Behaviour of Skin Fatigue Cracks at the Corners of Windows ...

R. & M. No. 3248;

MINISTRY OF AVIATION

AERONAUTICAL RESEARCH COUNCIL
REPORTS AND MEMORANDA

Behaviour of Skin Fatigue Cracks at the Corners
of Windows in a Comet I Fuselage

By R. J. ATKINSON, W. J. WINKWORTH and G. M. NORRIS

LONDON: HER MAJESTY'S STATIONERY OFFICE
1962

FOURTEEN SHILLINGS NET

Behaviour of Skin Fatigue Cracks at the Corners
of Windows in a Comet Fuselage

By R. J. ATkINSON, W. J. WINKWORTH and G. M. NORRIS

COMMUNICATED BY THE DEPUTY CONTROLLER AIRCRAFT fRESEARCH AND DEVELOPMENT),
MINISTRY OF AV.IATION

Reports and Memoranda No. 3248*

June, ±960.

Summary. Fatigue tests on a Comet I pressure cabin subjected to operational pressure cycles are described.
Cracks at window corners are the main subject of investigation. Results are compared with earlier experiments
on other Comet I pressure cabins. Conclusions are reached that appear to have some general significance.

1. Introduction. Fatigue investigations made on one of a number of Comet I fuselages that had
been specially provided for research purposes are described in this Report. In these investigations
the pressure cabin was subjected to pressure loading cycles only, and attention was directed mainly
to the initiation and development of cracksat the window corners. Special care was taken to avoid
unnecessary destruction, and cracks were repaired as necessary to prevent catastrophic failure.
Nine cracks out of a total of sixteen actually reached the stage where sudden extension appeared
imminent.

Results are examined below in the light of earlier fatigue work on a Comet pressure cabin1,2
subjected to wing loads as well as pressure. In the case of strain measurements, use is also made of
the results from a static test to destruction3 made on a third fuselage under pressure alone.

2. The Test Specimen. The fatigue tests were made on the fuselage of Comet I G-ALYR,
(Fig. 1). This aircraft was built in 1952 and had made 747 pressurized flights. (N.B. All numbers of
pressure cycles are totals, i.e., service plus test cycles.) The diameter of the fuselage was 10 ft 3 in.
and the pressure cabin was 70 ft long.

2.1. Basic Structure. T h e basic structure consisted of circumferential frames 21 inches apart,
stringers at approximately 5.5 in. pitch,, and the skin covering. The frames were of zed section,
2.75 in. deep and notched to take the witch-hat stringers which were bonded to the skin (Fig. 2).
The skin was generally 22 s.w.g. (0. 028 in.) except along the sides of the fuselage, where 20 s.w.g.
(0. 036 in.) skin contained the windows (Fig. 3). Skin material was D.T.D.546 (Appendix).

Attachment between the skin-stringer panels and the frames was usually by 2 B.A. countersunk-
head bolts at the stringer flanges only. In the centre section, however, the skin was additionally
riveted to the frames (Fig. 4), in the region of the windows.

* Previously issued as R.A.E. Report No. Structures 257--A.R.C. 22,270.

2.2. Local Structure at Windows and Escape Hatches. With the exceptions of the two forward

escape hatches, which interrupted a circumferential frame (Fig. 6), the windows and escape hatches

were positioned between the frames.
The windows and escape hatches were rectangular, their relative sizes being:

Window: 16.6 in. wide x 14 in. high, corner radii 3 in.

Escape hatch: 19.0 in. x 21.5 in. high, corner radii 4 in. (see Fig. 5).

The apertures were reinforced by peripheral members of zed section bonded to the skin with

Redux adhesive and additionally riveted by ~ in. countersunk-head rivets at the corners (Figs. 7

to 11).

3. Method of Test. Supported on its wing centre section in a tank, the fuselage was filled with
water and completely submerged in water, internal pressure being applied by pumping in more
water. The cycling action was controlled by pressure switches.

No loads other than those due to internal pressure were applied. The cycles were repeated
pressure cycles of the form given in Fig. 12; the peak pressure was 8.25 p.s.i, and the loading cycle
took about 65 seconds.

Detailed visual inspections were made at frequent intervals. When a fatigue crack was found its
subsequent development was observed continuously with an inverted periscope. When it was
judged that a fatigue crack would soon develop catastrophically the affected aperture was repaired
to prevent excessive dan:age to the specimen (Fig. 13).

By means of resistance-wire strain gauges, strains were measured at selected window and escape-
hatch corners before the fatigue test was started.

4. Results. 4.1. Measured Strains. Readings from strain gauges positioned at the corners of the
third starboard window and the forward port escape hatch were taken at increments of pressure.
Stresses were deduced for a pressure of 8.25 p.s.i. (Figs. 14 to 19). The highest stresses are as
given in Table 1.

4.2. Fatigue Test. A total of 11,319 pressure cycles of 0 to 8.25 p.s.i, to 0 were applied to the
fuselage. Fatigue cracks occurred in the skin at the corners of nine windows and two escape hatches,
sixteen corners being affected (Table 2 and Fig. 20). No fatigue cracks occurred at the A.D.F.
aerial hatches (at which the 22 s.w.g, reinforcing plates were subsequently removed for examination
of the skin underneath), the crew and passenger doors, or at the freight hatches.

The first crack was seen at 5,248 cycles at the third window on the port side, i.e., the window just
forward of the rear spar frame, and by 8,941 cycles all six windows in the centre section had fatigue
cracks at one or more of their corners.

