Laser machining system

Patent metadata

Field	Value
Publication number	US-11433478-B2
Application number	US-202016889949-A
Country	US
Kind code	B2
Filing date	Jun 2, 2020
Priority date	Jun 7, 2019
Publication date	Sep 6, 2022
Grant date	Sep 6, 2022

Abstract

The laser machining system includes a laser device configured to output a laser beam, and a machining head configured to emit the laser beam emitted by a laser oscillator of the laser device and propagated through an optical fiber, to a workpiece in order to perform laser machining. The machining head includes at least one wavelength selective mirror having wavelength selectivity with various values of reflectivity and transmittance according to wavelengths, and at least one image capturing device. The laser machining system monitors abnormality in a laser optical system leading from the laser oscillator to the machining head, during the laser machining, by reflecting light propagated from a side of introduction of the laser beam into the machining head by the wavelength selective mirror, making the light incident on an image capturing surface of the image capturing device, and detecting incident light illuminance distribution appearing on the image capturing surface of the image capturing device.

First claim

What is claimed is:

1. A laser machining system comprising:
a laser device configured to output a laser beam; and
a machining head configured to emit the laser beam toward a workpiece in order to perform laser machining, the laser beam emitted by a laser oscillator of the laser device and propagated through an optical fiber,
the machining head including:
at least one wavelength selective mirror having wavelength selectivity with various values of reflectivity and transmittance according to wavelengths; and
at least one image capturing device, wherein
the laser machining system monitors abnormality in a laser optical system leading from the laser oscillator to the machining head, during the laser machining, by reflecting selected wavelengths of light propagated from the laser oscillator at a side of introduction of the laser beam into the machining head via the wavelength selective mirror, making the selected wavelengths of the light incident on an image capturing surface of the image capturing device, and detecting an incident light illuminance distribution appearing on the image capturing surface of the image capturing device, and
at least a portion of the selected wavelengths of the light propagated from the laser oscillator is reflected by the wavelength selective mirror upstream from a laser beam emitting end of the machining head and onto the image capturing surface of the image capturing device before the laser beam reaches the laser beam emitting end of the machining head, the laser beam emitting end of the machining head being an end of the machining head where the laser beam is emitted from the machining head toward the workpiece.

2. The laser machining system according to claim 1 , wherein
at least the one image capturing device is a first image capturing device including a pixel having sensitivity at least to light having a wavelength of a laser beam,
the laser beam emitted from a termination surface of the optical fiber connected to the machining head is made incident to a plurality of the pixels of the image capturing surface of the first image capturing device, and
the first image capturing device detects the incident light illuminance distribution appearing on the image capturing surface of the first image capturing device of the laser beam emitted from the termination surface of the optical fiber, and thereby
the laser machining system monitors abnormality in the laser optical system leading from the laser oscillator to the machining head.

3. The laser machining system according to claim 1 , wherein
at least the one image capturing device is a second image capturing device including a pixel having sensitivity to light having at least one wavelength other than a wavelength of a laser beam, and
the second image capturing device is capable of capturing an image of at least a laser beam introduction side inner wall of a machining head body container.

4. The laser machining system according to claim 3 , wherein
the laser machining system is connected to a machine learning device, and
the machine learning device is configured to:
acquire, as input data, at least light output command data issued to the laser oscillator and data of the incident light illuminance distribution;
acquire, as a label, an evaluation value indicating whether the laser optical system leading from the laser oscillator to the machining head is in a normal state or an abnormal state;
prepare the input data and the label in a pair as teacher data;
learn relation between the input data and the label by supervised learning and build a learning model;
output, as a prediction value, prediction result indicating presence or absence of the abnormality in the laser optical system leading from the laser oscillator to the machining head, by using the learning model, with respect to new input data including at least light output command data and data of the incident light illuminance distribution; and
when the prediction result indicating occurrence of the abnormality in the laser optical system is output, stop or reduce the output laser beam to be output by the laser device.

5. The laser machining system according to claim 4 , wherein
the machine learning device is communicably connected to a plurality of the laser devices via a communication network, and
the machine learning device performs the supervised learning, by utilizing the input data and the label in the pair acquired from the plurality of laser devices.

6. The laser machining system according to claim 5 , wherein
at least one of the laser devices including:
a learning result recording part configured to record the learning model, wherein
the laser machining system records the learning model built by the machine learning device, in the learning result recording part, and
outputs the prediction value with respect to data input in the learning result recording part, by using the learning model recorded in the learning result recording part.

7. The laser machining system according to claim 1 , wherein
the image capturing device is a first color image capturing device including pixels having respectively different sensitivity at least to light having a wavelength of a laser beam and to light having at least one wavelength other than the wavelength of the laser beam, wherein
the first color image capturing device captures an image of at least a laser beam introduction side inner wall of a machining head body container, by use of the pixel of the first color image capturing device having high sensitivity to the light having at least one wavelength other than the wavelength of the laser beam, and
detects the incident light illuminance distribution appearing on the image capturing surface of the first color image capturing device of the laser beam emitted from a termination surface of the optical fiber connected to the machining head, and made incident to a plurality of the pixels of the first color image capturing device having high sensitivity to the light having the wavelength of the laser beam, by use of the pixel of the first color image capturing device having high sensitivity to the light having the wavelength of the laser beam.

8. The laser machining system according to claim 1 ,
the machining head further including:
an optical sensor configured to detect at least reflected light coming from the workpiece and being reflected by the wavelength selective mirror.

9. The laser machining system according to claim 1 ,
the machining head further including:
a third image capturing device configured to acquire an image of a machining surface of the workpiece by reflecting light coming from the machining surface of the workpiece by the wavelength selective mirror.

10. The laser machining system according to claim 9 , wherein
the laser machining system is connected to a machine learning device, and
the machine learning device is configured to:
acquire, as input data, at least image data of the machining surface of the workpiece and time series data of laser machining condition data;
acquire, as a label, time series data of an evaluation value indicating quality in laser machining result corresponding to the input data;
prepare the input data and the label in a pair as teacher data;
learn relation between the input data and the label by supervised learning and build a learning model;
output, as a prediction value, prediction result indicating quality in laser machining result and change in the laser machining result caused by change in laser machining conditions, by using the learning model, with respect to new input data including at least image data of the machining surface of the workpiece and laser machining condition data; and
when th

Assignees

• Fanuc Corp

Inventors

• Takigawa Hiroshi
• Kurosawa Tadashi
• Machida Hisatada

Classifications

• G06N3/098

  Distributed learning, e.g. federated learning · CPC title

• G06N3/092

  Reinforcement learning · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G05B19/4065 Primary

  Monitoring tool breakage, life or condition · CPC title