Practical applications on fatigue and fracture of aircraft structures

Transcription

1 Chapter Twelve Practical applications on fatigue and fracture of aircraft structures A. Dabayeh L-3 Communications, Aviation Services, Canada Abstract The fatigue-failure problem first gained interest during the industrial development age, due to the nature of repeated loading on machine components. The focus was then directed, during design stages, to the prevention of fatigue-failures and ensuring the durability of the component, in addition to the conventional prevention of static failures. The idea was to ensure a crack-free design during total service life of the component. With the development of metallic aircraft structures, the crack-free approach has shown to be impractical, especially because of the structural weightsaving challenge that has to be paramount if an efficient performance aircraft is required. In other words, aircraft-structures design was directed to having adequate strength and rigidity at minimum weight. The aircraft designer had to change the design concept and accept inelastic/plastic deformations at highly stressed locations during peak operating stresses of the aircraft. These plastically deformed materials initiate cracks and limit the service life of components, which initiated the idea of new design approaches. The damage-tolerance and fail-safe approaches are two of these new design approaches implemented to assess and deal with accepting cracks in service life. It is the purpose of this chapter to show some background and insights about the fatigue and crack-growth analysis methodologies typically implemented in the aircraft industry, as an example of the durability and damage-tolerance approaches. The chapter will also give some highlights about full-scale airframe fatigue testing and structural certification. Some practical examples from work cases on the CF-18 airframe structure, which the author is currently involved with, will be provided. doi: / /12

2 394 Advances in Fatigue, Fracture and Damage Assessment 1 Aircraft structures Aircraft structure could be categorised into two; Rotary-wing aircraft such as helicopters, and fixed-wing aircraft such as commercial passenger aircraft. This chapter will be dealing with fixed-wing aircraft structures only, although a large part of the theories, policies, and strategies are still applicable to both types. 1.1 Geometry and structural components Fixed-wing aircraft consist of some major structural components as shown in Fig. 1 for the commercial Airbus A-380 aircraft, and in Fig. 2 for the military McDonnell Douglas/Northrop CF-18 Hornet aircraft. Major components include the following: 1. Fuselage: mainly consisting of Formers/bulk-heads Longerons, and Covered by aluminium skins. 2. Wing box (es): mainly consisting of: Wing spars that could be as few as two (the front and rear spar) such as the A-380, or could be multiple spars such as the CF-18 that has six. The CF-18 also has inner and outer wing boxes. Intermediate ribs between spars. Transmissions to operate leading- and trailing-edge flaps. 3. Horizontal Tail/Stabilator: similar in structure to the wing 4. Vertical Tail(s): CF-18 has two vertical tails 5. Nose and main landing gears 6. Engines 7. Control surfaces: that include: Leading-edge flaps Trailing-edge flaps Ailerons Elevators Rudder 1.2 Materials Materials most commonly used in the aircraft industry are: 1. Aluminium-alloys: that include the 2000 series where copper is the main alloying element, and the 7000 series where zinc is the main alloying element. Examples of such materials are 2024-T3, 7050-T74, 7075-T76 etc. The 2000 series aluminium group is known for its good fatigue and damagetolerance resistance and is mainly used for locations dominated by tension loads such as lower wing skins, fuselage skins, etc.

3 395 Figure 1: Airbus A-380 geometry and structural FUSELAGE WING AILERON VERTICAL TAIL HORIZONTAL TAIL Advances in Fatigue, Fracture and Damage Assessment

4 396 Advances in Fatigue, Fracture and Damage Assessment Outer wing Aileron Inner wing Trailing-edge flaps Canopy Leading-edge flaps Fuselage Main landing gear Figure 2: McDonnell Douglas/Northrop CF-18 Hornet geometry and structural components. Vertical Tails Engines Horizontal Stabilators

5 Advances in Fatigue, Fracture and Damage Assessment 397 The 7000 series aluminium group is known to be best applied in compressionaffected locations where fatigue and damage-tolerance is not critical, such as upper-wing surfaces, wing ribs, floor beams, etc. 2. Titanium-alloys: Titanium is known for its high strength and light weight and is preferred to be used in aircraft structures where high fracture toughness is required. For example, plane-strain fracture toughness for Ti- 6Al-4V is 50 ksi in 1/2 whereas it is about 30 ksi in 1/2 for most 2000 and 7000 aluminium series materials used in the aircraft industry. 3. Steel-alloys: Steel is not preferred in aircraft structures because of its high weight and corrosion properties. However, in some locations where strength is required with no big influence on the weight, steel alloys are used, such as the engine mounts, main landing gears, etc. For aircraft that service in marine environments some corrosion protection or plating systems are applied on these steel components specifically and on some aluminium ones as well. Cadmium plating and Ion Vapour Deposited (IVD) aluminium coatings are examples of such plating systems. 4. Composites: Composite materials are known for their high strength and stiffness-to-density ratios, corrosion resistance, and their good formability to complex shapes. They also have good damping effects. However, composites are less ductile than the 2000 aluminium series materials and would have less fracture resistance. Common applications are upper-wing skins, vertical tail skins, upper-deck skins, etc. 2 Fatigue and crack-growth analyses The fatigue and crack-growth analyses of aircraft components are both important in defining the aircraft structural integrity. With these analyses and data obtained from different types of tests different measures are taken in design, such as determination of safe service life, inspection intervals, residual-strength characteristics, critical crack length and others. Fatigue-life of aircraft components usually consist of one or both of the fatigue phases; crack-initiation and crack-growth. 2.1 Fatigue-crack initiation and propagation Fatigue crack-initiation life could be defined as the fatigue-life of initially uncracked components that is spent in initiating a crack of engineering size. This size is taken as the crack length at which a small fatigue specimen fails or as some observable size. In terms of the progress of the fatigue process, this initiation phase consists of the following:

6 398 Advances in Fatigue, Fracture and Damage Assessment 1. Cyclic slip-band formation. 2. Crack initiation in a slip-band. 3. Growth of the initiated micro-crack to the defined initiation size. Not included in the initiation phase are the subsequent growth of the initiated micro-crack and the final fracture. Usually the cyclic-slip and crack-initiation phases are short enough to be ignored for service loading. The total fatigue-life can be represented by the propagation of a small crack from the size of a slip-band crack to the defined initiation size, which is called the crack-initiation phase, plus the propagation to the size at which fracture occurs, which is called the crack-growth phase. It could also be argued that there is no crack-initiation phase and that it is all one phase of crack propagation from the slip-band till fracture. This argument still stands, however, care must be taken in dealing with the fracture-mechanics of small and long-cracks as they do not follow the same theory. Having said that, and from a fracture-mechanics point of view, the total life could be divided to the following phases: 1. The growth of a micro-crack (short fatigue-crack growth). This includes micro-structurally and physically small-crack growth. A schematic plot for the threshold stress, for fully reversed constant-amplitude loading crackgrowth, on logarithmic co-ordinates of threshold stress versus crack length, is shown in Fig. 3. The figure shows regimes of short- and long-crack growth. In the micro-structurally short-crack regime cracks may grow temporarily at stress levels below the fatigue limit but will stop at barriers. The interaction of a small fatigue-crack with a micro-structural barrier, such as a grain boundary or a secondary phase, causes the following: a. When a small-crack approaches a micro-structural barrier it will decelerate or arrest due to the barrier and will grow quickly once it overcomes the barrier. b. The major retardation effect of the micro-structural barrier to small fatigue-crack growth is within the first few grains. As the crack propagates, the micro-structural barrier effect becomes less. c. The influence of the micro-structural barrier on the growth of a small fatigue-crack is greatest when the stress range is close to the fatigue-limit stress range. As the stress range becomes larger, the effect of barriers rapidly reduces, as does the fraction of the fatigue-life spent in the short-crack regime. In the physically short-crack regime the threshold stress increases as the crack length becomes shorter but falls further below the value associated with Linear Elastic Fracture Mechanics (LEFM), as shown in Fig. 3, until at the end of the regime, it reaches the fatigue limit for the material. In other words, these cracks can propagate at fatigue threshold stresses below the

7 Advances in Fatigue, Fracture and Damage Assessment 399 fatigue limit and the long-crack threshold stress given by Kth. Taylor [1] using criteria for the validity of LEFM, suggested that L 2 shown in Fig. 3 be defined by the greater of 10 times the metallurgical barrier spacing or 10 times the grain size. On the other hand, Kendal et al. [2] and Blom et al. [3] found that for steels and high-strength aluminium alloys, experimental evidence indicated that LEFM analysis was adequate if crack-closure was taken into account. The emphasis on LEFM applicability is not important if a strain-based intensity factor [4] is used in crack-growth models, which is identical to the LEFM for elastic strains and identical to J-integral at inelastic strain levels. Blom et al. [3], Kendall et al. [2], and others all found that measurable crack-closure began al L 1 and showed an initially steep build up that tapered off to a steady-state value for long-cracks at L 2, see Fig. 4 [5]. When these measurements were used to deduce effective stress intensities for short-cracks, data for long and short-cracks fell together on the A B Figure 3: Variation of threshold stress with crack length; Curve A considers crack-closure (i.e. threshold stress includes the crack-closure stress), curve B is the intrinsic curve with no crack-closure (threshold stress includes the effective stress only). intrinsic curve for long-cracks. These authors suggested that once crackclosure was accounted for, the long-crack regime could be considered to begin at L 1. The fraction of the fatigue-life spent propagating a crack through the metallurgically short-crack regime represents almost all the fatigue-life at the fatigue limit but rapidly decreases to a negligible fraction

8 400 Advances in Fatigue, Fracture and Damage Assessment as the strain amplitude increases [5]. Figure 5 [6] illustrates this trend for a low-carbon steel. Figure 4: Variation of threshold stress-intensity factor with crack length [5]; K th is the threshold stress intensity range, and Ki is the intrinsic stress intensity range. (Number of cycles to failure) Figure 5: Typical strain life diagram for a low-carbon steel [6]. 2. The growth of a macro-crack (long fatigue-crack growth). This includes the growth of a long fatigue-crack using LEFM principles. These principles are used to relate the stress magnitude and distribution near the crack tip to: a. Remote stress applied to the cracked component. b. The crack size and shape. c. The material properties of the cracked component.

9 Advances in Fatigue, Fracture and Damage Assessment 401 The stress-intensity factor K was introduced to define the magnitude of the local stresses around the crack tip and takes the general form given by K = β ( a) S π a, (1) where S is the remote applied stress, a is the crack length, and β (a) is a correction factor that depends on the cracked geometry and type of loading. Stress-intensity factor solutions ( β (a) or beta values) have been obtained for a variety of standard problems and are published in handbooks [7 9]. For a given material and a set of test conditions, the crack-growth behaviour can be described by the relationship between cyclic crack-growth rate da dn and the stress-intensity range K given in da dn ( K ) m = C, (2) where C and m are material constants. Figure 6 shows a crack-growth plot for 7075-T76 material [10]. The applied stress-ratio R can have a significant effect on crack-growth rates. In general, for a constant K, the higher the stress-ratio the higher the growth rate until a closure-free fully effective stress cycle is reached at which the crack-growth rate is maximum. The crack-closure arguments are often used to explain the stress-ratio effect of crack-growth rates as well as environmental effects on Kth. In addition, crack-closure theories are very important in fatigue crack-growth life predictions for components subjected to variable-amplitude loading. Elber [11] was the first to discover the crack-closure phenomenon. He observed that the surfaces of fatigue-cracks close when the remotely applied load is still tensile and do not open again until a sufficiently high tensile load is obtained on the next loading cycle. He proposed that crack-closure occur as a result of crack-tip plasticity. As the crack grows, a wake of plastically deformed material is developed while the surrounding body remains elastic. As the component is unloaded, the plastically stretched material causes the crack surfaces to contact each other before zero load is reached, Fig. 7. Elber then introduced the idea of crack-opening stress, which is the value of applied stress at which the crack is just fully open. He suggested that for fatigue crack-growth to occur the crack must be fully open:

10 402 Advances in Fatigue, Fracture and Damage Assessment Figure 6: Fatigue-crack growth rate curve for 7075-T76 material for remote and crack line-loading conditions [10].

11 Advances in Fatigue, Fracture and Damage Assessment 403 K = K K (3) eff max x min open K = K ma K, (4) consequently, K > K eff. (5) Therefore, an effective stress-intensity factor range Keff, which is smaller than K, should be used in fatigue-crack-growth predictions. Keff accounts for the R effect on crack growth rates. At higher values of R, less crack-closure results and Keff becomes closer to K because K open approaches K. min 3. Final failure. As the stress-intensity factor reaches a critical value, K c, unstable fracture occurs. This critical value of the stress-intensity factor is known as the fracture toughness of the material. The fracture toughness is dependent on the stress condition. It reaches its maximum in a plane-stress condition and decreases with thickness, through a transitional stage, until it reaches its minimum value in a plane-strain condition. The fracture toughness can be considered as the limiting value of the stress-intensity. 2.2 Fatigue analysis Background Fatigue analysis was performed in the early industrial age using the stress-based approach. The strain-based approach was later developed in response to the need to analyse fatigue problems involving fairly short fatigue lives. Subsequently, it became clear that the service loading of many machines, vehicles, and structures include occasional severe events that can best be evaluated using a strain-based approach. One example is the loading on automotive suspension parts caused by potholes, high-speed turns, or unusually rough roads. Additional examples include loading on aircraft due to gusts in strong turbulence, or due to extreme manoeuvres, especially in military aircraft where a pilot could pull 10 or 11 Gs.

