An Analytical Assessment of the Effects of Residual Stresses and Fracture Properties on Service Performance of Various Weld Repair Processes

This paper discusses analytical work done to assess the relative importance of residual stresses and fracture mechanics material properties on the service performance of various weld repair processes. In the first part of this analytical work, a model was developed which quantified the performance of various repair weld procedures. The assessment was based on fracture mechanical performance of two different weld/crack configurations as measured by fracture toughness and fatigue crack growth. One configuration was a crack located in the weld parallel to the welding direction. The other was a crack perpendicular to the weld centered about the midpoint of the weld. Both models were quantified in terms of both brittle fracture and fatigue crack growth mechanisms. Further studies were made assuming that the various weld procedures do not change the material properties of the weld metal, such as fatigue crack growth rates. Several cases were then analyzed assuming a range of residual stress variations in the weld. The results of these analyses quantified the importance of magnitude and distribution of the various chosen residual stress patterns. In the second part, actual material properties were obtained from the literature for the various weld procedures and used to analyze the structural performance in terms of fatigue crack growth resistance. Data taken from a Welding Institute report was used to make comparisons between six different weld metals and two different weld processes. A performance factor, Q, the ratio of fatigue cycles for the weld metal to fatigue cycles for the base metal, was calculated for the six weld metals as a function of residual stress divided by applied stress. Results of the analysis indicated that the highest strength weld metal gave the best performance in terms of fatigue crack growth resistance, while the lowest strength weld gave the worst performance as measured by Q. In addition, scatter in the fatigue crack growth rates was examined for the various weld repair procedures. Based on data obtained from several published sources, the scatter in fatigue crack growth rate was found to be a factor of 3. Allowing for variations in the magnitude of residual stress of an order of magnitude, it was shown that the scatter in fatigue crack growth rate completely overshadows the effect of residual stress variations which could conceivably be produced by the various weld procedures.