Document
FATIGUE TESTS ON BIG STRUCTURE ASSEMBLIES OF CONCORDE AIRCRAFT By V. P. N'Guyen Soci_t_ Nationaie IndustrielleA6rospatiale Toulouse, France and J. P. Perrais Centre d'Essais A_ronautiques Toulouse, France INTRODUCTION The Concorde, a delta-shaped-wing aircraft, has been submitted to numerous mate- rial, attachment and protection tests since, with its structural design, it is capable of reaching supersonic speeds (Mach number, 2.05). In addition, this aircraft has been tested in the scope of structural engineering tests performed on substructures. In this paper, only development tests on large structure assemblies and airworthiness substan- tiation full-scale tests are considered.
This paper is limited to the tests performed at the Centre d'Essais A6ronautiques of Toulouse (C.E.A.T.), France. The tests carried out in the United Kingdom are to be presented by the Royal Aircraft Establishment (R.A.E.). As a rule, the development tests achieved both in France and in the United Kingdom are usually performed on struc- tures for which A6rospatiale and British Aircraft Corporation are responsible. All certi- ficationstatictests are to be carried out in France and allcertification fatigue tests are to be performed in the United Kingdom.
EXPERIENCE FROM STATIC TESTS Two main sections have been submitted to pressure, mechanical load, and thermal static tests and are shown in figure 1.
/ I Fuselage Section 1 his.
The structure, named fuselage section 1 bis. or l(a), consisted of a 4.68-meter-long twin-looped cylindrical fuselage section including six standard frames and two main frames. On both sides of the lower part of the fuselage, rectangular structural boxes represented the wing assembly and its fuselage junction section. The purpose of this operation was to create the same thermal stresses over this area as those encountered I in flight. The skin panels (A-U2GN sheet) were attached in a classical way to the stringers and frames.
The aim of the tests was to observe the structural behaviour under the most severe flight conditions such as combined pressurization, fuselage torsion and loads on floor, and thermal stresses. Test measurements of temperatures and mechanical strains were also compared with calculated values of thermal stresses in order to (1) justify design methods, (2) make an analysis of the role played by thermal stresses among total stresses (to manage a test program of structures which will be tested in the future), and (3) perfect new test methods, especially in the scope of infrared heating and air-cooling units injecting liquid nitrogen. The tests started at the end of 1964 and ended in the spring of 1966.
This testing enabled the manufacturer to check for the thermal stress level in the fuselage areas hidden by the wing assembly and in the longitudinal stringers located at the bottom of the fuselage. (Fig. 2 shows the results of comparative tests on the heated lower part and the unheated lower part to simulate the presence of a fuel tank.) It was necessary to carry out tests, especially fatigue tests, by representing in a most accurate way thermal stresses where they are significant.
Section 2.8.b The test structure, section 2.8.b, was composed of a fuselage section (first defini- tion of the aircraft, 10 m = 35 ft long) and of main adjacent wing elements having an over° all span of 44 ft. (Refer to fig. 1.) This structure is a genuine aircraft element. The purpose of the test was (1) To check in a more exact way the aircraft design methods. Therefore, the test structure itself with its proposed end effects has been calculated by means of the same network as an aircraft (analog electrical network for internal load computation).
(2) To compare thermal stress distributions obtained from different aircraft mis- sions. These distributions are not easily obtained by computation.
(3) To evaluate fuel influence in the tanks on these thermal stresses.
(4) To study the superimposition of cabin and tank pressure, of air and inertia loads, and thermal effects.
(5) To prove the "fail-safe" characteristics of this structure by making some cuts to simulate cracks in the main spars, ribs, and frames, and then performing residual- strength tests.
(6) To familiarize test laboratories with exceedingly complex installations in order to proceed with the certification static tests on a full-scale aircraft structure (fig. 3) under satisfactory conditions.
These tests commenced in the autumn of 1966 and ended in the summer of 1969.
Results are too extensive to be presented in this paper. Therefore, only tests which made it possible to perfect the fatigue test programs are presented.
