About this Abstract |
Meeting |
2021 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2021)
|
Symposium
|
Special Session
|
Presentation Title |
Droplet Diagnostics in Droplet-on-Demand Liquid Metal Jetting Using Heterogenous Sensing |
Author(s) |
Aniruddha Gaikwad, Tammy Chang, Nicholas Watkins, Brian Giera, Saptarshi Mukherjee, David Stobbe, Andrew Pascall, Prahalada Rao |
On-Site Speaker (Planned) |
Aniruddha Gaikwad |
Abstract Scope |
The goal of this work is to monitor the characteristics of jetted liquid metal droplets in a droplet-on-demand liquid metal jetting (DoD-LMJ) additive manufacturing process. A millimeter-wave near-field sensor and high-speed video camera were deployed in a custom tin liquid metal jetting system. Experiments were conducted to capture the effect of processing parameters on the droplet characteristics, viz., size, velocity, and shape of droplets. Subsequently, predictive models were developed to monitor these droplet characteristics using the multimodal sensor data, and then predict droplet parameters from the millimeter-wave signals alone, to achieve a major reduction in processing requirements. We present a multilayer perceptron-based non-linear autoregression with exogenous inputs model to predict droplet size and velocity. Additionally, we developed an ensemble machine learning model to monitor the shape of droplets. Our models predict/classify with a statistical fidelity exceeding 0.9 (R2 and F1-score), demonstrating the potential of heterogeneous sensing in LMJ systems. |
Proceedings Inclusion? |
Definite: Post-meeting proceedings |