NAVWEPS 00-25-559

5 8 * & *
TIPS ON FATIGUE
r^atBU*nO& STATEWEHT A
D!S s Sits u 11
Belea
s e
Approved ror D"V
itpd
D'Ftribut on Unlimited
PUBLISHED BY DIRECTION OF
THE CHIEF OF THE BUREAU OF NAVAL WEAPONS
TIPS ON FATIGUE
by
Clarence R. Smith
Structures Des ign Specialis t
Fatigue Laboratory-
General Dynamics /Convair
Prepared for the
Bureau of Naval Weapons
DEPARTMENT OF THE NAVY
1963
For s ale by the Superintendent of Documents
U.S. Government Printing Office
Was hington 25, D.C. -- Price 70 cents
Reproduced From
Best Available Copy
2 0 0 1 1 1 3 0 1 2 7
TIPS ON FATIGUE
PREFACE
/Sooner or later, metal s tructures
under repeated load wear out. The prob-
lem is to be certain that it is later, rather
than s ooner./For this reas on, the Navy has
always encouraged res earch and develop-
ment to find ways of making s tructures las t
longer.
It has been found, however, that it is
not enough for s cientis ts and res earch en-
gineers to know the s ecrets of fatigue. If
des igners , s hop men, and ins pectors do not
recognize the s igns of fatigue, then the
purpos e of res earch and development has
not been realized.
To get this mes s age acros s , plain
s imple language and forthright pictures ,
are us ed, unhampered by s uperfluous tech-
nical jargon, theory, and detailed data dis -
plays .
This approach is one way to ens ure
that the findings of res earch become the
us able knowledge of the man in the s hop
and the man on the drawing board. The
premis e is that res earch and development
ard worth every cent they cos tif and
only if we make full us e of the new ideas
they produce.
Lack of communication between thos e
who know and thos e who need to know is
often the prime caus e of s tructural failure.
If the knowledge gained through fatigue
res earch over the las t 100 years were ap-
plied, many fatigue problems would never
occur (or recur).
Thous ands of documents on clas s ic
and applied res earch of metal and s truc-
tural fatigue literally bury facts by their
weight and profundity. Thes e documents
s hould be left to the experts .
On the other hand, the man at the
drawing board is a practical man. He
needs practical ans wers to s uch funda-
mental ques tions as : Is he continually
making errors that will res ult in fatigue
problems ? The fact that new airplanes
are s till failing in fatigue indicates this to
be true. Examination of s uch failures in-
dicates that many des igners are not even
aware that s harp notches are fatigue-
prone. Corrections for this one fault could
s ave millions of dollars per year and pos -
s ibly a few lives .
Hencej JJiis book propos es to be noth-
ing more nor les s than a s imple guide on
how to:
1. Recognize potential fatigue prob-
lems .
2. Rectify exis ting problems .
3. Avoid getting into s ituations that
may caus e problems . I
?-U'
1
in
TIPS ON FATIGUE
fpr: I Principles mentioned herein have to
do with (1) relations hips of one s tructural
member to another; and (2) paths of load
carried within the individual members .
This is not to imply that the more abs trus e
principles of s olid s tate phys ics would not
als o be helpful, but thes e are far beyond
the s cope of this work.] Before fatigue was
cons idered, airplanes were des igned to
withs tand a given s tatic load. This may be
in terms of the number of G's the airplane
may be expected to encounter during a
maneuver, or (in s ome cas es the larges t
gus t that may be encountered once in a
lifetime).
In any event, there was s ome des ign
number. Knowing the s trength of the ma-
terial, it was pos s ible to figure out how
much material was required to carry the
load. A perfect des ign was one wherein the
s tructure would carry 100 percent of the
des ign load and fail at 101 percent. Not
only that, no component would be relatively
s tronger than the next. Jus t like the "won-
derful one-hos s s hay that was built in s uch
a logical way that it ran for a hundred
years to a day." In his poem, "The Dea-
con's Mas terpiece," Oliver Wendell Holmes
(father of Supreme Court Jus tice) chroni-
cles that no part could fail firs t becaus e
each was cons tructed of the very bes t ma-
terial for the function to be s erved. Un-
doubtedly, des ign als o had s omething to do
with it.
. i n - *_w
) While airplanes are/s till Ides igned to
carry a certain s tatic load, fatigue pos es
the additional problem of es timating how
long the airplane will las t. Ideally, it
s hould las t as long as the des igner intended
it s hould.)
In des igning for s tatic s trength, the
des igner was given a s et of rules governing
the s tres s levels to which his materials
could be worked. This gave s ome uniform-
ity in des ign. The nominal s tres s levels
may have differed with location or purpos e.
For example, the compres s ion allowable
would depend on s tringer and bulkhead
s pacing, while tens ion allowables might de-
pend on the type of fas tener.
In fatigue, des igning to a uniform nom-
inal s tres s would not ens ure a uniform fa-
tigue s trength. A uniform des ign for fatigue
would involve the product of the nominal
s tres s times the s tres s concentration. Not
knowing the s tres s concentrations , this
would be an impos s ibility. Acknowledg-
ment is due to all thos e whos e encourage-
ment and as s is tance have made this book
pos s ible. While theirs is the glory for any
merit in the work, blame for any fault
herein belongs to the author alone.
As s is tance came from many s ources ,
all remembered and deeply appreciated,
though s pace limits mention to M. S.
Ros enfeld, Naval Air Engineering Center,
R. L. Creel and C. P. Baum, Navy Bureau
of Weapons for review of the entire book.
The author wis hes to es pecially ac-
knowledge the cartoons of W. Golds mith
and T. Adams ; the editorial as s is tance of
R. J. Prichard; the organizational help of
Ralph DeSola in the early s tages of the
work; and the technical as s is tance of G. G.
Green.
For the data and photographs that give
this informal work a concrete s et of ex-
amples , es pecially in Chapter 5, the author
wants to thank the following aircraft man-
ufacturers , operators and organizations :
-->
Aeronautical Res earch Laboratories ,
Melbourne, Aus tralia
IV
TIPS ON FATIGUE
Aeronaves de Mexico, S.A.
Aeros pace Development Center,
Wright-Patters on Air Force Bas e
Aircraft Plating Co.
American Airlines Overhaul Bas e
American Airmotive Co.
Beech Aircraft Corp.
Bell Helicopter Co.
Boeing Airplane Co.
Braniff Airlines
Bris tol Aircraft, Winnepeg, Ontario
Britis h Embas s y, Was hington, D.C.
Canadair, Ltd.
Ces s na Aircraft Co.
Aeronautics and Mis s ile Divis ion
Chance-Vought Corporation
A Divis ion of Ling-Temco-Vought, Inc.
Chapman Laboratories , Inc.
Continental Airlines
Delta Airlines
Douglas Aircraft Co., Inc.
Eas tern Airlines
Fairchild Engine & Airplane Corp.
General Dynamics Corp.
General Electric Co.
Grumman Aircraft Engineering Co.
Hiller Aircraft Corp.
Kaman Aircraft Corp.
Lockheed Aircraft Co.
Martin Co.
Metal Improvement Co.
Mexicana de Aviacion
McDonnell Aircraft Corp.
National Aeronautical Es tablis hment,
Ottawa, Ontario
National Airlines
National Luchtvaartlaboratorium,
Ams terdam, The Netherlands
Naval Air Engineering Center (ASL)
North American Aviation, Inc.
Northeas t Airlines
Northrop Aircraft, Inc.
Pan American Airlines
Republic Aviation Corp.
Standard Pres s ed Steel Co.
Trans canada Airlines
Trans -World Airlines
United Airlines
Wes tern Airlines
/This book repres ents an initial attempt
to make all levels of pers onnel aware of
the fatigue problem that exis ts in aircraft
s tructures ./ It is anticipated that revis ion $4.^
will be required in the future; cons equently
-
"
us ers comments are s olicited s o that a
meaningful revis ion may be accomplis hed.
Similarly, the photographs repres ent the
bes t illus trations pres ently available to
depict the problems dis cus s ed. Thes e
photographs were not s pecifically taken
for this purpos e; hence they are not all as
clear and uncluttered as would be des ired
Photographs that illus trate more clearly
the problems dis cus s ed herein or any other
fatigue problems occurring in aircraft
s tructures are des ired. Us ers comments
and new photographs s uitable for illus tra-
tion s hould be forwarded to:
Director (S-3)
Aeronautical Structures Laboratory
Naval Air Engineering Center
Philadelphia, Pa., 19112
San Diego, California
30 Oct 1963 C. R. SMITH
TIPS ON FATIGUE
CONTENTS
PREFACE
Chapter 1 The Fatigue Problem
Chapter 2 Bas ic Principles
Chapter 3 Joints and Joining
Chapter 4 Developing an Intuition for Fatigue
Chapter 5 Pas t Experience
Chapter 6 Making the Mos t of a Bad Situation
Chapter 7 Check Lis t
APPENDICES
A. Fatigue Tes t Data
B. Stres s Concentrations
C. Sugges ted Reading
Vll
COOLER
|..tHll(<
( IHIIIIV tiii r:iin tuniori
'Mijjliilj .,.!. r Mitny ait.t
iinith r.Mili-r liiinornm, H.-
tiinl ri|ii>n .in I'iiuc n-ll.
ItlOtt HNAL
****
ESTABLISHED IMI
SAN DIEGO, CALIFORNIA. MONDAY MORNING. MARCH 5, 1 9 DAILY 1 0 CENTS
Airliner Crashes
With 1 1 0 Aboard
j British Charter ,
| Flight Hits Near
Atlantic In Africa
LONDON'. Minvh 5 (Monday) *#*-A chartered Bril-
\<h ,ili-Uii<n- cniTj ing a crew of 30 and ebout 100 paftttO-
, Riu-s rallied last night in the west African republic
the Oimeroon, officials minouinvd early today. : - f)
Then* wns no Immediate,woni of casualtie*.
Thi* lour-vtijeiiut DC7C w"l__
!
TT"* ..
m>rted flylnij Irom Mm|BKa*'ll " " "".
hlque tri cut AJrlra tu LUX -
vmbourg on charter to Lon-
don firm. Trnim-Afrieart Air
Coach. Urt.
NKAB AUtrOHT
An Air Mlntwry upoluwwuiii.
Mid the cruh one wil bo-
Hoved to tw iwcr ttw Dotiala
airport not lar ffiem the At-
lande ooaat. He aid thtplan*
OUTtad 101 pMStKItftlM. ., .
A spokonaan lor ODMtotdf
an Aim*, : owner of *W
plan. Mkl thn unbar
uanaiaijirj u "'about 100.!*.',
Thrt* atownjrdottai' .wftt
roperta* to^h Oof h
The crath camt throe dayi
twr oao of tho wor*.i*M
Uon dUaaMr In Witory. i*rt
crath near Now York'* Ia%
wUd airport which too*: J
lives. - ' , " "
No tall ol t l IfcWB-
tan* ol Oi Camarooo
crash ljwre hnnadjalatx
available hair*.
DETAILS SOt'OUT $--4.J
Hot m aw Intern*
DB.UK nattomaity ot n', U B ,
era. Offidala aald aU MEftrtt,
were bolus mad to :**-..:*
Mils bom
caai wan alow.' -' ' "H' mM
Caledonian Airway, flyr
offices In Prtatwlck. S
, wiu mm***.*****
Ita*. lor iho leaae ol
. Douda* DCTCo ^to^itt*j|
-
:
"-- - ol a fleet tar **""*
ot fco to^rlfory ocUh^ J
BIGGER HEADLINE
IN THE HIKING ?
j tet Bounces
n Air; 1 0 Hurt
WKtO- SUrth s:iUonday>
m- ~k Pan; American World
wrnsa^.iarjot aldtner arty-
mg- rvtiemtata la
|tfh air after taantfni\Hono?
ijMiryoaiei
wi
::
rwitt
|tBta
rt<pfe<itUiadj oalJ treet-
. at a " " * " " " " " -' - * ' -
pat top. . .,. .. .S1,
J.:*VeaaNltNi4ywj
urt th bottoiMK9|M(l out."
BictHyw-4wtor mdaorA
Brooks 'aid. ' *mkwuat- ' tmW:
MOd, MV SlK^tVett up."';' ; ^ ^
'Brootn a*U neiber H.^ nor
no actrew iatte4 '" ti Stm-
cBon, Boff'ttwrowoBag com*'
panlon, Mrs, "t rm coraawia,
:
mi lajund.
:
*' .-.it
!
.a', - '<"-,: .
;^thiVe:7onot,- JPin,Amorlcan:
WABc
!
'ti*atIono- lr*tor hr;s
wft;;^,|llai^^.WaIl' not^da|ti>;
^^d^^'i|,^'U!at-j
apooatbie f
_Dkfa.ad. .
; nanof PHUtouigh ttdvotf bar]
TIPS ON FATIGUE
THE FATIGUE PROBLEM
1.1 A CENTURY OF STUDY AND
FATIGUE STILL FAILS STRUCTURES
The fatigue problem relating to metals and
s tructures has been inves tigated experi-
mentally for more than a century. In 1849,
Jones and Galton inves tigated cas t iron bars
in bending. They found that failure occur-
red in les s than 100,000 cycles if loaded
to more than one-third of ultimate bending
s trength. Similar work on wrought iron
built-up girders by Fairborn (1860-1861)
s howed s imilar res ults . Wohler's work for
the Prus s ian State Railways goes back to the
1850's when he made an extens ive s eries of
tes ts of various grades of iron and s teel
s ubjected to repeated direct tens ile and
compres s ive loads , to repeated bending
loads , and to repeated tors ional loads . Yet
we continue to read about and hear about
railroad wrecks , automobile s mas hups ,
airliner cras hes , and other catas trophes
directly attributable to fatigue in metallic
s tructures .
another report in much s maller type: "Jet
Bounces in Air: 10 Hurt." How many s uch
bounces can a s tructure s us tain before it
fails and becomes the fact behind an even
bigger headline?
Why an even bigger headline? The
firs t plane was a chartered plane on an
uns cheduled flight. It went down near the
