Meet Grey

Grey-1 is our next-gen AI brain for additive manufacturing. Already equipped with advanced error detection, parameter prediction and optimisation capabilities, Grey continues to learn with every print within our global network of 3D printers.

arrow_circle_down

Powered by Matta's global data engine

Grey-1 learns via deep analysis of industrial data, obtained from our global network of 3D printers.

arrow_circle_down

Pop the hood

open_in_full
Click to enlarge
open_in_full
Click to enlarge
open_in_full
Click to enlarge
Recycled PLA • Consumer product • Home decor item
Ultrafuse 17-4 PH • Engineering part • Helical gear
ABS-X • Product prototype • Cable tidy solution

arrow_circle_down

Moving on

Detects errors in real time

Grey-1 can autonomously monitor additive manufacturing process and find errors in real time.
add_circle
Learn more

Rapidly identify printing errors

Trained on countless hours of extrusion printing data, Grey-1 is able to find printing issues fast with remarkable accuracy.

Intelligent and robust, Grey-1 exhibits excellent performance across a wide number of geometries, materials, printers, toolpaths and even extrusion methods. We hope to train Grey-1 further to be able to find errors across a plethora of manufacturing processes, and to find new applications for its outputs.
Bed detachment
Poor first layer
Layer offset
Drag to pan through print
Drag to pan through print
error
User notified
Status: Printing

Part: XYZCube.gcode (432kB)
Layer: 41

Printing time: 32 minutes
Estimated time remaining: 10 minutes

Errors detected: 4
Print quality estimate: 60%

Recommended adjustments:

• Reduce printhead acceleration
• Check XY belt tensions
‍
Drag to pan through print

arrow_circle_down

Continue

Generates and analyses digital twin parts

Grey-1 creates digital models of printed parts using G-code, mapping errors to localised regions.
add_circle
Learn more

Supercharged quality control

Without expensive or time-consuming scanning methods, manual QC is limited to outer faces of parts. However, Grey-1 sees every inch of parts during the printing process, inside and out, making it a powerful tool for qualification.

Using a variety of data streams, Grey-1 maps identified printing errors to their specific locations on parts. Using G-code, Grey-1 then generates a digital twin of the part, and projects quality measurements onto each extruded line, ready for inspection.

compare

Slide to compare a real part with Grey-1's predictive quality digital twin
left arrow icon
right arrow icon

arrow_circle_down

What else can Grey do?

Corrects errors on the fly

Grey-1 learns not only to find errors, but to identify their causes. Combined with closed-loop control algorithms, Grey-1 can correct sub-optimal parameters on the fly, ensuring top quality prints, first time.
add_circle
Learn more

Manufacturing autopilot with parameter correction

Grey-1 can monitor printing parameters, such as flow rate or z-offset, and predict their deviation from optimal. When used in parallel with closed-loop control algorithms, Grey-1 becomes a manufacturing autopilot, continuously controlling parameters throughout the printing process to minimise the probability of errors and maximise quality.

In the video examples in this section, parameter-dependant printing errors were artificially generated while printing. Grey-1 can be seen to quickly identify the problem and its cause, then make the required adjustment to the parameter, returning print characteristics to nominal.
Flow error
Speed error
Z-offset error
Hotend error

Learns to print new materials

Grey-1 has learned enough about the fundamental physics of extrusion processes, that it can work out how to print new unseen materials.
Grey enables next-level batch-to-batch material qualification by discovering optimum processing parameters within seconds of printing.
add_circle
Learn more

Material profile discovery

Grey-1 was only trained on PLA, but is programmed to learn in a way that makes its knowledge transferable between materials. Watch Grey-1 find the optimum processing parameters for materials it has never seen before in this section.

In these examples, parameters are generated randomly at the beginning of the print. Grey-1 then quickly analyses the visual data, identifying and executing adjustments to increase quality until optimum print characteristics are obtained.
ABS-X
CF Nylon
PVA HT+
TPU

Read the papers

See how Grey was built

Want to learn more?

Sign up for the latest updates

Thank you!
Oops! Something went wrong while submitting the form.
Main pages
OSResearchAbout
CareersPrivacy policyContact
Contact us
hello@matta.ai
@mattalabs

Building AI copilots for advanced manufacturing

Matta Labs © 2022-2024