It was shown by the design calculations that the maximum thermal stress values highly depended upon the aircraft acceleration laws. This dependence was verified when a few wing panels buckled locally during tests simulating missions with high acceleration and low take-off weight. (See fig. 4.) (It was a case of a flight corresponding to a pre- vious definition of the aircraft.) The purely thermal stresses remain moderate in abso- lute value but are reversed, and their peak-to-peak values are significant. The presence of fuel causes the stresses in heavy parts of spars and ribs to be reduced. On the other hand, the internal skin surface is subjected to tensile thermal stresses when the fuel tank is empty. These tensile stresses add to the internal tensile stresses due to flight loads. The following conclusion may be drawn from this program. For tests on partial structures, great care should be exercised in simulating the temperature distributions over the fuselage internal areas (especially those areas hidden by wing assemblies). (The parasite end effects are very strong.)
Because of the high strength of the fuselage in the presence of large cuts (as required in the FAA fall-safe tests), fatigue tests can be safely conducted by using air to cyclically pressurize the fuselage.
A few "dynamic-cut" tests which were performed on the fuselage throughout frames ended the fail-safe tests; the data from these tests will be used for certification substantiation.
STATIC TESTS FOR AIRWORTHINESS SUBSTANTIATION The test structure is a full-scale aircraft. The test program consists of a sequence of tests to be performed under room-temperature conditions and including five different tests with loads on a part of the aircraft. All tests were conducted at least up to ultimate design load of the structure and some of them even beyond. The latter sequence of tests will be made under thermal conditions about July 1971 and will start with thermal tests only, during which several aircraft missions will be achieved under realistic conditions. In a first stage, to investigate ovens and cooling problems, C.E.A.T.
will use calculated temperatures which are being verified by means of flight measure- ments on the prototype. The test temperatures will be submitted to the Airworthiness Authorities for approval. Figures 5 and 6 illustrate different static-test sequences.
FATIGUE TESTS
These tests have beenperformed on many structural components,but the test pro-
grams achievedby use of big substructures 2.3.2 and 2.6/2.7 (fig. 1) are by far the most
significant.
Preliminary static tests showedthat it was necessaryto reproduce the temperature
distributions during acceleration and deceleration sequences. Whenthe fatigue test pro-
grams were initiated, it was foundthat this operation would require a test of long dura-
tion; the time cycle in the laboratory was almost equal to the time required for an actual
flight. It was absolutely necessary to compromise some part of the test program in
order to obtain some desired results for the structural behaviour within a reasonable
period of time.
Two changeswere madein the test program to compensatefor accelerating the
thermal tests: (1) To compensatefor creep, normal structural temperature has been
increased by 20° C (from 100 ° C to 120 ° C), (2) To compensatefor deteriorations dueto
thermal stresses, the heating rate d0/dt has beenincreased during acceleration and
deceleration sequencesin order to increase the stresses by 15to 20 percent, depending
upon particular components.
In order to accelerate testing, the time during which the external wall temperatures
were constantwas decreased. Figure 7 showsthat this decreasewas feasible since
(a) the same maximum temperatures were achievedas in actual flight for both external
wall andinternal structure, (b) the wall and structure returned to room temperature at
the endof the programed time cycle, and (c) the heating sequenceduring the time of con-
stant temperature producedsatisfactory thermal gradients during the deceleration
sequence.
On the test section 2.3.2, this requirement was met by blowing hot or cold air onto
fuselage areas hidden by the wing assembly. Ontest section 2.6/2.7, the same result
was obtainedby injecting hot and cold liquid into the fuel tanks, as required. These pro-
cedures are called "complementary means."
Determination of Cycle
Randommaneuverandgust loads were applied by lever jigs. For these develop-
ment tests to be performed, it was preferable to reduce the typical loading spectrum to
its simplest terms to investigate more easily the possible crack propagation rates. Pres-
sure loads, since they are actually known,havebeen usedat their flight true values; that
is, p = 736 mb inside the cabin compartment, and p = 250 mb inside the fuel tanks.
Thermal stresses were increased 10 to 20 percent, depending upon the area, to accelerate the observance of the deteriorations due to thermal stresses. By using this increase, an attempt was made to double the damage value due to thermal stresses.
Three mechanical and three pressure cycles were superposed on each thermal stress cycle. In one instance (A), the mechanical and the pressure cycles were applied simultaneously while the thermal stresses were high. In two other instances (2B), the mechanical and pressure cycles were applied simultaneously while the thermal stresses were small or nil (corresponding to a slow return to room temperature). This sequence of loading produced a threefold increase in damage due to the usual loads. Cycles C = A + 2B are performed one after the other.
