- Invited Keynote Speaker
- Book by Goldak & Akhlaghi
- University of Alberta Seminar 1
- University of Alberta Seminar 2

June 1 to 6, 2008

Pine Mountain, Georgia USA

Get the most current information on research and technology at Trends in Welding 2008 – where science and practical application of welding unite. Every three years, Trends in Welding Research brings together a virtual “Who’s Who” in welding science, technology and engineering. Experts from over 25 countries present technically intensive symposia focusing on both fundamental and applied topics related to welding and joining. Industry and academicians will present both experimental and ideal perspectives.

John Goldak, Keynote Address Title: Distortion and Residual Stress in Selds: The Next Generation

Abstract: As numerical algorithms and software for computational weld mechanics (CWM) near maturity and the cost of computing becomes negligible, we expect research in CWM to change rapidly and dramatically. Industrial people, experimentalists and modelers, who have up to now tended to work separately on relatively small research projects, will form more tightly knit teams that work on much larger projects. The goal will be to reduce uncertainty in product design, development, production and service performance for welded structures. The key technologies needed to manage distortion, residual stress and failure modes in welded structures are expected to be sensor development, data acquisition, material properties and statistical-probabilistic analyses. Software to support collaboration will be critical to success. Some very modest progress towards this vision will be presented.

«January 17, 2008»

Drawing upon years of practical experience and the study of computational welding mechanics the authors instruct the reader how to:

- understand and interpret computer simulation and virtual welding techniques including an in depth analysis of heat flow during welding, microstructure evolution and distortion analysis and fracture of welded structures,

- relate CWM to the processes of design, build, inspect, regulate, operate and maintain welded structures,

- apply computational welding mechanics to industries such as ship building, natural gas and automobile manufacturing.

Ideally suited for practicing engineers and engineering students, Computational Welding Mechanics is a must-have book for understanding welded structures and recent technological advances in welding, and it provides a unified summary of recent research results contributed by other researchers.

For more information, sample pages (springer), to look inside book (amazon), and available online:

«January 17, 2008»

Seminar by Distinguished Research Professor John Goldak

University of Alberta, Edmonton, Alberta, Canada

April 17, 2008 10:00 AM NINT 6-011

Computational Weld Mechanics: Current State and Future Trends

Computational weld mechanics (CWM) deals with the continuum mechanics in and near a weld heat source such as an arc, laser or electron beam. CWM involves a set of coupled problems. The first is the conservation of energy. In most cases, this can be solved to compute the temperature distribution in the solid outside of the liquid and plasma regions of the weld heat source. It requires a model of the weld heat source that adequately describes the distribution of energy in the weld pool. This heat source model does not predict the shape of the weld pool. More sophisticated models. solve the quasi-static momentum equation together with the conservation of mass and momentum are able to predict the weld pool shape including the effects of arc pressure, surface tension and hydrostatic stress. Even more sophisticated models include inertial, Lorentz and Marangoni forces in the weld pool. The ultimate in sophistication include the magneto-hydrodynamics of the plasma.

The evolution of microstructure during solidification, solid state phase transformations and grain growth can be simulated with density models that compute scalar fields, such as phase fraction, or synthetic microstructure models that have a representation of individual grains and follow their nucleation and growth.

Stress analysis is visco-elastic-plastic in the solid and Newtonian flow in the liquid. The thermal expansion of individual phases due to changes in temperature and to phase changes drive the evolution of stress and strain.

These models are 3D, nonlinear transient, coupled PDEs. We solve them using FEM methods. Because weld heat sources are often small relative to the structure being welded, they were first treated as singularities. Even today to highly localized nature of weld pools is an essential aspect of numerical models in CWM. Although they are computationally intensive they can now be solved comfortably on modern desktop computers. Managing the complexity of the geometry and the process is a significant challenge. To be useful to industry, it must be designer driven.

«April 11, 2008»

Seminar by Distinguished Research Professor John Goldak

University of Alberta, Edmonton, Alberta, Canada

April 17, 2008 3:30 PM ETLE 1-018

Designer Driven Holistic Fracture Mechanics of Welded Structures

A holistic simulation of the transient temperature distribution, microstructure evolution, stress-strain distribution during the fabrication processand possible fatigue crack growth with in-service loads is described.

This welding simulation uses the full 3D geometry of the structure being welded and typically solves hundreds of time steps in which each time stepis a coupled non-linear analysis. The software is designer driven, i.e, the software is intended to be used and is used by designers in industry.

Having completed the welding simulation of the fabrication, the behavior of the structure with its as-welded state that includes the initial state of residual stress, microstructure and distortion, the software simulates the application of in-service loads.

The stress, strain and material force distribution are computed for the in-service loads. (Material forces can be viewed as Eshelby mechanics.)A natural length scale that is a material property (see BarenBlatt for details on natural lengths in fracture) is assumed and the stress distribution is smoothed by solving a partial differential equation using this length scale as a diffusivity. This means that when the FEM mesh is significantly smaller than the natural length scale, then the FEM solution is independent of the mesh.

At crack tips the value of the material force vector is the J-integral when the mathematical assumptions required by the J-integral are satisfied. In other words, the value of the J-integral is a special case of the value of the material force at a crack tip. Unlike the J-integral the material force is always defined every where in a body.

The number of cycles for fatigue crack initiation are computed using the strain life equation. If cracks are present or initiated, then fatigue crack growth is simulated by solving the Paris-Erdogan equation for the moving crack tip. The FEM element size at a crack tip is typically of the order of 50 to 10 microns and typically cracks grow in increments of 50 to 10 microns per `time’ step.

The software makes no use of handbook formulae for fracture mechanics.

«April 11, 2008»