DAMAGE-TOLERANCE SUBSTANTIATION OF THE GALAXY EXECUTIVE JET

Presented to the ICAF 1999 Conference, Seattle, WA, 1999 DAMAGE-TOLERANCE SUBSTANTIATION OF THE GALAXY EXECUTIVE JET Abraham Brot, Shmaryahu Afnaim and Zvi Granot * SYNOPSIS This paper describes the fatigue and damage-tolerance substantiation program for the new Galaxy wide-body executive jet aircraft. The airframe structure was substantiated to the FAR-25 and JAR-25 damage-tolerance requirements. A description is given of the analytical and experimental work which were conducted in order to meet these requirements. The paper also discuses: p- version finite-element modeling, determining inspection intervals using a cumulative probability of detection method, the results of component tests that were performed on key structural elements and the philosophy and early results from the two-lifetime full-scale fatigue test. INTRODUCTION The Galaxy, wide-body executive jet (Figure 1) (1-3) flew for the first time in December 1997 and received its type certificates from the CAA of Israel and the FAA in December 1998. The Galaxy has a transatlantic range and a cruise speed of up to Mach 0.85. It transports up to nine passengers in an executive configuration and up to 18 passengers in a corporate configuration. The Galaxy is powered by two PW306A turbofan engines. The primary structure is metallic except for ailerons, elevators and rudder which are made from composite materials. The airframe structure has been substantiated to the FAR-25 and JAR-25 damage-tolerance requirements. In order to ensure a service life goal of 20,000 flight-cycles and 36,000 flight hours, a comprehensive fatigue and damage-tolerance substantiation program was undertaken, as described below. The analytical part of this program includes, fatigue loads spectrum development, finiteelement modeling, detail stress and damage-tolerance analysis of the various critical locations, and the determination of inspection intervals and methods. The experimental part of this program includes, coupon testing for material behavior characterization, component development testing of key structural items, and full-scale fatigue testing of the aircraft structure for a duration of two lifetimes. * Engineering Division, Israel Aircraft Industries Ltd., Ben-Gurion International Airport, Israel 1

The results of the Galaxy damage-tolerance analyses were translated into inspection requirements, which are applied to the principal structural elements (PSE) of the aircraft. A goal was established to schedule the threshold inspections at 50% of the design life goal (10,000 flights) and subsequent inspections at 25% intervals (5000 flights). The inspection intervals were selected using a cumulative probability of crack detection method, which will be discussed in a later section. The structural inspection requirements for the Galaxy Aircraft were integrated into the Airworthiness Limitations Section of the maintenance manual of the aircraft, which was written according to the MSG-3 philosophy and format. Approximately 130 locations for periodic inspections for cracks are specified for the Galaxy Aircraft. FATIGUE SUBSTANTIATION PHILOSOPHY The Galaxy Executive Jet is designed to provide at least 20 years of service including a life goal of 20,000 flight-cycles and 36,000 flight-hours. In order to accomplish this goal, adequate fatigue life has been a design requirement of every component of the aircraft. In order to insure adequate fatigue life, an encompassing fatigue substantiation philosophy had been adopted. This philosophy includes loading spectrum development, substantiation to the damage-tolerance requirements of FAR-25 and JAR-25, coupon tests to establish crack growth parameters under spectrum loading, coupon tests to establish truncation limits, component tests to verify the adequacy of specific structures and a two-lifetime full-scale fatigue test. Clearly, the philosophy was to provide sufficient experimental basis to the analytical damage-tolerance substantiation. Multi-site damage (MSD) has, in recent years, been identified as an important mechanism of many fatigue failures that have occurred in several aircraft. Galaxy structures, that may be prone to multi-site damage, have been designed to moderate stress levels in order to minimize the probability of an MSD failure during the design life. In addition, the two-lifetime full-scale fatigue test is being used to verify that MSD is extremely unlikely to occur during the expected service life of the aircraft. Composite structures are used for several control surfaces of the Galaxy. These have been designed to sufficiently low strain levels in order to preclude fatigue failures in service. In addition, the composite material elevator and rudder is being tested for damage-tolerance under manufacturing and service inflicted damage, as well as under barely-visible impact damage, as part of the empennage component fatigue test. The entire empennage is being tested for two lifetimes, followed by an additional lifetime of damage-tolerance testing, aimed at verifying the adequacy of the composite control surfaces. The results of the Galaxy damage-tolerance analyses were translated into inspection requirements, which were applied to the approximately 130 principal structural elements (PSE) of the aircraft. At each PSE, a threshold (initial) inspection and an inspection interval was specified, as well as the NDI method to be used. A goal was established to schedule the threshold inspection at 50% of the design life goal (10,000 flights) and subsequent inspections at 25% intervals (5,000 flights). The NDI method selected for each PSE was based on the crack growth characteristics of the specific location with an aim of selecting the most cost-effective procedure. 2

