Machine learning device, machine learning method, and fault prediction device and system

Abstract

The invention relates to a machine learning device, a machine learning method, and a fault prediction device and system. The machine learning device (5) which learns fault prediction of one of a main shaft (212) of a machine tool and a motor (214) driving the main shaft includes: a state observation unit (52) observing a state variable including at least one of data output from a motor controller (3) controlling the motor, data output from a detector (11) detecting a state of the motor, and data output from a measuring device (12) measuring a state of the one of the main shaft and the motor; a determination data obtaining unit (51) obtaining determination data upon determining one of whether a fault has occurred in the one of the main shaft and the motor and a degree of fault; and a learning unit (53) learning the fault prediction of the one of the main shaft and the motor in accordance with a data set generated based on a combination of the state variable and the determination data.

Description

technical field [0001] The present invention relates to a machine learning device and a machine learning method for learning failure prediction of a spindle or a motor driving the spindle, and a failure prediction device and a failure prediction system equipped with the machine learning device. Background technique [0002] Most of the failures of the spindle of the machine tool or the spindle motor (motor) driving the spindle are caused by the deterioration or damage of the bearing of the spindle or the motor. Here, if the machine tool is used in a state where the main shaft is completely broken down, for example, the machining accuracy of the workpiece is lowered, resulting in defective products. In addition, if it takes time for the spindle to recover, it will cause a large interruption time (stop time) of the machine tool, resulting in a decrease in the working efficiency of the machine tool. [0003] Therefore, if the symptom (deterioration) is detected before the spin...

AI Technical Summary

Problems solved by technology

[0009] For example, in order to detect the vibration of the main shaft of a machine tool, it is necessary to install a vibration measuring device on the motor or the main shaft. However, if the vibration is measured manually, the operator must hold the vibration measuring device and periodically measure the vibration of each part of the machine tool. Will create a greater burden on the operator

On the other hand, although it is also conceivable to automatically measure the vibration of the main shaft with a vibration sensor, etc., in this case, the vibration sensor needs to be always installed on the main shaft, which will increase the cost of the machine tool.

[0010] In addition, for example, in the case of detecting the abnormality by listening to the abnormal sound of the main shaft, the abnormal sound of each machine tool is checked regularly through human hearing, which still causes a large workload for the operator.

On the other hand, it is also conceivable to use a microphone to automatically detect abnormal sounds, but generally speaking, since the factory where the machine tool is installed is noisy, it is not easy to detect abnormal sounds properly, and it will also cost a lot of money to install the microphone. cost

[0011] In addition, with the complexity and sophistication of machine tools in recent years, the main causes of failures have also become more complex, making it difficult to predict the failure of the spindle of the machine tool or the motor that drives the spindle.

That is, in the failure prediction method performed according to a certain standard, sometimes it cannot be applied to the actual situation or lacks accuracy, so a technology that can perform accurate failure prediction according to the situation is sought

Images