The lead crack fatigue lifing framework
Introduction
Accurate prediction of the fatigue lives of metallic airframes still presents challenges, particularly for high performance aircraft. There is always a demand for lighter structures with reduced manufacturing and operating costs. This leads to relatively highly stressed and highly efficient designs where fatigue issues can arise at features such as shallow radii at the junction of flanges, webs and stiffeners, as well as at holes and tight radii. As a consequence, there are usually many areas that need to be assessed for their fatigue lives, and many potential locations at which cracking may occur in-service.
It is well-known that fatigue is a complex phenomenon that is dependent on many parameters, including the material characteristics (mechanical properties, microstructure and inherent discontinuities, e.g. constituent particles), surface treatments and finishes, the component and structural geometries, dynamic load histories and the environment. Nevertheless, engineering fatigue design relies in-part on baseline coupon tests to assess the many locations identified as susceptible to cracking. The coupons may be loaded by constant amplitude (CA) or representative variable amplitude (VA) load histories, and they may try to represent some feature of a built-up structure. The results of these coupon tests are averaged to give an indication of the life of the structure in a production aircraft. However, there are significant limitations to this approach:
- (1)
Experience has shown that, in high performance aircraft, the structural components have many features with the potential to crack, and that each of these features is typical of a single type of (more-or-less) representative coupons. Hence, the average indicated life of a component is equivalent to only the shortest average life from tests on several types of coupons.
- (2)
Even when the most critical feature of a component has been identified and assessed by coupon testing, the coupons are rarely fully representative, notably with respect to the surface treatments and finishes required for production aircraft. This is important because the commencement of fatigue cracking is primarily surface-influenced and therefore greatly dependent on small surface discontinuities inherent to component production, as well as any surface-connected discontinuities inherent to the material.
These limitations are addressed by other means. One way, which is mandatory for all modern aircraft, is to test actual components, part of the structure or even the full airframe, thereby including the effects of component geometry and production. Another way is to improve coupon testing by making the coupons optimally representative of the most fatigue-critical details, e.g. by applying surface treatments and finishes used in component production. This may seem obvious, but it is sometimes neglected or overlooked.
The RAAF methodology for lifing aircraft primary structures, e.g. [1], requires establishing the fatigue test life, under representative loading, of a full-scale structure or major component to a residual strength (RS) requirement of 1.2× Design Limit Load (DLL) without failure. Whether the test structure fails below 1.2DLL or survives, it is necessary to determine the equivalent fatigue life defined by the ability of a structural detail to achieve and survive 1.2DLL with cracking present. In other words, the test time to the critical crack length/depth (aRS) at the RS ⩾ 1.2DLL point is required.
For a crack that fails the structure below 1.2DLL the fatigue crack growth (FCG) life is assessed analytically and reduced to a time at which it would have reached the calculated aRS value for a RS = 1.2DLL. For those cracks that survive the RS test load some assessment of the remaining amount of life may be needed. This depends on several factors:
- (1)
During a complex full-scale fatigue test, it is often necessary to ensure the survival of the test article by removing or modifying cracked locations when the cracks are smaller than the calculated aRS values. These locations become the subject of fleet action prior to the overall life established by the fatigue test, but it may be possible to gain some additional life before the fleet action. This is checked by calculating the remaining FCG life to an aRS, thereby establishing a virtual test life (virtual test point) for fleet action.
- (2)
Although the test may in general establish adequate fatigue lives, it is often not possible to apply representative load histories in all areas. When cracks form at locations in non-representatively loaded areas it may be necessary to calculate the definitive FCG life to aRS and establish additional virtual test points. Such calculations require detailed knowledge of the FCG behaviour under representative and non-representative load histories.
- (3)
Finally, the load histories experienced by the fleet may turn out to be significantly different to the load histories assumed and applied during testing. As before, such differences may require further analysis of the cracks found during testing, in order to establish new equivalent test lives and virtual test points.