Final Test Conditions Final test conditions were based on and perfected from typical tests. During these typical tests, the actual flight real time requirements were met in order to accurately determine the required heating rates and thermal stresses during a flight. Based on the results of these typical tests, several short time cycles were tested and complementary means were used to obtain the desired temperature and stress evolution (especially peak-to-peak) at all significant measurement points. The complete time cycle of test 2.3.2 is shown in figure 8; whereas the complete time cycle of test 2.6/2.7 is shown in figure 9. It is easily noticed that with 1 hour's cycle (of which 40 minutes is thermal) for 2.6/2.7 tests and that with a 34 minutes' cycle (of which 26 minutes is thermal) twice the thermal damage and three times the mechanical damage of a 3 hr 15 min flight is produced.
Results Obtained on Test Structure 2.6/2.7 By March 10, 1971, 9900 cycles (A + 2B) and 10 900 additional B cycles (repre- senting purely subsonic flights) were applied. This stress history corresponds to the damage caused by 40 600 flights under mechanical fatigue conditions and about 19 800 flights under thermal fatigue conditions. The deteriorations that were noticed occurred on the (current) fuselage frames at the level of the cabin floor. They were due to a combination of pressurization and thermal cycles. As a result of these deteriora- tions, design improvements were made on partial assemblies representing the damaged area (fig. 10). In tests on these partial assemblies, a special fixture was used to simulate the frame warping due to thermal stresses. The results of these tests were very satis- factory, and enabled an excellent behaviour of the frames to be foreseen on series aircraft.
Results Obtainedon Test Structure 2.3.2
By March 1, 1971,14 000complete cycles (A + 2B) and4000 purely subsonicflights
were applied. This stress history correspondsto the damagecausedby 46 000flights
under mechanical fatigue conditions and about28 000flights under thermal fatigue condi-
tions. The deteriorations that were noticed confirm those which were obtainedwith the
substructure 2.6/2.7, andindicated that the samedesign improvements were required.
Someminor deteriorations were found in the door andemergency exit locking devices.
These deteriorations very likely come from local bendingeffects dueto thermal stresses,
andto defects in the door. A few cracks on metal sheetswere detectedandthe investi-
gation of the crack propagationrate is being made. Inside the wing fuel tanks, the orig-
inal rods fitted with clevis weldedby an electron bombardmentprocess did not have a
suitable fatigue life and havebeen replaced by conventional designrods.
Residual Strength After Deteriorations
Deteriorations, especially those concerning fuselage frames, were always found
during the systematic inspection of the structures, that is, following completion of a pro-
gram block including 1000cycles (A + 2B). The damagedstructure exhibited satisfactory
residual strength during the last cycles of the program block.
A flight limit load test uponoccurrence of deteriorations has just beenmadeon
structure 2.6/2.7; this test will be used for certification purposes. Figures 11 and 12
illustrate the test rigs 2.3.2 and 2.6/2.7.
CONCLUSIONS FROMDEVELOPMENTTESTS
The main conclusions are as follows:
1. On a supersonic aircraft whosestructure weight is a significant part of the weight
analysis, many fatigue and static strength developmenttests shouldbe made.
2. Fatiguethermal tests are absolutely necessary. Temperature andthermal
stress calculations, althoughthey are very developed,cannotforesee any fatigue failures
causedby distortion incompatibilities which are not easily evaluated.
2.6/2.7
2.8.h
FATIGUE TEST
2.3.2
STATIC TEST
FATIGUE TEST
FRENCH TEST SUBSTRUCTURES FOR
CONCORDE DEVELOPMENT TESTS
Figure 1.
STRESS_ __ER TEMPERATU /_ WITHOUT FUEL °C daN/ram 2 ksl - 115
STR,.GER I TE.P R TURE
JTEMPERATURE i _ ... - 60
-,o
S STRINGER 7 -80 2 /--- --- --'- _-"-- -- _'" _" -'_'_ - 20 1,oo_ 2,ooo 3,ooo 4,000 s,doo 6,ooo Zooo -2 \ ....... I TIME_ seconds • _STRINGER THERMAL I 20
z-4 I LSTRESS WITH FUEL i
_-6 I
_-101 i11 "'-8 I ... J I I "_ - - - o-lO_E I i_ STRINGER THERMAL I 1 1 I STRESS WITHOUT FUELJ -12 _ / z20 _" STRUCTURE 1 his-EFFECT OF THE FUEL ON LONGITUDINAL THERMAL STRESSES ON A BOTTOM STRINGER OF A FUSELAGE FUEL TANK Figure 2.