Nine fatigue cracks were observed continuously throughout their growth; six at windows in the
centre section, one at a window in the aft section, and two at the port forward escape hatch. The
fatigue cracks originated at the rivet holes at the aperture corners, not at the aperture edges, and
when first seen were usually about 0.25 in. long. Development away from the aperture was initially
about 1 in. in 500 pressure cycles, and, as all the windows were located between frames, the growing
crack invariably had to cross a frame when approximately 4.5 in. long. When the cracks had spread
2 in. or so past the frames, i.e., were about 6.5 in. long, they were judged to be critical in that the
application of a few more cycles would cause a catastrophic failure.

2

Differences of behaviour occurred during growth towards the frames and across the frames.

During the first stage, several cracks became critical when about 3 in. long, i.e., between the aperture

and the frame. This condition was noted at most of the windows in the centre section where the
skin was also riveted to the frames. The presence of this extra attachment appeared to have a strong
influence in delaying crack growth across the frame, and was clearly demonstrated in the case
where a crack extended a distance of 2 in. in one pressure cycle (Fig. 22). An exception however,
occurred at the port forward escape hatch (where the riveted frame was a partial frame only--Fig. 6)
when the crack at the bottom forward corner caused a catastrophic failure when 2-75 in. long
(Fig. 6).

The growth of one crack was observed in detail at a window in the aft portion of the fuselage
where the skin was not riveted to the adjacent frame. No critical stage occurred at 3 in., the crack
grew uninterruptedly across the frame to a critical length of 7.1 in. This behaviour was also shown
in a crack at a window in the same section of G-ALYU a (Fig. 30).

In crossing the frames the cracks behaved in various ways:

(1) The frame with the normal bolted attachment appeared to have no influence whatever on
crack growth.

-(2) Where the skin was riveted to the frames:
(a) Cracks passing between rivet holes were slowed down, but not stopped altogether.

(b) Cracks entering rivet holes were stopped temporarily; e.g., one crack was contained

for more than 1,800 cycles.

Development beyond the frames progressed for about 2 in. when it was evident that catastrophic
failure was imminent. Generally it was possible to stop the test before the fast-running stage, but
four catastrophic failures did occur either because of misjudgment of rates of growth or of the
difficulty involved in observing more than two cracks at the same time.

Table 4 summarises the data on critical crack lengths. Curves of crack growth are plotted in
Figs. 21 to 29; photographs of typical cracks are given in Figs. 31 to 36 (Table 3).

5. Discussion. 5.1. Origins of the Fatigue Cracks. All the fatigue cracks originated at the counter-

sunk rivet holes in the skin at the window and escape hatch corners. Those cracks which eventually
became catastrophic started at outer-row rivet holes. The few cracks that originated at holes in the
inner row grew inwards to the edge of the aperture and did not become catastrophic. No cracks
originated at the edges of the apertures.

As indicated by the strain measurements, the stress at the corner of an~aperture attained its peak
value at the edge; at the outer row of rivet holes it was about 20,000 p.s.i, or perhaps half the stress
at the edge. The presence of a sharp-edged (countersunk) rivet hole in a high stress field might,
however, increase the stress locally, perhaps by a factor of 3, and, in addition, there would be a
certain amount of fretting action, so it is reasonable to expect fatigue cracks to be initiated at the
rivet holes.

5.2. Locations of the Fatigue Cracks. The test on G-ALYR showed that fatigue cracks were

initiated earliest and most numerously at the windows in the centre section, and though the first
failure in G-ALYU was at a forward escape hatch4, fatigue cracks had also formed at several windows
in the centre section by the time of this failure (Fig. 37).

That the fatigue cracks should occur first at the corners of the apertures is easily explainable in
that the general level of stress there is some two to three times that found elsewhere in the fuselage.

The tendency for cracks to occur first in the centre section may be explained by a combination of
three reasons. First, the average stress in the skin at the corners of the windows would appear to
be some 20 per cent greater than at the corners of the escape hatches, as is shown by the strain
measurements made on three different fuselages (Table 5); in this connection it should be noted that
the radius at the corner of a window is smaller, 3 in. compared with 4 in. for the escape hatch.
Second, it is possible that there were aggravating distortions in the centre section of the cabin due
to the reaction of the internal pressure by the floor instead of by a complete cylinder as elsewhere.
Third, the effect of previous service use should not be forgotten, in that the shear stresses from
the usual flight and ground loads are highest in this part of the fuselage.

5.3. Propagation of the Fatigue Cracks. Many of the cracks when first observed were about

0.25 in. long. This is relatively short, but it is pointed out that the conditions for observing cracks
during the test were exceptional as they were anticipated at the corners of the apertures and the
paint was removed for easy inspection. It is problematical whether under normal service conditions
the cracks would have been detected so early.

Even if it were certain that cracks of this length could be found easily, the curves show that a
crack length of 0.25 in. corresponds to about 90 per cent of the total life when no remedial action
is taken. Coupled with this fact is the indication that the rate of propagation is probably greater in
service than on test, since the crack measured at a window in G-ALYU, with its more representative
loading2, developed 2 to 6 times as fast as the cracks in G-ALYR.

The delaying effect of the adjacent riveted frame provided a temporary barrier when the cracks
were about 4 in. long. Nevertheless the opinion was formed that special inspection procedures would
have to be used to ensure reliable detection. In this connection inspection would be greatly eased if
the fuselage were partly pressurized to open up the cracks.