coas t of Wes t Africa, and whatever happens
in remote places never s eems as real or
critical as what happens clos er to home.
The s econd plane carried more pas s engers
and was on a regularly s cheduled flight.
Its route involved the lives and emotions of
hundreds of thous ands of people locally.
When fatigue failure overcomes the s econd
plane, as it may in the cours e of time, it
is s afe to predict that the headline will be
bigger, the cas ualty figures more s tartling,
and the impact on the traveling public even
greater.
1.3 TODAY FATIGUE IS A BIGGER
PROBLEM THAN EVER
1.2 FATIGUE CAN BE BIG, BAD NEWS
"Airliner Cras hes with 110 Aboard: " read
the black headlines on 5 March 1962. On
the s ame front page, and at the foot of the
column des cribing the los s of the airliner,
her crew, and all her pas s engers , was
Airplanes in the pas t were not s ubjected to
loads experienced by pres ent day high s peed
aircraft. Als o, they were built of materials
whos e tens ile s trengths were s o low that in
order to s atis fy s tatic s trength require-
ments , s tres s es for s ervice loading would
automatically fall within ranges that would
provide an adequate fatigue life.
1-1
TIPS ON FATIGUE
The fatigue problem has ris en at an
alarming rate with pres ent day airplanes
having high s peed and performance. To
make matters wors e, the materials us ed
in pres ent day airplanes are s tronger, yet
have no better (and in s ome cas es , poorer)
fatigue properties than thos e us ed formerly,
The res ult has been an accumulation of
s ervice failures , s ome with fatalities
involved. Such s ituations have involved
the des igner in unfamiliar areas . Bes ides ,
more accurate methods of s tres s analys is
enable us to des ign s tructures with greater
efficiency and precis ion.
1.4 THE PANIC FACTOR
Des igners are us ually s hocked by s ervice
failures ; hence their s ubs equent des igns
and modifications of failed parts often
include a high panic factor. This panic
factor is likely to be far out of proportion to
the des ign improvement needed. The panic
factor is born of s udden fright. Some -
times it is compounded with ignorance,
and certainly its us e is contrary to all
the principles of good des ign.
Unanticipated fatigue failures caus e
des igners to become appalled at the
amount of information that s eems neces -
s ary to es timate the s ervice life of any
part or s tructure. The fact that their
knowledge of the ordinary mechanical
properties of materials - - ultimate and
yield s trengths , elongation, modulus of
elas ticity - - has failed them, leads them
to feel fully jus tified in us ing the high
panic factor. The natural tendency is to
"beef-up" the s tructure that failed, even
though this change may not be the s olution
and the weight penalty extreme, hi s ome
cas es , a removal of material might
s olve the problem, whereas a "beef-up"
may create a new problem jus t outs ide of
the "beefed-up" area.
1.5 THE VICIOUS CIRCLE
In aircraft des ign, every pound of s tructure
added requires additional weight in the form
of added power and fuel neces s ary to carry
the added s tructure. This quickly becomes
a vicious circle becaus e more s upport
s tructure is then needed to s us tain the
added engine and fuel required to carry
the "beefed-up" redes igned s tructure.
1.6 BREAK THE BIG ONE INTO LITTLE
ONES
As long as fatigue is treated as one enor-
mous problem, it never s eems to get
s olved. However, when fatigue is cons id-
ered as a number of s mall problems , the
s olution of each problem becomes appar-
ent. An initial approach to any problem is
to lis t the factors involved, s uch as :
1. What are the loads ?
2. What are the s tres s es ?
3. What are the s tres s concentrations ?
4. How much is the material good for?
5. Could we us e a better material?
6. Is that s hape neces s ary?
7. What about careles s s hop practices ?
8. What is the matter with ins pection?
9. Why didn't the engineer s ay s o if
that is what he wanted?
10. What "bird brain" called out this
heat treat?
This lis t could go on and on. The
point is , that in looking over even this
s hort lis t, the des igner, the ins pector,
and the s hop man can each find at leas t
one and probably more items that
he can pers onally do s omething about.
1-2
TIPS ON FATIGUE
1.7 IF EVERYONE DOES HIS OWN JOB
WELL ...
Accordingly, if everyone took pains to
correct faults in areas where he has
influence, a les s ening of fatigue failures
would s urely res ult.
Take the cas e of feathered edges . Jus t
becaus e des igners did not take the trouble
to call for corner radii, and ins pectors did
not reject parts having s harp corners ,
repairs were neces s ary on s everal recent-
ly built airplanes . While the repairs in
thems elves may not have been cos tly, the
interes t at 6 percent on an idle airplane
cos ting $5,000,000 will amount to over
$800.00 per day. Add to that the rental
value of facilities for repair and wages
of an idle crew, and the daily cos t is
appalling.
A fleet of 300 military airplanes was
recently modified to bring them up to de-
s ired fatigue life. It cos t 3200 man hours
per airplane to the tune of more than
$11,000,000 for the job. Down time
amounted to about 3 months per airplane.
1 . 8 ALL FOR WANT OF FATIGUE
RESISTANCE, THE BATTLE COULD
BE LOST
In the cas e of a military airplane, the cos t
can be failure to complete a mis s ion, which
in a critical s ituation would be impos s ible
to meas ure in terms of dollars and cents .
WM&
Figure 1.1. "Beach Marks " Identify Progres s ive Fatigue Failure (See Section 2.8)
L-3
TIPS ON FATIGUE
1.9 IF YOU CAN GET THEM WHEN fatigue des ign is s uperfluous and
THEY'RE LITTLE, FEW cos tly. Here is where the fatigue
PROBLEMS WILL GET BIG experts s hould be cons ulted. In
other cas es , the optional fatigue
The importance of cons idering fatigue des ign is s o s imple that little or
in des ign cannot be overemphas ized. no cos t is involved. Thes e are
In many cas es , it may s eem that mainly the items dis cus s ed in
the effort neces s ary for an adequate this book.
1-4
TIPS ON FATIGUE
BASIC PRINCIPLES
2.1 LET'S KEEP IT SIMPLE
This chapter will pres ent only thos e principles
that practicing engineers , ins pectors , and
s hop pers onnel can apply. In doing s o,
s ome of the more bas ic fundamentals of
crys talline s tructure, s uch as s lip planes ,
dis locations , and others will be omitted.
Thes e are beyond the s cope of this work.
While the principles dis cus s ed are adequate
for the purpos e intended, the reader is
directed to Appendix C, "Sugges ted Further
Reading, " for more detailed technical
information.
Knowing the behavior of metals under
load is vital to the unders tanding of fatigue.
Every day, common occurrences s how how
metals act when loaded. A thoughtful con-
s ideration of thes e examples is probably
the eas ies t way to s ummarize s ome of the
more bas ic principles of metal fatigue.
Have you ever noticed that you can break
a wire quicker by increas ing the bend angle?
2.2 IT'S NOT ONLY WHAT YOU DO,
IT'S HOW YOU DOIT AND HOW
MANY TIMES
2.2.1 UNIDIRECTIONAL VERSUS
REVERSE BENDINGRepeated bend-
ing is a familiar example of fatigue.
5 CYCLES
90-DEGREE REVERSE BENDING
A galvanized 14-gage wire that
breaks in four or five cycles of 90-
degree revers e bending .
2-1
TIPS ON FATIGUE
+ 90
. will las t for 15 or 16 bends
of from 0 to 90 degrees .
15 CYCLES
0 to 90-DEGREE BENDING
K bent only from 45 to 90 degrees , it may
las t from 60 to 70 cycles .
70 CYCLES
45-to 90-DEGREE BENDING
2 -2
TIPS ON FATIGUE
83,000 POUNDS
Similarly, the life of a s tructure may
depend more on how it is loaded than on the
total number of times it is loaded, or on the
maximum amount of the loads thems elves .
This is bes t s hown by loa xg an inch-
s quare bar of aluminum alky that breaks
at 83,000 pounds when loaded once.
ONE-INCH-SQUARE BAR BREAKS
AT 83,000 POUNDS
2 -3
TIPS ON FATIGUE
60,000 POUNDS (tens ion)
AAA
--) t-e I cycle
60, 000 POUNDS APPLIED 25, 000 TIMES
WILL BREAK THE BAR
60,000 POUNDS (tens ion)
~H * 1 cycle
60,000 POUNDS
(compres s ion)
Fatigue Failure
Identified by "Beach
Mark" Appearance -
See Section 2.8
However, if loads of from 0 to 60,000
pounds were applied, it would las t about
25,000 cycles .
Loadings from 60,000 pounds tens ion to
60,000 pounds compres s ion (commonly
called "plus to minus 60,000 pounds ")
would fail the bar in about 4,000 cycles .
REVERSING THE 6 0 , 0 0 0 -POUND LOADS
BREAKS THE BAR IN 4 , 0 0 0 CYCLES
2-4
TIPS ON FATIGUE
C/3
Q
ID
o
0 .
o
1,000 L0.000 100,000
CYCLES TO FAILURE
1,000,000
Figure 2.1 Load Vers us Fatigue Life for One-Inch-Square Bar
of 7075-T6 Aluminum Alloy-
Similar relations hips between re-
peated tens ion and revers ed loads are
plotted in Figure 2.1. Load is s hown on
the vertical axis (ordinate) and the number
of cycles on the horizontal axis (abs cis s a).
So that the lifetime s cale could be condens ed
to fit on one page, the abs cis s a has been
compres s ed in what is commonly known as
a logarithmic s cale. Note that each major
divis ion repres ents 10 times the value of
the previous divis ion. Schematic diagrams
of load cycles are s hown to facilitate read-
ing.
2.3 CONVERTING LOAD TO STRESS
The inch-s quare bar was us ed in the pre-
ceding example for two reas ons . Firs t, a
s quare inch is a s tandard unit of meas ure.
Second, when fatigue effects are unders tood
in terms of a s quare-inch cros s -s ection, it
is eas y to compare the load-carrying ability
of-our known example with the load-carrying
abilities of s tructures having other dimen-
s ions . In other words , load carrying ability
is then expres s ed in terms of pounds per
s quare inch of cros s -s ectional area.
Commonly expres s ed in terms of load
divided by cros s -s ectional area, the s hort-
hand or algebraic des cription for s tres s is
where
S
=
S = s tres s in pounds per s quare
inch
P = load in pounds
A = cros s -s ectional area in
s quare inches
730-755 0-642
2 -5
TIPS ON FATIGUE
2.4 FATIGUE SHORTHAND
Since fatigue life is not only dependent on
the amount of s tres s , but als o on how the
s tres s is applied, a s ys tem has been de-
vis ed identifying the type of loading, thus
R =
'mm
J
max
This is s imply the ratio of the minimum
s tres s divided by the maximum s tres s .
Us ing this notation, the curve for repeated
tens ion loading in Figure 2.1 would be
identified as R =0, becaus e the minimum
load was zero and zero divided by anything
is s till zero.
2.5 TENSION AND COMPRESSION
LOADING
According to convention, tens ion s tres s es
are always identified as plus (+ ) and
compres s ion s tres s es are minus (-). In
Figure 2.1 the curve for revers ed loading
would have a s tres s ratio of
J
mm
= - 1
'max
s ince S
oppos ite s igns
min
is equal to S
max
, except for the
2.6
COMPRESSION
S-N CURVES FOR SMOOTH
SPECIMENS
To pres ent fatigue data in brief form,
curves of s tres s vers us the number of cy-
cles to failure, (called S-N curves ) are
us ed. Since the dimens ion of the bars in
Figure 2.1 was one-inch s quare, applying
S = P/A, the curves s hown are als o S-N
Curves . Frequently, a whole family of
curves is given in order to s how lives for
other ratios of s tres s (R). Figure 2.2
s hows a family of curves . Appendix C,
"Sugges ted Further Reading, " contains
references to S-N data for other materials .
Curves for typical airplane s tructures are
given in Appendix A.
2-6
TIPS ON FATIGUE
80,0001-.
70,000
a 60,000
CO
f 50,000
I
g 40,000
w 30,000
R
20,000
10,000
1,000 1 0 ,0 0 0 1 0 0 ,0 0 0 1 ,0 0 0 ,0 0 0
CYCLES TO FAILURE
Figure 2.2 A Whole Family of Curves , Showing Lives for Various Ratios of
Stres s , R.
S = P/A
2.7 STRESS CONCENTRATIONS
In the cas e of airplane s tructures , the
fatigue behavior will be s ubs tantially differ-
ent from that obs erved in the s quare-inch
bar, becaus e free flow of s tres s is inter-
rupted by obs tacles s uch as holes , notches ,
bumps , and changes of s ection. Piling up
of s tres s at obs tacles s uch as thes e is
commonly called a concentration of s tres s ,
and the obs tacles thems elves are known as
s tres s rais ers .
2.7.1 UNFILLED HOLE -- In the
cas e of the one-inch bar, a s mall hole
through its center would caus e the
actual s tres s at the edge of the hole
to be about three times that away from
the hole.
S = P/A
Figure 2.3. Bar with Centrally Drilled Hole
2-7
TIPS ON FATIGUE
Figure 2.4 Photoelas tic Models , Showing Stres s at Edge of
Hole and at V Notch
