Residual stress reconstruction by amplification of limited measurement data via finite element analysis
Simon McKendrey, Xavier van Heule, Ranggi S. Ramadhan, W. Kockelmann, Harry Coules, C. Jacquemoud, David Knowles, Mahmoud Mostafavi
Abstract
Welded components contain complex residual stress fields which are important to quantify when assessing their structural integrity. Often such assessments involve finite element simulation of the components; thus it is essential to include residual stress fields in the model. While previous methods have been proposed to include residual stresses in finite element models (e.g. using iterative methods or eigenstrain reconstruction of residual stresses), these can be theoretically cumbersome and computationally expensive. In this work a novel technique for reconstruction of residual stresses in welds is presented, based on iterative stress imposition and relaxation, and using limited residual stress data from energy dispersive X-ray diffraction (EDXD) measurements. This method is validated using a combination of neutron imaging of small sections of the weld and finite element analysis. A root mean squared (RMS) error of 127.26 μ ɛ was achieved between the FE model and the EDXD measurement. Although the method is only viable for relatively simple geometries such as pipes and plates, this covers the most likely use cases in relevant industries such as nuclear energy. Reconstruction of residual stress fields can assist structural integrity assessments by requiring less measured residual stress data. As well as reducing measurement costs our method may enable less overly-conservative assessments, particularly for flaws that do not lie on a weld centreline. This work also demonstrates that neutron imaging residual strain measurement is a valuable tool for validating methods of weld residual stress modelling. • Proposed methodology to reconstruct weld residual stresses from limited data. • A minimisation based methodology which reduces required iterations. • Neutron strain imaging used as a validation tool for residual stress reconstruction. • Creation of an Abaqus plugin to simplify implementation of proposed method.