Each of these scenarios needs a framework of rules under which FCG predictions can be made with the aid of data from coupon, component and full-scale fatigue tests. However, before proceeding to this topic, which is the main theme of the present paper, methods of establishing the FCG lives are concisely discussed. This is because there is a major and fundamental difference between the method employed in the Damage Tolerance (DT) concept developed by the United States Air Force (USAF) [2] and the currently proposed and used DSTO method.
Both the USAF DT and DSTO methods assume that defects (cracks, flaws and discontinuities) are already present in new structures, and that these defects must be treated as cracks that are immediately capable of growing by fatigue under service load histories. However, beyond these assumptions there are major and fundamental differences.
For critical locations the DT method specifies initial flaw/crack sizes and shapes based on pre-service Non-Destructive Inspection (NDI) capabilities and the assumption that cracks grow soon after the aircraft is introduced into service. The minimum assumed crack dimension is about 0.5 mm, see Table 1. Soon after these requirements were introduced, there was some debate about their arbitrariness and unknown conservatism [3]. Subsequently, Lincoln [4] stated that more than ten years of data collection had validated the requirements. However, there is now a consensus that these requirements can lead to predicted FCG lives that are too conservative. This has led to setting up the RTO working group AVT-125 “Future Airframe Lifing Methodologies” within the NATO community, and with which the DSTO confers and participates.
For continuing damage1 and non-critical locations the DT method specifies much smaller initial flaw/crack sizes of about 0.127 mm, but allows the aircraft manufacturer to change these requirements if actual initial flaw/crack size information is available - which has rarely been the case until recently [5], [6].
Be that as it may, all the specified USAF DT initial flaw/crack sizes are based on NDI limits and not measured discontinuity sizes. For continuing damage and non-critical locations the predicted early FCG behaviour may also be questionable since it is derived from back-extrapolation of (a) VA long crack growth data or (b) VA growth curves derived from long crack CA data, with both methods using analytical models “tuned” to long crack growth behaviour. These issues of initial flaw size and VA crack growth determination from CA, together with potentially overly-conservative predictions of FCG lives for critical locations, constitute potential limitations to the DT method.
The DSTO method has been developed from many years of detailed quantitative fractography (QF) of fatigue cracking in metallic airframe materials and structures, ranging from coupon to full-scale fatigue tests, and also including components removed from service. The QF observations covered crack sizes from a few micrometres up to many millimetres and showed that the cracks originated from small discontinuities inherent to the material and component’s production. These discontinuities are discussed in Section 3 of this paper.
For high performance aircraft the detailed QF observations were - and are – essential to determine the FCG rates, particularly where most of the life of a fatigue crack is spent as a relatively small crack. QF data make it possible to (a) characterize the crack-generating discontinuities and their populations, (b) account for variability in small crack FCG behaviour, a notorious problem that is difficult or impossible to tackle in any other way, and (c) predict total lives from larger or smaller discontinuities. All of this information can be used to make more accurate predictions of FCG lives in-service. Furthermore, a key point is that the lead cracks began to grow shortly – effectively immediately – after the test coupons, components, full-scale structures and service components are subjected to fatigue loading.
Based on these observations, and applying a lead crack concept which assumes that lead cracks in production quality aircraft components and structures immediately begin to grow under service load histories, the DSTO has developed a service component lifing approach based on the lead crack fatigue lifing framework. Under (or using) this framework a methodology has been implemented as an additional tool to determine component fatigue lives for several types of aircraft in the RAAF fleet.
This paper summarises the framework, with examples from test programmes used in lifing RAAF aircraft. Examples of typical crack growth behaviour are also presented. Further details can be found at Ref. [7].
Section snippets
Lead cracks
If a particular region of a structure has the propensity to crack, it is possible that a number of cracks will nucleate and grow. The crack that grows fastest in this region is the lead crack. Since there will probably be a number of regions across an entire structure that will crack, there will most likely be a number of lead cracks across the entire structure and one of these will ultimately cause the failure of the structure.