ANALYTICAL SUBSTANTIATION The process of analytical substantiation consists of the following steps: loading spectrum development, finite-element modeling, damage-tolerance analysis and specifying inspections during service. These steps will be briefly described. Loading Spectrum Development: Loading spectra development began with the definition of typical aircraft missions. Based on market surveys and usage-tracking of previous customers, three typical missions have been identified having flight lengths of one, two and three hours. Each mission has then been subdivided into several flight segments such as: taxi, rotation, climb, cruise, descent, flapextension, approach and landing impact. For each of these segments, the load-factor spectra for gusts, maneuvers, taxi and landing impact were determined. Cabin and baggage compartment pressurization was added for airborne flight segments. The specific load spectrum for each aircraft component was determined from the load-factor spectrum by computing the aerodynamic and inertial loads corresponding to the specific load condition. The development of the Galaxy nose landing gear spectrum included an experimental study of nose-gear landing impacts during the first twenty flights of the Galaxy flight test aircraft.. It was found that the equivalent mass at the nose-gear actually was 3.1 times the static mass, which is greater than the value previously obtained by analysis. However the measured sinkingspeeds at the nose gear averaged only 0 41 m./sec. (1.34 ft./sec.), which was less than previously assumed (4). The resulting landing impact spectrum was felt to be realistic, since it was based on flight-test measurements. Typical buffeting loads caused by engine reverse thrust during landing were measured on a flight-test aircraft. Approximately forty cycles per flight are applied as roll and yaw moments on the horizontal stabilizer, based on the measurements that were performed. Finite-Element Modeling: Finite-element modeling is used in the fatigue substantiation procedure in two distinct ways: A finite-element model of the entire aircraft structure is used to determine the stress spectrum corresponding to a specific location on the aircraft. Special purpose finite-element models are used to analyze specific discrete structures in order to determine stress-concentration factors or stress-intensity factors for complicated geometries. NASTRAN is used for the first application while StressCheck, a special purpose p-version finite-element program, is used to generate the discrete models. Figure 2 illustrates a two-dimensional StressCheck model that was built to determine stressconcentrations and stress-intensities in a web of the Galaxy wing carry-through structure. This web contained several holes of different diameters in close proximity, at a location of the web that is stressed in tension under gusts and maneuvers. The relatively simple models that were built, allowed for accurate estimations of the peak stress levels and of the stress-intensity of any 3

