76 kui
b - 1.808
9 - 16.31 ksi
For the alternative design, the strength distribution at 106 cycles is found
from the table on Page 194 (Code No. 1i2) to be:
"133
K
Strength Distribution: Smallest Extreme Value

Strength Parameters: 8 - .8348
M - 19.98 ksi
9.1.2 Stress Parameters
In order to determine the parameters of the stress distributions (V =

Sequ and a - KV), a prototype of each design was instrumented and the stress
spectrums were recorded as shown in Columns 1 and 2 of Tables 9.1 and 9.2. To
determine the equivalent stress (Sequ), Miner's rule is used. From the S-N
curves of the material (Figure 9.1 and 9.2) the number of cycles to failure,N,
corresponding to the stresses in Column 1 of Tables 9.1 and 9.2 are determined.
These are shown in Column 3 of Tables 9.1 and 9.2.
Using Miner's Rule, as expressed in Equation (7.2), and the tabulated

data in Table 9.1, Nequ is determined for the present design:
N 1 x- ni
equ
Neq 1 x 2571035 4.05 x 10 6 cycles
From the S-N curve (Figure 9.1)the stress corresponding to Nequ - 4.05 x 106
cycles is found to be Sequ - 13.7 ksi. Hence, the application of a completely
reversed stress of 13.7 ksi can be substituted for the recorded stress spec-
trum in Columns 1 and 2 of Table 9.1. Similarly, using the data from Table
9.2 for ~the alternative
th the~e~ desimn:
lx 4.0 x 061c0le
168.80x135 1.51 x 107 cycles
From the S-N Curve (Figure 9.2) the stress corresponding to Nequ - 1.51 x 107
cycles is found to be Sequ - 13.5 ksi.
a/ In some engiteere n applications, the scatter in operating stresses is
va small, thestandard deviation of the equivalent stress can be
assumed to be zero. In those engineering applications where the scatter in
stress is appreciable, the standard deviation lies in the range:
O.Ol a a a 0.10U.
In the absence of any specific information, an average value of a - .05U can

probably be assumed. For the present problem, interference will be calculated
for the two cases: a w 0 end a - .05u.
Thus, the stress parameters for the present design are:
Case 1. U - Sequ - 13.7 kst and a- 0 and,

Case 2. u - 13.7 kei and a - .05u - (.05)(13.7) * .685 ksi.
134
DI
PRESENT DESIGN
Stress Spectrum Miner's Rule Data
Completely* Reversel Number of Cycles to n,

kxial Stresses, ksi Occurrences, Nj Failure, Ni Nj
1 2 3 4
11.7 240 1.5 x 107 16.0 x 10-6
12.5 217 9 x 10 6 24.1 x 10-6
13.0 176 6.2 x 106 28.4 x 10-6
13.8 150 3.9 x 106 38.4 x 10-6

14.1 110 2.9 x 10 6 37.9 x 10-6
14.9 75 1.8 x 106 41.6 x 10-6

15.7 52 1.15 x 10 6 45.2 x 10-6
15.9 20 1.05 x 10 6 19.0 x 10-6
16.0 5 1.0 x .06 5.0 x 10-6
j.i " 1035 .255.7 x 106
Table 9.1 Stress and Life Data for Miner's Rule
*Actually, the stress was not completely reversed. It was reduced with the
aid of Goodman Diagram to a completely reversed stress using the procedure
given in Section 7.2.
135
ALTERNATIVE DESIGN
Stress Spectrum Miner's Rule Data
Completely* Reversed Number of Cycles to ni

lal Stresses, ksl Occurrences, nj Failure, Ni Ni
2 3 4
7-
1i.7 240 5.1 x 10 4.70 x 10
6
12.5 217 3.0 x 107 7.24 x 10-
13.0 176 2.1 x 10 8.39 x 106

13.8 150 1.3 x 107 11.52 x 10-6
14.1 110 1.1 x 107 10.00 x 10-6

14.9 75 7.2 x 6 10 10.40 x 10 -
15.7 52 4.8 x 10 10.81 x 10 6
15.9 20 4.4 x 106 4.55 x 10 6

6 "
16.0 5 4.2 x 10 6 1.19 x 10-
"
-I1035 68.80 x 10-6
Table 9.2 Stress and Life Data for Miner's Rule
*Actually, the stress was not completely reversed. It was reduced with the
aid of a Goodman diagram to a completely reversed ntress using the procedure
given in Section 7.2.
136
IO
cn z
Ii K7 1 ~7
t;S t
I t
~-r-- u137
coo1
* -, Kiji~ij
U I
r-4
14j
* ,, . r 44
I... - I - ~."---1 4
_ S ~ i4.
Similarly, the stress parameters for the alternative design are:
•"~ I. - Sequ - 13.5 ksi and a - 0
and 2. V - 13.5 ksi and a= .05P - (.05)(13.5) = .675 ksi.
9.1.3 Percent Interference
Once the strength and stress distribution parameters are determined,

the Percent Interference and the Reliability (%Reliability - 100-% Interfer-
ence) can be determined.
Case (1) when stress standard deviation (a) - 0:
For this case where stress is Normally distributed with Ui Sequ and
a = 0, the percent interference can be determined as follows:
Present Design
For Weibull distributed strength,
Interference F(x) 1 - e
1 0 shaded area under Figure 8.1
where x - Sequ - 13.70 ksi

Xo - 13.76 ksi
b - 1.808
0 - 16.31 ksi
13.70-13.76 1.808
'e16.31-13.76)
1 - -•~63-~6
F(x)FW - 1
=0
Percent Interference - 02
It can also be seen, from Figure A-2.1 in Section A-2.2.1, that when S
squ is
less than X0 there is no Interference. In the case when Sequ is greater than
Xo, the Perceint Interference can be determined as:
Find, xu e~u'xolb
X JS - ý bo
Corresponding to the value of X read interference F(x) from the table in Sectim
A-2.2.1, Pages 371 or 372.
Alternative Desimn
For Smallest Extreme Value distributed strength,
e+8 (x'-)
Interference F(x) - 1 - " - shaded area under Figure 8.1
"139
where x - Sequ 13.50 ksi
B = .8348
M = 19.98
F(x) =1 -e- e 18348(13.50-19.98)
-- l-ee= -5.404 - -.0045
= .0045
Percent Interference = .45%
This can also be read directly from the table in Section A-2.5.1, Page 419.
Find X = -$(Sequ-M) -. 8348 (13.50-19.98)
=5.404
Corresponding to X = 5.404 read interference P(x) = .0045 from the above

table. Therefore, Percent Interference = .45%.
Case (2) when stress standard deviation (a) # 0: (in this case a-.0511)
For this case where stress is Normally distributed with a # 0, addi-

tional interference parameters must be calculated to determine the percent
interference.
Present Design
Strength (Weibull) Stress (Normal)
X-0 13.76 ksi - Sequ - 13.70 ksi

b - 1.808 a - 0.05u
9 - 16.31 ksi - 0.685 ksi
For the above data, interference parameters C, A, and B(x) to be used in the
interference tables were computed.
Q-X
c0 0 16.31-13.76 a 3.72
AC-----
a - 0.685 -37-37
3.075
AMXo-" 13.76-13.70 .. 07
o 0.685
(x) =b =1.808
From the tables in Section A-2.2.2 on Pages 388 and 390, the interference val-
ue corresponding to these parameters is read using linear interpolation (for
a higher order and more accurate interpolating technique, see Section A-2.1).
140
Interference = .0398
or Percent Interference - 3.98%
Alternative Design
Strength (S.E.V.) Stress (Normal)
.8348 I Sequ = 13.50 ksi

M 19.98 ksi a .05u - .675 ksi
For the above data, interference parameters a and y to be used in the inter-
ference tables are computed.
8 = (.8348)(.675) = .564
au
Y =(Sequ-M) -(.8348)(-6.48) -5.4
From the table in SectionA-2.5.2 on Page42 2 , the in~terference value correspond-

ing to these parameters is read using linear interpolation (for a higher order
and more accurate interpolating technique, see SectionA-2.1).
Interference - .0059
or Percent Interference - .59%
A summary of the results is given in Table 9.3.
Thus, the reduction in the stress concentration factor should, in

effect, result not only in lower stress but also in somewhat higher reliabllity.
141
7!;
oN ps
04 w% co
N. r44
e4n
ION5
U 142
JA
II
9.2 EXAMPLE PROBLEM NO. 2 (EFFECT OF TEMPERATURE ON RELIABILITY)
A part in service has been performing satisfactorily under the folloe-

ing conditions:
Material: Ti-4A1-3Mo-lV; .125" sheet; Su " 16.0 ksi

Loading: Axial Sy - 13.2 ksi
Surface Finish: Ground
Stress Concentration Factor
at Critical Section: Kt - 1.0
Operating Temperature: 4001F
Reliability Requirement: 99.992 at 5 x 106 cycles
Stress Parameters for the i - 28.2 ksi (Previously determined)
Critical Section: o 1.41 ksi
If the efficiency of the system can be increased by increasing the operating
temperature, how will the reliability of the component part be affected if the
temperature is increased to 600"7,800"F, and the tat p mftruremain thesame?
9.2.1 Strength Parameters.
From the table on Pane 303 (Code N s. 586 and 587) the strength
parameters for 600"7 and 800*F at 5 x 10?' cycles ares
Strength Distribution: Largest Extreme Value

Strength Parameters: 0 - .20,66
kal
M a 40.79
Strength Distributions Largest Extreme Value

Strength Parameters: B - .2836
H a 28.74 kstI
143
9.2.2 Stress Parameters
The stress parameters have been previously determined and are:
Stress Distribution: Normal

Stress Parameters: V = Sequ = 28.2 ksi
a = 1.41 ksj
9.2.3 Percent Interference
With the strength and stress distribution parameters established, the

percent interference can be determined at the various temperatures:
600_F
Strength Distribution: Largest Extreme Value (ý,M)

Stress Distribution: Normal (ii,a)
Interference Parameters:
Sa= 8 = .2066 (1.41) = .292
y = a61-M) = .2066 (28.2-40.79) - -2.6
Using parameters (a,y), the interference at 5 x 106 cycles is obtained

by linear interpolation from the table in Section A-2.4.2 on Page 4Ll(For a
more accurate interpolation method, see Section A-2.1).
Interference - .00092
Percent Interference - .092%
Percent Reliability - (100-% Interference)
- 99.908%
800_F
Strength Distribution: Largest Extreme Valup(a,M)

Stress Distribution: Normal (Goa)
Interference Parameters:
a - Ba - (.2836)(1.41) = .400
y " B(p-M) - .2836 (28.20-28.74) - -. 153
144
J6
Using parameters (a,y). the interference at 5 x 106 cycles is obtained

by linear interpolation from the table in Section A-2.4.2 on Page 411 (For a
more accurate interpolation method, see Section A-2.1).
Interference = .33426
Percent Interference = 33.426%
Percent Reliability = (100-% Interference)
= 66.574%
A summary of the results is given in Table 9.4.

Thus, an increase in the operating temperature from 600 F to 800'F
results in an appreciable reduction in reliability.
145
"4 W4
"40000
41* 104 0AC
OOs
4)U
Id
p 4
V4
0 04
0611
0 04
0w 0
t-4,-4 A I-KSTVt
146,
I
CONCLUSIONS AND RECOMMENDATIONS
I. A method was developed for employing Stress-Strength Interference Theory

as a practical engineering tool to be used for designing and quantita-
tively predicting the reliability of mechanical parts and components, made
of non-ferrous materials, subjected to mechanical loading.
2. This method is based on considerable empirical data gathered (Appen-

dix 1) and it also has sound theoretical basis (Appendix 3). This method
eliminates the concept of a Factor of Safety and substitutes Percent In-
terference (Probability of Failure). Tables of interference values are
given in Appendix 2 for a variety of stress and strength conditions.
3. Although a great deal of data was gathered and analyzed in the course of
the present study, no data were found to permit the establishment of con-
fidence intervals on the probability of interference.
4. In our past investigation( 1 ), because of the limitations of the graphical

approach employed, this method was limited to the following three cases:

Normal Normal
Normal Weibull
Weibull Weibull
In the present investigation, based on the computer approach, the method

was extended to the following three additional cases:

Normal Weibull (Extended-see Sec-
tion 4.2)
5. For these combinations of stress and strength distributions, interference

values were calculated by an Integral Method and these are tabulated in
Appendix 2. These values were found to be highly non-linear. Therefore,
for the purpose of interpolating these values in a given table or between
the tables, a higher order interpolation (Section A-2.1) should be used.
6. A computer approach (referred to above) was developed for determining the

statistical distribution function of fatigue strength at a given life.
When the raw data (conventional S-N type data) were fed into this computer
program, the computer printed the correlation coefficient (a degree of
fit) and the parameters for Normal, Weibull, Largest Extreme Value and
Smallest Extreme Value distributions, for any given set of strength data.
This method has the advantage of high accuracy and time saving over the
graphical method used in the previous investigation.( 1 )
7. The effect of type of loading, surface finish, stress concentration, heat

treatment, temperature, environment and other factors (Section 4.4) on the
statistical distribution function of fatigue strength of various non-fer-
147
rous materials was studied. These effects were expressed in terms of the
distribution parameters which were tabulated (Appendix 1, pages 155-350)
and the representative ones plotted (Section 6.7, pages 53-117).
8. In most cases, the effect of the factors listed under item 7 above was to
change the distribution function. That is, for a given material and sur-
face finish the distribution function might be Weibull but for the same
material and different surface finish, the distribution could change to
Largest Extreme Value. (This refers to the best fitting distribution for
each particular material and set of conditions, as described in Section
6.6). The only exception is the effect of life. Life was found to retain
the same distribution and change only the parameters of that distribution
(see the basic assumption given in Section 6.2 on page 31).
9. For the non-ferrous materials studied, the effect of life on the fatigue
strength distribution parameters is shown in the table below:
Fatigue Strength Distribution Effect of Life on Distribution

Distribution Parameters Parameters
Xo Linearly decreases with life on

log-log scale
e 1 Linearly decreases with life on
Weibull log-log scale
b Increases or decreases with life
depending on conditions
p Linearly decreases with life on
Normal log-log scale
a Decreases with life
M Linearly decreases with life on
Largest Extreme log-log scale
Value __Increases with life
M Linearly decreases with life on
Smallest Extreme log-log scale
Value Increases with life
10. As to the problem of stress, the stress distribution was investigated in

more detail than in the previous study(l) but no additional information
was located.
11. In order to verify the validity of the Interference Technique developed

in the present study, it should be checked against an actual life situ-
ation. That is, percent interference should be computed for an actual
engineering problem. These results then should be compared with actual
service failures.
148
e4
REFERENCES
1. Lipson, C., Sheth, N.J., Disney, R.L., "Reliability Prediction- Mechanical

Stress/Strength Interference. Rome Air Development Center, Griffiss
Air Force Base: Final Report No. RADC-TR-66-710, March 1967.
2. Lipson, C., Sheth, N.J., Pittman, R.B., "Designing Engineering Experiments,"

Engineering Summer Conference, Department of Mechanical Engineering,
University of Michigan, Ann Arbor, Michigan, May 1968.
3. Lipson, C., Kerawalla, J., and Mitchell, L., Engineering Applications of

Reliability. Engineering Summer Conference, Department of Mechanical
Engineering, University of Michigan, Ann Arbor, Michigan, 1963.
4. Grover, H.J., Gordon, S.A., and Jackson, L.R., "Fatigue of Metals and
Structures," Department of Navy, U.S. Government Printing Office, 1960.
5. Heywood, R.B., Designing Against Fatigue. London: Chapman and Hall, Ltd.,
1962.
6. Lipson, C. and Juvinall, R.C., Handbook of Stress and Strength, New York:
The Macmillan Co., 1963.
7. Johnson, Leonard G., "GMR Reliability Manual." GMR-302, 1960. Research

Laboratories, General Motors Corporation.
8. Gumbel, Emil J., "Statistical Theory of Extreme Values and Some Practical
Applications," National Bureau of Standards, Applied Mathematics Series
33, 1954.
9. National Bureau of Standards, "A Guide to Tables of the Normal Probability

Integral." Applied Mathematics Series, #21, Washington: U.S. Govern-
ment Printing Office, 1952.
10. Miner, M.A., "Cumulative Damage in Fatigue." Journal of AVilied Mechanics3

Volume 12, 1945. ' '
11. Miner, Mt.A., "Estimation of Fatigue Life with Particular Emphasis on Cumu-
lative Damage." Chapter 12 of Metal FatIuen. edited by Sines, George,
and J.L. Waeisman, New York: McGraw-Hill Bok Co., 1939.
12. Dolan, T.J.; Richart, F.E.; and Work, C.E., "Influence of Fluctuation in
Stress Amplitude on the Fatigue of Metals." ASTM Proceedings, Vol. 49,
1949.
13. Richart, F.E.; and Newmark, N.M., "An Hypothesis for the Deteraination of
Cumulative Damage in Fatitue." AST PrOceediLns.
o Vol. 48. 1948.
14. Grover, H.J.: Bishop, S.M.; and Jackson, L.R., "Fatigue Strength of Air-
craft Materialst Axial Load Fatigue Tests on Unnotched Sheet Specimens
of 24S-T3 and 75S-T6 Aluminum Alloys and of SAE 4130 3teel." WA Taph.
not& 2324, 1951.
149
. ..
15. Harris, J.P., and Lipson, C., "Cumulative Damage Due to Spectral Loading.'
Aerospace Reliability and Maintainability Conference, SAE, ASME, AIMA
Conference Proceedings, July, 1964.
16. Henrici, P., "Elements of Numerical Analysis." John Wiley and Sons:
March 1967.
ISO,
150"
BIBLIOGRAPHY
1. American Society of Testing and Materialst A Guide for Fatigue Testing and
the Statistical Analysis of Fatigue Data. Special Technical Publication
No. 91-A (Second Edition), 1963.
2. Bazovsky, I., Reliability Theory and Practice. Prentice Hall Inc., Space
Technology Series, Chapter 15: "Component Failure Rates at System Stress
Levels."
3. Bird, G.T., "On the Basic Concepts of Reliability Prediction." (Monte Car24
p. 54, Seventh National Symposium on Reliability and Quality Control, 1961.
4. Bowker, A.H. and Lieberman, G.J., Engineering Statistics. Prentice-Hall

Inc., New Jersey: 1959.
5. Corten, H.T., "Application of Cumulative Fatigue Damage Theory to Farm and

Construction Equipment." SAE Paper 735 A, September, 1963.
6. Corten, H.T., "Overstressing and Understressing in Fatigue, (Cumulative

Fatigue Damage)." ASME Handbook, Metals Engineering-Design, Second Edition,
1965.
7. Corten, H.T. and Dolan, T.J., "Cumulative Fatigue Damage," The Internation-
al Conference on Fatigue of Metals, I.M.E. and ASME, September 10-14, 1956.
8. Croxten, F.E. and Cowden, D.J., Practical Business Statistics, Prentice-

'Hall, New York, 1965.
9. Dieter, G.E. and Mehl, R.F., "Investigation of Statistical Nature of Fa-

tigue of Metals." NACA Technical Note No. 3019, September 1953.
10. Dolan, T.J. and Corten, H.T., "Progressive Damage Due to Repeated Loading."
WADC Symposium, Fatigue in Aircraft Structures, August, 1959.
11. Dolan, T.J.; Richart, R.E.; and Work, C.E., "Influence of Fluctuations in
Stress Amplitude on the Fatigue of Metals." ASTM Proceedings, Vol. 49,
1949.
12. Forrest, P.G., Fatigue of Metals. New York: Pergamon Press, 1962.
13. Freudenthal, A.M., Fatigue Sensitivity and Reliability of Mechanical Sys-

tems. SAE Paper 459 A, January, 1962.
14. Freudenthal, A.M. and Gumbel, E.J., "Distribution Functions for the Predic-
tion of Fatigue Life and Fatigue Strength," The International Conference
on Fatigue of Metals. IME and ASME, 1956.
15. Freudenthal, A.M. and Heller, R.A., "On Stress Interaction in Fatigue and
a Cumulative Damage Role, Part 1: 2024 Aluminum and SAE 4340 Steel Alloys."
WADC Technical Report 58-69, June, 1958.
151
16. Freund, John E., Mathematical Statistics, New York: Prentice Hall, 1965.
17. Grey, E.F., "Statistical Methods as Design Tools." Eighth National Sympo-
sium on Reliability and Quality Control. pp. 70-71, 1962.
18. Grover, H.J.; Gordon, S.A.; and Jackson, L.R., Fatigue of Metals and Struc-
tures. Superintendent of Documents, U.S. Government Printing Office,
Washington, D.C.: June, 1966.
19. Gumbel, Emil J., "Statistical Theory of Extreme Values and Some Practical
Applications," National Bureau of Standards, Applied Mathematics Series
33, 1954.
20. Harris, J.P. and Lipson, C., "Cumulative Damage Due to Spectral Loading."
Aerospace Reliability and Maintainability Conference. SAE, ASME, AIAA
Conference Proceedings, July, 3.964.
21. Haugen, E.B., "Statistical Methods for Structural Reliability Analysis."

Appendix p. 110-121, Tenth Symposium on Reliability and Quality Control.
pp. 97-109, 1964.
22. Haviland, R.P., "Introduction to Theory of Reliability." SAE Paper 343 D,

1961.
23. Henrici, P., "Elements of Numerical Analysis." John Wiley and Sons: March
1967.
24. Heywood, R.B., Designing Against Fatigue. London: Chapman And Hall, 1962
25. Johnson, Leonard G., "GMR Reliability Manual." GMR-302, 1960. Research
Laboratories, General Motors Corporation.
26. Kaechele, L.E., "Designing to Prevent Fatigue Failures." Rand Report P-

3022. February, 1965.
27. Kaechele, L.E., "Probability and Scatter in Cumulative Fatigue Damage."

Rand Report, RM-3688-PR. December, 1963.
28. Kaechele, L.E., "Review and Analysis of Cumulative Damage Theories." Rand
Report, RM-3650-PR. September, 1963.
29. Kececioglu, D.and Cormier, D., "Designing a Specified Reliability Directly

into a Component." Aerospace Reliability and Maintainability Conference,
Washington, D.C.: pp. 546-564, 1964.
30. Kullback, S., "The Distribution Laws of the Differences and Quotient of
Variables Distributed in Pearson Type III Laws." The Annals of Mathemati-
cal Statistics. Volume VII, Number 1, March 1936, pp. 51-53.
31. Langer, B.F., "Fatigue Failure from Stress Cycles of Varying Amplitude."
Transactions, ASH!. Volume 59, p. A-160, 1937.
152
32. Lipson, C. and Juvinall, B.C., Handbook of Stress and Strength. New
York: The Macmillan Company, 1963.
S• Lipson, C.; Kerawalla, J.; and Mitchell, L., Engineering Applications of

Reliability. Engineering Summer Conference, Department of Mechanical
Engineering, University of Michigan, Ann Arbor, Michigan, 1963.
34. Lipson, C. and Noll, G.C., "Design Practice." ASME Handbook, Metals
Engineering Design, Second Edition, 1965.
35. Lipson, C.; Sheth, N.J.: Disney, R.L., "Reliability Prediction-Mechanical

Stress/Strength Interference. Rome Air Development Center, Griffis Air
Force Base: Final Report No. RADC-TR-66-710, March 1967.
36. Lipson, C.; Sheth, N.J.; Pittman, R.B., "Designing Engineering Expert-
ments," Engineering Summer Conference, Department of Mechanical Engincer-
ing, University of Michigan, Ann Arbor, Michigan, May 1968.
37. Lipson, C.; Sheth, N.J.; and Sheldon, D.B., "Reliability and Maintainabi-
lity in Industry and the Universities." Fifth Reliability and Maintaina-
bility Conference. Volume 5, 1966.
38. Little, R.E., "Design of Fatigue Experiments in Research and Development."

University of Michigan #6637-118, Engineering Summer Conferences, Ann
Arbor, Michigan, 1966.
39. Little, R.E., Multiple Specimen Testing and the Associated Fatigue
Strength Response. The University of Michigan, Ann Arbor, Michigan:
January, 1966.
40. Liu, H.W. and Corten, H.T., "Fatigue under Varying Stress Amplitude."
NASA Technical Note D-647. November, 1960.
41. Lloyd, David K. and Lipow, Myron, Reliability, Management, Methods and
Mathematics. Englewood Cliffs, N.J.: Prentice-Hall, 1962.
42. Miner, M.A., "Cumulative Damage in Fatigue." Journal of Applied Mechanics.

Volume 12, No. 3, p. A-159, September, 1945.
43. Mittenbergs, A.A., "Fundamental Aspects of Mechanical Reliability. echa

ical Reliability Concepts. ASME Design Engineering Conference, New Yorkt
May 17-20, 1965.
44. National Bureau of Standards, "A Guide to Tables of the Normal Probability
Integral." Applied Mathematics Series, #21, Washington: U.S. Government
Printing Office, 1952.
45. National Bureau of Standards Handbook 91, "Experimental Statistics,"

August 1, 1968.
46. National Bureau of Standards, "Probability Tables for Analysis of Extreme-

Value Data," Applied Math Series, #22, Washington; U.S. Government, 1953.
153
t
47. Pearson, K; Stouffer, S.A.; and David, F.N., "Further Applications in

Statistics of the Tm(x) Bessel Function." Biometrika. Volume XXIV, Parts
III, IV, pp. 293-350, November 1932.
48. Pearson, K. (Editor), Tables of the Incomplete Beta Function. Biometrika

Office, Cambridge Office, Cambridge University Press, 1948.
49. Pieruschka, E., Principles of Reliability. Prentice-Hall, Inc., pp. 99-

115, 336-340, 1963.
50. Pirie, Alexandre I., "Linear Regression Analysis (Two Variables)," Sta-
tistics for the Engineer, SP-250, Society of Automotive Engineers, New
York 1963.
51. Popp, H.G., "Reliability Through Statistical Material Property Definition!'

SAE Paper 580B. October, 1962.
52. Sines, G. and Waisman, J.L., Metal Fatigue. New York: McGraw-Hill, 1959.
53. Stulen, F.B. and Cummings, H.N., "Statistical Analysis of Fatigue Data."
Proceedings for Short Course in Mechanical Progerties of Metals. (Matin,
H.U., Editor). The Pennsylvania State University, Department of Engineer-
ing Mechanics, 1958.
54. Weibull, W., Fatigue Testing and Analysis of Results. New York: Pergamon
Press, 1961.
55. Weibull, W., "Statistical Distribution Function of Wide Applicability."

Journal of Applied Mechanics. p. 293, September 18, 1951.
56. Weibull, W., "Statistical Representation of Fatigue Failures in Solids."

Royal Institute of Technology Transactions. Stockholm: Volume 27, No. 49,
57. Wilkine, E.W.C., Cumulative Damage in Fatigue. IUTAM Colloquium on Fatigue

Springer, Stockholm: p. 321, 1956.
154
• APPENDIX 1 TABLES OF THE DISTRIBUTION PARAMETERS OF FATIGUE STRENGTH
• ~155
I
- r •;
A-1.1 INDEX TO DISTRIBUTION PARAMETERS
FOR
VARIOUS NON-FERROUS MATERIALS
Distribution Parameters
can be found:
Material Page No.
ALUMINUM ALLOYS
1. 1100 Aluminum 175
2. 5 Mg Aluminum 176
3. 355 Aluminum 176
4. 7.5 Zn-2.5 Mg Aluminum 177
5. M-257 Aluminum 179
6. M-276 Aluminum 180
7. 2014 Aluminum 182
8. 2014 Alclad Aluminum 187
9. R-303 Aluminum 189
10, 2020 Aluminum 189
11. 2024 Aluminum (Sheets, Bars, Extrusions, Forgings) 190
12. 2024 Alclad Aluminum 202
13. 2025 Aluminum 203
14. 2026 Aluminum 205
15. 2219 Aluminum 206
16. 2618 Aluminum 207
17. 5052 Aluminum 210
18. 5056 Aluminum 210
19. 5083 Aluminum 211
20. 5086 Aluminum 211
21. 5154 Aluminum 212
156
22. 6061 Aluminum 212
"23. 5456 Aluminum 213
24. 7001 Aluminum 214
25. 7039 Aluminum 217
26. 7075 Aluminum (Plates and Sheets, 218

Bars, Extrusions and Forgings)
"27. 7075 Alclad Aluminum 228
28. 7076 Aluminum 229
29. 7106 Aluminum 230
30. 7079 Aluminum 231
31. 7178 Aluminum 234
COBALT ALLOYS
32. Stellite-31 237
33. S-816 241
COPPER ALLOYS
34. Pure Copper 244
35. 70-30 Brass 245
36. Cu-7.3 Al Bronze 247
37. 5.6 Al Bronze 247
38. Al-Ni Bronze 248
39. Phosphor Bronze 248
40. Muntz Metal (Brass) 249
41. Cu-6.5 A1-2.4 Fe 249
MAGNESIUM ALLOYS
42. 2.5 Al Magnesium 251
43. AZ31A Magnesium 252
44. AZ31B Magnesium 252
157
45. AZ61A Magnesium 253
46. AZ80A-F Magnesium 254
47. AZ81 Cast Magnesium 255
48. HM21 Magnesium 255
49. ZK60A Magnesium 256
NICKEL ALLOYS
50. "A" Nickel 258
51. FS-27 Nickel Base Cermet 259
52. Hastelloy C 259
53. Hastelloy-R235 260
54. Incoloy-901 (AMS-5560A) 262
55. Inconel ". 263
56. Inconel 713C 264
57. Inconel-X (AMS-5667) 265
58. Inconel-718 266
59. Waspalloy 269
60. Inconel-751 (INCO-X550) 270
61. INOR-8 271
62. Monel 271
63. Nicro-Tung 272
64. Nimonic-95 274
65. Rene-41 276
66. Nimonic-80 277
67. Udimet-500 278
68. Udjnet-650 280
69. 6 Mo-Waupalloy 281
158
TITANIUM ALLOYS
70. TI-A55 284
71. Ti-75A 288
72. Ti-150A 290
73. Ti-0.2 02 291

74. Ti-0.2 C 292
75. Ti-N 2 293

76. Ti-0.07 N2-0.2 02-0.2 C 293
77. K-151A Ti Base CeNmet 294
78. K-162B Ti Base Cermet 294
79. K-183A Ti Base Cermet 295
80. Ti-4 A1-3 Mo-1 V 296
81. Ti-4 A1-4 Mn 307
82. Ti-5 AI-2.5 Sn 308
83. Ti-5 AI-2.5 Sn-0.07 N2 308
84. Ti-5 A1-2.5 Sn-0.2 02 309
85. Ti-5 A1-2.5 Sn-.2 C 309
86. Ti-5 A1-2.5 Sn-.07 N2 -. 2 02-.2 C 3i0
87. Ti-6 Al 311
88. Ti-6 A1-4 V 313
89. Ti-6 A1-4 V-0.07 N2 315
90. Ti-6 Al-4 V-0.2 02 316
91. Ti-6 AI-4 V-0.2 C 316
92. Ti-6 Al-4 V-.07 N2 -. 2 02-.2 C 317
93. Ti-7 AI-4 Mo 317
94. Ti-7 A1-3 Mo 318
159
6'
95. Ti-8 Al-I Mo-i V 319
96. Ti-3 Mn-0.2 02 320
97. Ti-3 Mn-0.07 N2 321
98. Ti-3 Mn-0.2 C 321i
99. Ti-3 Mn Complex 322
100. Ti-3 Mn-.07 N2 -. 2 02-.2 C 325
101. Ti-4 Mn 325
102. Ti-8 Mn 326
103. Ti-30 Mo 328
104. Ti-13 V-il Cr-3 Al 329
105. Ti-16 V-2.5 Sn 334
OTHER NON-FERROUS ALLOYS
106. Berilco #25 347
107. Commercially Pure Lead 347
108. 0.5 Ti Molybdenum 348
109. Commercially Pure Tantalum 349
110.:0 W-Tantalum 350

111. Pure Tungsten 350
160
SO
411
CNI
A 00
0 0
-,44
"Aj ..
w
4
,-.4 ,"
0 e0
V4~ 4JA
c.~o a4 00
r4.
0.4 41
x
I;
I'0 4i "4 a4 t46
V- di 0 a 0
'-4 K' p4q4
"V
4
WA 4
xu
"L oilS
I J f*
0 rz
0
00
$44
0
44-
0 4) 4 4.'
410
0 tv
.74 V
.744
41 0
to 0
60J (A 4 1'
.744
wi 0 0s
C14 K K
OA~
Pm) .
00 I162
xx
00
tvz
41
i 0 UTUI%
0~4.1
5.44
w~u V
'4 0
fl) 41
M ca
2163
0x
'--4
0
V z4
-r4 t-4Ul
0 40
"4 4-r4
.04
P-4
1.40 "4
0 4)
U x "
"40 04
00
aa
0a 4.'
CA 0
'A4
fA
4.'*1- 164-
ommmwm m
u I.
4) .
0 p 41 0
Uw 0 -A
6)0
w 00 04 00 )
00 4-J 0 14 4) "-1
0 .ccH-' 4 4.1 4.1 4.J E-4
00 w0)
0 -AE-4
A 80 ,-q 0
0.4 HOG0 ) -H 0 -,4 a 1
04.1 4j '4.40)0 4j 4
PL V0 0 0 a* 4j1 r 0 0a8-I HO 0 b * .
V .00 ~ X w~0 0 H 4)04 w

41 -A 4 a) W 0 0 0 0 4-J Enl 00.11.C rG
CA 00 0~ HP4 0
9) 44 u
-H 4)Od w 82 HO
1 0 1 -H c 0 4
c 0 ci m
4.1 4 0 W 0 a 1 4
,.-I U U H H CO 0 0 0 m r4 H
0H 0.0 p. AH W : 4I 04)C
00
0 0l
0 04
U)0 0) -,
.946
to 1.4 A5
0 u). 4.1
0 v0 H U Q" 0w
P4 0 @1 O0 Rd
165
-I
A-i.4 COMPOSITION OF THE NON-FERROUS MATERIALS
I. ALUMINUM ALLOYS
Alloy Percent Composition
1. 1100 ALUMINUM 99.7 Al Nominal
2. 5 Mg ALUMINUM .21 Fe, .58 Si, 5.02 Mg, .26 Mn,

Balance Al Actual
3. 355 ALUMINUM 5 Si, 1.3 Cu, .5 Mg, Balance Al Nominal
4. 7.5 Zn - 2.5 Mg ALUMINUM 7.5 Zn, 2.5 Mg, Balance Al Nominal
5. M-257 ALUMINUM Not Reported
6. M-276 ALUMINUM Not Reported
7. 2014 ALUMINUM 4.46 Cu, .43 Fe, .95 Si, .8 Mn,

.44 Mg, .072 Zn, Balance Al Nominal
8. 2014 ALCLAD ALUMINUM 4.46 Cu, .43 Fe, .95 Si, .8 Mn,
.44 Mg, .072 Zn, Balance Al Nominal
9. R-303 uALUMINUM 6.4 Zn, 2.5 Mg,L2 Cu, Balance Al

1Nominal
10. 2020 ALUMINUM. Not Reported
11. 2024 ALUMINUM 4.5 Cu, 1.5 Mg, .6 Mn, Balance Al Nominal
12. 2024 AL(LAD ALUMINUM 4.5 Cu, 1.5 Mg, .6 Mn, Balance Al "oml.nal
13. 2025 ALUMINUM 4.28 Cu, .36 Fe, .77 Mn, .76 Si.
Balance Al Actual
14. 2026 ALUMINUM .5 Mg, .75 Si, .75 Mn, 4,5 Cu,
Balance Al Actual
15. 2219 ALUMINUM .2 Si, .3 Fe, 6 Cu, .3 Mn, .02 Mg,

.1 Zn, Balance Al Actual
16. 2618 ALUMINUM 2.4 Cu, .25 Si, 1 Pe, 1.5 Mg,
.08 Ti, .1 Ni, Balance Al Actual
17. 5052 ALUMINUM 2.5 Mg, .25 Cr, Balance Al Nominal
18. 5056 ALUMINUM 5.2 Mg, .1 Ma, .1 Cr, Balance Al Nominal
19. 5083 ALUMIUM .27 Fe, .15 Si, .06 Cu, .76 Mn,
4.48 HS, .01 Zn, .09 Cr, .01 Ti,
Balance Al Actual
166
20. 5086 ALUMINUM 4.0 Mg, .5 Mn, Balance Al Nominal
21. 5154 ALUMINUM 3.5 Mg, .25 Cr, Balance Al Nominal
22. 6061 ALUMINUM 1. Mg, .6 Si, .25 Cu, .25 Cr,

Balance Al Nominal
23. 5456 ALUMINUM 5. Mg, .7 Mn, .15 Cu, .15 Cr,

Balance Al Nominal
24. 7001 ALUMINUM .11 Si, .12 Fe, 2.1 Cu, 3.03 Mg,
7.59 Zn, .2 Cr, .01 Ti, Balance Al Actual
25. 7039 ALUMINUM Not Reported
26. 7075 ALUMINUM 5.5 Zn, 2.5 Mg, 1.5 Cu, .3 Cr,
Balance Al Nominal
27. 7075 ALCLAD ALUMINUM 5.5 Zn, 2.5 Mg, 1.5 Cu, .3 Cr,
Balance Al Nominal
28. 7076 ALUMINUM .6 Cu, 7.6 Zn, 1.6 Mg, .6 Mn,

.1 Ti, .5 Fe, .25 Si, Balance Al Actual
29. 7106 ALUMINUM .35 Si, .35 Fe, .1 Cu, .2 Ma, 2 Mg,
4 Zn, .1 Cr, Balance Al Actual
30. 7079 ALUMINUM 4.3 Zn, 3.3 Mg, .6 Cu, .2 Mn,

.2 Cr, Balance Al Nominal
31. 7178 ALUMINUM 6.8 Zn, 2.7 Mg, 2.0 Cu, .3 Cr,
Balance Al Nominal
II. (OBALT ALLOYS
32. STELLITE-31 4.5 C, 24.8 Cr, .42 Mn, 10.4 Ni,

.93 Si, 1.39 Fe, .01 P, .010 S,
.03 Al, 7.26 W, Balance Co Actual
33. S-816 .397 C, 1.12 Mn, .50 Si, 19.42 Cr,

20.62 Ni, 4.1 Mo, .018 8, .012 PP
2.86 Cb, 4.03 W, 2.99 Ve, 1.09 Ta,
42.9 Co Actual
34. PURE COPPER 99.98 Cu Nominal
35. 70-30 BRASS 70 Cu, 30 Zn Nominal
167
loPercent Composition
36. Cu-7.3 Al BRONZE 92,7 Cu, 7.3 Al Nominal
37. 5.6 Al BRONZE 94.4 Cu, 5.6 Al Nominal
38. Al-Ni BRONZE 81.75 Cu, .71 Mn, 2.86 Fe,

4.69 Ni, 9.90 Al Actual
39. PHOSPHOR BRONZE 94.05 Cu, .18 P, 4.2 Sn, .05 Pb Actual
40. MUNTZ METAL (BRASS) 60.3 Cu, 39.7 Zn Actual
41. Cu - 6.5 Al - 2.4 Fe 91.1 Cu, 6.5 Al, 2.4 Fe Nominal
IV. MAGNESIUM ALLOYS
42. 2.5 Al MAGNESIUM 2.5 Al, 97.5 Mg Nominal
43. AZ31A MAGNESIUM 3 Al, 1 Zn, .2 Mn, Balance Mg Nominal
44. AZ31B MAGNESIUM 3 Al, 1 Zn, .2 Mn, Balance Mg Nominal
45. AZ61A MAGNESI17M 5.5 Al, 1 Zn, .2 Mb, Balance Mg Nominal
46. AZ8OA-F MAGNESIUM 8.5 Al, .5 Zn, Balance Mg Nominal
47. AZ81 CAST MAGNESIUM 7.5 Al, .7 Zn, .15 Mn, Balance Mg Nominal
48. HM21 MAGNESIUM 2 Th, .5 Mn Nominal
49. ZK60A MAGNESIUM 5.6 Zn, .66 Zr, Balance Mg Nominal

V.NIKLALY
50. "A" NICKEL 99.35 Ni + Co Nominal
51. FS-27 CERMET 34.3 Ti, 50 Ni, 6.2 Cr, 9.5 C Actual
52. HASTELLOY C .05 C, .48 Mn, .64 Si, 56.6 Ni,

15.5 Cr, 16.2 Mo, 5.7 Fe, 3.8 W Actual
53. lASTELLOY-R235 .17 C, .017 S, 2.06 Al, 2.55 Ti,

5.31 Mo, 15.51 Cr, .27 Co, 9.95 Fe,
.22 Si, .03 Ma, Balance Ni Actual
54. INCOLOY-901 12.41 Cr, 6.14 Mo,.18 Co, 2.67 Ti,

.13 Al, 43.20 Ni, Balance Fe Actual
55. INCONEL .02 C, .13 Mb, 6.81 Fe, .01 S,

.17 Si,.06 Cu, 75.85 Ni, 15.54 Cr Actual
168
AoPercent Composition
56. INCONEL-713C 11.9 Cr, .86 Fe, .49 Si, .13 Mn,
.11 C, 5.0 Mo, 5.6 Al, .52 Ti,
2.1 Cb, Balance Ni Actual
57. INCONEL-X .05 C, .57 Mn, 6.77 Fe, .29 Si
.08 Cu, 72.85 Ni, 14.98 Cr, .72 Al,
2.54 Ti, 1.12 (Cb + Ta) Actual
58. INCONEL-718 .04 C, .20 Mn, .20 Si, 52.5 Ni,
19 Cr, 3 Mo, 18 Fe, .8 Ti, 5.2 Cb,
.6 Al Actual
59. WASPALLOY .1 Si, .1 Mn, 19.5 Cr, .55 Fe,
4.39 Mo, 13.5 Co, 2.87 Ti,
1.29 Al, Balance Ni Actual
60. INCONEL-751 .05 C, .73 Mn, .28 Si, 14.92 Cr,
2.5 Ti, 1.16 Al, 6.59 Fe, .03 Cu,
1.03 (Cb + Ta), Balance Ni Actual
61. INOR-8 Not Reported
62. MONEL 31.2 Cu, 66.9 Ni Actual

63. NICRO-TUNG .1 C, .05 B, .05 Zr, 12 Cr, 10 Co,
8 W, 4 Al, 4 Ti, Balance Ni Actual
64. NIMONIC-95 .08 C, .36 Si, .06 Cu, .31 Fe,
.08 Ma, 19.8 Cr, 2.98 Ti, 1.74 Al,
17.8 Co, Balance Ni Actual
65. RENE-41 19 Cr, 11 Co, 10 Mo, 3 Ti, 1.5 Al,
Balance Ni Nominal
66. NIMONIC-80A .08 C, .66 Si, .05 Cu, .3 Fe, .06Mn,
19.7 Cr, 2.4 Ti, 1.18 Al, 1.0SCo,
Balance Ni Actual
67. UDIMET-500 2.8.75 Cr, .57 Fe, 4.25 Mo, 17.61. Co.
3.09 TI, 3.15 Al, Balance Ni, .1 Cu Nominal
68. UDIMET-50 .61 Al, 1.05 Ti, 2.90 Mo, (7,5 Cr,
.1 Co, 37.2 Fe, 6.20 Ta, 2.75 W,
.1 Si, .21 Mn, Balance Ni Actual
69. 6 Mo WASPALLOY .136 C, .56 Mn, .31 Si, 19.68 Cr,
6.32 Mo, 13.5 Co,2.59 Ti, 1.04 Al,
2.5 Fe, Balance Ni Actual
169
"II OR
VI. TTANIUM ALLO•S
70. Ti-A55 .051. C, .058 02, .01 Fe, Balance Ti Actual
71. Ti-75A Not Reported
72. Ti-150A 2.6 Cr, 1.3 Fe, 0.2 O• Balance Ti Nominal
73. Ti-0.2 02 .25 02, 1.0 W, .061 C, Balance Ti Actual
74. Ti-0.2 C .2 C, Balance Ti Nominal
75. Ti-N2 .2 N2 , 1.0 W, Balance Ti Actual
76. Ti-.07 N2 -. 2 02-.2 C .26 C, .08 N2 , .1 Fe, .19 02,

Balance Ti Actual
77. K-151A CERMET 58 Ti. 19 Ni, 7.5 Nb, .5 Ta, 15 C Actual

78. K-162B CERMET 52 Ti, 25 Ni, 5 Mo, 4.5 Nb, 3 Ta,
13.2 C Actual
79.
80.
K-183A CERMET
TI-4 Al-3 Mo-i V

41 Ti, 32 Ni, 3 Mo, 2.5 Al,
7.5 Nb, 2.5 Cr, .5 Ta, 12 C
4 Al, 3 Mo, Balance Ti
Actual
Nominal
1
81. Ti-4 Al-4 Mn 4 Al, 4 Mn, Balance Ti Nominal
82. Ti-5 AI-2.5 Sn 5 Al, 2.5 Sn, Balance Ti Nominal
83. Ti-5 AI-2.5 Sn-.07 N2 .04 C, .07 N2 , .08 11, .12 02

5.04 Al, 2.5 Sn, Balance Ti Nominal
84. Ti-5 AI-2.5 Sn-.2 02 .05 C, .007 N2 , •1 Fe, .2 02,

.5 Al, 2.5 Sn, Balance Ti Nominal
85. Ti-5 AI-2.5 Sn-.2 C .2 C, .05 02, .007 N2, .1 h

.5 Al, 2.5 Sn, Balance Ti Nominal
86. TI-S AI-2.5 Sn COMPLEX Complex of the previous three Nominal

87. TI-6 Al .06 C, .035 N2 , .578 Al,
Balance Ti Actual
88. TU-6 A1-4 V .02 C, .18 Fe, .011 N2 , 6 Al

4 V, Balance Ti Nominal
89. Ti-6 Al-4 V-.07 12 Not Reported
170
Alloy Percent Compositon
90. Ti-6 Al- 4V-.2 02 .07 C, .005 N2 , .12-.13 Fe, 67-

82 H2 ppm., 201 09, 6.03-6.37 Al,
4.18 V, Balance TiActual
91. Ti-6 Al-4 V-.2 C o23-.24 C, .014-.016 N2 , .10-

.13 Fe, 75-92 H2 ppm .113-.120 02,
6.10-6.35 Al, 4.12-4.24 V, Bal. Ti Actual
92. Ti-6 Al-4 V-o07 N- .19-.22 C, .052-.053 N2,0
.2 02-.2 C .10 Fe, 62-63 H2 ppm, .194-.201 02,
5.74-6.18 Al, 4.12-4.30 V, Bal. Ti Actual
93. Ti-7 Al-4 Mo Not Reported
94. Tt-7 Al-3 Mo .020 C, .03 Fe, 7.0 Al, 2.7 Mo,
.004 H2,BlneT Actual
95. Ti-8 Al-I Mo-i V .023 C, o09 Fe, .013 N2 7.6 Al,
1.0 V, 1.1 Mo, .0l0-.014, H2,"cta
Balance Ti Ata
96. Ti-3Mn-.2
02.02-.03 C, .016-.034 N,.0
.83 Fe, 36-39 H ppm , .150-.175 02,
1.00-1.06 V, 3.9-'M2 Hh, .90-
1.04 Cr, 1,15-1.20 Mo, Balance Ti Actual
97. Ti-3 Mn-.07 N2 .03 C, .052-.053 N, .88-1.06 Fe,

43 H2 ppm, ..
122-11~4 02, 1.00-1.06 V,
2.50-2.90 MO, .96-1.08 Cr, 1.03 Mo,
Balance Ti Actual
98. Ti-3 Mn-.2 C .21-.22 C, .014-o017 N2, 1.08-
1.,12 Fe, 96-108 H2 p, 1-.40,
1.29-1.35 V, 3.00-3.05 Mn, .94-
.98 Cr, 1.35-1.36 Ho,, lkalance Ti Actual
99. TI-3 Mn COMPEX .02-.07 C, .011-.0122N 2 86- 88 Pe,
48-64 H2 PPM, .091-.082 02, 1.06-
1.24 V, 3.08 Mho .80-.96 ir
.12-
1.20 Mo, Balance Ti Actual
100o Ti-3 Mn-.07 N2- .11-.15 C, o056-.057 N, .98-1.0 NO
.2 02-12 C 54-58 H2 PPMO .1189- .21102, 1.00-
1.18 V, 2.5-2.6 Mn, .94-16'C~r,
1.~.35Mo. Balance Ti Actual
101. Ui-4 Mn .14 C, o02 N2, 4.20 Mo. Bal. Ti Actual

102. TU-8 Mn 8 Mn, 92 TI Nosinal,
171
103. Ti-30 Mo .011-.C16 C, .065-.069 N2 , 29.88-

30.17 Mo, Balance Ti
104. Ti-13 V-l Cr-3 Al .029 C, .14 Fe, .028 N2 , 3.0 Al,
13.5 V, 11.2 Cr, .010 H2 , Bal. Ti Actual
105. Ti-16 V-2.5 Sn 16 V, 2.5 Sn, Balance Ti Nominal
VII. OTHER NON-FERROUS ALLOYS
106. BERILCO #25 Not Reported
107. PURE LEAD 99.995 Pb Nomina.

