| Abstract Scope |
Introduction:
U.S. Navy researchers recently developed a high-strength, high-toughness solid wire welding electrode based on the Fe-10wt% Ni alloy system. The weld metal has demonstrated excellent mechanical properties for both the GTAW and GMAW processes. However, the toughness of the GTA weld metal is significantly higher than seen in the GMA weld metal. Previous research has identified the factors that play a role in toughness variation between these welding processes. The GMAW welds, specifically those produced in the spray transfer mode have higher oxide content and produce irregularly shaped weld beads. That results in non-uniform reheating/degree of recrystallization and tempering. The GMAW welds also have larger effective martensite grains and larger amounts of coarse-grain martensite.
The objective of this work is to reduce the toughness variability in Fe-10Ni groove welds produced with the high deposition rate GMAW process. The conventional trial and error approach in welding procedure development and optimization is time, material, labor, and cost-intensive, and frequently inefficient in achieving desirable weld properties. To address this, a computational design of experiments (CDoE) approach for process-microstructure-property optimization during materials processing with multiple reheats was recently developed at The Ohio State University. The work presented here will describe the development of a CDoE framework for process optimization and improvement of the weld metal microstructure and impact toughness of Fe-10Ni steel groove welds.
Experimental Procedure:
The CDoE framework utilizes a DoE, finite element analysis (FEA), post-processing modules, and a database of multiple reheat response relationships for the tested materials. The latter are developed by physical experimentation or thermodynamic and kinetic simulations. The CDoE framework develops a matrix of welding processes (simulation cases) defined by combinations of variable weld parameters. Each simulation case is run through the FEA module to generate thermal histories for all nodes in a FEA model of the tested weld. The post-processing module utilizes the predicted thermal histories and predetermined reheat response relationships to generate microstructure and property contour maps for the simulated weld. The outputs of all simulation cases are compared to identify a window of optimal welding parameters.
Results and Discussion:
GMAW and GTAW groove weld FEA models were developed in SysWeld. The models utilized calibrated heat sources that produce the overall bead morphology seen in real welds. Physical simulations of multiple reheats using the Gleeble 3800TM thermo-mechanical simulator were conducted to develop a thermal history-microstructure-property relationship for Fe-10Ni weld metal. The effect of reheating on prior austenite grain size was researched and an empirical equation for grain growth in Fe-10Ni weld metal was developed. The FEA models, grain growth models, and property relationships were validated using real weld thermal histories and properties. They were then integrated into the CDoE framework, and procedure optimization experiments were conducted.
Conclusions:
The conventional trial and error approach in welding procedure development and optimization is time, material, labor, and cost-intensive, and frequently inefficient in achieving desirable weld properties. To address this, a computational design of experiment (CDoE) approach was utilized to optimize Fe-10Ni weld properties, specifically toughness. The CDoE approach evaluates large matrices of welding process parameters and is time, labor, and materials efficient. Thermal history-property relationships were developed for Fe-10Ni weld metal and experimentally validated. Physical simulation experiments and metallurgical characterization were conducted to develop a prior austenite grain growth model for Fe-10Ni weld metal.
|