12 404 Advances in Fatigue, Fracture and Damage Assessment Figure 7: Crack-closure phenomenon. In the field of aerospace engineering, the need to reduce weight and the acceptance of inelastic/plastic deformations at highly stressed locations during peak operating stresses of the aircraft, dictated the use of the strain-based approach. The approach considers the sequence of loading effect on fatigue-life in cases of plastic deformations and behaves as the stress-based approach for cases of elastic deformations. The strain-based approach employs local stresses and strains as opposed to the stress-based approach, which emphasizes nominal stresses and elastic-stress concentration factors. The fatigue-life is estimated based on these local stresses and strains, by the use of a strain versus life curve, instead of a nominal stress versus life curve. Calculating local stresses and strains, and performing fatigue analysis on a cycle-by-cycle basis, results in more accurate fatigue-life estimation. Although most engineering structures and components are designed such that nominal loads remain elastic, stress concentrations often cause plastic strains to develop in the vicinity of notches. Because of the constraint imposed by the elastically stressed material surrounding the plastic-zone, deformation at the notch root is considered to be strain-controlled. The local strain life approach has gained acceptance as a useful method of evaluating the fatigue-life of a notched component. The American Society for Testing and Materials (ASTM) has recommended procedures and approaches for conducting strain-controlled tests and using these data to predict fatigue lives [12, 13]. Early fatigue research showed that damage is dependent on plastic deformation or strain. In the strainlife approach, the plastic strain or deformation is directly measured and quantified. Failure of the component is assumed to occur when the equally stressed volume of material fails, see Fig. 8 [14]. Because of this, strain-life methods are often considered initiation life estimates.

13 Advances in Fatigue, Fracture and Damage Assessment 405 Critical zone (equally Critical stressed zone volume of material) Notch Smooth specimen Figure 8: Equally stressed volume of material [14] Methodology In this section a methodology, using the strain-life approach, a commonly used in the aerospace industry for calculating fatigue-life, is explained. Fatigue-life is calculated using the following basic information: 1. Material properties obtained from strain-controlled fatigue tests that include: a. Strain-life data. b. Cyclic stress strain data. 2. Fatigue spectrum. 3. An approach to calculate local stresses and strains. 4. A cycle-counting technique to identify damaging events. 5. A mean stress parameter. 6. A fatigue-damage summation rule. The strain-life data or the strain versus life curve is a plot of strain amplitude versus cycles to failure. Such a curve is used by the strain-based approach in making life estimates in a manner analogous to the use of an S N curve in the stress-based approach. The equation relating the total strain amplitude and life is expressed by the following equation:

14 406 Advances in Fatigue, Fracture and Damage Assessment ' f σ b ' c ε a = (2N f ) + ε f (2N f ), (6) E where, ε a is the total strain amplitude ' σ f is the fatigue-strength coefficient E is the elastic Young s modulus 2 is the number of reversals to failure N f ' ε f is the fatigue-ductility coefficient b is the fatigue-strength exponent c is the fatigue-ductility exponent Fig. 9 shows a schematic for the strain-life curve components; elastic, plastic and total strains versus life and some strain-life data for RQC-100 steel [15, 16]. The cyclic stress strain data is obtained from the stabilised hysteresis loops peaks of the strain-life curve. Data could also be obtained from an incremental step test where the strain amplitude is increased in steps to a peak level and then decreased back to initial value in repeated blocks. Cyclic stress strain data is then obtained from the peaks of the stabilised block. Fig. 10 [16] shows a schematic for the cyclic stress strain curve obtained from the peaks of the hysteresis loops. The equation relating the cyclic strain amplitude to the cyclic stress amplitude is: ε σ σ = + ' 2 2E 2K 1 n ', (7) where, σ is the cyclic stress range ε is the cyclic strain range E is the elastic Young s modulus ' K is the cyclic strain-hardening coefficient ' n is the cyclic strain-hardening exponent

15 Advances in Fatigue, Fracture and Damage Assessment 407 (a) (b) Figure 9: (a) Elastic, plastic, and total strain versus life (adapted from Landgraf [15]), (b) Strain-life curve components for RQC-100 steel [16].

16 408 Advances in Fatigue, Fracture and Damage Assessment Figure 10: Cyclic stress strain curve obtained from the tips of the hysteresis loops. Three loops are shown; A-D, B-E, and C-F. The tensile portion of the curve is OABC and the compressive portion is ODEF [16]. ' ' The constants K and n could be obtained through several steps. First, nominal stress and strain data points read off the cyclic stress strain curve are used to calculate true stresses and true plastic strains using the following equations: σ true σ ε = 1+. (8) Since true plastic strain = true total strain true elastic strain, then: ε 2 ε 2 true, p εtrue true, e ε σ true = 2 = ln 1+ where the following are defined for (8) and (9), σ true is the true stress range σ is the engineering stress range ε is the engineering strain range is the true plastic strain range ε true, p ε true,e is the true elastic strain range ε true is the true total strain range. 2 2E, (9)

17 Advances in Fatigue, Fracture and Damage Assessment 409 A power curve fitted to the true plastic-strain amplitude and the true stress ' ' ' amplitude on log-log axes is used to determine n and K. The value n is the ' slope of the fitted curve and K is the true stress value corresponding to a true plastic strain of 1.0. Several approaches could be used to calculate local stresses and strains. Neuber s rule and strain-energy density are among the most commonly used approaches with the former being more widely used. The following is an explanation of the steps employed in calculating local stresses and strains using Neuber s rule: where, K t = K σ K ε (Neuber s rule), (10) σ Kσ = is the stress concentration factor, where S σ and S are the local and nominal stress, respectively. ε Kε = is the strain concentration factor, where e ε and e are the local and nominal strains, respectively. K is the theoretical surface stress concentration factor. t Equation (10) can be applied for both nominally elastic and plastic stresses and strains. For nominally plastic stresses and strains, it becomes: 2 K t σ ε = S e (11) K 2 t Se = σε, (12) and for nominally elastic stresses and strains it becomes: 2 σ ε K t = (13) S S / E K t S = Eσε, (14) and for locally elastic stresses and strains it becomes: K t S = σ (15) K t S = σ. (16)

