Cory J. Hamelin

The Role of Plasticity Theory on the Predicted Residual Stress Field of Weld Structures

Constitutive plasticity theory is commonly applied to the numerical analysis of welds in one of three ways: using an isotropic hardening model, a kinematic hardening model, or a mixed isotropic-kinematic hardening model. The choice of model is not entirely dependent on its numerical accuracy, however, as a lack of empirical data will often necessitate the use of a specific approach. The present paper seeks to identify the accuracy of each formalism through direct comparison of the predicted and actual post-weld residual stress field developed in a three-pass 316LN stainless steel slot weldment. From these comparisons, it is clear that while the isotropic hardening model tends to noticeably over-predict and the kinematic hardening model slightly under-predict the residual post-weld stress field, the results using a mixed hardening model are quantitatively accurate. Even though the kinematic hardening model generally provides more accurate results when compared to an isotropic hardening formalism, the latter might be a more appealing choice to engineers requiring a conservative design regarding weld residual stress.

Predicting Solid-State Phase Transformations during Welding of Ferritic Steels

The current work presents the numerical analysis of solid-state transformation kinetics relating to conventional welding of ferritic steels, with the aim of predicting the constituent phases in both the fusion zone and the heat affected zone (HAZ) of the weldment. The analysis begins with predictions of isothermal transformation kinetics using thermodynamic principles, such that the chemical composition of the parent metal is the sole user-defined input. The data is then converted to anisothermal transformation kinetics using the Scheil-Avrami additive rule, including the effects of peak temperature and austenite grain growth. Subroutines developed for the Abaqus finite element package use the semi-empirical approach described to predict phase transformations in SA508 Gr.3 Cl.1 steel. To study the effect of the cooling rates and the ability of the current model to predict the final microstructure, two weld samples were subjected to autogenous beam TIG welds under a fast (TG5-F, 5.00 mm/s) and slow (TG5-S, 1.25 mm/s) torch speed. Model validation is carried out by direct comparison with microstructural observations and hardness measurements (via nanoindentation) of the fusion and heat affected zones in both welds. Excellent agreement between the measured and predicted hardness has been found for both weld samples. Additionally, it is shown that the correct identification of the partial austenisation region is a crucial input parameter.

The Impact of Axi-Symmetric Boundary Conditions on Predicted Residual Stress and Shrinkage in a PWR Nozzle Dissimilar Metal Weld

Simulation of a dissimilar metal weld (DMW) in a pressurised water reactor (PWR) nozzle was performed to predict both axial distortion and hoop residual stresses in the weld. For this work a computationally efficient axi-symmetric finite element (FE) simulation was carried out rather than a full 3D analysis. Due to the 2-dimensional nature of the analysis it was necessary to examine the effect of structural restraint during welding of the main dissimilar metal weld (DMW). Traditionally this type of analysis is set up to allow one end of the structure, in this case the safe-end forging, to be unrestrained in the axial direction during welding. In reality axial expansion and subsequent contraction of deposited weld metal at the current torch position is restrained by solidified material both ahead and behind the torch. Thus the conventional axi-symmetric analysis is under-restrained in the axial direction at least during the early weld passes. The significance of this was examined by repeating the current simulation with the safe-end forging fixed to limit expansion during the heat up cycle. Contraction was however, allowed during cooling cycle. This modified boundary control method provided a significantly improved prediction of the axial distortion across the weld as well as improved prediction of through wall axial and hoop residual stresses.Copyright

Predicting post-weld residual stresses in ferritic steel weldments

The implementation of a semi-empirical solid-state phase transformation subroutine in the ABAQUS finite element package is presented to predict the influence of transformation strain on the post-weld residual stress field in ferritic steels. The phase transformation subroutine has been outlined in a previous study (PVP2011-57426), where it was proven accurate in predicting the phase compositions in the fusion and heat affected zone (HAZ) of an autogenous TIG beam weld in SA508 Gr.3 Cl.1 steel.While previous work focused on predictions of the steady-state material response using a 2D thermal model, the present analyses are 3D and capture the varying phase composition at weld start- and stop-ends. Predicted cooling rates at either end of the specimen are significantly higher, leading to a variation in the predicted microstructure along the weld line.To better understand the structural changes that occur in ferritic steels during a conventional welding process, a representative model of SA508 Gr.3 Cl.1 steel is discussed. The contribution of thermal, metallurgical, and transformation-induced plastic strain is highlighted in this example, providing insight to the key simulation variables necessary for accurate weld models of ferritic steels. Preliminary coupled thermo-mechanical analyses are presented that compare predicted residual stress distributions with those measured in SA508 Gr.3 Cl.1 beam welds via neutron diffraction; good agreement is observed.© 2012 ASME