size crack that could occur. This allowed the analyst to conclude that this location remains damage-tolerant, even with the interacting stress-concentrations due to the holes. Three dimensional StressCheck models are also used to analyze critical fittings and other similar structures in order to determine their peak stresses. Damage-Tolerance Analysis: The damage-tolerance analysis began with the selection of specific locations to be analyzed. These locations are selected on the basis of stress level, stress-concentrations, previous test or service experience and criticality of the structural element. The damage-tolerance analysis was performed, based on the FAR-25 and JAR-25 requirements, using software that calculates crack growth and residual strength. Initial cracks were assumed to be of a 1.2 mm (0.05 inch) size, and the number of flights to reach the critical crack size were calculated by the methods of linear elastic fracture mechanics. The crack growth material parameters, including spectrum retardation effects, were based on coupon testing that was performed specifically for the Galaxy loading spectrum. A table-lookup method, based on the Walker model, and the Wheeler and Closure retardation models were used to calculate crack growth. When appropriate, continuing damage calculations were made to calculate the progression of cracking in multi-path designs. Environmental effects, such as high humidity or fuel contamination, were accounted for empirically. Additional software was used to show, by analysis, that the fuselage could sustain a two-bay crack, which would be arrested at the adjacent fuselage frames. Specifying Inspection Intervals: The output of the damage-tolerance analysis is used to determine inspection methods and intervals to safeguard the structure throughout its lifetime. For each location that was analyzed, a threshold inspection and inspection interval was specified. The threshold inspection was taken as 50% of the crack growth life, as determined by the damage-tolerance analysis, but not greater than 50% of the design lifetime (10,000 flights), as is the FAA recommended practice in this matter (5). Various airframe manufacturers use different methods to determine periodic inspection intervals of damage-tolerant structures. The most common method used is to determine the crack growth life from the detectable crack size to the critical crack size. The inspection interval is then selected to be 50% of this value (5). The underlying philosophy for this method is to always allow two opportunities to detect a crack before it reaches its critical size. A weakness of this method is the lack of a precise definition of a detectable crack size. Dr. David Broek has shown that applying this method will not yield the same level of safety for all structures (6). A recent study, based on computer simulations of the entire fatigue process has come to the same conclusion (7). The two crack growth curves, that are shown in Figure 3, illustrate this point. Although the two crack growth curves are vastly different in shape, both share the same detectable crack size (5mm) and critical crack size (40mm). The crack growth life from the 4

detectable crack size to the critical crack size is also identical for both curves (4000 flights). Hence, the same inspection interval would be selected for both structures (2000 flights). Two typical inspections, having an interval of 2000 flights between them, are shown in Figure 3. It is, however, clear that the inspections corresponding to Curve A would always have a higher probability for detection than those for Curve B, since Crack A is always larger than Crack B during any of the inspections. Therefore, the structure characterized by Curve A will be safer than that characterized by Curve B. This proves that the degree of safety, in a damage-tolerant structure, is a function of the shape of the crack growth curve as well as the inspection interval something that is not corrected for by the commonly used method used for determining intervals. The inspection intervals for the Galaxy were determined using a computerized cumulative probability of crack detection method (6). This method computes the probability of crack detection for each scheduled inspection, based on the predicted crack size. The method calculates the cumulative probability, as a function of inspection interval, that the crack will be detected, at least once, during all the scheduled inspections. The required cumulative probability of detection criterion ranges from 95% to 99.5%, depending on the criticality of the structure and additional opportunities for crack detection that may exist. On this basis, the inspection interval is determined by a method that insures a consistent level of safety. Figure 4 demonstrates a typical output from the cumulative probability of detection analysis that was performed for a specific location. The results are plotted as the probability of not detecting a crack as a function of the inspection interval. The allowed cumulative probability of not detecting a crack (0.5% 5.0%) is selected, and the required inspection interval can be read from Figure 4. Often, several inspection methods (visual, liquid-penetrant, etc.) are investigated and traded-off in order to select the most cost-effective method and interval, as is shown in Figure 4. In order to minimize aircraft downtime, a goal was established for Galaxy, to schedule the great majority of inspections not more frequently than 25% of the design lifetime (5,000 flights). Since the Galaxy aircraft is capable of flight up to a 45,000 foot altitude, FAA regulations required reduced inspection intervals for critical locations in the main cabin, failures of which could result in rapid decompression. These locations were designed so that the cumulative probability of crack detection is not less than 99.5%. COMPONENT TESTING Cabin Window Fatigue and Damage-Tolerance Test: Figure 5 shows the cabin window fatigue and damage-tolerance component test specimen and rig. An improved, lower weight, one piece aluminum window frame is reinforcing the oval window cutout. The objectives of the test were: to show adequate fatigue service life, prove adequate residual strength, obtain data on crack initiation and crack propagation rates for cracks at the most susceptible locations, and to evaluate the appropriate NDI methods for in-service inspections. 5