18 410 Advances in Fatigue, Fracture and Damage Assessment The procedure for calculating local stresses and strains is as follows: Neuber s rule is used to derive the relation between local and nominal stresses and strains K 2 t = K σ K ε 2 σ ε K t = S e 2 K t S e = σ ε (17) (18) (19) K t S e = σ ε. (20) Note that stands for range (difference between maximum and minimum). A plot of σ ε versus ε is created. First, values of the local stress range, σ, are inserted into (7) to obtain corresponding values of local strain range, ε, Then the value σ ε is calculated for each data point ( σ, ε ). A plot of σ ε versus ε can be plotted from the ( σ ε, ε ) data points. The next step is to find the local strain range in the vicinity of the crack tip, ε. The nominal strain, e, can be obtained by substituting the nominal stress, S, into (7). Knowing S and e, we can calculate the value of S e that is equivalent to σ ε as given by (20). Using S e as input, a linear interpolation performed on the ( σ ε, ε ) data would obtain the corresponding value of local strain range, ε. The corresponding local stress, σ, is obtained by substituting the value of ε into (20). A cycle-counting technique should be used to identify damaging events. There has been considerable uncertainty and debate concerning the proper procedure for cycle counting. Among the many techniques proposed a consensus has recently emerged that the best approach is to use the rain-flow cycle counting technique developed by Professor T. Endo and his colleagues in Japan in 1986, or to use other procedures that are essentially equivalent. Each damaging event, stress/strain cycle, does not produce the same damage. There are two factors playing a vital role in determining the fatigue-damage of a cycle; the amplitude/range of the cycle and the mean stress. A cycle with large amplitude and a mean stress of zero could be as damaging as a cycle of small amplitude but with a high mean stress. Among the different theories available to account for the mean stress effect are the following: K t K t

19 1. The Morrow equation: Advances in Fatigue, Fracture and Damage Assessment 411 ' f σ σ m b ' c ε a = (2N f ) + ε f (2N f ). (21) E Although this equation is consistent with observations that: a. Mean stress effects are significant at low values of plastic strains where elastic strain dominates, and b. Mean stress effects have little effect at shorter fatigue lives where plastic strains are large, yet This mean-stress correction incorrectly suggests that the ratio of elastic-toplastic strain range is dependent on mean stress, which is not quite true because the hysteresis loop shape does not depend on the mean stress. 2. The Manson Halford equation: ' ' σ f σ m b ' f m c / b a (2N f ) σ σ ε = + ε f (2N ' f ). (22) E σ f This equation tends to predict too much mean stress effect at short lives or when plastic strains dominate. 3. The Smith Watson Topper (SWT) parameter: This parameter assumes that the life for any situation of mean stress depends on the product value of σ ma x ε. Hence, the life is expected to be the same for any completely reversed loading where this product has the same value, ' 2 ( σ f ) b ' ' b+ c 2 ε a σ max = (2N f ) + σ f ε f (2N f ). (23) E This equation tends to be un-conservative for compressive loads as the equation becomes undefined when the maximum stress is less than 0. The physical interpretation of this approach assumes that no fatigue-damage occurs when the maximum stress is compressive. There is no preference in the aerospace industry for one equation rather than the other as they are all sound. However, in the CF-18 world the SWT mean stress parameter is commonly used. Fatigue-damage of each individual cycle is calculated as 1 N f, where N f is the number of cycles to cause failure by applying that cycle only by itself. Total fatigue-life is obtained when the summation of all damages becomes equal to one. Palmgren Miner s linear damage rule is the most commonly used rule for damage summation.

20 412 Advances in Fatigue, Fracture and Damage Assessment Ni N fi = 1, (24) where, N i is the number of cycles applied that share the same amplitude and mean stress N fi is the fatigue-life required to cause failure by applying that cycle only by itself. Fatigue-life is affected by other factors that should be considered in calculating allowable stresses for aircraft structures. Among these factors, which are considered in the CF-18 fatigue analysis are: 1. Product-form and grain-correction factors: Very often the material being investigated is not of the same product form or grain direction as that for the material used to calculate the allowable fatigue stress. Correction factors are usually applied to account for different grain directions and product forms. Tables 1 and 2 shows some examples of correction factors applied to the allowable design stress. 2. Material-thickness-correction factors: Research has shown that thickness of the material when being heat-treated has an influence on fatigue-life. Figure 11 shows a schematic of heattreatment thickness versus correction factors for various materials. 3. Anodized-surfaces-correction factors: Structures that have been anodized have lower fatigue lives than unanodized structures by about 14%. A correction factor of 0.86 is used to account for the anodizing effect. 4. The shallow-gradient effect: Test results have shown that fatigue-life of components with a shallow stress gradient are shorter. This is due to the greater volume of highly stressed material near stress concentrations with shallow gradients. A shallow stress gradient factor is applied to the allowable design stress. Table 1: Product-form correction factors [17]. Product form Materials Correction factor Hand forgings, 7075, 7050, Die forgings (all tempers) Plate 7075 (all tempers) 0.92 Bar, shapes 7075, 7050, 7175 (all tempers) 1.0 (L-grain direction) 0.95 (T-grain direction) All Alloy steels 1.0

21 Advances in Fatigue, Fracture and Damage Assessment 413 Table 2: Grain correction factors [17]. Grain Materials Correction factor L 7075, (all tempers) L-T 7075, (all tempers) S-T 7075, 7175 (all tempers) Correction factor 6Al-4V Ti Aluminium die and hand forgings (L,L-T) Aluminium die and hand forgings (S-T) 0.7 Aluminium Plate (S-T) 3.0 Thickness at heat treat (in) 8.0 Figure 11: Heat-treatment thickness-correction factors. 2.3 Crack-growth analysis Background Engineering analysis of crack-growth using fracture-mechanics concepts is an important analysis in the aircraft industry. It allows the estimation of safety factors based on worst-case assumptions of crack lengths and material strength. It also allows the determination of necessary inspections, thus setting a safe periodic inspection interval. In many cases, during scheduled aircraft inspections, cracks are detected. The decision taken to deal with these cracks depends on many factors such as: a. Safe operation of the aircraft in service. b. Economic issues such as maintenance/repair cost. c. Scheduling and choosing appropriate times to ground aircraft for repair.

