“ELIXIR – Extending Plant Life Through Improved Fabrication and Advanced Repair Methodology” was a European Union FP5 sponsored project. During the duration of the Elixir project, much work was directed at providing the necessary data for the validation of numerical modelling techniques applied to residual stress generation and hydrogen diffusion arising from the welding process. The project focussed around four industrial applications, namely petrochemical, boiler, offshore and submarine. This paper presents through-thickness residual stress measurements obtained by the University of Bristol on two of the large industrial components. The results were obtained using the deep hole drilling technique and compared to Finite Element predictions provided by other partners. The components considered are a large P275 steel set-in nozzle, typical of a boiler application and a large S690 steel set-on nozzle, typical of an offshore application. The boiler application consisted of a nozzle of diameter 600mm and thickness 50mm, on a pipe of diameter 1100mm and 100mm thickness. The offshore application was a nozzle of diameter 900mm and thickness 50mm, on a pipe of diameter 1050mm and 50mm thickness. Both the longitudinal and transverse stresses measured using deep hole drilling showed excellent agreement with Finite Element predictions through the thickness of the boiler sample. On the top surface, a zone of tensile residual stresses, over a distance of approximately 40mm, was revealed, which was equilibrated by a zone of compressive residual stresses over the final 50mm of thickness. Results for the offshore application demonstrated that at the front surface, both of the stress components were essentially zero, but both the longitudinal and transverse components rose rapidly to maxima of approximately 500MPa and 220MPa, respectively. Tensile residual stresses were supported over a distance of approximately 30mm. Over the final 20mm of thickness, compressive residual stresses existed, which again fell to approximately zero on the back face. There is excellent agreement between measurements and the Finite Element predictions for the transverse stress component, but less good agreement between measurements and predictions of the longitudinal stress component.

This content is only available via PDF.
You do not currently have access to this content.