Machine learning apparatus and method for learning predicted life of motor, life prediction apparatus and motor system including machine learning apparatus

Patent metadata

Field	Value
Publication number	US-9811057-B2
Application number	US-201615247984-A
Country	US
Kind code	B2
Filing date	Aug 26, 2016
Priority date	Aug 28, 2015
Publication date	Nov 7, 2017
Grant date	Nov 7, 2017

Abstract

A machine learning apparatus that learns a condition associated with a predicted life of a motor includes: a state observation unit that observes a state variable composed from at least one of output data of a sensor that detects an operation state of the motor and data relating to presence or absence of a failure in the motor; an actual life data acquisition unit that acquires data relating to an actual life of the motor; and a learning unit that learns the condition associated with the predicted life of the motor in accordance with a training data set created based on a combination of the state variable and the actual life.

First claim

What is claimed is:

1. A machine learning apparatus that learns a condition associated with a predicted life of a motor, the machine learning apparatus comprising:
a state observation unit that observes a state variable composed from at least one of output data of a sensor that detects an operation state of the motor and data relating to presence or absence of a failure in the motor;
an actual life data acquisition unit that acquires data relating to an actual life of the motor; and
a learning unit that learns the condition associated with the predicted life of the motor in accordance with a training data set created based on a combination of the state variable and data relating to the actual life.

2. The machine learning apparatus according to claim 1 , wherein the operation state comprises at least one of a current command that commands a current flowing through the motor, a voltage command that commands a voltage applied to the motor, a frequency command that commands a frequency of the current or voltage flowing through the motor, a torque outputted from the motor, a number of rotations of the motor, an operating time of the motor, a temperature in a vicinity of the motor, humidity in a vicinity of the motor, and vibrations generated in the motor.

3. The machine learning apparatus according to claim 1 , wherein the learning unit comprises:
a reward calculation unit that calculates a reward based on the predicted life and the actual life; and
a function update unit that updates a function for calculating the predicted life of the motor based on the state variable and the reward.

4. The machine learning apparatus according to claim 3 , wherein the reward calculation unit increases the reward when a ratio of a difference between the predicted life and the actual life to the actual life falls inside a specified range and reduces the reward when the ratio falls outside the specified range.

5. The machine learning apparatus according to claim 1 , wherein the learning machine is configured to learn the condition in accordance with the training data set acquired with respect to a plurality of the motors.

6. A life prediction apparatus for the motor comprising the machine learning apparatus according to claim 1 , further comprising:
a decision-making unit that calculates the predicted life of the motor based on a result of learning by the learning unit in accordance with the training data set and in response to input of the current state variable.

7. The life prediction apparatus according to claim 6 , further comprising a notification unit that notifies an operator of the predicted life calculated by the decision-making unit.

8. The life prediction apparatus according to claim 6 , further comprising a notification unit that notifies an operator of information that prompts replacement of the motor based on the predicted life calculated by the decision-making unit.

9. The life prediction apparatus according to claim 6 , further comprising a change command output unit that based on the predicted life calculated by the decision-making unit, outputs to a control device that controls the motor a change command that changes at least one of the current command that commands the current flowing through the motor, the voltage command that commands the voltage applied to the motor, and the frequency command that commands the frequency of the current or voltage flowing through the motor.

10. The life prediction apparatus according to claim 6 , wherein the learning unit is configured to relearn and update the condition in accordance with additional training data set defined by the current state variable.

11. A motor system, comprising:
the life prediction apparatus according to claim 6 ;
a motor;
a control device that controls the motor; and
a sensor that detects an operation state of the motor.

12. A machine learning method that learns a condition associated with a predicted life of a motor, comprising:
a state observation step that observes a state variable composed from at least one of output data of a sensor that detects an operation state of the motor and data relating to presence or absence of a failure in the motor;
an actual life data acquisition step that acquires data relating to an actual life of the motor; and
a learning step that learns the condition associated with the predicted life of the motor in accordance with a training data set created based on a combination of the state variable and data relating to the actual life.

Assignees

• Fanuc Corp

Inventors

• Zhi Chao

Classifications

• G05B2219/37209

  Estimate life of gear, drive · CPC title

• H02P23/12

  Observer control, e.g. using Luenberger observers or Kalman filters · CPC title

• G05B13/0265 Primary

  the criterion being a learning criterion · CPC title

• G01R31/343

  in operation · CPC title

• G06N99/005

  Physics · mapped topic