22 414 Advances in Fatigue, Fracture and Damage Assessment Damage-tolerance analysis, DTA, is the type of analysis that, based on the fracture-mechanics approach, is used to deal with the issues described above in a, b, and c. In general, it quantifies the ability of the aircraft, or specific components, to maintain adequate residual-strength in the presence of material, manufacturing and processing defects, and damage induced during normal usage and maintenance. DTA thus enables the user to determine a safe periodicity of inspection interval until appropriate measures are taken Methodology In general, simple constant-amplitude crack-growth analysis involves performing the following steps: 1. Estimate the size of the crack likely to be present in the component when it is first put into service. The crack-growth life strongly depends on the initial crack size. Depending on the type of aircraft, whether military or civil, the appropriate standard usually regulates this size. 2. Identify the material data such as: a. Crack-growth rate data. b. Material fracture toughness properties under plane-strain, plane-stress, and transitional stages if available (otherwise plane-strain will provide conservative results). Fracture toughness is used to estimate the final crack length. c. Material static properties (i.e. Young s modulus, yield stress, ultimate stress, etc.) 3. Determine the stress-intensity geometry correction factors. The stressintensity correction factor is a function of stress fields (mode I, II, III), geometry dimensions, and crack length. It can be obtained from different sources such as: a. Handbooks for stress-intensity factors. b. Superposition of standard geometry solutions. This is usually for a single geometry subjected to different types of loading. c. Compounding of stress-intensity correction factors to estimate beta solutions for cracks in complex configurations using solutions from simple configurations. It is usually for complex geometries subjected to a single load. d. Green s and weight-function methods for geometries with non-uniform stress fields. e. Other FEM and boundary-element methods that can be used when the above methods are insufficient. 4. Substitute the stress-intensity factor equation into the crack-growth rate expression and integrate between the initial and final crack lengths to obtain the crack-growth life.

23 Advances in Fatigue, Fracture and Damage Assessment 415 da dn m ( K ) = C( β a S π a ) m = C ( ) (25) dn = C N = ac ai da ( β ( a) S π a ) m C da ( β ( a) S π a ) m (26). (27) In practice, such as in the aircraft industry, the applied spectrum is not constant in amplitude but rather of a variable-amplitude nature. Crack-growth life could be accelerated or retarded depending on the applied spectrum cycling sequence events. To quantify this effect different crack-growth-retardation models have been developed based on different concepts. Among the old models used are the Willenborg and Wheeler models. The Wheeler retardation model works by modifying the current crack-growth rate with a factor (a retardation parameter) that is based on the ratio of the current plastic-zone size and the size of the plastic enclave formed at an overload. Among the recent models, which have been showing good results, is the crack-closure-retardation model. Crack-closure models can explain the phenomenon of acceleration and retardation by using the effective stress-intensity parameter that requires predicting the variations in crack-opening stress. K = K K. (28) eff max open The crack-opening stress follows a steady-state level in the case of constantamplitude loading. An empirical model by DuQuesnay [18] quantified the steady-state level and showed its variation with values of applied maximum and minimum stresses. However, when an interruption to the steady-state level occurs by an underload or an overload the crack-opening stress changes. It was observed, by the author [19], that an underload would bring the steady-state level instantly to a low level and would require large number of the smaller cycles to bring it back to its steady state again, as shown schematically in Fig. 12. This would certainly increase the effective stress of the cycles following the underload and would induce faster crack-growth than what would have occurred without the underload interruption.

24 416 Advances in Fatigue, Fracture and Damage Assessment Stress Steady-state crackopening stress level Crack-opening stress variation Time Underload Figure 12: Variation of crack-opening stress following an underload. On the other hand, overload interruptions will increase the plastic-zone size at the crack tip and induce compressive residual stresses. These stresses tend to keep the crack tip closed in subsequent cycles resulting in lower crack-growth rates, than without the overload. Crack-growth retardation rates are dependent on the maximum stress-ratio of the overload to the subsequent cycles. Typically a ratio of 2.0 would be enough to cause crack arrest, while a ratio of 1.4 would show a very small effect. Crack arrest occurs due to the increase of the crackclosure stress to levels above the maximum stress of the subsequent cycles. This would lead to a completely closed cycle although it could be subjected to a tensile load. The flow chart shown in Fig. 13 presents the DTA/crack-growth analysis steps commonly used in the aerospace industry. The following is an explanation of some of these steps: 1. Gathering and reviewing of all available failure and crack-growth data: Usually after the design of an aircraft, and before going to production, a full-scale fatigue test is performed to validate the design and structurally certify the aircraft (i.e. confirm the safe life operation of the vehicle). Some of the cracks that occur during testing are left to grow for the following reasons: a. To collect crack-growth data that would be useful in calibrating crackgrowth models. b. To demonstrate the safe life operation of the aircraft with cracks. c. To determine the aircraft residual-strength capability with cracks. 2. Residual-strength: The residual-strength is the ability of the cracked component to withstand limit-loads without fracture. The residual-strength

25 Advances in Fatigue, Fracture and Damage Assessment 417 is a function of the critical crack length. The ability to withstand larger loads is inversely proportional to crack sizes. P xx is the load to be used in calculating the critical crack length. Civil and military standards dictate the value of the P xx load. As an example, for the CF-18 this value corresponds to 120% of the aircraft design-limit load, as defined in the CF structural integrity standard for continued airworthiness. The residual-strength is determined analytically by the minimum of elastic or plastic failures. LEFM is used for elastic failure and net section yield is used for plastic failure. Elastic failure occurs when the crack becomes unstable and the stress-intensity exceeds the material fracture toughness. The following equation applies: S ce Kc =, (29) β ( a). π. a where, S ce is the maximum allowable (critical) stress level for elastic failure K c is the material fracture toughness β (a) is the stress-intensity correction factor. It is a function of geometry, stress distribution, and crack length. Plastic failure occurs when the net section exceeds the material s yield strength. The residual-strength curve is a relation between the minimum of the maximum allowable critical stress for elastic failure and for plastic failure, and the critical crack length using (29). Figure 14 shows a sketch of the residual-strength curve. 3. Inspectable crack size: The crack lengths that are of interest in crack-growth analysis are as follows: 1. Initial crack size (induced or material flaw). 2. Detectable crack size. 3. Critical crack size. Figure 15 shows the location of the above-mentioned cracks on the crackgrowth curve. Initial crack size is used as the start length of the crack-growth phase. There are two types of flaws; the induced flaw that starts at physical damage due to manufacturing, maintenance or corrosion, and the material flaw that is due to material defects. Various regulatory standards specify values for initial primary flaw assumptions as mentioned before.