108. 0.5 Ti-MOLYBDENUM 0.46 Ti, 0.02 C, Balance Mo Accval
109. PURE TANTALUM .08 Cb,.01 W, .01 Fe, Balance Ta Actual
110. 10 W-TANTALUM 90 Ta, 10 W Nominal
111. PURE TUNGSTEN Not Reported I
172"
It
A-1. 5 TABLES
173
SI4
I
I 4
A-l. 5.1 Aluminum Alloys
II
!
.174 "
,w
0. 0 -NC4M*ic
cm0
10 m 0C4 t 4
HM %0~r- C4;t
>H ~ O~CU
to_
OH-S4
Ii r-4
Of%0 en% o0c c
%o
ONO n%OLMn
w4 0
H- 1; *.. C.* 4* . . *
0 dw
cnL
r%-4-T
n' - D 0C
'4C
'- 0H **. H
ta H
z9400
0 oH00 4) T00%
M- 01
go4 0)- ' q .4ýqc
Hý dm. 09 01 0
*b;zn cn 404e 4 jF4
1- UTL01, Uc
0, C) to % 1. oCA L f4
P-41o `4 PP-4 c V4 0 4
P-4M
0N*po+
I
I p175
i4 %D ON~ r,.CN2 0
%D O'K
*1 **1* ) Lri -4 CY) LtnOm
P~w4>
00~~%o O-s .ON ~'K'

W~o 41 0: U04 N r OHý
)-
9:0eqrI00 C N . 0_0_0___ e'4qr H ý

_H 4 1
OH ~r-
r-*co~'
%. o O' 0 - _.T A 4j m *14
ST -T C
04H .440
l 1-4P, CL*
C'J4- C'JH S,- .c 9 C"-0
0 r- L 0 00 0a cr) -ITj IT N
N -1r-4 r oom o4or0, '1 U-
d ~ -O 0,17 qDV
0%
dw4 09T cn
z - -o
Ln a)H - _____0
% C
14 - T0 Tc f- - L
>17
_4 r.1
01. 0 ~ 'T 00 e 1 %.0 C14 Cf) I-,
04
Cr) 4 C- O r-'
N-'.0i C14 r-
cN -40It
x N '0 c ..
C r. %0.o
-4 I-I *e4 .
H H CN- LA m-. 4
N-.t 0 -'D
co coJ
10-
HN- -70% 04 h0 IT m %
'0'0 NOr
CO H *
C%4 N-
n * - *~ H 0.' It 0N-.
U)Lf 104 *'H O'

lz o P4 H 4 9 14 H COHH Cý N Hr247
ý i-If
0~ ~ w~* *m ' 0* ~
Hr1C '00 1-1
C0ý V)It r- 0C14 00 r4l*No
x 1400 H -fN r4NN
0 wN
CL , ON 40' C4-C' 09 C4J'0N

1 r
-
10n (0(0 ON'DC-7N .' 7 OC N-I NI L .14
C4 mN aH N N N
ID T IC4 Go m C6 4
N zl C4 (N C'0OYCr aN- r1
0'0 ~ r4N
0 0 r,
(00 ' n0 cON ODmN-
0
Hn H4 E-4 4 r- C). -.? ?_4'*'0' or. *to
NO
CO(N 1(
N
z
Lnx- 04- .. 0.
r4-..LM*)I 0Ln 0
-I . N-
-H.40 0-IC4 '0-7 CV _ M1 cN-IO HrI
00 * *-40. N
HN- *%110
enN NO4 *4 0
.4 NII.-7 (
CO). 0
.0
1 1 cn . (" ,
1-
00)
ZJSI v *Ana AV' -
e~q ;; e 0
etq;N~~
z 00V
C, Wl ON
0i %,01 v . ,- V- T
-4 -
H
Ln 0 08 08)
m 08
-1D008 09% 09
Vi
W 4o
.*
"H -n
I u .
1-N It M.*0
C4
I n
-
.
enCDl
Il
N
O - -4 I -T
C- "OP-4 I
*O8P34 . . 01 7 1 .£-4e VE
177L
t.0 Cb0 :t co M 4M
0d?- OH '.oN
M~ 0%-q '4OT Cmen
-l. 0
Cq0% %0 00 M oeqO ~-.T 0%U
P- 4 H~
% t
'-T
N
H N .-
N*T r4 000
C%4; CcH0 -4- 00- CC*'0
r-. V0 r4 K
(n -
r0 rsD 0%O- 00 * 0% as% H

.,?%C 'O %
m Hl Hn 0- 0 co C.- ~ !-H0rýM
%H %0 "
0% N. 8, C* HH
10 COL
00N 0
04 o'4 r,4 M r-4 .- 0%H H00C4 H'O4V.
H 00
% 00cnrc 0%H14
) r-4 H"O
HH
ON ýH.0 CJh r-. HHL
0%;~
W (
CiLf LA4 10ONN ~ L
CO '.0' M~%'rNH HOl_4 C r 4 -
0~ 00r UcCn
0 DN O%'4 NH
mý
mO LA0i
.%0 H
H 0 00 '.H
YNNO
UncO
0 as Men 0% 04N % C4 ýT-l N N
%.0 ri -41 eNH %00 OCr- ( 0 COO0 r4%

CA 1--4 * 0 *. -1 * *M 4
en 0 * o*1 C
04~N M* 4 N 04 04NH
0 I EP- 0-Nr %01- 00% Ln 0m

7%. 0 . T . . C . 0 . . %DU;
Ln
j n % Ir- N N .O
MO ' ONN.0
'40 cl D %-Q nr
C4 0 r0'
%%NCh 0t (140
0* . * M -0 N %0 *1 0
en en, 014'C ML NU.
en N m
W Wt r4L n0% m1% e ý -* a

-l - -%
q -1 -4 -0 0O
00
0 44 0P1 ;
0.1 0T
Iq09 08 08 08 08 0R 0
Ou~a PeMSTTodoj2e39T
~L -JH
ý-L Z-~lH E*I V-3
178
0 r4N
0%H C- r-ft4 ý nMr-I 000 N 00
H go 1%C00 r4i
(Ii t-I U0JH
LM0 r-I 00I
w 0W H r-c Vc'n I-%N H -I It
ca4 Nt N4 '.0 C4 4 C 1
04 ON0 r- N4H -I4t N N C14CJ,-4

En 4j H4 Hc 0 U1 tj N.H4 01 n
%D Ce L~ri -
rII *A * * aa
N n
*
C1
0 0 C1 L (Y) NU-1-
W;1C CV44
0~ M O N'C1~~'
K.~~~~~- H %C1-O 10d ~ C
Kn 04.
-- aJ - -
4jc r40 N-U %DU

1044' '00 440 e 0
-%0n 1- -4' nC- 0 4ý
*
'4 NCo0r (n- P 0)
P~l-4 LION NC4O-4M
CA *H *n
- r- %
0 4 C Mi M-40 C.4 O Ha4. 0 Ce COf.
04-W0 M NLn0 N nC 'i -1 $5 .M M NN(M

P ~ M 0L 0 M 00 0%0 I-I C. # .0(0
H-.7P4 HNI0 C43)44 -,t 000%0
*rI
'TC H'.0n N~I en %.DA - *H N 0I M
44
N4 %c - nk r %
LA 4. IC
LM. N0 CO . .06.
S0 ari .00 NH'0 000 c4,1 r%*H
cq W4J a- - f0C)-0 -
o%0.0 . r4 InO '0 ' *

'6M vo
1*HN0 N H
00 *HIQ0 j-I4 t-I
- ai
-j -9
CO)
034VS%.0 re C10 r%0 4 %* C4e
-4 --- - -
ac a a0 NOca aM
I dues, 008 009 006 006 009 006
aU~ aH a
ý4a
a * a a ado
a Vlv
cn N
a 17~
si i8
ON*o+ N z T Zv
179
%ri
%.0 00 qc 4C0I
0n A 0- 00 0r
a)(7 '0
%D -4
1 or-
% M eni 0
r-I
~ u 00' %D a M 10
00H ,-I * * .O
'.0 *ý co cot 4 C4 C4
00 0 Leit %D c HLA0
% r
co ~ H-
OLf --Tr 004aJL tv-10
-. 0 C-I.N
00H *- *
cAC-. C cnAH4r 0 cn 00 A 0 i4
MC
10 a, _ I__
14. 0 00 m
1 -0 0 w 0~
LA.
f'a%0 %LAT
p.ý A * * 1 . .
eqJ eq
C4rIcn 00 %D Ae4
C1 en
00
o-T f'ON 0w-T4C4
00c 00a -4C) ' O

N~v-%
-:r v-IN
%0
H L -I V enC4 4 r-. C'4 -. ICM
H I%% H 04- 4A. r0

CO 'aI 4-1 en r- , 4V
40 -4 -- 4 -
1-4 .3 L ;LA
A-4ý 0) 41 ICCA400
cn '0C14%DAý
r-44 N P.1 ITcn 0 % . f

Ao 41 '.1 HO U). en
. 00N NHD I
W- H 4 O4 . LA * -"4 *L)
'IC %-44 - -n(
41 0 M P-4H" NH00
0-
cn t 4c
AV4 1-.nq1
N~l~Sk AV' ý 0 F '
CO - c %4 - ý%- 0 -
I duel, 009 006 009 008
H T wd- OINS - - -O
*v4 000~
P9392UTS u *JUT
ON apoo-.9zz8 6Z
180
o I-.
'0 *KV - 0
to-Tm o- -4 H -40 C4
W H L rI V)m H'Ot e'IsirO
H~II r-7
C4%J 4O~-4 1%7 O
tv ca cn %o. N' In~ CN04 7 H C14
c-400 NO mt m -41H co
- 00 0 I'D H - 4TN
7%r
0- -LE- --
0 n 04.4 Lr i C14~ ý -4*8.0 Un O%

T o r-o C1nf
-H 41* * a N 0 *
wI CI ý -l 9r -! -n -T *-
0- 0 l H CV C,4 -#00~%D ~ r% e 4
HHqr C4 4 C4
I *
-4? -4* .4 in * ,I
. . 0
0 *ý *ý 0ý Hu In
-40co H ItM% - o C444

% 0CV4.r7
a 0 ý0
*H H0'T
4.8 *U10
%.
en7
r..-4H
'H
*l
N0%O
c*i *
HD
**f-
-
4.8 m a. -n q- -; 4 441-41
r4
0i 1 n00 eqr Goor' k ~ V C%0 C) m q7
In0~%Q,
:3f 0'n' 00. P(7,
%~t0 4r.00O
0. C44 00 H -I %"D'
N.U
~
m w00 rr .N *N00c U0 00*n n G
HHC 7H. ' HO r-O
a- n. 0 -% V-4 Ar
-44 4
H4 0tn 00-- -
006 -
009 4 -M
00644 -
1 009 -e
10061
-n
be%~ 4) O% m-co ~ ---
oc0V-m0%
4,j0%N N -C% - DN% -% -
asial __________l4 lln~e

*OSPm W307 T~q~ U
.19181
x I E * 01'E W
,-4W~ Nv04 ('c
(D 00 *4 0 . .
rCi HIcn N H H ~ 4 174
A0 >
0) V %0 0 K
0H cOHnen I
H- HN 14-
. e 0
00 o*
41C' el *H *LC4 ;r n %0
r40%- CxJ
'4 8. 0%*
D 4* C4
a.s 04 In 1*o m NC4C4 HN -
-,?
C14 r64 c n*) Nn N N H-
- r4 * -
0 rr
r-4 --ej I%% 0%
t
0I n M % %0 0 eAn
~
1-..0% 0 . e
- 0 0 n - -nC1
z 14 4
ca.O
kn HO 411 0
iv)
adI DL 00 C%
- m0 9 A00C
1-.
V4J %0 . .
04 2I . n0 C4- IR nC4C
u-I'.O-1 *C c
Fia - )
S-4
00!P
0.!o 00
04 4 - - --
000 000!0
ica m
0~~ -f-n
m
LA d*
!:4Xv
5.4ý
T*1ei*A3S*OF
182M
.OC
~q
6
'0 4T .
Vr-I o) en41%D4
%00 n
tnJ cn4
Hn
Q40
asO4-C
-.1 r- In%
h-4 00
09400
%Cco I'
en P
rl
0 0n r-.1 0 0% en4 *.
0Cp4 HHr-P-4HH HH
IC 4 r4-HOH Hr-q N .4H
09
1.0V 'td *09. H-/ 0(40 c~-
"0w L 140 'l ý 10 rIrý-P1- %0"40 OH
r4J 41 e-I *0% L0 sM M - ~ .0 r-ci
o%
w DC4J r-iHH
0 H'4H a; 1ý H .- IH
0u4e.JC4 ý 001'4 - 4L
40 -4~ Hr'-I P44O Vr -4 P4
4
0: 9 4H *(10.81 .0-1
4 9 *
. 194 HC'J
N NH -CJ-
(qN V4NF4 -4 CnC" H -4
-1 *4 0% * 0%% %*0 0 FA
H N e N419 ,-4f
N'O~ m~ m P
H (n Ch oý4 r-4 C4000

&neDc P t n. P
P4 0. 0
q. . %T . %0 A $ H
n. * N .
MQ&f
X00 14~I 4% *04N4
000 T e
00Go mC'J1 "a N C%t ,-4 C'4 *C4
00-t .%~ HuIe Ch

40% . 0
(- H4 *D (' **4 -. ' 1. 0 *
o4
I& V- m. P4~ 40
%T M F-0 4
M N ~ **
0
C.. Hn r 0)e %Y- 0%
`4 4.) 01H C;e~A ' 4%N
4 0 C4 ~ 'i-e u.
0100
0 0%T
*T 461 N 0% 06 *
v4-A dul 80S 09 0909 0
bola OOO 0 .O000 .00
-M-S - - -
'ON 01O
it 0.1
VT v3~
H - - -18-
--- T C4.. 0 1
0ý u'o~ 4 Ac NH r4
r4 Cr.4 '.00 '
g-4 r--- r- e'.jý 0c%D 0 %or- c'? e

4) M 00
C-4'o % .1 *r- a r, 0r4*)%
0n Ai m . *OI . . .0
ýc
04c 4. O N.
c -- H-~r '.o00
H .c 40 *nC4r V,-4oN r4 -c ~
04
V4~ A %0#*C*~,*4 co
tocO' H~
to 0 N C- n 0000
A* a'. *- O *n . * .- . * * r-, *'.

to ~ u
~~
Lf)incr -
. ~Y
,
~ -C ~
4
-e
00r
0~ w (n 040'A'Ia.o 4 1 q r H r4Hr
O -ti a~C,- LfH4 Ch ,-4NC ,

ý! "a H CIC
H -It4
.* . r . . LII.0 .* N
f * n %'0
.O% *N'OH ~ 00 Ce) *0
M ~cn1e4 ULnv4-4 H NC14 44M-4C r-
0- 0 0 C
0' rLl' C-4~r 00c%
n1
* n-4. 0'
-4 c r- Ln~f 1- 00'N
nOý r4-
- -4 -0
H .SI >T~T LT0 c
"IicT
Tt0's % 4 *ý4 % NI n L~nq;
\ ONS
- \tr n0
-% -
4.-
r. * n- 4, --- nC -4 -ne
da08 08
N08 00 9 0
-n 09 09 09 09.
C6 -M- - -
0 Q0 e4*Jf
-- \0~ N.- - -- ? C" - -
484
0O. 0) CY 0 00C
r4
co' Cq
LCp. 7%mO-
-f 4 -t if
f41xCO-4 0. NAtL0% %D r-
LV %om-4v 0-2 O'%O
r- I . O'4 *L4AnC4
N H0'
0 m e O''O - 0O C *I~
C Ln 00 mO~
-o r4 -T ri r- -
O,
H* - c~-,T(D
0 r 4 eno
C,4 -.
-4 M -
H
xJ
OD 0 ýr f -4 nO
ONO
.4 oo
0%0 m N 00
'~L 14 0.00
3Nenen N n
zq. u %D co00H .-4A.
-7 *n -I -
n 00 .4 OD 'o .- 4
S10 .1r CY r
M-40 0 0 V -4
n nm
z m 4J
-
in.... 14 L LAI
04 10 a'-H0) (O Aw *n 0 'aý -I *O m
H ~ A' > " l 7 0m N o1 or. mCý
X inV - 0.CIC -
4 -N -4P -4- : -TM4
0T 0.
-4u -4
x
C4$a.'
;1 R
%
Aa
4 L
, I4
co 4c-
0- 0 Ln
cO %~D %?Co0 r-4C7
0HO HIAý
,4w H4 (%4U ý Ae P-4 en LAION % k J(naO
N~- ~ LAC4 en
H.-4 C4 HHHC4. .
r4 -0 C) %DW
0) 0)H 0OH 00(

4CoN M0r-- r- 0ON It tr-.
w 4j. H r- . '
ON %D0 Co C'4 * . CA). .I
Id C) H1 eJ .* . %D*- ON %. en H al
HHH 00
Lr)Ho ~ i C404
*'CN
H - HC-
0 0p
AO 0 HLn
Ln %0 Cn C
O
H4Jr ON. C'4C'4. r-o04 1 -
OCN C
]r-
0nrý
H. H N N O n "Nr
rz x0C o O 0 . ~ O Cn
Xý
E'4 ;A CO * *0. q LA
r-H LnHH *C'N HHnN
HC4 C r- 4NC
to In
C
:4 ý (7N H N, HNO-
c 0 ~LA N H' 0 It
W 0 %0 00 M0% N ý4 00 , 0 C14 .DON
W ) _ _o
01 LAn ~~ON'00N ) HIALA NO - -1H-

H
Ho ON n ,en' U% ý4 ~OOLn 0 -
CoHD
H C4 H p4 *A4
n H CA 1
en~C
0) I0
~
II
4j 0) C%0 -H
.,-C1 r-4
HH4 a \'N0 r
r-4 rf S'0$ 0 .. 0 (). r-. 'T .
Ln C enCA4 '.DN
4'.0 %D0 ONn 0%0
%D1 M
M
.0%0
enAC 14 -.T c
p _ _ _ _ _ _ _ _
4.4
ItO %0~ C 00_ nt __ %

0%
___I
4 0 coLA PNLP 1
co 10-T4 0r-r- .
C4A0~
0 C'-4 I- q4 -o n -- ý T
H4 n4 44 -4O-0N- - 09 09
44 \0 N c - n0
-nN.
090
a Ia YWO Inq9M ln~gf TnqOM186la
OD
4c
LA 1.K1 '.0 K0c % l %

r4
-4 C~o 0-T C14c'r) l
V4WiIO-4
xu r-4 . r-o
N cf . 00 r-.-.
aJ3 t-4 *.4 Lf a, * .
00 H ('C.J4ý4
P4 *r, 00.-H J,-4 co P-4 0 -
mX . . N .j CIq4
En C'J
Cd co r-I V)~CO %.0O
0 a%.0 N
C'er .-4 .. 4N44 -q--
So 9: 0
0 m0 04
0 a4 0 n r41' m.
w m .xj cnXJ .
W"%0O Ln 0~4c 0~Zpq 00' D
04c %0 0 K
E-4 Xn .. 4 00 *- r4 r- *I N oI
C', rn C14c' r
4~ -4 H041-1
fnw0L Ln r-Cn 0 '. m0

en %0 ON
r-4 C 0 )-t %QC -T ai(%0

z= dz . * :r . -%I ~
U-1 C) -- Cu Cu -A
10 m N CuE-N COIA%
C1 4 rICOoN
0). eQ .c'r V m.n -%0 C0G ,4..t 0 k,4Un
04
C4 >%
P01
0*.
0Eq T
z ~~ fr -.-
- co,. - .T -T 0 %0 M
M -00O '.Co
1 CJi-- O ~ON
m 00 0 00
00_ . _ .. i
to r4~' co~O
C4O_4~0* F_ ; C 144 o
H)I 0' c'JC C%-* r-4L'
coe N *4 C- -,t * 0- I
4A N* O4-4 f*r-i -4*
N 00o~0 - ~~ 0
0 00
4J4O0N m \0 00 o C14\0ONA IT m
No~- 10 0
m 4w - -o M
0u~ o0;
el ~ o N1' 0) .- 4O
NN~
r4- H I( q r-4 H H
a,~c'~' C1
00CV_\D__)% _t_ 0_ 0 L) C4
'44 *0 .CI-. ~ *r'- .o-

4.4 00
01-C~ C r4~ C1 o C ý j
N ý - ý-N ,C40
- 0-t
0H deIur. 0 \ .C) 0v.. -4'O1
m HH
> 4CAO 0C1 * n Ch. C4 tf OD -4 -TL)
el 4 41 0 c. -tNC
;r
1? 09 08 080
OQO
NOc n\O Nr
QOO nT D 0 0 T0
'0 ---
0-
ezn - -- DP4-4 O ný
GO0
089 9
0p~ 09 69 01
bou NT ZE 00180
UU-
0 4C ~.OH O ~
H 1
C1 000
ooc
x C* 00
0w a)-4- 4 %
~'0 oH
C)0 H 0 4
04 C, H.._
0 S
C*
-I N Ln P-4 Nt M
L G) %oH
En CV '- V ~ 00(%0
0 en
V H
V- -. 1
U))
-. 4en
E-40.
08 P-r-4
>4 T zfn
r-4 m r4 %
% 'o. . .,
AN
TI*2o LIn~
I qu UOS P189I
N .O-,
C''4 -- ~ rr
O N Le) C
4
~ C4~f
C'Ir9I~~ C'J- C
f-4 -ý r 0 r *C) L) -n D I ' 00
00
HH m C4OU
% H n 0 tr *r MC r
mr. C-*q *r r-I * N 4 Ch ý H -4 C %M'I
a% v- M 04 .C4 N M I0C4" -T(O . O

* 00( ; C frLi
!C 4J C40U OCt; ( tHC
cq44 HOn c-
- *0a ~ ~ ' ~ -
wN 0 0 00 'IT m CYtH
N nm CO
L)0 . 0 -
H*HMr- - rIV4 - HO'.-4 - 4 %.

C4 rO-4
_ _ _ _-4 '-
H' N
*O ~ 000J~ceJ~
E-4~ w0 .N -4 . *Cn )C) .. . 0D
En H 0 *ý 0 * mH. O* CON
- 0*)r4
H 1.4f
r*-T r-4 *ýi
V* rI -4 H -4 -'04 04r40
44
00
j 04 30 . r . . N C
AO 0) *c: Co 4 *'04r4
4 wI 0
P,. ilul
0l rJ' P4F. fl1 -A
H40w%. 0. N*
FA_
8J* 0
_-_4 _4 Cz 0
: e "4
HN eq 4
HN *NC4 C,4H4 I
0n
%414Ch d £-JS O-dV!S -a r-j Cý

'to
oe S
WhU A, e ~q*lqeM***
08 -Q8
0
08 ,'
-4
09 -A 09 000 N9
oM~pO 4 SL Sz 8L
190
mo. -I0-N
H N 0HCl H r 0-H ?-o rs 4 Ln
0) f4 . N .. . . . . . . .
COOH
Go* CY) r
~-T ~r-.
r-4C
%0'O
CY)Hr
H0
-T ONCI
0 r-.iIrT
O
OHC1
0 -
-
co
Co E-D
H
ol HO-
O-qH
04 HHD
r-4 r--0 o0.
r-4 O 00
O'HnN
-4
fN
0 X E-0
0O.' * C4.~ - 0 O-
irH CO * H*H
0 41 H 0r nr 0 %C4 0L)r I -0 07
U P4
H1 H%') IO0 H
Nrt*. a%, ~ r NON0 ('-O H-
n* -n -. oc
00ONN
r n 0 %0 'TH O%0~
ýt r, CI N % H
N.,. . Nf .N LM C
0 ~~ ~ ~ ~ 1 O,?
C LN *L
H 4 CN4I %*H Hw-C4C Cf)ft NHH0qH cnH

00 ý 14It
, 10F- -t 0O 0 CAJ CA~fC14 0 l o
L) .r co -. . -r n . . c -n
Xn .2 . .0 r 0 * 0.. H*4
. Ln - **0
H 0 eJ * oi*H Nt .
p M 1~4 %0- ( -0 0'40
0) -
P OCAL OH10O
HC;i
F-4 LA
LA NHC
04 0 d) ON V-N
CM P-4~c ~f H 4P4P
%IHNV4
r4 ~ 00. .tn%) HA l u0 r 0 Q LM . 0 L i
a4 -;
Li
W 8 0 D 0 *
0r 0%a O N 10 4N. 00
4)J m~C
PN'0 L4 9-4 m 0%-$A# 94%00
N~~ '- '44 C.4 Hm

r4H '44 u4N 0u4 N
N* X* -. 'K 00t
0220
M.. N.m N. M
. boa 9z~ .u1 vt Ic *rz H
29- C"4 C-
*or,08 -O - --
219
H a InC14f 'C11 L ri cnc
-C~ 0 H%-
% Y? ..
P40V400 N) cc-
ý NN ' rC cCn -4u- 14 r-UO)0 04-
aCH~~C Hc6' c4r-r 8' 2oC4cOC
(U 0) H1 NH4 %DCNHr4 CYC' N H -4-Lt) Lfr N' -H4r C

C'41 NO * 00~ 1c-~ 0c r- * L(4 *IC
00 '0 ND t ~CV) c~r, . .? rH- M- Ha-

H0 HH0
C14 r-- c -IH C14HC H bS.
H
'C) ca0
M 0-' c oa' H-I r-- 0 O0'
-4I ca% -T ~ -4 m
C4 0 ;r nr .C4 c . 'C .0- LA

r4s-r4 -*-UH
1-4 Mr40 *alc
HN r-4 I *-.? (7 -*
'?T.r 4N
%C.D ON
0'CD0 I.Tr- 00cc Hc0a i (Y) CU
w
0
M P- m T % oc.t It r- -0' 0't- W
r4 co(n r4
Hqr-
d arc 'D .r-j *'C

HH v-I-4H C1 rv4'a -4 4ri -4 (n -HrH NHý
-H- wA
tn ,~ occ0
N 00r- rcnN 00 - 0'H
.4 w HLp r-c0c--,
x~ wf r..*% . . . -4 .* . ON .-4i &n . . ,I . w
H G HHH-Ir4 - -~4H ,-N H-1 4 -4 C4
0 w 1nN nCtSO' Cn Ln Lf or-00 r--*'CN tl -

>%G0 -,-4 r-.U r4 '~r-
ýoH co aN -T -T-4 -1 0
-4N C4IH I-NH C NH

4J
Aj
- CO
r14c P* t- N . -e . . 0'P.. r. en* N *4 0
_ _ _ _ _ _ _ _ _
r -4 -0
_ _ _ _0 l 0Aj4
N4
I c_
-1
0_
-M C1 0-
V Iaý
r- Cl l 0 TN %.o ~ Cr-N 4%T i
a 0- - L - ý - 4 - - Ch
a-0 o - - -* N-- o- o %-o .
PiQ aa 4 4w
lop C-LS E-LI E-Y C-LIC - IS
011 *
I ;
04 94 1 C
192
P -%t-
-C4 4c -
Tc
oý r-..- U 1- 4O 00CN ~ .ýD N rC'J3t

D0 C4ý 1
0- H D 0' %1 nc nr4 V4-

(k) 'o
0 a. 0
V-OO 0 f 'ýr"
O
C4'J Lr-4tON 0% NI
00LrO
I r-4 -4 ON -4,- 0~O -ýr- 0 n- 0; q0 ~ n 4r
W* - MI?~ 0 C4
*-% (n' f, -* o 1?
.u ONt %
r4L N N L N %
- -TH~ N 0 H
0 - . -U-i
H- n1434 . ls.* ** . A . .t
* d
I- '.4 -1a -1A (n--* r'I 400 M.040 L -I a% C

Q -
0 - T(%c
-n r- 00-n00%
c) 0.4- 00-.
N~lI .cCmr"e -4.t4rl 0% 0a U-
4 IA * C) .*n 4 .. I A.D r% - F 41
~0 b- x 0 __M N ("I ___4 - -ý
go Ln 0 0
44r-4 4)
mQ w4
~O C*. 0-4 *n *oQ0. r4 . V40
:3 w4~ 00 ti) -.1, 1 Qfý q ) n 7 4 o
0- -- - -00#L
-n m
04-
\D- -% -a - -4
-4-4
x-
9 90 C4. 04
V40 0
0 -41 -4 4rIý --4 r- r-I 1- N - I
W.
2
HK
~ 1k) 9- - G -4 -4
- - C-1
co11a 44 14 N 4 LM in..
rý - - - -n a -.. - i4
n In 4 kn
-l %D Ln m-l -4 o P n-4 Iý 4.
4 * . 0 .. 3 No9
. . .1 1 11V
a M % - .4\.
'00 r- cl cn c'ii 0L 0ONm
a,, O D
C4'Joll
N-rc
N N
HA 9-1 a% D % C4C Hr- c'J
?-IJ~~ N -1
x N IIn 0~70To r- Ln r cv) * * r- C14 ce
4) v-I * * *w4
OH HA HO0 H1; HH Hf.
0)1 t-U
ýL) eC OC-) ON OAOH C14 n 0
CIGI
4)-4 O0H %D0 rH H -T L -Hr- NO0 L
It0 '
41 M0 % 00. 0 r- b, . 0
%D H *. 00
v-I *H *Hr -*C14C *e'J4 . T ý06
H H H H 0 -HHr4H r-IH HH HO'
0 0
0~
fzl H C4o
0 ~%D'N o v-la'
_ 00 0L
m H , cO **
.%O ý . * 0'ý * . r-. L
NH
4 rH4 O HH
r4H HHI
?-*rH r- %
HH
M e 0 0%C co H%D P-4 Cc

a L nL
i'-r.. Hcr. % o 4% O P% T N NON 0
0 .' . 00.* 0%. M0 * 0ý
0- -T -4 L oa - -0c
C N V- -4N u-4NN r-
P4 - --
%a V4 Or% 0%0 cor 00 r, ' r4 0M n -4

00.0-4 *I*l. 0N t. C'% co n
4t 4N ý4 P4NC ý 4
N4 C44
P.C%
-- P n 4 De
LSI a Tfn Nt4 N
cmlqS NAN
T'~ TI.IO
N V4 ' C4
4---- -- 4
C 41 0 P-dr- D i % V c-d God P.-km
m-d
09 P 0 09 0 09 0n
-------
lb. 09 0 0C-9 0O 9 0
u9 Pq3TTodo23*~T tITTodox31~3f
CUI E-I C-l .- C-l

* Is S 96 66 - - -- ff - TOT ZOT COT
194
L.. ........
00 qK 4c 49 K
,-4 mf rc-4 00 r-ScM 5c 0 irs-Ir-- MC) -4M
x 1C.0-4 a 6%0F
b - 0 *- * r-q Lf) C14 Ln - C14
U14J %9 m0C N-TP)-T

0 110 %, CO *r-- f10H Ln 0
Oi) H 000 - f
*0r- 0 .T
,-i *ý M*
- 0r ' - l
-oU3P - -ý
0K a H
.r4
to
'J. O
0~ V-C4' -1r-. C
0- -* --- -
t- %a . CO-
w P4 -n 4- %0 - -4M( -n -m0
LM M
z J 00 r r- a ' -T
C'4c N C-4 ID LA %, i r-4.c' c (
41) '.0 0 * H
'n N j*e NNa.
0X ot0 " -4O.f F-' N '0 Nc '0H'.

H'4.W 'oc N *.a. . r. . ~ .
0(~ A( ** 0 10
c -4aF 0% %C PHO% C'.%I
('..4 %
P-34 a 00 I 1 ý% 'Qr
0 0 HoL
CAr
w C 0 0L . %D '44 -4a -1-4

a- r%.Ae4
oQ -. -T - - . --
LM
04- 1%. 0.4 4ý 0- en.

NN.4 ell
- .4t2, s ('t4 *N
en k0 404t M-O O't
a .
Iai I-d
V!d
I* u.Sp PSTT. 1ý O*
0- -9 -8 -9 - -
U. ~SE
"u4t I0
0" &Poo.* IOT Sol 901 LOT got 601
195
ouI
r- MDC) .0 a -T V- -ý00
0)
Ht CO ~4
H o ; uH - r-
H -4 C, Nm t4
I-' *K0
' -- L)m '
to E-4 - c
-4 g4 C- a-, %0 '.0N0T
0 E -
-H H '
p0 -T h r- \r.- (Nc H ooN 0
H N0- '0\0 -ON r- m N
14 _
0H- O \ Nr-O m0
-4N ý'0 0%
Hn -T0 r-N r-'0 fnJ NH)
Ln CýUl
cn * .*. %D0 %r.P
NNHr-
N)~ CA4mJ 0IH Y-i H
CN~-0
n m -, )I
V) C) '0 4 n- 0J
%.Dcv cr-Lf 0Lt) Cl
- zO 0)4 ____ C) )
0E- L -44411r
. G
%ý N'0
1- V) a , Cr A ý 04 - 1 D0
H 4C 41 I Cu m
E-'~4 W~ 0 : 0() C '0-4.1 I
cn
H
~~ ()a
~~
W -HH C*
L
PA
r-4-44.
L1 -4 O 4--i
C44 H
*4
* 0, -C
*- N- rOI
U, 00 oý ,4 --
- -w - an\Dc
E-N4 0~
4Jzn --: I ' tL
--' a
Pk NpD OT ITT lT
M A u196
0 -
HLf %.0.
%D .0 m' H_4 00 C0
?-- 02 H1 00 ~ ,n0
-4 w2 00~Q In 2C0
-) C14 - -~
0D 0'I.C) N
(D4
10C 14 M0 m cfl0
(1) r .Am -*1
W a)H Lr LrI %D 0 4)H NH
ca cN r-A Nt c 0
-0% - I
HY)
04N H 0.'~ oP.
.H - 0. 00 H*1ChI 021 C7%.
Ch
"D .* *n ..
ý 0 -0'. W~ N
H C 1-C4N CNHP
7f r-i 00UC 1 0.0 0 C40

%D V))
* 0
C14 *
*~L (3CV)*.4 .
C14In H
H0 x
HH
N.1 02a M 0.4
E.O UPLI~t 0 en -1 .01-0
HO 4540~ Ln £zI 1 4%0 %

PA2 - ~ l -DC
0004 0r f
U 41.4%. 4
wN O N'. I? 0n m*l N r
InI a n nI
01 fn 0 0 -
a en a
%0.
U=A mO% t . co
A-- dWSI
TYMIO T~nfl0ZV=
-UT pUT)0ZV
dlw aD auo
""in punoi
ON~ sp03-* VTT STT ?A 911 LIT
'197
'0 0 K KK
a 0 m0 0C HC
4 no %Ln~*
o C')Q
M 0 *04C0
PI) cH -4 oM 000 "r)ýT040 . N %'40

CY C fl( r
Ct4) M ~ '
1. . 00. 1' is OH Lf) r-'
) :3 H - -T -1 h - - -
0
O1 .n0 ON'u
m H,
- c0%-I CO 0 - 0-er- IflHc
H %XT 000 %0-: L 0
r m C 90~
H O 0 0 . 0. H 0-
0 -4 -:T
n
H P4H4q - - -HHH co0r LO4
-- - - -x
4J 14 -4Tco moC%D V) u-I -en C '0 H .r -4H L)-
-H"0 H CD-It m*NH ý l C t I -.T~ CONL
HCA Hx~. HHH. 'H
C*H C C.
0~ -Q -n' -1 -0 -
0
HN 00
V -4t
0HC14
HCC.
r-4-a
UonV
~%CcY Ur H
0I 1
0Lý4 o - CD H V)
Zn * 0 N.ý Co-'~ 'T N~t' -I T ) 0
H4 -e -4 4NriHý -4 -q -1
H) C14H C'JNC14 Hn % C
LM-I 7%~C M~ NvI
za (Vq C'4. Co. H * H * 0
4J H4 HA CC HCJ HLCý H 1J4 H)
Ht; OC4on " C 14 r4 r- r1O

j c4 -4 C4'0 C
H ) 0C
410 -H
'H
m4. HmN
rO
~ *fH
-4 ýt N"-
O.'0 C1
i 'C'L
4 W O M M .. CO .. -j . . ..
CO H $4W 0 'Ch %D .e O N ONr% C"4%D Co
3 4 P% 0- C) -4C'
cq N 1- r4- -4 .- C
-:r -T -zL 1-1 - -r -. T

0- ON." 0 nN 7% O O) G
L)0 ( D - 00IO%
1.0.4 N'
NP4 Co .
'n N .* 0c . . W ** . Ný
iz)D '.4 0 LM Pe.J. 0% N H0
N
r4 %us
044 0V4L
0% - 08 08 00
n 409 Mm C4 N__ m m
1~o SII)1 0!-

- -i- -K-.-
9! 6-I ccM.PD .t.~ .T.p!
T1q 198UO 41113S*.

%At IC is
Or-%
H m
~ 0- 00 -%)
C C14*
%To. nC40
cI 00 a)
00 H
C)) o~r-ID -IO

"ct
1`4 IC)'T
014.
v4 IAN in 030
VO 00('%m
Ho 0 ..
% M 00 H H CNIH,-4H
44 m0
41 'nc'L -4.J 14.I0o L O%Q

c H4
0. CJH
OD Wi00H
co C)%.CCA r-
m OL4 .J lAOo
*4r C-4 r-4
H~O
C-4 In c r4.r
ULnU
N(~ M , C4 HH0
U) Xn 04 OD 00
N4 N~ r4
HA H uHr.
rý. 00
Sn r-CY.
d en. n 0coN
z4 L1n okNY ~~%
00 .4C CIN
HO .Qa
_)H ncn N C
tn0 0CY
0 n. 0A 0 00
$4. 04 .. w4 0. N .
*-'13 .4r- Z' 04 (n
HU£ HI. -14C
m01 93 ST V
U) -n4)0
08o 0CC 09'2
4 j*II r-4
03)0 N fn N9
Ro
m~
Go
0%ij
F4%
lZ0 %DOq e %DCM h r-%oo- L'O
0- en eq *0 .Ln t r_ N _t C1*0
C14 0* V--N
*T -#0%
0)v r0 C:H% O N -
bo-
0)1 ~ J O
- A0 -4tN N q O
N)Mý *CON N zr~ ON r~q I? 0'
U) *- . .~* *n C. )-
m
r Im -40 O N C4r- % CY) 4 1
0 04 _ _ _
04 V)~i 10
r- or-4- Pc-4 Vr-00r--
0 N*
C4 0 -n -tt * H
*IA Hr r- C-q t-4t C40
00 -r h0
pa Or- LI) o (% r-4 oC u-O
No
(ON *M
00 00ea * N) 0 %
w%
Hq CC). G3cr Nr- )a C4
zv . . :r . . I t 4V
*04 N0 ci ::
.4 m 00
-4 -4 0 r-1.t
z K
0~ ~ .
*n' . *v t 0- '
W 1-4 $4 4 %D C4
,js~ 04
1.nq4 ZP -1Ul4 Nq A3-1 eVZn *ALn
4 - --
n
43; 078 j C 09 08 08 08
.iz
ON4 &At
cuno4
.Tz.
~c
M LMt~ N m
BZT
L W
.i .z
a 20
m0 en CON D P-LnLn U'
C4 4 C0% L
Q). n .' 'IC K1

03
r.4 -.C C104 .ONU.
cn . 1..An
H 0 - HO14 H-H H' -C
H N HN4
4 C14H HNC 'NH-4
00
00 E-4 O
H IT %DLn -4rr4 f - n IH
mi0 0'( N ~ 04 0 0 (1
Hl . OH ~ P-1P'0
w H 4 C4 0 H V l 0C .00 VI i -t
o~~~~L H' H %~C ~ 0 0 %c0qi

ONJ
r_4JLn nc ie m T
0 ON1
0
-t(n
O-t
NO
CV)NN
ON o'O
IT N n,'; M N04C C
H0 N N~ .N
0% t_4 P0
A3SML~UlTMN CN_4
-ItZO eq 0
'T dR 'T c 0 MS '

- - -40en 0
w . - %L
z- 9::'a-L
M201e
- - _ _ - _ - _ _ _ _ - - - .
00 4c ac
0 LflH %, r-%c'4I 00 %D
0u00 0
,- ) r'-u
-T %Tj. o - H .- . N
:3 .7 t . N.
%C0' N-. 0 Lri . CsIJ
4 4 00n ý -4 AHHH-- 0 H*c
N- 4c Pk >'I
a a0 Un M 0 H %D -4 4) 0 0N-C 00 n 4 V)n
441 4
HN-ON 00 %D - 0) ( Hr-V).1- H -T nM0
00 0 __K
44 41 r" -I v-IH *0-4 HNN HHA

00 Y 0r
m -t *T N *0 *n N- . . 00 M C14
litM t*j 0' r.
a C %00 ON-t
r-O %D~C 00 0 V0 n
r- OD 00 -.1 * Cn*f N-.in M *0
-YA cn
*NC Cj .. * O.t
N-'~- .c -ý . 0
0Y NNHr-4 a, . 0. rC' %D00 . )
0 -I
n0%r 0 00~c C1 %D-T %D NCV)
m M
r GI iAr(-0 H %D 0
vnC4L-4'O On-H e t
.I q V- r-4 04 04 r-4 M0 *C11 HH4?-HO
0) CO
!`i .. .'u N-N . 00 . .
H4 0; *N4t 00N 00 - LNM
N) C- n V '
NO% rý 0Mr..%0
H x U C;
6 Tl;s nT' to n~ '.3 tnq
-4 a) 3 4O
04 04 a T C40
09 9 -ii- -ii- -i
C4n In ýDý a-
4L)& a P4-ne
ooc .pO ET T 6E1 O*l StT
I dwj 9 08 0 0202
'. (7 0 en K-
C4M 0K K% o0 7 o
O4ten C'n% rI0%4. O -O0%

nc o04'0nGo
b
r.
P, . ?. -* . , * 'T 4ý . . * .C ' T%0Lf.
(t 2H-a . N0 n 0H %0%' *- 40%C
00 WiAHIH H HIH ,-q H HrqH -4 n H 04P-
V) cn . . 0 . . In0 . . %*-.O
cn O Ln n -*O*'0 * 0rOCV
~ . C H. *C4 H-4
HH 4Cq'Je,-gN c
a0 W K NH
to 0H -'~
rAH 0 000
Aý q% nX
r- * . * * r *n . .4r *.O.
0 n-LnF4t $ n *CC'
-O 00~ V-4 0
. *
n- 'T
.4
t, A D 1N 04 C-C'e4 ý4 ?4 H H(4 H %0 %. 4 *,-r4
S0 000 C% r- ON~ 14P1 , 0- *'*C44

r.%~'14~ . L hC40%Lf
OjL
o ... ~ .. L,. * * * 4... .T
-~e HCC C1 I-i O
r4C
Kn 4C
-TC
r-4 c 4
't HLfS0 UM - 0 N H O cocn m
Wo -4 M o 00 N. M .. %-.O .4
H j 0 ,. -P % 0 L -- -
oon
4Ten.or.e Knc
.4'OC (n 04g V- m-e' N.1.'
'.O .4t4
Lei~U M % iM
a - %oe0
a 0 %.Ten I~~- .
Ln %o0 F4 % 0 *)o
4- oE4 K
- -e m* nL 0 0p 0%;rr- n CA 0
C4 *t OsHO; *Ci In*.,-N~

** ON
U0;
.410; k
01ý 44, LM~'

1(V4 C%IT~C H %00
lD -4 '0 Kn
MC O Kn0% %44 1- 00 m. V-4
C3%m'O 0f~f 0%00- 00 0%%oC)
m .44 %0
.7 *4 C
o
%TmQ 1TLT'
inqj
cnq tLnqFM pa wl &
0 q~j Go %
0
~~N d1~4 ~4fod.ST
0. :.e~PZTO
nqe-Tn-8InjjTIq*
O~T4nq:o En F nT£fl
oll OOT O*T203

OOTOO
L40
r-4jj%
.0 0 0 a%0 -T.-*0
60
N'~4
r
-hNr M'I .
*f -4.4 0 N.
0C4-
.-
_
*- c'''
n0 -
0at
'-H H '
W~' 0 -0LMLr OOM 0-

-. 1C4 C- r-4 M r-I C1 H
0- *
W? HH
0~PO t Hr- 00C
-M N . HCv,(n, eq C-4..
0 0r"f-M uq
*CJN
H r- 0 LA .
0 .1. 0
P- 0 *n .,
E0-.0 0- 4 c -4 -4 0 Cad I C
0 0000!01
-% -4 CI4 %.04 -- 00 q.
Ln -41 *de
NPH q
elan
- --- ... *
MU u Q
g
IV0unL 4 eoI oe
0K - 910
NP~ 64T N~%D TýP-
-P-
1.
X H *, 0-
HN
0O
cd 00I
Hl, * 0.
00
0 Ia0
0000
0 H
H co c'
4 *H
cn~ 0 0 0
0 D.
H r C.)4Cm
-'o * -
H
'z C:~
PQ-4 H)
0 .
%
.
(Ucco
zC ic%
ýA 0 0 CINN
10 41 ? rie 'C
r-4 4
H '' -N
4 ~ 40 00.
Tyw .oN T WW.ZON

.LSI(
IARIA
IC II p ; oz~j
205
%De H
Nr. 0%
%C)c.n
C-- .C%.O%
~ 0 0% tn o 0 I
0
cf (. .0 (. (e. C* r 0%1
in -4 *0 -4 .c 'A -4
4) co ýD I %0r% 0-
wf4 0 %0 .- O CJ-4.O
00 r
r-4 *0)- %0% In %0 0
a. 1i-f,- *,- *H -r-4
C W rl-4. -0ý
LM- ON0 4 LI
0~
%0 %D * '0 0 %0 0* (n- ut, *

Lt,4. C4 (OR4- en.1' f(%4
%0 -- 04P40 % 7))
0%..4% '.. ~~~P '
Ln ON1'
ON'f*
.% PC 4 .u4
Ln LM 4
4 0% f '.4 . rq1
c t, d0V~s pq% *(N *c a P. .

Ln4. N C-4 'K en4 4
COO
Hr4 ~ '? 'W.4 SO
I-%0- -
- -)
0t 0T0.1 0
due/) C9 080080 0 00' D3 009c
u 4c*1
N .d. d-
V4 A
*ON 9u ~ 1 ~ ~
44 Ai206
%0 01 00a "N%0N ' D00~ ýT 17 0 c" 0
.4(A2 '0
%D1 (' C14 NN *~j (Ti 0 m £NO '0
bo1- '. q %D' 4-.

