Control device and machine learning device

1. A control device configured to detect a collision of a spindle driven by a servo motor and included in a machine tool, the control device comprising:

a processor configured to:

learn a state of the spindle during normal machining without the collision of the spindle, observe spindle estimated load torque data indicating estimated load torque values for the spindle and spindle acceleration data indicating acceleration values of the spindle as state variables representing a current state of an environment,

learn a correlation between the estimated load torque values for the spindle and the acceleration values of the spindle during the normal machining with use of the state variables,

perform a detection of the collision of the spindle based on an existence of a number of pairs of the estimated load torque values and the acceleration values of the spindle in a vicinity of a sample of the state variables using a graph in which a vertical axis represents the estimated load torque values of the spindle and a horizontal axis represents the acceleration values of the spindle,

output a result of the detection, and

generate an alarm or control the servo motor for driving the spindle included in the machine tool based on the result of the detection.

2. The control device according to claim 1 , wherein

the state variables further include machining condition data that indicates conditions of machining in the machine tool, and

the processor is further configured to learn a correlation among the estimated load torques for the spindle and the acceleration values of the spindle during the normal machining and the conditions of the machining.

3. The control device according to claim 1 , wherein the state variables further include spindle generated sound data indicating information on sounds generated from the spindle, and

the processor is further configured to learn a correlation among the estimated load torque values for the spindle and the acceleration values of the spindle during the normal machining and the information on the sounds generated from the spindle.

4. The control device according to claim 1 , wherein

the processor is further configured to update a correlation model for determination on a state of normality/abnormality of machining by the machine tool, based on the state variables.

5. The control device according to any one of claim 1 , wherein

the processor is further configured to learn the state of the spindle during the normal machining without the collision of the spindle, through unsupervised learning.

6. The control device according to claim 4 , wherein the processor is further configured to

output a result of the determination on the state of the normality/abnormality of the machining by the machine tool, based on results of the learned correlation.

7. The control device according to claim 1 , wherein the processor exists in a cloud server.

8. A machine learning device configured to learn a state of a spindle included in a machine tool during normal machining without a collision of the spindle, the machine learning device comprising:

a processor configured to:

observe spindle estimated load torque data indicating estimated load torque values for the spindle and spindle acceleration data indicating acceleration values of the spindle, as state variables representing a current state of an environment,

learn a correlation between the estimated load torque values for the spindle and the acceleration values of the spindle during the normal machining, with use of the state variables,

perform a detection of the collision of the spindle based on an existence of a number of pair of the estimated load torque value and the estimated spindle acceleration in a vicinity of a sample of the state variables using a graph in which a vertical axis represents the estimated load torque value of the spindle and a horizontal axis represents the acceleration value of the spindle,

output a result of the detection, and

generate an alarm or control the servo motor for driving the spindle included in the machine tool based on the result of the detection.

9. The machine learning device according to claim 8 , wherein

the state variables further include machining condition data that indicates conditions of machining in the machine tool, and

the processor is further configured to learn a correlation among the estimated load torques for the spindle and the acceleration values of the spindle during the normal machining and the conditions of the machining.

10. The machine learning device according to claim 8 , wherein the state variables further include spindle generated sound data indicating information on sounds generated from the spindle, and

the processor is further configured to learn a correlation among the estimated load torque values for the spindle and the acceleration values of the spindle during the normal machining and the information on the sounds generated from the spindle.

11. The machine learning device according to claim 8 , wherein

the processor is further configured to update a correlation model for determination on a state of normality/abnormality of machining by the machine tool, based on the state variables.

12. The machine learning device according to any one of claim 8 , wherein

the processor is further configured to learn the state of the spindle during the normal machining without the collision of the spindle, through unsupervised learning.

13. The machine learning device according to claim 11 , wherein the processor is further configured to

output a result of the determination on the state of the normality/abnormality of the machining by the machine tool, based on results of the learned correlation.

14. The control device according to claim 8 , wherein the processor exists in a cloud server.