Method and apparatus for detecting processing abnormalities

Abstract

The objective is to provide a machining anomaly detection method and a machining anomaly detection device that can detect anomalies with higher accuracy by taking into account the state of the operating axis. [Solution] A predetermined signal obtainable from a machine tool is used as the target for monitoring for machining anomaly detection. A diagnostic model is created that calculates the degree of anomaly related to machining based on the signal obtained from the machine tool. A quantitative value indicating the state of the operating axis based on the signal is used as the axis state quantity. The axis state quantity during normal machining is calculated from the signal obtained during normal machining, and the axis state quantity before machining is calculated from the signal obtained during the test run. Based on the axis state quantity during normal machining and the axis state quantity before machining, it is determined whether the signal obtained during the test run is included in the learning data. Based on the determination result of whether the signal obtained before machining is included in the learning data, the threshold for the degree of anomaly calculated by the diagnostic model from the signal obtained during machining is changed. The degree of anomaly and the threshold are compared to detect machining anomalies.

Claims

1. A machining anomaly detection method for detecting an anomaly in a machine tool having an operating axis that performs an action including relative motion between a tool used for machining and a workpiece, A predetermined signal obtainable from the aforementioned machine tool is used as the target for monitoring for machining abnormalities. A diagnostic model is created that uses a portion of the signals acquired during normal machining, when the machine tool can perform the machining normally, as training data, and calculates the degree of abnormality related to the machining from the signals based on the training data. A quantitative value indicating the state of the operating axis based on the aforementioned signal is defined as the axis state quantity. From the signals acquired during normal machining, the axis state quantity during normal machining is calculated. The pre-processing shaft state quantity is calculated from the signal obtained during the test run performed before the aforementioned processing, Based on the shaft state values ​​during normal machining and the shaft state values ​​before machining, it is determined whether or not the signals acquired during the test run are included in the learning data. Based on the determination result of whether or not the signal acquired during the test run is included in the learning data, the threshold for the degree of abnormality calculated by the diagnostic model from the signal acquired during processing is changed. A method for detecting processing abnormalities, characterized by detecting abnormalities in the processing by comparing the degree of abnormality with the threshold.

2. The machining abnormality detection method according to claim 1, characterized in that the amount of change in the threshold is calculated based on the difference between the axis state quantity during normal machining and the axis state quantity before machining.

3. The machining abnormality detection method according to claim 1 or 2, characterized in that the axis state quantity is the average value of the feed axis load obtained when the operating axis is operated in a predetermined pattern at a predetermined feed rate.

4. A machining anomaly detection device for detecting an anomaly in a machine tool having an operating axis that performs an action including relative motion between a tool used for machining and a workpiece, A signal acquisition unit capable of acquiring a predetermined signal obtainable from the aforementioned machine tool, A diagnostic model creation unit creates a diagnostic model that calculates the degree of abnormality related to the machining from the signals, using a portion of the signals acquired during normal machining when the machine tool can perform the machining normally as learning data, and based on the learning data, The quantitative value indicating the state of the operating axis is defined as the axis state quantity. A shaft state calculation unit calculates the shaft state quantity during normal machining from signals acquired during normal machining, and calculates the shaft state quantity before machining from signals acquired during a test run performed before the machining is carried out. A learning presence / absence determination unit determines whether the signal acquired during the test run is included in the learning data, based on the shaft state quantity during normal machining and the shaft state quantity before machining. Based on the determination result of whether the signal acquired during the test run is included in the learning data, the abnormality calculation unit uses the diagnostic model to change the threshold value for the abnormality calculated from the signal acquired during the processing, and the threshold setting unit A processing abnormality determination unit detects an abnormality in the processing by comparing the abnormality level calculated by the abnormality level calculation unit with the threshold value. A processing abnormality device characterized by being equipped with the following features.

5. The machining abnormality detection device according to claim 4, characterized in that the threshold setting unit calculates the amount of change to the threshold based on the difference between the axis state quantity during normal machining and the axis state quantity before machining.

6. The machining abnormality detection device according to claim 4 or 5, characterized in that the axis state quantity is the average value of the feed axis load obtained when the operating axis is operated in a predetermined pattern at a predetermined feed rate.

Images