Second Failure Investigation
The conclusions drawn from the first failure investigation, i.e. that the failure was 'progressive' in nature, foreseeable and avoidable by maintenance personnel, left one insurance company carrying the whole financial loss. A second investigation was therefore commissioned by loss adjusters in this company, using investigators with known expertise in fatigue and fracture. By this time (some 6 months after the accident) all but three of the tractor coupling bolts had been lost, but the shear bolt was available for investigation. This second investigation therefore focussed on the fatigue cracking present in the shear bolt, and its consequences for the mechanical properties of the bolt.
Properties and Metallography of the Shear Bolt
The bolt was manufactured from an unknown high strength steel alloy. Chemical analysis indicated the following composition of the bolt:
This corresponds to the composition range defining AISI 8637 alloy steel, which was deleted from the SAE handbook in 1993. The microstructure showed a quenched and tempered bainitic steel with a Vickers hardness of between 364-374. Using the resources available from the linked internet site
, determine the likely range of tensile strength for this steel bolt. Additional resources regarding mechanical property testing and conversion of hardness values to tensile strength is given in the pdf file
from the Department of Materials Science & Engineering at Carnegie-Mellon University.
This data can be used to estimate the fracture toughness of the steel using typical data for quenched and tempered steels and then, using crack sizes determined from scanning electron microscopy (SEM), an LEFM calculation can be made of the reduction in tensile strength resulting from the presence of the defect. This stress estimate can be compared with the values of bending stresses which might be caused by the elongation of the shear bolt holes in the yoke of the aircraft drawbar.
The deformation in these holes was quite extensive, and the shear bolts were, reportedly, replaceable items. Thus the observed yoke damage could have been caused by previous instances of overloading and plastic bending of the bolts, which were then replaced with new ones. Eventually, a situation would be reached where the bolt hole was hourglass shaped and provided a snug fit to the bolt only in the central section. A significant bending stress would then be induced at the centre section of the shear bolt, leading to the observed crack initiation and, possibly, causing the failure. If the hole had remained snug fit over the complete bolt length, the mode of failure would have been shear of the protruding ends of the bolt.
Reference 1 indicates that the likely range of fracture toughness values for a quenched and tempered steel of this range of tensile strength, would be between 70-110 MPa m½
The maximum size of fatigue crack present in the shear bolt at the time of failure was hard to estimate, as significant corrosion damage had occurred to the bolt during its exposure on the taxiway. It had been chemically cleaned, which can also induce corrosion pitting. Equally, in high strength quenched and tempered steels, the micromechanisms of fatigue crack growth may appear relatively similar to fast fracture, particularly if surface damage exists. However, inspection of the fracture surface indicated that, as well as the definite small fatigue cracks present around the circumference of the bolt, other larger semi-elliptic shapes were present on the fracture surface. Figure 5
is repeated below, with such a shape outlined with the white line. It is quite possible that the fatigue cracks extended by progressive bursts of fast fracture before final failure occurred. On this basis, the maximum possible size of defect at the time of the accident may have been greater than 0.6 mm deep by some 1.8 mm along the surface, and several such defects may have been present. Consideration of the evidence of SEM fractographs such as the ones given in Figures 5 and 8
, leads to the conclusion that a number of defects with depths ranging somewhere between 0.1-1.0 mm probably existed in the shear bolt at the time of the accident.
Other cracks were also present in this region of the shear bolt, as seen in Figure 9 and their presence supports the above argument, as fairly extensive surface lengths are seen. The fracture surface in Figure 9 shows evidence of ratchet marks and multiple crack initiation sites, indicating relatively high applied stresses. Ratchet marks are described in the associated Fractography Resource.
Figure 8 Small fatigue crack on the fracture surface
Figure 9 Other cracks are present below the fracture plane
The activity linked
from this page allows calculation of the bending stress as a function of deformation of the shear bolt, and estimation of the fracture stress using LEFM. A sensitivity analysis of the fracture stress to variation in input parameters can also be performed.
Proceed to final part
of case study.