# US 2023/0273605 A1 · Machine learning device, additive manufacturing system, machine learning method for welding condition, method for adjusting welding condition, and a non-transitory computer readable medium storing a program

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2023/0273605 A1
- Document ID: US20230273605A1
- Status: Published application
- Kind: A1
- Application number: 18/005,574
- Filed: 2021-06-16
- Published: 2023-08-31
- Earliest claimed: 2020-07-15
- Assignees: Kabushiki Kaisha Kobe Seiko Sho (Kobe Steel, Ltd.)
- Applicants: Kabushiki Kaisha Kobe Seiko Sho (Kobe Steel, Ltd.)
- Inventors: Shuo HUANG
- Web page: https://patentgrep.com/patent/US20230273605A1

## CPC

- B33Y 50/00: Data acquisition or data processing for additive manufacturing
- G05B 19/41865: {characterised by job scheduling, process planning, material flow}
- B23K 9/04: Welding for other purposes than joining, e.g. built-up welding
- B23K 9/0953: {using computing means}
- B23K 9/173: and of a consumable electrode
- B23K 31/006: {relating to using of neural networks}
- B33Y 30/00: Apparatus for additive manufacturing; Details thereof or accessories therefor
- B33Y 50/02: for controlling or regulating additive manufacturing processes
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/09: Supervised learning
- G05B 2219/45135: Welding
- Y02P 10/25: Process efficiency
- B33Y: ADDITIVE MANUFACTURING, i.e. MANUFACTURING OF THREE-DIMENSIONAL [3D] OBJECTS BY ADDITIVE DEPOSITION, ADDITIVE AGGLOMERATION OR ADDITIVE LAYERING, e.g. BY 3D PRINTING, STEREOLITHOGRAPHY OR SELECTIVE LASER SINTERING
- G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G)
- B23K: SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- Y02P: CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS

## Abstract

A machine learning device that performs machine learning of a welding condition for manufacturing an additively-manufactured object by welding a filler metal and depositing weld beads, the machine learning device includes: at least one hardware processor configured to perform a learning process for generating a learned model using two pieces of shape data of a weld bead or a difference between the two pieces of shape data is used as input data and a difference between welding conditions corresponding to the difference between the two pieces of shape data as output data.

## Claims

### Claim 1

A machine learning device that performs machine learning of a welding condition for manufacturing an additively-manufactured object by welding a filler metal and depositing weld beads, the machine learning device comprising:
at least one hardware processor configured to perform a learning process for generating a learned model using two pieces of shape data of a weld bead or a difference between the two pieces of shape data is used as input data and a difference between welding conditions corresponding to the difference between the two pieces of shape data as output data.

### Claim 2 (depends on 1)

The machine learning device according to claim 1, wherein
the shape data includes at least one of a height, a width, an angle of a root part, and surface unevenness of a weld bead.

### Claim 3 (depends on 1)

The machine learning device according to claim 1, wherein
the welding condition includes at least one of a feeding speed of the filler metal, a travel speed, a current or a voltage of welding, a target position on a base on which the additively-manufactured object is manufactured, an amount of heat input at a time of manufacturing, and a weaving control condition.

### Claim 4 (depends on 3)

The machine learning device according to claim 3, wherein
the welding condition further includes the number of deposition passes or the temperature of a base material.

### Claim 5 (depends on 1)

The machine learning device according to claim 1, wherein
the at least one hardware processor generates a learned model for each control mode of a power supply used when the additively-manufactured object is manufactured.

### Claim 6 (depends on 5)

The machine learning device according to claim 5, wherein
at least one of a voltage value, a current value, and a current pulse supplied corresponding to the welding condition is different for each control mode of the power supply.

### Claim 7 (depends on 1)

The machine learning device according to claim 1, wherein
the at least one hardware processor performs the learning process using a supervised learning method using a neural network.

### Claim 8

An additive manufacturing system for manufacturing an additively-manufactured object by welding a filler metal and depositing weld beads, the additive manufacturing system comprising:
at least one hardware processor configured to
create shape data of a weld bead as first shape data based on design data of the additively-manufactured object;
determine a welding condition for forming the first shape data;
acquire, as second shape data, shape data of a weld bead formed using the determined welding condition;
derive a difference between a welding condition corresponding to the first shape data and a welding condition corresponding to the second shape data by inputting the first shape data and the second shape data or a difference between the first shape data and the second shape data to a learned model, the learned model being generated by performing a learning process using two pieces of shape data of a weld bead or a difference between the two pieces of shape data as input data and a difference between welding conditions corresponding to the difference between the two pieces of shape data as output data; and
adjust the determined welding condition using the derived difference.

### Claim 9 (depends on 8)

The additive manufacturing system according to claim 8, wherein
the at least one hardware processor derives a difference for adjusting the determined welding condition when the difference between the first shape data and the second shape data is equal to or greater than a predetermined threshold.

### Claim 10 (depends on 8)

The additive manufacturing system according to claim 8, wherein
the at least one hardware processor determines the welding condition for forming the first shape data using a database in which a shape of a weld bead is associated with a welding condition in advance.

### Claim 11 (depends on 8)

The additive manufacturing system according to claim 8, wherein
the at least one hardware processor acquires the shape data of the weld bead by measuring the weld bead using a sensor.

### Claim at position 12

-13. (canceled)

### Claim 14

A non-transitory computer readable medium storing a program configured to cause a computer to execute a learning processing step of performing a learning process for generating a learned model using two pieces of shape data of a weld bead at the time of manufacturing an additively-manufactured object or a difference between the two pieces of shape data as input data and a difference between welding conditions corresponding to the difference between the two pieces of shape data as output data.

### Claim 15

(canceled)
