Development of a Comprehensive Weld Process Model

This cooperative research and development agreement (CRADA) between Concurrent Technologies Corporation (CTC) and Lockheed Martin Energy Systems (LMES) combines CTC's expertise in the welding area and that of LMES to develop computer models and simulation software for welding processes. This development is of significant impact to the industry, including materials producers and fabricators. The main thrust of the research effort was to develop a comprehensive welding simulation methodology. A substantial amount of work has been done by several researchers to numerically model several welding processes. The primary drawback of most of the existing models is the lack of sound linkages between the mechanistic aspects (e.g., heat transfer, fluid flow, and residual stress) and the metallurgical aspects (e.g., microstructure development and control). A comprehensive numerical model which can be used to elucidate the effect of welding parameters/conditions on the temperature distribution, weld pool shape and size, solidification behavior, and microstructure development, as well as stresses and distortion, does not exist. It was therefore imperative to develop a comprehensive model which would predict all of the above phenomena during welding. The CRADA built upon an already existing three- dimensional (3-D) welding simulation model which was developed by LMES which is capable of predicting weld pool shape and the temperature history in 3-d single-pass welds. However, the model does not account for multipass welds, microstructural evolution, distortion and residual stresses. Additionally, the model requires large resources of computing time, which limits its use for practical applications. To overcome this, CTC and LMES have developed through this CRADA the comprehensive welding simulation model described above. The following technical tasks have been accomplished as part of the CRADA. 1. The LMES welding code has been ported to the Intel Paragon parallel computer at ORNL. The timing results illustrate the potential of the modified computer model for the analysis of large-scale welding simulations. 2. The kinetics of grain structure evolution in the weld heat affected zone (HAZ) has been simulated with reasonable accuracy by coupling an improved MC grain growth algorithm with a methodology for converting the MC parameters of grain size and time to real parameters. The simulations effectively captured the thermal pinning phenomenon that has been reported in the weld HAZ. 3. A cellular automaton (CA) code has been developed to simulate the solidification microstructure in the weld fusion zone. The simulations effectively captured the epitaxial growth of the HAZ grains, the grain selection mechanism, and the formation of typical grain structures observed in the weld t%sion zone. 4. The point heat source used in the LMES welding code has ben replaced with a distributed heat source to better capture the thermal characteristics and energy distributions in a commercial welding heat source. 5. Coupled thermal-mechanical and metallurgical models have been developed to accurately predict the weld residual stresses, and 6. Attempts have been made to integrate the newly developed computational capabilities into a comprehensive weld design tool