26 418 Advances in Fatigue, Fracture and Damage Assessment Detectable crack sizes are used to determine the inspection threshold. The detectable crack size depends on the inspection technique used, the geometry inspected, the access to the zone to inspect, and on a statistical evaluation of the inspector s capability. For the Canadian CP140 aircraft, for example, the initial detectable crack size typically assumed is that for 90% chance of detection with 95% confidence. Larger crack sizes may also be specified if it is felt that the crack will grow undetected due to inaccessibility or other considerations, such as a flush rivet in a part that will not be removed during inspection. The most widely used inspection techniques are either visual inspection (VI) or Non-Destructive Inspection (NDI). Visual inspections may include general, close, or special visual techniques. These techniques differ according to the equipment used such as general visual, or aided visual using lights, magnifying glasses, digital or video endoscopes, digital or video boroscopes, or remote video cameras. Non-destructive inspections include: eddy current, radiography, ultrasonic, magnetic particle, or liquid-penetrant. The critical crack size is used to determine the final crack-growth life or the termination of the crack-growth analysis. It is determined based on the residual-strength capability of the component/location analysed. The concept is that at critical crack size, a structure will not be able to withstand limit-loads. The crack then becomes unstable (fast crack-growth) at which point the component fails. Below the critical crack size, the crack propagation can be considered as stable crack-growth (slow crack-growth). 4. Inspection threshold and interval: Inspection periods can be classified into two types: 1. Inspection threshold. 2. Inspection interval. Inspection threshold is the time at the end of the period of crack-growth during which a crack is assumed undetectable. It is therefore uneconomical to inspect during that period. Inspection can begin after the inspection threshold is reached. Cracks at the threshold inspection are usually small in size. The inspection interval is the repeated inspection after the threshold inspection. It is used to allow the detection of slow (stable) growing cracks before reaching unstable growth and fast fracture. Inspection interval frequency is determined based on the crack-growth life between the detectable crack and residual-strength limit. We will call this value (N RS N NDI ). Appropriate factors are then applied to ensure adequate probability of detection. These scatter factors are obtained from appropriate civil or military standards.

27 Advances in Fatigue, Fracture and Damage Assessment 419 Inspection Interval = N N NDI Scatter Factor RS. (30) Gather and review all available failure and crack-growth data - Test inspection data - In-service failure data - Fractographic data Generate Rain-Flow counted spectrum Geometry Factors (β) -Handbooks -Green and weight-functions -FEM Identify Material Data da/dn, K c, etc. Residual-Strength - Pxx - Critical crack length Determine parameters for CG software - Retardation parameter - Pxx, etc. Crack-Growth Analysis Software Calibration or Validation Inspection threshold and intervals Apply appropriate SF Inspectable crack size Figure 13: Damage-tolerance analysis methodology

28 420 Advances in Fatigue, Fracture and Damage Assessment Residual-Strength Elastic Failure (LEFM) Plastic Failure (Net section Yield) Residual-Strength Curve Crack Length Figure 14: Sketch for residual-strength curve. Crack size a RS Critical size a NDI Initial size Detectable size N NDI Flight Hours N RS Figure 15: Crack lengths on a crack-growth curve.

29 3 Full-scale fatigue testing Advances in Fatigue, Fracture and Damage Assessment 421 The Technical Airworthiness Manual (TAM) requires that each aircraft should comply with the manual for structural airworthiness. One of the ways to structurally certify the aircraft is to perform a full-scale fatigue testing and follow it by a residual-strength test. The OEM of an aircraft should prove the certification of the aircraft for safe operation either analytically, which would require the use of analytical safety factors, or by test demonstration. The fullscale fatigue testing should then be performed for the following purposes: 1. Ensure continued airworthiness of the critical aircraft structure. 2. Determine the safe life of the critical aircraft structure through the completion of the test with duration equal to the test factor multiplied by the required service life of the aircraft. Test factors include different factors as will be shown in later sections. 3. Determine the economic life of the aircraft structure. 4. Develop engineering data on aircraft structures whenever possible. So many engineering data could be obtained from a full-scale fatigue test, some of which are listed below: a. Crack-detection times for critical locations by performing appropriate inspections. b. Monitoring of cracked components to collect crack-growth data to support fleet strategies for flying with cracks. Crack-initiation times and accurate crack-growth rates could be detected/interpreted after performing quantitative fractography of cracked components following test and RST end. c. Performing unit-load strain surveys and ground calibrations to understand the response of the structure to external loading. This would be very useful in obtaining transfer functions for spectra generation. Transfer functions could also be developed for hot spots or critical points identified during the test, and later compared to FEM predictions for validation. d. Certification of the preventative modifications that are to be applied on the fleet aircraft. Preventative modifications are applied to locations that are expected to fail before a complete service life of the aircraft. There could be other reasons to conduct a full-scale fatigue testing, depending on the purpose of the test. For example a test might be repeated if the operating conditions and the usage on which the airframe was originally tested has been changed, or if big structural modifications were performed on production aircraft after completion of the first full-scale testing and would require additional testing for certification.

30 422 Advances in Fatigue, Fracture and Damage Assessment To understand more about the full-scale fatigue testing and to appreciate the effort involved, it is thought that the best way is to give a practical example from real life. The Canadian Forces and the Royal Australian Airforce (RAAF) are currently collaborating in the International Follow-On Structural Test Program (IFOSTP) to conduct a full-scale fatigue test for the CF-18. The following sections will show some of the effort that is currently going on with some emphasis on the FT245 full-scale fatigue testing of the CF-18 wing. 3.1 CF-18 full-scale fatigue testing Introduction The Canadian Forces (CF) ordered 138 F/A-18 from McDonnell aircraft company (McAir)/Northrup for delivery starting The design was intended to last 6000 h under severe US Navy usage. This life was determined by the manufacturer after conducting full-scale fatigue tests. These tests were required to certify the safe life of the F/A-18 based on the USN operating conditions. The CF concerns about the representativeness of the McAir test with respect to CF operations has initiated the thought of performing a full-scale fatigue test other than what has already been performed by the Original Equipment Manufacturer (OEM). The most important issues in assessing the representativity of the OEM structural fatigue test are the structural configuration and the test spectrum. Having an unrepresentative configuration and spectrum increases uncertainty about the relevancy of the test results to operational usage. The CF usage data and studies have shown that the OEM test spectrum is not representative. Besides, the OEM test failures have resulted in fleet retrofits and in productionline changes that did not make it back to the test article for testing. These uncertainties provided a ground to the need for IFOSTP testing. The IFOSTP test is divided into three parts; the RAAF are responsible for the aft-fuselage, and the CF are responsible for the centre fuselage and the wing. The centre-fuselage testing was contracted to Bombardier Aerospace and has been completed with SFH. The Institute for Aerospace Research (IAR) at the National Research Council of Canada (NRC) was tasked to carry out the wing test with technical support from Bombardier Aerospace. Testing is currently going on with more than SFH of testing accumulated to date and is expected to end by the year The aft-fuselage testing was conducted by the RAAF and has been recently completed with SFH. The aftfuselage test have also been followed by RST testing and have passed it successfully. The CF-18 airframe was designed based on the safe-life approach. Consequently, the aircraft should be retired after full life as determined by the full-scale fatigue test. However, due to the severe usage of the CF to the airframe some major structural parts are expected to fail before the end of the service life. The centre-fuselage test (FT55) determined that the major 3 bulk heads will not