The s tres s dis tributions around
two different concentrations are s hown
in Figure 2.4. The left picture s hows
the s tres s dis tribution at the edge of
an unfilled hole and the right picture
s hows what happens around a V-notch.
Here, an experimental s tres s analys is
technique, called photoelas ticity, is
us ed to vis ually demons trate locations
of highly s tres s ed areas . This tech-
nique employs polarized light and
clear plas tic models in which s tres s ed
areas become opaque. This is an es -
pecially valuable tool for demons trat-
ing relative merits of des ign. Amount
of s tres s is directly related to the
number of opaque lines and concentra-
tion is proportional to the line s pacing.
As in judging the s teepnes s of the ter-
rain by the contour lines of a topo-
graphical map, photoelas tic patterns
tell the s teepnes s of s tres s .
2.7.2 WATCH OUT FOR OPEN
HOLES -- Mos t s tructures
have holes . Open holes are us ually
wors e than rivet-filled holes . Stres s
at an open hole is three times that
away from the hole. For this reas on,
open holes s hould be avoided in regions
of high s tres s . Where location holes
are an abs olute neces s ity, plug them
with rivets if pos s ible. Never plug
holes with weLd, as this creates high
res idual (locked up) tens ile s tres s es .
2 -8
TIPS ON FATIGUE
Vinyl Plas tic
i
Polaroids
& =ZZ S^S
Figure 2.5. Photoelas tic Model, Showing Dis tribution of Load in Fas teners of a
Clevis Joint
2.7.3 RIVETS AND BOLTS -- Riveted
or bolted joints always cons titute problems .
One reas on is that the load introduced by
the rivet or bolt increas es the s tres s at
points of concentration.
The s econd reas on is that it is
virtually impos s ible to dis tribute the
load evenly between rivets or bolts , irre-
s pective of workmans hip. This is becaus e
the s econd and s ucces s ive rows of rivets
cannot carry their s hare of load without
s ome s tretch in the s plicing material
between the firs t two rows . In fact, the
s tretch in the s plicing material s hould be
greater than that of the material being
s pliced at this point. Note in Figure 2.5
that the fas tener neares t the load has the
highes t s tres s . A s olution to this problem
would be to make this fas tener incapable of
carrying s o much of the load. Unfortunate-
ly, reducing the s ize of the fas tener is not
always a s olution. In fact it is an invitation
to trouble with the fas teners thems elves .
Perhaps a wis er choice would be to remove
s ome of the s plice material s o that it would
not be able to overload the firs t fas tener or
fas teners . The thinner s plice material
s tretches , thus allowing s ome of the load to
be carried by the s econd row of fas teners .
This is illus trated in Figures 2.6 and 2.7,
where edge views of photoelas tic models
are s hown. The model in Figure 2.6 is
very s imilar to the clevis joint s hown in
Figure 2.5. Being cut from one piece of
material, the model in Figure 2.6 clearly
s hows that a good load dis tribution cannot
be attained by providing a better fit.
Figure 2.6.
Photoelas tic Model of Clevis Joint, Showing Edge View of Load
Dis tribution Between Fas teners
2 -9
TIPS ON FATIGUE
Figure 2.7. Improved Load Dis tribution by Scarfing
2.7.4 SCARFING Figure 2.7 s hows
an ideal joint wherein the load at the firs t
fas tener is relieved by s carfing the ma-
terial. Practically, this is hard to do.
Other methods of relieving the load at the
firs t fas tener will be dis cus s ed under joints
and joining in Chapter 3. In the meantime,
this would be one of the cas es where the
s enior engineer or fatigue s pecialis t s hould
be cons ulted.
2.7.5 SINGLE VERSUS DOUBLE
SHEAR The photoelas tic models of
joints s o far have been of the clevis or
double s hear type. Unfortunately, aerody-
namic s moothnes s requires des igning many
airplane joints in s ingle s hear, the wors t
example being the s imple lap joint. The
major fault with a lap joint is that the
s heets being joined tend to align with each
other, caus ing s evere bending s tres s es at
the firs t fas tener. This is s hown by the
photoelas tic model in Figure 2.8. Figure
2.9 J.s an improved lap achieved through
s carfing. Lap joints will be dis cus s ed fur-
ther in Chapter 3.
2.7.6 NOTCHES - Like the open hole,
the notch cons titutes one of the mos t
haras s ing problems in fatigue. Figurative-
ly s peaking, a notch is compris ed of any
kind of a dis continuity, the hole being one
of the many types . Thus , it is common to
s peak of the "notch effect" when the notch
being cons idered is in reality a bump.
Accordingly, the term "notch"
can be applied to holes , grooves , notches ,
bumps , etc. Perhaps the mos t trouble-
s ome of all notches is the fillet radius .
Machinis ts s eem to enjoy machining neat
corners , ins tead of providing generous
radii at the junctions of two s urfaces . The
importance of providing generous radii is
s hown in Figure 2.10, where a photoelas tic
model s hows how s tres s at a fillet can be
relieved by providing a more generous
radius .
Figure 2.8. Simple Lap Joint Note Bending
2-10
TIPS ON FATIGUE
Figure 2.9. Scarfed Lap Joint
A proper radius is es pecially
critical in machined parts where only
a s light change could determine
whether the part were s atis factory
or not. Mos t important, however,
is to make certain that an additional
notch is not created by the machined
radii failing to meet the flat s urface
s moothly as s hown in the photoelas tic
model in Figure 2.11. Further
examples of the effects of radii on
fatigue are given in Chapter 5.
2.7.7 FRETTINGThe eros ion of two
s urfaces rubbing agains t each Other is
known as "fretting." The notch effect of
the pitted s urfaces tends to exaggerate the
effect of other notches s o that the combina-
tion is a s uperimpos ed s tres s concentra-
tion. Fretting is eas ily recognized by
Figure 2. 10
Photoelas tic ModeLs Showing Effect of
Fillet Radii on Stres s
Figure 2. 11 Photoelas tic View of Machined Radii Not Meeting
Stres s is Indicated by Number of Dark Lines
2-11
TIPS ON FATIGUE
powder that s ifts out from between the
s urfaces .
Since fretting is caus ed by rub-
bing, anything that will reduce the amount
of rubbing will reduce fretting. Equalizing
the amount of s tretch between material be-
ing s pliced and s plice material, as illus -
trated in the photoelas tic models of s carfed
joints , will help. Adhes ive bonding (s ee
Section 3.9) als o helps . Lubricants can be
helpful in s pecial cas es . Cons ult the s pe-
cialis t on this .
2.8 IDENTIFYING FATIGUE FAILURE
Thos e of us who have had the dubious
pleas ure of looking at the broken end of a
drive s haft or a rear automobile axle,
can never forget what it looked like. We
may als o remember the mechanic s aying
that the s haft was old and crys tallized,
and that was why it failed.
This explanation, however, is
not neces s arily accurate. All metal is
crys talline. However, becaus e
fatigue cracks propagate through the
crys tals ins tead of around them (as in the
cas e of the one-time loading, or s tatic
failure in a ductile material) outlines of
the crys tals come into clear view.
Fatigue cracks propagate at
various rates , depending on the material
and loading. The s tres s concentration at
the end of a crack, being extremely high,
caus es the material to fatigue locally s o
that the crack continues until enough fres h
material (not yet fatigued) is engaged to
res is t loading for another interval. This
gives ris e to the "beach" mark appearance
of fatigue failed parts as s hown in Figure
1.1. Sometimes called "tide" marks ,
they are us eful for locating origins of
fatigue failure. The beach marks left by
high loading are us ually s paced farther
apart than thos e caus ed by low loading.
In either cas e, the origin is us ually at a
point of s tres s concentration or nucleus ,
and the beach marks propagate in circular
patterns with the nucleus as the center.
Us ually, the marks near the origin are
obliterated by rubbing of fractured s ur-
faces agains t each other. Thus in Figure
1.1, the origins were probably at lower
corners of the hole; however, beach
marks do not appear until s ome dis tance
away. After fatigue cracking had pro-
gres s ed to the las t beach mark at the
right, there was ins ufficient remaining
area to carry the load. Static failure
finally res ulted as indicated by the rough
s urface at the right.
2-12
TIPS ON FATIGUE
3 I JOINTS AND JOINING
3.1 TOO MUCH STRESS IN THE
WRONG PLACES
There was no fatigue problem in aircraft
during the era when they were made of
wood. It was only with the advent of all-
metal airplanes , and in particular of
high-s trength metal airplanes , that fatigue
became a problem. Why? Too much
s tres s in the wrong places !
The reas on we have too much s tres s
in the wrong places is that we have thrown
away the s imple approach us ed in gluing
wood, and ins tead we res ort to the obvious
boiler plate cons truction. Now, boiler
plate cons truction is fine when us ed on
boilers , but we don't have to build air-
planes like that.
Wooden airplanes were built with
cabinetmaker techniques , and the cabinet-
maker tried to join his s tructure in s uch a
manner that the joint was not apparent to
the eye -- nor to the s tres s . When the
s ame technique was us ed on wooden air-
planes , the s tres s flowed from one piece
to another as if they were one. Maybe it
was luck, but the res ult was a continuity
of s tres s flow.
3.2 WHY HAVE JOINTS?
Joining, to begin with, is a technique us ed
only when the s tructure cannot be built in
one piece. Ideally, the load is evenly
dis tributed throughout the s tructure to
afford a continuity of s tres s flow. Accord-
ingly, the more nearly the joining res embles
a s ingle piece in mis res pect, the better
the joint. The cabinetmaker 's long-s carf
joint very nearly s atis fied this condition.
3 .3 TYPES OF JOINTS
The mos t obvious way to join two s heets
of material together is to lap the edges of
one piece over the other and fas ten them
3-1
TIPS ON FATIGUE
with s ome device. His torians tell us that
man's firs t attempt of this s ort was pro-
bably fas tening two pieces of animal s kin
together with a fis h bone. The res ult
would be known today as a lap joint.
4 ^
LAP JOINT
DOUBLE SHEAR BUTT JOINT
Another type of butt joint holds
the two pieces of material with two s plice
plates . This is called a double s hear butt
joint. Being s ymmetrical, it has a fatigue
s trength s uperior to that of either the lap
joint or s ingle s hear butt joint. The double
s hear butt joint is preferable wherever cos t
and aerodynamics permit.
A neater way is to butt the two
s heets edge-to-edge and fas ten each to a
third s heet (s plice plate) underneath.
This gives a s mooth s urface on one s ide
that is not only pleas ing, but als o is aero-
dynamically s uperior if the s heet happens
to be the outs ide s kin of an airplane. The
butt joint with a s ingle s plice plate or dou-
bler is known as a s ingle s hear butt joint.
* = f* = *
SINGLE SHEAR BUTT JOINT
LUG
A fourth type of joint is the s imple
lug. It us ually cons is ts of a clevis and a
s ingle fitting that is pinned between the
clevis by a s ingle bolt or other fas tener.
This type is generally us ed for moving
parts where bus hings or bearings are us ed
for les s ening friction.
3 -2
TIPS ON FATIGUE
Many variations of the above
joints could be mentioned; however, their
problems and s olutions are s imilar.
Several exceptions , s uch as hooks and
piano hinges , need a s pecialis t's attention.
3.4 WHAT'S THE MATTER WITH
JOINTS?
CLAMPS? -- Not s o bad if bending
is away from rivets
3.4.1 LAP JOINTS - While the lap
joint is the s imples t of all joints , its main
problem is mat, when the two s heets of
material are joined, they tend to align
thems elves with each other. This caus es
the s heet to be bent at the firs t fas tener,
which is already s uffering from too much
load (s ee Figure 2. 8). This offs et in align-
ment is commonly called eccentricity.
The logical s olution would be to
let the s heet bend, as long as it didn't
bend right where the load was greates t.
Ideally, you could clamp the s heet to
make it bend at s ome other point; this
would s eparate the bending s tres s from
the load-carrying s hear s tres s . A trick
s uch as this is frequently called "confus -
ing the s tres s " or "s tres s confus er." See
Chapter 6 for other s tres s confus ers .
LAP JOINT -- Sheet bends right
where it hurts mos t

