IP Library Granted Patent US 12697781
Granted Patent B1
US 12697781 · App. 18/483,958 · Granted Aug 4, 2026

Machine learning for real-time monitoring and control of additive manufacturing

Inventors: Devin John Roach (Albuquerque, NM); Andrew Rohskopf (San Jose, CA); William Derek Reinholtz (Albuquerque, NM); Adam Wade Cook (Albuquerque, NM); Leah N. Appelhans (Tijeras, NM)
Assignee: National Technology & Engineering Solutions of Sandia, LLC
B29C64/393B29C64/118B33Y50/02
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12697781
App. No.
18/483,958
Granted
Aug 4, 2026
Kind
B1
Abstract

A machine learning (ML) approach enables real-time direct ink write (DIW) print-parameter optimization through in-situ monitoring of printed line geometry. The method can use an invertible neural network (INN) to solve both forward and inverse, or optimization, problems using a single network. By combining in-situ computer vision and INNs, DIW printing parameters can be autonomously optimized to print a target line width in a matter of seconds. Furthermore, defects that occur during printing can be rapidly identified and corrected autonomously. The method eliminates user-intensive, time-consuming, iterative parameter discovery approaches that currently limit accelerated implementation of DIW and other extrusion-based additive manufacturing processes.

Claims (9)

1 . A method for in-situ process monitoring and optimization of direct ink write (DIW) printing, comprising:

providing a DIW printer configured to print an ink subject to one or more printing parameters,

measuring an output of the printed ink in real-time using computer vision,

comparing the measured output to a target output,

when the output differs from the target output, adjusting the one or more printing parameters of the DIW printer in real time using a pre-trained invertible neural network model to produce the target output, wherein the invertible neural network models both a forward and an inverse relationship between the one or more printing parameters and the target output, and

printing the ink on a substrate subject to the optimized one or more printing parameters in a single printing step.

2 . The method of claim 1 , wherein the one or more printing parameters comprises a nozzle diameter, print speed, linear extrusion rate, or height from the substrate.

3 . The method of claim 1 , wherein the output comprises a line width, line type, or line shape.

4 . The method of claim 1 , wherein the invertible neural network model and optimizing step finds the one or more printing parameters associated with the target line width that are closest in value to current printing parameters.