US 12,013,670 B2
Monitoring a laser machining process using deep folding neural networks
Joachim Schwarz, Kleinandelfingen (CH)
Assigned to Precitec GmbH & Co. KG, Gaggenau (DE)
Appl. No. 17/295,904
Filed by Precitec GmbH & Co. KG, Gaggenau (DE)
PCT Filed Oct. 10, 2019, PCT No. PCT/EP2019/077485
§ 371(c)(1), (2) Date May 21, 2021,
PCT Pub. No. WO2020/104103, PCT Pub. Date May 28, 2020.
Claims priority of application No. 102018129441.7 (DE), filed on Nov. 22, 2018.
Prior Publication US 2022/0011726 A1, Jan. 13, 2022
Int. Cl. G05B 13/02 (2006.01); G06N 3/08 (2023.01); G06N 3/082 (2023.01); G06T 7/00 (2017.01)
CPC G05B 13/027 (2013.01) [G06N 3/082 (2013.01); G06T 7/0004 (2013.01)] 14 Claims
OG exemplary drawing
 
1. A system for monitoring a laser machining process for machining a workpiece, comprising:
a computing unit configured to determine an input tensor based on current data of the laser machining process and to determine an output tensor based on the input tensor using a transfer function, the output tensor containing information about a current machining result;
wherein said transfer function between the input tensor and the output tensor is formed by a taught neural network,
wherein:
the taught neural network comprises an individual network for process data and an individual network for image data, which are coupled at least via a common output layer, wherein the network for process data and the network for image date are separated networks; and
the taught neural network is configured to map an input tensor for process data and an input tensor for image data to a common output tensor, and
wherein the at least one output layer comprises at least one fully connected layer,
wherein the individual network for process data has fewer layers than the individual network for image data.