(1 %T1
a4) C- 0t" MS' CiJ' -4''s

00%D04r-.V
J
%j -7
'T * .- P- * (%J
C- 00*;r
C-C- .0 *D0 . 0 4 *
HC-4 C4 r4Ni-4 .- r- 400r
'01-14"
0~ 0 --
-,4 ol ' 4c 4 4
04 %0 04 % r-.g
wI 'T '0
CC4 04 r-1 C;
44 C14 r-4('N .- 4 C14 C14H

P-1r-0% 0-f4-HOO
0 04
'0 en~s~ M- J
N 0 C7JO 00"*' 0~f4'0 ~ -% 00 % h y
U0 r-4 C1 04 t- r- 0- q'* -4
w10r2 .00 O%D

CAOI % *00- nN o m%
CA n as in*ON in a 0 %
0n
-a Go 'ýON0% CS0 C4t
F*4( 94ONO
P4q 00 41 'T * 0 - ON1 1 0 *ý tm C
~ ~ H1 11 0
Wz ~r. 0%' m4O0 er
Z"4 4. P-1
*r5
fn. %DJC Sq *N C4f-0
jT' qN e0
IM . %0_.__M
4-- - C
P4 ýT -- q*
'0 to *n 40*
f- C4in c
a ý 14 ý .- 4 0m 'T -4 LO).(% 4 MO
9-4 ) *1
0_i_0_MM__,M__C
w1 94
H0 ~ ~ 1 t~C5 45
4)
0Q
-en-
4Csf
Omf
C1 in0 D4-
44. 45
CA
4j
4H 4.4 m.-*0 f4 ( r-. o
00 W L90 *f '.4( .
M _ _ 1 _ _ _ _ _
94 4-
ca) a ca
4.5O P0LS03 O
p9S.ZOa POTO 2 38T1"

'10~ IBUMMJH 30H P1~01 p zi~'
gý dwet 08 08 09 08 9z O00V
f U SHE1 '7 -, ~ c Wal~$

Slo - '40TT~d SHE' $7 - ~ c
nsTVDTFUV920V TV)uvqu14*H
(nON (npLL.1.
Opo3+
9-1
09T
9-
19T
I 9-1
Z9T
[j1
£91
19-1
$9T
I-
~99
T9-1
207
1C 00
-. la' m.~f C'4\ C4 -1 -rO I
o -4 -; -V
0) riLn t , r r 01r HN
-1 D
1-4 0)r C~14C
L N-.4 ',%0 M r H~eH -

-It. NN
040
r. H0 Or.e~ Nt C~
421 r-H N-0 4j0r-N0cn C1
Lf0 *HX
*N *a' *L
N0 W~ OHn 'I" . .
L4 * -Y 0 I I 1 l ý
H* L-**
N' 04 r-r-- *oN\H
O'O NCn 00 r- L(C 00
N 0I 'D (3, lNm-- - Lr)L OH c. t)_T ýDC
-, 9 HG) *K)
Lf ' LjH Nc-4 ~r-4 Hc4 C4
0%~ a%~ CH\ .N- N

*Nm UC14L C C
H E-4 Pr
En H-
X_ _ ~ -4 -1 -
1- * -44 H *f4 ,t
T 44)0 N -K -1 NK
0~ m4- 0 dr- _ w CNO_.- a'-t ."
m L) 0
0> m
AL T-rqP~
%I Q~t mA2 CV)8 r, 2S
01 . Lr) oLi z 0.1 0.t 01
Hn *r- 0It 909 _4 1-4~
A a SWvia vuoN*8US pTn~a10dp A''

aaiS PTIOd 01EZT1tW-
20
baF C8 C s2 L 9
r-4- uUSEPqTt
0 4 -H 0 r~-Is (CV)cnC C14C1
H4 ClH 00 IT ON N NcM r- r-
m r-4 V) .' *ý ~ ? L( LI) ,:
Oý0
T0N %0 r 1 N a%~ r-~
CoX
GJ~-4 ~j ~ L ~N O Or N
I * rH ý .t' N. -.
4 Cfl HN 1 *uiO
o m. E-4---
HH C4 04C
o -IC-
H qc ~ I~I 0 0~*f
- 0 r-4 CN -r-,r 10cr .

E- "t . T . . COT $4
V) . :r. \aD ON ~ *
r-4- - Y "!c 0 MV N4
cn '* 'KTK -C 'IItCt0
-4H . Cf 14-. LIl ~C.fON,-t . CON H-
t C O
IV) r4 c,4 -4 cvK N 0 'Io 10
H Ln c- '-i U
:0 -4 E-4'
>' j ON mO C*;'i L) 0 T- 0
En
0-4c M c m -4 - OD 0 -I
- - : - > -
xt4c a) TUJO Irl-4

Ne 4 !ýnNs~
0 -4a -t --
Ln~~~~ABIJ
O NON 0. 0 )C
fr' o D
- (7) - - i -w e)0Le 04c
4.. C- -4 r 0 - r-
41'
xN ZLI
ON') 00LI
2919
.-.. ~ - " - _ _S_ _ _

0 ~- ~~ O QN% 4. -nt
N in 0f
Ot4% HWI0% r-4N 00%O
a) H* ý r
0.'C 4) :3-
ti r'H C'JH r- 0rIe C'JH

4 .-4 %0
ca1Q 0 )0 HO CH0C~
NZ0
0'0 06 intor 4)'0 '.4o a.No
Hn * H.c' 14
*- r- *r-4N r.
0~4 "K"0 "K

411 t 0% 00 fn " .OD rý-r-
00N0
O-t~9- N~ HO'I t
i, 0NI=e *tN0 nH0
-1N- Nr4 -HNH *H
O 0% 0 . rLrIO
co.
14t; 0%* O4**N
&k H0C *pH *'H U1
*0w4 r- .4. 0% N 0%0
ca0.
~
H-M ~
-0V
0.
~ $
4
0%
L
=4 *O4
e4 NHfN
a0 VIMN NC) Hc'N *
4)0 ___ w_
I" N 0 r-I-*LM LM 41 -D%9-

" "0
K in in"K90
oo 4( 0
0 0 L 0% * %
4 *
NolN
C4U 0
0M H n N
V- *H
H 444N'
0 - - - - 40#00
0 No -* -p
09T OT 09T N-OR

d0 O 09 08 08
uu LIfenOd
oam oo9+ U LLTOL
210
0 0%000L
'N-
%0 .4 r i
) 0 0
00a. 0H 4)V lH
C
JqF,-
N N
41 ~~0 ~ 4 4 900C
n0
H0 (LJc'
0% 0 4 "4H
0~ 00 __6
0 00 4c
HN H
0~ -N
COJ 0
HU U,
0 '0
"4 ~
N
0. 0
%T
*o~ ~ p
0 0%
0 *t
00
r-
0 C4 m 0
I'm 41 P4 '44
t
OD~ -4
0 0O NM
0T 4
r0T 0T0
LnCl04 08 08
00 Ch
C4-H C4-
0N) oft ap~. m8T
V44
co-
0 00 0'C Cý k nm mrI ý.c

( xur-.a *U 0 0 I .T IT -'tc
m
00oo a)V N- CN
-I~f T-4 co *C'i
-4N -I-400
0 0~ H ~ ~ 4(
00 0
0P- CYi 00r 1 ý oý 0 r ý nc
.34 H
00 c0f-u 4JIq r~ %JDY CON%D OH
r
H- f-C) cv m 00 4
CO)Lr) H tiC') ON)
.H %D*HL .- IA' *4 -
COON
ror- (Y)~N s,4
LF Hv *Y CY
,-. L . .
*w (% . N . . N . 0 L
caH 0q HHH
.. P4 *H *CnC t.c-r-L
H. -iCI U2r4C 14C) f V l
.
H1 QT 0w
AH 4 r- m L d)u 0 0 10 r~-~*
-It L-. -.
HPN'~-4
W- co cn . C-4
* -Ltf *. 1**-
Hn F- f-HN q t* .N f- n .Ce
rj
4
00 0
4ODI w
JJO _ H V)-'DC r, n n
4 -1 0 0 'vi 00r-. 0
NIO C -
H AiJ NC n.LANC CA)1;t
E-4. -4 u -IC ' C
PQ 00 00 q N
rTw L00 -4 m00ar
E-4- .0 .~~t .LAC 0
U1Z4 o _x caZ .0o I r ca4

.LI Trnq1~k 'AVS I~nq~sjh *A'RS AV'S
l~.0seqrnl seqnl
0o1 O* 0*1
H wo 08 08 08 08 08
*TA tl811d BU40UN PaT OII
9-1 9-1 9-1 9-1

-ON *POD+ VOT 98! L8T 891
212
c' 4' 4
C 0 o 'l . m 0r cvn-'I
-1 to fOi c H 0 - -V *% 0c
H) .0CH 00 1- nu
O0H - rOH OLtf A r-l00
N Niýc
a)4) H L U) OD cv (~NN O'N

NH41~HmN '1
~o g:6-
HH 100C H %. r -:r %Dc Nc
0 w INOC 0fC )
H ' 04HN %D H HC-q Hmn

,;%o~
4T
cou r- % nIL
-D
E-D -n cq f* r-4 en f- co*
H %0
HN (n0 00N N.C HN
0) CO
H CN COC)
0a H-4
r-O M 0n -4O
NOý V) 0%-
00E ~-ItC
CN c )
En H IT H H. . 4.l
IA H2 CgC~ -.ch'~0 C4 Hý *H
z ' HHq w 0 %0m