A section of the fuselage, containing the structure surrounding the cabin window, was repetitively pressure tested for two lifetimes in the test rig. The loading consisted of pressure cycles from zero to 10.35 psi. A 15% increase in the maximum cabin pressure is taken to allow for the exclusion of other types of flight induced loads. At the conclusion of the two lifetimes of fatigue testing, initial flaws, represented by fine saw-cuts, were introduced at three locations: at the window frame (at a cutout corner radius), and at two locations of significant skin bending due to bulging, near the top and bottom window sills. As shown in Figure 6, after 12,100 flights the initial crack at the window frame grew sufficiently to break the window frame. After a total of 14,600 flights, a continuing crack had emerged in the skin and grew to a total length of about 95mm. No growth was detected at the other two saw-cut locations. A residual strength test, with a maximum pressure of 11.9 psi, was then applied. Figure 7 is a photograph of the crack taken during the application of the residual strength condition. It shows that the crack grew in a stable manner, under the residual load condition, but had arrested at the fuselage frame. The test has confirmed the adequate fatigue and damage-tolerance life of the cabin window structure, including the ability of any cracks to arrest at the adjacent fuselage frame. Fatigue and Damage-Tolerance Test of Lower Wing Skin Splice at Rib 2: The Rib 2 area of the wing lower skin is a complex area that includes the splicing of four skin segments (outboard-to-inboard and center-to-rear skins) and change of contour around the landing gear bay. In addition, the rear spar is terminated and its load is transferred to the center spar in this area. The test-article included a lower wing skin segment from Rib 0 to Rib 3A and from the rear spar to the longitudinal splice, as is shown in Figure 8. The spars and ribs were, conservatively, represented by their lower flanges only, omitting the stabilizing effect of the remaining structure. A typical access panel cutout was included in the specimen. The specimen was clamped at Rib 0. Lower skin loads were applied by three servo-hydraulic actuators at Rib 3A, as is shown in Figures 8 and 9. The loading spectrum was the Galaxy gust and maneuver spectrum with a block of 2,000 flights, divided into three missions. Nearly 50 strain gages were bonded on the specimen and monitored during the test. Visual, eddy-current and X-Ray inspections were performed periodically. There was good correlation between the measured and calculated strains. A number of small cracks (3-10 mm) were detected at the radius of the outboard skin near the center spar after 30,000 flights of fatigue cycling. The cracks joined and grew into a 100 mm crack after 34,000 flights, as is shown in Figure 8. The causes of crack initiation were: Sharp edge at skin landing gear bay intersection with the thickness step at the skin splice. Small radius at the thickness step at the skin splice. Eccentric load transfer at the lap joint forming the skin splice. 6

The latter effect was more severe in the test-article than in the actual wing, due to lack of stabilizing effect of spars and ribs. At 34,000 flights a patch was installed in the cracked area and the testing was resumed. Design changes for the Galaxy fleet were introduced in this area, as a result of the component test. At 40,000 flights, eight artificial flaws were introduced, by fine saw-cuts, at points not influenced by the patch. A crack began to grow from the initial flaw at the access panel cutout at 50,000 flights and it grew, to a length of 15 mm, at the end of the test at 73,000 flights. The crack growth curve is given in Figure 10. The relatively constant crack growth rate shown in Figure 10 indicates that, as the crack extends, the load is transferred to stiffer portions of the wing. Main Landing Gear Fatigue Test The main landing gear fatigue test is shown in Figure 11. Spectrum loads are applied in the vertical, drag and side directions, as is shown in the figure. A truncated fatigue-load spectrum of the ground-loads is used for this test, after combining and simplifying some of the ground-operations. The six stages that were taken within each flight-cycle are: taxi, turning, braking, pivoting, rotation and landing-impact. The test spectrum was truncated to include approximately twenty events per flight. Groups of 1,000 of these flights were assembled into each block of flight-cycles for the fatigue-test load-spectrum. Using this fatigue loading-spectrum, the Galaxy main landing gear has begun its test to five lifetimes (100,000 flights). By early July 1999, the landing gear had successfully completed 10,400 flights. COUPON TESTING Fatigue testing of coupons was used for two purposes: determination of crack growth parameters for specific material and spectra, and to assist in determining spectrum truncation criteria for fatigue testing. For the principal metallic materials used on the Galaxy structure, center-cracked coupons and cracked coupons at a central hole, were tested to constant amplitude and spectrum loading. Various Galaxy spectra, which simulate various combinations of gust, maneuver, cabin pressurization and lateral loading, were used in this study. The crack growth behavior for these coupons were determined, for the various spectra and materials, by loading the coupons in a cyclic test fixture and measuring crack growth. By this procedure, the constant amplitude as well as the spectrum retardation parameters were determined for each material and spectrum combination. These parameters were applied to the damage-tolerance analyses that were performed. During fatigue testing of components or entire airframes, certain cycles of low stress amplitude, but having many cycles of load application, may be eliminated in order to accelerate the test. In order to determine whether these cycles can be eliminated, with a minimal loss of test accuracy, a series of coupon tests were performed. Various combinations of spectra were tested 7