Comparison of these findings with observations from tests on flat sheets and simple cylinders
shows at once that the flat-sheet tests were unrealistic because of the absence of radial pressure.

There appears to be some measure of agreement between the window-corner cracks and those
induced in 12 ft diameter cylinders in that critical lengths and numbers of cycles to failure were of
the same order for a roughly comparable nominal stress cycle, but upon consideration of the differing
conditions between the window corners and in the simple unstiffened cylinders, such agreement
is perhaps fortuitous.

5.4. Interpretation of the Fatigue-Test Results. The meagre data make reliable analysis

impossible, but certain features need comment. First, the initial failure in G-ALYR occurred at
approximately twice the number of cycles as that in G-ALYU. Second, a closer grouping is evident
of the failures in G-ALYU compared with G-ALYR (Fig. 37). Third, rate of crack growth in
G-ALYU was about four times that in G-ALYR.

From this evidence it appears that fatigue performance is adversely affected to an appreciable
extent by other-than-pressure loads reaching the cabin, an effect already noted by Walker1,z. This
means that where accurate life estimates are required to be obtained from tests the general flying
and landing loads should be reproduced as faithfully as practical considerations allow. Furthermore,
since perfection in this respect is unlikely, an allowance must be made for inadequacies of
representation.

Y~

6. Concluding Remarks. This Report contains material from which various conclusions may
well be drawn, and especially if combined with later work. Three points, however, appear to be
established.

(i) The simplifications in fatigue loading which are generally accepted to make a full-scale test
practicable are likely to give a longer life than would be realised in service.

(ii) Tile attachment of reinforcing material inevitably introduces its own stress concentrations;
the example of the countersunk rivet holes at the window corners illustrates this important principle.

(iii) Nearly all the fatigue life associated with a particular crack may have been expended by the
time the crack first becomes noticeable.

No. Author REFERENCES
1 P. B. Walker Title, etc.
P. B. Walker
P. B. Walker Pressure cabin fatigue.
H.M.S.O... 5th International Conference, Los Angeles, 1955.

Pressure cabin fatigue.
Aircraft Engineering. Vol. XXVIII. January, 1956.

Static strength tests of a Comet I pressure cabin.
A.R.C. 18,359. December, 1955.

Civil Aircraft Accident.
Report of the Court of Inquiry--C.A.P. 127. December, 1955.

6

APPENDIX
The Materials Used in the Structure

. D.T.D.687A.--Clad, high-tensile aluminium-alloy sheet.

(i) Chemical composition (nominal)

Copper 0.4 per cent

Zinc 5.3 per cent

Magnesium 2.7 per cent

Manganese 0.5 per cent

Aluminium The remainder.

(ii) Heat treatment
Quenched after 2 to 10 hours at 455 to 465 deg C.
Aged 4 to 30 hours at 120 to 140 deg C, or appropriate to suit requirements.

(iii) Strength properties
(a) 0" 1 per cent proof stress: not less than 27 tons/sq in.
(b) Ultimate tensile stress: not less than 32 tons/sq in.

2. D.T.D.610

(i) Chemical composition

Copper Not less than 3 •5, nor more than 4.8 per cent

Iron Not more than 1.0 per cent

Silicon Not more than 1.5 per cent

Magnesium Not more than i. 0 per cent

Manganese Not more than 1.2 per cent

Titanium Not more than 0.3 per cent

Aluminium The remainder.

(ii) Heat treatment

Quenched from 500 to 510 deg C at 2 to 4 hours.

Aged 5 days at room temperature (W condition).

(iii) Strength properties

(a) 0.1 per cent proof stress: not less than 14 tons/sq in.

(b) Ultimate tensile stress: not less than 24 tons/sq in.

(c) Elongation: not less than 12 per cent on sheets up to a in. thick.

. D.T.D.546B.--Clad, high-tensile aluminium-alloy sheet.

(i) Chemical composition. Same as for D.T.D.610.
(ii) Heat treatment

Quenched from 500 to 510 deg C at 2 to 4 hours.
Aged at 155 to 205 deg C for an appropriate time. (WP condition.)
(iii) Strength properties
(a) 0.1 per cent proof stress: not less than 20 tons/sq in.
(b) Ultimate tensile stress: not less than 26 tons/sq in.
(c) Elongation: not less than 8 per cent for sheets thicker than 12 s.w.g.

7

TABLE 1

Highest Stresses Measured at the Edges of the Apertures
for an Internal Pressure of 8.25 p.s.i.--G-AL Y R

Third starboard window Forward port escape hatch

Aperture corner

Top rear Bottom rear Top rear Bottom rear

Stress obtained by extrapolating to 47,700 35,500 40,000 32,700
aperture edge, p.s.i. 32,800 33,250 27,500

Highest 'strain-gauge' stress,, p.s.i. 40,800
(see Figs. 14-17)

Rear aerial hatch

Port forward Starboard rear

Stress obtained by extrapolating to 33,000 34,400
aperture edge, p.s.i. 30,450

Highest 'strain-gauge' stress, p.s.i. 32,850
(see Figs. 18-19)

8

TABLE 2

Chronological Occurrence of the Fatigue Cracks at the Aperture Corners

Aperture Crack
length
Windows I Side Corner Origin of I Pressure O00 0a /
between of of fatigue when cycles . , ~ 0 Oc
. spar frames Other Escape crack first seen
(Centre section) fuselage aperture ~ '0 0
rivet hole (inches)
windows hatch ©

