P J Bouchard

Measurement and prediction of residual stresses in thick-section steel welds

Significant levels of residual stress are developed in the production of thick-section steel welds. Residual stress can be measured by a variety of methods, although few offer the ability to determine the spatial distribution completely through the thickness. This paper summarizes recent developments in the deep-hole method for measuring residual stresses in thick-section welds. Experimental results obtained from a variety of welded steel components are described. The measurement method has also been used to provide validation for finite element simulations of residual stresses in a welded cylinder-to-nozzle stainless steel component. The paper describes the finite element weld analysis and compares the results with measurements obtained from several locations in the complex geometry. Overall, there is good agreement between the predicted and measured distributions of residual stress, but the magnitudes of predicted stress tend to be greater.

Effects of Repair Weld Length on Residual Stress Distribution

This paper discusses residual stress distributions induced by repairing a stainless steel girth weld in a 19-mm thick pipe of outer diameter 541 mm. In particular, the effects of repair weld circumferential length are examined using finite element modeling. Results for three different repair lengths are presented having circumferential angular spans of 20 deg (short repair), 57 deg (medium repair), and 114 deg (long repair). A special 3-D shell element model is used which facilitates the simulation of multi-pass welds in 3-D piping components. The results shed light on a number of important 3-D residual stress features associated with repairs. Outer surface axial residual stresses in the weld and adjacent base material are tensile along the length of the repair reach maxima values near the arc start/stop positions, and then drop into compression beyond the repair ends. The short repair develops the highest axial tensile stresses due to the overlay of start/stop effects. The circumferentially remote residual stresses are unaffected by the repairs. At midlength of the repair, profiles of axial stress along the pipe show tensile peaks at 40 mm away from the weld centerline; these peaks decrease in magnitude with increasing repair length. However, the medium repair axial stresses show the greatest range of influence along the pipe. The pre-existing original girth weld residual stresses have very little effect on the repair residual stress characteristics. Finally, residual stress measurements on mock-up components are discussed which confirm the validity of the finite element methods used.

Slitting and Contour Method Residual Stress Measurements in an Edge Welded Beam

Welding is known to introduce complex three-dimensional residual stresses of substantial magnitude into pressure vessels and pipe-work. For safety-critical components, where welded joints are not stress-relieved, it can be of vital importance to quantify the residual stress field with high certainty in order to perform a reliable structural integrity assessment. Finite element modeling approaches are being increasingly employed by engineers to predict welding residual stresses. However, such predictions are challenging owing to the innate complexity of the welding process (Hurrell , Development of Weld Modelling Guidelines in the UK, Proceedings of the ASME Pressure Vessels and Piping Conference, Prague, Czech Republic, July 26–30, 2009, pp. 481–489). The idea of creating weld residual stress benchmarks against which the performance of weld modeling procedures and practitioners can be evaluated is gaining increasing acceptance. A stainless steel beam 50 mm deep by 10 mm wide, autogenously welded along the 10 mm edge, is a candidate residual stress simulation benchmark specimen that has been studied analytically and for which neutron and synchrotron diffraction residual stress measurements are available. The current research was initiated to provide additional experimental residual stress data for the edge-welded beam by applying, in tandem, the slitting and contour residual stress measurement methods. The contour and slitting results were found to be in excellent agreement with each other and correlated closely with published neutron and synchrotron residual stress measurements when differences in gauge volume and shape were accounted for.

Residual stresses in P91 steel electron beam welds

Abstract This paper describes the characterisation of residual stress in electron beam welded P91 ferritic–martensitic steel plates (9 mm thick) by neutron diffraction and contour measurement methods. The novelty of the work lies in revealing the residual stress profile at a fine length scale associated with a ∼1 mm wide fusion zone. A characteristic ‘M’ shaped distribution of stresses across the weld line is observed with high tensile peaks situated just beyond the heat affected zone/parent material boundary. Measured stresses close to the weld centreline are significantly less tensile than the adjacent peaks owing to martensitic phase transformation during cool down of the weld region. The effect of applying a second smoothing weld pass is shown to be undesirable from a residual stress standpoint because it increases the tensile magnitude and spread of residual stress. The results are suitable for validating finite element predictions of residual stress in electron beam welds made from ferritic–martensitic steels.

