Assessment Techniques Using the R6 Procedure: Investigation of Limit Load Approach for Pipe Branch Connections

Abstract

The R6 Revision 4 Procedure ‘Assessment of the Integrity of Structures Containing Defects’, states that the use of the finite element ‘global’ limit load derived for pipe branch components can be non-conservative when used with the Option 1 and 2 failure assessment curves but that ‘local’ limit loads, based on the spread of plasticity through the pipe wall, should lead to conservative results. The current advice of R6 is based on separate studies by Fox and Connors of pipe branch components with fully extended surface defects. Their studies provide two distinct methods for calculating a suitably conservative local limit load. However, there is concern that these two methods may provide an overly conservative local limit load with therefore a less realistic prediction of defect tolerance. Furthermore, typical defectiveness is perhaps most commonly characterised as a semi-elliptic surface defect and it is therefore necessary to adapt both these methods in order to accommodate such defects. The purpose of this study was therefore to investigate local limit load approaches for pipe branch components with postulated semi-elliptic surface defects. A typical pipe branch component was chosen for assessment during this study, as part of a series of separate studies on a variety of pipe branch components. Local limit loads were calculated using two approaches. The first approach adapted the ‘Connors’ method by applying an adjustment to allow for the semi-elliptic surface defect; this is referred to as the ‘Modified Connors’ approach. The second approach used cracked body finite element analysis and evaluated the local limit load by consideration of the onset of plasticity at the crack ligament. The global limit load was also derived from the cracked body finite element analysis. Assessment points were developed using global and local limit loads, both obtained by cracked body finite element analysis, and also by using the ‘Modified Connors’ local limit load approach. R6 Option 3 failure assessment curves were produced for each limit load approach in order to investigate the extent of any non-conservatism in the Option 1 and 2 failure assessment curves with the chosen limit load approach.