3rd Port Bottom A 0" 20 5,248 O
1st Port forward O
1st Starboard B 0- 50 6,542
• 3rd Starboard BottOm - OF
2nd Port rear C 3 "409 6,90i
Starboard °ou
Top D 0" 17" 6,901
forward Rivet ,holes .at which the ,cracks .occur.red.
C 0" 06 6,901
Bottom
6th forward E 0-14 ~ 6,901
(Rivet hole
Bottom oversize)
rear

Top
rear

2nd Port Top C 0.10 6,959
forward

3rd Starboard Top C 1" 62 7,692
rear

2nd Port Top A 0" 04 8,564
rear

Forward "Port Bott6m B 0- 08 8,564
forward

4th Port Top A ,0- 51 8,941
rear

4th Port Bottom F 0" 70 9,225
forvcard

6th Port Bottom A 0" 06 9,350
forward

2nd Starboard Top A ,0- 7i ;10,016
forward

]Forward Port Bottom A 0- 31 10,016
l%rward rear
'(2 '0-t0 1t,286
Starboard l Top

rear

* Between rivet hole and edge of aperture.
]" Crack had spread to frame before it was discovered.

(83824) B

TABLE 3
Growths of the Fatigue Cracks

Location of the fatigue cracks When first seen When the crack had reached the adjacent frame Final details

Window Crack Additional Did crack Growth Delaying Additional Growth Total Remarks and action taken
length cycles spread into in inches effect pressure
Escape Corner Side Crack Pressure per cycle Length cycles in inches
hatch le(ning.t)h cycles (in.) fromwhen a frame of frame (in.) from when per cycle cycles
first seen rivet hole? in cycles first seen
Between Outside

spars spars

3rd Bottom Port 0" 20 5,248 4"40 700 No 0.009 90 6.25~ 794 0-10 6;042 Catastrophic failure imminent.
1st forward Port 0" 50 6,542 6.75~ 417' 0.06 6,959 Aperture repaired.
1st Starboard 3" 40" 6,901 4"70 359 No 0.021 60 6.41~ 2,040 0" 05 8,941
Bottorh Catastrophic failure imminent.
6th rear 6,901 3 "40 0, had Yes Stopped >1,820 0'03 Aperture repaired.
3rd 7,692 3"15 reached at rivet 240
Top adjacent Catastrophic failure imminent.
forward frame when hole Aperture repaired.
first seen

Top Starboard 0" 14 No, frame About 2,040 8,941 Catastrophic failure requiring
rear not 12 873 8,564 major repair involving three star-
feet board apertures.
riveted
5.85~ Catastrophic failure imminent.
Top Starboard 1" 62 168 Yes Stopped Aperture repaired.
at rivet
rear.
hole

2nd Top Port 0.10 6,959 3"35 1,501 Yes Stopped 890 15"0 2,449 Instant- 9,350- Catastrophic failure, only stopped
forward at rivet aneous by the patch around neighbouring
failure
hole window. Major repair.

Forward Bottom Port 0"08 8,564 i About 2,682 .Instant- 11,246 Catastrophic failure from front
forward aneous spar frame (18) to between
Crack spread catastrophically before reaching the cir- 15 failure
cumferential frame--was about 2"75 in. long when feet frames 8 and 9. Major repair.

failure occurred.

4th Top Port 0" 16 8,941 3' 40 669 No, frame No discontinuous kink 7'10 725 Instant- 9,666 Catastrophic failure in 9,666th

rear - not observable in crack aneous cycle running from 7.10 in. to

riveted growth curve failure about 12 feet. Major repair.

2nd Top Starboard 0"75 10,016 Crack grew to a length of 3' 50 in., but was not allowed to 3 '50 229 10,245 Likely that crack would have
forward reach the adjacent frame, as the aperture was reinforced. grown further•

Forward Bottom Port 0"31 10,016 3 "35 1,189 No 0.1 When this crack had reached the circumferential frame at 11,205 cycles, a reinforcing strap was
rear appro-% riveted close to the frame to prevent further crack progress but at 11,246 cycles the crack at
the forward corner caused a catastrophic failure.
i

* See Figure 25. Approximate critical length.

10

TABLE 4

Critical Crack Lengths

Aperture Location Critical crack length Critical crack length
between the aperture attained during final
and the adjacent frame
development
(in.) (in.)

Port escape hatch (Bottom for- Forward section 2.75--at this length the crack caused a
ward corner) catastrophic failure

Port escape hatch (Bottom rear Forward section 2.80 No result here because of

corner) the occurrence ofthe above

failure

First window, port side Centre section 2.45 6"75

First window, starboard side Centre section Crack had spread to the 6"41
adjacent frame before any
measurements were taken

Second window, port side Centre section No observable critical No final length as this

stage crack spread catastrophi-

cally from the frame

Second window, starboard side Centre section No observable critical This aperture was repaired

stage before crack had spread to

the adjacent frame

Third window, port side Centre section 3.10 6"25

Third window, starboard side Centre section No intermediate critical 5'85
stage occurred

Fourth window, port side Aft section No intermediate critical 7"10
stage occurred

Sixth window, port side Aft section No intermediate critical 5"60
e (G-ALYU) stage occurred

A previous fatigue test in which wing loads were applied as well as pressure loads.

11

(88824) C

TABLE 5

Measured Stresses at the Edges of Apertures in Three
Fuselages, for an bzternal Pressure of 8" 25 p.s.i.

