Session: MF-03-01: Welding Residual Stress and Distortion Simulation and Measurement - 1

Paper Number: 84779

84779 - Electron Beam Welds in 316L Part 1: Weld Production, Residual Stress Measurements and Predictions 

There is enthusiasm for new nuclear plants in the UK to adopt power beam welding technologies, which could offer several advantages over conventional techniques. In particular, reduction in the time taken to produce and inspect a weld, and thus the cost of manufacturing components. However, a blocker to the adoption of these technologies is a shortage of accepted methodologies for demonstrating the integrity of these joints, which forms part of the generic design assessment within the UK regulatory environment.

Residual stresses can contribute towards crack driving force and thus should be accounted for when assessing the integrity of a component or its tolerance to damage. Whilst bounding residual stress fields are often used, it is often desirable to have more realistic estimations that capture the through-wall residual stress distribution, which also allows them to be decomposed into membrane, bending and self-equilibrated components to aid stress classification.

Material-specific weld residual stress profiles already exist, for example Level 3 profiles in the UK’s R6 procedure. However, they are for arc welding techniques. This work seeks to provide a framework for the generation of weld residual stress profiles for power beam welds and is split over two papers:

1.    Weld Production, Residual Stress Measurements and Predictions;

2.    A Methodology and Example for Parameterised Residual Stress Profiles.

A series of welds have been manufactured in 316L austenitic stainless steel thin plates (3 mm, 6 mm, 12.5 mm and 25 mm) and pipes (outside diameter 170 mm and wall thicknesses 7 mm and 15 mm). Welding parameters, thermal histories and weld macrographs are presented. A suite of experiments were undertaken to measure residual stresses in the weldments. All weldments have had the in-plane components of the residual strain tensor measured across the welds with high-energy energy-dispersive synchrotron X-ray diffraction. All but the thinnest, 3 mm plate have had the weld longitudinal residual stress measured on a pane across the weld. The full residual stress tensor has been reconstructed for the 12.5 mm and 25 mm plates using a sample rotation method with X-ray diffraction and supplementing the in-plane measurements with out-of-plane neutron diffraction respectively. In contrast to some reports in the literature, high levels of stress triaxiality were not observed in these welds.

A computationally efficient method of modelling residual stresses in 2D is presented. The thermal model is tuned using the weld macrograph and thermocouple temperature histories with both near- and far-field measurements. The geometry of the flux is tuned to match the shape of the fusion zone. The modelling assumes heat-transfer is by conduction across the plate and radiation from the plate, with no account made for any weld-pool convection. The results show this is a reasonable assumption for electron beam welding in a vacuum but does not capture the behaviour of laser beam welding at atmospheric pressure with a shielding gas, where a computationally efficient method of modelling weld residual stress would also be desirable. The thermal histories for the electron beam welds are then solved mechanically, using a mixed-mode (isotropic/kinematic) hardening model. The residual stress predictions are then validated using the residual stress measurements.

Presenting Author: Graeme Horne Frazer-Nash Consultancy

Presenting Author Biography: Graeme Horne is a Consultant at Frazer-Nash Consultancy, with an academic background in residual stress measurement and prediction, and their contribution to fracture. His PhD was supervised by Prof David Smith and the University of Bristol.

Authors:

Graeme Horne Frazer-Nash Consultancy
Andrew Moffat Solar Turbines Europe S A