and the effect of load truncation on crack growth life was determined. These results enabled a rational decision to be taken for load truncation for the component and full-scale fatigue tests. Following a component or full-scale fatigue test, fractography is used to track crack growth. In order to correlate crack growth with the number of flights, certain marker loads are introduced into the loading spectrum. The coupon tests were used to check the visibility of these marker loads, as well as to enable an evaluation of the effect of these marker loads on the fatigue life. FULL-SCALE FATIGUE TEST The full-scale aircraft structure is being fatigue tested for a duration of two lifetimes. The Full- Scale Fatigue Test (FSFT) is being performed in order to fulfill the following objectives: Substantiate the airframe structure for two lifetimes of fatigue loading which is representative of anticipated service. Detect critical locations where cracks could initiate. Establish more accurate crack growth rates at critical locations, which account for redistribution of stresses and interaction between structural components. This data will be used to update the damage-tolerance analysis. Define effectiveness of Non-Destructive Inspection (NDI) methods and set criteria for detectable cracks in the analysis. Obtain data on the extent of secondary cracking for multiple site cracking analysis. Demonstrate the effectiveness of any repairs or redesigns introduced to rectify fatigue problems. The test-article for the full-scale fatigue test consists of a standard production aircraft, having all the structural members of the fuselage and both wings. All non-essential components and systems such as: cabin interior trim and furnishings, radome assembly, aft cone, mechanisms, and landing gear doors were not included. The empennage is being fatigue tested separately, as described below. Figure 12 shows the Galaxy Full-Scale Aircraft mounted in its loading fixture. In this test, the empennage from the static test-article is employed for load application purposes. Dummy engines, landing gears, actuators, and the left-hand Kruger flap are included. The system and equipment supports and brackets, that are structurally effective, are included in the test-article. The remaining smaller brackets are omitted, but the fastener holes used for their attachment are drilled on the airframe structure. The test aircraft is mounted to the test fixture on rigid constraint supports, as follows: Nose landing gear: constraint in the vertical direction. Engine mount fittings: constraint in the vertical and lateral directions. Fuselage skid: constraint in the fore-and-aft direction. 8

The reactions at these support points, throughout the tests, are monitored by means of loadcells. The fatigue test spectrum loading consists of randomly selected flight-by-flight sequences, reflecting the anticipated usage of the aircraft. Each flight is arranged to consist of all the various airborne and ground events that the aircraft experiences in service. The spectrum is arranged in blocks of 2,000 flights having three types of missions: 900 short range flights, 600 medium range flights and 500 long range flights. The 21 events included in each flight (11 air events, 3 transition events, and 7 ground events) are performed in the following sequence: No. of Events / flight 1. Taxi before take-off 2 2. Braking before take-off 1 3. Take-off run / Rotation 1 4. Climb gust 2 5. Climb maneuver 1 6. Cruise gust 3 7. Cruise maneuver 1 8. Descent 1 9. Flap extension 1 10. Approach gust 1 11. Approach maneuver 1 12. Landing impact ( 4 steps ) 1 13. Reverse thrust / 1g after landing 1 14. Taxi after landing 2 15. Braking after landing 1 16. Transition: landing to takeoff 1 Total: 21 Randomization of events is performed within each stage, and events are picked randomly to form one flight. At the end of each block, marker loads are introduced consisting of the 21 highest events in the block. Testing started with a strain survey, which will be followed by two lifetimes of spectrum loading. Approximately 600 strain-gages were mounted to the test-article. At least 200 strain gages are monitored every block, and 60 strain gages are to be monitored continuously. Wherever cracks may appear, the crack growth is monitored using crack gages. The test was designed so that the test-article is initially under zero external load, with the dead weight of the test-article and test rigs extracted to give zero reactions at the engine and nose landing gear supports. 9