Aerial hatch* Port forward Starboard forward Third window
escape hatch escape hatch

Fuselage

Strain Extra- Strain Extra- Strain Extra- Strain Extra-
gauge polated gauge polated gauge polated gauge polated

G-ALYU 28,000 32,200 23,000 t 26,500~ 34,500 No 35,100 38,000
G-ALYR 32,850 34,400 33,250 reading 40,800 47,700
G=ANAV 30,400 34,000 34,000 39,600 No 38,300 43,500
38,800 reading No
reading
31,600
36,300

* The strain gauges attached at this aperture were cemented to the 22 s.w.g, reinforcing plate riveted over
the skin.

t" These stresses were measured after the escape hatch had been repaired following a catastrophic failure
and may differ, therefore, from those that would have been measured in the original structure.

12

ATTACHMENT TO WIN~

AT F'RAME6 18 AND P-6

AERIAl HATCHES

I0 II 14, I',5 I 7 IS 18 25

C.3

C~EW ~OOR

F

FIC. 1. The f

/
WINDOW8 ~.3.4.5 AND "7

5 86 27 ~ ~9 31 32 33 34- 35 36 3"7 3S

FREIGHT HATCHES
fuselage of Comet G-ALYR.

,......o..,,.,!20SW.Ct. O.T.D.610. IL ~Os.w.cq. ID.T.O.687A.

FRAME RESTS ON =uC
THE STRINGER O3
FLANGES,
ELSEWHERE ~'X'REDUX 3-OIINT.
THERE IS

----.;WLCLEARANCE.B' ..~;

S~KIN

Fro. 2. Typical frame and stringer sections.

/ SPAR F R A M E 18 / SPAR FRAME ~6

OF 5KIN JOINT. C~5 ~6

__\ 13 ~ 5 ~ C]'

~ "~¢:... OF ,SKIN ...TOINT.

THIS SKIN PANEL ~05.W.G. D.T.D. 546.
FIG. 3. Skin panel containing windows and escape hatches.

14

ge

(No.:,,)

ledio~
I
ment
bolted at
r fkmge$

hat'

r

Fzc. 4. Attachment of the skin to the frames.

Escape hatch frame Skin removed
Window frame

FIQ. 5. Relative sizes of window and escape-hatch apertures.

15

(83824) D"

. . ..

Circumferential fram~

FIG. 6. Inter
Failure at p

16 S.W.G . - U-section
partial frames

rnal partial frames at forward escape hatches.
port forward escape hatch at 11,246 Cycles.

LOCATION OF SECTIONS

~SKIN I II

.................... / ~ DIA RIVET C'SUNK -90°

/ . . . . . . APERTURE FRAME
i6s.w.¢- ~.tD. olo)
~EBUXED TO THE ~ I N

STRIP REDUXE m
TO THE FRAME

m

.L

FIG. 7. SECTION THROUGH A WINDOW FRAME.

/SKIN /~"'OIA. ~wT, C'SUNKso°
A ~. A P E R T U R E FRAME

] /.~ SEAI:IN~5TRIP REDUXED

r

FIG.& SECTION THROUGH AN ESCAPE HATCH FRAME.
Fits. 7 and 8. Sections through window and escape-hatch frames.

17

4 B.A. BOLT AT 5TR. FLANGE. EDGE OF REDUX JOINT. 4 B.A. BOLT AT
STRINGER FLANFaE.
EDGE OF REDUX. ~TOINT \ O0 " " -..
la_
- -/- - -O- - -O- - ul \
0 0 o-O." < \ J
h_ z
J

b-

Z

OO ,, o 0\\ <
/
06 \ THIS ADDITIONAL
p ADDITIONAL RIVETED ATTACHMENT AT
Qi ~L~ ATTACHMENT AT o THE FORWARD

I WINDOWS P-~3 AND ol ESCAPE HATCHES
I
| 4 ONLY. I ONLY.
DT" 0I

~-0"

WINDOW CORNER ESCAPE HATCH CORNER

J_ ii

RIVETS ARE B DIA. CSK, HD.

FIG. 9. External structural details at window and escape-hatch-aperture corners.

" " °-: .
FIc. 10. Local structure at

Fuselage skin Circumferential frame

h

Redux/
joint
between
aperture
frame
and skin

frame

Section X-X
a window corner, viewed from outside with skin removed.

O.f -S Q
:i I . . ~
Q'.

f
/

-- / /
//

Outei- row of
rivet holes

k--~nner row of

~ rivet holes

Inc
FIG. 11. Local structure at an esc

R4zdux joint between a p e r t u r e

I. . . .

0

2
3
-4
-5
6
ches

5~ction X-)
cape-hatch corner, viewed from outside with skin removed.

LEAKAGE OF FU51ELA~E

8 PRESSURE RELEASE

t ZO 4 0 60 80 I00
~-
I
o
Z
6-
O
-..>.

_i

n, 4
D

if)
If)

M
nt
O_

l.d

.J
ILl

LO

h

0
0

TIME IN 5EC01ND5

FIG. 12. The pressure cycle.

21

18

tO
ix)

£,.!;

,~,.,& - o,- ° o ."

I . . . .: ~ * .~ .

S.W.G. P late 16 S.W.G.Plate

" Or

-

',.rid

. . -~- -°

,~ x/.. - ',.at

FIG. 13. Typical window repair.