Many observations at the DSTO e.g. [8], [9], [10], [11] and by
Ideal and realistic (service and production quality) conditions
Under more or less ideal conditions, typically for carefully prepared specimens4 tested in laboratory air, a significant period of fatigue-induced microstructural damage can precede fatigue crack initiation. This may also be the case for highly finished engine components and low-stressed parts in secondary structures and helicopters, although service conditions can allow other mechanisms like corrosion, fretting and
Approximately exponential crack growth
As mentioned in Section 2, many observations have shown that approximately exponential FCG commonly occurs for naturally-initiating lead cracks. In fact, the observation of exponential FCG has a long history, going back to the 1950s [25], [26]. This behaviour is described by the following simple relationships:
These relationships mean that the FCG data appear to be well represented by straight lines on plots of (log) crack size versus life and (log) FCG rate versus
Framework
The lead crack fatigue lifing concept has been developed on a framework of observations to calculate virtual test points from full-scale and large component tests, unanticipated service cracking, and sometimes from cracking in representative coupons. The key elements of the framework are (see Section 2.1):
- (1)
Lead cracks start to grow shortly after testing begins or the aircraft is introduced into service and subjected to flight loads.
- (2)
Lead cracks grow approximately exponentially with load history.
- (3)
Conclusions
This paper has presented a lead crack fatigue lifing framework and methodology for metallic primary airframe components. The framework is based on many years of detailed inspection and analysis of fatigue cracks in airframe materials and structures, ranging from coupon to full-scale fatigue tests and also including components removed from service. This framework and its exploitation provide an important additional tool for accurately determining aircraft component fatigue lives from detected
References (31)
- et al.
An experimental evaluation of fatigue crack growth
Eng Fail Anal
(2005) - et al.
A comparison of crack growth behaviour in several full-scale airframe fatigue tests
Int J Fatigue
(2007) - et al.
Understanding crack growth in fuselage lap joints
Theor Appl Fract Mech
(2008) - et al.
A method for evaluation of in-service fatigue cracks
Eng Fail Anal
(2005) - et al.
The effect of dynamic loading on the fatigue scatter factor for Al 7050
Int J Fatigue
(2008) - et al.
Prediction of fatigue life with interspersed mode-I and mixed-mode (I and II) overloads by an exponential model: extensions and improvements
Eng Fract Mech
(2009) - et al.
Experimentally derived crack growth models for different stress concentration factors
Int J Fatigue
(2008) - et al.
Flight-by-flight fatigue crack growth assessment
Int J Fatigue
(2007) Fatigue crack growth from flaws in combat aircraft
Int J Fatigue
(2010)- et al.
What is fatigue damage? A viewpoint from the observation of low cycle fatigue process
Int J Fatigue
(2005)
Observations of fatigue crack initiation in 7075–T651
Int J Fatigue
The propagation of fatigue cracks in test specimens
J Mech Phys Solids
Characterization of pitting damage and prediction of remaining fatigue life
Int J Fatigue
Cited by (94)
Equivalent initial damage sizes (EIDS) for type 1C anodised aluminium alloy 7085-T7452 under variable amplitude loading
2023, Engineering Failure AnalysisPredicting fatigue crack growth through the small and long crack regimes for a military transport aircraft loading spectrum using FASTRAN
2023, International Journal of FatigueA Building Block Approach to Sustainment and Durability Assessment: Experiment and Analysis
2023, Comprehensive Structural IntegrityMicrostructure and surface finish influences on AA7085-T7452 small fatigue crack growth rates
2022, Engineering Failure AnalysisThe effect of crack length and maximum stress on the fatigue crack growth rates of engineering alloys
2022, International Journal of FatigueCitation Excerpt :Jin et al. concluded that fatigue cracks in 7075-T651 nucleated at pre-cracked Al7Cu2Fe constituent particles when the specimen was loaded in the L direction and crack initiation from this phase has been observed for high strength aluminium alloys by many authors [63,85–90]. Barter et al. expected immediate fatigue crack growth from these pre-cracked particles under high cyclic stress and observations on crack nucleation from corrosion pits indicated that crack nucleation from less sharp stress concentrator occurs essentially immediately upon the application of cyclic loading [90–93]. Therefore, it is expected that crack growth from pre-cracked or cracked Fe-containing particles occurs from the first loading cycle for highly loaded smooth S-N specimens.