The test-article was divided into 26 loading zones and four reaction constraints, as follows: No. of Zone Number Description Channels No. of jacks of Areas LOADS 4 1,2,3,4 4 Forward fuselage 3 5,6,7 3 Aft Fuselage (F z ) 3 8,9,10 6 Fixed wing, inboard (F z ) 4 11,12,13,14 8 Fixed wing, outboard (F z ) 1 15 2 Inboard flap (resultant) 1 16 2 Outboard flap (resultant) 1 17 2 Kruger flap 1 18 2 Aileron attachment (F z ) 3 19,20,21 6 Slat (resultant) 1 22 1 Horizontal tail (F z ) 1 23 2 Engine thrust 1 24 2 Main landing gear (F z ) 1 25 2 Main landing gear (F x ) 1 26 - Fuselage pressure Total actuators: 42 REACTIONS 1 - - Nose landing gear reaction (F z ). 2 - - Engine mount support structure (F x ) & (F z ). 1 - - Fuselage lateral support (F y ). Loading is applied in the vertical and fore-and-aft directions. Each loading zone is independently loaded during each event of the spectrum, using servo-hydraulic actuators. The zone loading for each event was determined using a constrained least-square error method which minimizes deviations of the important structural parameters. In addition, the passenger cabin and baggage compartments are pressurized during the airborne events of the spectrum, to a pressure of 9.0 psi, using compressed-air, NDI is being performed at specific intervals at all major structural items. An evaluation of the inspection methods and results in the NDI program will be reflected in updates of the Instructions for Continued Airworthiness, per FAR 25.1529. 10

Upon completion of two lifetimes, additional testing may be carried out as required. At the completion of cycling, a teardown inspection of selected areas will be carried out, and fractographic analysis of cracked areas will be performed. By early July 1999, the test has reached 8,500 flights out of a planned 40,000 flights. As a result of strain-gage measurements and minor cracking that occurred in the full-scale test, design improvements were added to the test-article and were already implemented in the fleet. These improvements include: Adding reinforcing members to the corners of the nose landing gear bay, reinforcements to the baggage compartment aft pressure bulkhead and reinforcements to the doorframe of the baggage loading door. EMPENNAGE FATIGUE AND DAMAGE-TOLERANCE TEST An empennage fatigue and damage-tolerance test, including the horizontal stabilizer, vertical tail, elevators, rudder and a section of the aft fuselage, is being performed. The aim of this test is to substantiate the empennage structures for fatigue and damage-tolerance under a combination of symmetrical, asymmetrical and engine thrust- reverser buffeting loads. The test-article and its installation to a rigid rig is shown in Figure 13. The aft fuselage section extends up to the aft engine support frame at fuselage Sta. 15891. The elevator was installed at an angle of 27 upward relative to the horizontal stabilizer reference plane. This angle corresponds to the most severe fatigue event for the elevator. Three types of loads are applied to the test-article: Vertical loads on the horizontal stabilizer and elevator. They include lift-off, gust, maneuver and ground loads. Side loads on the vertical tail and rudder. They include lateral gust and yaw maneuver loads. Buffeting loads caused by engine reverse thrust during landing. They are applied as roll and yaw moments on the horizontal stabilizer. The loading spectrum is based on a block of 2,000 flights based on three missions; it includes nearly twenty gust and maneuver events and, approximately, forty buffeting load cycles per flight. The loading set-up for the empennage gust and maneuver loads includes five hydraulic actuators on the vertical tail and six actuators on the horizontal stabilizer. The buffeting loads are applied by a separate set of four hydraulic actuators: two for the roll moments and two for the yaw moments. There are a total of ten independent loading zones. The residual strength test to be performed at the end of the cyclic loading will include two limit load cases: A combined horizontal stabilizer / vertical tail limit load case. A limit load case for the horizontal stabilizer and the elevators. Approximately 150 strain gages were bonded on the test-article. NDI inspections are performed periodically during the test, at fixed intervals. A zero inspection was performed after the assembly of the test-article. A teardown inspection will be performed after the test is 11