INNER EDqE OF AFT
WINDOW FRANE
O- --0
0 00

~z a,

I © so,oDD
47,700 EXTRAPOLA

¢, I0 4O)OO0 IqHEST 'STIll GAU~E'3TRE~

Z~'3oo,ao~ i

z_

l0
NX~-- - - INNER Er OF
WINI~OW FRAME

O I 2.
RADIAL DISTANCE FRQP1

EDgE- INCHE..S

FIG. 14. Stresses in the skin at the top corner of the thir
starboard window for an internal pressure of 8-25 p.s.i.

oo

w Noow AME

~S 4o,ooo I i --...~ o o

- Ore--

EXTRAPOLATED

EDGE STRESS.

II AFT
"~
__ HIGH~'ST "STRAIN

30,OOO ~--GAUGE" 5TRESS.

O2

] \'~
z

-- \ ' t

~uJ-IO~OOO \ \ [

~ \ I . / I N N E R EDGE OF

M" I WINDOW FRAME

O ]\\1
RADIAL DISTANCE FROM
EDGE -INCHES.

d FIG. 15. Stresses in the skin at the bottom corner of the
third starboard window for an internal pressure of 8.25 p.s.i.

Ag'T >

0~ ~ /.INNER EDt~EOF ESCAPE
HATCH FRAME
40,000
OO
~oo EXTRAPOLT!ED
% \ \ Eo~E ~TRESS 0

30;ooo ~H,OHES.'~'T.A,. ;

\ \~0A°OE' S,~E,E.

20,000~ .

10,OOO ~ . ~ ] ~ ' ~ j . NN-i - ~ENER EDr4E 0F ]tu" ('~I

~ 2 ~ ~ E HATCH FRAME. / q" "Ji
i
4OtOOO - - EXTRAPOLATED ~0 C)
EDgE .STRESS I :~
/ /
co 0 RAO,AL DISTAINCE.FROM /~o0 ~ II
,I iob~ EDrqE- INCHES.
---HIqHEST ..~TRAIN~AUr~E'
oo/
8o,o o o STREE6 0

i O

.z_ "--_____.
~0~000
0 0O
z 0

I0,000 -- - - -

j l N N E f EO~EOFESCAPE 0 ~"~.
HATCH FRAME
~
OI a INNER EIDGE OF

RADIAL DkSTANCEFROM ESCAPE HATCH FRAME.

EDgE- INCHES AFT

FIC. 16. Stresses in the skin at the top corner of the FIG. 17. Stresses in the skin at the bottom corner of the
forward port escape hatch for an internal pressure of forward port escape hatch for an internal pressure of

8.25 p.s.i. 8.25 p.s.i.

J~ S,?'50

INNEREDgEOFAERIAL
HATCH FRAME
&oooO

/© © , ©© ©

O © O 0.00

" ~ 0 O 4- ~,~oo

/o/0OO// \, \

/oO \

IoO O

b~

4O~OOO

~/ ~XTRAPO-, EDGE ~TRESS

30)000 Y / A

20)000 ~ ~%

io,o o o

INNE~ E£)~EOF AERtAL
-- HATCHFRAME

O
RAOIALDISTANCEF. ROM
EO(~E- INCHES
FIG. 18. Stresses in the skin at the port forward
corner of the rear aerial hatch.

°°°°°~1 I
34,400 EXTRAPOLATED
EDGESTRESS.
a: 3O'ooo

,

Z

~ ~O~DOO

IO~OOO G
\ \ ,-j ] O/
/C) /
oe
RADIAL DISTANCEFROM
EDGE-INCHES.

- - ~ _ + ~ - © ".'~'-~ ~F-"/, ~ AERIALHATCH

\ FRAME.

° °-L°~ 0/"
0 O ~ "~ ,?

FIG. 19. Stresses in the skin at the starboard rear
corner of the rear aerial hatch.

FRONT SPAR j CENTRE SECTION REAR 5PAR
FRAME ~6
FRAME 18 ~ - - ' -

10~\016 i8,941 OI

E.H. 10~245 8~,564 7:'i9~ 8~94-1 8,941
/
I 8,941 6,901 STARBOARD SIDE

I

[bO ESCAPE HATCH
E.H.
6,9\59 ./9,666 1941' WINDOW

11~'46 6~959 9,350 8164 J 9;~?_5 9~666 /J 97666 97666
6~54? 6~04E
8,564 10~016 9~350 PORT SIDE
6~901 5~74-8

NOTES:-
(I) NUMBERS AT CORNERSARE THE PRESSURE CYCLES

WHEN THE FATIGUE CRACKS WERE FIRST OB3ERVED

(2) NUMBER5 INSIDE ARE THE PRESSURE CYCLES
WHEN THE APERTURE WAS REPAIRED OR REINFORCED
AND THEREFORE ELIMINATED FROM THE TEST.