Accounting for Phase Transformations During Welding of Ferritic Steels

The numerical application of solid-state phase transformation kinetics relating to conventional welding of ferritic steels is presented. The inclusion of such kinetics in weld models is shown to be necessary for capturing the post-weld residual stress field. To this end, a comparison of two approaches is outlined: a semi-empirical approach that uses thermodynamic transformation kinetics to predict phase morphology; and a fully empirical approach that directly links local material temperature to the present constituent phase(s). The semi-empirical analysis begins with prediction of TTT diagrams using thermodynamic principles for ferritic steels. The data is then converted to CCT diagrams using the Scheil-Avrami additive rule, including austenite grain growth kinetics. This information is used to predict the phases present under varying peak temperatures and cooling rates. In the fully empirical approach, dilatometric experiments of steel samples are performed during heating to simulate expected welding conditions. The constitutive response of the sample is then used as input for the subsequent numerical weld analyses. Input derived from each technique is transferred into weld models developed using the Abaqus finite element package. Model validation is carried out by direct comparison with neutron diffraction residual stress measurements on two beams of SA508 Gr.3 Cl.1 steel subjected to autogenous beam TIG welds under varying torch speeds, heat input and preheat conditions.© 2011 ASME

A Comparison of the Constitutive Response of Austenitic and Ferritic Steels under Welding Processes

The need to accurately measure and predict weld residual stresses (WRS) has led to several examinations intent on developing best-practice guidelines in the assessment of welded structures. The present investigation examines two benchmark weld specimens; both specimens are autogenous edge-welded beams, with welds deposited using a mechanised tungsten inert gas process. However, one of the beams was made from AISI 316LN austenitic steel, while the other was made from SA508 Gr.3 Cl.1 ferritic steel. Considerable differences in the cross-weld residual stress profile were observed between the two beams, prompting a detailed examination of why such differences exist. Computational weld mechanics was used to assess both processes; model validation was achieved using previously reported WRS and micro-hardness measurements. A comparison of the numerical solutions indicates that the shape misfit resulting from a sharp weld-induced thermal gradient causes significant longitudinal tensile stresses in the heat-affected zone in both specimens. The presence of influential solid-state phase transformations in the ferritic specimen leads to the formation of compressive stresses in the weld metal, while the stresses remain tensile in the weld metal region of the austenitic specimen. The compressive stresses in the ferritic specimen serve to offset the tensile stresses in the HAZ, leading to a reduction of the self-equilibrating WRS present in the ferritic parent metal.

Predicting the Creep Rupture Behavior of Austenitic Steel Using Finite Element Analysis

The creep behavior of structural materials is often measured using uniaxial tension creep rupture tests. Unfortunately, the time required for austenitic steel samples to rupture under ideal (i.e. elastic stress) conditions is prohibitive. To accelerate creep rupture in these samples, a tensile stress in excess of the material yield strength is often applied and the post-load deformation is assumed to be largely creep-based. There is currently no method of measuring the creep deformation separately from the yield-induced plastic flow that may occur during such accelerated tests.Using validated finite element models, the effects of creep and yield-induced plastic strain have been decoupled for a series of accelerated creep tests using 316H austenitic steel. The influence of continued yielding after the initial sample loading was predicted to be significant, which suggests the diffuse necking in the samples due to creep is responsible for stress intensification and further yield through the tests. These results suggest the initial plastic loading in accelerated creep tests may significantly influence the measured creep rupture time in these samples.Copyright

A Validated Numerical Model for Residual Stress Predictions in an Eight-Pass-Welded Stainless Steel Plate

Welding processes create a complex transient state of temperature that results in post-weld residual stresses. The current work presents a finite element (FE) analysis of the residual stress distribution in an eight-pass slot weld, conducted using a 316L austenitic stainless steel plate with 308L stainless steel filler metal. A thermal FE model is used to calibrate the transient thermal profile applied during the welding process. Time-resolved body heat flux data from this model is then used in a mechanical FE analysis to predict the resultant post-weld residual stress field. The mechanical analysis made use of the Lemaitre-Chaboche mixed isotropic-kinematic work-hardening model to accurately capture the constitutive response of the 316L weldment during the simulated multi-pass weld process, which results in an applied cyclic thermo-mechanical loading. The analysis is validated by contour method measurements performed on a representative weld specimen. Reasonable agreement between the predicted longitudinal residual stress field and contour measurement is observed, giving confidence in the results of measurements and FE weld model presented.

Finite Element Modelling of Welded Austenitic Stainless Steel Plate With 8-Passes

The current paper presents a finite element analysis of an eight-pass groove weld in a 316L austenitic stainless steel plate. A dedicated welding heat source modelling tool was employed to produce volumetric body power density data for each weld pass, thus simulating weld-induced thermal loads. Thermocouple measurements and cross-weld macrographs taken from a weld specimen were used for heat source calibration. A mechanical finite element analysis was then conducted, using the calibrated thermal loads and a Lemaitre-Chaboche mixed work-hardening model. The predicted post-weld residual stresses were validated using contour method measurements: good agreement between measured and simulated residual stress fields was observed. A sensitivity analysis was also conducted to identify the boundary conditions that best represent a tack-welded I-beam support, which was present on the specimen back-face during the welding.Copyright

Prediction and Measurement of Weld Residual Stresses in Thermally Aged Girth-Welded Austenitic Steel Pipes

The current paper describes finite element simulation of the complete manufacturing and service exposure history of girth-welded austenitic steel pipes fabricated from ESSHETE 1250 material for the STYLE Framework 7 project. The simulation campaign examines the impacts of prior quenching of pipe material, fabrication of closely adjacent welds, variation in mixed isotropic-kinematic hardening material constitutive models, and high temperature (650°C) service exposure (thermal ageing). The predicted residual stresses are validated using measurements made with the deep hole drilling (DHD) and incremental deep hole drilling (iDHD) techniques.Copyright