Injection molding system for the calculation of optimum operating conditions and machine learning therefor

Patent metadata

Field	Value
Publication number	US-10416617-B2
Application number	US-201615221697-A
Country	US
Kind code	B2
Filing date	Jul 28, 2016
Priority date	Jul 28, 2015
Publication date	Sep 17, 2019
Grant date	Sep 17, 2019

Abstract

Disclosed is an injection molding system including: a state observation section observing, when injection molding is performed, physical-amounts relating to the injection molding that is being performed; a physical-amount data storage section storing the physical-amount data; a reward-conditions setting section setting reward conditions for machine learning; a reward calculation section calculating a reward based on the physical-amount data and the reward conditions; an operating-conditions adjustment learning section performing machine learning of adjusting operating conditions based on the reward calculated by the reward calculation section, the operating conditions, and the physical-amount data; a learning-result storage section storing a learning result of the machine learning by the operating-conditions adjustment learning section; and an operating-conditions adjustment-amount output section determining and outputting an operating condition to be adjusted and an adjustment amount based on the machine learning by the operating-conditions adjustment learning section.

First claim

The invention claimed is:

1. An injection molding system provided with at least one injection molding machine and having artificial intelligence that performs machine learning, the injection molding system comprising:
a state observation section that observes, when injection molding is performed by the injection molding machine, physical amounts relating to the injection molding that is being performed;
a physical-amount data storage section that stores the physical-amount data observed by the state observation section;
a reward-conditions setting section that sets reward conditions for the machine learning;
a reward calculation section that calculates a reward based on the physical-amount data observed by the state observation section and the reward conditions set by the reward-conditions setting section;
an operating-conditions adjustment learning section that performs machine learning of adjusting operating conditions based on the reward calculated by the reward calculation section and operating conditions set in the injection molding system and the physical-amount data;
a learning-result storage section that stores a learning result of the machine learning by the operating-conditions adjustment learning section; and
an operating-conditions adjustment-amount output section that determines and outputs an operating condition to be adjusted and an adjustment amount on the basis of the learning result by the operating-conditions adjustment learning section.

2. The injection molding system according to claim 1 ,
wherein the learning result stored in the learning-result storage section is used in the learning of the operating-conditions adjustment learning section.

3. The injection molding system according to claim 1 , further comprising:
a measurement section, wherein the physical-amount data observed by the state observation section includes at least one of a weight and a size of a molded article measured by the measurement section, an appearance, a length, an angle, an area, and a volume calculated from image data on the molded article, an optical examination result of an optically molded article, and a measurement result of a strength of the molded article, and the physical-amount data storage section also stores other physical-amount data as the physical-amount data on the molded article.

4. The injection molding system according to claim 1 ,
wherein input of the reward conditions to the reward-conditions setting section can be implemented by a display device provided in the injection molding machine.

5. The injection molding system according to claim 1 ,
wherein, when at least one of stabilization of physical-amount data, reduction in a cycle time, and energy saving is attained, the reward calculation section gives a positive reward according to an extent of the attainment.

6. The injection molding system according to claim 1 ,
wherein, when at least one event from among destabilization of physical-amount data, an extension of a cycle time, and an increase in consumption energy occurs, the reward calculation section gives a negative reward according to an extent of the event.

7. The injection molding system according to claim 1 ,
wherein an allowable value is set in advance in the physical-amount data, and the reward calculation section gives a positive reward when the physical-amount data falls within the allowable value.

8. The injection molding system according to claim 1 ,
wherein an allowable value is set in advance in the physical-amount data, and the reward calculation section gives, when the physical-amount data deviates from the allowable value, a negative reward based on an amount of the deviation.

9. The injection molding system according to claim 1 ,
wherein a target value is set in advance in the physical-amount data, and the reward calculation section gives, when the physical-amount data comes close to the target value, a positive reward based on a deviation amount between the target value and the physical-amount data.

10. The injection molding system according to claim 1 ,
wherein a target value is set in advance in the physical-amount data, and the reward calculation section gives, when the physical-amount data deviates from the target value, a negative reward based on a deviation amount between the target value and the physical-amount data.

11. The injection molding system according to claim 1 ,
wherein, when a state showing a molding failure occurs, the reward calculation section gives a negative reward according to an extent of the molding failure.

12. The injection molding system according to claim 11 ,
wherein the molding failure includes at least one of a burr, a sink mark, a warp, an air bubble, a short shot, a flow mark, a weld line, a silver streak, a color irregularity, discoloration, carbonization, intrusion of impurities, deviation of an optical axis of a lens molded article from an allowable value, and a failure in a thickness of a molded article.

13. The injection molding system according to claim 1 ,
wherein the operating conditions subjected to the machine learning by the operating-conditions adjustment learning section include at least one of mold clamping conditions, ejector conditions, injection dwell conditions, metering conditions, temperature conditions, nozzle touch conditions, resin supply conditions, mold-thickness conditions, molded-article extraction conditions, and hot-runner conditions.

14. The injection molding system according to claim 13 , further comprising:
a robot serving as a molded-article extraction unit in which the molded-article extraction conditions are set.

15. The injection molding system according to claim 1 ,
wherein at least one of the operating conditions is fluctuated within a prescribed range to be learned by the operating-conditions adjustment learning section.

16. The injection molding system according to claim 1 ,
wherein each of a plurality of injection molding systems has a communication section to communicate with an outside, and physical-amount data stored in each of physical-amount data storage sections and a learning result stored in each of learning-result storage sections are sent/received to be shared.

Assignees

• Fanuc Corp

Inventors

• Shiraishi Wataru
• Yamamoto Kazuhiro

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G05B2219/45244

  Injection molding · CPC title

• B29C2945/76949

  using a learning system, i.e. the system accumulates experience from previous occurrences, e.g. adaptive control · CPC title

• B29C2945/7613

  Weight · CPC title

• B29C45/766 Primary

  the setting or resetting of moulding conditions, e.g. before starting a cycle · CPC title