FIC. 20. Distribution of fatigue cracks at window and escape-hatch-aperture corners.

v • °/ [ FORWA
b--.-)I -7
!o °"o\ / /
© 0~"

SKINFATI~LJE CRACK A /'/

O

8,~00 PRESSURE CYCLES - -

/ <6,000

5,400

!I

4 3 ~- , g,aO0

o
DISTANCE FROPI CENTREL-INEOF RIVET'A t IN INCNES

FIG. 21. Crack growth at third window, port side.
(Bottom forward corner)

ARD

7

>O @,

\/0 0 O~

~ ' ~.~~--_~_.____ - - ~ ~ ~ o'

O

io PRESSURE CYCLES f

,,,f

~6,E

I

- - 6,400

3

DISTANCE FROM CENTRE- LINE OF RIVET ~J~--INCHE5

FIG. 22. Crack growth at first window, port side.
(Bottom rear corner)

I

O e~

~) O~ 0 ~ ( ~w O

: :::o; % ,~ "\ B IO
/ A T 8,478 CYCLES
O O
+ AILED AT 9;350CYCLE3. kJ / ;-O' RWARD O
A
CRACK OEVELOPEDI 9,500
97000 FINAL LENCTTH OF CRACK: 3"50
"]--~ CATA5TrO FHICALLY (WINDOW REPAIREOAT THI.S~TA~E)
AT 9~350 CYCLE& --IO~;

, CRACK REAC"EO d
I RIVET HOLE , ~"
bO BISO0 0
CO

7~500 --Io~ooo

i°° DISTANCE FROM CENTRE- LINE OF RIVET ' A ~
INCHES
G~500
I FIG. 24. Crack growth at second
DISTANCE FROM CENTRE - LINE OF" window, starboard side. (Top for-
RIVET 'C *- iNCHES.
ward corner)
FIG. 23. Crack growth at second
window, port side. (Top forward

corner)

O5 o1 ©

1 ~- ©

' .<

FORWARD DLENGTH = G'41"
IT~ocR!CAL

/ol

° lba

" ' 8,500 o

~J <
8~000
7,500 o

o

<

__-2°1°o

l~ LEdtEl"i~H OF CRACK WHEN FIRST 5EEPt =3"40".

/ 4

I T3
O]$TANCE FROM CENTRE - LINE OF RIVET 'C' - iNCHES.

FIo. 25. Crack growth at first window, starboard side.
(Top forward corner)

//____<>FATIG~UECRACK IN 5KIN AQ O
O
CI~ITICAL LENC-I'I'H = S ' 8 5 "

©

©

®

I

I 600 PRESSURECYCLES
8,L O0
\
00--
\ O0
O0
\\

"4-.

,q- 3 7,6o0
~. I

DISTANCE FROM CENTRE-LINE OF RIVET C IN INICHES

FIG. 26. Crack growth at third window, starboard
side. (Top rear corner)

O FORWARD

% AFT /O o
Oc
O •

O

O ,O

© Q B0 \_ d xx
, ~D--n

-- __ ~ ~ CATASTROPHIC FAILURE I

"~----~-.-- TOOK PLACE" A L O N E / ~ \
// ~ THIS PATH /

~RITICAL LENGTH ~.75"APPIOX.

CRACK i0 0 " ~

AT 1l,~46 CYCLES.I "-"'~""k i, o

1
)7i00- -

9,000 ~ - -

4 3 ~: 8~500--

DISTANCE FROM CENTRE-LINE OF RIVET 'B' -- INCHES. t O~

FIG. 27. Crack growth at forward escape DISTANCE F R O M #E.NTRE- LIN~" OF RIVET ~A~- INCHES
hatch, port side. (Bottom forward corner)
FIG. 28. Crack growth at forward escape
hatch, port side. (Bottom rear corner)

A AFT i

0c SKIN F
CRA

AL LENGTH = 7"10"

9,700

9,6Q0 .jl fJ

rJ / A CATASTROPHIC
u OCCURRED AT TH
PRESSURE CYCLE
o9,SOO ~"

M

9,400

I

9'TO0 I
9~200

4

OFDISTANCE FROM CENTRE- LINE RIVET

FIe. 29. Crack growth at fourth window, port
rear corner)

FATI6UE ()
ACK.
o
JCYCLE 9,66
I <

FAILI ()
HE 9~G66~.
E. SKIN FATIGUE CRACK

I CF, TICAL LENGTH = 5"6"
7--
I

56
'C'. - INCHES.

side. (Top

Forward

• ]:

/':. ! •f,4,-xl

iJ

~. ~i~,~~. ?
•% , ,,

.~.~,?~ 2 .~:~

Fzo. 31. Third port window corner, 5,951 cycles.

Crock running -- Forward
between rivet holes

\

FIG. 32. Third port window corner, 6,042 cycles.
32

- - Forword
• ..;~" ..¢j

FIc. 33. First port window corner, 6,901 cycles.

Forword

w

•' • o

•N

oE

Lit. ¸

I '"
B °• •

o

FIG. 34. First port window corner, 6,959 cycles.
33

ForwQrd = Crack stopped at rivet hole

e, C

Fxc. 35. Crack stopped at rivet hole in frame, 6,901 cycles.
First starboard window.

Forword

D
i,m

~Q

LL.,-
u

• * -4" 4 u , ..

FIG. 36. Cracks stopped at riveted frames, 8,564 cycles.
Third starboard window.
34

.q

ESCAPE HATCHES + +

WINDOWS IN AFT SECTION +
-t-
WINDOWS IN CENTRE SECTION + "1- --th- e
t"

0 IIII I

I~0 0 0 2~000 8)000 4",000. S) 000,

FIG. 37. Distribution o

SERVICE HI.STORY

I ~- ALYU -- IP_31 FLIGHTS
r~- ALYR -- -/4"7 FLICxHTS

[EACH POINT REPRESENT5 THE
FIRST FATIGUE CRACK DETECTED
AT A PARTICULAI~ APERTU/~E ]

+
+

i- • @@

+•
w

KEY J + = i~- ALYU

L e = CT- ALYR

6~ 000 I I I I I
"7, 000 8~000
,9,000 I0,000 I1~000

PRESSUt~E CYCLE8

of fatigue cracks at apertures.

