10:50
Session 5: Fatigue crack growth and life prediction methods II
Chair: Antoni Niepokólczycki
10:50
20 mins
|
Utilizing condemnations, retirements and mods to improve the structural risk analysis of the T-38: turning trash into treasure
Marcus Stanfield, Laura Hunt, Isaac Grothe
Abstract: With the T-38 Talon entering its fourth 20-year lifetime, engineering has increasingly relied upon a robust risk analysis to ensure safety of flight. This presentation will outline how the T-38 engineering team has implemented a successful risk management plan that utilizes teardown results from fleet condemnations, retirements and major modification programs to achieve structural safety.
Along with teardown failure analysis of large cracks from condemnations, T-38 engineering has identified and executed numerous retired wing teardowns to measure and record the smaller crack findings as well. In addition, simple teardown and failure analyses were conducted on structural items removed and replaced during a major modification fuselage life extension program. These efforts were able to capture valuable data on fatigue critical, life limited parts that would have otherwise been discarded. A thorough failure analysis of detected cracks allows an analyst to build a robust Equivalent Initial Flaw Size (EIFS) distribution dataset for probabilistic risk analysis. This marked a move to a more proactive, rather than reactive, risk management plan. Combined with maintenance records, T-38 now has an extensive library of thousands of findings covering several wing and fuselage Fatigue Critical Locations (FCLs). This allows the creation of FCL-specific EIFS distributions rather than a general detail (e.g. fastener hole) or purely material-based distributions.
The risk analysis process, including gathering teardown failure analysis data, building an EIFS distribution, and PROF (Probability of Fracture software) analysis, will be shown in detail for three locations. For the first case study, a set of condemnation data was used to develop an EIFS distribution that resulted in a Uniform distribution versus the typical Weibull distribution. The second case study at a wing FCL had teardown findings contrary to previous damage tolerance assumptions. Incorporating the teardown findings produced a probability of failure much lower than previously calculated. The third case study will show how historical teardown data was used to enhance the risk analysis for a previously unknown fuselage FCL.
|
11:10
20 mins
|
The lead crack concept history 30 years on
Loris Molent, Simon Barter, Russell Wanhill
Abstract: Since the early 1990s extensive quantitative fractography (QF) of early nucleation and growth of airframe fatigue cracks that have resulted in accidents, or led to threats to structural safety, have shown that the largest – or “lead” ‒ cracks often show approximately exponential growth. These observations have been formalised in the lead crack concept and development of the Lead Crack Fatigue Lifing Framework (LCFLF) [1-3]. This has now progressed to a stage where it provides a robust and appropriately conservative method of assessing crack growth, enabling the setting of inspection and maintenance periods for a variety of in-service fatigue cracking problems in aircraft. The framework has been used extensively in airframe life predictions and or life extensions by numerous airworthiness authorities.
The LCFLF relies on determining the equivalent crack-like sizes of the fatigue-nucleating discontinuities and the crack depths at known points in the fatigue lives. This method is flexible in that it may be pragmatically combined with fracture mechanics models of crack growth, provided they have been verified by actual measurements [3].
This paper summarises the current knowledge state for the LCFLF and provides some new examples of the framework directed at determining the crack growth history from limited quantitative fractography or in-service crack length measurements. It has increasingly become an important tool for aircraft sustainment and fatigue failure analyses.
[1] Molent L, Barter SA, Wanhill RJH (2010) The lead crack fatigue lifing framework, DSTO Research Report DSTO-RR-0353, DSTO Defence Science and Technology Organisation, Fishermans Bend, Victoria 3207, Australia.
[2] Molent L, Barter SA, Wanhill RJH (2011) The lead crack fatigue lifing framework, International Journal of Fatigue, 33: 323-331.
[3] Wanhill RJH, Molent, L. Barter, SA (2019) Fatigue Crack Growth Failure and Lifing Analyses for Metallic Aircraft Structures and Components, SpringerBriefs in Applies Sciences and Technology, Springer Nature B.V., Dordrecht, The Netherlands.
|
11:30
20 mins
|
Fatigue performance and DADT certification of powder-bed additively-manufactured Ti-6Al-4V: defect assessments, EIDS distributions, and inspection limits
Matthew Krug, Reji John, Sushant Jha, Patrick Golden
Abstract: Powder-bed fusion additive manufacturing (PFB AM) of metals has reached an intermediate stage of technological maturity. Parts manufactured by PFB AM are considered by OEMs and end-users for a growing set of applications. To date, however, few of those applications have been fracture critical components due to uncertainties inherent in durability and damage tolerance (DADT) certification of AM metals. Certification considerations specific to AM include fabrication system-to-system variability, updates to hardware and build-parameter-sets that occur at a frequency similar to those for personal electronic devices and software, and difficulties in non-destructive defect inspections of complex geometries to name only a few.
