The COD design curve: An assessment of validity using wide plate tests

Abstract

This paper examines the validity of the COD design curve approach for specifying defect acceptance levels in welded structures. The degree of conservatism built into the approach has been assessed from a statistical comparison between the allowable flaw size predicted from small-scale COD tests and the critical size at fracture in wide plate tests. From a survey of the literature, a total of 80 sets of COD—wide plate test data were obtained and carefully evaluated. The results indicate an average safety ratio of 3 in the allowable crack size predicted by the COD design curve and a 96·7% probability of the critical crack size exceeding the maximum allowable. A brief insight into the significance of the following important variables was gained by using the safety ratio as a comparative index: 1. (a) The influence of residual stresses in the design curve prediction for as-welded structures. 2. (b) The treatment of through as compared with part-through cracks in the design curve approach. 3. (c) The relative merits of using minimum and mean values of COD, as generally measured from a series of three specimens in the laboratory.