The Prediction of Residual Stresses in an Autogenously Welded Ferritic Beams

Abstract

The continued safe and reliable operation of plant invariably has to consider the assessment of defects in welded structural components. This often requires some estimate of the residual stresses that have developed during the welding fabrication process. Increases in the computational power available to finite element analysts have made the prediction of welding residual stresses using finite element methods an increasingly viable option. Consequently recent years have seen many advances in the field of residual stress modelling. However, relatively little work has been conducted on the accurate modelling of welded ferritic components. This is largely due to the added complications associated with the solid-state phase transformations that occur during the heating and cooling of such steels. Against this background, a programme of work has been undertaken in order to investigate the effects that phase transformations have on the development of residual stresses in ferritic components and develop methods for their simulation. To facilitate this, generic components of increasing complexity are being studied. Reported in this paper is the first phase of this programme which considers SA508 beam specimens. These specimens have been subjected to autogenous TIG welds using two different torch travel speeds. In order to predict the resultant residual stresses, simulations have been conducted using the commercial finite element package SYSWELD. These predicted stresses are then compared with residual stress measurements conducted using the neutron diffraction technique.

Numerical Analysis of the Effect of Phase Transformation on Residual Stresses in an Autogenous Beam Edge Weld

“…The predicted temperatures were in good agreement with the measurements, which was achieved by calibration of the heat input parameters using thermocouple measurements and macrographs. The predictions of the phase transformations for the slow weld were in agreement with earlier published work on this topic [5][6][7]. However, for the fast weld the predicted phase transformations are not in agreement with all published work.…”

“…However, for the fast weld the predicted phase transformations are not in agreement with all published work. In [7], a large bainite proportion was predicted for the fast weld, but in [5] hardness measurement data was published, which indicates that the fast weld consists of mainly martensite. Figs.…”

Residual Stress, Texture, and Phase Investigation of Autogenous Edge Welds Using High Energy Synchrotron Radiation

“…In previous investigations of optical micrographs [5,6] three metallographically distinct limits were observed, namely the limit of the fusion zone at 4.25 mm (slow) / 2.75 mm (fast), the limit of the HAZ at 10.25 mm (slow) / 6.25 mm (fast), and a limit dividing the HAZ at 6.75 mm (slow) / 4.5 mm (fast). The different extensions of the HAZs into the parent material have their origin in the differences in heat input for the slow and fast weld.…”

“…Stresses were calculated assuming plane stress. FE modelling macroscopic stress results, as published in [5], are plotted as solid lines. FE modelling strain results were calculated from the stress values.…”

“…In addition results from FE modelling are shown in the figure as solid lines. The FE residual stress modelling results were published elsewhere [5], but the corresponding strain values were back-calculated here for the purpose of direct comparison with the experimentally observed in-plane residual strains. hkl weighted elasticity constants were also used for this reverse calculation and the contribution of the austenite phase was not taken into account.…”