Allow me to introduce some fundamental principals of failure analysis used by engineers as I recall
from my younger life as a mechanical engineer
. The bolts in question as mounted on the hull are subjected to multiple stresses. First, the bolts are stressed by the tension introduced by tightening the nuts. Second, the bolts and mounting flanges/backing plates are subjected to the forces necessary to resist the torque of the load on the end of the arch -- this bending moment expressed in units of torque (ft. lbs.) is transferred to the several mounts and met with a combined counter torque by the bolts in shear once the friction between flanges/backing plates is overcome (if it is overcome). Third, the dynamic force of the load (dinghy) acting through the cantilevered arch as measured by the Mass x acceleration as the arch load cycles in a seaway. The vectorial sum of these forces acting on the cross-section of the bolts is the principal stress vector, which will be oblique to the cross section of bolt as it combines lateral and axial stress components. Interestingly, at the time of installation
of the bolts, an additional torsional stress is introduced by the act of tightening the nuts, and is often the time that fasteners will fail if that torsional component of the principal stress exceeds ultimate tensile strength in combination with other loads then acting on the fastener.
The principal stress, once determined is then compared to the yield stress of the material, discounted to accommodate stress concentrators such as cut threads and other abrupt changes in geometry, and deformation of the bolt shank if any, should the mounting flanges/plates have shifted. Cut threads are a severe stress concentrator as compared to rolled threads. The cyclic stress (such as the dink suspended on a bobbing boat) introduces the doctrine of fatigue, which can be quantified using empirically derived curves showing number of cycles at given stress levels. As I recall
, carbon steel
has an infinite life when subjected to cyclic stress less than 30% of the yield strength -- unlike all others such as aluminum
and stainless which have finite cycle life at any level of stress. Cyclic stress militates changing fasteners from time to time in order to avoid fatigue failure.
cracking, among other sources of introducing significant stress concentration, drastically affects the cyclic life of a fastener, and will lead to failure as the crack propagates toward the center of the fastener from its outside surface. Far less stress is required to propagate a crack to failure than would defeat a smooth surface. Torsional stress and bending moment related stress are transferred to the outside surface of the fastener, the center of the bolt being its neutral axis. Tension force stresses the entire diameter uniformly, but surface cracking or sharp deformations will diminish the useful loading of any fastener.
Fatigue failures generally show concentric rings across the diameter of the bolt working their from the point of highest stress inwards, and leaving a portion of the bolt that shows either a shear or tension failure when its ultimate strength across the diminished cross sectional area is reached.
I won't pretend to analyse the failure at hand here, but perhaps the foregoing principals will be useful as you consider how the deck arch mounting hardware
is loaded statically and dynamically. Consideration of one or another of the stress components is inadequate as a basis of analysis. It is the vectorial sum of all stresses acting on the fasteners that must underlie any investigation of a failure -- and should govern design and installation
means and methods.
With Best Regards,