Simulation Workflow Development for Additive Manufacturing (2020-355-001)
Bath Iron Works
HII-Newport News Shipbuilding
NSRP ASE INVESTMENT: $127K
Given a lack of insight into manufacturing process physics, current methods for the design of additively manufactured (AM) parts require costly trial-and-error efforts to identify print parameters that result in adequate part quality. These gaps in understanding of process physics also drive current part qualification approaches that require exhaustive testing of printers, processes, and parts to develop a statistical characterization of performance. Under this effort, the project team—led by ATA Engineering—developed an integrated workflow for physics-based predictive simulation of mechanical response of AM parts, with the goal of improving insights into process physics and identifying opportunities to reduce the cost of AM part design and qualification.
Simulations of material deposition, crystal microstructure evolution, and mechanical response were connected through use of automated data mapping and translation tools to make predictions of as-built part temperature histories, microstructure features, mechanical response, residual stress, and net shape. The project team demonstrated the efficacy of the workflow by simulating directed energy deposition (DED) of 316L stainless steel. Simulation methods and results were validated against published research including both high-fidelity simulations from national laboratories and experimental results. Key innovations include the following:
- An open modeling architecture that gives users control over physics and material models for easy extensibility to additional alloys or calibration of models to available empirical data
- A modular physics modeling architecture that enables future implementation of surrogate models built with machine learning to rapidly sample the design space and identify optimal print parameters, or implementation of multiscale methods to enable scale-up of feasible part sizes at acceptable computational cost
- Direct connection of print parameters to part performance through simulation provides insights into AM process physics, with insights from this workflow uniquely embodied in a “Digital Twin” part finite element model representing the as-built condition suitable for simulation of testing or operational loading
Note: Distribution Authorized for U.S. shipbuilding and repair industry members or NSRP representatives.