completed. Embedded manufacturing defects and barely-visible impact damages were incorporated in the composite material elevators and rudder during their production. Artificial flaws will be introduced on the metallic structures, and visible impact damage will be applied to the composite structures, after two lifetimes of fatigue cycling. The test includes the following stages: Static calibration consisting of a complete strain survey under a calibration load. Fatigue test consisting of two lifetimes (40,000 flights) of fatigue cycling, with calibrations and inspections at fixed intervals. Introduction of artificial flaws at critical locations and one lifetime (20,000 flights) of damagetolerance cycling with crack growth and defect measurement. Two cases of residual load testing. Tear down inspection. By early July 1999, the empennage test has reached 8,000 flights out a planned 60,000 flights, with no structural problems noted. SUMMARY This paper has described briefly the experimental and analytical elements that were part of the certification program for the structural fatigue and damage-tolerance of the Galaxy aircraft. REFERENCES 1. Vitaly J., Shaul A., Rubin T., Development Approach of Galaxy Business Jet, Proceedings of the 38th Israel Annual Conference on Aerospace Sciences, February 25-26, 1998. 2. M. Boness, Israel Aircraft Industries Business Jet, Past, Present and Future, Proceedings of the 38th Israel Annual Conference on Aerospace Sciences, February 25-26, 1998. 3. Ghilai G., David A., Reisel P., Galaxy Composite Structures Certification Methodology, Proceedings of the 38th Israel Annual Conference on Aerospace Sciences, February 25-26, 1998. 4. Chester, D. and Brot, A., Development of a Fatigue Spectrum for Nose Gear Landing Impact, Proceedings of the 20th Symposium of the International Committee on Aeronautical Fatigue (ICAF), Bellevue, Washington, USA, 1999. 5. Swift, T., "Verification of Methods for Damage Tolerance Evaluation of Aircraft Structures to FAA Requirements", Proceedings of the 12th Symposium of the International Committee on Aeronautical Fatigue (ICAF), Toulouse, France, 1983. 6. Broek D., Chapter 11, Fracture Control, The Practical Use of Fracture Mechanics, Kluwer Academic Publishers, The Netherlands, 1988. 7. Brot, A., A Critical Review of Damage-Tolerance Methodology by Means of a Probabilistic Simulation of the Fatigue Process, Proceedings of the 19th Symposium of the International Committee on Aeronautical Fatigue (ICAF), Edinburgh, Scotland, 1997. 12

FIGURES FIGURE 1: Galaxy Executive Jet during a Test Flight FIGURE 2: P-Version Finite-Element Model of a Web of the Wing Carry-Through Structure 13

60 50 4000 FLIGHTS Crack Size (mm) 40 30 20 CRITICAL CRACK 1st INSPECTION 2000 FLIGHTS CURVE A 2nd INSPECTION 10 DETECTABLE CRACK CURVE B 0 0 1000 2000 3000 4000 5000 6000 Crack Growth Life (Flights) FIGURE 3: Selection of Inspection Intervals for Two Crack Growth Curves 1 0.1 Probability 0.01 0.001 0.0001 0.00001 Visual Inspection 0 2000 4000 6000 8000 10000 Inspection Interval (Flights) Penetrant Inspection FIGURE 4: Cumulative Probability of Not Detecting a Crack For Two NDI Methods 14

a) Top View of Test Specimen b) Section AA of Test Rig FIGURE 5: Cabin Window Fatigue Test Component and Rig 100 90 80 70 60 50 40 30 20 10 0 Crack Size - mm 0 5000 10000 15000 Flights After Introduction of Initial Flaw FIGURE 6: Crack Growth at Window Frame and Continuing Skin Crack 15

FIGURE 7: Cabin Window Component Test Under the Residual Strength Condition FIGURE 8: Wing Lower Skin Fatigue Test Component and Rig 16

FIGURE 9: Fatigue Component Test Performed on a Section of the Lower Wing 16 Crack Length - mm 14 12 10 8 6 4 2 0 0 5000 10000 15000 20000 25000 Flights After Introduction of Initial Flaw FIGURE 10: Crack Growth at Wing Lower Skin Access Panel 17

FIGURE 11: Main Landing Gear Fatigue Test (landing gear mounted upside down in its loading fixture) FIGURE 12: Full-Scale Fatigue Test-Aircraft Mounted in its Loading Fixture 18

Presented to the ICAF 1999 Conference, Seattle, WA, 1999 FIGURE 13: Empennage Fatigue Test-Article Mounted in its Loading Fixture 19