Characterizing liquid reflective surfaces in 3D liquid metal printing

Patent metadata

Field	Value
Publication number	US-11958112-B2
Application number	US-202117356627-A
Country	US
Kind code	B2
Filing date	Jun 24, 2021
Priority date	Jun 24, 2021
Publication date	Apr 16, 2024
Grant date	Apr 16, 2024

Abstract

A three-dimensional (3D) printer includes a nozzle and a camera configured to capture a real image or a real video of a liquid metal while the liquid metal is positioned at least partially within the nozzle. The 3D printer also includes a computing system configured to perform operations. The operations include generating a model of the liquid metal positioned at least partially within the nozzle. The operations also include generating a simulated image or a simulated video of the liquid metal positioned at least partially within the nozzle based at least partially upon the model. The operations also include generating a labeled dataset that comprises the simulated image or the simulated video and a first set of parameters. The operations also include reconstructing the liquid metal in the real image or the real video based at least partially upon the labeled dataset.

First claim

What is claimed is:

1. A three-dimensional (3D) printer, comprising:
a nozzle;
a camera configured to capture a real image or a real video of a liquid metal while the liquid metal is positioned at least partially within the nozzle; and
a computing system configured to:
generate a model of the liquid metal positioned at least partially within the nozzle;
generate a simulated image or a simulated video of the liquid metal positioned at least partially within the nozzle based at least partially upon the model;
generate a labeled dataset that comprises the simulated image or the simulated video and a first set of parameters; and
reconstruct the liquid metal in the real image or the real video based at least partially upon the labeled dataset.

2. The 3D printer of claim 1 , wherein the model comprises a parametric model.

3. The 3D printer of claim 1 , wherein the first set of parameters comprises:
a relative weight of a function that describes an oscillation of the liquid metal;
a constant offset;
a constant representing a steady state shape of the liquid metal at rest;
or a combination thereof.

4. The 3D printer of claim 1 , wherein reconstructing the liquid metal comprises inverse mapping the real image or the real video to a second set of parameters, based at least partially upon the labeled dataset.

5. The 3D printer of claim 4 , wherein the second set of parameters comprises a subset of the first set of parameters.

6. The 3D printer of claim 5 , wherein inverse mapping comprises training an artificial neural network using the labeled dataset to predict the second set of parameters in the real image or the real video, wherein the liquid metal is reconstructed based at least partially upon the second set of parameters.

7. The 3D printer of claim 5 , wherein inverse mapping comprises selecting an entry in the labeled dataset that is most similar to the real image or the real video using direct nearest-neighbor matching, wherein the liquid metal is reconstructed based at least partially upon the selected entry.

8. The 3D printer of claim 1 , wherein the computing system is further configured to extract one or more metrics from the reconstructed liquid metal, wherein the one or more metrics comprise a carrier oscillation frequency, a pulse-to-pulse covariance, a waveform decay rate, or a combination thereof.

9. The 3D printer of claim 8 , wherein the computing system is further configured to adjust one or more parameters of the 3D printer based at least partially upon the one or more metrics to adjust the liquid metal.

10. The 3D printer of claim 9 , wherein the one or more parameters comprises an amplitude, a frequency, or both of a waveform from a power source that causes the liquid metal to be jetted through the nozzle.

11. A three-dimensional (3D) printer configured to print a 3D object, the 3D printer comprising:
a nozzle;
a camera configured to capture a plurality of real images of a plurality of drops of liquid metal while the drops are positioned at least partially within the nozzle, wherein each drop comprises a meniscus; and
a computing system configured to:
generate a parametric model of motion of the menisci of the drops positioned at least partially within the nozzle based at least partially upon the real images;
generate a plurality of simulated images of the drops positioned at least partially within the nozzle based at least partially upon the parametric model;
generate a labeled dataset that comprises the simulated images and a first set of parameters;
inverse map the real images to a second set of parameters, based at least partially upon the labeled dataset;
reconstruct the menisci of the drops in the real images based at least partially upon the second set of parameters;
extract one or more metrics from the reconstructed menisci of the drops, wherein the metrics comprise a carrier oscillation frequency, a pulse-to-pulse covariance, a waveform decay rate, or a combination thereof; and
adjust one or more parameters of the 3D printer based at least partially upon the one or more metrics to adjust the menisci of the drops positioned at least partially within the nozzle.

12. The 3D printer of claim 11 , wherein the first set of parameters comprises:
a relative weight of a function that describes an oscillation of the drops;
a constant offset; and
a constant representing a steady state shape of the drops at rest.

13. The 3D printer of claim 11 , wherein the second set of parameters comprises a subset of the first set of parameters.

14. The 3D printer of claim 11 , wherein inverse mapping comprises training an artificial neural network using the labeled dataset to predict the second set of parameters in the real images, wherein the menisci are reconstructed based at least partially upon the predicted second set of parameters.

15. The 3D printer of claim 11 , wherein inverse mapping comprises selecting an entry in the labeled dataset that is most similar, in mean-square error, to the real images using direct nearest-neighbor matching, wherein the menisci are reconstructed based at least partially upon the selected entry.

16. A three-dimensional (3D) printer configured to print a 3D object by jetting a plurality of drops of liquid metal of onto a substrate, the 3D printer comprising:
an ejector comprising a nozzle;
a heating element configured to heat a solid metal within the ejector, thereby converting the solid metal to the liquid metal;
a coil wrapped at least partially around the ejector;
a power source configured to transmit voltage pulses to the coil, wherein the coil causes the plurality of drops of the liquid metal to be jetted through the nozzle in response to the voltage pulses;
a camera configured to capture a plurality of real images of the drops while the drops are positioned at least partially within the nozzle, wherein each drop comprises a meniscus;
a light source configured to illuminate the nozzle and the drops as the real images are captured; and
a computing system configured to:
generate a parametric model of dynamic motion of the menisci of the drops positioned at least partially within the nozzle based at least partially upon the real images;
generate a plurality of simulated images of the drops positioned at least partially within the nozzle based at least partially upon the parametric model;
generate a labeled dataset that comprises the simulated images and a first set of parameters, wherein the first set of parameters comprises:
a first relative weight of a first Bessel function that describes an oscillation of the liquid;
a second relative weight of a second Bessel function that describes the oscillation of the liquid;
a constant offset; and
a constant representing a steady state shape of the meniscus at rest;

inverse map the real images to a second set of parameters, based at least partially upon the labeled dataset, wherein inverse mapping comprises:
training an artificial neural network using the labeled dataset to predict the second set of parameters in the real images; or
selecting an entry in the labeled dataset that is most similar, in mean-square error, to the real images using direct nearest-neighbor matching;

reconstruct the menisci of the drops in the real images based at least partially upon the second set of parameters;
extract one or more metrics from the reconstructed menisci of the drops, wherein the metrics comprise a carrier oscillation frequency, a pulse-to-pulse covariance, and a waveform decay rate; and
adjust one or more parameters of the 3D printer based at least partially upon the

Assignees

• Xerox Corp

Inventors

• Price Robert R
• Bala Raja
• Korneev Svyatoslav
• Somarakis Christoforos
• Shreve Matthew
• Lew Adrian
• Ramesh Palghat

Classifications

• B22F10/80 Primary

  Data acquisition or data processing · CPC title

• B22D23/003 Primary

  Moulding by spraying metal on a surface · CPC title

• B22F10/22

  Direct deposition of molten metal · CPC title

• B22F12/90

  Means for process control, e.g. cameras or sensors · CPC title

• B33Y30/00

  Apparatus for additive manufacturing; Details thereof or accessories therefor · CPC title