31 Advances in Fatigue, Fracture and Damage Assessment 423 meet the full life target and the CF is considering major structural modifications to the centre fuselage or implementing a Central Barrel Replacement (CBR) to part of the CF-18 fleet. The wing test (FT245), which is currently going on at NRC, will determine and certify the possibility of taking the CF-18 wing to the full service life. The wing and aft-fuselage vertical stabiliser are multi-spar structures and consequently are considered load redundant. One of the approaches that could be applied to the wing and vertical stabiliser for structural certification, and would lead to economical outcomes, is the damage-tolerance approach, or safety-by-inspection. The aircraft safety for certain components would then be managed using fracture-mechanics to ensure that cracks will not grow to final fracture before being detected through periodic inspections. Simpson et al. [20] presented a summary of the benefits from the collaborative IFOSTP work giving some background about the aircraft test set-up, usage spectrum selection, test systems and more. This paper by Simpson et al. would be a good source of more in-depth discussions about the IFOSTP testing Wing testing The FT245 full-scale fatigue test is planned/expected to go through the following phases [21]: a) Definition phase - During this phase, all the test requirements and criteria are clearly defined, including load-spectra development and test-rig design. This phase has been completed successfully. b) Test phase - During this phase, the actual full-scale test activities shall be performed to the specifications defined in the previous definition phase. This phase is currently in progress; and c) Post-test phase - This phase encompasses the post-testing activities of residualstrength demonstration and tear-down. This phase is expected to be completed by Load-spectra development It was agreed upon between CF and RAAF, who are involved in the IFOSTP program, that wing loads are considered as being represented by accurately defining the interface load components at 12 locations. A list of these interface loads is given in Table 3, in approximate order of importance, while their corresponding location is given in Fig. 16 [22]. In this section a quick overview of the load-set development is given. The usage spectrum for the FT245 test is called IARPO3a. This spectrum was obtained from the Maintenance Signal Data Recording System (MSDRS) of the CF-18 and consists of data from 236 unique flights. The data was arranged to form a sequence of 279 flights corresponding to 326 Spectrum Flight Hours (SFH). IARPO3a is a compromise spectrum for CF and RAAF usage. The manoeuvre sequence in the IARPO3a spectrum was generated at 10 Hz and contains about 10 million lines. The full sequence was

32 424 Advances in Fatigue, Fracture and Damage Assessment huge, many more than 10 million lines, mainly due to dynamic loads, and consequently a 1% rise/fall truncation was applied to obtain a baseline sequence. This sequence was named 99LD and contained about 25 million lines. The 99LD is considered an accurate representation for the CF fleet operation. The 99LD was subjected to several consecutive truncations and filtering to achieve a representative/reasonable sequence, in terms of number of lines, for fatigue-testing purposes. The following are the processes applied: 1. A rise/fall truncation for each of the 12 primary interface load components of the 99LD at various ratios. The 99LD lines were reduced to lines and the new sequence was named 99LD-Residual. 2. The 99LD-Residual loads were distributed, balanced, and delivered as 90LD. The 90LD load set included new derived interface-load components, such as the shear, with a total of 43 components per wing from which only 20 were used. A list of the additional 8 derived interface load components used is given in Table 4. The 90LD sequence contained the same number of lines as the 99LD-Residual, lines. 3. The 90LD, which contained unique load lines, was binned, by grouping similar loads into bins, to reduce the spectrum to less than unique lines for testing purposes/limitations. The binning process resulted in unique load lines. When these loads were inserted back into the sequence some load lines were redundant, as they were between a peak and a valley, and were removed/filtered out. This filtering process resulted in a spectrum with lines that was named FT245 target. 4. A jack-load optimisation process led to substitution and removal of unrepresentative load cases. This process resulted in a spectrum that was named FT245 jack loads. 5. Applied loads on the FT245 test were fed back during a block of testing and the measured spectrum was named FT245 applied. Table 3: Wing interface loads. Interface load Right Wing-Root Bending Moment Right Wing-Fold Bending Moment Right Inboard Leading-Edge Flap Hinge Moment Right Outboard Leading-Edge Flap Hinge Moment Right Trailing-Edge Flap Hinge Moment Right Aileron Hinge Moment Right Trailing-Edge Flap Outboard Vertical Lug Load Right Aileron Outboard Vertical Lug Load Right Wing-Tip Torsion Right Inboard Pylon Rolling Moment Right Wing-Root Torsion Right Wing-Fold Torsion Abbreviation RWRBM RWFBM RILEFHM ROLEFHM RTEFHM RAILHM RTEFOLZ RAILOLZ RWTTOR RIBPMX RWRTOR RWFTOR

33 Advances in Fatigue, Fracture and Damage Assessment 425 Fwd LEX X Y 0.00 Z Mid Wing X ; 487.0; Y ± Z AFT Fuselage X Y 0.00 Z Fixed LEX X Y 0.00 Z Aft LEX X Y 0.00 Z Wing Root X FWD Fuselage Y ±38.50 X Z Y 0.00 Z ILEF Inner Idler X Y ±59.87 ILEF Outer Idler Z OLEF Inner Idler X X Y ± Y ± Z Z H-Tail X Y ±41.90 Z V-Tail TEF Inner X Hinge Y ±38.40 X TEF Outer Z Y ±47.42 Hinge Z X Y ± Z Aileron Inner Hinge TEF Deflection Limiter X X Y ± Y ± Z Z Wing Fold X ; ; Y ± Z Aileron Outer Hinge X Y ± Z Wing BL217 AFT Launcher X ---- X Y ± Y ± Z FWD Launcher Z OLEF Outer Idler X X Y ± Y ± Z Z Figure 16: Wing-interface loads locations [22]. Table 4: Derived wing-interface loads. Derived interface load Right Wing-Root Shear Right Wing-Fold Shear Right Inboard Leading-Edge Flap Shear Right Outboard Leading-Edge Flap Shear Right Trailing-Edge Flap Shear Right Aileron Shear Right Wing-Tip Shear Right Wing-Tip Bending Moment Abbreviation RWRSH RWFSH RILEFSH ROLEFSH RTEFSH RAILSH RWTSH RWTBM Test inspection Inspections are the fundamentals of structural fatigue-test programs. They play a vital role in completing the test without long delays due to failures and collecting valuable durability and damage-tolerance data. Thus the aim of test inspection can be summarized in the following [23]: 1. To support management of fleet aircraft by detecting and monitoring fatigue-induced damage in order to obtain crack-growth information, validate/certify modifications and repairs, and obtain engineering data. 2. To evaluate standard NDI methods that can be used for in-service aircraft. 3. To validate load/spectrum severity.