Spatially Resolved Materials Property Data From a Uniaxial Cross-Weld Tensile Test

Application of electronic speckle pattern interferometry (ESPI) is described to measure the spatial variation in monotonic tensile stress-strain properties along “cross-weld” specimens machined from a stainless steel three-pass welded plate. The technique, which could also be done with digital image correlation, was applied to quantify how the material 0.2%, 1%, 2%, 5%, 10%, and 20% proof stress varied with distance from the center-line of the weldment for parent and weld material associated with the first and final passes. The stress-strain curves measured by the ESPI method correlated closely with stress-strain data measured using conventional test specimens. The measured results are consistent with the hypothesis that thermo-mechanical cycles associated with the welding process work harden previously deposited (single-pass) weld metal and the surrounding parent material. The stress-strain response of the heat affected zone adjacent to the first weld pass is consistent with an accumulated (equivalent monotonic) plastic strain of 6.5% and that of the first pass weld bead was consistent with an accumulated plastic strain of approximately 4% greater than the state of the final pass weld metal

The Appropriateness of Residual Stress Length Scales in Structural Integrity

Residual stresses exist in most engineering components as a consequence of material processing and service history. It is well documented that such stresses can vary across a range of length scales and that different measurement techniques are sensitive to different components of the stress. Needless to say for reliable structural integrity assessment the residual stress field must be accurately quantified over the appropriate length scales. A classification system for sub-dividing residual macrostresses into long-, medium- and short-length scales is proposed to aid this process. A series of examples are presented that illustrate how the length scale for residual stress measurement and modelling studies should be determined in terms of the nature and size scale of the specific structural performance concern. The examples are related mainly to the residual stress field induced by a weld repair in a pipe that has been analysed by 3D finite element simulation techniques and measured by neutron diffraction.

Multiscale 3D analysis of creep cavities in AISI type 316 stainless steel

Abstract A sample of AISI type 316 stainless steel from a power station steam header, showing reheat cracking, was removed from service and has been examined by a combination of microscale X-ray computed tomography (CT), nanoscale serial section focused ion beam–scanning electron microscopy (FIB-SEM), energy dispersive X-ray (EDX) spectrum imaging and transmission electron microscopy (TEM). Multiscale three-dimensional analysis using correlative tomography allowed key regions to be found and analysed with high resolution techniques. The grain boundary analysed was decorated with micrometre sized, facetted cavities, M23C6 carbides, ferrite and G phase but no σ phase. Smaller intragranular M23C6 particles were also observed, close to the grain boundaries. This intimate coexistence suggests that the secondary phases will control the nucleation and growth of the cavities. Current models of cavitation, based on isolated idealised grain boundary cavities, are oversimplified.

Residual Stress Driven Creep Cracking in Type 316 Stainless Steel

Specially designed AISI Type 316H austenitic stainless steel 25 mm thick compact tension specimens have been plastically deformed to produce significant tensile hydrostatic residual stresses at the notch root at mid-thickness. These specimens were thermally exposed at 550 °C for 4500 hours in order to study elevated temperature creep relaxation of residual stress and the development of reheat cracking creep damage. Residual strains within the specimens were measured using diffraction techniques before and after thermal exposure. A three-dimensional finite element model was developed both to predict the residual stress within the specimens before and after thermal exposure. No reheat cracking was found near surface, but due to the reduced creep ductility with increasing hydrostatic stress, significant creep cavitation was found mid-thickness. A previously developed creep damage model was applied to predict the onset of reheat cracking. Good correlation has been found between measurements and finite element predictions of strain and stress before and after thermal exposure. The extent of creep damage has also been assessed through destructive examination, providing validation for the creep damage prediction model.

J-integral and local damage fracture analyses for a pump casing containing large weld repairs

Three-dimensional J-integral and two-dimensional Local Approach finite element studies are described for postulated crack-like defects in a large repair weld to the casing of a light water reactor circulation pump. The repair weld residual stress field is simulated and plant operating pressure and thermal transient loads are applied. Crack tip constraint effects are quantified through detailed analysis of the cracked structure and compact tension fracture toughness specimens. Fracture initiation crack sizes are shown to be larger than conceivable fabrication defects that are detectable using modern ultrasonic inspection techniques. The Local Approach study demonstrates the benefits of quantifying crack tip constraint conditions, compared with conventional J-estimation schemes and cracked body J-integral analysis. The method of introducing the crack into the finite element model is shown to have a large effect on calculated crack tip fracture parameters; a slowly developing crack in the residual stress field being more benign.

Evaluation of through wall residual stresses in stainless steel weld repairs

Weld repair is used widely in many industrial applications. Knowledge of the through thickness residual stress distribution is of particular importance for accurate assessment of the structural integrity of repaired components. However, the distribution and magnitude of residual stress after weld repair is not well characterised. In this paper, results from a programme of residual stress measurements on a butt-welded stainless steel pipe, with 35mm wall thickness and an outside diameter of 432 mm arc presented. Measurements of the in-plane components of residual stress through the wall thickness were obtained using the novel deep-hole technique. This technique relies on measurements of the distortion of a reference hole drilled through the wall thickness of the component. Residual stresses were measured in the heat affected zone (haz) of the butt-weld, prior to and after introduction of weld repairs of different depths. it is shown that the introduction of repairs significantly changed the original as-welded residual stress-state. Finite element (fc) residual stress simulations for the original girth weld and a deep repair are described. The measured results show good agreement with predictions both prior to and after weld repair.