NN~ %qr
> 'u
oo 0%%4
~ OH
Nt14ck 0 - 14 0.C140n
~~~ 08 09 08
.. 0' -. -0 nf 0
r% P
~flU%
213H
V4~~~~~ V4 80 da 40
0D (n 00r,0 N -4-
-%D- Lr)r- ~riýr~u mr-
00J H- F
C14.>4N.
~ orI o4C4r-I
H
(n C14'O.
.
t"H
. *
r-o 1
0
Y)C)
C
C
H H N N C4N Mcq H _T "c'.C N %D V) CN %.
*rH- *rH *-.

H 41 H7% 4N00C14N C '4
)NN V ItN C%D'O' 0
NN.N
u-,o o.N N N mN eq v-N'C1 NO~

%DHý q
N Go C) *n M~ * ON It. *~ 0C.0 1%O
14 *q N* N 'T **4 0 . . . .
0nNC Nkf o~r4LM)u. %DvITDý uDr- a n

rO
'I. CO~ - cn 'T
SH M en n cn N* H. n * 0. - H
go~ H .N- *Or . 0 *O
W~ 0 N *0 *n * 0 **
m
H 0C4 -
' H
14 '-'4 .0n .H '? * . . 00.. . LI0 r)

(D ( CO *04-t
M 0CCVT
n M : m0 *N O'I r-o
In- - c-ý T 0r- -T -cn % -r
N~ ~ D- N-4 . . -4 . . H * . H . . f
0 0-t *NN 'NH

- 0 q N r--
44 r-@ V)~ aNe % 4L O

nmr4 .00 nýD L )
;a ( n 0 * ý -ý0 .
1 0T %D
m1 0T%
£n
MV Ifue 08
: 08m 008 07 08 a08r
oV~n Uvu~pn
HO'O' ______ _______
* *' pS Oll
9'T p3 T
O-CPOiR
x ~I" LI
*O doPD £6! 0 06 096 061 86!
boa,&OE Oc OE OE214O
00.
r- 0coo %0 0 - e~IOD-4 %000- 0 %D0 m %

H) Hi O N *r M*OlO * . *0ChLf
P-4ej .4N r- . D c'cic'
C124-HiN HNIIT .%ON %0H C4I

H *0 *o
cd0 H -. 1 ON~ . . . -L
r4 04.4rcn- rV-4 C'LI)
N% -. NC14
C4 C4 9
S0 4 ____ ____
-r o %V .
HO-
ILn ~~ f-zN
HaHý
~ .-~
04
HCO
0 ul
LflO
~ zzI .z.
%O.C'
VP
.4
4
-C
MN04C *O.O'H0D D
%f~0 CC'JOn--
.0L,
O
ON r-4
OLO
, 4.
-0
C * 14
HCOI
4
HH
-
I0
1010 ULIT C'4 Mi Li, c %OO.0r-4 C4IO 14 00

0-?. %,o Ln . *or C4 a%.
0'* '%,(
r-I 4 . (4C4%. .
%o NU 00A'0U'io CIHO D 0al %1 C' cr)L

'.-r
COO .C'4 .40 *0
C 14C 14'J C
M.. Ci
eHn- O O)HO
Mi 04 HH4T
n - -
w H 4) 0 'f-?. IfIrO OU 1C)A
a 00 C' t * . O.*
0 u 1~NHC~C1 * 1WCI
p.Cu c % -I ON W__ m__%D__t o0%0 H
w~S. --. c
o )1 -fr4 c HC MC T r NP 44 O-t. M
enNC .00
-4-
H4 H H C 0 0 *OC *NO 0% q4nC
00 NO
o 14
o) ic 4 4 cJc4
pa, (4 aK 0 - . *K
P4 o -
o4 0".CenI
CCI
.
On
-NiCA%0
eni %0 -T
.
N enN
U0'.
# 4P4
O'.C%0 '
UM
In
.
aLsia TVIILON TVIUON jjnqT9 iTflqTOj TInqTeM A'

i~q:10TVTpn ASUVIIL TwUTpfl3TSTZO' 0,3zAguWZL
E-4
Hwe 08 08 08 08 08 09
p~ns__ _ _ _ _
0 1 pollzL I)Ta pov.ZOL1 PUVR

:3~~~ ~ 1 L1SI.L-Jl Z-
-ON Opoo.. 66T uuo0Z o o
215
0%DY 00 ODc co I-0 00OCAIt0
Mc r-i-- co .D aAoo F4 c-IT'
00. r
Q) vlrz H0%
- H C; *ý LNC
to H HHq NN1H W)I HN
W 4) %D C1N O 00N Hr- I Lfn

W 41 * H * rH-- CO
H4 HNC'4 CNN4 Hf H C1
4J1. 0T frN
m 00 cO-N
00C OCO 0L
KO 00 '.O H1
j * 1Nr- C
0N w0 Nn Hc N
04
K 't0
I' '. ýo
o- . " 00
N M~
*n N N r-cc r-
In. I-
zi H c.
I.4 H. N- 00 0.L
o.-CO
C1 Mo CN-t -l
HcQ Or-r4
pqH ~CH HL0 -t 0'
H 4) 4 00 ON It r-4
4.)() * c 0mc
04c HH 00~c N-r NH H0 %101,
CO O
E-4c co Cr
WL H~A O)41c
S.4
r MO CO 44*q H-- m
-4 O (7%- N IDN cn 0 O cVnL
O 0) ~A o-l
0;Cý
orc COHC
49 1 C6ON
Ol*
C
r-
H '44 r
44 %D 0'.4 U Ln~
- c - -0-T
%4 -
094
',-
c09a o- 08 -M 09 *
o~~e
*U. '.T su
0 pe .Z0 CD L
lo v l a*
,,z-i *OkjTvu£0 oN p03-.
T-L
9O~
I nqL-J
0TV9M
I iau o
60
216pjTuo *O
020mC H 1
eJ~ %00
H) HO .n
erý %Ds tC 14
0 I.- C1 C ; r V
ca ~%c4 4
J 4c
10E-
01 0o r- %0- -'tH
N'~ H N00 c4 ; 4
0O C.1cs E-4c',
-Hcoj
-H'J\ r'"Hu% ý N
HH0 oq
o ~ CCJN
1 %.0 LiQ'T ~
co rJ-41 cl *ON%D
0%ý C 2%C1 C r 0
Ho1%PN
N 0U2 M 00
00 'T C14N-i,
O%0v-
H4 0 * r*
Nn. 6 .
.zeluao
~C P10IJ0
H 9enj 9 0
*4 (1) 41 -t -
>~ C3, C4'I.IL -4 Pý %
HnC) 0
0NI 3 Up aIý0 q 217D0
,-I ~ ~, o%M ný - . -l r-q0
0n c'oO
f- ( %o ' -
wv.4 *Q,; * 4H ýL
v.4'
enr. N n* .DQ C4 o D
0 0
r.. 4 K
0 cn 0 H M % GVAoDr4 c T4
P-4)C.4 N r4 L M %D~4 ' r-
4a 4;L __A o Hrl

__4 K
u Ln' r-.l r-4 (n
'.4 00" cn o r-
0n
) %D-4%. r4 ON Ol
D N
to4 0 - -
K z* 14 0.
in -H
10
' r-l. -
r-4P-1OD V4 r4P'1 4
0E u d
-
PODP -)ýinv4j0f- -T4
-- 4
-4 L 0 -e 06
4 - -- - -%%
3-1 TnTt
'ONe ~ z POTTR
2180
0 %DN --4 - oL -4 -o -D0
0.
CO v 04 n O -T gn C-1
1-4H
0ýr!%Q en. %0 M .en % N%
4)~iI *c O .
eq-4 Ln-gco
-> -- - --
0r-4
O m r-4'C' Nr0.r4- cof- enr4
4) 0) co-c - -n
0nen 0. - ý
to 4 r-
K . K n .gn -
id
rlu%0 L ONIl n %0e-4 N OrO' rit
00
CL 0 K
V4 ~w.
jt-4 o 'CO Il--t N 4o
W 4 c iN' co. V4' mA.0 %
k4F.n0%%
ITr -% n . j. ' r
to X .. -0 - -%- -T -
0 ..
rA - ~~ ~ ~ t4 D .4 4O
j; 80i,
;6 t;4
i
F-4I ON ~
V-4qT
-7
-N
q'*
%npf
00' TIZO
-- -en
r-404P-
-
*n
M -4r-
00 -4 - % r --
D
11-1 . .
AM0l-0 n 0..
%
.% -M0a
3t Iz
P4~!T 40 -0 m
m mC4V
P N 4
4) -1P4 -4 ý4"421I9C
-.. --
Lj'0 cN 0OM C14%D O N~ 0 mQ
Q)
. I *l 0 *r
H . 1
4)~ ~ 100 1)ODH C140 cq0

%0O'J c'.11 C4
M %0
r-4C. Lfl %DCJI
cqf. 0 00 - cC4r
4) c H
(H0.eJ
0 0
-t () T( 0CY 0C40 -TrL)H 'CH04 0
.~ %D.W 0N0f
*'4 ItH H- %D
P.
-4 C cO C") " Li")*
I. Ix~ -ITcO
00 * *. *t 1-
'.L -j 00O0 0 . I m00 *D . 7*

-1
.r-i O'.O-
Z
"'f) ii 04
H;N. HIý ___ r-4 __r -l _
go0 H0"0 -H .L)
Ho HD co H - -IHC\0 w 'D* %0 D
HC4c4-
.' M .' .
'O'Jq' C4 IMC"0\ I
p"' 8$4 v
Mo -r-L. I
H 4j Ln l I~ MDf)
H-4
>L)0-4
.IJ v-4' i -
- *r'J 0
*rLf - lIm H C
i Ct) r 4H HD %0HCN'.J00
HH (dON 'C" H'C'
%D
Ln dHaj
An 4J
H
')
r-*0 HI)
z fl
1-4
C
44~~.
C08 00
-
-i
'D)C"J\.D'
-
\.
7ZML %%
%
*C
1e--
-
h
4
os
) N4
1
U
)
'K n
C
'
1 08U
--
ILNL
1
0TI
t
V
a n 1.
I
qc
NC Inc 7
(7" I t : i I NC Ie
22 I-
f- W VH Ln %0H
r_ 0 HIc
P-C4 -4c' H 7'-4r-
oeq-. C1 04 It r
14 N C140 N
0 0 C'4C'4*C'4H *-fr-
H -o
oCV ..
0
r r- OT--
%
p rý
cCe C H,4 14
*C14 C14
C14 H N
-U2
1-0 ON0* 1 I
H LA 4 0-:r4 'Tr-e~
IQ H -4 *L O *fC
00 00
V- 4C- ý 1
C) n08 08
L
""q 0' - -0
:n (n %'o -
en N -4enC
.221
A en-
0f oo. VI r.11 Lf' ci 00 r 004 00tf1-4C
Li~ C, LA 10 I
H eJ. N C1 Vr) C4H4 C -4 H *HI C4
00
P4 0 Lr-4 0OH c n i C Got OLn'
H~~ CV)00
cn )w D . . %D . :r * 0. . ~ 00 . ý
Cd;2 ~r-4 N.-4' N *4vAcr
-1O C.4 -H I C1H
$4~ 0H cqcl( -t H 0 7-4 0 0 0r-r
0.w P1I H cq NSI N Ln cn N CAJH H'J *H
r- O -4I c Ln ICfLf%D .- %0' %0r- HI

r4~ 0% um LIMO oL N'M~DH- NN o
-~ *0* n 0*
V-1 n 4 e m ~ 1tnMC)N eq-qN- N ~ -
goa P %::
-1 e-o %dC
M.o-
%O ~H
t; en m e enC~ N.t N NN .q04
cq
HLn-tLNN O
0 *LA C-4
M~ NT.N. *Lf4 L)r 4 C)%0
O-H N 'OO) f'Tr -C)% t ODeju

*1H
Ln 00 -.OI -0n 0 CY) D .Cl
0 a*44 0Of4 01%rirjr Lf V)

%Do~ .00 - r0-4c
x W *n * 0 M . N
C14' MH( Hr.4 t0
Hr N '4'
goN 0I ~ N-H ra4 '

2
r-.. r. r" a%
4 -- ---- V r
'j -'"'rq'nN " oX r" N
.01 0T 00O
- a.. n -. -1 -= -ý -N"
-0 - -n -N -T an
-0 -c- a a
dUvuo
gln~o M T'nqTe'~M ---
Tu~rl31T vu
682*'4 nr 'aeosz 3 PosV 4918i W02; *bfj

PD+TE
30 ZE E3EZ SE H
222
ce)L 0 0 r-4 It L) 0 n N -4C-
0r-
Lf 0Lr)0fr
L( O
00 ~H ý 1 00o
OH'.
Lf OH ON.f %0
Ulf4N fn
:1~i 1rc 1 M-Lr--Ic' co H
Hc4N -4 vo~-4 N C4 f-4 'cLnNor.-.

w
) Ha 0 0o %0 r-i r-. * Le)

*cf) 0 .c-
o04-) 00MC ) 00
cfc
H 00 *r
dn H *Un H *LrO
cn ?-A c
C'4'~
Ocq 4OH
' eqr C,4.
H H
W4
0~
wA H. * O* 0* O
H- 4~4 -- ~~~~~c
r-4LAC
cO2 )N-C.
Hf Ln CvO N Ciý
-M
CI - -I
tI
-1
o
C
Ln
o. a GoP ý % 0L
M 0) 0n ' 09 009-l % 09 - 09
r-4~ C)II r - - -011
* 91m ~~ 04g~
T.R in~l
____ ___
u,(n ____ ____ ____ ____
'ON, e r- O 6C,
8C14O
' N L 0 T vr-
.
Lf) I . 2230 c7
H C14
O% .eIt ON CY
0 CH C4CJ Cý 1
0H cqJe
aNa cc)0
Lfl4 ;2 H H O .ý
H 'OH
C4 C4 CN 4
0 -
0-
o4 cccc 'uc
w 4 . . 4 *ý -..
1ý
H~r * r-4** - 00
-% OO'
H .?OC
L' HaT
r-4 M C
E4c~ -44C 0J4 OC'
H1 HI N~ NO *CH~$HN\ "
4JH
0 ONnIi"Hr-%VI
V)
o ~OIIC -
4
~ w t
-ý
CNnr
-
1
0.4n q -O.C
C4 q Q Cq
H N O~ .-- i H0. e'j>
M uu~j 08 108 U0c 08i
ca - - -
O)f- PO44 u)O -4.

r-I c N r- N CI 22C4
fH U H? Lr0O ON~a
00 (r)4~ CYOmCJe
'AH
w~t
HC
~~v
00 TO C14JC. '04(
)'C *. CI* 0 0CsN C
0 o CIC -4 -NH
ai H-Nen.ar0C0- mH1- N wwMH .

to0Lt "'0r C - ~ w'H'.
co -nT CO Ln -.T to4. 0...
C CO
r. 4 or.
0 PC00 -
P.
p Na0, OC _ 4.J~$4O
m CO C-
x H
x 00 :r~c' co- x,~. Hr-.c C.I CO-
0 C'0. C ... 0OW*
a*-I H0. 0, f
*Q C P**C14
N~x .14 -. 0
Co C zr 0 r o
%. 708 7Ný
M c'r
r_4 m mO *n mCO4c 00 *n
_ H 0CC . . 7* . N *ý. 0
H
-I n
*
C %D00
0 ..
~ H O m
%D
Lf7. .
r-4 0
H *
L CNO
X ' O LN0 *C"I N.
*0COo CD
zu, r,
N q H~C~
H 10 _H
'.
0 NzI
H4
" . * Le) Le) *-00 *C) *N
Ch00 .nN _
wN r7
ca V N O N
.zLI) 0 C4Qp D Tr- pýce)0x f

0)w-I t 1 i I t )e
oc 9 OScn
ire *u-A
m 0 Ln
HN
CONc)rý
II %D ) ý0 f
44~ - - N
N I0T17 IT r I C
44 N "4
LnZ 8n L0S
422
x Lý
Lr% :ý4
4enneI~
0
cq
CD
Z
0- -0 n r 00V Tq mL)L) -
r-4oC) 00- I 04 CY. 0*.r

H'4 r_4 004C H~L4 t',4 H
C1H~ cf)C1C4
o) 0- 4 H J~' ýc 00r -4ON HHn

ItHf 0 I'O
e..L
;0
r4 .
*-N *H
*o0. .. 'T
1
r..
C14H~
*-C4 *nI
0'
1I 0r n N0 N~-irl(7%r-O C14
N r-('LI)e, f-ON'
CV) *f * *C> *0-r
0.-
00 r1crH 0~ HCrr- tt 'r r- \oo\r

ON%0 VO)c 0*r tC% r-C4
f_ -I %0 CY) _ _ _ _ _
00 0 L-:r f--N ID cr-I00 .

0ID Q -4L f -4\L
m -co --- r4 9~ co.9- C6 1 -jr--
f-I e-I * tft *Y -4 0 1e * It m 0
Zn 0
tI-00o0 _4 4 T 00~
m `4O Q) in0 o r- cq I,
Kn $4 C) r
1t z0
0_ W_ __ 0__ _
0: 00 % c %
0 )N~ U H 0N 0 ri -
-.Tgnr n( Lf n r- 0 *M w. 0- W~ IT
*coH 4.4 *1'-
10 % 4t - - -) nV -r 0 7 - 4 r
H- 4) -q - .' -n0 .N - . - -n
-% -n nr4-- -T -
mmI
--
I 24 in m g 0
-
9
nr
(4
. popnaPOD 9TS
2263paaajqna
C- - 0-- % 4 f-0
0j HC 10 04 cn ~ I f 4N N.0 C'J *0c

%0
:3 en %0 . 1 00 * * .
HC'IH -4C C\J4 - 0
r-HH-O C~nC4 r0
Vbqfl
H 00 -T - f - -4.
*v- O 0N) CY0 'iC H C n % t14 0 I Dc

4J$.I~ coo. . oCc oco -' L oH
c~-C
COX U3
'1-
r* ON 0
r4H
H
C4 4Ci
N4
*LAC
l*,
-
4
.'LA
~
41C 140 0 o 4C C'1'cHH 0e -.
04
w 0LAOL0A cnc cMrH 0 (n P-4 L
ý400en h )%%,oA CN c rCILA

% N@'r- -0P
.* Ln . Hý . %0 . . %,0. * Z .
E4 xLfA
C'4 C'4s ~ .C n L ..? . a% %D LCO %
0- %D -N 04C- - I % L -%
M
0 ;.o '
'C-n' t r-4en -..A j r- . C)O c
M -4n - '0-0oo tO C'H -1

14 0t44 *n *i LAV. * N
cN P', N
10 HN H
mN *CA * .cI
431 4 001,
- -I --
-- w
0227
0~ ~ o00 D%
P4C
0)~~H N n
__H__ý1. 00 I
H~
a~
H -1c~
o~H ~
~ o
0'~
.1, %0 cn m 0 1
41 p 0 1- Y)N * '
0~~ r.-44- ,400%
H uýHo enLf aýc

C) 0n H9 V_ o . J
H i i-I -4' cn -4 00
a, ToLmo 0 Nr- Nrc

Oi I%4
0 H* 2 )
4Jý *ý C
00c 0;-.
0)2
r_4 I *i :
r-4 *n C44 r- r
0- r-4 -4 r4 C~-4 u0r,

in~ om.
r .4 L JD02 -H
E- 08
#-ueC 80
4444
N-c Cc .I IcK.
i-IH
r.04 04 Lr) HOO4
v-C)H ýoCN CON-% 0 tO-o 0 enH coCO0
4)x m n o . %o~6,. C140%u 00C
ccH coHN H* uC14.
N C1 *
OCUVNu 00C1 0 o
4) d) H N MN NN0
00H 0 0 .o
0n
41- HOý0c CO4 HC
a1 H HN4 (" 4 ONqN
CO ~ C CD en Or
0 E-4t Ca. *C O- * C Hf.
"0$ Lfn HC -Ivn-40

Q) COC COo 0
0 'K C;cnC
C14 H C4
N~ )H (
E-4~ o4 *~ . M -CV -
H-M 0c COO
r-4 %0 0o~ 00cv
C. * N %D . . N _t en P* H
On 00 O4C %D. M
)r 0W
*4 N C14
04 * H~%0 r. Le)
z U HOt ~ r
Co 'T 0K0
U n 00 n 00~ N-.. StrN-* . M
Hr N-O HnC~j N> . 00c~
* N
z NCN C4
MO~~ '
If0FZ N-
CV %7
E-1 -I0 H. r-4.0 Cq ~0%N C O a.n
r- >% : ___H .H
0.
4) H ___ _
44 'K4 'K 'K

>4q HN N...~-.~''
H- E-94 u .O 'nON
N- r"- kn o o 0c * H.I
0 44 en mcvCO v-4. %? N a -
" H
"4C N 00 N % .4p
Ul N HA ~ ~ * ~
~~L6I0~%
*AS ,%~~A3
be
_ _ _ _ I
po l d Put D A
dwl8 09
OS 0 08 0
-bUOL00 L91 00z 191 0
r-4 UTJ[ Swu T q;flOd TW;TU9400N swR T -
0
* N*pO. L~99Z r9~7t TLz
229 .
r-4 en co mw00'
010 r-4 ONOm 1) a)a NO
cd CQ rýFC4. Cý l*.dL
H-'J M -I C _1
0 ~0
H EI4 r-c' k~oa H4%I1ý
0-N PQie C'!'JN
%D- *0Do ct 0'

r4 N CO N NH
w fl 0 r 00
2~~~~N o Ni-ir
%coc
V)0
to H l-
P-I to -Tl
m -4W Ic
Go - H9
cc - n .t
V-1
r- VI04 N C1 OH4 1
44 -r en N -i- 0n( L
9Ln
Go 01
t CN 0%0O 0 rLn~. -. &tl%
Lnf~ %D
)H
H
0
nH q00 iar~-I crý -4~ ~O
m4 C C4j* v. C~Uo
W)% N -~ M .s
0C~v- H *1 *- mi
,-4% 0%. rILn Lr N?

Lnt-1 .O
%00-4
00
030 D -.. %0 %D Ln CI4 Cn(~
en (n0%
) . u'). 00 . 4 . * )
to 41o0
W) rlrl Go *4- 1-4r4 00 n 4 0 C4Hr.
r..40
A.-44'1ON %D %OO'
(cC4 en'O NO%
%0r
.H4 *j. % .0.. t.1 en
In c'14- . P~-4 mI 0% M "
wO. 0.%.00 %0%'*'0
CO
CA' P4 *- .-4 * O. -4
** *.4 * ~ rN r4 N
LM4 0%
C'o'0
NO C c -O C 4C
C' '.
%D .. A. . O**
d- -O a0 -a
VLI %DI' *A'A%

C4'V1 %
%0%D kv Cq
e~q~s 0 L
a
in4----
41 00 4t494
V-4.C4r 0P4en 09m 09 9 00 0 9
a" a aI 94e - Va 1-
0 " 0 9-t u 9-.4 %Di

9-i. wF
r.i
'4D -4~) Co 9L 4AZ P4 s
2310%t 04
f
0
0- C
40
v-
4 0 " P404 N
N-- It OT
C)% IDLf)OC Do L -%
ri m ~Co ~m O
La'IHO
C
00 W) %0
H1
C1 q
0r-N cl
Ci14 V)00
Hj
%DCI
0 . 4
IcN *N1 00 ~lr-

*e' *Hi HHý 'iO
% f L) 10L C1r0
ON - 00rH
QH00'
*N4
r HHH *N00 *HH
0*CNr TIT 0
O0 ;2
or-C' r-. r-4N

H Cc--- H q C,4r-
f . t- * -4 . cYc .
to M w all 4N -1 tnC)%c
o nLf 0N Y
0 NO64i~HO ~ '
0~c O\N H r- r-- 0

00 0
N~ 0
E4 x C) . %0 . . H) .* . * N
HHH MC *ýn mCNH %D~*
HnCf) HC LflOr mNOO

lL) n0 0 UP)C
CoC
1-4-H 14 *CYOn(7 -40 "o

OLn4 C ~ 0 0 4 '-
r r -4 MLI -1 C'14M HW
H 04 Z 4.
r-4
NCY) 4 4'Cj
0 N0 0 %DO 00''CV)'
o- Q0 r Cf
r4 ON. -IO m C4 C* MC14* 000
Nf H m.
M r -i OC14 a..'TN 4 NO -- -
E-4H~C * ~ * J * vI'I'
0* .l.Do
p3N 4I P
o~~~~~ -l
ýLC4 nV - 0r40
-
4
~
4
0.mj
me
0 08 0N 08 V
084 -e 08
M n-I D --
nOND TD m T-4
-4
C1-
Nn tnNO
A1Z g8Z 98Z 98Zva

12811 19s
Tv~ni~o
232 .14
q-0-ION-.01--------ON
______________ON--01-
0I
U0
0N
04
OIT.
cn.
MN
H ..
CN14v
O)
O-Lf)H
coo -
1t
H .T% .r, C)2 . .u .Y
Nl KN .1Kq N
0 M HO0
" %0 IT
caw -N 4A .
Hq lkC14N14' M 04 C4 0
0'
0
coH
O-la,0
Cco
~
c~CO
[I MC~ H M CIC4 en*q%

.H '-I 0 00 Iqco 0 OD 00
0 ýo V- NOGo ONON O
x1 'o
C . . . C) * ON r.
o- e - -y
7%G . -o
N-
0 t 14.C
1-4 Lr W HO C4) N 0%%0
00 H~ OH
w %. 0
r.4U -4N enI 0LM
w On f -
. m*n
O4 . mr 'n * ..
1.4 H ý% *4ML
N in
(' PQ D VK 0 LM%0c
00 '~
C14 .0
r4 en M
H. cn -T
NN) .c . u
(OD. '0.40
U L,(Da
m1m1 .gh. LM-.LM
3':1 ~ ps71
%A3 en 'Do(
N'.O
5.ZflsC1 *d4
-"
1W
9-L J3 Cb a
09- . (a
UIIfl *A*3* Tlrtoe. TVIUZ O
Tlnq6~
P610,
MVJl
pavaOil
L23
f'0I co 00O_:tC ~ M~OF- N c
en *r%-o
. . r .0. r
00 *N Ln 04 C1 *4 H c H r-
H coNO :r~ )
LE 4 %
CA~ CO . '.H H *Nf-
UIN .L'LUPC)N -. 'H 1 cn,-I
'-I ON en NO~C' .4H q 7
41 000 t I.THN 1- NLM~ H -4 co

0
CO .4 . CO .
o LM 1f1T.4r .NNr (Lr
I . m LU m N qH cn4HH
r`4
'-I NO HOO r- H
* M Nn ,*a C 7 I
%DH H H 0 % % Ic*N%D
9-4 *4 0* 0'
(O H'I .
P. - 0**O ("I
t0 W)
'0 *0O 1.0
W4. P~ N000O r- en %0 ON %0NN'0
e 4 Lm 0-. enP-c r
.4 0 N
t0 *MC4 Cn-4
M 4 -
- 09
Goe P 09 09M 090
-M - 4- l-c0f
Ul 9..J~ 1-z49
94 UPmm
41 2:.0 a p ca23/0
%0r- 0c . C'Jt
"4 ~ C
r-
rLA
Cl
LA
00 c
.- o*
. H
* -4
*
co~4' ClY-4
w 00 -4 NO Ln H coH -1 - N %0
cnn. *n Hn r. * IN H ON
cdL~C% .CH lO
41W0 00r- C-4LI) It r4 N t- -4 co
y44 H- C:) ST r-4H %LA CV N %0Ltn
0 w LA LA) enC A *r~ - *-t
CL .*CV *.4 LA)cnCV Cl)H r
r-O OH4 HOO%

0 C len
H Cf) co ~LA -T -- C; lHHC
j-CO H * * r-4 r4
CLNj *LON
col *YNOm
(n tCD --- -D
HO LAO enA cnlLr-4 r
zK K
H 0 H in 0 ----
1-4- -H~ ki % r 0%.
u- c .4N
NO .' . H. .
-4 w n Cq '-0 n
cl n .%10
LstIUýCN AU; I 0%~ Cnq.s C; -Tfl-f
n P-4%.Q08 08
'44
4) N r- L - 00 C
V-4 44 (n(
~
. ~ ~ C Pp-49 ~ L6 6
CO ~fne MC
4
A-1.5.2 Cobalt Alloys
236
Mn 0 01 0 0 C14 .o NNiA0%4
r- 00
x
P- t CY 'o%T - 0.CN N 00..I
:-I
*n *%
a) N.jM1J
co
0)
r4
c . 0Le:
VL
r-
. 00
0C M*..
. g
%0KIr
*e) .MC0
H % uli4
.,
m
. 0
C4 -I
*1
-4N
C1 %U
zC
A 1U4 'T L 'CN)0 %T

IAf0 O L)r- P -a
0 -4- q - % - - 0.-T -
0 ) 4N f- - %() I U)H1 -IC14

OVu-
0O CVIAQ 0'r*
4- 0 T, 0 Ln
MV. *
ON , *C110 V 0c
H HON 00C %D MO M % 0Nn00 r 1
M
3 rI *IN -
001 N-.)
. w *. .I 4.8w' .0 0 -- -
*00m
H f m. V.H %0
r.O7' ~ N
CJ44 InI ~ODm
r~
"44 N. *f C14. N A r* .- N~I 0..
04*r-
H P --1 00
.OC %DU
-- 0
O'co= % V4
H %.0. 0.4* . % C
M) CV rK - - r V - 1
M1 CN0 0 V-* - *04L On . u'.

as
w 14 r'a M. L** (h' 04* .4 0 - 0
D - .Mr
C)0. '.0. r, W at. co. CV)

E-*1Ný. H%C. HN'.
m.. $4 . 14
co 4 0 - f- 0, -- )4 - 0
0- - - q > %r nM 4e -
Cofq~ *AV TtMq. fWTA4

TWnO in 0%
go 0
044% or. Nr m0 T 4co c

-eq0 % % M N-. f- -V 0i
due u4 O in~ r4 QOTM nS %DM
N % 7%NnTg Pj PWt
coN iuno
00 - -
1 dwe OOZT 9!E L.OOT 09 0I 09'.OE
'ON *POD* 66Z OOE TOE ZOEC D
237
-> NUN
%0 K0
4c iKK
1-4 0n %-~0 CY)r.. 00 oD 00rF -.T-t mc
4)-r0 F4 O ON L1nN m04H H %D
m4 cn- -j r- r-I -
W 4) 0 00O
(D fY-4 It o cnJ~aON I 'r-.
Nd * .. U) 6 -1 . . .* )
.. H * Ln
% '.,4
% *H)rIOý *0 L
9:004 4e N.A r-4 mc I-4 N'4 --TN H
0 '0m
"4~ 41 H 'C00
%00cn, N L)00
'.O 0 00 '0
Ow ND '.0.
%D * 0 *o 0.%.O 040 -
0 P4 0 . e 0*-~ .(% HC) . . (D 4
0f . " 4. c'n r - r-4 C) r- 0
ý..OLAC,4- n UA qA N, C.U
4 c O-,TN-
%,D . 0% . 0 . H*I . 0:
%DU n LC 14,'0-
*H ' ýL n'0 en
'.0C
%D m*C') MN
00
C ý
H 10
Nq-* cn *c. q HC4 HNC' N It q r
%D.0 . 0 . 0). e
t; C'.-.t *0\0cn . C-4 Hci 1; 4 C
Ln -T mOo' N~ '4 cn N 0 4 I r-!-'
C
0q
04 4 -N -%Dn r - oT
- r r
N cn H
-* -4.O 0C4 r *'40A Un 0,U,
ý4 I' '.c.Ia c ~ O0 .~
Lg 0. 1 , 0 n
'4t"m4.4
4)wN '0.t- 04 IT t N --
04
0~f -NP- -0 M-
414. N~-4.
%o co 0 c0%0%0C
o C
44 %.0. IA . O .
44H *I%D co 4 4 0%. F-4O '0

r% a
!.4-\c) f *4 0C 00A r4NH
Ln ~~ ~ %r- ~
~ CA.IO .~ 40
0 O
.5GT~nqT9M *A3*7 OA*S TTnqTOk1 TTflqTef
04T I01EVz
SdOUZT OOZT OOST 00ST OOST
WI*ol09 09 09 09 09,
-U -m M.O-1- o -O
uo!
"ON1 SPoD+ COE 9t lot E 0
238
1) NCI4
C1 LM0
UN %DN Cn r
X 0 * ~ . * H N **
ONC (3N NH c m0
0 0eb
cn en N n n 14r
cn 41H rý o ,C\'N o % ým nO
4 'T
j.O. * Hn
.0 *V .-
d4J 0 (3Nt- 00 % m O'%N 04N )000

co0L *D%0U *n H 0 00 HO4mL
CUn C-4 't. n N *-*ON 1
0M%D* '0. n 0I r",*n %0 %D H%

Hn0 0 *H e en n*4 .0%% %DM
r0 - % C4 it1r .C;O
o*. ' . 0 . 0. C~..
H ) w. a% Lri co%D1 -t C-) Go N0%c

C1
02 to~ 0 N r (n- ON~ \0

~j H .Nm > n f4 1
00) No r'C Ln N O-NC

H14
4 I'-\ý O NO Oo0N C4O% .t co 0
un * er-
cn 7%4 *T
H
a) 4)m e M-I -
130 oe 0 0c .
d H 4 0 N* OU(nI*%o' *n
*00 0 %V Go H **
N l%IM %I On4N
011 ax ca~x ba .
LIU TVW3ON Tlnq~oM A*Z3S TIEZON TTnq;SA
OTO*T 0T 01T z l
due 09 OU~T OOST 09 GOST
[4634 09 09 09 09 09
** .Jn1SH 01 d1 HIt - p
.ot~ POPO 661q (K

4W3Vy A*.OCT
(1) 0 **0 ltH%
*-
60 P-4-4Ne'4 *Ho, c4
wa) .o % n L 4 N%0 r -
..l4 H fO l- l PH
O L I F
0He-4CN'N *e'I -e IH
4J 14 % .I
0n %0.O00 Ocvl 000r
00. C4 rCn c* 0*
p. - -
0- NHC14 N -.LINOr
1
-4 C140 Cl'0 IT 04 l
0' H
mrI,- M% N r-4n
o.0 r- ccI
0 0ý-IM
-n , % - - c .
N~cr Lb-1 -
S4) 1.4.' ý %0 - -4 * I-,..

- U) H3 H) *O-t
0C VOf *D10Mco
H i ci -% -T -DL)e
In ul 0 00NIC nL
H
r4 H-4
- LbH0F N O 0 Dco0
14)
)44 0 .. eC . ..
cnl"4 - * -:r %OLA1 - M r
0) w -4cnN C4*NN -
CAi
40 r nC14' *(4 M N r
MLI
ctqo *A.I
-4 V4 0 TM)
llq~ IA3 Inq9
~ 08 [OUT j 00T
bo&09 L 09 09
SIU OT -ptmoss
aY
Oy"94 0S 'A*00CT
'ON opoa* si 9TE LTC
240
i- C0 N~jc oCIY' ' 00 c
oun.
0 H HUV)0.0 CI*
NH
bo rr~r%%
HH .ai - 0uo
OH
W0J
0) 0
0
~~COL1'
U1 M 1 1 -
m W ON *
00 * . o a n. '0 .
H %'0HI *r- (D. 0e
00
'0 0 oI4
(i m
0 x C140% P4 rorN I orý.r
0 w 0 0. H .t.
U) OnLA
cf O- COCV*
H-I
en~ HNCA
nNCO r-i -4t
%D L 4W '0*
C
r-or-.
d- 0 en VA-1 4e r q% a-
r- -ItO 00 ~- CC
0% NN '0.rq-
E40H r--I en 00. "1 0 OD
Ho -4 x 0. ,4 4 o 0
1-4 -4 .4 z P, Y00-4 LAN 0n0* 0( 0 ý
rnIA r-, f-4 00-IT 4 *1%
HA to I H I-
'T 00 - 41 CC 4c W
000 * . CO C4
'0N
00 Ln rý H..e L HIA-
0- H4 0 0r-4- * r-
%O -.0 N~.
1
CIA L4 Ln %0
n 0 .
-- 4 ~~ 0 M c
P4%00 "40
SA
0-%0 C4I
10H
LSIU ~TtnqTvP TTnlqTD I~ 3I
000
10~ dual OOZ 00qT

*4 boaj 09 09 09 .~.e
f1 UTJ m0 Swd0 SL
-~C .9 -4 tC4
OTJHT-H ZT-.tH ZT-IH 1

~ ITC
8O4 61C OrC TZC
241
-o --
AD00- OLn %D 0 a 0 44 tI
%* O0 ri
-f. 4
4** -4 . 0 . I ~ A*
.o r-
* 00 ON' .(cnON rý n 0Co.0oI-r
%0C4C~M-C4- vi- 4r4H %rW j LHCý -I CJHL
IZ~c4-U - ~ON . r-s'4c .~~~' 't
ca* C-H~r ~ -~CJC~

ONJC4r 0 I 40 (no LMfo
$4 0 I00r 1 Orýf Dm0
41 0 '-r r--I V C -4 c -%
n0
w '4 0'O %C GO,- . '--cO. ru .-a .r
0 n . 1 - -100 -0 - " Y -C)0
1%, On -.0r tn 11 -4 *OO

enCV crCA P- cON -4C4o
U
P0 0. N r -t
ON- - c n0-7 -0-z - -
0 n 'co r-4 s4 Ch *-fi0 r- n -4 C1

. %Is
4 * ,444
0 -N .'%4 r-4 a%
N 00 % 4ý n Go~f 'tA0 r- '0"41I
n-
.ON I 4 .o4'ml'4 .oH@oM
No 0
t;~ 0CD C %0 .c --- l
00 '0
'11 D0 n 0ýht 00 L
Ai~
(n 0
-
09.
-M
O¶
.O
-
. 0.9
-
0M
09H D . 09 0 9% 09 09 h049
04 ;3~ cn %DOAT4L-T jc 8u Nuoz
Na
o q eu N
1n
0 %D-~f
LM T1 7 0P0L lo-z
00 ýc
Ot-ZE1.
caz
242o
C4
0 - P4
"D c 4 ': l%t %0N N .
IA-1,
5.3 Copper Alloys ' -
4.-'
243
arcn Lfn~I
Nc~~I~
'~D-
H N 0
OD 0i nrý LtH '~-! 1ýLf rý
r-40'O r H 10H Y)'i H r-4a, -- %o ,
rr- NN (~- t -i- H -4LI) \D

4) m D.
" Hn - .- m*N \,
H00 -H
ca - ' r4H
0 CL~ L0H*NL~
r-4 0 ON0 0 N-) * Lt \D-

' C4 t- LrCi0 %0 - *hz0V
0 w nNLn0
o. 0ý NN- 1 08L~ C Nl LC4 lN-

4H - * e~ N *
ri H _
0u * ~ ~ HHH -H*
rH o Y 00 ICY *r N-oa Ln r
4J0 - C\ r0L
L~N 0
00C
-tr'- -4 r-
r- Cu-u N-.4 ' t - *l f

0 ) 0U u Le)
*N - -.T* r--Tm -J ,* 00. r)0
o 00*04 r--
OHt *N-IT- .u -H . IH
U~ '-' CCn C4
C14 -44 C) C1 _ _ _ T
z U)
E4m VI ' ý nYI )

.w
to ~ -4.
r T -- -T. co
p- rVa .4 N .0
41 0 0
ma .
MTq~ A' (D nq. C)..e
o
-b - C -4ItrN-
--
H'T .j 'T44244
-4-zr ný C14 0 y)0 - C1 I

00 c
N tkOO V)
ýOO10 H
(7%S.OHDr Lf) I
W 1 % esjq x c'-11 00rP- r- Ir 4N o
Cio Ný c-.IC
X3 Ný . . r .* . -4 . %V * . LV)Lf) CI
06 r-4 *f N. ) . ,T~en . *f . H*-

**o to H r- 4- ;-4f 17U ý4t *-*' *0O
P4J >'r 4H H ri
coWý V r -0 CY. ýo r--- - 1,0)C) nr
H o1, 0 - CD- n rlýo mr ,0
V)4J4
(a0 ',t~ r-j N
. . O* .
InC ~C L)U,0,-4
Hn rl-0 H* In* II c*
-- in -T* . 0 0
1''~H'J C'JH
O-r-4
m~-r,-4r-4 .-4 r-i -4,-4-
o, -4 C ) e ý 7-ý n I -oc
In x~ m- 00 enr 0 *rNDPi 10C1 e DI -cl

*ý 1
IC *IC .I;
P4 H 04 r-l. -q 0Or-
OC 1,4o MC * - -1
'K* r-4f " 1`
0
1- *Lýo~ ON *~ Oh ~ *N-
v-4 *CV), 00m0 0 -.rH 4% - e

x 1.4 0. *f * *1 Nq . . . 1f4
0',' 00 r- . 0c -*r-4~~%
00 .004
-He a% cj NCJ' r
cr-.I'J 'I-4 H J4N41-
H -r-4HH
M -I - -
12J M
.~~e. CNN0 0 %
C -4 4
pfh11X0 W1liQ
PL0u~Q 0 . rý _; WW
1
H .C D% -)N
Na
04 -1 - -k -, -
C) l T fl
tI
PTTOITHPlODHP4-4
O P-4 uC u C
424
I'Lr t4C4 H 4C~r~
co
U1 t en~ -40 * %r
P4 N *n ONi'
'n4 r-4-
.j'L400
1)D 04 LnLf
-
b-I IH~ rI LnL

"4ct0
Q'Dm GoNr- u l 0 N oo
4J Hq N m . . Hl %D**
W
* * ' 0N C4 -I 14
4J*
-H
14C
ca
n % nC r
C
D
0C4
00
m14I
z 4
-,4~ ~ p 0 00r4
-8 -4 00
_____ -
9
H4 4.4 rIr _ -_Iq
N -
H-
GIW
o o aX
I %D
N-dZM P-
CAd14
%D on )I l
- -
HI -H-
V- (A~ 08ýo- 08 c 08)I 0 8 ý08 :r%08

4
.r
V- C aaa
'AI I- kn 4-I
goTIU PUI 246.I

PD0O ~Tt~
cc OD-xC) qc
%D0
-Icc
0
9-4 '-4H
4
" CV)1u O"o
0'
LQ0-'s
~ C14
N-4 0T0
1- 00H
nNH HN H
W 0M 0 PIC1 nC4h
0) ~~co o'.).
0% a 00N0 0
W 1C 0. . Ln . . W 4 0
H 94l
oH cC*C,4C,
(1)~ . HC 44~%4
0.
0 0
rzH 0 O . H m.' O
-H 1 H-T CIN iroc C4 -H4 un%
V)c
tomX D . . V) . . r
H.H- c~n - C4 C1
0 0
Cl)~~r- 0 I~L~. 0 0 %.
0) *
00 C4 0 0 L e
HO Ln O '0T - h
C4 0 0) f D-4H N
H > 0'0- CO -4 .
H-
Hy ON) 0CN
r-4 cn*LA
zz
H Ln0 ~ 0
C *0N *coM o O I
0) 0 4 . z 0
~O 0.H
. NN 14 ~ 0'N.
z0 0 . .j . .,.
()
-
HH 41 tn0 r- .N -N H
K0
0 0
W3 7% 0 0
41 001 0.n
M~r 0) -,
~~en
m CO
> . . O
NNC14 -1
24
co OD00o
Vý00cn m O9v-
4) OD - ) OH
4 3H *U4
Le 0 W
HO H-~H n oHr
93 00
H0ft4 00 .
"q %0 PA to
0~ ~ r 1 0.) HOl
0 0*
H *'On NH
*H
z ~ ~~
aHO.I Nr
04 0 cn H
%D 00 -
N 08 08
'C
C
ý24
0p 0Ln 000
0)00 03 C4t
0) %0 W 0
0)0 -4 0Ln 030 -
C- Cu
r
0 00
0 000
0 M4 0
Goor
x " 00
C)N.e1
0Z *-1
C14 ~ az Z 00 O
n " l -4. :.1N 44 0
C)0
<n r- 90 r- Lrnr
L14 0 < -4 *
U,
U)
C) u
r, V
E-
0)0
L4<
U1UWdcnZ ca.
* I0N-A 0.1
r4-
n 09 09
HH.0 UcTA cA
CIj n
ON~ i2POD--.- OSET9
249
1.
A-1.5.4 Magnesium Alloys
J 250
) . n - -- -n
(n m "K
0 K
4)
r44 4N
%Q N CJ0 -cINH
N?0 0-4r4 O
Or . (
0-0
0ý ui) ~ON
0%I 0C% .40 N! *'0
r-4 r% Hcno M (%en 'T V-1 H
In C4 C4 C4 4 C4 C4 C4
0 CaC
en rK-
4. gA
-4
-0-.
(n 0en
ýD L
.I 4~
)-
04
-.?
0
0%
r,
-hCm
0 ()
(ACO!-
4 A
. n 0%
.nO *L cor
4C
0a- 00 -: 0G n - ý -d -.0r4k
zf d~r4 4. (n
-n
c'4I I
on
14 ~
K4eI
4
Le)
0C
00 mm
"K
r4
4~ '"-PL4
A
4 "K
C4
C 4
0P
40*-4 .9.4 4c 49'
LM 0)K r- "K K- %0 KnL

0r l
o0 r4 0Pf
IAO% V-1 ee0e0%
1- N.
I41 "-
00 eNr 4CO% r4 V!-4%
-m - - hi-7
N--
~LSI 1
1IIO AVS
-M
2V
-
TiUO
-
A0- 1 ;1PA
1-
go 40
1- 09 09 09
0-9; G09 0N
P43
r-
eq
-
1i r4
-
11P4 ..
0r4
-
mn
1
-
94 4
%-
U
.140.wcl ca ( Z t251a
ID0 HL".Hn '.% 0 C0NOm
9 ~ ~-4 oo M
0cO
%0
U
0)
Nn
41 '"A
0 %J'
1CVCý
4) :3
00 Ocn N 0) 0 n00
0N o* w 't G e
%0 00..
4J( 0 ccCJ
.
LMo- f-4'O v-4 'T H CO
-r4 41f- . -4v 0H.'
0 '
00
LA en -T*14 0-It
P.4 C) cn~O Un H'170 00Hn I
E4
e: .i-t 0 `4 O
C; 00 %
OCd 0 _; * C*
to) H ': . . C .
OpV-4 ý4 H
0 40
PO -V 4 r-
me 0
* ~ VIN
c7-%
fne~
'OH kn0 P4 r-P 09t

P% N M kn
LA fn252
0 C-40 Lt) r, r- aN -? 1 ~
fnCNO a% 0 0 -Toa 0%% H 00 0
0) O L0c' 0r %t
r- H0 9l H0 H 0-4
cn -- -4 1% a%-
Q) N'r- 00 OýD Lf.4
01 4) H1 10jCJ
H *Cd * nL
r--
**-4 . - lH
i c.ý C4j CNJ Le) C4 H '
0.
0 E-'H4~ A%
.H Hct\ L~(H bf re
AHH 0 N nV H 0% 0*ir- rq0
0 -t 0 C-4 4
X4
00
rý 0 0
Hr'.
N
0C4C-r 0 C-4 N Ln 00
SI .4 *4 'T *.t .*T . N
e
C4 t% cý N H Lý C4J H4 H4 H
n n1 q
5
W0 C140%a
.I10 I %M
.0A .Tr 0 0N
E- Id H M N~ r-4 t-. r- 0aLo a
LMN H-4 HH H
a) " 00 ýD r-9 - 00 -, - M0 ) 0 0
V-4 ~ -4 r-M r *d 4 ý n 0 0L Ie Mr%

x-N ýAU . 0 . 0 cOýH '4000
1-4 -4 z A..C)a 40 ný
ý-4 r4 " ~0 4 P 4 % 4 L P4'~ ~4
0 7 U *N 41~-
H4 14 14* 1*
m ~ ~n . M ~ P- ~41 S
e - MN C4 a)r r4-
LA3 4c -.
)I 0 WT 009m
4dnen 1
N0-ET0u
r-N-084 n09m c 09 09 -4
x9 %W 0 - -n
b~j A --
1%0 W 6Z
"4NN £9! -
u u d
O -4 4
epo
t9C4
25
4
9 3 -
O*T 66 *T*T 0
00 Hr
I OFF
*o
4) 4) , -4 iI C
~f41 H cn* o%
0o
H %
D cn H
0.r P4 ,-4*'
-4x
j. 0A*
A. 4 -jc
2 Ln
14q
oc4c
000
-4 -4c)
4 0 )
F0 9 0
N 44
i-4 04'
1-4.
0c
CA4) 00%','
ý %04
4 19
A0
In wcn
A3S
r4 0 r-4
P4O1
00
4
\
00
02
F4 02
C'-4
254
enC
Do r'-o %Oh- cc
>1 0>4 P4 4 ~
p.4 C4 4D
00 ~
.~ . 0 r - c 'j 41 go V-
H *- *l m - 4L
0. m 0 00
0r rý cri 0. 04.
H. O 0.
C1
rHO' 00 H07
T (0
T4"in *HQe'.l 14
Hn Cj4
ý r - 4c'.J0
%-4H
rr4 N) -4%D rqP- r4r-
z0 w
~
E-4 ~ mt!0 W
>, 0 . U4
N P r 4ý
04 0*4
0. 0 " . . NV 4 z 4
uiM u
LSI 1-4 CIA
Ji
O4ý4N r-4 P4~
-a -
.~ - 6C 0L~ I'
ito -4_
f1
00-fro
04:7;- -414
- -4
-- ?-
I-)
4)Nm TN 0CD
41 4or 00-
to o V-4 M
o0
0 04s.
P4 -CN rq m H r-L
0N m3 00'T-.
-E-
WL -It 0 IC CoHOI
z r-f C14 H * o00C
H u. 'V) -1 - -0
H Lf "a J c HUH
HI, (D 4- a 0 o a\C
$4 14 'DC) a'0 Cýc~

a) C') ON'4
>( *N4 H -HH-M
1-4
0. q m0 oM0
-P -
%o
W% W0-4
0. 14 -'D -T
MP,Or N.0
4.4
%0 OZD
- km. - 1. O
cn
- .A4N
N---0..
itt
A-1.5.5 Nickel Alloys
2"57
03~ H
OH HUVLt
03~03 00 c %0
03041 0 OD
co0' cr-
0'
HH H
41 C44n 0
o Ln r' 0* (7 0 C1
00 r-r ()C1 .
a%03D* N 00 C4 8
to
~H 1 1
00 NK'
oo H0 OHO 03q S iO'.ý
MO
00 mN wN2
I- 14 ;~ 0*1
. 0**9
H~~C *Ne' ~ 1 c' -4
1
H r-4~ H) 00r- /0-
0 04 02 00 02 'T - e
H C- 1 1 -
0 30
,-II c 1 -'T a%M

0 0 0 - 0) 0 T
du 0%v9n n0
V-4O~ 0 94. .
en58
* 1
00 00 0
co~~( coto1
00) H LtCI O .Ir'"

11a r-4 CY) C'
%0 1 HHH Ha 0 0
4JH
I C) -ýt
. Hr acv) 00 u-:t 4 $) LI OGON
0 N0
-H ~ 4Jt00 m H-
0 w"N V)(% N0 Am0
*~-4 r- mr
0 C'00-4 r-.e0j
04 * H 74*~ H a0
4. Lf)N * N
02 H
HQ~.
(0)0- I
t) 4N N 0
m~ w H 0C3 u U)NH r_ N
H ~ > w N HQ00
00 0)4* *n H
;A "c 'T.I0
0
HO ) 2U Hn C"
I. H
ý4
Hq!-4
4 V-4 N N
r_40 M 0%
C4
Or-4
'o- ) :
ý4
4~
ON~'
0) 03
.c3
e
:
*tn
'Oo)
-
e4 . -4*4
N kC co A
N1
U)
41i0
0 HC"N
r- *-
ýN
cv)~ a 44i
0S
*C-**0*;-T
V
n
* l V)~
OU'
%0e
%D
%D
H
Q mN 0
4) V.4
UI A4j
'ON4 *7 4poONC
u%
25c9
J
a)r
0) *f -It .- i0 tr
.CC'
Ia' --
0Noc '
0 - 0
4c i
'-H CCO 0 - 00I~ (ONtfCi -It
0co r-i 00C) o m~c 4i MorCIAC
-H 4-1 r-4 0 ) C -L) C)r-i C,),jc

0 CL) 0
ca MN cn . r . * CN .*
0- w -n -. r 1.4 . o0-,
0A4
LP) M 0 % r- 0 00~f *00-tI N-:

EC14
E-4~~~~ . , Ij(Nr
-0 - 1- -1
ý a
f)C)c
000T 0T 01"C 00.!) c 001 -
- - -- - -
Is - n -Y--- 0 r,0 4 -- :
Vo coV -. tI -
C40Ui
-
c -r t -
E4-
Ln cl 14 n .2331r I
- n of,-l 0 0r l
4
r44 rP ~ ZE~9~
2604
00 co 14 N CKrI L - l
N00
N I-.'l-t H
j-4Lf HNIV
a) C31 0 . N *
0) * a
cm Hr-
4 *f L c
M__ 4 C4____i 4
00 -1
0l r-r. C14%D0 Cf) O L P~-4

U))W 0r-cN . C* - 'D.
-Hc.0.0 H *
0H- 00 000 r-4 Mij C4C) 1$l
41 p E
0 0 enH Ct-U OHLn %D 00 Lt A N
Xn . CiI!
0I in
w0U *-111
H*c,
r.. cn M N -It H 04 N
0. C) 0
r=-' NO CO' o ~ o o r 4 0
3 .
tH na 00 00 00 LnH t'-Cco cor'Cen
0-H)00 0M-0 0r ID
Nn * co C)*n .. N- P*
H4 -T, N~ -ItI, -N C4LIN4.
0Z
r, 0 00~
00
C40
c* O 00 0* C-
-tO
r-4 i
OH
n.
0*
0 .O H -Us NT HC-4N a)
CQ H~ L "10)C
o ) CIn00 Lne - mf% l 00 '
H
H4 f-O 00U4
%4.. ON *
e 0~
n.
~n
W0W
Hn HUo
km 4-U HN
a 0. -) -
Ln'' 00 r- LM V'-0% GON 0% 0o1. L i

r4 r Or- in'c~n
C;0
X CAJ~f
7-4 LM N?~ cn 4 N
H C'Jq
-C' LM %D e Kn
-0I
m .~ 0'* . " . N. - LM *
.4 %fO
V r%.%0 A0. -4V4
u In. DLn I n 4 X
44* 4 4) 4
kd 44 r
-n - - - 4
M -: - -D -
"0"9- -
9 09
- A
0-
r4
~i IF
626
mc -1
00 0 K% CO
CD4ýCý U HOA OH
L40 O'O
4)4.. Hm r- m0 * '. m P-4 --0JOn

En x 4 ý .. r LD .4
-Av-0L 0 C-HO N~
Lrn H(' H~Ln0r-
*. x4' HC. V .4 H4 rN 0
r-C- _
o Ln - 0% 0-ý
IAC,4't r-4 M 4' ~ ' O
%D00r-. ON 14 CN -4 %DLn N''oCh

r- r-*4 * 'A C" 9* .,4 00r
* rr= on c' IT
cK
(%~'I o. N( -4 Hr
- r-04 00
9A* in co r 0o Nt 0 N
ONr- . -* C4 C' *0-
LnC4 04 N-4 *'A H4 cOH
0 COH O
4 14O C'AC 'AA 40 O
4 00
cn* H
co *
0 4 N H- *f
% LNW 0' in*
0 %Z . 0 c rC
' *00 -
0 0 N
0O ON - r- a
'-3. 4-' r ChW V4- co C14 NO'
4J4 cIO
q co NW% Ul -.0nH('t
00 r-O NO
- 4 V-4V-4 COLA r4
40fa ,-3C
ca Zr4
llnq~n N IA'3-S At-3S IlinqlSA
3Sq2peoxo 10 psuazIJ *'4p'1
OT1 0-1 0I~ 2ZL

0- 0ZL
14 du 09 009 000T 00's! I 009
-3v '69A4 tiz laostc 3
'oi *P*3O. c6C V6 6C 96C
262
r- 0%0 -D0 0N
04 Hr0 r-'c)%0O
r-~ r-o %oriX % 0 e
(iC4 .* ý0 0 ) N * c'n *.
. 0 * N .'%P. 3
boH
r-404r-f C1 cn M")r- HNI 1 I
1% 'K 4c ' 49'

4 00 0 0cor-. -cr=
rlm 1 4 oen -Qr0% 0
c24! N 3% * ONfe 0 . 0 . 0 .
to ;20u cN0 NNI Nr1 r-I 4
0 0 wi-_____
00 'K' K K'
-r--. too le H~I'

9) m ,.. C H. 'C~0
mO 0T caO
C) N '
0 4N T N04tý ao.) 0L
m~H HC 1. 1 .A
.40
u 44
%0%
OH 0r'-%1 I. M en * 00
MO.-C
.~~ 4 *0 4.%0 001. ' N .%*
10
,* 0. *0 (O
ci 0% P-40 N i .40
to D. - -I- V- - -e
r4~0 0'
m 1 H-4 Mu-A- C4 .4,0 Go 4
In mh -4
r4a Go -C--"
LM b ' 04 1V 04'N
N. N 0%.. N. 00 Go %m
En4 H~ x J *1 4 41
to% .4U0 fl1%
0% A .-
H .3 4 -U H .4 m' W-4P4 $4 s-4U
fn6 &M 'e"q
4W
vi4lGoDV4 en r'. P- G P a
V-.0' 04 N-~ 40I sn'n20
4
z x. a)* .. . E' ý 6
V%4 e l - 410 C-4 '.4 N
rN.- w HN
ý4 tok A - -n%-%
a- -w
4
0 i 40-Nl
C4 C4 V% 0 j% 0
b4
P 4
O0 In(
*A1*
PT'SUUVm 16A3 1 *3
it~~~m " O 01 m OT i1 0
Ao~*o.L6 6 O
005 051 g OZ26300
0 ý4 0 .C~j *- 11 0ý ' -4 0 *ý
0)i- 0 HI0 .0(: ) nr

H0~- C4 H -4 0O4 H N C14 *C'
0 M0~ , 1
Al >
0v4. -tON I cor-4 rJ 0) ~ ~ n( l14

U0%
4mx
*C' a .C cnZ *C4 I*.-'O
r 4 ~%0 NC14 HHr4 4 N C
00 4'
0 m'0c 00*r~J
41W0 nC14 C1 -3() 4 J (nC *CN

C
4 -% w 00 -4 n
0 * m cn 0 4'n 0 D -
S0. 'C*4 N rIr4. ý4 04 4 (1
C144
N.IC144 0-4 1J 4-4 '0 0
w - 0% L 00-L r-I1
_0 "44
14 10 e4 *.i 4 Cq
) A *0 ONr-
, . . -r .w '4 -- . . ' -
tou.
0 o. ) 00G 1 0 Or *00
(n tI-. 0
-I -4 -
*11 S-44 -4 -4 -nM

e
V% C14 - -m
.0 CD ImeI CC
32
*o~~O'
.:£0 WI P O*U 24
I~dmu00C
OOSOOL1.OU
0N H %00 0%0
4!)
V (h. -~
C'C -I
*C'
0'
0 Na N0 0 C4
to4 -W 0 * %.0 *
-L- H GCO
0 %0 - .r'.
m. 04 Ln Lr%%0P4
C.)-
w %Do'00 00 u'I
z dr . * , . '0
'-4 t;L 40
w 0 j ý Go0 Q7% 0) 00 -4
i~b~g
.. 00 en (n' ~ w
004 -IL' m'. I. LQ
r - -40 010 r-
%me
.40 N,4 *04-4 P,*
- 4 co 0) t0 r
CvN r4 00 1'00
m-
N.4 4)' be0
C4 I t
0 .0
-4
4j-
*-4 . ~- CC*TV
z
NO *Qr4N *
UZ V
A31
COOZO
m 61 U 0
L1843( ITOW uflfl*A
~ .41
11 0* 011n 0-*1~I
jr *P0uo. 0O9 60¶ O9
beq Ee E265
0 H %o 0 ,-I T
MH -It-1 O0 c l
4) m 'eJNC . C
:3H H*0C) C14MoýL
0ý -r 0) - V
10CV 0D % NnH
q r, 00 4
0)0) H %O\H C Nc Nr, C1HO4
ra41M -. * . o0 ' ' C
0)H r-cýC-4 C1 a
4Jt 14 i 0 4' %D4'(, D 0 L

41 -.TC'1' ,-4 0' % %
OOONU
*r
Ltn'r
-: Iýa%
aor-
(7N
H 2
'-l N't
*l *.r o
00
N CON 0'C N 00 r,'0
r-.NN% *A 'f.? a% 4C
z C. N CO .0 H * Lý
U)~ '0 ~~~~ Hn
OC'tC Nd C
goH H * It ~HDf
__0 ___ ___ o)
HO
c0) -4L)I Y 0
H4
HHr4
ul
E-44
z'-.
0a
N'CCO14r
U',0* H1 %0*
oo
-
00 n
-,t
4
ifi~~ HOm1'H 00 ~ 00 0
0) ___ 41 04 _ 1~.
0i 0 .o
O! 4' -L 4H
0 ) 0 ~O'rc
r-4Oco' .10
Hu H) ?-I0*C)
H EI 00 0-0 %0.
r Y%
.C ii .~ a%4O re4
0 . 1.0 ' - jcn -O L .
nV4 V u LM~
A-S qe t
0-1 04 0)
.? ON 00%09 %0 % 09 r-. i'o

0
-4 00 en,
M . H
H L
L1 L9M 0 ~ O4 r-
ca' Z -L V-
I#e N NT*
*ON i~ **pO,+
T1 IZ
aw266
-4
ol o*T
0-1
x 0-1 .
0'0 rý coo %0 0 .c rn Inm' c 0'H'f-% -T I
r-q
or-.- 0o.o
- . OD- -00'- C O- ,- 4.iL I
0 3P4r'-r-- . 000 *'In n- 00
- Pý en* *
60OOI 10r4P- - c fn0
H.CC' r-OH
Hv-1 N In M. ror-
m I
-a - -
a)1 C14 -4 . OH r-- r% co. 0% In -4 4T H C'4 %T M0O

En)J 01 *) -. .
*T '.% . A..
%0 . O
0% . . 04 - %
d W) rH -t .4 * M I-. 0'NH I *-c N
*'NJ
H 04 '.N--i4 H-t M -* 4 H- C%Lu r (OH-4co
0 0 N
0. CL'K'K'
4N 4 %D ON LIN Oa '0 O'.0%Q- In 'r-. 0 KTInu,
mCO 0 'n . . T *. .en . .M . .en-00
0 wIn 1 COO4 f%.JIf)L C7%.'. ON . O 0 - 0'. C
en 0
0. A4 %0-It N'N *U40 1 nT
0 *COC e *Tf
qCh
C14 .'0 '.04 C14 n % %-4

0 m.-l mN INL0T co0
C~-t
tfn . tf . . 0 . . 'M' . . -
Nh.
%DI
Nr- 'TD.CN N H- 0 OD-T04 - r-4 SO - aN
'0*.0 0
%0 M'' -4 104 CJC04
*nýt %S
'0-4 COenH-1
w r-
-N %D ~ 40 ý ~ M-T n* ý 1 o I -Q0
d
z 4 Is .- %0 ..- (n
' . .r 0* 'T *..
H ~ ' r-en'N0 f-400-It '0000

C4 Nuv%0 Nc P-4
001 H- i .. .0 '0 . 10c0..

H n ii 41 .6%'- i
'.'0' *'00 0 Cn
.- 4NN 'nOD0 co C4oN
o4
0 01' '.t0 000! '0rP4-n
o4N r I -40 r-.H Nr%0
H x -W00 %0 0vI N-4 n o%0 - 0 0% oCOON
r- - M '0'
%V- 0
-4O% a.0 00 , % &OO.0 N- in I
z0 P-4 %.e. 0ue- Me *tj 6
! 4 - %. r -10 enr.Ll -z -n
0! 0.1 0E 'i-~
0
duwL 44 0 MN8 %0 l 08 000 0 v4
00!
r- -104 C-- 0o)e C40r- -q% r-m
("4 in -o - --
.29po~, ~ 9~ n T
odl
06TOoE
SE O*
"T W
44 09 9 09 09 OO267O
S-.-----
o-4
0____0______________0___OS"
_______
U7.
j --
0o m'.0 %
'.'.C' M'. O
m~~ cn ON
C-404 .O
Hf HH 00 M coo0r- *.oD c
000 -1t
OH -4 H0 4..Lf) *Ch
C40 4 *N 3% 0
04 - -
0%0 o cr-.
OD I I
CA 41 0 0 N r'I0 N CY or) C4Me
NCC -'.0
cc
;20 m 0 . * M.D. 0 . '. *.
- -H
c'n- Lnt CY 0 ei L)r-

u V)
Lt'0 Lf
-It-T-t-Tz
LnCf)-1
Lf)
10 U '. W *D *.'0. CV)'Oco
0~ w* ). Pý* .-. * '.
04 %D4 C14 OuD % NLf4

.OO( ON '-*
%D *N
I
14 '.0ON
* *;--ý ' T 0r
- i T ONt
Li: V)
0* N 5N.
0o enrýeqco**0 0' o.Z. co.L
0
d
o- C4
N;
o
-
N00- m
M .~r .
. O0%
I -
I .-
'.UrN.
Hii,
- mc
.
'r- cor,
nr
rO%0
cn-l
H -. t LIN-T N..c%Co *Li,- *t
z 04
41 N 0 cv) 't W
to) ON.. 0* 0* N-'r0ocC-
$4 %01 -Ic0
JM
'0
w 0 L
00U 4j
00 N 0*NC
HL %'.
-t MU 0'm
&IiU
ChU
Hr
I-lU w1 %D'IMN al nI4. I
co 4) 00 m - 4 Ln 00
- co N M-It.
H- N' 0* 0 00
41e 1-%QC - co - -n -
4SI *ftQ* T
toZO qTOM fLnq,t
I M %0
LnQ uzC4
I-I(" C) 0M
nOl : 0'!
t n 0 O Dz 0'z
44C 1 00 . 08 000 00 O0
'4%
C4 0
V44 - c fl
.n 00rý M2h68
'0 0 0%a'c C14
c.r-e, a o"D%
'-~~~~~ CoH'00 H H4 Ln% 0 r OL %
r4 -00r Lý4 .'0 o
(30 P-4L ~ H.4
H- riO -r-I
r or.ý 00 0)0 '0
cd m __ __o to
oL P. ~.C
'0A K l0 L
0 0'0 IJ -4uL' (0. 0
w X Wo CO4
. ~ .o 0 '04 H
0 P.
")HC cn e'j
M 00
U).f
H4 m N H0 \00 00 f 0
H ~ CO. C *0~~O
ý',
0 %n
cnF-I N00m 1 N00-0tl
U4 HH4 tf~0 r- NrH 0
E-4 nH C. c 0oý
0)T. 04
0.. ON n m ON It
0 00 inuH
H- C.) r L~n (n C1
e' L-4 0
ONfr
14 ~ CC* ~ c .C4. *h
)
00 H4
'n fcv
WN U f O%
E-
10SI fTlT 0L IUUO aNC
ZS4 m 4 *an4
Z r H
4>
0a *;znE- n r.N 0
m 0) "1 pSAS
% . C4
WNa)ooP-4 N 4 w-q
269o t
44*
sO0 N -11 r40 Hcn C1
r_4
to N t- O~0Itrm
C
q) o) * f . C
4)0H . U) c -I
C c
c- -
>
9: 0
H cH
m~4N H .'
*f
.C-4 *C-4r4 lJ
co ON Nr 0c 7
-It TV1m
,ýON ?01 Lm
c 14 0 * - ,T -,* T
N_.T 00 _ 0m0 nV4L

w ~ - k1.4r* c
H (D2 0)r- 00
PLf-4
u cn .C*
O 00
r4 -) Y) -4 0
Hn rq H). t-4 F1D '.HC
nen
m )0 (0I. cn 440 C I
C
to P4 -
H4 C)0 N - )c 0 Ho4
r
%DO %o
IA H'4
en LA
O P4) Fl.c Tc
4 44 8 * 04
U)
KW P4
W
4 E0)
0I%0
9Z-H 9-,Ll 9-I 0 O
ON l~ -po Z.1-
- - .270
a'~ cc 41a'N
%D
00
W. 00C a)r -4 00
pP4 .u 'ocH U
:3
to cN M * (a ca*
('8r- .cc a) Vc
rA Ai OH
crn~ Ci
E-41
02 ic ~ H ~ cjt 4J $4 H 0 en
HH 0r 0 ~ LmU ON' 00)
(A cr nw 0.) n p
L)
w NHr- coccn HOq H
'.0 '.oC or- T 00rý
H cot -" GoH
in '.. O Om' C~
.4.. c', H C4'. 00
a) N )m -
Ci. 4 co
o O-- H' rN00 0)

Woc H.c .61 Go ,.4H
oo *4 I) *1 '4rC4 ' m
Go G~.o %0r M H H t C
NO HO'. H 9-4
to
U21. 44 QU
a4 'N -N
0 .0 .-0 ' cto toa
OSA
48I N~ N.OO ON
%0. ci
-cL - -
G N ý
if- - ;L
-V O -V Y O Tflq;sf
00! 000!
OOC! O IO
- -- .4q4 0
UTZ Su~q.;TioL
*ON OPOO* VO S 9EV L
271
0 ,. ~ Ln 00 V)Nc'MHC f %0C 0ri
-.T 0%
n CJ 1
0 C-4
f-.c04 m~ 0 00C11 r .Ln 01
N~ . .
*ý CS - .* c . . I
p 04 r-r-qr
C14 047 r4c r r
1co000 0) cr-
00 HcI .o ~0Oý
oID .
4) ow
cou o'1'm
C'4 '-T
00 '. %0* .%o.0,0F-
cn 0 )c4N
M . I4
. o)r-L
U241 04 rý ca%N-
C1 0HF-.-C'4 r as0~
rIT Ln cc CJH % C 4 r-4rC4-C' c -4~
0 oow- ~ r
HA cr-% COH '0C r-.c

-jH-0
fn' 0 C* .CN4 % .- o l 1 'T 004 m
0y Mz IMC . 1
014 ~ II
-. '
KnN
Cv4 C4C4 c
ONN
P14 00~ N.* mH(Ns. ~ c
CV)0 (D .7 H1 . Occ .- .Or- O'O

u. I n *j en. cn H 4 cn
co Im m-~r %D4I
H~'~- CDC) e4M
*0 0q
0) cal-1 c %.i n 04*(TC H C r
bdIz. H4 - CO o cn . . . Cc.. Ho.

a%
t HO O'TO ON' %0 r
2; L nr C nN (4O%
to 0
'0 44V r. O0 0f
0 1 r - r) r- H r-. 0%-0* 00 %0 u ui
9- o o M n1n -T- .o~cn0 D ý0c
0 T enOT0 00
t V- 41 0O ~OOZ as~ 00I u 0C
04biz 0 1 eq004 n *
-n -4 D "4aP- -N - -
(A (5-1 0e %nmm l- N
cof r,:)r- P4 L
.1Ln LA0 n -4 %D14027200c
j c . 0 O-
100
-T -- HI.00L
-M -n 4N%
% -)
H r 1 00 en -~'T~
0 D n r- IA 6 0 N
e4 C14
* * 0 *T * -. *
w3 H Ln 0 n 0 *CMO o~ :r
.- I%&r tco
l. C14 *C'j H 'H *N *C"IC%4 *C
> r-r oc .c~o~

qc-\
-t Ln C1 CN4
*ý - en. c) *- a% * oP.O
Jmc Hý 14 H r---1 HC00 % 0r- 00Olt

03) ) r- * 0 W .. %0 . M~ . *
E- Hr *T CO * * -* en'~ *
m ITLr CJ'j .*N cj 'n 'A 0, 4 f C4 Np
'0n C"JU -4 ~ ~ ~ O
w~r- N %m O'O
C)0V1 000. ICU
%D%. ( 4
Wo 0 j en NOm 001 -4(r 00.I
) u
0Ho .-
Hen
H~ I. *'n.ý m4. C--4~ - * C

C4 4 ------
fWL'
t 4 M 0 0MI?'Z --TN % C Ln 4%DM00 '10M
O HW r-49 .* CO . I . col . - r
- - Y---- 4Iqc
4j O01 NOL
001 $ O0 00 i V0011ý
0-v- - --- o0
-n oc -4%
4O fn0POD.O %
z 4)
ca-:~~ Az( :c%

a% %Dr 4 K0
(p I (0 N.e C J4 04 H 0%C4
0CJ~ ON
VCN aN0 -T 00L 0 4 *e'4
O r -C 14 r -
O , N
04 1C
011 cn4 ~ 'O A .~ (n JHcn 4 r-cn
o ~
41 n0 -4 - ý
co a -. . 'K 0C
HeJI C% OL tf0
*3 m.Lf,
LM *.cn Lf*C0 *
'T 4 f t . *?.Ic
P4f,%DO l 14 0%%I
mO n HU%
f- f o0
Wre %D r-000 tO % 1 -4~

C-4 1- C' O
0 C%0- 't . 0
$~4
.f%%
-M - VSnq'MAS S .4
0 0*4
:r 0T- r 00 c %0%0 00 TLm
0- r-4 u
09T 9 n009 04 09 "q9

r49 cn r
'T%-, 'o
-D 40 " -o
-- 0-omr
274 1 .1
_____ _____ __c_
V ___in
I- to
%0N
H
0-C
0- *1 10 .4
r4C4C1
c i 0
*N,1
%
NHC.
me
. CI
-I
NnL
*~
*4C4
,n -4
-N-
L
r-4 C) r On M4 e -cn04 (
irN NOO04
0A' 0,) E-4~~t0 O ~ *-'
on0 N0O H~C' .C:) n ý 0000

rý 0N00
4J H 11 0 C '4 *ý * .4
4 0 *- en ..
U)iI M~C rC'J..N *r- . .C
0*T 0.4 80 CO,4(nO' 94f C4
oG .1c0 00. o4O 0e) 0

UH H~ OUý0 Hn 0%P4
T14
.4 en.00t 0 a%
%D-4 LM H'MO
14 V _MG _-_ C-4 -%
nr% % N-n0
U)) -1 -; %10- K
in N0 C14 ený0 0 hLn 0 0 C4

'-' 0 DC V4I o 0 r-4(N co0 *
0 1. U
t;~~~~0 r4LMI '-4 (
1-- -- -9 --
H0)$4 L)!, 00 10 0 o0 NV1 Go - OT

C muj O9'T 009 009
Oo9-40T L
0ýo eTr
Oo
*- - 09r-4t LM0 it Le%: I O

0.LH tC-4 vc-404
(ON
U)00LnP 4)
A9ý 'I
AiO.. C1 0 -L4 e A
n - Tc
-F 4-
%D0 0 r-H. 0 m H C, -.4 0mP
00H -T1
'-D Hn H~ * N I-1
7
oil 1HLnr C C4H
r% t N C"oCNH
0 0 ,-4r-r.- 1n --co -T
00 .4COr- 0'0%a Ln ; %
CieILn o0Cf) r-4 O3)C

C3 0 * C4
-W w fl%0% CnH0
%D r 4
0 HCr 4 .. -4 0ý CU r-N _ý0
H r ,nr-4 %DifcC tCN H

ONO
%C -q ItI0ý -T
%D Iz IOr' C'JNI'. I % cOIH
ONr' 0 n M-~ cnC, 4

C14 - f)c
Kq
H ~ 000 0 H~C~~0
LM
0
kiU' 1 0 'K
-t. m en 00 eq.I
.4
rHz' HE-4jh '. 1N00 1 4
I .4 C-4 co %De 0%. HT oný
V- -4 E-4 LO Drý c 0 *- 3~
%.0 1...O P r.. '0'O
Hý14- C** *. *. 0
m 00( 0, CO4 40% %

14 'D 4A .4 .n W
A ýTl 04
0 ')A aUo'ý' '0N4 LL4O
- H)
-4 -9 40 ti
'94s14 IPq~I Tn~M -lqT1*MN
00f/T 08 4::4
jd 4.
a&08 00L
I
0
IA
SHW0O6 S34 06
P91TTh POTIT
276
0. W'C. -- n f -L)0U)4~
0 1 * 4) It . i)
ba-40 1 -4 N M r1-4 CN N r4
4)0 (74% 00) Lf'U)%0 uo O.1*

CI4. C-4 H0 *1 N (n4 r4 % P4
C).
r.0 w - 0)% 0 ~ )
0o E-4 0 0.--
.1o10 -K -H c
41H*nc "4JA* In%Z
CýIT -4 P)-tr-
to x 0m0 _ . _ý 00 .. 0
0 n.rj 0 en %D .0 00 f . (7 f%.
w n 1 5 4-
m%0 c '? r-4C
ir. H 4
Hof m- C4.'IV-4* CN e C1
0 -'T ad 1L4 r4 U)4

k n Q?-TO
0 -1 N N 400 -% -
t;0 in
-- - 4 MN
to V )
H ~ 0 1' 0~~0 N 1 0 rý N Nr4
*.4 *tf C'Jt- Ch4C
U)
ca a)U 1"0 -4 M.
-- 0
> %n -- -4 It
-~4 to ~ ") t'% ir ~
OsU) 4) N~ 04*)
M). INm'ne
LA ) 4 ý rý C ;.0 0 ý .00
.. 0 toS 41(% *'0

0 c7%
8 i 00 % %C&A
C01% LI 01
ON L1%.
0%0 0%
LA. .N
.1sial IMUON 3' '0T14

-I qT
'I- -
Sde09 WSEI ("ST. 0091

b; 486 86 96
0-4 #-4
U)l -~ -9-
am
cc-uis
277
0' M 0 140oý
1- 0CO4 P, C1 f 0 H *cn
4 3H .CY *0 *D C4(* 4
60O HH
rI 1 H m 0C'H0M
r4 N'0H()%Z 1 0C
0~C,
CON LI CO,;~
004 C4.- C.,r-* N r-4 t N .. r-
w' H *N N I: *' C
0 E-4 i
MNC H 8r.-' O-.' H
0
V)KON0
l CN ItC1
0
H
HL)r
Lf N'0
3
,o
-
P
*1
14-
-
M 00 . * H w
W~ 0 nsJHnH .Tr- CN N N~~
P_ .1 _ .
0" fn : O
IC .H .4qH'co
44C C fN '
HN Ln a)
0 >l) 't %DNt
H - 0 1N-. '. N, 00 m*i%

'0 C
to ;r 0b 0 N..
o 4 -I 'T -
Hl PH H co > 00 0n *'.N *nc
0 -4
H ~ ~H- '0~ HC ~ OO 00 ~ C N
o )W N *N ('n "Z Cn H. .
- *N ~44 C40. ~ *-70~

LC.I0' 0 0 ItC' 0N C4N N enN . )LIn
E-4 r-i
0o -r C.*- 0 (1 H
r4 C InULf I T 11
Otf1
00
H4.4
z.C Ln* .- . Nmm N C
HO *Nq NC.~~ N-7m en
r- -. L
'00 HN C Nu-1
co0rl%
00 1N0mO "
V-4O (
HO 0. 4 ** H *.
C4 * Ow
ON A A *O -1 T. *Lfl P
1)
*LIsU VO T1W=0N T~qTa T~qq u
asip:o BUT3900 P83Vd PeOe1U

ON amo.aU -.'uWLn~ L *
-P4~~ -H 4 z
L- -!009
.2A 00T 001 0
-0ll - WI P4
UM' o
-o -4 U;,O
K-HO
'ON
opHN-I TO Ut El
27871
00.
01 Kn KT M
Cir-c~ C- r- 0% *t H.. %
Ht HH N*H nH C4 Hn
nc~n
0c 0 l DM L T 0NO P-I
r
H 04 U) 00I *HNC
H O
%aOO
HHc DLn L D G 000 C
00 0 r~-. Ho
4.J WH coD 00r0 I H- C400 00 MN "
mq C14 Co*j C0 NN
I~ -n -n C
H 0
0 0 D 0
% H0'0 Nt %Oq - A
xi C:) . .o. 0 0.
0 )0 T%. nOD - -4H -410 %0 cn "

H~c~
00LI0t C,4 0 -
NCv * H I-** A ý
H n Ci -N en
I0e 0I4 0060 AA-
CH Ln 0. 0.) NC, HCN CSH *O

paH>HN un r *oI%:r *A 0~ 0% H'mmC~
c%
Hwr C NI - - -
Ca "0
)4.
H
A
HI4
>% 0
I'-' M4
'*-4
IA(I
04'. MIC Nccn M r
0 )0 14 H
0 CON Ln MO Ln CO0
ST 0 % 0
%C. N. 0
11.
Pw N -4 N4- * 41 IN HD
PNHl
V) (n NH4
fnO .'* HIIf cL
C-~ it. -c -c V
Z
enrs LA' .*H14 0 0

ý41
44 OOL
%00% 11 *n O4 1% H
N 0
n 10
a a InZ am -A am a
j duel 0091 0091 009T 09 OOSL OU~T

co P% *e OE ZET ZET ZET ZET ZE
co4 -%Iaa
00r
OflI 9V,9V, 9EI CcV, 6L'
279
co'
t0 V')O rý (fON
:3C H %0 4) CO CVH
H 60 F-4
od Cco (
0 ; H *-H cU CN r-
H CY HL
0 , 0 a.
-HC~C
~ JCN C
14
41 $4 4QIC
U'- H1 L0 ,-H
4-r0
0 9 0 % c CNH C-4C cc
S i
OfC'4 c0oco
14 1 . . 0) . 14
-~O 0' % H0c

1-4 r. 44
-z C"C7 U
Hn CO %O.Cý M0
44 * m 0 "
Uq H H i .- NZHn
~j H 0 ( H-1-f H1 *H(n
-t4 0 %0 -4t cn
et~n . '0 ) CqC~ .
OnH2d
o0 ~ G) ~ . .- Iq CDU 0
IZ1 H T-t - H
0~' 49w
Ln I A
1-4 I P4qa
tha
.
mqHq
~~rducOn10
LtnI m Sen
Ln280
41
cl) C*-4000 C7 'JH
1. C14 C00ON q 0a CO~N
00~
r- 0 1 % . I. )
(t I o IC 1 0 0 1
'Lf¶ 01
0 CnIc7
C1 M I Nr- *0 ( *O7%(r-
:3H r 4 A
w -4 -4_T C14Hr-1 n0 - C,4H r4C,
P-1c~J
04
V 4 * V-4C
00 Lný
*10' - %
0 *o00 DO
4)a
co -4Lf - r-C)' -ý
IfeJ *'-0 rrc)' 00*'
g4.MC4 rC 0 .C4 C1 r-.e .CAi
0 w~
ITco .LfO-
-jOM I
*r- n* Nc .i 0% NLAm
a0 I r-rl.
CI %D I I O(MI-' 1-i 1 1 0.. I
0 f)0 *ar c 4, r-1u *C14 Olk *oo *CCJ
00 %0f ~ 0? C14 f0 i I'D ~UIn

r- -It 4 C-4co
(0 I? -1 nW
HV'-T
ON'~
*t
- I
I-
I C
-f
t
I
I 1
.O
0~
I *1
1 ~0 'T
14
IA' 1
*O *
0 C **1
*'C cn
tn cq CN N cq N unuN
Ln N Ncq1
-o cn -Drq L- 0 ) - n -ýc
'4 V-r4 1 C4 J
ON~0 N'-1O
r4I~ 4 0%
~ I * m. CN N C'JN
N* n N NN
H 4-1 CN, 4 r) i *
rWN - - -o0
H0 W
CA~ 0 MC
0Wl
nSt~ a) PT)~ 1w
0 0 C 0
m 08
08 1 00 ý004 a
1Oa, 08
:3 )03 L) L- IN 1 -4 N~dr SW) O.1
--- D
--
091 -u
090 9 9 0
-D : -4 T Irc -e 1& - o I -D
6GJ0
*0)
E-4I '
LA M8' 'IV M£81N'-; QL L9-4
14 W281
-_ .--.- ----------- - 4.-..-.-----.--

______PA_
'00 aOLn' 4'.0NC 00 .0 00 LI) M -.T M 4
7-4 9) M0C
rH %0 r - m0 r 04 '011
-t~f *NG%4r-
.ULf) fHf)-4 v-4C.T *N C14C4

*C'l H-
00~ 00enC C%N m -4- (7%Hr 00-T0

AOL 0
H)')H %
0 rIfLO00 CIN%D Of) o 11,00
Ca4.1 H.. '0* . H * V . 00 %0*
H1 *Nt *~r- .% LtHI L -~--.?t U-
ý:
H-4D V0 HLn t - MC .CY C*N N N 0
r . -0
0 040
-H %0 HO1
14Co 0 CO% r-N H0 it, L.'.D
40 H4
0H*q%- 'I 0. 0. ID* 0N0 r- %D *n a
cO-tr HC'0 C4'O 4H C4
0 n00000%v 9 ý a C;-Lt LM'. C0;
irn ''
-1 VPr
H'.0i H-: .4 *C4 N NN0
0%
t-, 00 00 H-1
00 C- . 01ON 0%U r-4t C) %D C
. 4J Cl; *4 0. * % *
H- CoC* r- *L%0
Co 0 m o 1.0% 4 4 m :r" 00
HO
- gC4 C40L *-.C4t *t NN
14 ~ -I nr
Ln r.4OCO.-4 .H %0 H''.
IAA 0%T. N
tn N 04
Ho enI %O0%QT -4 000%o ON HO-* 0n%D

n00
0) 4*0.. % N *- %.0 * I .
44J *Hn.r.C .%. %DC .tt .~t rz
$4 rH '0.0 F-4 .0 Ln *.T _ _ _ N
'.0)
%_0O '4'0I '.% H -IV H %
0%000Co
4'.--T 00%O e - .. 00.
ON
% . '
r-42
a)
H-
**NH. %
* 'COP
,.
0N
Hq.
N0ensNO
M ~ U
4)~
m- (a - DL' nC NC
4f- --n D0 a%. -1 q -q H -%

%0
0 r-IG *N - n n -4 -.T-
Cý10%
H u -1-4i N rI 0.1 0.1 De

9 O~E 0~
NC
044~ 0~
a-
*bs1zj 09 09 09 09 e4
09 09
r r-4 -% M:-4
[-I'
.0 CD
. Ca2a82
A-1.5.6 Titanium Alloys
283
N -1
00 H I
IN >
10 Lnr
1,-. ,-Ln r- N.- r-0 D n0
00 V-4 CN c -1ONr r"
NLt OC) unLr- ON c' 0
1 H %*o 1 (7% 1 1 m 1.
H4icn~ Lri-? *M
*r 0
0
04 ~ IC*CI
LfH00 H0 00 O
4J 0*40L r 000%. cN %0 * ON
9-4~*O -- r H rIt Cor-4.)L) - L)r
-
0 -; ___ __
r-4 4
HNN MC' N.c''o -.
-t0 * T0%. 0C H0 00* %
M0 ON~c*n c c
r4Ln ,t N o,' .Lt 4fLn cn
N-H -.- 0%ucl r
-- L r r-%1, nr- - one
m H; %q 4 *n .. -It
tnON%-0 Hrei c r-,0 0 0. c'tr-.

H 0r4 -CA N r-c 0 0r- acn 0
O
En~ 0%.. *. H l H* CO*
H- fn %D r-44
Hr.,~ HH)
r- *C'4 f ' o
H In 4) -4J________
E-1 : F
tr
41i%0.f
-O
- 0
H4N
.tC r- 0%
- N4
04o : ONn 0oc
01 01%
co
'rV' f .0 - Hn C.*
ON -4 *D ( 0 0*0
m 0 4
P w .0 .-
qt N-0 NcO 44
'.n NiT
.%
HO4
a T IN LMf%0 .tH
1-40N Oc00c0
*O4H 0 ~ 4
a, .o0 0 aca>1
ZqOl idd 81 uidd 06C
08 08 08 08 08 09
ISH dusj
-~ ba~ OOLE £9 40914

I-t -M - --
'.4TVUV u Ulva SmO

40-S 49 *111
2101 PSTV*UI1PWnov
U- - - - - ----
00
a)r Nr %D nV
H - 00__ %. - MC4 0__ 4 -r
0 0
-,4 4-00Lf)I (14 %0 - i I r-qr %DH Lr) en

rn NL1 . I I .q
I*,I U
0 ~~ ~ ~ J~
W - *r--)
10 C *HN
CH
*N C ONt
ONC'C~ HC
Kn
N0 It IK q-t0 l N
~
9 ~ .4(ý- ~ N~ HCNCir
H n Lf H 0 iL~
W 0 0 n e'j
00 r* 0
P- D H..%
DCt
HC'JN NC y ' 1 n-o04HCM C4 e'J
0 H L
H W
o00- N-ItIDON% 0 0 0%
H
1- HA z ) Ht N a% r- N
0 q 0 --T1O ) H'T
- ýc 0%D
ON D0%0
H, 0.f r-4 '00
%D~O a% NLn - 0 0-
JE-4 rn 0
A4 >4
w HenP-4 00 n H ..;
Min
%"
044 ONc~i-
14%P:O0 *Ln 00
44 Ntl N-tI-t 00%-4 ne
ýo0o - w c -0 ý-D
H 0 Nrý
H1N H-4 . I N I* j r. 1% 0 1 H 1
8C. *Ht
N.D0%- L o m oi
0 0
XLSIU ~nqTOM T~nqTefi iTv=OM~ TV=iNz InqTOMb
wdd 8! wudd 06C
H dwal 09 09 08 09 09
P4 %D
UTJSW 9-tsflOdl
P;DTVBUUV mnnosA
'ON Op03, 00o£0
285
00
A-4 >2
0 r-crnJ ONJ110, LnC

Hr-~v or)
(z)r-
H Ln
H~0 __ .n% 0%
__0
41Lf *Hl _ _ _ _%D_
__4
H
4jI C4 a r4 00%
-t~~~
*c** -* F 0. 0sJ
0 %0
-400 0ncl
(n
.1% r-0
) *' 0 r4
1 HH 04Hi%O4
-ItLh Hn HN
-n - r4~T - - -4e
V- -f -n %D-a
H
N 09
0uej
C9 ý 09 LA 08 0
L~~
cc 41 (nC
%. ON00 L 0P~ 'T C.4 OD~ 90 60 ý cy
Hw -r 00a% -C286
~~~~~~~~ _______________________________ 00 ~~ N____

_________nen
1-4_________________________________
HI
4) IC . r . cq
to' - -. N-nc) -
0I 4c
0, ~ -co r-. Ol en~ 'nH'
H 41 0 00' W) 4 H Cfl o
*ýý H
___It
04 H
N cn r- QT U0% LA N L
*n 14 N C O ** N
0%L cn -T
Hn m NC~
(nC'NC
V.) - - ---
H~si (~nqP-4
rAa' 0 e~nq0 n fl-q.
H~~~~'C1 7 14
o-Z4
f09 r- 0- 09 008o \ 09c
-r f4(I -n % -V
Hg SD -U13ldT'T"
P4 41
287o1-
0 c -4 inrq0
00c4 t -T-4
4)
P4 >
,-4tr .tU tn enN 00uLM' N0

r-T a0 . r- 0 en C % V
to 'C C14Z
-H Z~~ *I co
't 00N CV) CY) C.j 00O L(n 4 1HoCYN CC
00%DC 0 m~' L
.4UC -DmV 00ACn
004
CO. c
en . * . 0). \0 .U.
00 cLc C4
C1iH CON -:t
0oýr- %OO
r-4 .IO r-1 H
.OLAHHU CO4
cfo 0 -4C
*0 m
Pf ON %0 r- C -T -
"o'l I. 0X N* q=
*C.10
0 1
N
CO'. : (N IT
In *'t '
eq r4 -N
N 00 T
%.0 4 *No mN 00c
C'4QLI - n
Z -Iut -t'4 LMenONHIOu0

'n-A . *O C4 N P% %'D
ýi~ t;
-. ?)- *NT 'D O'LM Q r P-

*0%0 C ~LM
*0%i %O. -*NO
P.4w
00
.4NC'IC' r-4 *CM
en-.-
00 - . -
-
-0 . Du
43N
en co (n-c 000V-4 4 rn C
O%0%I %r-, O0 O NO mI
a.' 00000 N
co~ L'
00~ C 0
a i -A -o
tw~ coo
C"QT? %n
P4'l' M %0
(), * . . 0 0 L o'r0
00! 001 00! 001 001c n\
r4 a0 - -; L -: F -4 -D
r44
rn -. . .f-4 PQ
oG
al 0
00 H
ioc 4c ~ 4c4:4
H.t CV). Pt' . NH ý- % r%C *4t 14
o " 1-4
cc
-H
.1-
Weo'
41 V-
00 Ln *xou ri00
a q 00J~
*0Hcs0
'T
04%
0% ý
4
-Tc40
0
00a
r34 0 ~0 ' .- 4~ eno IfI N?~e L0-4

.I0 '. 00o
-C
%D'T
4)I cn ,
~ ' 4 %-T0l 0c
H*UY4
In0t lu Cr *
IT. m Go V) U UoH
0LI~~C
P.~U~~
Hn *sLn H1(--T -IT NT .
L) 00 H ~ r N D
Hnt Q -%%
P4 W)~ 4L3O ýP3It
due 0 09 09 09 09 0 09
14co99co %0_ 04 £t 00 00 100
4 Z U4 - 4C4 1
1-4 lzUU P-
4TWUU4LnLA0;
0
!rr*pon rlNr4 P-4
A ~~~ W-CI-.-.--.-0
60 i
>0
r- ýC)I - r
0) 10
14V -
1O'C%.DO m1 Vra O-i 0CN
cJG
0)(
H1 n cc'eu.Qr. ioe
-I.-
r. ; nM r -*CO) M NLr)
1 40 Lr) cl)0% oo Z ý Ur mCN\ ~CY)

00
r-4HN
H4"ý 0 - 4r o-It0 0 -:r
w mC ~ .. 14 . . 00 . 0.
0 w 0n ~ - ~-.t
00 - ONV
H r-
*~
~ ~i 0
l
C'J-t
ONHIt10
r
HD M~H
't . . CV) . . T.4 .

x
H . r- r4 *-4 00 -~ .O
0 4
I'Dt CNC4
C14 %t It cor--
z 0
po--
1-4JC ~,-C ~ H CUCrC ~
94 0 0) r- n I l to mNr-4 en~¼ ~C) 00

H 0 %~
-4 . (V
rq %T-.to.~
H a H 4' . .4' v- * 4 ' o .0 * 0
H
4 -T r-IT-4 , nCV1. -H --tr-- HO1-
0 r
LA ~ cn 'Tr. a*'44r-
%,oHO 0Cl C
r-4 O -H..0J 44 %D . - CD
0o )I -ý
zr r-~o
-t
00o r r4mc
0.N
Hn
Hr dmei 08 0848
w 4T - -4 -W0+
Cfl0 - -l%
0C
. -- - 44___ %0 * c -0 o 0 4 -C- -O I-
5. 00
60 H
0 COO %~0nomC4G
4) H- U) HC14 %0
r-4, D
U34- c. C . . %
CU fn H * H T0IT4
o- -
CL P4 C'JLf C404C1 c
r=X -.t m
DO 00 %D
* (D C -It~ %0
0-IC cýC
%O C1 en %
N n C4ONc C4 ooL c
z 0*,
9Q H 0 -. * ~
r -: ~ T Qr-c ,
rn 0 e'C)xO. HLV
e'J4fN IN en
PO 0-
.4Lr Pr-4o c 't15.*
04U c -~ %D0 0OHC.4
c
0 cný 4e n 0
-4 H 00 h Go r CAco-
HE43 4 4 mCI
0 r-. NN -4 *. n nI
94 *r *J
00
"P4
LSI IUION TTnl~eM tTTqTOA "0
o tp 01!
~ELO*T OsE
N 0n
o.- u;D 43110 0zos
%O *34 OV
"ON OPD TES E E
291
00
P41
8CH4
00~~0 H cJ 0'
A4 . .4 1-
C41 J H_ .
1-r CO
H 41 NLfn t 00HH
S P4 Hn r~riH 4
~ 1--4 N
~LA -
a r-c %:Tco
0
z 000
U Hu 0
c0r
00 00
en~,' O w 0- o
CV
tH m M~~C4 .N0 04AU

-
to .P-
(rD ,-cenncn,
4 14 A r4
C.))
00 en ~P0 Tm
0 0 Cý 0 0
enCO -J '0
-H ~~0-7
~44 Go %Tc %0
H0 14. '- c 04
e"n 4..V)
x' 0 ' 06 011

0~o 08 0 09O
9 b Ot-IH
CU3 1 Hn 3r1 4T'%
I
O*A3
*ON VS SESTI 91WZO
-29-
O ~0 0
C)4U
0 boj-
oo IO.o
coco to0
0 ~~
cotv
Ha * IV -t *N ONft
0 P..
4i0,1 -
44
o- HJ0
Ln C'4
c
~ -H0
0
Fi14 irnfA
--
-It 0'en
c -L
Ln -4 C'J
_o_40_40
0~ H
HC0'J00
14 -It
o . ~ 0C H
En In C14C
u 0' 0 *
0-
-NN rýDH0' C
H Ln %D
H~ Ho OH
110 C
H 4)U m~omH N IfD

E- C4 j-4. * ( 0 ..
. 0 041.~ r-4 ON ' r-43
1-4
4)H 0 0
z : ) C; .b1-4
0 0 n nInG
1- 8
%D0
00 0 -444 k
H ao %0 9C-
m-
V4~ 4)lO 4c
*~e IT 003-t v 0
O0 1p03-g f-
*N 293P4-t '
Z*a (
Go 0
bc
be -4 F(ac
(U CO -- -n Co-
rO' iC PAHH(N 0 -i r
-0 0 04
4J 0 Ln~
%Dt c T r4JJ Lf
14 n lcc
ri -*. - cn 0 wC 4 I - ýL
pa Ln
0 0-M L
ir eq00 r-r- cr
%;*ý * *
-
Vccn
Ln* q e -
m~c r-q r-i
Ln tr-Ttn
M0 c n c -t-'
'o tn00 CAm-%
-0 %
j-4 4
Nn c o cn cv, L) *o
-4 -.T mv ITcci F*
Ln .10-1 M - -4 E--l -Dc l 1
. .4
(n -W .
E-44
E-40 4 Lnc- ID Ncn0

0 -n%
N o
I' o *
U1nI . .- .P o0 1
V-164-W - -, -aui
ni
0 w 1 W, 0'T
co tA1(
4 C002 0O A09;1 0 0
In e U~,j punca!Dl puno~t!

_ jans4a TVUTpfl3T3uo' N.41upnTu"
P-4 aN M TSO
'ON 3P034K i C l
294
1-t 0 TrIr - e0O
0
LU0 - --
0 ~ ~ ýO mH00 crN. 0H
m . r-. 00' .
H *r-e
q q2 n*r-I r: u,.
-T o r- o1 D N -r l
.11
*n 0 .g - .ýLr- -I
HHHr- "IH .cý r-SH
04 '0 * n It %D-~o
H
9Qo A > r ;HO -ir!
4) C) C)vc c g O o I)C
w (N CA -4HO e C
o 'T Mr~C
%DNO o I
004t" 0-*0 '0'C . , 0
co en 0- -4c
,...44-W .. OO- 14 . 14
.4r
00 V C~ 4N
4 H)%
Q) 0 "1
4ý -444 co* P0*e 0. 0% %T%*
00 .K 40K
-T en
4 "H U* ocn
T~nqT'O TTnq;omA o*Ai1 Tnq;.M
II - - -
*O SPj O 5 s 1S05 St
295
o c
4) 0 c
60 H,
to ca
-N4c
00 ro 00-4~ItiCLf) L(IHr
4)J LtI H. * Ltr m. V0*
0o H nL~.
*Cý1 -tOTC
r0 a.
140 L()'I %D CY' C'4 Ln rH
C14J
C'Co
00 aCIN0r- H-O w.
tal c)~ * TMr-I (C14
0 o0
0- C
o- Lf) r- ir) VI 0 '1 cylr.- E-
0C14 N
0H r.- rCo00'T--T -- It
E-4r .. . -'D.0 r-, . . - * 4..)
H - C1H *- r- - HOC C14' 0a
r- Ln -I M-tC1
%D m C CO* (D N t*
(D -I CN0 oC*t*t C'1 1 Ot C
H'0'
LA~C 0; CCFt*-tN
0H C) %0
OH L .j.w Pr-4 cZ-Z CY-It 0u T-- T C14 CoLm
H H 4- C1 ITCo.. a)-'Sm --T~rN 0*- 01

EH O
ad_)p__ _O .
0 04 r4W%
-1 DH0 t )N I - 0 r
o- co 4)J4.J
a)0E-) C0 Co4. '-H- ~ Co)H
H H4 CoHC' -t$H4-r- Q
C,
p M*om co w.1ar-4. MHL
00I co 0qN Lf*. M0 .tL-c *-lir
___ ___00
o o , no0
CAO 44r HLf'. CLrX
H m u *jC14 *H 41 *re -4 5-Lf)

,-44 _ 0)_
00*44 H OH
- 0
- - 00
r14H
'U* SI*t*UT OZOO

LA
H'
ca4 -W -m -
~IUJ~08 08 006 006 --
9- bipa cO -cO
-0 - . *n
UIA 4TUHp~b
HN C'4
pn,
296
0 00
Hrn
: >J~
VW
H- c' - T-4 0
C4-4...
%'0
4
.*co H
\a
C\
OC-4~
Ch It)
~ '.4 ONc
%0en
-4.r
* CC)c
mCJf
C Lt
*.
Nc0
..
4)2 c.I C1T H .cql
*-i
9-
on4. -It Q - .C'J
C4 "C hC
d -C414.C0Ir-
v-
HH-C
00H
C4I
0 J 9-404
Or* U C4
$4r4
-oc
CHd
HH
- 10
-t
r-A-~ ONJ
m-F-
H*r-co *. LCC-D
H
0* %
-U
c'lq*
*ý
- IOrC40
4cv H
t
.4) -OD-4
-c)
Ci4
1 -Lr 4C1 -4r, 6

.'* -' 4- (.- - q m
E-4~
0.CH
~ ~ r I)
__ _)__ . . %0__
C:)v- C.C'c
r-I 00O 001 . *.4.T'.)
-c ,
-4.. N- r- 0M c nc 111 *N C11 eq * im
CYN-It0' 00~ O'.

w0 %.0-4. M o
Cý'OO O%O
Hq
0 o-4*o . 0** N C ~ " H;. M 1-
%0'. 0,
00 In . 0 . . I' . C.J
u0'H
H .LSINU4 tcn
TcnT en
(I;-o cq~1-MN
I M~V
1- dCsj 9c08 09i r

tr 'T~
08%o 00~ Q - 00%Cm
$ 00 -. -z r- 3 - -N
u
1-4 1.4m (1 C 1 ~ lm
31O m I4nC)CiC 1
> a)
r297
0 1 D.' - ý
:> ,c
oc
00
a% m c- -: . , -0' . . ýP Y
41H ONLF) C4 . -. T --.tLf) *C'cr *'Lf) Ih

.e'JCN C%4CN -4 CV)(N H CYC'NC H *-
00
C HH C14 C)C4C 0)C4 )'J0 CDr-N
4 rH':) r- .-
C)'c q
4J $4r- * C% CV ON m P, "Cý l*1 n'.M - 1
0~zW *ee C9JCqJ'- C'NHq C4 1-4 *Hr-
ra C) H cn0 H C0 rI-~J im
HCN.r- N ON0 cf) -
E- (31* It* rH* n i tf
- J. I, \D- N '* n L
I%-, . r4*. 0 *NT O .0 * .O -
MCV4 N cq
'1 N -I-Cs
IT~ HC'JH *
nLn0
o U~Cq O
mO(1\m %~.cn MC~I4 c'JIJ c'o~
V-4 C. *'It' %:r .TONm Hn
t4L *n a i ) .
'z-# *cC1 .~' C00 C'f
CV Cf, V4 M (Y) mCI Lf)CYC' iV) C14 *CII
00 H~ o)
Un 0) 0U rq T- 0 H
ý4 '.c0J -i 'T P, 00J HC-i
HM~ ~0
0) ..
r4H
H.
f 1 CV)CtLf r-.i C)r- cn
tf
00.0
00 \
H ) 0\~ 0 u) *Or- N'r-. CO .- rini-, en
4
Ho N.N~
IT C) 0.ON 0 cn e C'I4r ONCI4- T r.C f)C'
o- - 0.
4J a'c'
l I I4 OP-4 %nc *H
ena *c?) m C14
H1 (n~- -. - -
T,,U4 0 .c
It 4
u0 ea 10
'ON 'LI 0 N V) HU lt
It ~ t ) ; C14~
Zo'-U 0)'.. 0 * u . * r. *f a'. * .
Ib44
i-i CM%D
* n CH4 en r. %.o 1) \* P '.. c o F
00 r'CJ%.* . t,- . ()o*Nh 0~I V-% .
nq~vj T~nq;TBj TTnqTOA TYMIOtN

j~~a~ T~flq19o AogVI
'T OW0 UJ E90' UT1 SZT'UT 0O0. 'I*T £90'U UT
dul009 009 0J09 008 008 008

-M--
0£ 0£ 0£a 0£ 0£ O
0£oOE
04 HV4
I1 'uTA
r4.
N 0
co co
4 00 Cf)0%"
ONH NCtO O0%
CH . 4 0HHe H40tn I)
"4J.M 0 N HH NO0 Al MH %0.~ 0 NHr

1C Hr 0%D0m NO0 co U) -T enc r~)CO(
CH
v" t. 8 1* %O 4 Li~ *ý 90** 0%ý
:
0 pq Cf)04 C* P-1V- - *foi
rIý m - .c
0 '
t-4H HN- Ln LI Lm 00(n HON0 l %r

%10. H*
'T*
Kn
x H *
-- LI)-T zN
OO'O NNi-q4N
C4 -4
0C)
W 0~ N% O 1r-4#
C)0C
*NI- .0 *Uý 0*N .LN
41 H 0 ( W Le n 7 0O Lm- N NN0 C4 ()o0%

0 '4 14 ** Ny *o 'O* O.
U~H
l)H H44 >C n *Cfnr4 c C14 HC NN 0
4 %0%~ * '
0)
H C NH O L-0 '
~
.~ 1N
4J n a e -O cON t- m40
0%OC Nc
H 41 ao qe -- -:
. %%'o . ,
0
tn
00
C.)
%0
C4%
Lf
H Hu-
NH
H ~
em,: 0 C 0 7- 0%M -0
"4 c U41
N'
U~HHO.4
-) -Tc aq -P4G
--- 4
ms1 006 006. % 00 ON~

.0
-) aa
HZC UT OUT MT OU OU'.Zpa p UTo90
OR I- - --,1
CL
H U-TR Ct-3M Cu-Jza
299
-G-
00
H
-, - NI - ,IO
0n 0n 'or N~- r-
0. -a_
0nLnL 'IL n HCfor40 -4 Co H-
p4) .0% aN r-
fr-.
0. . ) .N.L o -4-
04 rU-4 *nC14 l *11
CoCq nC HNINL ýor

0n ON *H .r-- . r--O.
ta H-*( 1*c
to N 00'.1t0 LA N
r-',D ~ r
Ln Jr. 04- --
8 0 CHN
NN Nmj - n O ,
4j 4)to 00 0-0 %0 O n M 0LnO'
u % IN M -n 1-r cOn
---
du 009 009 N 009N -009e
0 "- . 4Qf-0"
4) -4DLH %D V4% -4t CLM-J.T
SpA~ jO 0t ?Ln- hm
1-4 4 ý 0300
___________ ".4
COS 00Ir
0,
ca 0
O~r-4
-4 - Hn
fn- N-n
-3%tO dCU r4
0H r-
('4.
%00 .4.0r
U14J _4
.3%n00e
en r4*
%D0 C4D P.
M.Lf i~
.4C~ r4C'9- .C (U~4-1 oo '
0 - C'
c~ Qý**~ C-4~o ~ - %4 '04
N
H~c r-c'n '0% 0C- '3
.44 H'N 0-4 CNr-I * n

0.
*0' *Cý. 34 *N4 0

-- cl
r-CnC'T M' en'N .M
3' .~0' %4 (
c'~~~~oc N'i
r% %T (4 00
on 0%*
%0' %0. Lf caa
** in 00. n In C4 e CO.
0K r4
"o.3 CL tlf *'0 . r. 4

Z - , ..
. MN .4%.94 NO ýT N"0'
0%hIO
"LM M CnO4 Pf-4C7% %0 %0C1 %oO

14
rq
.0C O*4* I . O ~
0- 00 ClC
in x -4
H~~0 *NC' 1O1
r . O%4
AC P.
-ilj
in4 . TI .T . .T . . vupLM'I;3 0'

£nN
I.%
4)fA4 -4f -? &e
.oT -
8ZN &o0r- 0'M
cn~(0 N 008 MP Nc
0084I
1. %..o -w ..-
- P4- i -
.. ixlnihu~hhIiMEi -4 Irmmmm mIm0
23S ~C'? CC CR%
F301
Go Lnrft n LMm___
bo H
'-4c
r.0H
0I 94 Nn M 0
M" 4 .- H C)
(0 H) *e'J' 110 cr- .1to .r-ý00

ý-OC HH - *Y
N (1 C',H)r- C
O41 - c
-4 -44- D
z cOoIn
4 1 7 . C.cVn' C)OH
0H D 00co m %D
H4J H aýor- N000
coL., ,
H'.o # a r- O 4
-I -41C1
*f 0 4
Lf% C4. riC .4 -1 Ln
;L 0 p O '0 ' Hr- 0,%'

c0r HODIDr4
0O- x W4 o00 . o .c o NO . D0ýuý 4H0 Ho'

u W H r- a) . M.r - l - nC . %0 N.
%_ H 0. -* -. 1H C C41 N 'T *C -
'4m 4 y%
4 .4-O HNHocO'0.)orý' *( CrNH

x
L0
0
o) w tInC-4 M4C-4J *l co
Mo
* -i V, ca.0 k.
H OJi *C4)*C ~ H ~ ~ (
00W__ P____omm' 0 0r- 0 4 0) (Y__ _ _0

00L
-t 0n.t(%0
44 Cao'. 00
C14cONLf 0 m %0
0 0) ) C- r NCC4 00'
H H ot-4H ' 0%N ca.' C 414 L(4N
_0_ o - 0
E- 4 H O4 0_ 0_ 0_C
_ _00
- -a 0 - -
V___NP -
0'.0-44 4 '0 m ON M V).
MWAVC 4L~. 0%0. cao' >IFo '~o
E4c Hco. Oc o -o O
Hw 09 O H *I' 4 00N*
00 \c OEN
N
IH'cc
ca*%T
Hý
'0*
o'.co
pu
CC HO'.0'.
C . '0oa N
r4
H-I .40%H
E'O c IH C ".tH 0UIH HO-
'ONC- HUIC.* .0'. CTU'. HH HHs
-0-
9-44
G)- "a 00 f- ' n '0 I

P- 0 0
0 0Jo-4 .r-4 A i O)L %-Tm
1 %T*u o
N' *ý H * A'*
H4 N ~ .')C14 .*(14
*cnL .4.'N4-
H N
.'*c
'N
S0 A J -E--4
-A M.
Aj w mt - NT H0
H&c -t t-
-A H1 H
enr 0 - HN A
C,4O
%a%1 cr.4
w ?-L OD" C.. NO
-N
0'4 wO( (A.N
*TC4 t-4
M H 04v- .C14
oz r-C H.. 0 N
-4'O N U
N'4 OM4
'.10 NO%
U; *-t C *CA'4
W H '
LM C,4 * 7%ONCOD 04C7.*
00 C4H
aA COCO0 M LAO
H4 Hn '4C4L
-I NOlN 0'0
0- 0) 3 e N N* L 0' OD *en Go
ut 4 - % - - -1 -;
cn H COCO'4
*00 N'C4 m
'04
1-4 14 H' 0'* 0.. N CA* ( ..
0'D COO .00

c%-1 00 0% 4 -
I 4 -4 ene C*0 IAn' 0 HP-' lAC
'
.Ar'
mT'q A3' rp'T' AS'A
A~ 0. jj0"! j-0'!zC0
44u. 009 008- 4 M%0 006 04 km 00
E--- u l 4 - 0 n t4 -I-
'-444
CN coo~ *8%C~ 69 4 06
C
303 . 4 %%D t
'Ca - az j bm2
000
04 __ _ ___1__ _ _
p. -O O Ln r- ( n"
4
NOH 0~~
'o1o ONa'
ON o
0 0ON cn 'ON
F-. *4 Lf4 CO F . N
-a LN -.T
*qNC' mCe4 C') HI00 H C4 1 r-I NHr_ N N r
(a PaN U)~ '?NO r.4 ON~- 4N
0a 10oe0a - r-.-.-
Hc.- "T~0 Ur-
.
$3 r-HC,4 NL~ NO *r-0
mN r C - c *0'O7%
H4 cn C) q c - . C nID r Cn
N 1-i r4C) N
V
anC4 H
cn(Y)v
- C~ 0;~.
1
Hi
,Tc
L-- rhu 4
1 r-
NLI~q
~.-
Ui
V__
r- -N0ý
OF
(n~O
ON CO
CH1
4 Cf CN
Ln00-N N*C4
C)CI 0( %D NCN
00DH- n00 000ONaO
0l HS11 r-N'. NL 't- - HNO
00 cFuNO ýt n m
ri_ "1 _a'*.-I..

m
LA a'D N *H *0 aCH *'C
"n_ _ _ _ _ qfi _?- _ 4 _ " m C4 _
N _()_ 1_
bG
94 W0 n 0P-14 TlfLn NO'ýO -4 ~Ln 1-UO'
009tcl l N00 008 8 )ý 009a

- 00- -z ION0r- r -)C4
to~ T6n I 9
430
C 00
VV
?-~I r-4N 4 -
o4) HNCAr 0cn
Ho(nC4H C4 C C14rCN
mN~
O0E
o ** . I
*
02 ta
0 ~ r-I 0~' 0 *C0
co 4 4 C)J .- H- ~ %D
0 C4H
4CO~ H ~ H C' N
,-1-4C4HH~ ~~
aC HHi
14-4 00~' 0 0
N
H1
0 %.* .* 1.
x * *
r-4t m -4e C e V
a) r
HC14 -4
Ha) N_0% _ __ 000_)___ ýoc V
nV C40 - m ' P
toa
% . . fLe *
z- 0' * *
"L0 - a *f0 *4%0ti
0
N- C14 m O
C14 cnMO
H
PQ~ 00t *-4

0\*rl'4* *N4
rA
HITT A~ 0~ULAs'
0 4--1C1 r
yf) M5c
E-4 c CO) -4 rS
\o 0.
.4.
ODf
a). 0
41 CV)
0c 11),~ 00% M ~ Ln )r-N 0Wn

H . N- *nCj c
P4 V** i. *n
.H41H .-44 H-T HC .%~. *C'
m.5.HH ' .?1H . O. Lfnf C H %

LA I
0 4 11 0' i4 %ý** 10* c * nc*.

H rI tcn rI-t HC1enN 1 C
all P4 *H4 HHH
ý0Hr-I HNr4CIV c C',N Cj 0 ID 00

W 0 encr,r-
CI4N r 00 M rI r- rýcO 0 00 LAC4 L 'D4

Hr
) . %0 . . H** . .* CO Nr
8n Nicn r, -
o% T 00 0%i.
'00
% 0* nC 0
*n U 0 c* %0 C7
m 04*
H- '. )0 C
r. . 10 . . 4 .. .0 NN . .N CI
0 ) $4 V- C14o 1-4c 0 NcN - , n%
Z"H
0%'. N4%'.DNATN
GO0
H-. " ~0 . .* NID . .
to
( -o r- fn O q -N I N0
H Pi 00n O0%0
e4 (.~ Ln H -. ~ nN
*-0%0 .00. 0% "c O Nc 0 %0

N LM n 00 O c q %0,' H'T
V4* 0n *4 s*- 0
- - - -NIW
in
(A
a)SI aV
3
enq;
---
-
mUUO t7nq;
N~qTf-AVI0 r Q 1 0
--
0 00
009C4 N-
6 0%0n 1?
009 %Q
00 006N
. -? t- -2 nr -0 rI P- t - -1
%g)
-M I4DI
-S
*O~
PO ~ N £09
40 9920 c0
I^ -4tncn 306
-1 o P-
~CH 00
00r-4 -t'? T
%4
00
D
0 Mc"
LnO
or-.
0 *
0)
4) 41
(V 4) m. *n t .
Ncnc N -cItCq un N H
r0
ca o
0-.om,
0 m -H4
-H C r- N 0 14 H Le) D 4
"J 4 7%.1* I -t-
o *
. '0
H4 0 n 0 -t 00 0C . %D 4U l
InI-4 _ 'l
N~r
_ _ H'0
Hn -I
%.-I
CH 0 n *0n0
to a%;r00 Hr'- 0%t)
C40-
It 0 cn n0
z H.0% %0U)
"1- LI)C4 0 O: r%N C4N iNl NN
HI~
coT~~
41si LnUZk O 4cl;. A
0 4 0 0 14I C% 0 0% 001% 0 14L %0e

'94 4-I - --
due 09Lý O. t W 009 C Os 09
t H~ -:rt - 4.1 H-4 c -%0
En
'10 14 CI- C 4.IjI 4 L * 41 r%
.
'0 0TC4
f~~~~ 0 n0% %c
-4 g4
O0 0
0b H
Pw >
N) Wid 00)GH Lr Lrn
0 u- vo0 P0
~co _ 4
5* 1:144. -H
*4-W 0~4 r .D
N-
0 P4 *l cnN
w* 0 - N.
0.
0 NU,
%z~ ON C4 '
r- D0 04 14 -- %
"t;m' .0 MN v.
N,- 4
Ln- C4
%,- W40.8%D -
LQ 0u N80 ~ 0
C14~~ 0 oNI - 0
4fU)enr
0t4 3PO~~ ct9VL £9
r-I
4jN308
coI
m mi0
bo tjH
ca Cu Auu_ - _ __
>14
0o r- Ci)H 1)0)- m - 4
40 N H (7%H G WO 4'
Hý r-4 Cf Ln ONC 4 co
0o - r.
0 r-4.c 0i Ia.c
H C'Cj' .,4 TC%4D NOc
- ~0K
000
co 0
9:*0 N IT *n .*
- r f4 M n r- Hr.- '0
H UJ ccjli' r-
* L*
C -4 z____
1-4 o T41 0 -I C) r
o 4 %cn d4 *oK40 rl0
Hn x
0- ' r-.o H
Lunr-~o,
1-4 Cu ,-43 ;T
N C4 jv
N ON "H ~ ()*o
jf ) LM N~ C4~0 a~ 4 4 Ha'
00G
.*.j C4 a') n 'TU, 4L
Ai 1 44
N %nl00 4 cON
Nf
E-'. ~ f 44. r.- Go)
N~ - -n
0a .0* (br :le ca- Hca 0

7jL-a *44.4 I1.~o * *- A*~ a
N ~ Cz . r~ 4') .~. 04
*H-- ,*1 0N%

H , E
-" --
A- - -;n
t- -
ON OCO-0919
T9 99 61
ZS409
000
0r 0 fn.r04 00 cM cja
4 1-4 0C0f NT IT
H14r-'0
N r,
M1*C1 1 3
4.J 0- 004 0 cf)
t;~ N %, 0'T
z -
g.. H tn -
H 0 -W00N 0 0 r%0
E4 m $ *C0 4- C.J
4ZC" - c'4 C"4 "1

H C30
-1 co 1T
00 ()0%I 00 C
en ýI- -4 a
o 0 r j-
N 0 ' CON --
Go t
ul CO4 N P4H
Hsi THO'cO * 4**
r-. O*c 00C
ý44
U) 0 N'0 00'
%44 V4 bell 4
UT
%0 ~ - pn ~ KV*uq
ONH.I *'0+O9
Z
I 31-
o@00
to-4W '4 C1 -1C4 I
44
-n M L
g., 0 0%. * - N C14*
4J$4C140 *@04 N. @%Te~fl *-

Lfn u~LM 4@0t
Tf T n . 0 r-I ON04N H In (V
0 4 wI~f 4N' 1% r-4 . %0 C P.4Ln
-2 ITf. -r n c -Cý0
K40T N N (% GOu 440 ONU NP-40

E- ON'~
q.
ONC ON %O4C-OO@ NWr~
0C1
c'J.. *C
M'0 0'*O--4M
* -T~ 1.0r- *@0. %C 'T .. h
4c'00
M~ %
r.4r) l
00 -l -. -
@0 -4~fn@ 0% u-40c
-
~
ca ~ en0N ~ na
,0 en%M0 %0 I @0@
NP 0 "4%0 -T0
* L.j 00 oN 0
M-c4 %7.j
C)
*n @
(71 . GO c~*. n .
4 NP40
o~ ~ ~ %D000
~ ~
Cr4 c H 'c-4
K -4Nr
1. - -H .ICNOC. 84
1 4 0
-.
f-. *C .
%"I.
4
UC.
6. -. C4"qfM -
I' 'T '%TLn V- r
9A (44
'T N t% 0 ucn e
C., r%0U% @0% Nn
m0 @0 ** -*~ -00 M
54 -d '.i Vf,'c
M4 &M .o -.
.:. . ..
I- c-4.0'4N Z, .. P-4 NT m r-4 *0% *0V' *%*T*uc
%
pa 4.8 4inl
0 C1 4c o U
i 4
cnZ 'o co th In MK No'K0 c
O4 a)~ "; ~0-: 4.4 0%.D 'm
"SN- ~ ~ I
.NO 0~~ V-4p4 ' p
T-
9-4.0 co~ % 0s- 4
CLI coT C4 r-. %0 41qT nn;0 T q
-4
ON4Ps3o
enu~
POtj
N 1-3O
A
~O n
0C A)N
O!
0.1
r- V4 f%-% -T oP
-tP
ed.. -
09 09 oeam
-~q~t - _~o
*A3o I~q.buln;V TzjO
B31UO PoZTI p020og 03puw
333M I ~ ,U
pt~
*Ofl *IOT4 IT-2 OZT9 olog -
fnpn 49TT.1 UM pn311

00p
-ITW
0 r
-- - -
A!
4C
N.c t Cv 4% %C ICY)0
V) r- L n *r *ý *- T
NI.)
w 4 0 *Oq% N- *CI C-4 OH
0C4~i *r- HNPý
ACY'JC4 I' HH-
C04 C4 C4
4Jn Hn £f1U'
M Lf14~
r-44 -
04 C n *. (n1
tfC -T04
14J %cN W1 NT 'o *ý H r. ON VINCS)
X CO'.O ON .a C c',r.4
I-i HH* 0
'CJC HO4
o -4-
H H 44c 0 ý
9- CO 0%N (2 % % .-
in 1.~ 0c rI '4
44 C qc I
. Ln-
en O'T . -4Z, 0 C TN %
i 2%*2 ,4
0.tn- 0% ell
44 %-Ius* n e?
c* FN 0% - ( *C C* T
en 0-40 Ch P-4 00.P% I~
nq~ae
TJIU~ *AOS tnITO"Ai ITnfqTaM
0_9 _ _ _ 0
6 ol 08
OR.
1S
N 4 UTI 4TT pufloaq Sri~ purnol
Hp
312
oI-
P4
> GJ' aien co r- 0O P-10% 0%cV
1-41
Hn L'- N mr-. 0CO~
C oir iiL V
"; ( 4 .Ln V-4
0 0. .
0r MC ON ON 0~'. c'0IHý cf
*N(;i 00 C4'
I- - -A -ý -
-) ----
zr . r- -1 -o
H n I)i Lo 0.O) as. r- in4
1-4 .T m
M -~L fM. ON en
z 00
t- - -
'.4-
I - ~" *
'00
Lnc
.0 r4 wj
in
r' . ( .. 4
1-4~
144 'A e n0 04Ln'
- -tr -1 -n~
Aý4 -
% cn Da T*4 -Tx
-'. * 8 -4 '0
-, c
-*
- - - P4
0- -, 1- 00-1r
04C H,-4I H NF1- NH \CY
04 C) N Y **a e N C)
40m 00 0 .) N 0
4.1 rci __O__
n~a NO *.
0 wO 19~
.4r-04,
C4r. M N~ MrN LfILf
w 0,% n HNI
4
r0(V 00 c'.O0fl-M.t I'Dlm
w- r-4 -1 4.. C14 '0 *ýr 00
Q N CnN -40N 4Nr I CN C~n0V
0a% OADr-. CD It."(% 0 L(ICM

0r-.ct
H ,~cy) .0 r-4f m
NH04 -4 C') .0 'AN1
'0** . . 'T .* . 0C',
H nC4-104 ='T '~
N06L
`4-yAI 41 T Lfý\
Z 0
0)4n 14I*LnrL- -0~
' 00 cn
H
P4~~~ ~ ~T paO
0 0 nr- 0r* *'.tr- 0 0
N*
q- w~ cc -,I rJ m N.. N
0) j * C- Ln MH c . oC *NILr) Ln r-
4J W~ H-
_
_ _
0_ n v_
LC'.. FI m .T %.0 P1 00
0 %0 -T4..
I1NID
1
1x4) 0 '.0 . n . I .* '.0
c . r-4
4 IrnC4 O . 0 z-'.T
4%. C40
V)IN
%0f 00~'0~0~~O
'.0 0-4. ' 0** N.! ' 0* '.0 *. 0.
*4Cv
N Tcor
v C-4'. N-TN
5-4 CJ 44C
""1C
c H'4LI
T nN 0UC' m 00I*C'L)
'.0*
. C *4 aý '0* 0 T
r-4
.M;vl00
xsa TTncq1ai TTn'qTq TUuuoN jTfqTe.M *A~grS
o x 01 01
0', 011T I
jdmaj, 008 008 08 008 0001
00 boa& tE iE'VE T ýE1 KT

- -
v.-4
010 UT1 ssp POeITI 99SP3 P8TTTW
ON DP03.L-9 8t 6C9 0o9 T
.31.4
iiiQ)"o
bc
0d
M-J
(dco 1-
x
rC
4¶0-4r
--
in2
Cr
C%
.I __
C4
*-1
L
04
CcH
-
H
ON
C14
0c~e a E-4
CI Y -4,T. .. C - 44.
'0 1
LH c'4 %DC4
-f) -I . C14. C4c .
CN
Cqcn 04C4J'
0 4c
0% %0 OCN ý% 5 - TL
H Kn
.4 ; 4 ,r- O 4 ne
H C4~cn N C1 *H4 H AIi
9QH It~ W OLn Hc *r *n 03

C)1 mI *-U
A!
E4 x p 4 'T . CV) * r,
H 4 *HO rr-%
.n *N HY
In Lro~ 0ir0 0%
-ý
N
r-I
;> r- I -00,c
ý
x 1*j~
00 008 0 D0
0)04
8- 0,E-
8 ) 044a : D0 \ :ý( 0 %n 0-
41 11 I P4ILnJ -
P910UU
po 44 00 r 0 4
coa I%
t T04 4cie
all Ný*
%0 3151
en . 9-400 .w %, " ý
- --- - 0 -0-0%-.
,-I
co
T-
0o C 0or
to dFr cHc
ir) cn 1HLr,.-4 W0 00 r- -00

00 00 Lrl .- I-1
00 H) C)ON
c*O (1)0
41 Nn c-40 . r-T M
v-4 MNN .40) CJC)C
H
0 00
HJ. o ~ %QOaN ..44. ;r- uL N

m 0 - -- r
N N
1- M M r-0 -
Hr- *NH
C- c%DO
.cO c'nI *c D C
0 ~ C4~* NH C;a
z 0
0n HO4 _ _' 0 __Y)
H C
I ~
9z~ WN N N
00-I0 ON*0% Ln PO 4*O

> -4 0 HO *', -T0 0,0M14 NO *NO-T '
5-4 0 -*H oL~ H N f
4J)
-1 Cf -4 _ --t _ M.
>4-)40 00 C4 r- A4E 'ID I, c o

:D 44 U N'. j '4. ON NMON C)
'4 tN 44.
;5f~* H 4
r4 ~ q~ '
OHN*
VE- N 0 1 N
*tu u ~- -0
-n m -0
-j N Nr
dmD 0n Cý 08 08 08
0
H 0
-ma -8 -9 - 0
'0
E- *U~ tl9TTodoaj9L '4STTodo.1W8 H
NN 0-I.H
T%
In
1-L
N
l 19T3 t-LH
-1H
ON~ spoo-. - 09
-9 9 09
316
borI
1D 10 >
0 w
0l 00o0
00
4c
coH aoH
p M r- r-O 1L
00~U c)0 m
cu Hn
CA) a, CO4
.~I
CN -- 0DC
"0 uOD
N'
Hl (ON.N
0e
no H
0 H IJ
E2H rX ca
w
r-%
r-4C-4
r'. %M H
I*r-4 -
41 - 4c
H 0j 00 C1 rN en
0 " 0"q H .H
u r4
~~44J
- %D C4 g
r4 E44. 00 00 -0H
H
N 00 j4 r-4
ISM
dmaH. i 08 O'O*08
NN (N. 0 ~dqTodaD
EHO
rI -o VD~tl HO
61-IH 0-IH 5-11
't
' *l
BIOD 059 T99 Z5
- a17
50 I
H4 r
""u L -T -n H- r -4r%0HI
41t~ oH
H- Hc*1
r, 01 cLflo' nTU n f o r u- -4
w rc'! r-a ~ *aC4 *l oý 'eoo
Pl8.o c V) -o __ _ r- 0%-- ) %
-.
oNr% 40e r U,0' Lr-qr~-4 cLflcr- oCJ4

l C4 U
-TC
v-4 r%. 00 .* . O *. H *0
. . ~ *-
Ch -* %0 .%C4 (: . L) cn - CN'T '
V3(% L)% m~-u

0oc -ur,0 -4-T cr- C r
%..004 -n -4
0) LI) en
)4.1 :r 0 IT Ch Cr-4 0-
4 L . HON
0c0c -. 0r- '
o ~ tn00c) ~ : L - %D L- n LCo0 %.0 a.
O ý P-
SOc~O C (1 - C4 i~' U C;J 110r H
00 r- ID 00.
.ý Q*. ýq co r0. aor
n *r-I~-' CO00 r- 000 N -% 0

r-4 m - ig
r-4__ 8
o '4 m-~ 'D I r 4.4 r- ýO00 cocO
0C'0 m V4 nr%- P
%Q%
0** %14 L It'4 .o f*
Le) %.. ON C* 0*
0)
£44
oYi Go *0co u C14oc
OD0 V4coc m0 o *%0
* a
r34 4.4 - --
Qo eq4r4 C%c00 0 04 NC' U C4 ýD% co N r-.4 '.t -4t
aa al 00
OD 44
*40
3oa 09 09 01 09 09 091 *H
au~ a a6a a
1-E4
0N'POE90S9 09 0S9 059 0S9
318
A
41 %. :L cG
0-
0 tn 4 00 -4-I
W a r- r 01- -M a -T%0L
in 0 - *T c -Ii'o.tn '0. %0
U) H~ ~ r * C.l;(9
0 coc -z -
GOJ n -* o L -)c
o~ 'K ý' r-. a% mKý(, 0
0 P
-tA- r4JCý 14'*H; I'.Cý
ne
H r,- It V41T 1-0P 0 4 n
0'
in OH C6-4 O-
v *cc
-t* %D,4 'o i .0 HOý
H Li ;C~ HO . "
U)i. > K q 'IC N94 4 %T
0cc No oc% Noc0 00%N 0P4 G

"4 O- 4 )
4 a-
a
c..
-'
r.
enO~-
oTen (N
-*H- -
4
*in
0e*
c
l %0
m Hwf- *r4m U~. % 0 4 .0% "
- -
UK 04
C" 4) 44
H
U) D3
.t0%
~ -4NH~~c #H 1-t
0
*0'
-c4
0
~ -
*H H
'o0 OP m r ZLN c
H Ncl T' * N T~q*tiC *A* '.Tlqe 0'
T**O O S
0,~-To .0cJ 0*c rN ) *0

Hmo 09-8 -s -s 041
*.?HiC-IH g-'IH ES-11 4.J *H H IH

ON 6HpoP #-t9 099 C
t'99
Z9 UEH9 .0
- - - - a, -19
P44 >
0 OC't '~D~r- H
r. r 0 - .__I
0 4 00 .
- . HccI~- -HJ m0O
41N p 0 4 M.
V)X 41 w 40 0
0~ w 4 C)-. C
H
Hwh~
C~4.0
~ I 0
t4C.
-IHL00 -It --
:
00~ '00'I
x T *-CO -ItcI,- -It . ** *,

H .- n . * -I~' (sf-
c
N- m 0 ON ý 0N 0, *n M
n -K-
0f -4 --T r- P-*r 00 O
(IN N *00 H
00Le
z -ý- -
Ha -a IT~
%D. cor mC '4 '*T- nn V
H- w %. r-
C4 OH
- H0 ON %0 NT
0fI mCH0 M C~' '0oCl
- r- W - % - 4.0,O
44 -n I - - --
Ln 0 4 E-444 CO - 1 0'-
- -, - -
T oJ ;T3 H
29
(z00J~
I. 00 'N ON '4ON
W ON
320
00
Gi0 H 00
11 >14 -
x ()o . . tn 41%.
0.00
-H
41 p %00 ýT H-- c csJ c C
1 )00%4eWr-
rh cO . . m x%,
0 cc'CN e.4~)
en? H
e~r%D cn % %%0% 0r4--

H %0 . . r.O.. LInO cy%'OH
%*. D r- Ne 1
r\oI *r0 Ln
*
t~~ % %G -
ccccr-0 0. Nccr co
r-0
E z. 0 cc. 0
Wdc' N(0 r-4
C4r4 t L TN N ' 0sc NtH~
-? -
z H 0n
m'. H - ý Z4 0 -ON0
0N Cý P, N CV) 1-40C3CUV
Zj0uj *NH LAv- eJ4

LJ N*0C)-G
04~ 1- 2
1 c%' 00-:C1.% (n A
- -0 -n
K
1.1 44 C4
0 0 %O - 0U N % ocn
(D) 44
%4 ~ tK r%D 44 co co
r-s~~ n 4.4 r- %C)coc
W. N.0
k - Sn~ " 4 *4 M-0
0z Hed 0%..N
V0-4co %a *%t
Ln
HD cc.pc Nb. be*H Db
I.SMU IjnTfl Lflq~so 'A3*S [TnqT*
1do 08 08 08 0
-;a~ 09 092p o9 0eeS Pdq

09P
- -P"4 .%-
:1 0 98VSIH K- H H -I
' Nk*699
;p 019 T19 Z19
321.
00 H
10 CIs P,~(f.
r-r0%
COIJ *l cJ 00*A 0 *% N
%D
Nl -~- It f
unF N co HN
o30
4A
to
0~
~ NH
*r . 4 o
Cj0
*n '-4 O 4
*I% -o
. HqCN *cn
0r
t ''
4 l 0o0*
enH H'-40 V-1 m "cnt-
Go C)
en en r-. 0%
Nn LM0
-4C .* IT'-.~ 0~e

MO r-4 M 0CN ~ ' 4r- q
U) ' ' N **N '
NO 4 %0
V0' -
tQ 0 JJ NOr
s'*4 .5 .O N 0' .4c*. .Or.

-4 -r-4d' .4 .0H0- C4fN~
Z -ttL ~ 0 L I~
'0A
C* *'
NOND0 % 0 oc 00"t
1- 1( in *Cc* N 4 7s
- 0 - CO - -,j.0
w0 -.
04 (n. Go N1 N
C4'
E4 8 %0 - -4go LM
-Pý I40 Na' '4* '0 .4

"C"4;AoN "C -44-- l -4~
6
4.4 4
~I41 LON UN -4Nr
0"0 en %
04 UN 0 U'O N % b:1
&Mc "4 Dr. N .,. 0
.%0 &M1 4
V.0 N u4 m'
000
A3*~nq~s*A' V rTTnqT'lM~
09C
WE OC 06 0'T 0'
0 809 09 09 08 9~
'I'li ;i 4STfodo3p~fU 43TT0 oSJIDI"

I .pa pedtIqS
I. 1
%Dn T"9'034y
00 00 £~-LH95-IH LS-IH~-t LS-.IH ?'

