Computational Tool Development for Weld Sequence Planning in Major Assemblies

Abstract

For the U.S. Navy, the use of computational simulations is prevalent for structural finite-element analysis but not for shop floor fabrication during construction. However, prevention and mitigation of welding-induced deformation creates a significant manufacturing challenge during fabrication of major ship assemblies, especially for thin-plate steel construction. The objective of this project was to improve weld sequence planning (WSP) capabilities for major ship assemblies through the development of a quick and user-friendly WSP software tool. An approximately 5× reduction in analysis time (from model setup through solve time) was realized through process automation, development of a weld joint database, and two-step weld sequence optimization algorithms. Physical testing of tank-like structures validated the computational tool, which established high correlation between measured and predicted distortion results. Sequence optimization analysis for an eggcrate structure showed a 43% reduction in maximum distortion from the two-step optimization process within the WSP tool. The end goal of this program is improved confidence in, and use of, computational weld mechanics techniques to more cost-effectively serve the U.S. Navy enterprise. Within shipbuilding, the construction of major ship assemblies (e.g., foundation tanks, bulkheads, and deck plating) can result in significant welding-induced deformation, especially in thin-plate steel construction (Spraragen & Ettinger 1950). Prevention and mitigation of this distortion typically creates a significant manufacturing challenge to the fabrication shop floor in terms of impact to cost and schedule. In addition, the skilled trades do not typically have weld sequence and clamping plans for major structures, instead relying on trade knowledge (i.e., prior experience paired with trial and error) with limited documentation across successive (but corollary) builds. A more rigorous approach using computational weld mechanics (CWM) techniques would involve finite-element analysis (FEA) of the welded component. CWM techniques enable better documentation (possessing a digital component) and sequence optimization for distortion reduction. However, current FEA tools using a transient heat source model and an implicit solver require days, weeks, or even months to prepare the computational model, run the simulation, and analyze the results for major ship assemblies because of their size relative to the size of weld beads. This lengthy calculation time is not feasible for use in a shipyard environment. In addition, the users of the transient heat source FEA tools are typically highly trained, with PhD.-level experience in computational simulation, and are not typically found on the production floor of most shipyards.