STHFSS CONFUSER
STRESS CONFUSER
3 -3
TIPS ON FATIGUE
While us ing a C-clamp would be
highly impractical in an airplane, the s ame
res ults can be obtained by driving extra
rivets through the edges of doublers .
In fatigue tes ts , an ordinary lap
joint with two counters unk rivets as in A
s ame problem as the lap joint. One of its
advantages is that the doubler can be made
thicker than the material being s pliced.
This reduces the effects of bending, but it
creates an additional problem: the rivet
neares t the doubler' s edge now carries
mos t of the load jus t as in the cas e of the
clevis joint s hown in Figure 2.5.
ZBZB2ZBB02.
A=55,000 Cycles
>>->>^>.>.>.-*VVVN--V^
1WMM
L
^s^^s
B=248, 000 Cycles
Same As "A" Except Extra
Rivets Are Driven Through
Edges Of Splice Plates .
To vis ualize this , s take the end of
a wide rubber band to a board with three
thumb tacks and pull. Note the amount of
deformation required at the No. 1 fas tener
before the No. 3 s tarts to carry the load.
This would indicate that s ome provis ion
mus t be made for the s plice plate to s tretch
failed after 55,000 cycles of repeated load-
ing. By driving extra rivets through dou-
bler edges as in B, the fatigue life was
rais ed to 248,000 cycles . The s econd joint
las ted longer becaus e the edge-driven rivet
could take no load other than that caus ed by
s heet bending, thus pas s ing the s hear load
on to the next rivet, which was thus re-
lieved of the bending load.
Thumb Tacks Rubber
Wood
3.4.2 SINGLE SHEAR BUTT JOINTS -
The s ingle s hear butt joint is really two
lap joints facing each other, s o it has the
EXPERIMENT -- Showing load
divis ion between fas teners
i:
n
HEAVY SPLICE PLATE -- Reduces
bending, but makes rivet carry too
much load
if the No. 2 and No. 3 fas teners are to
carry their fair s hare of the load.
Some degree of deformation can
be achieved by thinning the doubler
material between the firs t two rows of
fas teners s o the s econd row can carry
3 -4
TIPS ON FATIGUE
Very thin at
fas tener No. 1
Us e extra fas teners
if neces s ary to
carry s tatic load
Bending s tres s
is relieved
at firs t fas tener
and als o s ome
of the load
TAPERED SPLICE PLATE
s ome of the load. Since doubler material
mus t s tretch in order to do this , the
thicknes s at the firs t fas tener s hould be
les s than half that of the material being
s pliced.
The value of one-half is arbi-
trary; the point being that, with this
thicknes s , the No. 1 fas tener won't over-
load the s pliced material. High loads
would caus e doubler material at the No. 1
fas tener to yield in bearing -- which is
good, if the remaining fas teners can
carry the des ign load. Thus , for s tatic
s trength it might be a good policy not to
rely n the firs t row of rivets but, ins tead
to provide extra fas teners for the job.
Theoretically, the doubler s hould
taper to almos t nothing s o that the No. 1
fas tener carries an infinites imal part of
the load s omething on the order of the
cabinetmaker's s carfed wood joint. Gen-
erally, the extra machining is impractical,
s ometimes , however, the weight s aving
does make it worthwhile. Almos t as good
as the thick, tapered s plice plate is the
thin auxiliary doubler next to the material
being s pliced. The auxiliary doubler
s hould be long enough to engage an extra
row of rivets outs ide the main s plice area.
Here again, there is a compromis e between
the practical and theoretical optimum
thicknes s of auxiliary doublers .
3 -5
TIPS ON FATIGUE
AUXILIARY DOUBLER
TWO AUXILIARY DOUBLERS
For s plices in aluminum alloy,
an aluminum alloy auxiliary doubler about
one third as thick as the s pliced material
is about right. Where two auxiliary
doublers are us ed, the one neares t the
butt s hould be about one fifth as thick as
the material s pliced.
double s hear type als o has the problem of
load dis tribution between fas teners . (See
photoelas tic models in Figures 2.5 and
2.7). Scarfing, or providing auxiliary
doublers as for s ingle s hear joints , will
improve fatigue life.
Auxiliary thin doublers when
properly us ed will increas e the lifetime
more than ten times .
3.4.3 DOUBLE SHEAR BUTT JOINTS
Double s hear butt joints are s uperior to
thos e of the s ingle s hear type. This is
becaus e the s ymmetry of the double
s hear type eliminates the bending effects
found in the s ingle s hear However, the
3.4.4 LUGS The lug is a s imple form
of the double s hear joint. Since the joint
has but one fas tener, the problem of load
dis tribution between fas teners does not
aris e. Whether or not the fas tener (us u-
ally a bolt) fits tightly has made a s ub-
s tantial difference in tes t res ults . A loos e
bolt tends to bend more and will s ome-
times fail in the middle of the tongue, or
male fitting. It als o introduces an ex-
3-6
TIPS ON FATIGUE
WIDE THIN LUG -- eas y on bolt but bad on fatigue
tremely high bending s tres s on the corners
of the clevis , or female fitting. Thick
lugs with clos ely fitted bolts or bus hings
will have twice the fatigue life of lugs with
s loppy fits . A good interference fit
will improve the life many times . See
Chapter 6 for more on this .
Becaus e of the bending effect,
it is a good idea not to s tint on bolt
s ize. Indications are that it would be
helpful to have the bolt even twice as
s trong as the lug. This keeps from
overloading the corners of the lug.
NARROW THICK LUG -- better
than wide, thin lug, but bends bolt
too much;
Als o, yom won't have to worry about
the bolt.
Shape of the lug's cros s s ection
is very important. A wide thin lug,
while relieving bolt bending, caus es the
s tres s at the edge of the hole to be many
times the average s tres s away from the
hole. (See concentration factors in
Appendix B).
On the other hand, a narrow,
thick lug requires a long bolt,
which bends and
overloads the
lug edges .
3 -7
TIPS ON FATIGUE
Tes ts have s hown that the lug with cros s
s ections with ratios of A/B between 1
and 3 would be bes t where the bolt s hear-
ing s trength is equal to the s trength of the
lug.
3
m
U
U
1,000,000
100,000
10,000 i
Calculated
if Bolt
Didn't
Bend
Tes t for lug having
s tatic s trength
equal to that of bolt
Stronger bolts
permit us ing
larger A/B
for better fatigue
performance
A/B
3 -8
TIPS ON FATIGUE
3 .5
FASTENER SPACING
Much has been written about s pacing
fas teners in a joint, mos t of it from a
s tatic s trength point of view. Accordingly,
handbooks recommend fas tener s izes and
s pacing that leave as much of the area as
pos s ible in the s pliced material. For
s tatic s trength, the net cros s s ectional
area is us ually no les s than 75 percent of
the area away from the s plice. (Net
cros s s ection is the area remaining after
removal of material for holes .) This
would give a fas tener s pacing (commonly
called pitch) of about four times the fas -
tener diameter. To further enhance the
s tatic s trength, the firs t fas tener neares t
the load is s ometimes reduced in s ize.
This Row of Fas teners
has 5.7-Diameter Spacing
o 0 ^
Frequently, fas teners are s tag-
gered. Joints of this type have never
proven to be any better than the tandem
pattern. Staggering, however, is de-
s irable for fuel s ealing or other us es
where joints s hould not leak.
Thes e Rows Have
4-Diameter Spacing
TANDEM PATTERN
Good Static Strength Joint
STAGGERED PATTERN
OK for leak prevention
730-755 0-643
3 -9
TIPS ON FATIGUE
Des ign practices us ed for
optimum s tatic s trengths s eldom
apply to fatigue. Again, in fas tener
s pacing, the bes t practice for s tatic
s trength fall s far s h ort of good
fatigue de s ign. As has been s hown,
the s tres s at the edge of the firs t
fas tene r hole is th e mos t c rit ical for
fatigue. Accordingly, the bes t
fas tener pattern would be that which
would les s en the s tres s here .
In terms of the average s tres s
away from the firs t row of fas teners ,
the pattern can be likened to a s eries
of lugs . This would indicate that
the optimum s pacing for fas teners
normal to the direction of loading
would be about 2.5 diameters . In
terms of s tatic s trength, this would
amount to a joint whos e s trength was
60 percent of the s tructural s trength
7.0
U
<
tu
Z
o
"*! S 0
w
u
z
o
u
3 .0 - JL X
0 .2 0.4
d/D
0.6
away from the joint (in area clear
of holes ). Tes t data indicate that a
s pacing of 2.5 diameters between
rows is als o helpful.
While it is unlikely that the
s pacing indicated will be us ed
deliberately in practice, the idea
of reducing the s pacing might be
helpful if the s tatic s trength margin
were more than adequate and there
were a need to increas e fati gue
s trength.
* See App endix C fo r other
s tres s concentration factors
3-10
TIPS ON FATIGUE
Edge dis tance s hould always be
cons idered along with fas tener s pacing.
Except where the fas tener is us ed through
the edge of a doubler for clamping pur-
pos es (s ee Section 3.4.1), a minimum
s pacing of two-diameters from the center
of the fas tener to the edge of the doubler
is recommended.
The number of rows of rivets is
us ually governed by s tatic s trength
requirements and available s pace. For
fatigue res is tance, it is des irable to us e
two or more rows . Never one! This is
particularly true where s plices are made
over rib flanges . Often, one row is us ed
on the excus e that the flange is too narrow
to us e two. You can always find room for
another row away from the flange, or you
can s tagger a row.
T^n
TWO OR MORE ROWS OF
FASTENERS ARE A MUST
3
NEVER USE A SINGLE
ROW OF FASTENERS
STAGGERED ROW
3-11
TIPS ON FATIGUE
3.6 COUNTERSUNK FASTENERS
Counters unk fas teners are us ed for
aerodynamic or hydrodynamic s mooth-
nes s . Few other cas es will jus tify
the us e of a counters unk fas tener
not even the doors in the galley. Where
material thicknes s permits , holes for
fas tener heads are machined. Thes e
are called "machine counters unk
fas teners ." Where the court ers ink
is provided for by dimpling a reces s
in the s kin, the proces s is called
"dimple counters inking."
3.6.1 MAC HINE C OUNTERSINKING -
Us e of machine counters inking is not
recommended where the material
thicknes s is les s than 1.5 times the
depth of the head. The ragged edge
caus ed by the counters ink tool creates
a s uperimpos ed s tres s concentration
that invariably res ults in fatigue cracks .
(See "Feathered Edges , " Chapter 5)
BETTER
Sharp
Edges
POOR
BEST
3-12
TIPS ON FATIGUE
DIMPLE COUNTERSINK OVER
HEAVY SUBSTRUCTURE
DIMPLE COUNTERSINK OVER
LIGHT SUBSTRUCTURE
3.6.2 DIMPLE COUNTERSUNK
FASTENERS -- are to be
us ed wherever s kin thicknes s does
not permit machine counters inking.
When as s embled over heavy s ub-
s tructure, part of the s ubs tructure
is machine counters unk to accept the
protrus ion caus ed by dimpling the
s kin.
When as s embled over light
s ubs tructure, the s ubs tructure is
dimpled als o.
While mos t dimpled joints
have a s uperior fatigue life to thos e
with machine counters ink, extreme
care is needed in forming dimples
to prevent cracking around the peri-
phery of the dimple. This is us ually
caus ed by too s harp a dimple radius
or .by not s upplying the proper amount
of heat on forming.
It is a good policy to us e a
more generous bend radius than the
minimum s pecified in handbooks .
3-13
TIPS ON FATIGUE
3 .7 WELDING
Books have been written on welding and
welding proces s es . No attempt will be
made here to go into the many ramifica-
tions of welding: the advice of a s pecialis t
s hould be s ought when des igning welded
joints . However, a few pertinent facts re-
garding effects of welding on fatigue are in
order.
To begin with, it is not in the
cards for a weld to develop fatigue s trength
equal to that of the parent material. Many
vendors of s pecial equipment and materials
will make s uch claims ; however, their
claims are bas ed on laboratory tes t s peci-
mens , rather than on full-s cale s tructures .
Welds in s pecimens don't have as many de-
fects as thos e in the field. Defects s uch as
voids , s lag inclus ions , and dirt are good
places for fatigue to s tart.
Since fatigue failures always s tart
at s tres s concentrations , welding s hould be
done in a manner that affords the leas t con-
centration. Figure 3.1 s hows a compari-
s on of various weld configurations for
s tatic s trength, fatigue s trength, and cos t.
For purpos es of comparis on, a butt weld
having complete penetration and a flus h
s urface is cons idered as 100 percent (not
neces s arily 100 percent of the s trength for
parent material). Photoelas tic models in-
dicate s tate of s tres s at the welds .
Comparis ons of welds in s tainles s
s teel are pres ented in Figure 3.2. It will
be noted that the butt weld has very nearly
the s ame fatigue s trength as that of the
bas e metal. As previous ly s tated, caution
s hould be exercis ed in us ing data from
s mall s pecimen tes ts for es timating
s trength of fabricated s tructures . How-
ever, the data s erve to illus trate pos s ibili-
ties of good welding.
3-14
TIPS ON FATIGUE
Figure 3.1 "Welded Joints Compared for Static Strength, Fatigue Strength
and Cos t.
g 100
w
a > 80
to k
CO hH
s
S
OT
40
Bo
60
H
0
BASE
.METAL
BUTT WELD
FILLET W LLD
SPOTWELD^
^. ^_
10 100 1000 10,000 100,000
CYCLES TO FAILURE
Figure 3.2 Fatigue Strength Comparis ons of Welds in Stainles s Steel
3-15
TIPS ON FATIGUE
HEAVY LANDS FOR
ATTACHMENTS
\
EROUS
RADIUS
SUBSTRUCTURE
3.8 SCULPTURED STRUCTURE
In keeping with the Ideal that a perfect
s tructure wouLd be built of one piece,
s culptured s tructures (machined from
one piece of material) have come into
us e.
effect of concentrating the s tres s . At the
time of the pres ent writing, no good ap-
proximations are available for the right
thicknes s of land material needed. In all
cas es , need for extremely generous fillet
radii and good workmans hip is paramount.
3.9 ADHESIVE BONDING
Thes e s culptured parts are put
together in erector-like fas hion, extra
thick material being provided at attach-
ment points to negate the effect of holes .
For s tatic cons iderations , a 25 percent
increas e in thicknes s might be s atis fact-
ory, but for fatigue cons iderations ,
the amount of buildup (land) depends on
the s tres s rais er. For the s imple
hole, a buildup of three times the nom-
inal s kin thicknes s m ight be s atis factory;
then again, it might not. The reas on is
that the thick land builds up a mis align-
ment at the center of the material and
the res ulting eccentricity accentuates the
Finally, getting back to our cabinet-
maker's technique, adhes ive bonding for
metals has become quite common. When
us ed with cabinetmaker technique (s carfed
very thin at the edges ), bonded joints have
s uperior fatigue s trength. One reas on it
is not more widely us ed is that it is difficult
to ins pect for defects . Als o, mos t ad-
hes ives deteriorate with time and environ-
mental expos ure. As a res ult, a combina-
tion of bonding and mechanical fas tening is
often us ed in primary s tructures , s ort of