'ON IDPOD C197 W9 M' 9L9n LL9
322
OCCD
41Q ONAL
. . %%
h o1 L i
rq ~C1.l%OLM %aOL
mi%' 0 0
rII4 o1ocn ItcM * I
0m * 00. *0t CJ
r-I
%0.d 00 gsC4
n a !0
NO U U)N0'4
H r *Lf%(, HU0 *'AeNC

C
CvOn FNI C4 00%

'O%
*t . (4 C.'l*'0
14C~4 . '4.* ~ 4 *
c -4 U'T"- P4 0 -- OHC
0~~~~ U))*A0 )O. '0
a m IA)
= H.
. 7 a%
fn a=~ .!
IT'0 4 #"' it' '4
V- .n00 m H %m 0m ON
.. ( - co0 I C V4.%D
T -461 M UNV"A31 1tlW

a4 a - a4
PA 0!Ch0 0
H49 fn
~O tTTnqTOf 'A3* ~lTTnqodOZAP3T
I dus1 09 039P 09 09;,~q

a - a a
JA*n IS 9-Li{ 9O-J.H I-J.UVI DSL
w) U) a
,ON *POD- Rt9 609 099 T99
323
F0
II
,o r
w V0 N '4r P-~C4e om 0 0
N %D ri 00 *l No *
"T
-4 1ý
*.OLA
-
0-
0"- 4 ** 11N. *ý 04-1.ýP

*ts( NO
,! , m% f
P% H H a r-9~4
-%0
00 C4 " 4 . .C1 0 *- 0
0 1. 0-4 C1
N T% -N
NC0
'1 C4 4LI)%D r%Ch
co r4 M N. * I r-4 %
"MN1-4' 04f ~ '
[0
04 0 ~ 44
t.. 00 H0 n o*1 -C?
"4..
"43.4 "4 AJ P-4 00 'O4
IT m'~ *c
008 r- V*? A*4 r- .4

N. %0 Iýt
O%4%
@ 'AO NNcNP
. lxg t 0 w - n 4
0 0ao
K" m 04 0
to
go
'1
-
-m
d0 eq
-4
,; 0
M 0IA o0
f3
-n
w4
324U
No
.(o
- -
-
-4
V-
0 t sN
co-
0t
-oc%
u1
cc 0c I
. rI0 r-. 24C4rl
4) 0)M 1 D n0 hO
U) 4100 . . M 41 ~ r%
0 0 - 0
o'r%. 0-4 0~ 0.
~
4-1 ~ ~ w lO
00 cn% 1 nc
jH fr
nD cn Lt)I- V4QC4N
m X x 0 . . I . mN.C
C1 4%c
1 00
-4 eqaM1
- -o
Eu 080 0z
0 r.4 Z Aj 14 C
9 90 eqk o4) e
999 If Np~) .9r, I
3250
N 0
HID
H
W- 41 00 .IC. r . . 0
Af\O 0
i-H CCId-.U cc ~H
4jH H0 0 -Ic 0 *000m ODC14
* cc
*cc H c ~ c
oq 00 -n me 4Cmc
-t
.ID. 9 oC;
\0 C4r 9 St
r-11
aA N(O*
* -T n m
"I HT Ht
m 0% H T N 0c c'
C4
H H)M ccnc
- -Tc "t
E-4 X4 \
H N NcVcLf
H mN a)N
r4(~
zN - - T 00-
L)r
1- J tNN C O N 0'T 'T
00)4 Lnf00 * N * N
0 'Tm N r0 '0-t , HN
E- HOIT% q Lf)*
V' . %
E-40 .H HLrI) *f)Lf)
HN0 HO m*C.J'
% M ( 0
F4 CO) . rf o r.
0)0
ý4 09 09 09 08
boj99T 99! 99T 99T
In wl U. SlOE 9- PU14UITOd
699 "ON06 tPD19%
326
J-- -
Cd 0
4) 140ý 0 L % n0 -
0JH
(0 P.-
woc 01 H in ~ 0 O D m00 ItC

-H 41 H 00 00 % 0 O Nc
C) .*Lf0 0 C',H
W0 H
in~ V)e P4a% 0
.crt H- CV) M~ 00 CV
'0 0 C-)Z V40 1

Hir
%0 C4~' f-4
Pa .0T
O HCl) *00m
ccc4 00~' 00~-4 a -4

M r- enenr- CV) 0
z 'I0D( r- 4
H' H CT%
0
pq H~ -Hf
H 0ý 4.1 O-1 M 0 %0 V-4
H M4 o LMLn rO.~0
. . 00'
IrI I .
8 HWU H HI Z 4j
~
HOýa
*H;
- 0? -.
H-4'
r
Hf-
*'- O
n-
*~ H
C)) ____r-
4-I U
m - - -
-I0 r-0% t0 HCn O Lf
9 8 440
oý
HO -.t.
'0-0
Lv.I
'ý00
C4
rt. 0
H 0 F-Ir- HIA.
Ch*H *0 -tc
E-4 E-N. 4.. 10 * UI, 0

P4 friJ KK
44 00 rh%0I 0 LA*ý Go00
44 000 Ol%en % M 0000%

0 44 Hn * tM
O 0 %T f-*
04 %a .4 1.C
saic[ *VS T~nqT*M *A93*1 *A*3I
0'T 0T 06C
inLN - - --
PqT
C"4 ;.Ins S
33fS-JM SC-J.U 6SZIH 6S-IH