34 426 Advances in Fatigue, Fracture and Damage Assessment The FT245 first passed through a pre-test inspection activity to perform a baseline inspection of all test articles to be used throughout the test. The pre-test inspection was also meant to serve as a comparison tool for structural discontinuities found during testing and to ensure the structural integrity of the aircraft components prior to test start. To build the FT245 inspection card deck the areas of inspection were first identified, based on failures reported on previous tests, the suitable inspection method was then selected, and then the cards were created. The baseline inspection card deck for the wing test consisted of 175 cards. Few additional cards were added during the early course of the test during inspection cards review and rationalization. Different NDT methods are used for inspection that includes visual inspection, ultrasonic, eddy current, radiography, magnetic particle, and liquid-penetrant techniques. A description of each technique is provided in Table 5. Inspection periods can be classified into two types; inspection threshold and inspection interval. Inspection threshold is the first inspection to be performed while inspection interval is the repeated inspection after the threshold inspection. Inspection periods are used to allow the detection of growing cracks before reaching unstable growth and fast fracture. Inspection thresholds and intervals can dictate the speed of testing. Too many inspections will require prolonged downtime periods, while too few inspections can risk the test safety with fractures and long downtimes for repair in addition to overloading adjacent structures and losing test-data information. Also, too few inspections does not allow enough preliminary crack-growth data to be obtained and used in fleet management early during the test program. The term early was used because the crack-growth data could always be obtained from quantitative fractography, but later after test end and RST, which does not allow the early use of data if needed for the fleet. This case only applies when the fullscale fatigue test is being performed while the fleet is in-service and is depending on some of the test-data to support future strategies. The baseline FT245 inspection card deck was passed through a detailed analysis for the critical cards only. The detailed analysis was aimed at minimising inspection downtime, while ensuring test article and transition structure safety, as well as the adequacy of the gathered information. The effort done through the detailed analysis of critical cards has shown a substantial relief of the FT245 test inspection burden with a total LOE decrease, for the critical cards reviewed, of 47% [24]. The FT245 card deck was also passed through an optimization process to minimize downtimes by grouping inspection cards and performing the inspection for those within a certain area/group at the same time. As the CF-18 aircraft is designed to withstand fatigue-failure using the safelife approach, and as it is approaching its demonstrated safe life, structural management using the safety-by-inspection approach (Damage-Tolerance) could be implemented for certain regions of the aircraft. The approach could be considered in many circumstances, and could lead to avoiding costly repairs and/or component replacements. As such, the requirement to use DTA in managing the wing and aft-fuselage has increased recently, especially for those

35 Advances in Fatigue, Fracture and Damage Assessment 427 with redundant structures. The crack-growth data obtained from monitoring cracks through inspection intervals on the test and/or fractographic analysis of the fractured surface after tear-down inspection would be a key to aircraft management using safety-by-inspection. Table 5: FT245 test NDT techniques. Technique General Visual Inspection (GVI) Close Visual Inspection (CVI) Special Visual Inspection (SVI) Eddy Current (ET) Radiography (RT) Ultrasonic (UT) Magnetic Particle (MT) Liquid-Penetrant (PT) Description A straight visual wide-area inspection looking for gross deformation, large cracks or broken fasteners Inspection of limited area using aided visual inspection equipment such as lights, magnifying glasses and digital endoscopes with a magnification of 100 Inspection of a specific location using video endoscopes, boroscopes, video boroscopes, and remote video cameras A passive method for detecting surface and sub-surface flaws in conductive materials Used in cases of limited inspection access to the wing structural components A passive method (with couplant) that can be used on any material. It is the primary method of inspection for composite skins A good real-time method for finding surface and subsurface flaws in ferromagnetic materials Was used as a means of confirming damage found through other NDT methods Test-data types The FT245 test consists of a series of repeated blocks with 326 SFH each. The test is monitored through strain-gauges attached at different locations, mainly on the wing with others at the center fuselage. Several types of data are generated and recorded during the test. Each type of data provides useful information about the test article. These data types consist of the following: 1. Strain-gauges a. 512 channels on MDAC (low-speed recording) b. 60 channels on IDAC (high-speed recording). Strain-gauge data is used for trend monitoring and transfer-function calculations. Data health checks are performed for every block of testing, that helps in identifying any structural loading changes through the test course.

36 428 Advances in Fatigue, Fracture and Damage Assessment displacement transducers 3. 6 reaction loads command and feedback jack loads. Fig. 17 [25] shows the different locations of the 63 load jacks. Load is distributed on the wing through a whiffle-tree ending with pads attached to the wing upper and lower skins. Fig. 18 shows a picture of the whiffle-tree arrangement of the FT245 test targeted and applied jack interface loads Strain-gauges on the FT245 test are very useful in support of fleet aircraft and test management. They are useful in many applications such as test-trend monitoring checks. If measured strains, through the test course, are not within the predicted limits (based on values recorded in previous blocks), the test operator is notified. The trend monitoring can serve as a protection for the test from sudden failures, and as a warning for component cracking and/or drift in strain-gauge data. Strain gauges are also useful in strain surveys and influence lines or unit jack-load surveys. The data obtained is very useful in the calculation/determination of transfer-functions that is greatly supporting the Aircraft Structural Integrity Program (ASIP). TEF AIL AFT1 FA9 FA8 WA WT02 WT FA7 FA6 FA5 SWF6 SWF4 SWF2 SWA1 SWA3 OLEF ILEF 4 1 FWD1 1 LEX1 FA4 FA3 SWA5 4 1 OPF IPF FA2 FS+ FUSE STUB WING PYLONS WING FA1 BL+ Figure 17: FT245 test 63 jack load actuators [25].

37 Advances in Fatigue, Fracture and Damage Assessment 429 Jacks Fuselage Fwd Outbd Pads Figure 18: Whiffle-tree arrangement on the FT245 wing test. 3.2 Structural certification The safe life of an airframe or associated structural components are not considered certified if it can not demonstrate, by test or analysis, sufficient residual-strength following the completion of fatigue testing. Therefore, one life time of fleet service can be structurally certified if: 1. Aircraft components have demonstrated no failures after being tested for not less than one lifetime of usage amplified by the test factor. Test factors considers uncertainties in manufacturing, material properties, fatiguedamage monitoring and corrections for test-loading inconsistencies. 2. Aircraft components have demonstrated sufficient residual-strength following the test end. Usually, usage spectra are very long and can not be tested as is. Therefore, the spectrum has to undergo various truncations, filtering and balancing processes to size it down to a usable size for testing, as explained previously. The test-applied spectrum would then have to be interpreted to evaluate its severity at different locations on the aircraft and project the required test end for structural certification. The interpretation process could be extensive and would include the following: