Polishing tool wear amount prediction device, machine learning device, and system

Patent metadata

Field	Value
Publication number	US-11822308-B2
Application number	US-201916268043-A
Country	US
Kind code	B2
Filing date	Feb 5, 2019
Priority date	Feb 6, 2018
Publication date	Nov 21, 2023
Grant date	Nov 21, 2023

Abstract

A polishing tool wear amount prediction device, machine learning device, and system capable of predicting a wear amount of a polishing tool unit of a polishing tool during polishing are provided. The polishing tool wear amount prediction device includes a machine learning device which observes polishing condition data indicating a processing condition of polishing as a state variable indicating a current environment state and performs, based on the state variable, learning or prediction by using a learning model which stores a correlation of the wear amount of the polishing tool with respect to the processing condition of polishing.

First claim

The invention claimed is:

1. A polishing tool wear amount prediction device for predicting a wear amount of a polishing tool in polishing on a work by the polishing tool, the polishing tool wear amount prediction device comprising:
a processor configured to
observe polishing condition data indicating a processing condition of the polishing as a state variable indicating a current environment state,
perform, based on the state variable, learning by using a learning model in which the wear amount of the polishing tool is modeled with respect to the processing condition of the polishing, and
output, at each unit time during a polishing operation on the work by the polishing tool, a signal for controlling the polishing tool on the basis of a result of the learning, wherein the processor is configured to
predict the wear amount of the polishing tool based on the learning model learned by the processor, and
output the signal to a robot for correcting a position of each axis of the robot which holds the polishing tool during the polishing operation and a rotational speed of the polishing tool,

wherein the processor is further configured to
acquire wear amount data indicating the wear amount of the polishing tool due to the polishing as label data, and
generate the learning model learned with the processing condition of the polishing and the wear amount of the polishing tool in association with each other by using the state variable and the label data, and

wherein, as the state variable indicating the current environment state, the processor is further configured to observe, in addition to the polishing condition data,
polishing temperature data indicating a temperature of the polishing tool in the polishing, and
at least any of
polishing sound data indicating a sound occurring in the polishing,
tool type data indicating a type of the polishing tool, and
processing environment data indicating an environment in the polishing, and

wherein the temperature data are acquired by estimating based on a combination of (1) the tool type data and (2) how long the polishing tool has been continuously used for the polishing.

2. The polishing tool wear amount prediction device according to claim 1 , wherein
the processor is further configured to
calculate an error between a correlation model as the learning model for estimating the wear amount of the polishing tool from the state variable and a correlation feature identified from teacher data prepared in advance, and
update the correlation model so as to reduce the error.

3. The polishing tool wear amount prediction device according to claim 1 , wherein
the processor is further configured to compute the state variable and the label data in a multilayered structure.

4. The polishing tool wear amount prediction device according to claim 1 , wherein
the processor is configured to output an estimation result of the wear amount of the polishing tool based on the observed state variable and the learning model which is learned with the processing condition of the polishing and the wear amount of the polishing tool in association with each other.

5. The polishing tool wear amount prediction device according to claim 1 , wherein
the processor resides in a cloud server.

6. A machine learning device, comprising at least one processor configured to
observe polishing condition data indicating a processing condition of polishing on a work by a polishing tool as a state variable indicating a current environment state,
perform, based on the state variable, learning by using a learning model in which a wear amount of the polishing tool is modeled with respect to the processing condition of the polishing, and
generate, at each unit time during a polishing operation on the work by the polishing tool, a signal for controlling the polishing tool on the basis of a result of the performed learning,
wherein the at least one processor is configured to
predict the wear amount of the polishing tool based on the learning model learned by the processor, and
output the signal to a robot for correcting a position of each axis of the robot which holds the polishing tool during the polishing operation and a rotational speed of the polishing tool,

wherein the processor is further configured to
acquire wear amount data indicating the wear amount of the polishing tool due to the polishing as label data, and
generate the learning model learned with the processing condition of the polishing and the wear amount of the polishing tool in association with each other by using the state variable and the label data, and

wherein, as the state variable indicating the current environment state, the processor is further configured to observe, in addition to the polishing condition data,
polishing temperature data indicating a temperature of the polishing tool in the polishing, and
at least any of
polishing sound data indicating a sound occurring in the polishing,
tool type data indicating a type of the polishing tool, and
processing environment data indicating an environment in the polishing, and

wherein the temperature data are acquired by estimating based on a combination of (1) the tool type data and (2) how long the polishing tool has been continuously used for the polishing.

7. The machine learning device according to claim 6 , wherein the at least one processor is further configured to:
output an estimation result of the wear amount of the polishing tool based on
the observed state variable, and
the learning model learned with the processing condition of the polishing and the wear amount of the polishing tool in association with each other.

8. A system, comprising a plurality of devices mutually connected via a network, wherein
the plurality of devices include a first robot provided with a processor for predicting a wear amount of a polishing tool in polishing on a work by the first robot using the polishing tool,
wherein the processor is configured to
observe polishing condition data indicating a processing condition of the polishing as a state variable indicating a current environment state,
perform, based on the state variable, learning by using a learning model in which the wear amount of the polishing tool is modeled with respect to the processing condition of the polishing, and
output, at each unit time during a polishing operation on the work by the polishing tool, a signal for controlling the polishing tool on the basis of a result of the learning,

wherein the processor is configured to
predict the wear amount of the polishing tool based on the learning model learned by the processor, and
output the signal to a robot for correcting a position of each axis of the robot which holds the polishing tool during the polishing operation and a rotational speed of the polishing tool,

wherein the processor is further configured to
acquire wear amount data indicating the wear amount of the polishing tool due to the polishing as label data, and
generate the learning model learned with the processing condition of the polishing and the wear amount of the polishing tool in association with each other by using the state variable and the label data, and

wherein, as the state variable indicating the current environment state, the processor is further configured to observe, in addition to the polishing condition data,
polishing temperature data indicating a temperature of the polishing tool in the polishing, and
at least any of
polishing sound data indicating a sound occurring in the polishing,
tool type data indicating a type of the polishing tool, and
processing environment data indicating an environment in the polishing, and

wherein the temperature data are acquired by estimati

Assignees

• Fanuc Corp

Inventors

• Zhang He

Classifications

• G05B19/404 Primary

  characterised by control arrangements for compensation, e.g. for backlash, overshoot, tool offset, tool wear, temperature, machine construction errors, load, inertia (G05B19/19, G05B19/41 take precedence) · CPC title

• B24B49/003 Primary

  involving acoustic means · CPC title

• B24B49/14

  taking regard of the temperature during grinding · CPC title

• G06N3/02

  Neural networks · CPC title

• G06N20/00

  Machine learning · CPC title