*ON *OPOD £69 069 C69 969
327
Go -
"C4
0 t0
-Y
00 ný r -- n
V)
H04o
0d - H
r
Lf)
t-a"r-
H Hr4
r-
C C) ~ -1' CV) fl,
r .
C1- l
0- *H.f *c10H (L4
kH 00 It.)0(%HHMr-I
C'1N *'o nC 1
0 ~
94 .
00 kInV 1-401%
110O 00~ a m0C D
Lt-T
00 r*
N..) 04 CY *n 0, N0 nr
* l
~~~.ý HH 4.LC*- C14 H
.Oý * v-ý 01 * *- C. . Y * v- \-
0v NU)H 14CN rcv-4 - 1 - oL)0
C)~~~~-N
I r l f 000
1-4 HNT co m0~
00 C11 CO Hr-V)col)H
z Lr-u . -c r-co
ýi-.0
) 14 C4
*-4O 0ý e'j-?40ý C;J C14~ U)
HH
0
44 4 0 4r 0 . U) 00 *It
NC 0C 00 4
0- -d
E44ue w'r N N Cn 00-I ko 00 -ý
Ml~j~o
r-..
crlI
H0 *0 z-O
a
0 C: 14C
C~H l
H en M00 cIenj 04 V4 - c enic.C14
H% r- 00 r- r- I 0 C~H
%0 r4 t-- N
1a 4 0 4 r- M - -- -d
H~H
)4 4 4OnO
40 co.It
r- feq
~* H.i * *
04tA H'Oi 'D HvDDIN
-o - -T 0
44-T0%ON kn 4PIH
IT r n 0'!-4r 0'
-- OC 0C 0o
.f)08 08 0te9 094 0
-ni -V r-4~ r- 0C-4n -00T
00o 00! oi 010 1 0 1
09-LH400 OiiPV
LH
*o~06 6640 Opol YZ'

069 o
WE9
328,
- *0
9- - - * - . .
0 0- -
030
NO0'KYI 4Lr -
0T N~O1m r4 NLI-r
mI' 10
-410 M
' H. . Ný * -1I . 0'O
0
O
00
0 101 O
0K
0 N4
tf0 .or- H .
HD 'K'T V)C C)
C)C'C Hr c r) Lef'o
H~0
H OH
0'H
- N., C10LON a%0111-4 ('4-
tncO N o)
00 r-~c
*o :~.NQ HO7%
E- 0
0 o-T r- co r- N - e'r
C4 cooo
0 0%
(nN 4.
j . r . r.t o
z H c HcOC.
N0(
z
ON U)
m00HN *44 M
cON .1
N . ' . L
H N03
O- NZ -410P-4 oLn Li
0'T0 4H-IL
H H44
: n n
Cý~~~cO -T N.0n
4C)1
0
W 4 14
C%'LM
.ltf ON 0%4 OUtk00n-
100
go 0
0 H N-4
Nýr~N *O
('4%1
-P0
L . r Oc H.4
%0Go 0%10 N r4 C HNRD
ý4 u
04 %-&J H *ý 0. .* . ~ H- . 0 en
00 44N .~t CU %0 0
en CU 0 U~
rm r14 0 & m -4..1-
TJ AC SA
&M. L'* .
NI in -4 %C C4F0.0 %
40 .4N1 ('.0% 1 V~
c-
.0% Xl CA0 N.0(b 0
SL 00 09 0 0
09 00
kA---
Z-I
T9-M 19IH Z9-tH 9-H 44I
q EIIG4 VO co O H
ON*POD+.41
329
60 r-0
40 H, a.OCn V)~ H-r04~ c14M r-.oLr eq n o~
rol 'II -II~

*Oi *HOý NCY 0 o *I
04
00 ~ C1 ~
~C) ~4 - *er
N.ia%4o)r-I 0 C.r-O0
*ý - .. . o .. in& - *
*'O0ý- *O4. -t1 *1.0
M. -~00 o
%tv~M~ ~
M 04HI HCU4
DL) Mr-' %0 rIcn ce bo
94 ~r-.00 r%. %D-t c T-4 11 H-I afn c C14O -.Lf1N

d -4GOC4 N 0 r4 %. M~C cý-jo ocn NcvC M
M _ C _ M _-r M0 %%0 r _ -
__- r
H* 0 t_
9 ~ t '1C4 k ý0 0 0 o 0 M N% . 1.4
CA8 0.t D D m-A l-O00C

-am ' 4 Ln ~i.-.
0
1:0
DO0 *.O C4'N ( 1CJ t- %o0!
Hf-S Jm-t~4 0 00~ur% ri CV)
C4 0 %o 00 0'.T coH%0.- N H! - V~-4

.1.484
P42f~~ It 0H- r*-*. IT N M
LM -T.IM 004 N C
0
0**N O . 0 . un
in~ 000 ýto N 0
-~~ C o 00 It00 *H C4 10
O U(in C~jMNH P %D ~ M00-
-4 00~ ý. -4 -t r
v-4 0)
V3 -V0%inf Go004 ini 7-

00u~ M0 l e T
n .. 0 . H
4. . Ln .
> .0i44 %Dr4 0 N
0 0%C4 -%o00*0 r r,
%W P4NiN - % HMco
I*- ON%P4
Q% . n*
~N k CO0 V4 In tn 600 0r% %
u0*0
M
MN -46-4.0 UMNH
P4t%
sow
,LI Tnq;.t jTTnqT.AWT~nqT*.j TUW.ON TTnqTOM Tjnq;.I
nwp39dS 30q 29g p94o9 Pug posuznj 94291
306T 06C 0.! OO

0C T WE 1
Sii 09
0SL 009
~[ .09 09 09 09 09 09 .
4gtpZ pufloigPOT4
z9?la E9sZH E9-10R E9-.1R E9-.H j 9-Ill
3W0
rI r
a)~
a)V0 Oa% N in Ln 0% NLn

Cq r-
Q) -ItLnc NOO N T mI-
(1) 4.) '4r.. C4*
Hl IU VJ2 04N N~ -It
4J 4 f- 000I- -0 0 -_
HH4 ' - CV) O Um '0 ko Ln 0OC'4
w M xLn . r-I . . 'H . * ,:t. .
0 L)-I i) . t. %o*r. -C1
0 X I
Hnr * * q-= in %0 0 H. 0% P -4 %o .
&4 -1 4 .. -4-
H.. .. .
Cj2 mO.
olC qI m 0 )
CA
or%0Cn00 O Ifb C- -4'O 00 0 m
1- co(f N - . *n u-I r- OD S1 C (I
U,4( .0 tI4
- -4 -Y
OO%
M# C4'.O C14O( m C1 '#
c- -IT ON4co'4
ci a d P4 N 3
.LSIQ~~~C 4jlqS
A0Sfnq.%Tpq
Xuapd
tAW~dw H s4LAL
1b. 009LnG 09 rý;

09J 1 094
Ot 090r.
'T C.4 P
'ON
*pOo r r.% " 0 ITt 0I N6W4
X -444 C4 . . k3*
- >U r-- 0- n A MuI 0 c 0P

NO
0..
0.4
44'0
0%,-, HH CY'JCN r'Tr-.c -0 nC14 0L
1.6) 0
H *- On 0 ý %0r n r- r-
N HHOLf, %0'
LA *o 'T 0
r-tL ý ~ .- .o-' .. Lf
L
000 0' *n *1 I-r
. * n H 00**
. NC% LALA.
.OA
CO rH
*OO tf0 0
%O - Pc - ' P-4I
0 ~~ 4 %D-4en -DC' 00 LO
W r-4C'.
r4 00IC14 tmto en
A.0
ý ..
*
'04 1ý
* V
(07 ..4Cý %
COO)
. N.*%O'4 iA-.'i8 r-4
In S) -n - - -, %-
H C14
C4 ; W.. -
* '0 ý *ý
km *')
V- 4 -0 -- -
a N
t.E4 CM r-I' 0*
Ln ~ Cq O'0H
CI 4rý L COH
V- ?0 - %0 L q D-0e. c* N
4) 41~ e ... I.''z . .o~. T
NtL -) -) V% ý 0%- 00 raC
-' - -
-ýC 4- - -
094 09 09 09 09 0
..Sp pu44 supj4I
N l 9 D Ar c %- - V-1
0 -n -%t - -- 4 %0 0P
U
'ON~~~~~ L~L U~ Cp)
r4 i~33I
04 0
V 0 coc' -#o.T0 - 1 nL
a)oO 0'40 -d CN0 0
En 41 *1 He..
"1
Hd PO
.. ~ 'A H4 C14 '~
H4 m
w~- 0 .?P -- cp 0r.

N jnr-Ie rn( C,4
LF fý 20 r- 17 0 H
0Hcn4 LI . Or- O
-4 - 000- ý
to. 00
- -T
HC' GHC%1
H- HO4 tv n( -nc
go H in -4 K
1- ___ * SA * . C-4 . .
0 04 H% *.4
-0 * en
H r4 -n L nI
0 %yO 'OIT 0004

d. 4 9 In %n~ o04
Lm ON- m . rls % .
4 114 N r"'40% * O
r-i's
-n C
m
iiSI m1.o LI;,
rNqs
(n1 plio
en
- V44-.-t N
09
O- 0 06C
Uld pa';pow ev
Iwo E9--M E9-,LH £9-LUH

-II
ON4 *Poo+. 9Z
333
N)1
41
0
0
-i
-H - 4
-T
D
-n H
0m
-o H N N1n
W
-
N%
Hd HnC1
H0 Nc 0 H0 .M
0l 0 -04--
Ow~ 0 .49 ON '40 H% 9

00 I- Ho
%a1 qO0
~ M.0 4N #) 4NW -. 0 c HIOM N' %0%4
'0.1 *- N N N I*1 9 H; H
0 ON 0 en- oC H - tc
ccO N5Qne 1-4 1
IC ICrcoO H C
0%4
H# coI
Hr ITmD e o o1
0%4 HN-"4. '00 CO 449.', 1-%0ý
H%
0. r-4 *
H4 N
0H 0Qf~. 00 HN
G -4V-
a% n -4 M
d~f a9 4 e* n nG .4S
r)% t NHc
W% to. r. 4m r.H AI-Hn OH
NHm HO
- o % % 0 4.C1 Go 0 co-
%* Nl
0% 00 * n kn* l% o0 -
b-I)
I. 0of .4 .4 49 .4 N.
&j NJ N. r-4 -49 VU4 n 44
DI 49 *- 4- Hn * N en0
.
) P-0 0% N-- cc 10 1c-n - %P

V-4 %a. C4P%
*41 049 4 NOcc 00 n e
- 1fl4 41 tA A& 0 O
o"D C4%D
b~ A~ 0 X. 4n 4 bU a0100A D
09 N
OA' S.. NPH IC C
H *t' 0 - - . 49 *:. I
@ur4
011 0-. 96Z 001 .44
0- 099. 001 O"4 009 N009
coca
0% IBM+ UL Oc TEL Z~EL EEL V
U
334 ;o~
c'j 00 M .
0 P.
,-.4 *LM *# *- .(n% .. #rr,
(A %0- ONi D GoW-e

' V; C ; 4 OmVC
0 V- &I'j'
la
.t-gT N . 1 4
U C O.* ' . C', J. :. . 4
H- W 41 4 -Nn *lNm M N 0nLnc

W4~~~~~ M ]C'4 *0 .0 4wO\ 0V'aIn.AC1
93 H~U . Or40 0n..
%0 O. r40
6.4 $4 *4 * ri-4%
(
0 . -40 "40 enO *-
0 l-- m co &A r%
r- U (L-tr4
CV
% I~Cr4
NN( en en4,I
%D- 04 l0 -4- C4
w KC1
un to VI *4 Ul k'f* '0 .
10 (A4 el ul N en fn 94 in.4 *
00
44 M . . en . . .c . r
N PW C~srN C'~-4 4UC .. C 4 %n
I CO ~CO
C ~x A0 .0 a' m A CD No b.
TO T
r~~S 00q' 00l. 006~lf 00
In -'cc
rb*'.v *C c o
.C9b4 .0 OUEo
ASON A00
.LSI -C E-E
~~n~eTt"lsft~3I13O
600
0 O r-I r-4Lr) Lrn C'.)r-c0 0 ON'c~

M 0
"000
ocaC4 Cf mV
("ONf
NN %D'' 00)C)-r
V) 41 Cj * Ltr g . " 0m
*4 C
'd; .r-
n
40*C",
H C')N C *r-N
"N ~C C4*iC4 .Ncq
C4cq-
e ,CN cn n r *cn W
L~0C' m ý
0 00i Iz.11 -4
$J4
&,4J 4P r.-c00 rI'- ,-Ior-- C-4 0 MC'c"I
r-
*ý c * . . in~ . %D .-
04
P
C"~~
HVN 1 0n 1.
'-N n NH*%
0 .
r co 0N'
0% M~ ocIaLr- -tn c
in * N i00 * r-* ir o o n 0 m It -
-THcH 10*Ln~ *LfN * O .00
Ho
nr . . r., . . '
Z~c .
Ht0 LAC4
m~C~ m .4T
H-N' C
~N cnN
LHNO ~ *0n m mr-4 * H
0) 4j -~-T0 00C14H r, LnOC en"%

H 0
fr0 ~~ ~ ~ ~ %D.'
1 .~t NN0'-c0F-C~
I .A 00 . Le) * . %0 . *
03 $ .
0 W3 41 cHH i-TC 4 .L C~I ~1 I
N
-AO
10e I
-0C'4 r
1
HO Cf4
ro tH.n 0 N~f
LAH H00m 0 CA0
-1D 0O
H
C .) '.O
a N"OO - LnrN 0r f' M4 .n -4CA
r- * n 00 .*~ . in . L .. . .
K0c m.4TD
0n .- *o1,n Or--L .00 *caca4
0
Ln~s w~- ko*1

cC4~ : nD.1e
CAHL
ý 4)~*0 0 * Or' C** LC
H
14
1-144 COM . 0 £90
- . 0L - - -LM000
H-
Hk H-.i
du4jr4 08 -)
08 ooroo009 L -1-c
9P0
0Mvvc ;CL 0 ZL .0 te 0
Isl TULON TVONTI M TnqaMjjqT336nT9

Cl
i.
- -
0 0
.H0 H LC' -00 mvI Hrjo n V) n (

tv. %o . . M . . U, . .
Hn C14~e
LOC' . *iAO-Hr-
r
HHrIc- H M N w-4 H-I
0 40
0CLA C-4(N
Itm%0 m
N'-' CLDad
rA 0 * t . tn . . L AM
LA Cr4L r. - M~ NH Vr-4Nr_4
p.~r_ *1 - '
A %D -V) ON.
H '
Lna)C LA0f % 4 O HH m N r
r. 4 x W 0 * r- . . A *n . LA
4 )W C'4H4
r_ -Ln ( *I.L0 n O
V) ~~ ~ - ~ C *. . LA .
C4 4) to Lr O r_4LMAcn -LA"C
H4 0l HI %0HP4JCQt
v,
.oI4 0 * r- .. . C. .
%DX
14mj 0 08
% 008 006NM 006
- - d -- nr 9 4 r-Ir
0& w .p~ . 9zL c 87L

camV- - .. M 3no %0 0
C\4C
41* 0-9P
(n C14 N H H
enN~ r%-rlH
HM4 Hr
41w 0%d C14 00 0% 0 %O

C . N%~0 C1~4~~ 0 0
rn
-H 1 C C14Lf . %D fi %0 .- Hl 04-
CV
L)('4% *tý r. .O f fO
NNO4r- NT
C% *r- 4 r-1- CO)- CN)
04 0- Gor- V).0%0~ir' r'-0%c'C~ C %

11l o 470% r- M 00 r-H0V OHCh.-lr-i rn%1
nO N (nN NO V1-4N *.4I-;

- .4.N
-- - -
In N enN H"N
enf %:t4 *NC
H0
in -o --
4)4 n0 % nG n hmC ý0 - nr
41 ( Co Co 0% in Go Go m 0 0000 00
3.e3 $4~W f *n
- . .. in
W). .
41Q r4 k .C- ONO HUGo aýC4 air 0%fQ CVIl :
*%0
LM Nin .% r .l. .
o 1- C0 &%-1 % *4 ent ' 01.- .

K *1 *C4* 4'N.n%.
Q W .0! z - LA A *l% en r.4 'n m
C1 p. :. -! -. n r - - 0C 4% M
IK
,-4
In 0 *ur e D~ 1-4 d%ur- 0 0
%~o& N(n
*4r4 enH~ (10% V4AC ~ 0%0

0 0 OCn
NW '0% %01%4 O'*U
r4'0
P4 s-4 PCO'
0,' C r, ' C4 0
41 -U -. V
CnJ 00i C 4 t C
1I.
00Z Go0
sW 1.oo I C
P4
CT 9 0L a 0~ 0ia -oC
Icadu
OS. OS
0 O0v o0v' 009 009
OE OC OC OE OC OE
39 4M '*U~u OE '4601V
338
60 0
0
c 0 --- 0
co c7% 'IC
NC
00.%0 0 I~0 0f
.41 4 enr~-co~
Mu~
Kn C . . 1-4 CO 4m .
C4-en F-4 MC8I,4 *4H r-1r4-4H
P% 0 'InC wC 'IC r
ca r_ 'N O G
*,-.INa 1.J0 .n
oC8m
9C CýN 'o*,- IHC4
H en r-4~O
m N rOCN -
H~r- 4j ~% - ~ 00 oc ic 0104N
H JN CO H r-4 fn ,-UJ
P- O %. -0%,lr -4 -4rý
'4
u0 c
N4 .
N~
.~' C4r.
*T
*. %0 N* ' 0
H 0N v - q
C4 (n
0- 0 -D % n P -n
W-4
4.
W en 'OH NcC 0%0 0%
0, ___ ___n
k - -u -
E-.4 u-) OCI-oN '0

c 40 ON % c NO :r
44 o -* co ao coi
*n 0* * 0-
S 4J4 *vL,o *NC' V)40 . f
N '-40
u- ~ ~
*8 04 *M0 00 Z 1-0 &I
.- T T-99q OU T'
OT%8' alO 0IO 9ON
00 m 0 ~ '4 .~o *"C
rI0* 0tI
SH -~a $sa aOT panl%
Pu
duq 000 008CL O006

t -10106
-O -p:) ac M
-s -S
0*339
3-. 0,0
liV 0
0ONOr- 00C' HO%

0 c00
100
14i 0-
C % r-H -rr4 ~N m
Q-
.-H4 r3%r-, m C) r- ON r- N
or eI nsJ .HcMN -,t M-
*n -;r NH30 00 *n 0 0%
0%00 rH OH4 OH.? -.-

r- CO en .. . -T
Ne-Iu
C.
0n t.4
05.4 *M -. ?nITI
e-J r4Cvm
~ Lt00C
N* * ' *
H: H H'0
M. * A 00
.? "' 1 V -q en N
U to . ~ N0 f3
LA3 %t A~
4J
v H% .0 L-. M LnN -4IN
H bi
01.'
r- C% %0 N 00 O -400 O-.%0?
ON cn . -4 - . D. .*
19 4.) 9 0IG
4 4 .0
*IM * ?tO
-N -.
%
%n~ %. 00 *o 04. (n
1" 0%H OH 000 .0%00 as 1 '
14 C; .~ 4'I'D oi
9-4z 0%
en
j40" c
40 0 %?
-%T
P4- P4 %0~p pinoen

Mti
0%~T
09tTOI T9L
**po*.
6AOHL
TMON q9L
434
i H001 00I00
009
amol
C'4 c
Ný 0
H-1 H
00
O) 0 sti HO cnC00 ,4 q H %D
0)0a) HT4 00Ch ON %0' Cq00r-4zt o _
HO H1
C4 . 00* ON M o H
-H .CC- 04
NH 'NHHf- (n 14 % N
0 1-4
4.11 0%0 .4 -iV)% 'HC m
ItU cn en
A"H ONc' %0 00 -1 %D 0U4Cr) M Oa07
14m 0 00. H** C,4 .%0 - .
0~ ~ ~ wH..
N NHr
T .
. *MC4-T-r
y rqHP% N rH
*Lin
N NC14N
0 M' 0'O
O%00 r-.c
0 c4 CO00 T -- c %D Un O 00'
*N4 n C1 C4 1 rNH CenNOat (nr-
w - -- f-00 %0P-r-4H C'ON,

n0 %
NO
CO 0o H4% CO0-
o00 'T~ 0-400 PH4ON
H; en n N C4, *q *- N *4 cnN
00 .0 c4~*t
SH C 0 *4)
ItM mHHrlr4N . 0% 0% 0 co
04 O ) $4 Cn tcn co *U 10
.LnO -44
,.1 &4 xz U) C . NGo H*-
r!*.
4 1
.A HCU4V ';A *,4 A
C!A
.4 H 4ao *cn N* cna
N ~ -4 -t.nm
01 0 00a o0 nr
r4 Hm
m Ln Go00V40 %4-1o~ 1,0 C-
""
HE-4
wA -I -ILL 0% '00%
H
4
4)
-,0
4rmf
-
'T . N * * col
12.
.
4n . A.
o1- ..1 Ln Lm Wco CH NC4oat GoC

;D % 0 rA 0 -F4 COCM
00S- P-4 00
'o
o
H
r4
H
0 -*
0%O.rN
C4N . en
-0
' f-%
I4- i -W -1 Mc -
1-4 o 4.4 00 4
co 1-H CO0000co
UNH"i
G
m
m
O- T
O't
0H-
t NN
"0 "0 A CD'O x

.00 0 A
A~'3I
'LIG TMIUON I~qou TInuoN
TTnq~ei 3AT~~
29qzo 394S *UT OZO*
i dusjx 08 I00~ 009 008 006
H
5.4I *U*J 698P3 POTTH
1%
I *3.z 9 '4*066 '*UTUI OZ 'A.OOVT
S 014*Poo, V9t S91 994 19L 99L
* 341
0 a)I- o MO n 0 : nM r -a
.HH
0 OH Hn
.fýLr) 0 r4 I Ul)
CO -
Xn * H CV) Cb M 4 0 n
H1 r ON 0 Ln*C4r- ONC N*m 10
E-1
x. . r . . )C*. . .Cv) HC NH.
IFc - - - 0 7 lL
0 In It MOLt NIOC
1:TJC -4I M N
A .0 C)r0Ltr *Htf Nu'

ý ý Nr_ 1
cn H c'o-I oML TC3 r-1''

m (AC4
m0 H ON
0- M % * 3
HOO-
c *U T *COO-4 H
-4-T )
3% CO3 OCl -. 3%3%) - C
aO. m~j~If P, It)C m 00 V

.'00 . . *n . \3% . .Y
HCH'Jm0 *O- *NH
4 H-4C4 03
0 o
0~ ý4J P,' -cn otOnL O-AItH
3%4\.H LA.Cel
Ch. - _T O NO0\ OHLA
0
n CO. a\ r- r- In - 0 It *1 a
ZO M
,-, E-4J CAH .O . NC
. . InHN' N
Hn 0101
(9 ,-43 H' - c '.4e \.a - \
o~~~I
r-'. 0HA H.& ' 'A %cn
CI4 H r) m~ A N'% HA -
ca HE4 * %.9L
> Ha
m~ 0 . . 0 . . i . 0
%D04 r ID 00. '.ON0 It) tm 'OC A3%L
OCO I.~ 11 111 *4 * N -4 N

Hn~~0
1-' *OA . 0 en'DO
HO*
4J H 0C QT CO-.?
.C) M
- Ar1r -- \ -0 0-4 I
z 4 '.O
O O'T( O* NH O*%o
x-0 ~ ~ -
Hd .a 00 0-P 009l 009 00
o- o -- -j
490M~ ~ 9
~4c
d) rar-4 D Cf'f-~
co 0-0 r- eq W) V N 0r-.
d) 0) C14-4e C % 4C'Jo
%I4~%%a cO4
0000r- n 0
041
J 1- ,9 4,; I"C "1j C6-.' C4A
C4%' Ln
_zrC-1 Fn v.4 N C4 H cr*Av-4

H pi-1rI -
0-
~
H ~ ~ .O rIC4 (00-4P
0 NT10C'4 C0o
I". '.0O%
n
'T (ON
C H C 1C4 H- mNO 1
HCh.
04 c(Ct cýo
4 r4 oi%D C1 Mo t-r Ntrý0
(0, t . . t- * . rý . . '0
n o .00 - -D0p
Lncn O
e enOrN-4 '.C4 n04 r-4 4 4%c-jI%
K M LI . . r. . r- * . itt..
.10. en. No
*c-jr- ~ 00
c tU 4
00C'HCN O
H o4 1-o00N %.00
V-.co M 'at%4 0%N-4
wut0 4L 0 &, mH'O.-' C4 C ' o ý4
H.r-w '0 . *L-4n 0c U *CO
.00
0)4 to % 4 4 N.C14
It-. H Cc-4 Pi #~-I
-I E- V- -A -
%w~.a'0 00 %0 N ' r4H(n~ 1% u r %Cn

0041 .4 @00 *0-4 *0
r4 P. - N LM 00'0. c -'H04C
04 ' * %a P. at-4 . %Dm m
%0 * I4 * 0
%0 kn 'I N4 -44m f
TVszu
!uoiT~nqT9A ITnqI'TTnl'1 IT~nq~ej.t
00 m. -
P4 - %4
V,1
174066 '*U$U St '40oTVT
*ON aOP... ti St L LLL RLI
343
________________________ ________________A"
9*-00
Nr-
en% .cn *
NNr-i N~
41 ,-IOM~N 4 ONO4L
0OI 0 cn. n % I
44 r_4 -4 r_ en, C1 P* %DVIr
0P NO0 NH LM - 4Ci 13;J e'J 1
m4 N CA
ITD
1i30- r- co raN 4.L, c4
.I~- n0 ..
0 0 r4. HL00'r.-H . a, -tL
W r-n
00 %. or- r- v-4 C4. 00 OIDc
Ic Im * *n.. .
_ . DN.
z- H-
r4n -; - ,n-_ r4i
2
H WE- H4
M c . 0r-H co D Nq CON
H A A J4 HN II inL
NC~ H e
Wn ,E D
- - N-l )%- -ý
"Ar-I N r-cd c 4 ?_4n t- CAO'

N fa LAH N0 %00cn
r4 w r-4 .04 Hr! i Nr N-

4 _4)
0 ' ' 'T ID L

~~0 * 4n
OH ON 0''0 c0a,
OJISIQ 'A*R'1 *AO3¶1I nq;3kM T~nqTeBi ..I

399SMS OUT Mt
0T 0-T 0-T 0-1 0-T

Id 09 00# 009 009 006
MeM E O OE OE O
88SP3 POT11W PUB PUnrIOZ
~ *;~ *.ui'&,SL~6 'O"'P OE '40B'T#
'ON OPOD+ W1 09L M9 M ES
344
1' 4)
0.)
0o
T
nC4ID
-l
--
-% L ýý
%0
-e
00 -s0
-n
(~ic
0( W.~ . f~ . OHT r-
H 0 0% OIA
ULn Ký * 00* *- r- . n r- H0
* I M
NNHý HHH
_ C~-4 Cn -4r4 H4r4HH
S04c-
C4 *K *K r-
r_4fn .* ' . .* e . H.4 CO~

LA *-.T .H- Ii0 *HN4 *0~
P4 HtH .'N CnCJ,r4
MC -4C -4
N M ~N HCC4
enC)N N
o 0) 14 %0')JC~ n 'nc N
nP 00H00 H
W 0''L
cn NOO OH'A ~-
cc *NO .%00, M ,H0 P04 ar-
l N ýa N f_4 ' *M 0)en0,cnN r4N 0
en - - -q0
0 44 0 N ON P-4M % ~~%D -4 enLAN Ac'o M
N~00
0i0 H'(- N4
*C'0% ma%
n 4 4' 4
'.OH
r_
%0LA nco MN
"D co 0ON*
0'* 0 .' N * H * 0.
.MN
(7 LMU INa% -4'-
.0 C b)eX CA : i .00
-~q9 -~qo -A31TnTMTT TS
l~q~o 39949 'UT S

- ---
B- - -Z
d-* 09 001 009 008 I006

bojo E OC OE OC
* '311 V '1*916 4#Up3 OC '4o0PTV
-ON epo_-t SeL 991 Lot SOL
345
A-1.5.7 Other Non-Ferrous Alloys I
346.
on r 00
*'0
M4 60 ý4
P40>r o-c 003
o. -.T a
000.
0. 0.
04t.1~
UV ' m-
0,
Hz~ CI
0 -x 0
cn~ ~ t --. Ln C4
w~ LP) r-ecr, 5-i

-1 C) r.-I 0 i
:3 z
go H~Ln Ir
oz~~ ~~ rI%
F4t t0Oi M fl. rl0
" ucv-'. 03 'C'JC) tn r'.U,
(J 4 5 Z
;4 e4Ln 0 I~'~ N~
Us-8 ~0
4 I0-ri CI
:4 o3 0 -Oc 0w - u r-
0 0I: % - ccrP nc
% %00-4-4
0, <- 2r *u4 00 en ~o 00ro -
E-4 1 rtC' 01 44 ' -t'
w- U
-ccf~'-
T Z -occ
-:t-~ r-4mr
IT 44 ~
ý4 nw -ý a
%aL 00 0 0
x ~ 0ý4 O ý 0 "L 0i
00
IIT .b*j
O0Z
*8 -ba - '
.f4
00 00
to U) C4
U~4J 0* 13
0~
:31
H*
Izc) -
410
H
0 r- tr-
U W
%-
H pH
"434
_- ar _D
- -I-4
H-MC1 --
4)o . 4 .
%0 %D Hr--T Lef
00 H- HC14
m ~~ Lf -4
> K4 0H
OCH
4)~ 1t-.r ON (n -t. HN4 .N%
V). 00 41 44j asV440 44
2.. 0d m CY) 0 hI dN
C))Hf-4 cn 44) r
rr.4
Hr
CC\ %D *f enN
,44 .11...1
1--1 0 r -- 14ý 0 % -
oA o" c000ý r LAr C4C

"~ Hi If 7, : *n ,1( n"' *ý 0I
A. H *c1 .u5 ,*00Lf
04
wo00N 000 04 t- C4N ~ ~ %-

a,0 oM D 0 145oL
NO1 Old .IT O e 01%
-) -0 en) -
C'SN
LI) 'O csi. L! m'~
4 .* ;r . %D . us.nChe
-4 5.4
C40 mC-4 -
0- ___ -1w _ C4
'n 4)4J *- .-4 * 0* P-1 *4*

* N
0. - 04¶
-4
H r- m co f%% 4 IU
Ex0 0%-t
-N L -0A
a CNN %-
44 44
:D 01 0T 0M 0.r0 0T un
v- 00 0 en 08 0 m 08 09 c
0 'm I- -j - " %1 - - a
lI C) 0~~~
%D4
0) n q 4 0
* 349
%0 It0 e N-
a) 4
N0 -. -4
0 Ch
:j
H.
04H
>-- NN >O
4) 0%0 0 % %.D N d )%
0) 4) 4 O r-~ 00ITc )( r
41 w 1N* )4 0Lf,
LI) fn.H .. Hc
CIf-H HH CD
oH 4 C1 I o4 Je
w- mi~am r-
(D C* N * oN r-4o
*-
~
*or-,
N nON - eqMC1
cn 0d or-4n0 *'ON0a a4%
0041-- 0 Y4C
Ea H V4 or-l I M~f
N
nCh4 0 H.0
r4 1
*NH
en4 m n0 c -
4 z -. A4J 0
Pr4
T~nq;,M T1UUON 13
lqq3O UTPUDR BUTpUDO

*TSUO'I *A6U.1~
-r ---
I dw 09 09 08
A *u ogp3 punoigz nSGP

.Ispus
¶34% podvq 0 pno!
POTTON PTOD
09 opo:4. 66L 009 T09
350
PAGES
ARE
MISSING
IN
ORIGINAL
DOCUMENT
A-2.1 A METHOD FOR INTERPOLATING INTERFERENCE VALUES
The values of Interference given in Appendix 2 are highly non-linear.

As a result, when interpolation is required to obtain an interference value, a
higher order interpolating procedure will be more accurate than a linear one.
One method for interpolating at a higher order is to construct what

is called an"interpolating polynomial." The interpolating polynomial gives an
nth order interpolation using n + 1 interference values from the table.
A-2.1.1 Higher Order Interpolation Procedure
There are several types of interpolating polynomials, one of the easi-

est to use is constructed using Neville's Algorithm.(16) This method and the
symbols used are:
X is the value of an interference parameter which does
not appear in the interference tables.
Xo,Xl,...,Xn are any n+l consecutive values of the interference
parameters which do appear in the interference tables.
d=0,1,...,n are the consecutive indices.
Pk,O(X) K-0,1,...,n are the n+l interference values correspond-
ing to the values of the interference parameter X ,Xl,
0
Pn,n(X) is the desired nth order interpolated interference value

for the interference parameter, X.
Pn,n(X) is the last entry of thi triangular array of polynomials

Pk,d, (as shown in Table A-2.1 where Pkd(X) is definted as:
(Xk-X) Pk-l,d(X) - (Xk-d.l-X)Pk,d.(X

Pk,d+l (X) - k-Xk.d
Xk -Xk-d_1
where dm0, 1, ... , n

k - d+l, d+2, ... , n,
It is important to note that the value of the interference parameter,
X, for which the interference must be interpolated can lie anywhere in the
interval Xo d Y 4 Xn. However, the value of X may also lie outside this inter-
val, in which case the problem becomes one of extrapolation rather than inter-
polation.
When the interpolation is required for only one interference para-

meter, say X, it is called one-way interpolation; and when the interpolation
is done for say two or three parameters, X and Y or X,Y, and Z, it is called
two-way and three-way interpolation respectively.
353
_ _ _- - - __- - - _ ...
_ _ _ _ _7
Interference
Values From
Values of the XP 1,o Pj11 (X)

Interference p20 p,(x 2 2 X
Parameter2 P20 PX) 2,()
* Inth order inter-

Xn n.o Pn~l(X) Pn, 2 (X) ' Pn~n(X) polated value of
_________________________________tinterf
erence at X
Table A-2.1 Construction of the Triangular Array of Polynomials for nth Order
Interference Interference
Interference Values Front When a-.025
Parameter Table _______and Y--2.5
(d-0) (d-1) (d-2) (d-3)
.- YmXk PkVO Pk,l ___k,2____ k,3
0-1.0
0X P0 , 0 -. 3079
X -2.0 P10 -1266 P11 -0359
(X-2.5) x1-3.0 P1 0-08 P 1 1`
-. 7 P2 - 74
x
40 P3,00m'0182 P3 ,1 -.0638 P3t2-.0817 P -.
"*0782
Table A-2.2 Polynomials for Third Order Interpolation
354
The following examples will illustrate the application of this higher
order interpolating method for one, two, and three-way interpolation problems.
In the example chosen, linear, third order and fifth order interpolations were
used but this method applies equally well to even and odd order interpolations.
A-2.1.2 Example Problem for One-Way Interpolation
A particular part has Smallest Extreme Value distributed strength

with parameters (0,M) and is subjected to Normally distributed stresses with
parameters (P,a). The calculated interference parameters are:
a - $a = .025
y = M(p-M) - -2.50
What is the % Interference?
From the tables in Appendix 2 on Page 422, it is noted that a - .025

appears in the table but y = -2.5 does not; therefore, the value of interfer-
ence must be interpolated. The values of interference as read from the above
tables are:
Interference
-Y When a=. 0 2 5
2.0 .1266
3.0 .0486
Linear Interpolation:
If the quick and easy linear interpolation is used, the interference

corresponding to o -.025 and y - -2.5 is found as:
Interference - .1266 + (.1266-.0486) (3.0-2.5)

(2.0-3.0)
- .1266 - .0390
- .0876
or Percent Interference * 8.76%
Third Order Interpolation:
Using a third order interpolation, by constructing the appropriate

polynomials, Pk d(X), the value of P3 3 (X) is the interpolated value of inter-
ference. Where'P 3 , 3 (X) is calculated'as follows and the results are tabulated
in Table A-2.2:
(XkdX)pkldLX) -(Xk-dl-X)Pk.d(X)
P (M - - -%-
k .. . Xk Xkd.l
P (2.5)- (2-2.5)(.3079)- (1-2.5)(.1266) * .0359
211
355
! 1.
iP2,1'2.5' (3-2.5)(.1266) - (2-2.5)(:00486)

"2,(.) 3-2 2 .0876
i .i.4.-2.5) .0486) - (3-2.5)_(:2018ý2 03
P3,1(").
2 5 . 4 - 3
i( . ) = =.0638
(3-2.5) (.0359) - (1-2.5) (.0876) =

P 2(2.5) 0 1747
P3,2(25
n 4 -2- (2-2.5)(.0638)
r.5 e (4-2.5)(.0876)
( - .0817
SP33("5)=--(•'5-(077)4- - 1-2.5)(.0817) -. 078
08
Interference =P (.)=-ý--5)L--71
3,3 4- 1
or Percenat Interference = 7.82%
Thus, it can be seen that the third order interpolation gives signifi-
cantly more accurate results than the linear interpolation.
Fifth Order Interpolation:
Using a fifth order interpolation, by constructing the appropriate

polynomials (P (X)), the value of P5 M(X)is the interpolated value of inter-
ference where P 5 5 (X) is calculatea as follows and the results are tabulated
in Table A-2.3.
Pi(2.5) - (1-2.5)(.6321) - (0-2.5)(.3079) =-.17840

l,1 1 -0
P'2,12.5) . (2-2.5)(.3079) - (i-2.5)(.1266) - 0.03595
2, 2 - 1
P3,1(2.5) .- (3-2.5)(.1266) - (2-2.5)(.0486) - 0.08760
3,1 3 -2
P4,1"")" (4-2.5)(.0486)
4 - -3 (3-2.5)(.0182).- 0.06380
P (2.5) - (5-2.5)(.0182) - (4-2.5)(.0067) . 0.03545
P2(2.5) . (22.51(-.17840) - (0-2.5)(.03595) .0895375
PP3,2(25
2.5) - 3
(3-2.5)(.03595) - - 1 (1-2.5)(.08760) = .0746875
P422.5) - (4-2.5)(.08760) - (2-2.5)(.06380) 0816500

4-2
P5(2.5) . (5-2.5)(.06380) - (3-2.5)(.03545)= .0708875

5-3
C
Ci
LrA
_:: Ln
U-Co
ii a 00
PL4L
Ln LA)
Hr Hý Co4
0 0 %D a
4J
cri Ci CA H
cn IT Ln
t-T -0 04
NI. co I- -
co a - Co00 -1
ml000 %D 4
A,A J Ill. 'J 0 - 04
H N CA N -o
u 0 N -M
VF4 LA00)
44 H) (C) 0 0 CA
H0 H4 H 0 H
4) 'I ^. *;c . n 0^
14 W C A --T LA
'.0
('I Co o '0
41 0 041 OC
H0 CAH l 00
HO * 0 * 357
I
P3,3(2.5) (3-2.5)(.0895375) - (0-2.5)(.0746875)1= .0771623

3-0
P42.5 = (4-2.5)(.0746875) - (kt2.5)(.0816500)
4 -(1 - .07816875
P525) •5-2.5)(.0816500) - (2-2.5)(.07088751
5 - 2 - .07985625
P, 4 (2.5) = -- 5---(4-2.5-(.(0-2.5)(.07816875) • 077791h
P,(2.5) (5-2.5)(.07816875)
5-1-- - (1I2.5)(.07985625) 07880156
0805
Interference = P 5 (2.5) - 5) 0777914)5_- (0-2. 5). 07180156° 0783
or Percent Interference 7.83%
Thus, the fifth order interpolation does not significantly improve on

the third order result for this example.
A-2.1.3 Example Problem for Two-Way and Three-Way Interpolation
Two-Way' Interpolation:
The preceeding example demonstrates one-way higher order interpolation

The method can be easily extended to the problem of two-way and three-way inter-
polation where values of 2 or 3 different interference parameters are not in
the tables. [For example, a and y or C,A, and B(x)]
For two-way interpolation of order n, the value of interference cor-
responding to interference parameters X - X* and Y - Y* (as shown in Table A-
2.4) is determined as follows:
(.) Determine the interference values l0*,Il*, ... , In* for X - Xo,
. o"t A-.2at
Y - Y* by performing the n+l one-way interpolations as discussed in
Sction A-2.1.2.
(2) Determine the value of interference, I*, at X - X*, Y - Y* by
performing the one-way interpolation on the interference values IO*,Ii*, ... ,
In* obtained in (1).
The next section on three-way interpolation contains several examples
of two-way interpolation.
Three-WaY Interpolation:
The need for a three-way interpolation may arise when the fatigue
strength is Weibull distributed and the stress is Normal distributed. If the
S- • - - II !!
__ _ _ _ _ _ _ __ _ _ _ _ _ __ _ _ _ _ _ _ _
CI.
0 0
00
* C
4. o
r-4
CD4 r4 9
H44
C)
0 .9
U H H .10
X -
Ci359
interference parameters have values, say B(x) b*, A - A* and C C*, and
none of these'values are availabliE In the tables, then a three-way interpola-
tion is required to obtain the interference value.
(1) Determine the interference value I ** for A - A*, C - C*, and for
a given value of B(x) = b appearing in the tah1es by performing a two-way
interpolation as discusseg above. By repeating this for various values of
B(x) bn,bl,b 2 , ...
b b , appearing in the tables, n+l interference values I
I* .. , In, will be generated as shown in Table A-2.5 (a) and (b).
(2) Determine the value of interference I** at B(x) - b*, C - C*,

A - A* by performing the one-way interpolation on the interference values
Io**, I1**, ... , In** obtained in (1) as shown in Table A-2.6.
As an example of three-way interpolation, consider the Example

Problem No. 1, given in Section 9, page 140, where the interference
parameters are b - 1.808, C - 3.72, A - .0875. None of these values ap-
pear exactly in the tables, so a three-way interpolation is required.
As it involves a great number of calculations for high order

three-vay interpolations, a second order interpolation will be performed
to demonstrate the technique.
For the three-way second order interpolation the following in-

terference values were selected from the Tables in Appendix 2, pages 398
through 392.
The first step is to perform a total of three "two-way" inter-

polations where each two-way interpolation requires four "one-way" in-
terpolations. The necessary calculations are as follows.
(1) Two-way interpolation within Table A-2.7(a) for b -1.0,
A-.0875, C-3.72
(i) One-way interpolation for A - -. 2, b - 1.0, C - 3.72
(interference
values from
tables)
Xk - Ck Pk,0 Pk1 Pk,2
Xo- 3.00 .1372
.1210 .113872
(X 3.72) X1 - 3.50
X2 - 4.00 .1081 .115324 .11491744
360
_ _ _ _ _ _--a- _ _ _ _ II ___I
II I
qA
UH
i.i 1~4
H C--4 .
^ 0
4- 4t4 - 0
"'All
UU
x 0
41361
m mm mmmw m
i Interference Values
When A = A*
B(x) and C =C*
b0 I0"
bI Ii**
S(I**)
b2 12 *
bn In
Table A-2.6 Determination of the Interference Value

Corresponding to A - A*, C = C*, and
B(x) - b* by One-Way Interpolation.
362
*1
I!
(a)
3.00 3.50 4.00
-0.2 .1372 .1210 .1081
b = 1.0 0.0 .1095 .0963 .0860
0.2 .0852 .0749 .0668
(b)
A3.00 3.50 4,00
-0.2 .0639 .0490 .0386
b - 2.0 0.0 .0477 .0364 .0286
0.2 .0348 .0265 .0207
(c)
A 3.00 4.00 5.00
-0.2 .0359 .0166 .0088
b - 3.0 0.0 .0254 .0116 .0062
0.2 .0176 .0080 .0042
Table A-2.7 Values of Interference

as Read from the Tables in Appendix 2
363
Pk,d+1 (X) - (Xk- X) Pk-l,d(X) - (Xk-d-l-X)pk,d(X)
Xk - Xk.dI
P'(3.72) = (3.5 - 3.72) (.1372) - (3.0 - 3.72)(.1210) .113872

3.5 - 3.0
P21(3.72) - (4.0
21'- 3.72)(.1210) - -= (3.5-3.72)(.1081) .115324
4.0 - 3.5
P 2 2 (3.72) = (4.0-3.72)(.113872) - (3.0 - 3.72)(.115324)
4.0- 3.0 .11491744
In a similar manner the following results are obtained!
(ii) One-way interpolation for A - 0, b = 1.0 C = 3.72
Ck Pk,0 Pk,l Pk,2
3.00 .1095
3.50 .0963 .090492
(3.72)
4.00 .0860 .091688 .09135312
(iii) One-way interpolation for A - .2, b - 1.0, C - 3.7Z

Ck Pk,o Pk_ _ Pk_ 2
3.00 .0850
3.50 .0749 .070456

(3.72)
4.00 .0668 .071336 .07108960
36
i 364
M
(iv) One-way interpolation on the values obtained in (i),

(ii), and (iii), to obtain interference for A = .0875,
b = 1.0, C = 3.72
values obtained
in (i), (ii),
and (iii)
Ak Pk,0 Pk,l Pk,2
-0.2 .11491744
0.0 .09135312 .08104375
(.0875)
0.2 .07108960 .08248783 .082081676
(2) Two-way interpolation within Table A-2.7; A = .0875,

b = 2.0, C = 3.72
(i) One-way interpolation for A = -. 2, b = 2.0, C = 3.72

Ck Pk,o . Pk .1 - Pk,2
3.0 .0639
3.5 .0490 .042444
(3.72)
4.0 .0386 .042444 .042444
(ii) One-way interpolation for A 0, b - 2.0,

0 C - 3.72
Ck Pk,0 Pk,l Pk,2
3.0 .0477
3.5 .0364 .031428
(3.72)
4.0 .0286 .032968 .03253680
(iii) One-way interpolation for A - .2, b = 2.0, C - 3.72

Ck Pk,O Pk,
V Pk ,2 'r •
3.0 .0348
3.5 .0265 .022848
(3.72)
4.0 .0207 .023948 .023640
165
(iv) One-way interpolation on the values obtained in ()
(ii), and (iii), to obtain the interference for
A - .0875, b = 2.0, C =3.72.
Ak Pk,0 Pk,l Pk,2
-0.2 .042444
0.0 .0325368 .02820240
(07)0.2 .023640 .02864445 .028520123
(3) Two-way interpolation within Table A-2.7(c) for

A = .0875, b = 3.0, C = 3.72.
(i) One-way interpolation for A =-.2, b= 3.0, C =3.72
Ck ~ k.0 Pk,l ~ k,2

3.0 .0359
(3.72)
4.0 .0166 .022004
5.0 .0088 .018784 .02084480
(ii) One-way interpolation for A - 0, b -3.0, C - 3.72
Ck Pk,0 Pk,l 1__k,2_
3.0 .0254
(3.72)
4.0 .0116 .015464
5.0 .0062 .013112 ..

01461728
(iii) One-way interpolation for A - .2, b- 3.0, C -3.72
Ck PkO Pk,l Pk,2
3.0 .0176
(3.72
4.0 .0080 .010688
5.0 .0042 .009064 ~.011036
3566
(iv) One-way interpolation on the values obtained (i),
(ii), and (iii) to obtain the i~nterference for
A - .0875, b - 3.0, C -3.72.
Ak PkO P k,l ____k,2
-0.2 .0208448
0.0 .01461728 .01189274
(.0875)
0.2 .01010336 .01264244 .012431587
In the second step, the final result is obtained by performing

the one-way interpolation for b - 1.808 on the interference values cal-
culated in part (iv) of (1), (2), and (3) above.
bk Pk,0, ~ ka1 Pk,2
(1.0810 .082081676
2.0 .028520123 .038803941
3.0 .0'-.2431587 .031609122 .035897 -. 0359A
Thus the interference for b =1.808, A =.0875, C -3.72 as ob- A

tained by a three-way second order interpolation is. * .
Interference -' .0359 or,

Percent Interference -3.59%
Interference obtained by the three-way linear interpolation

(as shown in Example Problem No. 1 in Section 9, page 141) is:
Percent Interference - 3.98%
Hence, a s~ubstantial improvement is provided by using the higher order
interpolation.
367
A-2.2 STRESS DISTRIBUTION - NORMAL
STRENGTH DISTRIBUTION - WEIBULL
368
•I.
A-2\.2.1 tress Standard Deviation -0
'N
369
'I
TABLES OF INTERFERENCE IF(X)]
Stress Distribution: Normal Strength Distribution: Weibull
V - Sequ (Equivalent X Depend on the

Stress) 0 material and the
a 0 b operating conditions
9
Strength
f(X) Stress
x b _x =S (r~si) -• x
0 equ
Figure A-2.1 Interference with Standard Deviation of Stress
Equal to Zero
Weibull density function is: b

b
f~x) b-1b
A.0 x 'o
Cumulative distribution function, F(x) " f(x) dx
0
Sb, b b-
370
,I- __ _,
: . ... . . 0, ' .
6O%0tN~0%"a r- % 4 r-. 4 NC4 * 4
U%
%0 0% en r- L D 0%4
% 00
'00000000 *-N fn m' 4M in ' of
* . . . . . . . . . .
pa, o"co 4%D ,"4N cy 4 "a4~ P,0

00 "N m8 M 4 mm a0rO 'I I00 ta0-01 cr 00 C
0% 00 o Vn00O
en NVon cn o M 0 uiA M~ C 0 0r%
'0o o. o In A44cr N 1% o0 o0 C IA 4 00 1%- - 0
V- 4 I A 000
Go 0 LiA
"a (c
N44 0 r- en 4
00 Ien 00
A '
0%ap o in0% P- 0 t' 4 0% ON (4 r- 00 "a4 (c1 Go P

LM
IAII
O 4MZ4A m m " %0 r% w0 N w 00 M~ N4 o0 r
"8
9-4 14
fn o r 4 0ý 4 N4 0 D C4 00 vn rý "4
00000000 * N4 m nV .4 .4 sri LM o
. . . . . . . . .
1. t-4N LM o r-00 0% 0 Vn N4 0% o0 N4 5%ý N4 P, "

cc 0 0CDO * "a N 44 44in A'
. .. .. . . . . .
r 0%
ch0% 1% %l 0 C4 tnoo
U%0%I 40
800 oo * -C N N m 4 4 LM
~~-~ 0*.
.i~ . .. .. . . . .
f%
m Ln 0 f- 0% LM "a U)4"a.
MN N
V4 4 f
- P40 0%0
VN
In %Dr- r- co
Vw V4
"a
N Ci
LM "1
Go 4
.0 Fa %D
Co
0"0
V" CA
H
~O
Q O~OO OO * -4 N N on 44 I I
U31
WI 0% 0 IA " N %0 in U
0p% 4"'
f-%000
"aN A A05 0ao 0. % 0% fn 44 in
0n N
;s01 8 I
P0% tonI "In
- "o N 0%0*5
r-F l 4 P .
Ft N0 w*% 'l *
ool N Goo
371.
co tN Ch 0% C % 0 0% 4m 0%
ol %
mA CA 0% m0 00 (A
0%0 A 0 N~ N 0 % %
C%0A
C4 C4 N 0 0% 0% 0% %
0 4 40% f 0% %0 00 CA 0% 0%
L%0 %0 Nm 0ý m m
0ý cn N4 co m n0 % n
0 l 0 * 4 0 Ln %D 0 0 7% 0% 0%
0 Cn Ný m %
mV 0% A
(% 0% 0%
0l 0%~ . . 0: '0: 0I
Nn 00% (A 04
IA 0% 0 % 0% 0% Ch r4 0
o0
IA
0 N
N
co n
4
0% 04
O
m
00
%D
0%
0D
0%
0%
m%
0%
m
m oo 0% C4 00 0% Ch in P- 4% %00
00 r% .ý m m.o t % 0
tn 0% 0%
00 N
F4 Ik0 r'* A 0
m0
* 6r 0% %0 00 0% m 0
in 0 %0 at 0%m "4
go 0 V b0 0 MN0
in r% m% Pa
en OD a N
In en f0% r4 0%W %
00 0% P4 44N91
P% m %0 a a
NN
I A
"40I A 0 qr4..
N'0
N 0 A IA (A N 0%fn
0at
do
0 . . * 14
O U.4~
N N 372
A-2.2.2 Stress Standard Deviation • 0
373
373
I
TABLES OF INTERFERENCE
Stress Distribution: Normal Strength Distribution: Weibull
P - Sequ (Equivalent X Depend on the

Stress) 0 material and the
o 0 b or B(x) operating conditions
OI
Strength
Stress
f (x)
Measure of
_ _
Sequ, x,ksi
Figure A-2,2 Interference with Standard Deviation of Stress
Not Equal to Zero
374
0* ~ mml('
~ ~ f 'U.J% r.m 4 0*
0* NAOU
' -4u0 0 0. In* trn' IC 4 lojo
2.' ) 0 00a0 0 0*ý acooOLJOC4IOO
0 ct 0 0 0o 00 , u. . . . . . .
A*~ C-)I'('.
V 0 0* vWOcju 0 0 0
Q0*Uu * C-100 ty, Q* 00006 j aj0
0*~ ~~~~ 0 14.

0*
aU'~L
a 00O .
IN
0 nP-fl-
(T M 0 --4 ,q CD 0 Inr- fn 14 .00 '4 0I n IVj
m*.0 0 000
04
0 (o~ wooo
0>, 0 u0u
!.( -4 (p 3*-4 i0
mi~ 0L9.3
U00~ 00 .;
060Cie000 0 000900 4
0* 0* 0r~ oo 4N
0*~ .. 0* -4.~~%00 K4010

i *. V
0l* l n. L - 0*tt% w% o
It I N .- UW
4 .J '.J U W 0* -40 %j UV j OWw
Q C) 00j0 0 a 0. 0 UM* 0 0 0 0 0JUOOO
*n
* 7 INUP(.%0 4000%,ntn
u UI -ujj ~
4 VU
U
0 ~ t .j -40 14~ k5 w~ w v W* uf~O~&
0 *q 0 *C w tf
375
0 * n " N ý 0 - 0
0* "1 - 0 * 0000 000
9 400j- 00 000 0000
:L
01gge O S 0 0V
03 0~ 00* 00 0 0 000a
00
Iva 0 000 0
001 t)e ... *
101 b ~ CIO.
0 0 04
0 0 r-4Q'~ in 0L%~ ~ 0A
UU* 000 C.) U000
0* CJOP0 0) N 0 0 0
U 0000Q 0 0 00
*000 a 0 0 00U0
L) .
0 000 *ec~~
u-*000 000 *
*~~~
00 @
jq utfC)N 0
Qj IV-4'J. 1wV ;14 ellN .4
0 I.0~.0~
Ow ~r~t~J
.No
e') Ono*
00~. .O JC
0 0 0
.* 4I~o0 * . .. .
'u* 0~ ~0
0 0 0C 0
0* Q Q
1-4'C L j a-
*ou
t~ J Q 0* C0 0 0 D-
0*Q . O 0 00
iA*00 0
L0 CO)00 000 0 0
00000 0 0 0001
0
0* 0000
*I *
-4 fn OWNr uj0 ww Q4 0.
tW rw to %J 6.jo
a* %J 14 v u t-
* g0 0 0
v0. %J u .
C0 *
6.140
4 4v l
*i f P0N -4
a
. ýr L. UgP0 (P
N Q No M* il '16*00
4d . Io.
0 j 0woo t u376
0* N~ -4"4 ' 4 0 0* N'O 44
* 0*
01*
'4 ~
00003D00
0 000 * (000000 00
0*00n000
1* bt. 0T1. 00 .
Q
0 4 Dt -74 U *
0* CtC)CT CU 000 00 N 4 0 0
00I * 0000 0 000 0 0
0*0 &00 0 0 0 0 01 (4 0000 0 0O0
0*~ . ~~ UQ 00
1 ~ 0 1 0 * N( ~ ~ .
( 000 0 0 6*' 0 0 0 00
0 0 00 0 0 0- 00 0 a0a0 0 0a
* 6 O ..... 6 * 0 600DO000
4.J* 0 0 0 0 0a0 0 0*0~ ~
0*00 00 0 0 0 6*0000 0 0000

'0 0 0 0 0 00 0 0 0 0 00 0Ol
** 00 0 0 * 00 0 0
0 6 5 0 0 65 %r 1
0* A.- Vo -44 uJ' u~I N

0* -40O jQ
6* 1-0 00) D 0 m% 0 00 00
0 00* '
0W 00 1*t 0
* 0 0 0 0 06 0 6
l.A ty-r71i
0 " 0 0* vi m psai 44 Y-
-4 M*I-4 644aJ ,%L 6* 'T
C11 -4 V 0j
10 c ý v 0 Po :3 0
4* U 60 0 0 .0
0ytQ0 m OC3~ co r# c
*~ P4 .00 .*
[ 377
0*N*4 0 0 00 0* N"-400000000
0*00 0 00000 I *0000 00000
0I 0-
M~ 0 0QQ000 0000
(Vlb 0u 0Q0 0
ij *000000 0000 0 * *.oooo0o
00 0 0 0 0 0 0
0* 0'.~
0 *N 0 0 ,0
0* 0 0 %0
s 0- -P4MW 0 ouo nN0 oo
* ST N 0 0 0 0 0 0 40 0 0 0C30 0
* 00U00 0000000000
0* r IV' P-4 0
m~ 0* 0 04 00
0 0* 000U0 * 0000 0000
* Igo Q 04 044 0 00U 0P 04 r 0- 0N4-
U'N 4*J- 0*1 fl- fn P-i- Q'

-41
'4( '-4 U U ci Q.2' r-40000 . U i %j 4
0 k) I N -*'U~~l~ 0 000 * JO
fP4-0000 U U
.47N 4 O~-MO m
in
P%0 00 04 4
00 00 0 0 0. 00000
10 00 Nfl 0 u0o 000
wo o JU Ufw. 00 OL.5V
0000 co 0000 0000o T 00 0000 Mr
*' %
0 0 0 000
** 0000 0 00'0 01* .4000 00000
0*0000 0 000 0 0000 00000
Q*00 0 00000
*1 *
- r V, u/Il 1 (V r4 -4 a* -wN#1 4c
0~Iu
u D Q 0 0
Z * 000000000
CO* U00 000*
0* 0 N4
0 0 0 * 0* N-400DN -400
r-*00 0 0 0 0 u0 00 0 0 0 0o
:
r* 000~0000000
.. 0 0 0 0 0 0*w
w00u * Nr
V,U%
'0 0 00 J0 00 00 0
*~~ * 0 00o0
0** 0 0
00 00 000.0
t* 00 0 00 0
0* ' o
.~ 0 0I0 00000 ff .000
0* 0 0 00* 4N -I 0 0 0 0
1113 0 0 0 00*
C0) 01 0 0
w*t Qwt*o 4-
o 0 0 0 0 un 0 0- 0t 41f4-
A 0rM
*0 0
Mu
'J- 0TI AWV'--
V0 0 'I3 .Aa.
00 0VUV V.J Q. M... tV" % N
0000 * 0000O00000
0 0* 0* e 0
0
U% 0 j%%0 0n 00
0 ~ 0 0 r- vi %nN 0n p%
0 0S (I0 0
0 IV
0 4* - NI-I -% 0n-r In &n

0 0 0 ~.U
0 C
111"4,.i 0
U) -4 U 0
*~~ 00 0 0 0.o
0*t'
N coo r~
09 04, N 4' ' 00. 1w
379
* -4000000 o 1o 0 * C, 0%..4 - tNI.-4.4
0* 00000 000 0 * N-4C~cocc0c
; p0000 000 * 0 Q~~c
I* 0 0
0 u00 0. -4 000 * 0* C c
Q~~ )~~ Q*
0uýC Coo ,0l 0C
00 0 0 0 000~ 4* O. C-CO~cCO
01 t) a 0'0 0 0 0 0 0 0 0
I
U* '00C0000C0 OC
99 *9
0 00 0 0 0*
*
0COoCCOCcco
0 10 0000000,
U* F- -40~ Q
Nj0 0UU
Q Q 0QU 0 C, 10 C'
00 00 0 0 0. 0 0
jjý
* r 00I*nO Ar 09 00
*
00!0Q
MiN1 0 U QU U
Coo
*;,
* .
c;
0C C
m.NK-~ .r-ccO
0* 0 0 00 4* 0 0
1 M x ; C N* 4 CN'0uc
W
-40 0 %.)C.)U IDf %t4 N i.O C
a 0 a U0 c c c C0
*ý *
0 u Q 0 u* u jL . i. .zmN-CC0
*n -O
O O CO 0 4- 0* O 'cc C
*0 0 000 0 - 0r. CýI
0 0 0
a 019 0
Q.
* ) P- 1,ct1
*r* %*:
Q* I... CA
0 4 u u D 004
Ni mNO 0 Oý
1C
tj 0u u (ijwuQN - c 0 pcd 1. r
0 00
0 60 00 3800 , Cc
_ _ __ _ _*I__
_
0 0 tn
CC 1c,* C , o, 2 CC
C o* ~ ~ 1 0-C o O0 0* 0000
0000 o0
0 000
c0oo~
,i- CC c~ joccj
0
'ojc) c0000 rjC 0 000O
c 00
Aci
0000 C0)~
c * 0000o00 00
Xi * ýC c C*
cC0- c -0 C 0* o oo 4
0 * *' c1* N 0. .. o c 0* 0 J T~
0: 0 0 m cC Ln cmn 40U0 00
C *MC,10P4000
*0 c~ 0 c
0
u10
010
0000 000
C0 000 0*, 0000 0 000
IC C. FC. 00C
-1 0.000c
0ci c~ 00,000 IC*l 000000 0000
0 0
-4C.
00 0C.
c~W .E4 ~ ~
c c- 1c,O .0c c
C', 00 0 0 00
Ol*! ýý' 0 00 0 ac 00
0 !0 c- 0 0* N. 000
00c O C!CCo0It* O
r- ) Cc cA 0 0 0 0
' I 00ý 0 D0) 000
*'
*, C C)coQ* -
* 4UC
0* F 4 .c
o ~ C0
c~ CC)L~ o 0* 0
~~'= ~ NOIC 0 N 0010 . 0D
381
0*0o 000 0* 00 04r 0 00 0
0* 0000000 0 000 * 00 00000
4.0000000000 0* 000 000000
U*
oo0 0000u 0 0L 000000
0u
0*
*
wV LnIc -4 00
0
~I 0*
0 I 0
C'C) 4r4Qu0 Q0
U*00 0000000 '.0 0
q* 00 0000000 OOU*
000 0 00
wI * 0000000000 000000
0~
0 * I1
0
0* 0 r--rN 4-4 0 00 0 ( n( N 0004
oc
0* -4000 00000l0 0 UO0
0 0 000
0* 0 000 UOC
00 0 0 00
r- 00000 0 0 * *000000000
0* 0000~%v- 0 000 *lf J- 0
o* 'V4Q0 0UU 0 00 0 * 40 0 0U0 0 U00(

0*0000000000 0*0000000000 4
*0 0 0 00 0 0 0 0 0 0 0 0 0
z I
0* ('J .4 0. 0 V0U 0* '4u-'Mv4 04 0 00
0* on 0
1* In 0
%0 000 o -.
Ul 00, 04 04 0 0 u* 00 0F000
m 000000 0000 0000 .000
0* 1-r P *nJ-F4Qln
-4--400
0 4o 00 A* I'D M o .0 0 0
v04 %0 U0 0 U0 U 0* 00C. 0 u0 0 03
Q ~
QQe 04 a0 0 00 0u 0 0) 300
ý0 o -
S14f0. 0 00 0 00 a
o N co -I 00
i~ taI ~ 9 1
382
* 0*00*.000 0* **..O 00000
00000000 U0000 00000

00000C000 *%* 0000 00000
U*000 0W 000 0 u* C.o 0 u0 0 0oo

ci * 0000000000 01 * *
0**0**0
J 1
U
0*
SO
*
rlA
0*0000
U~*00
U00 000000
S0U000...
N P-4 40
00000
0000000
0 0 I 0*
0*
0-
r-*~
Q0r4v -4 0)
00
00
00
0
0
0
0O0
0
0
C3000000000..
000
0
000
0
0
a* Ono -I N -4-000 *
0* -4"40 u 0 O 00LQ 0*
0* CJ0 -40 CL000 0
0 0 COOQ 0*w 00 0 0 0
* *00000000.
*00000
0*~~~~rvý-
4 00 * t VOU4 JV3
0 * 00I0 0 00 0*
i0* 00to000 -00 0 0 0
* 0
0 0000 -r.. N*-
4- 0 0 0
-400~uu
.)(J V C) cj;)000 * c 0
* AD00 An-#00 0* 0000. Am"00... 0
0 ry 4 no% T0 to*0 T'1 -40 0 % )I
4 no 441
0* ýN.44 0 0
CI
00 0-' 0* 000
0*0000 00000 .0* 0000 0000
4*0000000000 0* 0000000000
* . . )O00 0
0~ 0 0 0 0 0 9 00
C
0 %r Nrr Q ui 0* * 0 0
0e'4 V. *)
*0 ) N 4-4u 0
0* V0N-
0* 00000 U3.)00~ 0* 000 0D 0 0
000000000 0 OD*0 0 00 0 D0 00C
0 04 01 *.cf DU
1t0
Ln I N 4 ý(VkN 4ý400
0* 00 00000 0 * *00000400 Q~
U U0 0 0 O O0* 0 0 0 0
0 0kJ 09
w. #0%JwO,)(j '
01O00 0 ~*~ooo
T n n .40 : l rNf
0*
0*
Q*
00Q
00*1
.0 0
000400 0
0
u t.)10u
C) 0~0o
0
: -r-tU 0 .4
o o~
ifl~ ~~ n C.)~)~ ~ C,
v u *j IIwt
1* *1) JU 0 0 . 0 0 e 0 a Oto slip
0* WJ 0 "1
!"n 0t 0J l0 kI V V
ou "T Nt ioo
0 6b 6 . C f, 0) A
0 9
-4 I " I I.
%D* 4- 00 * ýZ 4,O
0A 4-.0uQ v u u up t
0 0 0;60 *~, s
0 oo 0:s o ,
* 1 T- (J Ao14. x
4D 0
* *10~~ 0~ 6eo. ae..
384
000000000.*00 000
0 0~
0*, 0* C% '- 0'fJ-u0 0 2 00
0* o 0* wavO 0 0 00
0 C30, 0 0 00
0- 04 ~C"No D*0 0 110
4N4 0 0* 000 0O0

0*00 00000000DC)
00*0000000 * 00
Q 03 0 000C
-W
*
00 000 y
* 0 0 D 0 0a
0* C.000)u )0 0 0~ t
00
uo Lo 0 (1 00oo 0 0 0 0
ONO n 000 P
0000 1 l 9A Oe a F
I It
N0 -T'- ý 4 0 i 0 iiJoJ-40U J
0-
0 0 u 0: 0* w* *1a
*I a *l
*i
o~* ~~mJ~ 0*~'~%J~4(J~jL(J
'6 to 0 1* 0 0 0
000 000 IN ;4c
385
-4.-4ýO0 0 0 0 00 0* 04 0 00 00 0
0*0000 000001 0* 00 0 0000ý0
a*000000 0001 0*0000~0
0 00000000
A
1I
0 0 000000 0=0 000
1 0* ( -4 0 Q- 0 u U u 0 0* 0 0 u- 0 .j00
0*00 0 00000 0* 000Q
0* 0 C.) 0 u00..000 0* 000
0000 0O00 ci0 0 0 0 0 00C
09 0 I 0 0 0 0 0
ol^ OD
U*0000* 00 f,0000U
0 00 0 0000 00
00 0 0 00
1 -* 0000j ý ýpo- 000 1 - 00 0 0 0

0 000000 000 *0000000 000
o0* o - C-3~0 0* 00 0 0 000

0* 0*00 000 O000
*4 0 CO *00 0 a 0 Q 0 0 0 a 0I 0
* r03(7 * rN 0-003
* 0 0 * 0 0 00 00 ;
n 000 000 IA* 00 0 00000

600000
*Z I
0* 1)-' Co Lk 0N w 10 N- V 0 000 Iz
ri0 0%. u-rN14-40 0 0 00 00- 0 0 000
*0.00000000 * 000000 0000
aLf0 000. 0 00 0*0 0 0 0
00 (q"`)u0 0 00 %nf)r-40 0C0
0 0 00 0 00 0P 0
*lJDý 00 0 0 'a0 0- rnWN001zODrN

%n* Ir *o )N% InNý CANý
I In
06 0 0 00 0 *000000
386
0* 000000
0* ~000000
~0 0 0000000
0 0* 0 0 0
N 0 pr4r-l 0 00000
440* 0000 0 00
0* C 0 ci0 0 0 C,0Q
Ln 0 0 0 0D 0 0
u0* -4 0j 0~ 000
0* 0 0 0 0 0 c
$ .. * 0 0 4C 0n 0ný 0
r-* 00 0% 0 000
0* 0rN r00 60 6I"0r0
* Nr~ 1-1 )% U. uIL
00.
N a3 0 0 0 r 0 o 6
*v
38
0* 0''O~~~.t' 0 C)~I~0
r''- %t o
INCt-~ m - 0, 0 N NOD
* w' .4. ~0
NK No I*.I* wlm
1o b0 0 00 0 *0 6
* ~~iO~"4 N40 W0 - M 4" P-4cc-
I 0 11ý
6 le 0 0 a 0 *ooi oooo
91V 0'0
*I
"M 0 coo P- w N t aDC -'te 00 f

001-~ lo0 im 0 0, a,~
'-N -- '
0o 0 0O aDt'q"+4
%DO~L m 0 CMP-0 M 4 0 Cis -4N
0 000 000 0 0,0 0 0 0 i10 00 0 9 .01 0
0a 40 0 0 0 90O0 0 .
fN 0 m NJNP4N m~~~. m ,asmo
00 0e 00 o00 0 70
,0 000 * 000 *,. 00 0.
0 IN C 0% - o0 -!,04 M 0 4 ' o D -W'4640 P- M D4IN

Wi Ot o CkI ilN NM r-P4 C% I- r ý N LA4OD 0 -1 M oP-- ~
0t*L U'WiO'CM,
inr~P 0-Ot40 AN.4 0 so O- CON 4
10o N ' 00 1- 000(C000, ,00 0,0 (7
U3u
IN a' Nr~U 0'rO'- 0 7 %OOt-4
o )0 m% 7 '
0OM.'P4UN4 0%PN,MPr - 4ýD0"40 kdM-c Ma Q
* * 000 0
-4-lNINMMsi...10 ' 01
Pi-000i00 or.
00 %Ci0 0 ,i 0
0* 0 NtCD Q 0 -4W1-ML r%t O4jl-t % f- 400elO

i u h ,a
k(1 Y
0* 0 P4 INO
InNMlt4 00f c ,0 v,0 U(7 ~0 ,O Cp0,
0 i 0 0 010ti 0 0 0~ 0 00
Nt0
Will P4 fnNfC#O% 0 in0U~:
388
m 0 n M 1 .
0*re00'D" 'C ., 00'.
00 P - )of
0 4
" 4-'N
" rfltU'A 0
or 4 *'ON
t 0 LA0..OrM OD'CY%
0OS 0~ 010
LC%* M cU )U4 OOi NýJ1f'- L N'4f NmO' P0 a,~
co . 0 D a . 4 IN N I M M M .i 1
* g 0 OO 000ef l o 0 00 0 9~ 0 ! ,
,Di 'c .wJý
*- *: t 0 0 00 010ý 0[0 000 *j *00
IIla
*s 0,
AAr- 0 1 NI'C0r~o~'C
0 a0m. ,N00
:0,0
1C
'0* *N
00 4' ~0
64NU, IN
*pw !C - :O Ir 1 -
* * N M4 -4. 'D fk 0' - Im 0 U) 10' cc N 0 0': N in IOdN LA cofm

'0 * 00000 0 -A N NIN In~jf
4 %+
4 10 WNL~L C C' -i
:.- 0 0W t0 000 N0 C00co 4 , U)Nfn% 0
* 0- 0,- Ln00P N 4c ' 0 0 , 'C
Q ZZL Od 01 0'fL
4~0 OD10MN
NA N M 0 LA N - N 4' 00 Q"4 'Ow ' N 0n
.0- '01
N !' 0U to94IS
0) LAN
N D
* NC N f~L 0 00 N4o'C. P- 'C It
4~' 0'N
%t 0'
In!
ILA 6000 9 %0 r- CD 0COW
;9000
-- p4NN 1W4 4' of' s oAleoAC'C' '.~
* . ~ ~ 0.00000
:0000*i ~ 0* 04I
00 0001000
0A *4 Io INN mM% t in N % P-O
0 *9
- I~,38*
6r 0In Mi ON6 0 opN4U 00'0 g
11b ,n166AQo-i 00 "0 P4.0 13 i
0=*t).~ H 1- NMM VA r ON Q
'4' 41 000 'N IlO 0 0 041 U
"r " 6000b00 p" 0006 q 6!Wow1

'0 nN1' c oI n M 0
01*
; Coco0' Q-i
P4. cy: r A n!0
N0 O3 G9 c
9990
*1 0 00 ~ JWW
I t)
0*C Ir.-oi O 9`11C4?0 10%o~o

jP4 N iO r 0 r N % 14
U 0* V?'4;o 0 inDON#W% INO0 ~
M*~~~ tf
i 0000i"749 ~
In0 w00 0 % Q6, Qý
010 -40,0 tN % c
F in IN0 ;p %U
(
e* 100 coo '0 0M.4 NF.4
c I"%d0i0 1- 1-OD do kcl 04 ig 40
M4o 0pN 0w 0% 10 0 Oh@~

NO" ONP m t
w ' -in% 0 014 0 *0 M.0 000'a 0in (kP
0 -10,
1
kn,*T F- P0
c Q 0ý~
q 0.- 10 :6Q
IF
F w - n COM-T~ ;~un0 s
0* elNI
%rQw P-1 ~rP- P CV 0-4f40 w % 'i'ufm - cvu~'u
w HWr
M W aD 61
#-Io so
Plot00 o .0
41 4%I- N mI InP
0*0 0 0 ton W 000 Cn
Poo Q00 0,0 P-40 N

NI~@ N*m t co 940
41 040 I1
0*~t 4 0-ru M r~~-e

w~o~ da8 1
1l41 'On. 0N 0 in (P 08
d0* Min a v% ln0 in!

0:%0O~. N" Mjh 04 OW
6a.P o ;i, or! %
to4j~wp lO g 0 0
8
9 ~. F~$u'.M..0 U~U~~Q03908
in 0404 In - 0GQIlN!No0 4 r 14 ON 4- MM Ii ~
0rcc&Win
~o~ Q0 0 0. 0O0 0 0 0 0'0 0P4 P4 ~4 M MBlf,'

b * I ,0 0
0* l,.44111. C4 %r.4C wt tIQOMQ1p i N T
MOI F .0 0 0*e 0 0006

OF Ii *e l* 0 Ole of0 a00 Ole
0 00 1 0 00_ YA W l in nQP
O¶
0*~~~~ 00 Nf~~
0.eI~l O ~ 0 so'.7Nf
0* c -1 . V ( 0c 40 MF:CK)-iiOl n a N.O 4 ,e c 1,0 !.
0* 0 .INO-0
4It0 P- to~ 1"0 iO I 4'O N M INE'mflv I
NE V 04IS.4flIN
Pf* Ym1," 0%INUiW
IV (k - P4Q 04 YP-P
ON tIcy0
I~*i ~ j 00 0 0 *I 0 . 00
090
F w In i 4 N f *
:* 0 0 0
0101 10...0 0!. *0Oe 0 0 0 0 Ol@e@
0 * 4 Ill IV I*MP caI1 "i I f %aago0

fn0*09nF OIfN4Ok
40N it M W% 104 ty ICY
M4E~0,0~
Wil ~
0fl 0~I00=4 C),rcol~f"lot 4014t OI P4Mam
*a a P N00N M~N'0
I"!* V%0 4 140 cc
0* 0 10P4W4NN.9I9 #in4 o9vicQ N Co j-W I4a og

wl*
0* 0000.4 6m0906i''1640c'0NFIe4soig
00
I In. S...........'
* * 0.0 0a00 (30 0 Ij- .74%
C4
T w a- Z, 0!-4Q V
' 0 i M
LT OO00Q
r .! (4) 0C Va0 o -41-4 ;tv ,j(i mui r-
10. Os*** 5 0 0-1 0 1 O OSO
Mo1 INI
k
i( iL
r- n"
ji
Kll U L.)
rK( 'a r-
Jj 'A)-I.--NJ P1 -4,.r-
U C3 -4 4 ~'
1`1
t-a4 IJ<L f-- U' r-NI
3o lu-, i. r-
L30ILM LV L)
-4- j -J
-J*
1 L"I)*4
-T
-r-m
-q :X)J U
'u'T (N ;i l
)
k) -4 ýr- -J~4
rFr'
s 4,
( * .-4:44N '3 .~).u 11-:\ N0-4 J

L JI- M ~ -41W I'J~JT
Q n -O N L) nOI
L.~ H.
Li) Y)
,ft --4 nU
jh 'L f 4 s'-3 'U,% :;l u !:,I n Y) j'.1
4 -- Zr ýO 0' !P- 3J -
JD ,jft
** C1 s Clge ell C41 'e .
*~ I
Q3 ui ll%ý -i 1 -1%At J' 1- I -4 1"J(k- 4 !"1 W IQ %- :'I i.J1 Wi 0 'J ''X (-~I 1-4
'f .* U.ý -4 N r . '.l4Jý'J X 4ý 1 PV %. ) 5-
.-4... . -4
(IU j.N4j n IU
u%14. V I M jýu ; JV V . j .392 kj.

* **ý 0; 0 e * 0
0 * i0 0 0ý0 8 1 l 0 O e 000 O
LI ~ .)C~ 0 CC) 000 0ý0 C" 0- N M .4 t(I'

0 f NN . 04
C 1*000 0 )-4 #N
t ýW !t" M -tO' N4 or 1,- DON 04 1 U'ý-t C~cdD
4 -ZC
0 0 00 0 Cr t4 .00'
NC', 0 4:" 00r- 7 CN r- ,ice
* )
0* 0 0.0CCCN 4(Y,0.0
N'0% N..4
*set o090 #0 0 00 60 00,4090 60 9;0101
C* 0- c 0 N0 .4 c
a" N- o0,0%t'i M
U- 0 M' uT
(7 01 Na 0!0 0 C
* W C CDC,00ff0P- 00000'
1.0 000 0
* 1NI 1 ( O% I % 1QQ CýQQCr
* I pK , ( tf k ,c%0
%t Lmý 0w0.1o , )CC0'0
0*c
~ - Go~0 %C1 on 0 It
%ttI a0
.* C 4
cc0*
0* Nj -0~~.0~0' 'O"03930
0 C)000CD0 COCON10 - C' *010NOimclC MT-m C)0J~
ý% %
0 * 1c 0 C *00 w0 1-
C,90 r l0 009 INco ýý
N * 00
0 01~C -4 %tW CC
OCONN OD~c
-~00000 O60ococ O N %+0c ý t. kr
0~ ~ 001 0i90ac0 0 @0 ý0 ý0 09 'r P1
r-: 0 0 l 1 i l ' , * 00i
;o0oo oc~o * .iN"40% 0 Nr%%tIOý I
r'O00
10
0 010 0OC09000o
0
a ,9 a00 MNmu Dc 'o
see
.* 001600 00. 000u.N Pt 00 dc,% 'e0'o tc '

%* 0*~~
C0100
00.D.a 0 ~ '-0UN
000 0 ~ 0N.VA+
U 0N N;U0Cj O0 t~'' 0
e-CD 0, 0iy
* 0 Ol 0 lo eoot!& 0 *e* *'0io~q~ fe~
9009
: 1'
0* 0 0a ~ o s
C49,4 0 C) 0!0 0 1 N10,11C-
cc(, ,C'0:
if000 0 0 Oe000 O 000
*00 00,s 00 9 s!O *0
a * U u* c. %- t-- t'P4 0 o0
0,t0-)C P 0 .0-*v*N f& " 0
*IC
W0t-0 k 0-:00
0* om-o 00 NDN4 @.NNNOII0I eVc00

oa,%C 0.0,
SN M 4.4P~-a ta @ok 0 0 ap0 00
C) 0 cc 0 CD 0 cc
0 0 C0, C0.0C0.
0 * ADp**
** itP-*&*,M w- %WNa*,*y*'**%
***** 4 000 0C.
14 1om
a0 . mP .V &0 .
0 0 I ,MM% aa Wa 0 7 ,V Q 0 0
394
0* ~ Co
COo O tO 0P t N
11
0 C, 0 0 0 00 000 fie
0 - 0O *000 0 '00
* 00 ,0.0j
Os.... 0 0 0000J Mir% ,000000CC0 C *-,% * %c
0,0 00,00
00fe000000fefe00o 4 Nfie 0ir
b *
up c * Coo, 0000 00 .4
M4 000 ~ l M I 4U 0~ N 1-~ 0 M
N- IOr.
0 * 10cOC0ao00OO00O '0O00-Nl40OOO
* * 0 0 0:0 0 1-000 0'0 C0 0 00 fnIt0C00D0e0!
0* c' 0,- N~ f 4m, P-a 0 M- m- cr w4 c a .4'It N:O - .4-

'0*C0 0'C~ C Q0C. - -4M `*00W'O4 r- UN----- N N r 0
'c) 0c C; 00- -4
*0*.0000008 e 000e000
e 600000000
C 0*-4N% 0 %OCO UN
C0C N M P (I0 Ile 0, 4- aP 00
ac
*, c 0 *0 00 000000000000000c
0 UAct NI C p~r030
i* 0 ,a C
-
W 0P C0. 'C - 10
0* 00 e 01 0 0$
0* C0-
A*
C)* 04 "CO 1I-N'0 C- rC

N ' M M4-t 7 OW C P0 0 C
o'
0C 00 0COO000jm. O, (n - 0 r , 0, 0, 0401 1C0
4 P r*a .. V 00I@ m.-0rWN4qEV', z 4t'~UN 0 -0C'0 cc

N cc P- j~c- '-I PIM.-
NA ft'aO m0'&cp0 000 00
C*C0 00000100 r m00 4 w 1000%&Oaqa 00aa 0 00
* I
a* fl- 4*N w%,0 aP-0 N 40 1- a. 00 C 0

U~l
* 44~ ' -f~ ~i4 10 0' NPI-0 'ClIoc C oe 000
#,a0%MI %p F- 0 0 0 00 0
0 * 00 a 0 0 q 09 0 0s 0%
*w ad 0"I y0 "' -t
.0 395
',0C c.Cc~ CC 000C 01~~z~
CcCCo 0c c c-C~, C0 0 (;Cbc 0 Cc C,-N r.%
00 0 0 0 0 0 0a * 0 0 0
c~0
roII 0 0 0 0 0 c - 4 c04t c 0
o CC co cloNacc c~c~cOpr-.w& rr~ 'cc

0 a l
*0 00 0
Oc ~C)c 0oC-C0) ~C
C) r''%
C) ',c ~CCO0C, 0 C0 0 -~
0* C 0 0 0, c , 0 0 C c Il 0 N C' Il0l 010" w t- 0 0024

* 4r
W** crc * rcCO cc ~ f~
Icccc tI F- ,cý
0' C 0C'0 .4AfI4I 4WNP j14P ý 0r[ýf rC
C . N0 y ,%
C)0 o~~
,NP 4Oý w4 0v
Nff 1 ccN
ýe
-c c'
L: ,Cg.c
w, Cf N 0 N0 N t. W q 0 CI 60 * c
u: a
r. 0 * 0 N.0 a a Sl a-* a 0 0! 00 n 0: 00 n
c-c'r. 0 N0 Q alC' 742cvorc.c
1* 0CID 01 -t w~ N 0 0 44S0 0 4 0 0 00,r 0 0 0 0
3w ey t~o
000 ; -
0y fy 0%f 0 e z
F-9
.O
0C 0 0 0 a ,00 10 0C 0 050
N OC~CO 0-N-awN'
xl * c ,C c000Cr.r 4MM 0-r
* *o0co c cco Coloo..o o~ ..
* II
I*
b
0* co Cl0 O0OGOC. 0CIC 0 -M lIn 1CC WCC NW% 4 OJ
c 0 0 C'C 0 0' 0 c 0 0Co C' C .1+
P-4 N w U' c- It %0'
01
C @0
CG0 0 0 )010 0 0C cOC 010 05; 00Mi.* 00
01 c 0CCCC
oC o cc'zNU . Z P E ccZP-C
0.OC'C C.10nC's 0CIC
I:~~~~
C C c~r 0 I(CrP
M CwCz
Lt* C Cc-c'c Oc ecc~c-c.c IC Nr' l r0 ,0
0 00 00 C CO O C C O 0'
cI C'0CC ' - N it M 0'- 4 'C 0' O M P- it''00 N

W C.
c;
00 CINC
0 C. 0t
000 C 4 0if
P- P-?
0 C,
C C'Coo 0.0 C Cotoo, coC t CO C (70
-'4
AN
C Ncf I0 0 C
C CVI
C' rC C CICi
0000 00 0 10 0 0 e00000000f I- 10000
c~i
00 U, 4. 0 o w'~a. 01 a~ec-i'0 0 0 C c"
000~ ~ ~ ~~~iCOO 0010C 00 c~ IcaCC
C 0 -IV ui
'r:0 t' a,
£' cc
0' 0t C, 0 c C~_docC -
0 t4 N i* o N Q if* 00a N P ~N
0 P 0% O 'C itcit 0 00 0
41 4
3971
C,
C~i 40 C.CC 0
C c cc 0 tP
c c opc'CcooccIC 00 c CCC,0 I-~C,C) ol-cc
.* co~o oc c CoC. mc~- nC ý

A0l a** 0l e . 0 o0 0~ '1 * 0l
* a 0 00 0 0 0
0000) 01i6 )o Cc NMD CC
00 00 000 0 0~ 0 * ' 0 0 0 0 ' 0
!000 0a o C-,c0 C.,-44 oO' 0 c

c C)cKoccooo- CC
C,: C)00C'
0 oo'
** a is 0,0 *-0 0 0 a 0 0 0 via 0 0 0 0 0 0 ; $0 '-
C. CL 00 C.'.c
C*= C- 01 C C-,' - Ný 'fL rrO 0, C
C)00 C)c 0C- )C C -CC 4Ne, L l cC
*. * a0a0 0. 000 00 0 00 0
0*
0co
OCO
0C.,
M.U o4 C
0O00
-
-t r- f.O
C oNM0MO
IC4o
, C
4 c 0000C) 0 OO
000c C cC N m s 7 COOcf-
*y -4el C-aý
*1 D 0 G0 odN4a AC r 7 0(10000
*~
Oc~**S~ ~ ~ 0 00 ~ 00L0 r 000
%Cm*t 404ev, 4 04(e,%t co fcoMI oCum
P- -4aC
%c 11111,0100000gg~o ** 000
0*m N4~00000000
x:~ NJ0 EI'C P4~O ~ C~C0
0* *@~N O~O~N ~ ooo398o

C)
I i)
01
Ii C1*
o*
C- c In0 C) 0C 0C,000" i C 0 C' 0-4 N-4 a? I- N 4- Qr-.4 Ut1M '~
0 C' In C 0 0:0 0 0 0,0 0 0 0 0 0,0 0 P-4(41.'. ý-iC p40 t '0
C,C) cc~n ,00oC, C.0c ooocooo C, oN() C C'a
000
C Co 0 **C C: C C. 0 . , M.0 -4 Lg C' -I f- t N
0. t
0;. 0 '0 a
C CýtrP-;c
CC cjc,-C) 0 0 W.0 't coC, O iI
COC00.0 ~~~c*)C C-0Ic m'0..ro.~
0* CCC 0 00'-N'LCc C C 0 0 01 f o
C 0 c0 00c0 0 0:0C LnNU.',4lý%

U~: r~'0 cco
0 CCC'C 0r- OrUCU' C' rei.0.t q 0 , mo Cc a~ n
c* -4 Im% )f- 04--0-P n% 1 y - (,rC 0,C ,

c -0,0 t~o0%*0I ()I.- u~'N
r-000lo, Ca
C'N
." N4 0ý m r-0 0 r 00w 0%0 00 0 c 00
0 10 .+ P- t-o U ' c a0 '-SI- -
: I' I- f"10%r- 0' U% N NM,#

rt t'C-t Ow- 0 0, 4fa10.
IC Mn -W OAý.. 4 c.0 000a0 0 000
0 0, a0 C,0aCa'
* aC t4C i*
.I y WC a W
0 i I
* U399
STRENGTH DISTRIBUTION - NORMAL
~I|
400
A-
A-2.3.1 Stress Standard Deviation - 0
401
Stress Distribution: Normal Strength Distribution: Normal
Uy Seu (Equivalent )x Depend on the

q Stress) material and the
0 a 0 ax operating conditions
Y
Stress Strength
F(x) Interference
X-Sequ x,ksi*
Figure A-2.3 Interference With Standard Deviation of Stress Equal to Zero
The interference is the area under the standardized normal curve corres-
ponding to the value of the standardized normal variate, z, determined from
Equation (2.1) in Section 2.
Z -xI ISeu - I
V x y
The value of Interference corresponding to the value of z can be ob-

tained from Table 2.1 (in Section 2) of a vs. Ke where:
a - interference
K s
402
Ii
\ A-2.3.2 Stress Standard Deviation # 0
403
TABLE OF INTERFERENCE
Stress Distribution: Normal Strength Distribution: Normal
- S (Equivalent III Depend on the

equ Stress) material and the
a 0 ax operating conditions
Stress Strength
Measure of
f(X yy
Figure A-2.4 Interference with Standard Deviation of Stress Not Equal to Zero
The interference is the area under the standardized normal curve corres-
ponding to the value of standardized normal variate, z, determined from the
Equation (2.1) in Section 2.
The value of interference corresponding to the value of z can be

obtained from Table 2.1 (in Section 2) of a vs. K 0 where:
a - interference
K -z
404
I.
STRENGTH DISTRIBUTION - LARGEST MXTREME VALUE
405
• .. .. .. %,_ 4
.1STESSADR
'-. EITO
4 1
7?
/.
4016
• nl•
ini li i IleI R i mi ln i I-miie I
TABLES OF INTERFERENCE [F(X)]
Stress Distribution: Normal Strength Distribution: Largest Extreme Value
Sequ (Equivalent Stress) M

U
Depend on the material
and the operating conditions
o O 0
Stress
i Strength
Interference
equ xksti+
Figure A-2.5 Interference With Standard Deviation of

Stress Equal co Zero
Density function of Largest Extreme Value is t

L ,
f~x =•"~x') ••"-6-(x-H)
Cumulative distribution function, V(x) M

f(x) dx
Lot y--- ) A dy - +1 *-S(x-M) dS (
F(X a)
- where X *B(Sequ4-
407
X F(X) X F(X)
0.00 .3679 1.15 .0425
0.05 .3495 1.20 .0361
0.10 .3312 1.25 .0305
0.15 .3129 1.30 .6255
0.20
0.25
.2948
.2769
.2593
1.35
1.40
1.45
.0211
.0173
.0141
I
0.30
0.35 .2419 1.50 .0113
0.40 .2249 1.55 .0090
0.45 .2084 1.60 .0071
0.50 .1923 1.65 .0055
0.55 .1767 1.70 .0042
0.60 .1610 1.75 .0032
0.65 .1473 1.80 .0024
0.70 .1335 1.85 .0017
0.75 1204 1.90 .0012

0.80 .1080 1.95 .0009
0.85 .0964 2.00 .0006
0.90 .0855 2.05 .0004
0.95 .0753 2.10 .0003
1.00 .0660 2.15 .0002
1.05 .0574 2.20 .0001
1.10 .0496 2.25 .0000

Where X -
STRESS DISTRIBUTION - NORMAL (a - 0)

STRENGTH DISTRIBUTION - LARGEST EXTREME VALUE
408
IJ
A-2.4.2 Ctress Standard Deviation #0
409
Stress Distribution: Normal Strength Distribution: Largest Extreme

Value
4 Seu (Equivalent M Depend on the

equStress) material and the
a O0 operating conditions
Stress Strength
Sequ x,ksi+
Figure A-2.6 Interference with Standard Deviation of Stress Not

Equal to Zero
410
crca
0 m- ) o
I0
CYatI 0 C4N -41n0
0* S0% 0%n 0
0~ 0*0
-,O*4t~~no00
.*%o-0ý 40 co O%
0 P- P40 0
'o fl.0 0 OD ~4000
t - 000
0 *" 0000 * MI 4 0 go...0
0*0L 0*a o0C

Nl- r--40 -* 40--4.O00
NOsNo 9* 00%0%oo0c
(n 010 0 ý4** .4 0 0 0 0
0* " l 0* 000 t t
0*' %010~ 0 ODUSMOO'000
0 (n 0 00 ý*'rI-4000000
0*
* 4O t- 0* 0N40;0
o* i- . 4O 0* toOL00 a
'.010
NO 0* OU~00
*l M-0@0000,-
UN 0
0F-'.0 0 co inmN0
aoc0 #00f 0 00
0* i 9 ) O*fn00000 0)
ow0*fI O0 ). 0* P'4OOGOO 0
o 0A IO0 * en 0000000 01
8*s~
10 0* fIOdO
P. 'D
*0 !I!!!:
01
0%G*.of-oo
toS
0 M 4- 0 0 00* 0Oi~400
0 *00 a 01
60 00
N)*.tN'.ID0000mP40CO
** fil- 0 0 0 0 0i
0*0
0*
N0* 0e-t
4 N cr
%+ 0 00
4*4 ,tQr-40NjO0O/
N* %+N.,400000,00O0
0 N 4:p0 0 w 0' 0 04 0 00
N* %tN.-4000000000
* 9 *1 0 0 0 0A 0- 0IO a
6* ArjN um 0 0 010 0 0
4N:* 4000001000
* 0 $ 0
o 0s
s 0 0 ' 00
0*
0 *- r-
M I, P- .4:0ooo
0* 0 cp1 N t m40 0!0I00 0
0 ^P* -t..4000 0100 0
* 0 0 0 0 0 0:10 0 90
o*0no'"".00O1000
0* w":0 sO *ON* c~*oa *

Q* H
0* ,
:.* P400400.00,00 H
P*MNP4000000000
*0 ¶~412
< C'Cc
u
er-.- c. ;
ca fCu r r-' Q. c- r- C'. C
0*
0 *1~ C:. U' K,¶ fr'm lf - X'C~U r - \I'N - . 0
I-fl tr C) ?'- P` -41"-. ..T C n LC M\- C: c_ 'C
% Lr,
C
NT P, (P'1en1o U, CIr -4,-._: C'. c C' G 0 C. ZD C) c'
4 4 M N~ - 0C C- 0 0 C) 0 C) 0 0 0. 0 0 C lr0 0
S S el* 0 0 . *
410 t0 *S
0 * n CC,~4w fv4 P- C. Lr' 0' f'.4C 0~I) C-)~

C' rV.
ILr -.I-c -a:c Zt- c ''~\- 0 o o o
* ~'I' -~ C. f Q'-C
0'C)J 0 0.
C' * r'46
*5
("I ..- 4 .-.
U'
'I55 6e g
0
c_ C.-Q
0
-0 4C) C ( , C
C
~ "
c*4uw~~trs61 0 0~ * -
c-C1C ' Cj C, 0 Cý C', C' C.

c n W,
-r fdr' 'c j-c,
IN c' .~
CIS f0~ P4~

a p- LI 4I
c'i t.?
(. al o o v ..
* L 4 4n IV 0 ODI Ar P-1 Q0C C00 C0 Ci C.)
*I- *S, *o oig 0 * :6 ~
0
0 S, 5S
C. _ ý 6 0
0* N w IV, j
N". LIQ ~.
.4 00 C 0 U 'C -7
i& CLr - IN w ci
Q~ 000000Cio'c.
*100 0
0 0 0 . CiC0 c
*I'
0 *Iqc fl
aCC
co r. P-4CC N0to0
C~J 0# a0 00
g~IN
!00 0
0a 000
Cc"0a I 00CIO oic 0 0 0 ..
V0 0lQ 0
0 0 S110t "Ps00 - - - -
0*'13
I
A
MO
0 0
l ON P4N Ln-4 0 4-4 a 0
P4 LI 0 - N a '-P 0,0a,ýtN
0o 0%0%.00-~ 1-o
o 0 lo90 %a
%- 0 ' NNN0d1--
Ln 4-N O 0mO1-
tn m
ca 04000094000 00 00
I *
0n* o m. m m 0,•• N- ý4" -4 "4" 0

.IiN 0 01 0I 01 ,0 o
I0 * ° I'°°'4' N '''''''''°'
0o 4 r- c. 41,tN • ,0 N
w,.Pl0r- m..rl-I,4" fn %t"%00 ~ 41
l 0*
0• !, -t" 00* (qN
01 4N-N,,-
% --3% o - 0%ODMI f" n 0
-
I-*NNN-0 % M0 4 NOO0 000oU00OMNNP
*L 0pIN 0 00 , 0LN t 0 00o0
N ,'w 1. " ,,1"1,00000
0* In In'0 Inr0
wa m 4NIM
- 0 0 w0t-O.
0 .4fn
0 00
U"IcO
"4 0 >lM
m4" 00 I 0 W 0 00N 0 I- a ,t"00 0,-
S-t0 0 •t-I -ION f- % 4 M N 04 - - 0
,0 "I W*
0* N-. 0% - %O,
. ft0 00 ,I 0 - 0•
I..- P4,0r .IS 0 N-i0r.•U 0

st M N N-4,4o00 i
lnt M ,0% o -4-4
, 4 0 0 00o 0, ,0 010 0
Io *tý 0,1. %D,-4P-,t

.0 N
-6 W %0 N PN .4 .:!
0
00
1o 0* o 0 14
-I,. No owo 0o0 In
4,0,i .-,,-,40, 0,,0%!
0,* W,
in *a0 01,. .4o .0' a a 00

*0 i
0N ' 0Nr enN01
' 4'
N' %t
' 0 N'P-
'0 -t f" I" No.44u.i00O~0009O9OO
*4p 1. 4r 4p Ww
414
'~ 0 a 0N0r - 4000000o
ca . N 041-4O0 0 00 000 010 0 0 010 0.10 0
0 000O0Oooo999000j99919900
6M * L9N 0999ln M 9N 9000000i0000

0 * "4400 ~N 0~-0000ojo000000
0** 0000000 000
0!000O 0~o~
0* 9 91904M 0 00 00 0 0 0 0 0 0 0 0
*090*0100000 00000s00
0*
0* 0 1
LA*W-~N' 1-11')000 O0 0 0 0
.*00000 00 000000000000
0000 00 000 00,0000
10 100 0 0
I
S
09 10 . .
. 0 * 0 0
*
0* InN4P4, 00000 -00
0*LA(N000
1000 0000 000 0~00
lw000010000000ee... 0.
I~104N - -tN'I 0
NI 0 000 0 0o0oo~
.*-Ala 0!000 00010 00 a00~~ 000
r0*Oo&0000oO a0aJ 0 1a
0 0 oo 0 I
*e aj 01eg 0: O 000 0 .08
0*
0 OoicolrofioooaOpiO00ooo )
P-000000000000;01
0*0010000
* Oa 000 00 0100091
*0000000oo 10oooDOC
00 ' 01
oj-000000 0 00olocroo
01 1001.000
~~- 0.4fI&lOoOo.J
:0 0 0 0!0
a a 1a INN.NNN9NI(E $@ *1
iii ~i e gj i~g iggg
415 g
I

STRENGTH DISTRIBUTION - SMALLEST EXTREME VALUE
,i~i•. •!416 "

_______________ ______________________________________
A-2.5.1 Stress Standard Deviation - 0
417
Il .l I I II "... .. • a
TABLES OF INTERFERENCE [F(X)]
Stress Distribution: Normal Strength Distribution: Smallest Extreme

Value
u a S (Equivalent M Depend on the
a - 0 operating conditions
Stress
f(x)
Interference
xUSequ x,ksi-
Figure A-2.7 Interference with Standard Deviation of Stress Equal

to Zero
Density Function of Smallest Extreme Value is:
f(x). = e+(x-H), e--e8(X-M)
Cumulative Distribution Function, F(s) - f(x) dx
lot y -. +O(x-•) dy - a +13X'tl dx
le(z , F~.S
$ eSu)- - .(Sequ 41 ) - a7 dy a* .-orY - 1 -
+(eO~( .e(Sequ 4 I)
-X
FP(X)m-1 -e~
where X -B(Sequ--)
"418
X F(X) x F(X)
0.0 .6321 2.8 .0590

0.1 .5954 2.9 .0540
0.2 .5590 3.0 .0459
0.3 .5233 3.1 .0441
0.4 .4884 3.2 .0400
0.5 .4548 3.3. .0362
0.6 .4224 3.4 .0328
0.7 .3914 3.5 .0298
0.8 .3619 3.6 .0270
0.9 .3341 3.7 .0244 4

1.0 .3078 3.8 .0221
41.1 .2831 3.9 .0200
1.2 .2601 4.0 .0182
1.3 .2385 4.2 .0149
1.4 .2185 4.4 .0122
1.5 .2000 4.6 .0100

1.6 .1829 4.8 .0082
1.7 .1669 5.0 .0067
1.8 .1524 5.2 .0035
1.9 .1389 5.4 .-
0045
2.0 .1266 5.6 .0037
2.1 .1153 5.8 .0030
2.2 .1049 6.0 .0025
2.3 .0954 6.2 .0020
2.4 .0868 6.4 .0017
2.5 .0788 6.6 / 0014"
2.6 .0716 6.8 .0011
2.7 .0650
STRtESS DISTRIBUTION4 N ORMAL (*rnO)

STREWC-M DISTRIBUTION . SMALLEST EXTRF2E VALUE
419
-2.5.2
2A Stress Standard Deviation 0
42011
I
Stress Distribution: Normal Strength Distribution: Smallest Extreme

Value
S
S= (Equivalent M Depend on the
o 0 0 operating conditions
Stress Strength
f(x)
Sequ xksi-
Figure A-2.8 Interference with Standard Deviation of Stress Not Equal

to Zero
421
0 0
0,l MOI M'11- 0-t -4- 0 1 '00A-0 0 ý.t%-ý4 0
I f Nn'Lný '4 0,0 0 00 0 cýO- 00asI.Oi0 - -4!0 0 0 0.0 0
.0 :, M-40 000000Oo* ý4
-~'~ 000 00 0O0
0 to 'M N 0 10 0 00 a -- 0 is,
* POf-m~00 -4 it- -4 00 :F-4 0
0EL* -4(q)Ln-4 O0 0 0 000a00000 -4IM-N!000000*0

* *;0o O*90:0
O 0 OseC0 * In N 0!e 0s *0 0,0
0*M -4 1 Nl 4'-4 00o Oio 0 0 0 P- Co fl " 000

1"t 000
-0 0*0M)1-40,0
00-000a 0 )c--cC 0000
*: c= e: e,
00 CCz~ fl e 0n ee.0 10 C
-4 4N-430C0;
*-M 0. -0C: 3! k.0 0 Z) 0 'D e N C 0; 0 0

41. E-4
* eec0*00
0 0 0 Q* -4ge g ee
isN - ' 4 4 0)
N -4 P 10.i N
t-! t % o): ~c
* r- C0 4L; ce N 0 S e" 0 10 0* .4 f 0 0 (4 -1 e., C C

* 4 n, C.0C0 )0m .1 n0 f , $
0 o 0, 6 6;0) 0* 1 0
V. 1-0 r.I **%m ,

4s~~4 f'0*tN-0o0...d '4Nk%4N
ag4e geesoc * e0
q* $'o 0 a a 0 00 0
4 aN* .144Q- 01 ONF 10* 4fCU~ 0 .4 .iO 0
*s e e ee is icS s ; e
* Ii
422.
Nr. 001 04 oL ;AN~~0
o 0 a4
%t ju0itNO64 0 00i00l
N -0000010
LAA4~ 0
~*0 ~ 1- ~-t
0%0'' 0 0!0i 010 0
'N* L4 pq00;0.O0 0 0 010 0
LpAIN vi%4. f'An N1.4 0

7 0
MA
N *n 0 rO o~
0t o x ýAi
-10 0o0
1*00N 0 ý00 't4N 0000010 0
ýN *lLA t4N -10 0:0 0 0 ii010 otO
10*
0 41 7 p 4dý IN 0% ~-4-. 11
CON~L
N- L01 M-4 n0 0010 010 0
40 *n ' t N 0~ 0 1N -0
0 0' 0C)
f0 0.
00 tie 00 1000
'N *~- -~-410 C00 0Oi0
0 'LA rl- N lonýONN.o 1-1-- -400

.*0 4 ^-4111 ., le u No0I 0ýO00
A, (7 LA.4C%,+ .I-t)i 0 0 0011)0 1 00
Ln N --4! . 0! a * *:)i * 0
N ~~ 4-4N 0D
* 0% ' 'AN
-00N10O0 0 00
^4 01 @9
*o~ o9 oI 0 o
:1 *i 0
0: 0 **1
* 0
0 N Mllil'%
. N N t1OlN -!00 4
-4 - 0 1 n I 1
t o o ai .0 *1 6 9
a )IP CIO 011
423
0 3%Iina t0r-J N - ý ý
0 .0 0 Our
9 -N
0 0: al 0ýN-40O0 00
*- (fN U,'40 0 in00 0 00 C
0n N10% 4,-0
0 if aI )F ' oL N-100 -4 ,0 0io0
.*LjvrN-0.-0000000I00
UIN 00 0 00 0
00 00*-
0 *t N O -4 0 0 0 ý 0) 0l
0 0
* * Je* '~so* 96 :4 0 0
go's 000 *00
if3 4'J *N -000000 N 4r Nt 0 00000 D000
*if 10 t 9- 90 94
1 Q NM Sc 0. ; 04 C) 9- r 0. 0l, C-- ;0, = 0
0 4 N~ r fn 0 '47% --i
I'D 000 c
~
S 0-D:0 ~ ~
9 - , S -49 0
-4 5 5 9
0~~
ONj% ~ 0'N -y0"00 *0~00 020
In 0 0 I' 9) 0.1 :**n
if
* 09 0 0 a0 0 0.00
0 0 0 :0 0 0 0' 0
~0 9 0 09 0 9 0 S 9, 09 0 s
if ~ N 5 Z) c9: 9 r-9 0 9) D 00 0, C
if f* I- V '~~~ N 0 00c
c 0 c C) 0 " U
* lp900 "1 00 9ý 9n 00 0l 0 9 01 0 0 9
0*9
I0 9:
MOP N-0C)I 0 010 044,'
r- I0(1 C a0 ýO a0 00=42400
if~~ if-* 4 -f
if ' k
8~"~J 'r C T T0 NM- Dr%,m0 4r
0L M M Ntr 4U 00ý 0f - 0
%0 0 4 0m0m0r- AN00P
4j8
- 0* m 4P oG Ac %Q 0% Xr
*N~O o I o a 9 # 0 0 0 0
0*
*0 *0t 0000 M0-( 0NN

*oLn. tc o' o.
0* - 0n4I
* nP- -- 0 1 0 0-NNý
0
~0 0*n
0%* L wN P- 4.
m~ Q0 N. %0N A w4 P-t
0~ ~ A ~w
* rP-8Y4clQ
z r0A4~'44pNQr0
4n% 40 0 PI-4
* M N0000o
N 4 4 4 * 00 00
LA*I-U '+'00~o
w~.V Nmw L- 0 -4Ut*¾-og.t
m
F* % . NN Qt 4 0~4000rtN 440 0O
z
0I 9
0 0 0 00 to 0 o0 c oQN(%P4-4D IO 0ma
A4ýr %0N .4.4... 0oo1oo1
0 NO-P * mA 0 r* 0 %t m N P-4 P-4 0
0 144
ý00 .000 0 0 0 0 0 00
ýo
P4*
.4 P-104P 4P- - wP4P
0
0*2
0
S0001000000 00~00000000o00
t- w*0 Ne 0 r- SeeS..
M Nee4 00 00 ee
I * Ni.--4P4e0000000010000 00j000
0001400-0000000000000000
0*
Q*-O-00000000000000000
4*r000000 00000 000 000 0000
0*
.*400000,000000000000000
Z*OOOO00.0OO00oooooooo00
O*WMNNiO000000000000 0 I*
0000004100 0000000000000
*~~~~~~~~ 0 0 05 ,
4*j0N0000000000000O0000OO
*00000iO000 0 ol00000
4*000 00000000000000d0o
O*.OOO0001000000000000 00
0*. %005 SOSOSS 05
5 5 000
0 * co S 0
ON*.wO"O0 '0
01 00 0 0 0 0On.
000
000 800 000 00 0p000oz
* 00 ee...
0000 0000 e
000000000 8 0 008
j*,**** .0.0 9I99900Q000c000
In OO CY0.4 OOOQ114O
0
NNNNNNNN mmmmtl.mefmm
426
APPENDIX 3 EVALUATION OF THE INTEGRALS BY NUMERICAL ANALYSIS A
427
APPENDIX 3 EVALUATION OF THE INTEGRALS BY NUMERICAL ANALYSIS
A-3.1 THE PROBLEM

In this report stress/strength interference theory was used to pre-
dict failure probabilities for several important cases not included in [1].
The two cases of interest here are (1) the smallest extreme value distributed
strength and normally distributed stress and (2) the largest extreme value
distributed strength and normally distributed stress. The integral
Pr[Failure] - Pr(X-Y) - F(y)g(y)dy (i)
was evaluated where F(.) is the distribution function of X (the random varia-
ble associated with the strength) and g(.) is the density function of Y (the
random variable associated with the stress). That is, we suppose that Y has
a probability density function:
1 1 12"0 2
y(y) e dy, "-y (2)
and X has a probability distribution function
Fx(X) - l-ee ,x--< x - +C4 (3)
if it is distributed as the smallest extreme value. If X is distributed as

the largest extreme value, its probability distribution function is:
-e(x-M)
One sees that if z.- 1-Fx(x) is plotted on a lnln scale it will be
linearly related to x on an arithmetic scale as
Z - B(x-M) or z - -B(x-.M)
Hence one can call H the mode of the distribution and B (or -0) the "slope" or
"shape" parameter.
Inserting (2) and (3) or (4) into (1) will give (5) or (6).
Pr(X&Y) " 1 "e Hi1 1 0 dy (5)
428
2
Pr(XWY) -l e VF-e dy(6
A-3.2 SIMPLIFICATION OF THE EQUATIONS
Equations (5) and (6) are quite difficult to work with numerically
since each one involves four parameters. It is desirable to define a new
variable of integration and new parameters. To this end we let:
u a Y- I a - $a ,and Y- (0-M)
Then (5) of Section A-3.1 becomes
Pr(X-Y) -Ia e L / e -e e T(
2"
and this further reduces to "2
Pr(XSY) - 1- e e du (7)
Also, (6) of Section A-3.1 becomes
12
-e-(x~) T
Pr(XIY) ee du (8)
Inspection of (7) and (8) will show that the integral to be calcu-
lated is of the same farm in each case except that the uppermost exponent is
negative in the Largest Extreme Value case. Therefore, to further isolate the
numerical analysis problem involved, define new constants A and G such that
for (7), the Normal-Smallest Extreme Value integral,
A a a and G a y
but for (8), the Normal-Largest Extreme Value integral,
A - -a and G a -y
Then, the calculation of the probabilities given by Equations (5) and

(6) of Section A-3.1 reduces essentially to the integratioft of
Se"
X0 - 2
a2
Au4G;
du (9)
429
.,'
I Since this integral does not exist in closed form, numerical techniques must
be employed. The ranges of a and y of practical interest and the associated
values of A and G are
A - &awhere a ranges: .001,.005,.010,.025,.05,.075,.l,.2,... 3.,
3.5,... 9.5,10
G - Ay where y ranges: O.,-l.,.... -40
Note that a itself is always positive and y is always negative (or

zero). A negative a would imply a negative standard deviation. This situa-
tion is without mathematical meaning. If y were positive, the material in
question would have a mean strength less than the mode of the stress. Such a
situation is of dubious engineering importance. (8 is always positive.)
Note, also, that for the Normal-Smallest Extreme Value case, the pro-
bability of failure is one minus the number given by (9).
A-3.3 NUMERICAL METHODS CONSIDERED FOR THE PROBLEM
The integral (9) of the previous section could be evaluated in sever-

al ways. Its form immediately suggests Gaussian-Hermite integration. This
method can be summarized as follows:
Au+G
Let F(u) - -e and define v -u/12
then 1 2 F(u)dum e-v F(v)dv
and by the Gaussian-Hermite formulation,
- v2• (v)dv" 7,n

7 AiF(vi)
where the arguments vi are the zeros of the nth Hermits polynomial
t_
n V2g -v2
()a d,"R
and the coefficients AI are
The numbers v and A are available in tables. The approximation error using .
this method dipends in the higher derivatives of F. For the n-point case,
430
_____________________ I..----- N -- --.

S1H'
error = nF
2 (2n)
< < 4-
II
(The formula for computing the higher derivatives of F is given in Table A-3.1)
Errors in integrating (9) of Section A-3.2 were computed for the n -

2, 4, and 8 point formulas for the entire range of the parameters. It was
found that the n - 8 formulas would give acceptable errors for small W and
large kl. Theoretically at least, this scheme could be used to integrate (9)
for any A and G if the number of points taken were sufficiently large.
Despite its advantages, the Gaussian-Hermite integration methoi. was

not used to prepare the tables of this report for the following reasons:
i) The higher derivatives of F become quite difficult to compute

and hence the errors involved are difficult to evaluate.
ii) It proved useful for most values of the parameters for which
the approach presented below proved even simpler.
iti) It was decided that using only one method (that below) was more
efficient than programming different methods for different
values of A and G.
If users of the tables wish to compute probabilities of failure for 4's and
Y's not included in this report, the Gaussian-Hermite method may prove accept-
able, but the method employed here is recommended.
The tables of probabilities included in this report were generated by

a FORTRAN program run on the University of Michigan's IB4 360/67. For every a.
and y, the program:
i) Determined limits of integration following the logic of

Section A-3..5, *
ti) Used Slapson's rule to integrate (9) within these limits, and
iii) Added to the above value the Integral of the normal curve
neglected by using the calculated limits.
This approach is Illustrated in Figure A-3.1, Its major advantages, which

will be made clearer in later sections, can be listed as
i) Automatically preventing computing difficulties by 110iting-

the exponentials to be computed to acceptable values,
it) Having an easily ~omputed error,
iil) Reducing computing time by using che error function subroutine

wherever possible and by automatically recognizing where the
integral is approximately zero, and
431
Au+G
DERIVATIVES OF e
Au+G A+
F =e-e P -eAu+G
PAn
F(1) 1AF [pl]

F(2) 2 1 2
F) AF (P + P
F(3) 3 [P " + 3 P2 + p3]
A3F
F(4) AF[p+ p + p3 + p4]
F AF (P1+P+6 P
p2 CkPk+ CnnPn]
F(n) AnF [ClnPl + C2nP + ... C ..
If we let Rn+)nT= 0, then the kth coefficient is given by
Ckn =k'Ck(n-1) + C(k-1)(n-1)
-u-eAu+G
12
FOURTH DERIVATIVE OF e
1 -e2
j Au+G -e Au+G
"
U=e F=e
uV 12 3-6u2+u4)F + (12u-4u 3 )F(I) + (-6+6u 2 )F( 2 ) + (-4u)F 3 )+ F(4
UI= U i(3-6u2+u4 +
(-eAU )[A(12u_4u + A2(-6+6u + A3 (-4u) + A

(~2AuG(2
(+e2(Au+G)) M 2 + 3A 3A4
A (-6+6u2) (-4u) + 7A4]
(-e3(Au+G) M A3(-4u) + 6A 4
(+e4(Au+G))M A4]}
Table A-3.1 Expressions for UIV and F(n)
432
im" ~
Normal-Smallest Extreme Value
Curve Approximately Normal Rein ntrad

Using Simpson's Ritle
Curve gives approximate.

ly zero contribution to
point
--
+wu Axis
-5Us 0 uS2 +5
PR(Faiiure) l.-(total area under curve)
INormal-Laryest Extreme Val-ue-
Itzero Curve gives approximatelv

contribution to inite-
gral below this point.
Regior~ Integrated
Using Simpson's Rule
Curve Approximately
f (u) Normal
-5 ul0 UL2 +5
PR(Failure) -(Total area under curve)
Figure A-3.1 Limits of Integration
433
iv) Being applicable to any range of A and G simply by changing the
Simpson's integration step-size.
A-3.4 SIMPLIFYING APPROXIMATIONS
The functions comprising the integrand of (9) of Section A-3.2 are

quite well behaved and bounded between 0 and 1. Their exponential form leads
to a number of important approximations. Let U equal the integrand of (9) and
note that U is the product of the normal curve and a "double exponential."
x 10-6
The integral of the normal curve is equal to 1 within .5734
over the interval -5 to +5. Therefore, since
Au+G - 112 Au+G 1 2
0OS e-e S 1 implies e 2 ee • e ),
X Udu Q Udu to at least 6 places. (10)

•- Au+G
Let F - ee
If the function F differs from one over some interval [ul, u2] by
less than 5 x 10-5, then the integral (10) can be evaluated to four places
simply by integrating the normal curve (actually the FORTRAN error function
subroutine was used). That is,
112
e )(l-.O00005)du 1W e du where the
12
worst error - .0005 x e du - .0005
5 x 10-5
The combination of u, A, and C which causes F to be within
of one is easily determined:
Au+G
-e .99995
aAuo G 5x 10" 5
for
Take aAu+G 1 X 10-5
Then rA;W < -11.5 (11)
If tk, function F differs from zero over the interval [-5,+51 by less
than 5 x 10"5, then the Integral can be approximated as zero to four places.
The above argument concerning the error applies here also; that is, the
434
__ _ _ _ _ _ __ _ _ _ I'
SI
integral differs from zero by less than 5 x 10-5 when F differs from zero by
less than 5 x 10-5. It will prove useful to know the value of the upper expo-
nent for which this occurs.
-e Au+G
_e < .00005
Au+G •
Take e-e < I x 10-5
eAu -11.5
Au+G > ln 11.5
AuG .44 (12)
A-3.5 LIMITS OF INTEGRATION

Equation (10) defined the limits of integration for our problem as -5
to +5. The discussion below -shows that it is possible to further isolate the
interval on the u-axis in which numerical integration of (9) is necessary.
A-3.5.1 Limits for Normal-Smallest Extreme Value Case

For a given A and Gexpression (11) defines a u-value, call it uhl,
below which the integrand can be approximated as a normal curve.
-G-11.5
1 A
Likewise (12) defines a u, say u8 2 , above which the contribution to the inte-
gral is zero.
u82 -G+2.44
A
Therefore, numerical integration need only be used on [us 1 , u.2 ]. This is il-
lustrated in Figure A-3.1. In some instances ua1 and u5 may be outside the
-5 to +5 limits established by (10). The limits should ehen be taken as -5 to
+5. Note that:
When us, > +5 the integrand can be approximated to sufficient

accuracy by considering it to be the normal density function.
When u82 < -5 the Integrand can be approximated by considering
it equal to zero.
A-3.5.2 Limits for Normal-Larmest Extreme Value Case
The results of the previous article apply here directly except that
for this case A * -a (recall that a is always positive) and the senses of both
435
(11) and (12) are reversed. This makes
-G+2.44
ULl A
and for u < uLl the integral is "zero".
-G-11.5
also, UL2 = A
and for u > uL 2 the curve is "normal".
When uLl > +5 the entire curve is "zero".
When uL2 < -5 the entire curve is "normal".
Of course, in those situations where uL < -5, the lower limit should be taken
as -5 and where uL2 > +5, the upper limit becomes +5.
A-3.6 DETERMINATION OF SIMPSON'S RULE STEP-SIZE
The truncation error for Simpson's integration scheme is given by
4
lErrorl - ( UIV( )(Au) , uI • u2 • (13)
The fourth derivative in this expression should be evaluated at the • which

makes UIV a maximum. The complexity of UIV (shown in Table A-3.1) made a com-
plete analysis of its behavior impossible. Study of the form of UIV together
with careful scrutiny of a large number of computed values gave what was be-
lieved to be good estimates for the maximum of UIV for all combinations of A
and G. These values were substituted into (13) and Au's were determined
assuming a desired error of less than 5 x 10-5.
Actually, only three Au's were used: .1, .05, and .01. Although
the decision of which to use for a given set of the parameters was based upon
inspection of calculated values of UIV, an understanding of the behavior of
UIVmax with changes in A and G Is important. The double exponential coeffi-
cient of UIV will damp out any influence by the other exponentials to make UIV
large. For A-ýO, UIV will be small. In fact, for A small, UIV should behave
similarly to the fourth derivative of the normal curve since G only acts as a
"scaling factor" and does not effect the "curvature" of U. Tables show the
fourth derivative of the normal curve to be approximately 1.
As AO, UImax 1 and
15 x l0-51 = f+5-(-5)] MlOO4

180
(au)4 9 x 10-4
Au W ..17
436
Secdnd Quadrant:
Normal-Largest Extreme Value G
1+40
" 0o -o +30
3.3xlO-l 2.6xlO-1 0 ( 15
A
3.9x10• 2.OxlO 1.0 1.8xll 6 . +5
+.7 +2.0 +7.0 +10.0

l.lxlO 2.OxlO 2.2x10 1 - 12 2. 2.OxlO' 4.3x1O0.4x10
10. -7. -2., -. 7

-.
5 .3.0 3.0 9.3xl02 2.6x10
-15 3.0xi0- 3.3xi0- 1 1.5 2.7xi01

-3Au2010 .41' 1 :oxo
-30 0.- I0-
Simpson's
Step-Size I Au-.05
-40
Au-.01 Fourth Quadrant:

Normal-Smallest Extreme Value
Note:UIV is only given for the interval where

UiAX
Stmpson's integration was used.
Figure A-3.2 U;Lx

IV for Various A and G and Simpson's Integration Stap-Size
437
.p
d
For large JAJ's (neglect G) the UIVmax will behave as A4 . Then for the lar-
gest A of interest, A - 10.0:
4 9 x 10-4 -8
(Au) 104 -9x9 10
Au ; .0017
It can also be seen from the form of U that for large IGI's the value of UIV
will be driven to zero by one of the exponentials. (Note also that the beha-
vior of UIV is only of interest on that interval where Simpson's integration
is to be used.)
Recall that for each of the two cases being studied and for each
pair of the parameters a and y, the integration of (5) or (6) of A-3.1 requires
use of Simpson's integration on an interval identified by Section A-3.5. The
error produced by the use of Simpson's rule depends on the derivative UIV and
the fourth power of the step-size. Figure A-3,2 gives the value of UIVmax at
various points in the A-G plane and shows the regions in the plane where the
three step-sizes (.01, .05, and .1) could be used. In the A-G plane, the
second quadrant is the region of interest for the Normal-Largest Extreme Value
case (A - -a, .001 X a • 10.; G - -y, -40. < y S 0). The fourth quadrant is
the A-G region of interest in the Normal-Smallest Extreme Value case. In
Figure A-3.2, choose an abscissa and an ordinate; the number indicated is the
maximum of UIV found by calculating U for many u's on the interval in which
Simpson's rule was used. The cross-hatching defines the regions where the
indicated step-size will maintain the error less than 5 x 10-5. (These re-
gions were delineated more carefully for the actual computer runs.)
Example:
Assume that we wish to find the probability of failure for the Nornal
Largest Extreme Value case for a - .7 and y - -5. For this case, the para-
meters A and G in the integral (9) of A-3.2 become A - -a--%7 and G -
+5. Figure A-3.2 shows that for A - -. 7 and G - +5, UIVmax I 2.0 x 102. The
cross-hatching shows that a step-size of Au - .01 is acceptable for the inte-
gration.
A-3.7 ERROR ANALYSIS
The possible errors produced in generating the tabulated values are

listed below:
Error Source Maximum Error
Truncation of integral on [-5,+51 -6 x 10-7
Approximations of Section A-3.5 -1 x 10-5
Simpson's integration truncation *5 x 10-5
Computer round-off *1 x 10
Round-off to 4 places *1 x 10-4
The tabular values are correct to 2 x 10-
438
4-
1i
A-3.8 SUMM~ARY
It was shown that calculation of the probabilities of failure re-

duces to the numerical evaluation of a single integral. The integral was com-
puted by using Simpson's rule and by employing several approximations.
439
APPENDI X 4 FLOW CHART OF THE COMPUTER PROGRAM USED FOR
DETERMINATION OF THE FATIGUE STRENGTH
DISTRIBUTION
440
~~oA
I
APPENDIX 4 FLOW CHART OF THE COMPUTER PROGRAM USED FOR THE DETERMINATION OF
THE FATIGUE STRENGTH DISTRIBUTION
A-4.1 EXPLANATION OF THE SYMBOLS USED IN THE FLOW CHART
C1 , C2, C3 , ... , Ci, ... C• are the lives at which the distributions of
fatigue strength are to be determined. In the oresent study, these
are 103, 104, 5x10 4 . 105, 5x10 5 , 106, 5x10 6 , 10', and 10 cycles.
m is the total number of lives which in this study is equal to 9.
U is the code number for a mode of stress, for example;

U = 0 is for the alternating (completely reversed) stress,
U = 1 is for the maximum stress,
U = 2 is for the range of stress (maximum-minimum).
US is the identification number for the units of stress, for example;

US- 1 is for the units in psi,
US - 2 is for the units in ksi,
US - 3 is for the units in short tons/in2 ,
US - 4 is for the units in kilograms/mm2 ,!
US - 5 is for the units in long tons/in2 .
UL is for the identification number for the units of life, for example-,
UL - 1 is for the units in cycles,
UL - 2 is for the units in reversals (two reversals - one cycle),
UL - 3 is for the units in hours where frequency (cycles/hr) was
known.
Sm is the mean stress which in the present study was always equal to
zero,
Sl, S2, S3 , ... St, Sm are the values of the applied stresses, and,
L1 , L2 , L3, ... Li, ... Lm are the values of the life to failure corres-
ponding to the above stresses.
i is the subscript to designate the ith data point (for example, Li

and Si are the coordinates of the ith data point of the S-N type
data).
n is the number of data points.
MRn,," ,,,Mn, ... , MR..n are the median rank values for the sample
size of n(for example, MR5,, MR 5, 2 , MR
5 , 3 , KR
5 , 4 , and MR
5 , 5 are the
median rank values when n - 5). "
Zn,l, Zn2 Zn, 3t..-, Znn are the standardized normal variates corres-
ponding to the median ranks of sample size n. (These are also called
the median ranks z values as given in Table 6.1 on Page 49).
441
I
YWi is the linear mode variable of the ordinate scale for Weibull dis-
tribution for the ith data point, and it is also the same for
Smallest Extreme Value distribution (See Equations 6.24 and 6.25
on Pages 40 and 45).
YEj is the linear mode variable of the ordinate scale fcr Largest Ex-
treme Value distribution for the ith data point.
YLi is the linear mode vriable of the ordinate scale for Logistic dis-
tribution for the it" data point.
YNi is the linear mode variable of the ordinate scale for Normal distri-
bution for the ith data point.
a is the intercept of the least squares line on the Y-axis. (See

Equation 6.3 in Section 6.1.1 on Page 28).
dN, aL, o1, W, and aW are the intercepts of the least squares lines
fitted to Normal, Logistic, Largest Extreme Value, Smallest Extreme
Value and Weibull distributions respectively.
0 is the slope of the least squares line. (See Equation 6.2 in Secton
6.1.1 on Page 28.)
ON, OL, OM, OE, and OW are the slopes of the least squares line fitted
to Normal, Logistic, Largest Extreme Value, Smallest Extreme Value,
and Weibull distribution respectively.
R is the correlation coefficient. (see Equation 6.6 in Section 6.1.3

on Page 29)
RN, RL, RM, RE, and RN are the correlation coefficients of Normal, Lo-
gistic, Largest Extreme Value, Smallest Extreme Value, and Weibull
distributions respectively.
Ir is an identification number for each data set.
LCj is the logarithm of tife j at which the fatigue strength distribu-

tion'is desired.
LSTji is the lc.garithm of fatigue strength at a given life j for a

given data point i.
SN is the scatter of fatigue strength at a given life.
DEV is the standard deviation.
DX and DR are the differences between the two consequtive assumed

values of Xo for the Weibull case.
ADZ is the absolute value of DX.
442
DR is the difference between the two consecutive values of correla-
tion coefficients corresponding to the two assumed values of X
for the Weibull case.
A-4.2 EXPLANATION OF THE PLOW CHART.
The flow chart of the computer program used to determine the fa-
tigue strength distribution was made up of three parts. In the
first part of this program, the scatter in the life was converted
into the scatter in the fatigue strength, and the y-ordinate of
non-linear mode variables into the linear mode variab es for each
distribution (For example, P was converted to in in for the
Weibull case). In the second part, the distribution - functions
(Weibull, Normal, Logistic, Largest Extreme Value and Smallest
Extreme Value) were fitted to these strengths data simultaneously,
and the one with the highest degree of fit was the best fitting
distribution. The part three is an external Function Fit which
was used in part one and part two to fit the least squares lines
to the data. As shown in Section A-4.3, blocks were made around
different sections of each of these three parts. The specific
functions of each of these blocks in the total program are discus-
sed in the following sections.
A-4.2.1 Part One - Conversion of Life Data to Strength Data and

Non-Linear Mode Variables to Linear Wode Variables.
Block 1: The function of this block is to store all the

necessary information and the various symbols used in
this computer program. The program will be rejected in
either case when number of data points (N) is larger than
40 or number of lives (m), at which the strength distrib-
ution is determined, is larger than 11. (See the flow
chart in Section A-4.3).
Block 2: For some data, where the units of stress (US) V
values are in tons/in 2 , killogram/in 2 etc., and the modeas

of stress (U) in terms of range of stress, etc., these
are converted to the units in psi and the sode of stress
in completely reversed stress. This is done for each
data point in a given data set.
Sk: Non-linear mode of y co-ordinate of each data

point (in this case the median rank value () assigned
to each data point) is converted into the linear mode
of y co-ordinate (e.g. In In . for the Weibull case).
This is done for all the four distribution functions.
Blok 41: In some data if the units of life are in terms

of, say, hours or nuaber of reversals, these are convert-
ad into cycles. This is done for each data point. The
443
values of stress and life are converted into the loga-
rithm of stress (log S) and logarithm of life (log I).
Block 5: Using the External Function Fit (as discussed

in Section A-4.2.3 and the flow chart showm in A-4.3.3)
1 least squares line (Y - a + qX) is fitted through these
data points log (S) and log (N) found in Block 4,. and
the values of a and ý is determined in terms of log a
and log 3.
Block 6: Using the values of a and 8 as found by Block

5, and the procedure as discussed in Section 6.2, page
31, the scatter in strength in terms of logarithm of
strength (LSTj,i) at a given life j is determined.
A-4.2.2 Part Two - Determination of the Best Fitting Distribution.
Block ?: The values of logarithm of strengths [log (s)]

at a given life (as found in Block 6) are arranged in an
increasing order, and they are then converted from log
(S) to S.
Block 8: Using the strength values as found frcm Block

7 and the information from Block 3, the distributions
such as Normal, Logistic, Largest Extreme Value and Small-
est Extreme Value are fitted (as discussed in Section
6.5) by aid of the External Function Fit discussed in
Section A-4.2.3. The values of the correlation coef-
ficients, R, for each of these distribution are then
computed.
Block 9: As discussed in Section 6.5.1 (on Weibull

di3trlbution,) it is necessary to determine the value
of 70, the lower bound of strength, before fitting the
Weibull distribution to the data. In this block, the
initial value of : is assumed in the nanner shown.
Block 10: Other values of Xo subsequent to the one as-

sume6 in Block 9 are assumed, and the magnitude of the
new values of Yo depends on the magnitude of the pre-
vious ones, as shown in this block.
Block 11: Once a value of 7.o is assumed, the values of

1l n(xi - o - in (SNi - -) are found. Using the
value of - and YWi (as found from Block 3). The Weibull
distribution is fitted to these data by External Func-
tion Fit and the corresponding values of the parameters
a and 8 of the least squares line (Y - a + OX) and the
correlatidn coefficient, R, are computed. This is
444
repeated twelve times and twelve values of 7- and their
corresponding values of a, $, and R are computed. (Note:
The number of repetitions, twelve, was chosen arbitrarily).
Block 12: Out of these twelve values of the correlation

coefficients which has the maximurm value (the best fit)
is selected, and Its corresponding value of is taken
io
as the "true" value cf the lower bound of strength _7.
Block 13: The fatigue strength distributions (such as,

Weibull, Normal, Logistic, Largest Extreme Value and
Smallest Extreme Value), the values of their paraireters
(in units of KSi where appropriate) and their correspond-
ing values of correlation coefficients are printed out.
From this information computer also prints out the name
of the best fitting distribution.
A-4.2.3 Part Three - External Function Fit
Block 14: External Function Fit is used in Part-one and

i Part-two
(Y - a+ ofrBX)theto flow
a setchart to fitpoints
of data a least
wheresquares line
the values ,
of a and 0 are computed using Equations (6.2) and (6.3)

given on page 28. The corresponding value of the cor-
relation coefficient is then calculated using equations
(6.6), (6.7), and (6.8) as given on page 29.
445
'.•
A - 4.3 THE FLOW CHART
4 6
'k-4. 3. 1 Part One
ad data, C,... US, UL, SM, L
Cm v Ln, U
START U 0 s,... S . n, ID,
n '%,l-" MRnn' Znj... Znn
F-
PrinL "m or n is too large' n > 40
M > 11
Set the dimensions of Matrices BLOCK I

mmý
IINS
FIT. (L, S, a, B, 111,11,
Ar B470000'CK7
Print Results
1 S S - Sm 0, 0, nID
IF i
S S /2
U 2
>n
US
US 2 S 1000 Si
F-
>m
US 3 2000 nj+j
1 +1
Fr
US 14 LC logio Ci
S 1'1!122 j
4,
LIIIIIIIýNcus E51
S 2240 S EST Ji i +
Li 10910 it BLOCK 4 YW i - In In 1
It L YE - In In I
- L wh. L
L 0.5 T. or UL 2
IJ YL In
III, M I
BLOCK 3
F-I ý07 YN
Zni
447
A-4.3.2 Part Two
- - -Nl -h-1
- -
i~n 448
u-~ -S ---..
MP-----...- - --
pni TMP
LS I.
- D-R - 6 ]R
I t
e0 +xok
DO DE"-t+X k
I
Q
ýdft
dý400
60
ýIXNW I
42MMM
ýIFý
_______________________
k- m - -L
Ir Iýre449i
A-4.3.3 External Function Rif-
IE. Ic~SY~S
I
I nSSX)-(Sx)2 I
4R-0
sxys~x+_
_ _ _ _ __S~-k
RETURN
UNCIASSIMD
Security Classlflcstton
DOCUMENT CONTROL DATA - R &D
(Security clasificetaion of title, body of ebstract end Indexing annotation must be entered when the overae Ieport Ii eclaelfled)
1. ORIGINATING ACTIVITY (Corporate author) a10.REPORT SECURITY CLASSIFICATION
University of Michigan wUCL
Ann Arbor o4 ab. GROUP
S48
N/A
B. REPORT TITLE
Relsabil~ty Prediction-Mecbanical Stress/Strength interference (Nonferrous)
2
4. D ESCRIPTIVE NOTES ( ,rpe of report and inclusive dotle)
ina~l Repot - Marmch 1967 - August 1968

5. AUTHORISR ( middle Initial, last nime)
amsreflee,
Charles Lipson Ralph C. Disney

Narendra J. Sheth Mehmet Altun
S. REPORT DATE 7.. TOTAL NO. OF PAGES 17b. NO. oF REVS
February 1969 I45O 16

Se CONTRACT OR GRANT NO. Se. ORIGINATOR$ REPORT NUMUER(S)
F3o602-67-c-=24
b. PROJECT NO.
5519
-.Tas5k No. 96. OTHER REPORT HOM)t (Any adber .Imbee OetoWa be ;Weelae
"epolt) -68-4.03
551902 •..' ~ Ohie
~~~RAD)C-TB' "'' •
10.I:DSTRIMUTION SYTAT•EMNT
Ttiis document is subject to special export controls and each transmittal to foreign
governments, foreign nationals, or representatives thereto may be made only with prior
approval of RADC (D(ER), GA"D, NY 13440. _,_.,_O ____* ___,_,_______T,_V,_
II. SUPPLEMENTARY NOTES 1. SPONSORING MILITARY ACTIVITY
Rome Air Development Center (ERM)

