Fatigue Analysis of Needle Peened Welds under Variable Amplitude Loading Conditions

Regarding the analyses presented herein, the following comments can be made: 1 Both models show a decrease in the treatment benefit (fatigue life improvement) with an increase in the stress range under constant amplitude loading for all stress ratios. 2 Under constant amplitude loading, the strain-based model predicts higher fatigue lives for the untreated weld and a lower treatment benefit, regardless of the stress range/ratio. 3 Despite the lower benefit predicted by the strain-based model, significant fatigue life improvements are still seen. For example, at stress ranges less than ∼450 MPa at R = –1 or ∼250 MPa at R = 0, a fatigue life improvement of 100% or greater is still predicted. 4 Both models predict that the effect of the periodic compressive overloads on the fatigue life of the untreated weld will generally be less significant than for the treated weld. 5 For the treated weld, the fatigue life can be seen to decrease as the magnitude of the overload stress increases. This is also reflected in the calculated treatment benefit at the various stress ranges. Interestingly, the drop in the treatment benefit (in terms of the fatigue life improvement) is most pronounced in the results obtained using the LEFM model. When comparing the results obtained using the two models, it may be noted that the largest differences can be seen at the lower stress ranges, even though the influence of the non-linear material behaviour is the lowest at these stress ranges. The reason for this trend is that the stress intensity factor (SIF) calculated using (10) considers the effect of the non-uniform stress distribution along the crack edges, whereas the SIF calculated using (3) does not. As the crack grows out of a weld toe notch, this effect becomes more significant and the influence of the non-linear material behaviour diminishes. The reason that (3) has been used successfully in [11–13] to model the behaviour of small cracks emanating from notches may be that the non-uniform stress distribution plays a much less significant role in this case, in particular when the level of yielding is significant. One possible way to account for this observation would be to use a hybrid model that uses (3) for the smaller crack depths and then switches to (10) when the influence of the nonlinear material behaviour is seen to become negligible. Despite the apparent limitations of the models employed herein, the following conclusions can be drawn from this work: 1 Under constant amplitude loading conditions, although considering the nonlinear material behaviour appears to reduce the predicted treatment effectiveness, substantial benefits are still seen — even at relatively high applied stress range levels. 2 Both the LEFM and strain-based fracture mechanics models predict a significant decrease in the treatment effectiveness for load histories that include periodic compressive overloads. The reduction in fatigue life due to the compressive overloads is generally greater for the treated welds than for the corresponding untreated welds. 3 On this basis, it is concluded that, although residual stress-based post-weld treatments have shown strong potential as a means for improving or rehabilitating the welds in existing steel structures such as bridges, it is important that the influence of the true load history be considered when exploiting the benefits of these treatments.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.