LIQUID N ITROGE N 2.8.b THERMAL TEST - COOLING BY 2.8.b THERMAL TEST - WING OVENS Figure 3.
TEMPERATURE> (__ o C
GLAG_E A TOP WING
F//_,/////7/////H_ SKIN
G_E B
_/"SKI N TEMPERATURE (INTERNAL) GAGE A 000 STRESS
seconds r o
F°z
GAGE B -5t-4 _
HIGH ACCELERATION -15f-12 o
k s i da N/turn 2 TEMPERATURE; e_°C f I TEMPERATURE STRESS
1,000 2,000
0, i , 0
-2
_---'_- TIME) sec°nds 5_-
-4
GAGE B
-6
LOW
ACCELERATION -I0
ks1 daN/mrn = 2.8. b STRUCTURE - THERMAL TESTS.
Figure 4.
MAJOR STATIC TEST-
GENERAL VIEWS.
Figure 5.
\ •
THERMAL STATIC TEST.
INFRA-RED OVENS ARE BEI NG INSTALLED AROUND THE FUSELAGE.
Figure 6.
TYPICAL CHANGE IN TEMPERATURES
ejoc
DURING AN ACTUAL FLIGHT
EXTERNAL WALL 100 - TIME 0 12,600 s
1,5oo
3 1 hr
TEST SHORTENING ACHIEVED DURING
e, °c A THERMAL FATIGUE TEST
CONTROLLED HEATING OF INTERNAL AREAS 100-- # /_J_ ICONTROLLED COOLING TIME 0 1,500 _PHASE A] LH_ASE__BB l 1 hr FOR A AND B PHASES THE TENPERATURES CHANGE AS IN THE ACTUAL FLIGHT.
THE PRINCIPLE USED TO SHORTEN
THE THERMAL CYCLE DURATION
Figure 7.
TEMPERATUR_ GUSTS 10GUSTS 6 GUSTS 4 GUSTS e_ oc e =130 °C I i ' CABIN PRESSURE [I J
L/
I I
' f
5O
\
IDDEN p-- -- TANK PRESSURE \ O 2,040 TIM_ SECONDS NOTE : GUST LOADS AND PRESSURE LOADS ARE NOT SCHEDULED IN THIS PLATE.
TEMPERATURE AND FATIGUE CYCLE OF THE CONCORDE 2.3.2 DEVELOPMENT FATIGUE TEST.
Figure 8.
TEMPERATURE
0 o¢
TWO SUBSONIC FLIGHTS SUPERSONIC FLIGHT oC ENGINE AREA
150 e
100 _ /I _
,,if, I ! iIl
, _--ll'ti';/ U_!
0 1,000 2,000 3,000 3,400 TIME_ SECONDS
2.6/2.7 CONCORDE TEST STRUCTURE
TEMPERATURE AND FATIGUE CYCLE.
Figure 9.
NOTE-THE LOAD L WAS DETERMINED TO OBTAIN BETWEEN B- AND E- SECTIONS THE STRESS DISTRIBUTIONS THAT WERE MEASURED ON FRAMES OF THE 2.6/ 2.7 STRUCTURE DURING THERMAL AND MECHANICAL FATIGUE TEST.
L VIEW FROM F
F RAME SKETCH
/ ,
F
/
I;:::LE
/
\ /
\ /
\
SECTION 1 - 1
DEVELOPMENT SPECIMEN TO APPRAISE IN A SHORT
TIME IMPROVEMENTS OF THE FUSELAGE FRAME DESIGN.
Figure IO.
GENERAL VIEW OF THE 2.612.? TEST RIG.
OF THE 2.612.7 FATIGUE TEST.
CONTROL ROOM
Figure ii.
2.3.2 FATIGUE TEST RIG.
THE WING PART IS VISIBLE BETWEEN TOP AND BOTTON WALLS OF THE OPEN OVEN.
ran|| L mmmm _ mmmmm
!
GENERAL VIEW OF THE 2.3.2 FATIGUE TEST RIG.
Figure 12.