ThDonald W. Fulton - Project Ingineer Gritfftis AFB NY 13440
1.AOSTOIACT
-This study we addressed to the developmnt of a prsaticel engneering tool d *A W

I
stress/strength interference theory to be used for desuiing and predicting the mll-
ability of nonferrous mterila parts subjected to fatilg lo.ein.
Fatige data we gathered for 111 nonferromus =tsterielso ie we acaverted to
strength data to obtain the scatter of strength at a given l1fe, X ca~uter proprm
was used to determine the distributiou of' the daa\( eiu 0. FNoMI, LOBO ~sto
Lf end Ov).
The required stress distribution Is obtained by converting the stress speewtrm,
known or asumd, to a zero sen stress s•etlia using the Oomban diapm. Me
required stress spectrm, e.pressed in term of equivalent stress is ebtaind using
ineIr's z.la.
In this struo the interfroen of tw distributions Is eaorossd as &a intealm
*th we eval• ted using an IN Coauter 360. Tabua valuos are given fa 3NIml/-
Weibull, ft/ol/LUrgest zxtr Vale
mad
AMal/•lL t zu v ae d"smtt s
Un order to illustra1te the OPlieSUtn of th interffete'e taetlaqis tiM eO les
an presented and solved.
o0.......--
DI6,06.1473
S. -,
.... ft'lW MGM=,.
uly tete
-w
-NOW.•
./-
UNCIASSIFIED
Security Classification
I4. LINK A LINK'S LINK C
ROLE WT ROLE WT ROLE WT
Reliability Distribution
Mechanieal Components
Stress/Strength Interference
Fatigue
II
AP•IR -rrll u. A~i~'W .Jea(?~,

Ad 0856021

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ad 0856021

Uploaded by

Copyright:

Available Formats

UNCLASSIFIED

NEW LIMITATION CHANGE

USAF ltr, 27 Aug 1973

THIS PAGE IS UNCLASSIFIED

Final Technical Report

RELIABILITY PREDICTION - MECHANICAL

This document is subject to special export

Rome Air Development Center

Do not etewn this copy. Retain or destroy.ý

RELIABILITY PREDICTION - MECHANICAL

This document is subject to special export

This report is the result of a study performed by the University of

The following contributed to the various phases of the project: Messrs.

The distribution of this document is limited under the U. S. Mutual a

This report has been reviewed and is approved.

Approve* e;I .era ~LL

Chief, A4 onced Studies Group

This study addressed itself to the development of a Stress/Strength

In our past investigation ("Reliability Prediction-Mechanical

Early practices in stress-strength relationship dealt almost entire-

In the present investigation, an approach was used which attempted

The prevailing practice is to use the mean values of the calculated

In our past investigation(1) on ferrous materials, the problem of

In the present investigation on the non-ferrous materials, a com-

Mechanical Properties Data Center in Traverse City, Michigan, was

In order to present these data in a pictorialand thus more easily

The problem of stress distrilltion (Section 7)was investigated in

In the present investigation, the required stress distribution was

In view of the serious limitations of the Monte-Carlo ;.echniquea

In our past investigation on ferrous materials(l), these integrals

Stress Distribution Strenath Distribution

In the present investigation on non-ferrous materials, data analysis

Stress Distribution Streuxth Distribution

These interference values (tabulated in Appendix 2) were found to be

In order to *how the application of The Interference Theory Tech-

Section Page No.

3. Objectives of the Present Study 11

5. Analytical Expressions for Interference 17 *

6. Statistical Distribution of Fatigue Strength by Computer Approach 26

7. The Statistical Distribution of Stress 118

8. The Interference of a Stress Distribution with a Strength

9. Application of Interference Theory to Design Problems 133

Conclusions and Recommendations 147

Appendix 1 Tables of the Distribution Parameters of Fatigue

•!• .' ].•. I, II~A_ .. . _____IL

Appendix 2 Tables of Interference Values 351

Appendix 3 Evaluation of the Integrals by Numerical Analysis 427

Appendix 4 Flow Chart of the Computer Program Used for Determin-

A-4.2.1 Part One - Conversion of Life Data to Strength

,,c.... , ,.. . ,j... t•'• . .

1. This study ms concerned with the development of a practical

2. The study has provided a prediction technique which is almost "cook-

Relability Engineering Section

A complete cycle of reliability is made up of four stages; 1. Speci-

Early practices in stress-strength relationshis dealt almost entire-

The definition of the factors of safety varied from user to user,

Some of these definitions, found in literature, are listed below:

Factor of Safety = Ultimate Strength

Factor of Safety - Yield Strength

Factor of Safety - Yield Strength

Factor of Safety = Computed Strength

Margin of Safety Strength-Stress

Design Factor Strength