Publications of the
Aeronautical Research Council

ANNUAL TECHNICAL REPORTS OF THE AERONAUTICAL
RESEARCH COUNCIL (BOUND VOLUMES)

I94z Vol. I. Aero and Hydrodynamics, Aerofoils, Airscrews, Engines. 75s. (post 2s. 9d.)
Vol. II. Noise, Parachutes, Stability and Control, Structures, Vibration, Wind Tunnels. 47s. 6d. (post 2s. 3d.)

I943 Vol. I. Aerodynamics, Aerofoils, Airscrews. 8os. (post 2s. 6d.)
Vol.
IL Engines, Flutter, Materials, Parachutes, Performance, Stability and Control, Structures.
z944 Vol. 9os. (post 2s. 9d.)
Vol.
I. Aero and Hydrodynamics, Aerofoils, Aircraft, Airscrews, Controls. 84s. (post 3s.)

II. Flutter and Vibration, Materials, Miscellaneous, Navigation, Parachutes, Performance, Plates and
Panels, Stability, Structures, Test Equipment, Wind Tunnels. 84s. (post 3s.)

I945 Vol. I. Aero and Hydrodynamics, Aerofoils. t3os. (post 3s. 6d.)
Vol. II. Aircraft, Airscrews, Controls. I3os. (post 3s. 6d.)

VoL III. Flutter and Vibration, Instruments, Miscellaneous, Parachutes, Plates and Panels, Propulsion.
I3OS. (post 3s. 3d.)

Vol. IV. Stability, Structures, Wind Tunnels, Wind Tunnel Technique. x3os. (post 3s. 3d.)

1946 Vol. I. Accidents, Aerodynamics, Aerofoils and Hydrofoils. z68s. (post 3s. 9d.)

Vol. II. Airscrews, Cabin Cooling, Chemical Hazards, Controls, Flames, Flutter, Helicopters, Instruments and
Instrumentation, Interference, Jets, Miscellaneous, Parachutes. I68s. (post 3s. 3d.)

VoL III. Performance, Propulsion, Seaplanes, Stability, Structures, Wind Tunnels. i68s. (post 3s. 6d.)

t947 Vol. I. Aerodynamics, Aerofoils, Aircraft. i68s. (post 3s. 9d.)

Vol. II. Airscrews and Rotors, Controls, Flutter, Materials, Miscellaneous, Parachutes, Propulsion, Seaplanes,
Stability, Structures, Take-off and Landing. 'I68S. (post 3s. 9d.)

I948 Vol. L Aerodynamics, Aerofoils, Aircraft, Airscrews, Controls, Flutter and Vibration, Helicopters, Instruments,
Propulsion, Seaplane, Stability, Structures, Wind Tunnels. z3os. (post 3s. 3d.)

Vol. II. Aerodynamics, Aerofoils, Aircraft, Airscrews, Controls, Flutter and Vibration, Helicopters, Instruments,
Propulsion, Seaplane, Stability, Structures, Wind Tunnels. xxos. (post 3s. 3d.)

Special Volumes

Vol. I. Aero and Hydrodynamics, Aerofoils, Controls, Flutter, Kites, Parachutes, Performance, Propulsion,
Stability. x26s.(post 3s.)

Vol. II. Aero and Hydrodynamics, Aerofoils, Airscrews, Controls, Flutter, Materials, Miscellaneous, Parachutes,
Propulsion, Stability, Structures. x47s.(post 3s.)

Vol. III. Aero and Hydrodynamics, Aerofoils, Airscrews, Controls, Flutter, Kites, Miscellaneous, Parachutes,
Propulsion, Seaplanes, Stability, Structures, Test Equipment. x89s. (post 3s. 9d.)

Reviews of the Aeronautical Research Council

I939-48 3s. (post 6d.) 1949-54 5s. (post 5d-)

Index to all Reports and Memoranda published in the Annual Technica! Reports

19o9-I947 R. & M. 2600 (out of print)

Indexes to the Reports and Memoranda Of the Aeronautical Research Council

Between Nos. 235z-2449 R. & M. No. 2450 2s. (post 3d.)
Between Nos. 2451-2549 2s. 6d. (post 3d.)
Between Nos. 2551-2649 R . & M . No. 255° 2s. 6d. (post 3d.)
Between Nos. 2651-2749 R. &M~ No. 265° 2s. 6d. (post 3d.)
Between Nos. 275t-2849 2s. 6d. (post 3d.)
Between Nos. 285 x-2949 R . & M . No. 275o 3s. (post 3d-)
Between Nos. 295x-3o49 3s. 6d. (post 3d.)
Between Nos. 3o51-3x49 R . & M . No. 2850 3s. 6d. (post 3d.)

R . & M . No. 295°

R . & M . No. 3050

R: &M. No. 315o

HER MAJESTY'S STATIONERY OFFICE

from the addresses overleaf

~o & M o No. 3245

.'q

~ Crowncopyright x96z

Printed and published by
HER MAJESTY'S STATIONERY OFFICE

To be purchased from
York House, Kingsway, London w.c.z

4z3 Oxford Street, London w.I
I3A Castle Street, Edinburgh z

xo9 St. Mary Street, Cardiff
39 King Street, Manchester z

5° Fairfax Street, Bristol i
35 Smallbrook, tRingway, Birmingham 5

8o Chichester Street, Bdfast
or through any bookseller

Printed in England

~o & M o N~o 3248

S.O. Code No. z3-3z48