like wearing belt and s us penders . In s uch
cas es , however, the bonding is us ed for
fuel s ealing, vibration damping, or other
s pecial purpos es .
Mos t adhes ives become brittle
when cold and s corch when hot.
Cons ult the s pecialis t on this .
3-16
TIPS ON FATIGUE
DEVELOPING AN INTUITION
FOR FATIGUE
4 . 1
SEAT OF THE PANTS TECHNIQUE
We have frequently heard the term "flying
by the s eat of the pants ." This has more
merit than s ome of the ins trument s pecial-
is ts would care to admit. Certainly the
number of G's to which the airplane is
expos ed would largely be a function of the
comfort or dis tres s experienced by the
pilot. Yet the feel for the amount of trim
required for s pecial s ituations is largely
a res ult of practice and intuition. Simi-
larly, one can develop an intuition for
fatigue res is tance.
4.2
NATURE'S WAY OF DESIGNING
Boiler plate cons truction is man-conceived.
You never s ee tree fronds connected directly
to the huge trunk of a tree; nor do you s ee
branches attached to the trunk with huge
gobs of extra wood. Yet man, in s plicing
NATURE DOESN'T DO THIS
NOR THIS
4-1
TIPS ON FATIGUE
two pieces of material, often us es a
doubler the s ame thicknes s or thicker
than the material being s pliced. Con-
s equently, mos t of the load is dumped
on the firs t row of fas teners in the
s plice.
Intuitively, we s hould know
that the s econd row of fas teners can-
not begin to carry load without s ome
"give" at the firs t row. Naturally, we
don't want the "give" to occur in the
s tructure that we are trying to protect.
Intuition s hould tell us that the doubler
material between the firs t two rows of
fas teners s hould be thinned down s o it
will s tretch without overloading the
firs t row of fas teners . See Chapter 3
for more detail.
Similar to the heavy s plice
doubler, the huge gob of material that
is machined away to almos t nothing in
s pots s hould rub our intuition res pons es
the wrong way. Sometimes the s mall
fingerlike extens ions are made for
attachment purpos es . There is nothing
wrong with this . What's wrong is that
the fingeriike extens ion ends abruptly
in a gob of material where s tres s can-
not help but concentrate.
4.3 SHARP CORNERS
Sharp corners , either internal or external,
have produced more fatigue failures in
metallic s tructures than any other one
common fault. There is nothing aes thetic
about a s harp corner. In fact, any work-
man who takes pride in his work .takes
s pecial pains in s moothing off s harp
edges and corners . Examples of what
happens when s harp corners are left un-
finis hed are given in the next chapter.
4-2
TIPS ON FATIGUE
4.4 DON'T FORCE IT
A bigger hammer is rarely the
ans wer. Intuition s hould tell us
that when s omething has to be
forced, everything is not well.
An example of this is trying to
fit a heavy lug into a clevis that
is too s mall. This fault is us ually
obvious . Jus t as bad, but not s o
obvious , is the cas e where the
clevis has been s prung by bolt
tightening becaus e the s pacer bus h-
ing was too s hort.
4.5 ROUGH FINISH
Similar to the s harp corners is the
rough finis h. Like the cheap
automobile paint job s pecial, the
rough finis h reflects s loppy workman-
s hip. Intuition s hould tell us that an
exces s ively rough machine job is not
right. Where our intuition may need
further training is in telling us how
much roughnes s is tolerable. While
s awtooth finis hes are never tolerable
anywhere (s ee Chapter 5), we s hould
s ens e that long s traight machined
s urfaces would not have to be s o
s mooth as fillet s urfaces . A good
rule of thumb is to us e exceptional
care at changes in s ection s ometimes
even to the point of hand finis hing.
4-3
TIPS ON FATIGUE
4.6 POWER OF KNOWING
Intuition has been defined as the gift
of knowing without recours e to infer-
ence or reas oning, but it takes a
lot of obs ervation and logic to get that
gift.
Previous chapters have pre-
s ented s ome of the pure reas on of what
fatigue is and how to avoid it. The
next chapter will pres ent s ome examples
of fatigue.
While it is hoped that the
s ugges tions pres ented herein will
prepare the reader to better cope
with whatever s ituation aris es , the
reader will have to acquire his own
experience.
As the father who tries to
prepare his s on for the pitfalls in
life -- none s eem to be like the ones
des cribed.
Experience with s imilar
s ituations , however, s hould help.
The important thing to remember
is not the individual problem, but
the logic us ed in its s olution. Later,
you'll get to the point that you auto-
matically know what's right -- that's
intuition.
4-4
TIPS ON FATIGUE
5 PAST EXPERIENCE
5 . 1 STRUCTURAL BLUNDERS AND
HOW TO AVOID THEM
As the Pars on s aid, "Everybody feels
what's right, but don't always do right. "
The feeling for what's right can be s timu-
lated by object les s ons of what's wrong.
That's why the Pars on preached
fire and brims tone on Sunday -- and
that's why this chapter is devoted to
thos e wrongs that are known by s uch
terms as blunders , butches , botches ,
boo-boos , genfs , mis takes , and what
have you
When we s ee the cons equences of
wrong, we are driven to do what's right.
Take blunders , for example. There are
des ign blunders , s hop blunders , and that
horrible group of atrocities that are con-
ceived as des ign blunders and perpetuated
as s hop blunders . The des igner who s ays
"I don't need to s pecify details , the s hop
man can take care of that" would be horri-
fied to hear the s hop man s ay "If the fool
des igned it that way, I'll build it that way.
5-1
TIPS ON FATIGUE
The photographs of fatigue failures
collected in this chapter have been furn-
is hed by various manufacturers and air-
plane operators who hope that s eeing will
make believers of des igners , ins pectors ,
and s hop men.
Examples of des ign butches
s tart off with notches in various forms ,
including des igns with no radii, s hort
fillet radii, and s harp bend radii. Special
cas es of notches include s quare holes and
feathered edges . Feathered edges may
more logically be blamed on the s hop,
s ince mos t s tandard s hop procedures in-
clude rounding off all corners . Never-
theles s , where s harp edges are antici-
pated, s pecial pains s hould be taken to
s ee that the drawing clearly s pells out
rounding off all corners . This is es pe-
cially critical where careles s handling
might res ult in nicking s harp expos ed
edges . Poor load dis tribution is another
des ign botch, and clos ely related to this
is the cas e of s uperimpos ed s tres s con-
centration in its various forms .
5-2
TIPS ON FATIGUE
Shop boo-boos involve poor
craftmans hip, failure to recognize des ign
goofs , and failure to alert engineering
when a des ign or s hop mis take is dis -
covered. Es pecially critical are dings
res ulting from careles s handling practices .
While s hort rivet or bolt edge dis tances
can us ually be blamed on the s hop, fre-
quently engineering drawings do not make
allowance for s ufficient s pace to ins tall the
number of rivets or bolts required.
No matter who was at fault, a
good crafts man would never leave ins u-
fficient edge dis tance or mis matched
machined s urfaces , even if the blueprint
called for them. Even a poor crafts man
would blus h if he were s hown the failure
that res ulted becaus e he had ins talled a
bolt of the wrong length.
5-3
TIPS ON FATIGUE
5.2 NOTCHES
5.2.1 SHARP NOTCH - The very
exis tence of the s harp notch is a res ult of
s tupidity, careles s nes s , and unwis e
penny-pinching.
The manufacturer may aim at
s aving a penny, but the us er will always
pay heavily, and often the manufacturer
will pay later in the form of warranties .
In any event, machine time s aved is
peanuts when compared with overall s truc-
tural integrity cos ts and, in many cas es ,
us ing machine cutters with proper edge
radii will avoid the s harp notches at no
extra cos t.
Thous ands upon thous ands of
fatigue failures s imilar to thos e s hown on
the oppos ite page could have been avoided
by good des ign and crafts mans hip.
every tool s ales man with the need, machine
toolmakers would s oon des ign cutters s o
that s uch crimes would be impos s ible to
commit. The s harp radius , as well as the
means for making it, s hould be outlawed
the way they outlawed gun s lingers . Tools
for making s harp radii s hould be kept
under lock and key for us e on nons tructural
parts , s uch as as h trays .
Similar failures can als o occur in
well filleted parts where the fillet has tool
marks or s cratches .
5-4
TIPS ON FATIGUE
730-755 0-644
5-5
TIPS ON FATIGUE
5.2.2 DINGS -- Dings can be intentional
or accidental. Upper left s hows the fatigue
failure res ulting from an accidental blow
by a blunt object. Lower left s hows failure
res ulting from identification marks , and
upper right s hows how marks left by a
rivet bucking bar caus e a wing s par to
crack in the radius .
Failure at lower right is s imilar
to upper right, except the tool was an
impact s crewdriver us ed for removing the
wing fuel cell cover attachment s crews .
Precautions s hould be taken in
handling parts to avoid denting and nicking.
Should dings s uch as s hown be dis covered,
they s hould be called to the attention of the
foreman and fatigue s pecialis t. Thes e
can frequently be corrected.
5-6
TIPS ON FATIGUE
O
o.
5-7
TIPS ON FATIGUE
5.2.3 FILLET RADIUS - Think big
when it comes to fillet radii. Radius of
the lug in the upper left photo would s eem
ample, but it only had a fifth of its ex-
pected life. A s imilar lug in the upper
right photo has a radius that was increas ed
by removing s ome of the material. This
lug with the larger radius had a fatigue
life eight times greater.
The lower photo s hows fatigue fail-
ure through a fillet that would have s eemed
to be OK, but the fact that the part failed
in s ervice indicates a larger radius s hould
have been us ed, es pecially s ince there
was no problem of clearance in this area.
A good policy is to us e as large
a radius as s pace and practicality permit,
es pecially at changes of s ection thicknes s es ,
5-8
TIPS ON FATIGUE
Original Life = N Cycles
Redes igned Life = 8N Cycles
5-9
TIPS ON FATIGUE
^
5.2.4 SHARP BENDS OR NUISANCE
FAILURES -- The clip in the s ketch cos ts
about 10 cents , but, buried in the ma^e
s hown in the photograph, it took 32 man-
hours to replace While failures of clips
s uch as this may never cons titute more
than a nuis ance, the replacement cos t is
appalling.
The problem here is that s omeone
thought the s harp bend in the clip radius
looked cleaner and was within the allowable
bend radius for the material. Unfortu-
nately, s pecifications for bend radii were
bas ed on the ability to form without crack-
ing at the time of bending. Little thought,
if any, was given to fatigue performance.
Us e a bend radius that is as gen-
erous as pos s ible without des troying
functional performance. In s ome cas es ,
us ing a s tainles s s teel clip that is one
gage thinner than the original aluminum
clip would be in order. Cons ult the
s pecialis t on this .
5-10
TIPS ON FATIGUE
5-11
TIPS ON FATIGUE
5.2.5 SQUARE HOLES-There is us ually
a purpos e in making s quare holes . Some-
times holes are made s quare for no pur-
pos e at all.
The hole in the upper photograph
was made to accommodate a rectangular
duct. Solution to the problem was to make
an elliptical hole encompas s ing the original
rectangular hole. Fatigue life was five
times that with the rectangular hole.
The hole s hown in the lower photo-
graph was made for no purpos e other than
to s ave les s than two ounces of weight. It
could have been round, better yet, forgotten
altogether. The part failed des pite the
s eemingly generous corner radii.
"Rectangular Duct"
*ssga
5-12
TIPS ON FATIGUE
^S
5-L3
TIPS ON FATIGUE
5 . 3 FEATHERED EDGE AND SHARP
CORNERS
No one in his right mind plans to have
s harp corners or feathered edges . They
us ually occur becaus e the des igner didn't
pay attention to what would happen when his
two- or three-drawing views were inte-
grated. He would be s urpris ed to find raw
edges of the kind that caus ed s ervice fail-
ure in thes e examples .
Bes ides being a s tres s rais er
in its elf, the feathered edge or s harp
corner is eas ily nicked. Thus , a
s tructure employing s imilar parts
may fail mos t frequently at the nicks
on feathered edges , and s omeone
might wrongly infer that feathered
edges were all right -- it was only the
nicks that were wrong.
However, If there were no
feathered edges , there would probably
be fewer nicks . A nick on top of a
feathered edge cons titutes a s uperimpos ed
s tres s concentration. While every con-
centration does n't caus e failure, why
have concentrations that could be s o
eas ily avoided. The time could better
be s pent reducing the effect of concen-
trations that can't be avoided.
5-14
TIPS ON FATIGUE
l
-*&
5-15
TIPS ON FATIGUE
Don't s end a boy -
5.3.1 THIN FLANGES - This part
has faults s hown in the preceding example
becaus e the s potfaces res ulted in
feathered edges . Thes e feathered edges
might ultimately have caus ed a fatigue
failure, but an additional faultthat of
us ing thin flanges to carry load in bend-
ingcaus ed the failure even s ooner.
This does not mean that it is
always undes irable to have areas a long
dis tance from the center of bending. The
oppos ite is true. Take the cas e of the I-
beam. The flange is remote from the
center of bending; but, there is enough
beef to more than compens ate for its
dis tance from the center of bending.
In the example s hown, the thin
flange added little to the s trength of the
forging. This caus es the s tres s at the
outermos t flange fibers to be higher than
it would have been in the plate s urface
without the flange. Whereas thin flanges
attached to heavy main s tructures can
s ometimes be condoned on the bas is of
increas ed rigidity, this is done at the ex-
pens e of increas ed s tres s and s hould not
be attempted without the approval of the
s pecialis t.
The flange on an I-beam is an
example of s upplying material where
needed- Jus t as important is the remov-
al of material where not needed. Thus ,
removing s ome of the material provides
a generous fillet, (as in Section 5.2.3)
and improves fatigue life. Another
example of the need for material removal
will be s hown in Section 5.4
to do a man' s job!
5-16
TIPS ON FATIGUE
,#SW*
mH
|j b
t?i
THIN FLANGE
CRACK
5-17
TIPS ON FATIGUE
5.3.2 TROUBLESOME SPOTFACE -
Spotfacing is a common caus e of fatigue
failure. While not quite as bad as the exam-
ple s hown on the previous page in that no
vulnerable outs tanding legs were left the
s harp corner combined with a feathered
edge is a s ure caus e of trouble.
5-1!
TIPS ON FATIGUE
*..:.
,
:*^
l
' . * " ' . "' ' "' . -
:
.':;' y.V'' '^':\/
:
'
V
0.V
;
.V''''V''
;
', "^'V^
:
;..
%
it
i .vi; - " - , ' '.. .,-< * .,': ' ", .' ' "' ' ', $* # '
V
:
- F' ; ' .
:,i
-
/
' . "' A > ' "' ' ' - * ' *i- ' ' -, l'
;
" ' ' ' *
W
:

:
- ' - - ' ^i
jgjLiV v,V-H? :' "J ''"' '" -<-*\V i:'^;,,-
jfi
H^/? .' ' "' si' .' ' ' &
C
^H^,' *
r
'' ' '''.ykilr
^^^L r
-
;
".'i
*
jiH
:
-^*
'" '-^M
^^^^^KnTTnT A -A
^^^^H^HHK
^^^^^^^^^^M > **
5-19
TIPS ON FATIGUE
i .4
BAD LOAD DISTRIBUTION
5.4.1 STRINGER FORGING-Thes e
s tringers failed becaus e there was too
much s tres s at the end fas teners . Since
the s tres s here is caus ed by a number of
items , including axial loading, bending, and
fretting, a reduction in any or all would
improve life.
An attempt was made to reduce
the s tres s here by us ing s maller rivets .
As you s eeit didn't help. While us ing
s maller rivets at the ends of s plices is
common practice for s tatic s trength, their
us e for fatigue s ituations has never proven
helpful.
One s olution is thinning the forg-
ing (s ee Section 3.4.2) to relieve the bend-
ing effects of the s ingle s hear attachment.
This als o prevents overloading the firs t
fas teners and at the s ame time reduces
fretting by equalizing the s tretch in
s tringer and forging.
Another s olution is to replace the
end rivets with interference fit fas teners
(s ee Section 6.3). This is es pecially help-
ful where an eas y fix is wanted to bring
exis ting s tructures up to required life.
Replacing end rivets with s omewhat larger
diameter tapered bolts will ens ure remov-
ing fatigue damaged material (provided no
cracks remain vis ible) bes ides providing a
better-than-new fatigue s trength.
5-20
TIPS ON FATIGUE
%
->-'-~\r\$ et
^
730-755 0-B45
5-21
TIPS ON FATIGUE
5.4.2 SPLICE This s plice contains
the s ame problem as the dagger fitting on
the previous page. Being more compli-
cated, the problem is not s o eas ily diag-
nos ed. Thus , the us ually propos ed
s olution is to increas e the thicknes s of
the material being s pliced. While this
may s olve the fatigue problem, the extra
weight penalty may be extreme.
Solution for this problem is the
s ame as for the s tringer attachment
forging, except that the loads on all firs t
line rivets mus t be relieved. A common
practice is to cut finger-like extens ions
to the s plice doubler as will be s hown on
the next page. This , however, is not s o
effective as tapering the doubler or,
better yet, cutting the fingers and als o
tapering the doubler.
5-22
TIPS ON FATIGUE
ste-.'
:
''' ' -
J
mm
U- ^i 1
5-23
TIPS ON FATIGUE
5.4.3 THE CONTINUOUS MEMBER -
It would be fine to have all continuous
members with no intervening s plices .
This would end many of our fatigue prob-
lems . Since we do have s plices , the idea
of making only a partial s plice at a given
location leads to trouble.
This example would appear at
firs t to have been caus ed by us ing s plice
plates that were too thick*, as in the pre-
vious example. The primary failure,
however, was not at the end rivet as the
photograph would indicate. This failure
happened after the continuous s tringer
broke directly over the s plice area,
dumping its s hare of the load on the s kin.
What happened was that there
was too much give in the riveted joint, s o
that the s kin was unable to carry its fair
s hare of the load, caus ing the s tringer to
fail. The s kin failure was s econdary.
* They really were too thick at the end
rivets . This might ultimately have
caus ed the failure as s hown had not the
initial failure been in the s tringer.
5-24
TIPS ON FATIGUE
mmmmmm
5-25
TIPS ON FATIGUE
5. 4. 4 HEAVY FORGING -- Thes e are
typical s ituations where heavy pieces
are attached to relatively thin under
s tructure, res ulting in the failure of
the s tructure
Wherever s ituations like thes e
aris e, the s olution is to feed out the
s tres s through s ucces s ive layers of
thinner material or to provide gradual
taper to the firs t point of attachment.
If this is not pos s ible, perhaps
the heavy piece can be s plit normal to
the direction of s tres s s o the under
s tructure can have "Breathing Room,"
To lower the s tres s fluctuation res ulting
from the hole its elf, taper bolts at
points of attachment will help.
5-26
TIPS ON FATIGUE
I*
k,
fe
CR^**
1
*"
0
tfc
.-:.-;-;;
si i pi Pi pspi si i i Pi
w&
5-27
TIPS ON FATIGUE
5.4.5 NO BACKUP STRUCTURE - One
of the fundamentals of s cience is that for
every action, there has to be a reaction.
Here, a hus ky control bracket was attached
to a flims y s par web s tructure with res ults
as s hown. Solution is to provide s upport
s tructure (backup s tructure) having
s trength equal to that of the fitting. This
differs from the examples s hown in 5.4.4,
which were known to have adequate s upport
s tructure.
This is a typical problem where
one pers on des igns a component while an-
other des igns the s upporting s tructure. In
this particular cas e, it is doubtful that
much thought was given to s upporting s truc-
ture. Rather, the bracket was attached to
a s tructure that was des igned for another
purpos e.
5-28
TIPS ON FATIGUE
5-29
TIPS ON FATIGUE
5.5 SUPERIMPOSED STRESS
CONCENTRATION
As though notches were not bad enough,
they can be made wors e by s uperimpos ing
one upon another. Thus , we have exam-
ples of feathered edges that als o terminate
at heavy s ections and fillet radii that don't
match the res t of the machined s urface.
5.5.1 HOLE PARTLY THROUGH - The
oppos ite photo s hows how fatigue can
s tart at a hole that was not drilled all the
way through. The s olution is s imple:
finis h drilling the hole.
5-30
TIPS ON FATIGUE
-K * v . '.--
'c wJ PI
' ./A.-
^^^Sw
" " . * * x
>^yH
1 1
,
'."w'
-;^
* > " ~
' v .' 'viv''c
K ^ ' " -I * _* * '^^^1
'.* a-* ?'-
'*m
' * >>..'- *
* ^
I ... ""*
'i * .
5-31
TIPS ON FATIGUE
5.5.2 INTERSECTING HOLES-Here is
a cas e where a hole was tapped into the
highly s tres s ed region of another hole in
the tens ion flange of a wing beam. This is
s imilar to the cas e of a greas e fitting hole
for a bearing.
If you mus t have a hole, the idea
is to move it to the leas t damaging pos ition.
A little local reinforcing (beef up) is als o
helpful.
POOR
beef up
ALSO ACCEPTABLE
5-32
TIPS ON FATIGUE
f /
:
,?:J 'f?Pii
FORWARD
I
1
5-33
TIPS ON FATIGUE
5.5.3 RADIUS AT CHANGE OF
SECTION There always has to be a
radius of s ome s ort at any change of
s ection. Likewis e, there has to be a
radius at a change of direction. However,
you don't have to make one radius right on
top of the other. The radius for change in
s ection s hould have been made at another
location. If this were impos s ible, both
radii s hould have been enlarged to permit
a more gentle trans ition.
5-34
TIPS ON FATIGUE
' J**'
-;f.
*
Vi
$>
5-35
TIP N FATIGUE
5. 5. 4 ROUGH SURFACE FINISH - A
finis h s uch as that illus trated here als o
cons titutes a s uperimpos ition of s tres s .
Where a s tres s rais er already exis ts , as
in this cas e, it is foolis h to let a rough
s urface like this get by, es pecially with
tool marks normal to the direction of
loading.
5-36
TIPS ON FATIGUE
730-755 0-64--G
5-37
TIPS ON FATIGUE
5. 5. 5 CROSS GRAIN -- The parts
s hown failed for two reas ons , rough
s urfaces , and the material's grain
s tructure is normal to the direction
of Loading. While it might have been
pos s ible to avert failure by machining
a s mooth s urface, the wrong direction
of grain makes s uch a s olution highLy
s peculative.
5-38
TIPS ON FATIGUE
5-39
TIPS ON FATIGUE
5.6 AUXILIARY ATTACHMENTS
The part s hown was from a landing
gear as s embly. Failure occurred through
a rivet hole us ed for attaching a s chafing
s hield. Similar things happen to frames
having attachment s crews to s upport hy-
draulic lines , uphols tery, or what have
you? It would have been better to tie the
part on with rope (als o try adhes ive bond-
ing) than to take chances on fatiguing as
s hown.
5-40
TIPS ON FATIGUE
5-41
TIPS ON FATIGUE
5.7
HARD PLATING
Here is a typical example of where a part
was chrome plated to make it more wear
res is tant. It was n't more fatigue res is tant.
The cracks in the plating act as s tres s
rais ers that eventually fail the part the
plating is s uppos ed to protect.
Shot peening prior to plating is a
common inhibitor of fatigue cracking in
chrome plated parts . It is unwis e to
chrome plate paints for dimens ional buildup
or wear res is tance without the help of the
s pecialis t.
5-42
TIPS ON FATIGUE
5-43
TIPS ON FATIGUE
5. SHOP BLUNDERS
While mos t of the examples previous ly
s hown can be blamed directly on des ign,
a number could equally well have been
caus ed by s hop blunders . Thus , we have
the hole that was not tapped all the way
through and the rough s urface that was not
s moothed. The following examples can be
blamed almos t entirely on s hop practices .
5.8.1
NO

