The influence of NDE and finite element modelling parameters in the integration of 3D NDE to FEA-based defect analysis

Recent and ongoing developments in 3D non-destructive examination (NDE) techniques (such as full-matrix capture ultrasonics and X-ray CT) have created the possibility of capturing the actual shape of sharp defects in pipelines. Such advancements could theoretically enable engineering critical assessment (ECA) procedures to be performed based on accurate defect shape information, as compared with simplified, idealized shape assumptions. A recently concluded EPRG project has shown promising results in establishing the feasibility of integrating defect shape data from 3D NDE techniques into finite element analysis (FEA) models, highlighting further efforts required for in-field application. Additional to accounting for defect shape, the numerical modelling approach allows to directly assess (multiple) defects without requiring flaw interaction and recategorisation rules, which are known to be of potentially excessive conservatism. A follow-up project funded within PRCI (3D NDE-FEA) is ongoing, with the aim of further developing the approach of coupling 3D NDE data and FE modelling for ECA. Within the abovementioned context, this paper reports on an investigation of influences of FE modelling and NDE parameters on the numerical accuracy of the developed approach. Regarding FE analysis, influences of model parameters on the accuracy of crack driving force output were graded both qualitatively and quantitatively. These influences were compared to those resulting from material uncertainty as inherent to defect assessment, providing a reference to justify potential improvement needs in model accuracy. As regards NDE, specific attention was given to how and which scan variables influence the representation of the defect shape which is fed into the FEA model. Metrics were devised for quantitative influence grading across the different categories of variables. Finally, concrete opportunities are outlined for further improvement of the coupling algorithm, both in terms of accuracy and time