Accelerating Part-Scale Simulation in Liquid Metal Jet Additive Manufacturing via Operator Learning

TL;DR

This paper introduces an operator learning approach to efficiently and accurately simulate the complex physics of liquid metal jet additive manufacturing at the part scale, significantly reducing computational costs compared to traditional methods.

Contribution

The study demonstrates that operator learning can outperform kNN-based reduced-order models in accuracy, data efficiency, and generalizability for simulating droplet coalescence in liquid metal jetting.

Findings

OL requires fewer data points than kNN

OL achieves similar prediction accuracy to physics-based models

OL generalizes beyond training data

Abstract

Predicting part quality for additive manufacturing (AM) processes requires high-fidelity numerical simulation of partial differential equations (PDEs) governing process multiphysics on a scale of minimum manufacturable features. This makes part-scale predictions computationally demanding, especially when they require many small-scale simulations. We consider drop-on-demand liquid metal jetting (LMJ) as an illustrative example of such computational complexity. A model describing droplet coalescence for LMJ may include coupled incompressible fluid flow, heat transfer, and phase change equations. Numerically solving these equations becomes prohibitively expensive when simulating the build process for a full part consisting of thousands to millions of droplets. Reduced-order models (ROMs) based on neural networks (NN) or k-nearest neighbor (kNN) algorithms have been built to replace the original physics-based solver and are computationally tractable for part-level simulations. However, their quick inference capabilities often come at the expense of accuracy, robustness, and generalizability. We apply an operator learning (OL) approach to learn a mapping between initial and final states of the droplet coalescence process for enabling rapid and accurate part-scale build simulation. Preliminary results suggest that OL requires order-of-magnitude fewer data points than a kNN approach and is generalizable beyond the training set while achieving similar prediction error.