Predictive Process Control: Residual Stresses in Welding

We are collaborating with LANL and UCSD to improve failure prediction capabilities for beryllium welds. We are utilizing a combination of experimental and computational methodologies to identify factors that contribute to residual stresses in beryllium welds and to explore processing paths for mitigating those residual stresses as a means of lowering the likelihood of weld failure. Specifically, we are designing and constructing a flexible weld testing station for investigating the effect of welding process parameters on weld quality, including residual stresses. We are also selecting surrogate materials to test in place of toxic beryllium and characterizing the microstructures of welded specimens. Finally, we are creating thermal models of the transient thermal field in the braze welding zone and validating the models with experimental data.  We are using the models to support predictive (model-based) control of the welding process. 

Ray Ely

Sanjiv Shah



Los Alamos National Laboratory

Related Publications:

Shah, S. E., J. C. Lee, C. Rios-Perez and C. C. Seepersad, 2009, "Comparison of Analytical and Computational Thermal Models for Gas Metal Arc Braze Welding," ASME Summer Heat Transfer Conference, San Francisco, CA, Paper Number: HT2009-88491.

Backlund, P., J. Lee, S. Shah, C.C. Seepersad, J. Howell, E. Taleff, and M. Prime, 2008, "Thermal Modeling and Experimentation for Residual Stress Prediction in Braze-Welded Beryllium via a Surrogate CuZn10 Alloy," IRCS-8: The Eighth International Conference on Residual Stresses, Denver, CO. 

Howell, J., F. Franca, and C. C. Seepersad (Editors), 2007, Proceedings of the Pan-American Science Institute on Predictive Process Dynamics in Manufacturing, Porto Allegre, Brazil, June 2007.