# US 2023/0390824 A1 · Method for detecting defects during laser additive manufacturing, data processing device, computer program and storage medium for implementing the method

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

- Number: US 2023/0390824 A1
- Document ID: US20230390824A1
- Status: Published application
- Kind: A1
- Application number: 18/250,003
- Filed: 2021-10-22
- Published: 2023-12-07
- Earliest claimed: 2020-10-22
- Assignees: Arianegroup SAS
- Applicants: Arianegroup SAS
- Inventors: André CHANDELLE; Guillaume REDOULES
- Web page: https://patentgrep.com/patent/US20230390824A1

## CPC

- B22F 10/80: Data acquisition or data processing
- B22F 10/28: Powder bed fusion, e.g. selective laser melting [SLM] or electron beam melting [EBM]
- B22F 10/85: for controlling or regulating additive manufacturing processes
- B22F 12/90: Means for process control, e.g. cameras or sensors
- B23K 26/034: {Observing the temperature of the workpiece}
- B23K 26/342: Build-up welding
- B23K 31/125: {Weld quality monitoring}
- B29C 64/153: using layers of powder being selectively joined, e.g. by selective laser sintering or melting
- B29C 64/393: for controlling or regulating additive manufacturing processes
- B33Y 50/00: Data acquisition or data processing for additive manufacturing
- G06N 3/045: Combinations of networks
- G06N 3/0455: Auto-encoder networks; Encoder-decoder networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/09: Supervised learning
- G06N 3/0985: Hyperparameter optimisation; Meta-learning; Learning-to-learn
- G06T 7/0004: {Industrial image inspection}
- B33Y 10/00: Processes of additive manufacturing
- B33Y 30/00: Apparatus for additive manufacturing; Details thereof or accessories therefor
- G06T 2207/10048: Infrared image
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T 2207/30164: Workpiece; Machine component
- Y02P 10/25: Process efficiency
- B22F: WORKING METALLIC POWDER; MANUFACTURE OF ARTICLES FROM METALLIC POWDER; MAKING METALLIC POWDER (making alloys by powder metallurgy C22C); APPARATUS OR DEVICES SPECIALLY ADAPTED FOR METALLIC POWDER
- 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
- B29C: SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING (making preforms B29B11/00; making laminated products by combining previously unconnected layers which become one product whose layers will remain together B32B37/00 - B32B41/00)
- 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
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- Y02P: CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS

## Abstract

A method for detecting defects during laser additive manufacturing, comprising the following steps: B21) a first image is captured, the first image being the image, captured in the infrared spectrum, of an upper surface of a layer of powder exposed to laser scan; B23) the first image is processed using a first convolutional neural network of self-encoding type, in such a way as to produce a defect mask indicating the location of defects at the upper surface of the layer of powder. A method for manufacturing parts during which the presence of defects is detected using the preceding method. A data-processing device, computer program and storage medium for the implementation of this method.

## Claims

### Claim 1

A method for detecting defects during laser additive manufacturing, comprising the following steps:
B21) a first image is captured, said first image being the image, captured in the infrared spectrum, of an upper surface of a layer of powder exposed to laser scan;
B23) said first image is processed using a first convolutional neural network of self-encoding type, in such a way as to produce a defect mask indicating a location of defects at the upper surface of said layer of powder;
wherein during the capturing step B21, a second image of the upper surface exposed to laser radiation is also captured;
the second image is an image of said surface, captured in the infra-red spectrum, wherein for each pixel, the intensity of the pixel represents the maximum luminous energy received by the pixel during a capturing period; and
during the processing step B23, the processing carried out takes the second image into account as input, in addition to the first image.

### Claim 2 (depends on 1)

The method for detecting defects as claimed in claim 1, wherein said first image is an integrated image of the upper surface, wherein for each pixel, an intensity of the pixel represents the cumulative luminous energy received by the pixel during a capturing period.

### Claim 3

(canceled)

### Claim 4 (depends on 1)

The method for detecting defects as claimed in claim 1, wherein during the capturing step, a third image of the upper surface exposed to the laser scan is also captured;
the third image is an image of said surface, captured in the visible spectrum; and
during the processing step, the processing performed takes the third image into account as input, in addition to the first image.

### Claim 5 (depends on 1)

The method for detecting defects as claimed in claim 1, wherein said first convolutional neural network is configured to receive as input the first image and the second image.

### Claim 6 (depends on 5)

The method for detecting defects as claimed in claim 5, wherein the defect mask is directly produced by the first convolutional neural network.

### Claim 7 (depends on 5)

The method for detecting defects as claimed in claim 5, wherein
during the processing step, the processing is carried out using a set of neural networks comprising as input, said first convolutional neural network, and as output, an output neural network, particularly of self-encoding type, configured to receive as input the output of the first convolutional neural network and to produce the defect mask based on this input.

### Claim 8 (depends on 4)

The method for detecting defects as claimed in claim 4, wherein during the processing step, the processing is carried out using a set of neural networks comprising:
as input, said first convolutional neural network, and at least one first additional neural network of self-encoding type configured to receive as input a first additional image from among said second image and said third image; and
as output, an output neural network of self-encoding type configured to receive as input the outputs of the first neural network and of said at least one first additional neural network, and to produce as output said defect mask.

### Claim 9 (depends on 8)

The method for detecting defects as claimed in claim 8, wherein the set of neural networks further comprises a second additional neural network of self-encoding type configured to receive as input a second additional image from among said second image and said third image, and which is other than the first additional image.

### Claim 10 (depends on 1)

The method for detecting defects as claimed in claim 1, wherein said first neural network comprises connections linking neural layers of index y to neural layers of index n-y, the total depth of said at least one of the neural networks considered being equal to n.

### Claim 11 (depends on 1)

The method for detecting defects as claimed in claim 1, wherein the capturing step is carried out using a photographic camera or a video camera, during the manufacturing of parts by powder bed laser fusion.

### Claim 12 (depends on 1)

The method for detecting defects as claimed in claim 1, wherein in the processing step, the defect mask is directly produced by a convolutional neural network.

### Claim 13 (depends on 1)

The method for detecting defects as claimed in claim 1, wherein during the processing applied to said first image using said first convolutional neural network and producing the defect mask, said first image is transmitted only to an input of the first convolutional neural network.

### Claim 14 (depends on 1)

A method for manufacturing parts by powder bed laser fusion, in which at least one part is manufactured by powder bed laser fusion, and during or after the manufacturing of said at least one part, defects are detected by employing the method as claimed in claim 1.

### Claim 15 (depends on 1)

A data-processing device comprising at least one processor as well as a memory in which instructions are recorded which, when they are executed by said at least one processor, lead said at least one processor to execute the steps of the method as claimed in claim 1.

### Claim 16 (depends on 15)

A system for manufacturing parts by powder bed laser fusion, comprising a machine for manufacturing parts by powder bed laser fusion and a data-processing device as claimed in claim 15.

### Claim 17 (depends on 1)

A computer program comprising instructions which, when the program is executed by at least one processor, leads said at least one processor to execute the steps of the method as claimed in claim 1.

### Claim 18 (depends on 17)

A non-volatile storage medium readable by computer, on which is stored the computer program as claimed in claim 17.

### Claim 19 (depends on 4)

The method for detecting defects as claimed in claim 4, wherein said first convolutional neural network is configured to receive as input the first image, the second image, and the third image.

### Claim 20 (depends on 2)

The method for detecting defects as claimed in claim 2, wherein the capturing period has a duration greater than 0.1 second.
