Fillet welds are widely used and therefore we encounter many cases where these welds are subjected to fatigue loads. Since ultrasonic peening is a fatigue life improvement technique it would be reasonable to propose the treatment to enhance the fatigue performance also of fillet welds. However, fillet welds normally contains only limited penetration and therefore it always will exist a risk the fatigue crack could develop from the root. Furthermore if the features which makes the weld toe prone to fatigue cracking, crack-like flaws, sharp transition and high tensile stresses, are completely removed by the ultrasonic peening treatment the fatigue crack will unavoidable develop from the root of the weld which will remain untreated.
However, the ultrasonic peening treatment of a non-full penetration fillet weld will increase its fatigue resistance but certainly not as much as if compared to the treatment of similar weld geometry/load configuration done in a full-penetration weld with no root gap. Furthermore, a fatigue crack growing and developing from the root will definitively increase the difficulties of an early detection during scheduled inspections. Early fatigue crack detection is one of the most important reasons to carry out periodical inspections since it will allow enough time to repair while the structural integrity of the installation still allows for it.
As a result it would be advisable to apply the treatment in full-pen weld as much and as long as is practically possible. However, if the weld in question is not easily accessible or if the work required to produce a full-pen weld is beyond the practically acceptable it could be resolved by and additional weld bead. Reasonable hand calculations could give a first indication about the required a-design throat to carry out the ultrasonic peening at the weld toe with enough reliability to predict the crack to grow from the treated weld toe.
It is also important to take into consideration the stress range to which the improved fillet weld will be subjected to. A recent study carried on improved weld toes of longitudinal stiffeners shows a clear shift, from toe to root, of the fatigue crack initiation point as a consequence of the stress range level. The risk for the fatigue crack to develop from the root gap in a non-full pen weld is higher for lower stress ranges for the improved welds. As result, if an extra weld pass is added and the expected stress range for which the weld connection is designed is relative high, the risk for root crack would be small.