The US Air Force relies upon the Equivalent Initial Damage Size (EIDS) distribution to characterize the quality of an aircraft structure upon manufacture. Although several decades of experience with this approach inform certification of structural components produced by conventional fabrication modes (cast and wrought parts), there is to our knowledge currently no available published EIDS data for PBF AM metals. In this work we report on our measurements of EIDS distribution data for PFB AM Ti-6Al-4V. EIDS distributions are in development for several hundred fatigue samples across which fabrication modality (laser- vs. electron-beam energy source), sample inclination to the build direction, surface condition, heat treatment, and other variables are modified. To assess the nature of the crack-nucleating defect in a given specimen, and to relate its size to the measured EIDS value, a series of destructive and non-destructive characterization activities is undertaken, including computed tomography, profilometry, and serial sectioning, among others. The results are interpreted in a certification context by comparison with data developed in parallel for mill-annealed Ti-6Al-4V plate forgings.
|
11:50
20 mins
|
Can we predict fatigue crack growth without the help of Paris?
René Alderliesten, Jesse van Kuijk
Abstract: Since the introduction of the so-called ‘Paris law’, named after Paul Paris, many engineers and researchers have successfully predicted fatigue crack growth in metallic structures utilizing the Paris curves. After this introduction, various versions of the Paris relation have been proposed and used, of which the NASGRO relation and similarly the Hartman-Schijve relation are famous examples.
One can debate, however, whether this practice really constitutes predictions, considering that always first sufficient number of fatigue crack growth tests must be performed through which these phenomenological relations are fitted. Essentially, these predictions constitute interpolations between pre-existing fatigue crack growth data, which are in most cases calibrated and validated with the same test data used to generate the input parameters.
To really predict fatigue crack growth, i.e. without pre-existing fatigue crack growth data, requires a better understanding of the physics of fatigue crack growth and the role plasticity has on this phenomenon, are needed. The current body of empirical and phenomenological work does not allow for such physics based description.
To demonstrate that truly scientific predictions of fatigue crack growth are possible without using Paris curves, this study developed a physics based description of fatigue crack growth in metallic plates, through developing an energy balance between strain energy dissipating mechanisms. In the end, it is demonstrated that essentially only quasi-static stress-strain curves are required to enable the prediction of fatigue crack growth.
This paper explains the essential steps necessary to describe the physics of fatigue loading (application of strain energy or work) and the corresponding dissipation of energy through the formation of fracture surfaces, and plastic volume around the crack tip. In fact, the method allows to generate as result the Paris curve using stress-strain data, rather than using it as input in prediction. It is foreseen that further development of this work will accelerate introductions of new materials and alloys in aerospace, simply because material qualification and structural certification require substantially less testing and analysis.
|
12:10
20 mins
|
Multiaxial crack growth prediction
Bemin Sheen
Abstract: Bladed disks are subjected to foreign object damage which results in the initiation of cracks. Blisks in aero-engines are affected by a combination of loads which grow a crack in a multiaxial stress field. Depending on the local stress field, a crack may propagate through the blade, or more undesirably grow towards the bore of the disc due to hoop stress. The overall trajectories of cracks initiating near the base of the blade are particularly difficult to predict. Classical crack trajectory criteria, which are often based on the strain energy density or maximum principal stress during a cycle, are insufficient for predicting the growth of cracks under non-proportional and multiaxial loads. Additionally, a combination of inter-related factors including mode-mixity, cyclic plasticity, and crack closure are also influential on non-proportional crack growth. Therefore, an improved crack trajectory criterion is needed.
To create this criterion, a series of fatigue tests are being carried out on representative Ti-6Al-4V test samples to build a database of experimental results. The test rig and a cruciform shaped test samples have been designed to capture the key features of crack propagation in a blisk. Specimens may be loaded in three axes using a biaxial machine and hydraulic fixture. This arrangement aims to reproduce the combination of steady and dynamic loads in a crack in a blisk. The ratio of the applied loads has been determined from fatigue simulations in FRANC3D, a software package which simulates crack growth using linear elastic fracture mechanics. The results of the ongoing tests will help to update the crack growth model to gain an agreement between the simulated and experimental results of the test specimen. The final crack growth model will be used to accurately predict the trajectory of cracks initiating in blisks subjected to non-proportional loading.
Simulations on blisk geometries in FRANC3D have shown a good agreement with existing data from literature. Results obtained from the completed fatigue tests have been presented and compared to the results from the simulated specimen.
|
|