EDGE
DISTANCE - The
engineering drawing may not have s peci-
fied the exact locations of holes for nut
plates ; however, s tandard s hop practices
s hould be s uch that this would never
happen. As s hown, there was ins ufficient
room for nut plates to be s paced in s uch a
manner that holes would fall between nut
plates . Als o, note that the edge s urface
finis h was nothing to brag about.
5-44
TIPS ON FATIGUE
5-45
TIPS ON FATIGUE
5.8.2 MISMATCH-- Even excellent
machinis ts often machine a curved
s urface that does n't meet its s traight
counterpart, leaving what amounts to
a s uperimpos ed s tres s rais er. While
it is not s o bad where the two s urfaces
are convex, the concave ones us ually
res ult in failure as indicated in the
illus tration.
Many drawing room manuals
s pecify die maximum allowable mis -
match. It s o happens (not in thes e
cas es ) that fatigue failures have
res ulted where the mis match was
within tolerance. Care s hould be
exercis ed in permitting mis matches
in critical areas -- even within
s pecified tolerances . While it
would be virtually impos s ible to
define the.amount of mis match that
can be permitted in every cas e, a
rule of thumb is to us e extreme
care with concave s urfaces .
5-46
TIPS ON FATIGUE
l
%
^' M4a*i'
5-47
TIPS ON FATIGUE
5.8.3 EXCESSIVE CLAMPING-The bolt
on this part was tightened without having
the proper s pacer bus hing. Fatigue failure
finally s et in, as might be expected. Make
s ure you have the right length bus hing and
the right length bolt and THINK TWICE
BEFORE TIGHTENING (s ee 4.4).
5-4*
TIPS ON FATIGUE
5-49
TIPS ON FATIGUE
i.9
ORTGTNAL FITTINGS
A. WRONG GRAIN DIRECTION
B. SPOTFACES
C. INADEQUATE FILLET RADIUS
D. SHARP CORNERS
E. FEATHERED EDGES
F. ROUGH SURFACE
G. UTS=240-250 ks l
MURPHY'S LAW
Murphy's Law s tates that if it were pos s i-
ble to botch up a job, s omeone will s urely
find a way to do it. DON'T UNDERESTI-
MATE GROUP EFFORT ! As s hown in the
upper left photograph, the combined efforts
of engineering and s hop very nearly s uc-
ceeded in doing everything wrong.
Botches in this one part include
s harp edges , bad s potfaces , s mall fillet
radii, rough s urface, cros s grain, and
others . In addition, there were s igns of
hydrogen embrittlement due to cyanide-
bath cadmium plating. Hydrogen embrit-
tlement is a term us ed for low ductility
caus ed by abs orption of too much hydrogen
during proces s ing. Thes e nice big words
REDESIGNED FITTING
H. ENLARGED FILLET RADIUS
I. ROUNDED CORNERS
J. SMOOTH SURFACE FINISH
K. NO SPOTFACES
L. UTS=210-220 ks i
s ound authoritative when us ed to explain
failures for which no real reas on (other
than poor des ign or workmans hip, which
we hate to admit) is apparent. Decarburi-
zation (another mouthful that means los s of
carbon due to poor proces s ing) was als o
apparent to a minor extent.
Proper proces s ing was ins uffi-
cient to bring the part up to required life,
s o it was neces s ary to perform a major
overhaul. This included providing a better
s urface finis h, removing s harp edges , and
providing more generous fillet radii. The
reworked part is s hown in the upper right
photograph. Tes ts on s imilar parts re-
vealed a life of approximately four times
that s us tained by original parts with no in-
creas e in weight.
5-50
TIPS ON FATIGUE
MAKING THE MOST OF A BAD SITUATION
6. L
BAD SITUATIONS
6. 1. 1 THE CASE OF THE
FLAT TIRE -- Have you
ever had a fLat tire while driving
along the countrys ide and dis covered that
your s pare was als o flat? ... No pump or
patching material ... s o you finally
decided to drive it flat. Then there was
the joker who pas s ed you and yelled, "Don't
you know you got a flat tire?" And you felt
like s ticking a big s ign on your rear bumper
s aying, "I know it's flat - - s o what?"
If you have had s uch an experi-
ence, then you can imagine what it's like to
have a s ervice failure in an area where
there jus t is n't enough room for replace-
ment with a hus kier part. Maybe you are
already us ing material as s trong as you
dare.
What now? Shop is s till turning out
parts (like thos e that broke) by the barrel-
full, and you're faced with the need for a
quick decis ion. You have three pos s ible
decis ions : (1) you can do nothing and hope
that the res t of the parts won't be s o bad;
(2) you can try to fix it up and hope that
the fix is OK; or, (3) you can fix a few
s amples and tes t the parts to s ee if the
fix is any good.
6-1
TIPS ON FATIGUE
6.2 DECISIONS
Oar decis ion on a s tructural fix s houLd
be bas ed on the facts we've learned
s o far. The purpos e of this chapter
is to arrange thes e facts in s uch a
manner that our decis ions can be
eas ier. What are thes e facts ?
fatigue life. In Chapter 5 we found
what happens when we vioLate principles
of good des ign and fabrication. As in
vioLation of principles of good health,
the corrective meas ures may be s low,
painfuL, and bad tas ting.
6.3 SIZING UP THE SITUATION
Some of the bas ic principles
were pres ented in Chapter 2 Als o
s hown were methods for reducing
s tres s at fillets by providing a more
generous radius .
The s ubject of joints was
introduced in Chapter 3. Of partic-
ular importance was the fact that a
s mall change in bas ic des ign could
res uLt in a vas t improvement in
As far as fatigue is concerned,
remember that a s tructure will
never fail except at a s tres s concen-
tration. Accordingly, Let's worry
about s tres s at the concentration and
never mind about what happens els e-
where, at leas t not for the time
being. This s implifies our problem.
The next thing is to vis uaLize what can
be done to this particular s tres s to
make the mos t of the s ituation.
DECISIONS -- ALWAYS DECISIONS
6-2
TIPS ON FATIGUE
We s tart out with a s tres s
like this
zero
s tres s
which could pos s ibly be s hifted to
til LSo o o e o
zero
or again it might be changed to act Like
this ...
or s omething might be done to make it
behave Like this
zero
F
21.
6-3
730-755 0-647
TIPS ON FATIGUE
While there may be others thes e
cas es will s uffice to s tart. Suppos e a part
were loaded s o that s tres s at the concen-
tration fluctuates from 0 to 60,000 ps i
(R = 0). Such a part could be expected to
60,000
J __J
90,000 Cycles , Stres s
Range 1 60,000 ps i, R = -1
90,000 cycles (curve for R = -1) and maxi-
mum s tres s =30,000 ps i, Figure 2.2.
30, 000 Cycles , Stres s
Range = 60,000 ps i, R = 0
las t for about 30,000 cycles , according to
the S-N curve for R = 0 s hown in Figure 2.2.
Where the s tres s range is defined as the
difference between the maximum and
minimum s tres s , total s tres s range would
be 60,000 ps i.
The introduction of a compres s ive
s tres s at the concentration s ounds like a
nice trick if you can do it. That is , the
compres s ive s tres s s hould be a permanent
affair - - locked up s o it can't get away - -
and s hould be in jus t exactly the right s pot.
Such s tres s es are commonly known as
res idual s tres s es . Res idual s tres s es can
either be in tens ion or compres s ion.
Similarly, if the s tres s could be
made to behave as though it were cycling
6.3. I MOVE THE WHOLE STRESS DOWN
If it were pos s ible to do s omething to the
s tructure locally s o that the s tres s at the
concentration would be 30,000 ps i in com-
pres s ion (-30,000 ps i) when the part was
unloaded, the original loading s hould caus e
the localized s tres s to fluctuate between
-30,000 ps i and +30,000 ps i ( 30,000 ps i).
This would corres pond to the s ame s tres s
range as before, but the life now would be
200,000 Cycles , Stres s
Range = 60,000 ps i, R = -2
from minus 40,000 ps i to plus 20,000 ps i,
a lifetime of 200,000 cycles would res ult
(curve for R = -2 and maximum s tres s =
20,000 ps i, Figure 2.2).
6-4
TIPS ON FATIGUE
When a part having a s tres s
rais er is loaded in tens ion s o as to
caus e the material at the concentra-
tion to yield locally, the permanently
deformed material mus t go into com-
pres s ion when the load is removed and
the part s prings back. Such items as
hooks lend thems elves to this type of
correction. Generally, the amount of
overload is critical and s hould be
s pecified by the s pecialis t.
A practical method called
"s hot peening" is us ed to introduce
res idual compres s ive s tres s for a
longer life. In the firs t ins tance, a
compres s ive layer at the notch,
amounting to 30,000 ps i, will do the
trick. In the s econd, a layer of 40,000
ps i would be required. Both are eas ily
achieved, it being common practice to
introduce res idual compres s ive
s tres s es as high as two thirds of the
material's compres s ive yield s trength.
Other methods of introducing protec-
tive compres s ive s tres s layers include
controlled mechanical peening, vapor
blas ting, s urface rolling, and a proces s
called "coining." Surface rolling is
es pecially s uitable for cylindrical ob-
jects s uch as bolts . Figure 6.1 s hows
how thread rolling was us ed to im-
prove the fatigue life of bolts .
Fatigue life of a part can be
improved by providing a better finis h.
This is particularly true when the
original part failed becaus e of machine
mis match. (A mis match occurs where
the two machine s urfaces do not meet-
s ee s ection 5.8.Z).
2 0 0 ,0 0 0
1 80 ,0 0 0
( Bolts having threads rolled prior to heat
treatment have about the s ame fatigue
s trength as bolts with machined threads )
1 0 ,0 0 0 1 0 0 ,0 0 0 1 ,0 0 0 ,0 0 0 1 0 ,0 0 0 ,0 0 0
Figure 6. 1 Thread Rolling
6-5
TIPS ON FATIGUE
Figure Z.ll s hows how s tres s es
can be very s evere as a res ult of this .
Simply s moothing out the radius would
be a s olution to s uch a mis match.
Other cas es may not be s o s imple. We
may find a part made with a s harp
notch s o that a bearing could s it clos e
Let's take the cas e of a Lug.
As s ume that the machining direction
were normal to the direction of the
Load As we have aLready s een, this
would cons titute a s uperimpos ed
s tres s concentration. However, if the
machining had been paraLLeL to the
POOR POOR
Direction
of Machining
BETTER BETTER
to the s houlder of a s haft. Even s o, it
is s ometimes pos s ible to improve the
dis tribution of s tres s by removing
s ome material to eliminate the notch
and s till have a clos e fit, as s hown in
the lower s ketch.
direction of the Loading, a s ubs tan-
tial Increas e in fatigue life wouLd
res uLt. This not oaLyappLles to
rough s urfaces but wouLd be equaLly
appLLcabLe to any s urface finis hing.
6-6
TIPS ON FATIGUE
6.3.2 CUT OFF THE TOP OF THE
CURVE -- For ordinary rough
s urfaces , s ometimes s moothing will
do the trick; but the manner in which
the s moothing is done will have a lot
to do with how s atis factory the part
will be. For example, the part that
failed in Section 5.4 might have
been s atis factory if the direction of
the machining had been with the radius .
The importance of removing
material to improve fatigue life cannot
be over emphas ized. This not only
res ults in a lighter s tructure, but one
that is trouble free.
Our examples s o far have been
for removing material in fillets .
Equally important, perhaps more s o,
is removing undes irable external humps ,
Photoelas tic models in Figure
6.2 s how how a beam can be reworked
to lower the s tres s at the concentration.
Except for removal of material s hown
in the cros s hatched area, the lower
model is identical to the upper. Being
identically loaded (one agains t the other)
the revis ed model had the s tres s
lowered to two thirds of the original
value. This would increas e fatigue
life more than ten times . Removing
NOTE: Stres s is directly propor-
tional to number of fringes
Figure 6.2 Removing Material Lowers Stres s
6-7
TIPS ON FATIGUE
0 STRESS (REF.)
Figure 6.3 Analogy of Stres s Cycle as Influenced
by Interference Fit
a s mall amount of material makes the dif-
ference between s atis factory and uns atis -
factory performance.
6.3.3 CUT OFF THE BOTTOM OF THE
CURVEAnother method of s hifting
the s tres s cycle is by means of an inter-
ference fit fas tener or a pres s ed-in bus h-
ing. Here's how it works . Turn back to
Figure 2.2. Let's imagine that our part
failed becaus e the actual s tres s was cy-
cling at R = 0 with a maximum s tres s of
50,000 ps i. If we could make the s tres s
behave as though loading were at R = +0.5,
we s ee that the life would then be about
10,000,000 cycles . In other words , if we
could block the hole and keep the s tres s
from returning to zero, the load could be
reduced to zero, yet the s tres s would
remain at 25,000 ps i in the example we
have s elected. Practically, this is eas y to
do. Figure 6.3 illus trates the principle
involved in the us e of an interference fit.
Where the deflection of the s pring is
likened to a s tres s cycle, it can be s een
that an introduction of a hos e s egment to
prevent the return to zero would lower the
excurs ion without adding to the maximum
deflection.
Figure 6.4 s hows the effect of
various amounts of pin interference on the
fatigue life of s mall lug s pecimens . While
fatigue life does not s eem to be lowered
appreciably by exces s ive amounts of inter-
ference, look out for pos s ible s tres s cor-
ros ion. With tapered bolts , now commer-
cially available, the amount of interference
can be clos ely controlled. Amounts of in-
terference s hould be thos e recommended
by the tapered bolt manufacturer.
6-8
TIPS ON FATIGUE
1,000,000
H
CO
/ f
t
o ,
3/8-In ch Di ameter
o
Jj; Taper Pi n
H
B|[
\-4
100,000
e|
|; )
M
"l
*
o
j
CO
a
o
i n .mn
1 7
0.000 0.002 0.004. 0.006
TAPER PIN INTERFERENCE - INCH
0.008
Figure 6.4 Effect of Interference Fit
on Fatigue Life
Figure 6.5 s hows a graph for
large lugs with interference-fit bus h-
ings . In order to achieve the interfer-
ence s hown, the bus hing mus t be longer
than the lug thicknes s , with exces s
length chamfered 10 degrees . Exces s
length is ground off after pres s ing.
K^
L
v
j
.75-
10
W
n
<
h0.5
O
H
to
W
u
u
Bus hing with 1/8-inch
Wall Thicknes s
Bus hing with
1/4-inch Wall
Thicknes s
When more than an 0.002 inch inter-
ference is us ed with an ordinary bus h-
ing, the bus hing tends to broach the
holes .
0 0.005. 0.01
BUSHING DIAMETRICAL
INTERFERENCE (INCH)
Figure 6.5 Pres s -Fit Bus hings
6-9
TIPS ON FATIGUE
Examples in Chapter 3
s how how fatigue life could be im-
proved by adding edge-driven rivets
to relieve the effects of bending in a
riveted joint. The problem area in
any riveted joint will us ually be at the
row of fas teners near the load.
Figure 6.6 s hows how this
can be achieved by providing a larger
hole in the doubler (at the firs t loaded
fas tener) while us ing an interference-
fit bolt in the material being protected.
This acts very much like the edge-
driven rivet des cribed in Chapter 3:
tight
loos e
Figure 6.6 Interference Bolt
Ins tallation
axial and bending loads are s eparated,
carrying bending loads only at the
firs t row and pas s ing s hear loads to
the s econd row. Both devices are
examples of what are commonly called
s tres s confus ers , but the interference-
fit bolt has the advantage that it can
s till carry s hear load in an emergency,
which the edge-driven rivet cannot do.
Fatigue tes ts of joints
s imilar to that s hown in Figure 6.7
s how a fatigue life increas e of from
117, 000 cycles to 1, 371, 000 cycles
over s imilar joints with 1/4-inch
diameter rivets . Us ing an inter-
ference fit fas tener without the over-
s ize hole in the doubler res ulted in a
fatigue life of 3Z3, 000 cycles .
Figures given are averages for five
tes ts each.
6-10
TIPS ON FATIGUE
7 CHECK LIST FOR FATIGUE RESISTANCE
7.1 A GREMLIN'S MENAGERIE OF
COMMON OVERSIGHTS
In the light of what we have learned about
fatigue, let's review s ome of the mos t
common overs ights that caus e fatigue
failure.
When things go wrong in an air-
plane, it is common practice to blame it
on a gremlin one of thos e mythical
foot-high, ill-humored imps that have
been dis rupting the works s ince man
firs t took to the air. In the cas e of
fatigue, however, it would s eem that
even a gremlin would need s ome
help. It's as though the gremlin has a
menagerie of common overs ights , each
"animal" with its own s tructural char-
acteris tics that only a gremlin could
love.
Since no one loves thes e critters ,
it is common practice to s ay they belong
to s omeone els e that therefore they are
s omeone els e's res pons ibility and that
if they are ignored, maybe they will go
away. The critters never do go away.
In fact, if let alone, they tend to multiply.
There are ways to breed this
menagerie of overs ights , and there are
ways to reduce it. When the des igner
feels he is being pes tered by a s eemingly
s tupid ques tion from a s hop man who has
found what the s hop thought was an over-
s ight, it is recommended that the lofty
engineer lis ten with interes t, res pect,
and appreciation.
Maybe this time the s hop man
has n't found an overs ight, but it's a
cinch he won't keep looking for one if
his efforts aren't properly appreciated -
if he comes in ten times and only the
tenth is an overs ight, it may be the one
that s aves the des igner's reputation and
100 lives .
7-1
TIPS ON FATIGUE
On behalf of the s hop man, there
undoubtedly will be cas es where a poor
des ign has been allowed to get by becaus e
of cos t or other excus es . In areas of very
low s tres s , this might not be s erious , but
always remember that NEW AIRPLANES
STILL FAIL IN FATIGUE becaus e over-
s ights have been allowed to get by.
Only a few of the critters will
be s hown, but enough to relay the idea.
The thing to remember is that whenever
one is found, the time to make the change
is now. You may s ometimes change a
part that might have las ted its required
lifetime, but it's better to be s afe than
s orry.
The gremlin's menagerie is in-
tended to jog your memory by as s ociating
related ideas . Numbers in parenthes is
at the right refer to appropriate s ections
in the previous text.
While there may be many other
items that need checking, the few
mentioned here s hould s uffice for s timu-
lating the alertnes s required to catch
mis takes before they become s erious .
A good plan would be to make a
game of lis ting all pos s ible des ign and s hop
errors . Try to categorize them. You will
find many times the few lis ted herein. Als o,
by this time you will know what needs to be
done to fix it. Good luck'.
SHARP NOTCHES
GOUGES
NICKS
ALLOW STRETCH
BETWEEN FIRST TWO
ROWS OF FASTENERS
SHARP EDGES
RAPID CHANGE IN
SECTION
CHEWED UP SURFACES
FAT GOBS OF EXCESS
MATERIAL
LONG SPANS OF UN-
SUPPORTED STRUCTURE_
DANGLING MEMBERS
HIDDEN MISTAKES
HARD PLATING
TROUBLESOME
SPOTFACE
EXTRA HUMPS
7 -2
TIPS ON FATIGUE
SHARP NOTCHES (5. 2. 1)
7 -3
TIPS ON FATIGUE
GOUGES (5. 5.4)
7 -4
TIPS ON FATIGUE
NICKS (5. 2. 2)
7 -5
TIPS ON FATIGUE
ALLOW STRETCH BETWEEN FIRST
TWO ROWS OF FASTENERS (2. 7. 3)
7 -6
TIPS ON FATIGUE
SHARP EDGES (5.3)
7 -7
TIPS ON FATIGUE
RAPID CHANGE IN SECTION (5.5.3)
7 -8
TIPS ON FATIGUE
CHEWED UP SURFACES (5. 8. 1)
730-755 0-648
7 -9
TIPS ON FATIGUE
FAT GOBS OF EXCESS MATERIAL (6.3.2)
7-10
TIPS ON FATIGUE
LONG SPANS OF UNSUPPORTED STRUCTURE
7-11
TIPS ON FATIGUE
A. - -... - .yi-
i
)
DANGLING MEMBERS
7-12
TIPS ON FATIGUE
HIDDEN MISTAKES
7-13
TIPS ON FATIGUE
HARD PLATING (5 7)
7-14
TIPS ON FATIGUE
TROUBLESOME SPOTFACE (5.3.2)
7-15
TIPS ON FATIGUE
EXTRA HUMPS (6.3.2)
7-16
TIPS ON FATIGUE
APPENDIX A-FATIGUE LIFE OF TYPICAL STRUCTURES
10
S
T
R
E
N
G
T
H

0
0








o

o






o

-

^wX.
P5IDWi n g
O Wi n g
& Verti col Tai l
D Hori zon tal Tai l
i-i 7 0
H
< 60
to
-
Horizontal Tall
* * * * *
i i Range of Test Data
J ^N. ^-P5ID Wing
H 50
W
O 40 $

W
^ 3 0
I
a
Wir
g ^ *
/-Vertical Tail
R0
J ?n
M
A
X
I
M
U
M

o

-
i i i 11 i n i i I l 1 i ll I 1 1 1 Mi l 1 1 l Mi ll 1 1 1 1 Ni l
100 1000
LIFE TO FAILURE
10,000
- CYCLES
100,000 1,000,000
NOTE: Load values are given in per cent of s tatic failing loads . Note that the
vertical and horizontal s cales are both compres s ed. Except for the vertical tail,
where R = 0, cycling was from a load repres enting 1 G to the maximum value
indicated.
TIPS ON FATIGUE
APPENDIX B -- STRESS CONCENTRATIONS
W 4
1
r? W
3
O ^
H
O
<!
h
i i
H
w
O
o

CO
CO
W
tf
H
co
L
< J 4
W

< i
<^
S 3
en
co
co
W
CO
0.40
TIPS ON FATIGUE
APPENDIX C-SUGGESTED READING
Following is a lis t of references On fatigue for thos e who wis h to purs ue the s ubject
further. While there are many other valuable references ,thos e lis ted repres ent a cros s
s ection of the thinking during the las t decade.
STP 9, References on Fatigue
STP 91, Manual on Fatigue Tes ting
A.S. T.M. Special Publications ,
Publis hed by The American Society
for Tes ting Materials , 1916 Race
Street, Philadelphia
STP 203, Fatigue on Aircraft Structures ,
1957
STP 237, Sympos ium on Bas ic Mechanis m
of Fatigue, 1959
STP 274, Sympos ium on Fatigue of
Aircraft Structures , 1960
STP 284, Sympos ium on Acous tical
Fatigue, 1961
STP 338, Sympos ium on Fatigue Tes ts of
Aircraft Structures : Low-cycle,
Full-Scale, and Helicopters
American Society for Mechanical Engineers . Conference (International) on Fatigue of
Metals -- Proceedings , 29 Wes t 39th St., New York 18, 1956
American Society for Metals . Metals Handbook , Vol 1, 8th Edition, Metals Park,
Novelty, Ohio, 1961
Forres t, P. G. Fatigue of Metals , Pergamon Pres s , New York, 1962
Freudenthal, A. M. (Editor). Fatigue in Aircraft Structure (Proceedings of the
International Conference Held at Columbia Univers ity), January 30, 31, and
February 1, 1956
TIPS ON FATIGUE
Grover, Gordon, & Jacks on. The Fatigue of Metals & Structures , U.S. Government
Printing Office, Was hington D.C., 1954
Harris , W. J. Metalic Fatigue With Particular Reference to Significance of Certain
Standard Aircraft Fabrication and Finis hing Proces s , Pergamon Pres s , New
York, 1961
Hey wood, R.B. Des igning by Photoelas ticity, Academic Pres s Inc., Publis hers ,
New York, 1952
Heywood, R.B. Des igning Agains t Fatigue of Metals , Reinhold Publis hing Co.,
New York, 1962
Peters on, R. E. Stres s Concentration Des ign Factors , John Wiley & Sons , Inc., 4
New York, 1953
Plantema, F. J. & Schijve, (Editors ). Full-Scale Fatigue Tes ting of Aircraft
Structures , Pergamon Pres s , New York, 1961
Society for Experimental Stres s Analys is . Handbook of Experimental Stres s , Edited
by M. Hetenyi, 21 Bridge Square, Wes tport, Connecticut, 1950
Sines , George and J. L. Wais man, (Editors ). Metal Fatigue, McGraw-Hill,
New York, 1959
Weibull, W. Fatigue Tes ting and Analys is of Res ults , Pergamon Pres s , New York,
1961
y
U.S. GOVERNMENT PRINTING OFFICE: 1964O-730-755