Acoustic detection of tool break events in machine tool operations

• US 4,636,779 A
• Filed: 10/24/1984
• Issued: 01/13/1987
• Est. Priority Date: 10/24/1984
• Status: Expired due to Term

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Yerazunis, William S., Erkes, Joseph W., Thomas, Charles E., Lee, Minyoung
• Primary Examiner(s): Rowland, James L.
• Assistant Examiner(s): Hofsass, Jeffery A.
• Application Number:
  US06/664,189
• Time in Patent Office: 811 Days
• Field of Search: 340/680, 340/683, 340/679, 364/475, 73/104, 73/660, 73/593, 73/658, 364/507, 364/508, 364/550, 364/551
• US Class Current: 340/680
• CPC Class Codes: B23Q 17/0957 Detection of tool breakage detecting failure of drills B23B49/001
• CPC Class Codes: B23Q 17/0971 by measuring mechanical vibrations of parts of the machine arrangements for measuring vibrations B23Q17/12
• CPC Class Codes: B23Q 17/098 by measuring noise
• CPC Class Codes: G05B 19/4065 Monitoring tool breakage, life or condition
• CPC Class Codes: G05B 2219/37351 Detect vibration, ultrasound

Abstract

A system and method for monitoring vibrations of a cutting tool produced by tool break events, and for interpreting them to detect tool breaks of sufficient magnitude to endanger the machined part. The signal generated by a sensor such as an accelerometer is preprocessed to attenuate low frequency machining noise and detect the energy in a higher frequency band, then sampled, and the digitized signal samples analyzed by tool break detection logic. This logic is triggered by a positive-going signal transient, and prevents false alarms on minor tool break events that do not mar the workpiece and on noise from other sources.

Claims

1. A machine tool monitor for detecting cutting tool break events while machining a workpiece comprising:

a broadband vibration sensor which generates an electrical signal representing vibrations at the tool-workpiece interface;

analog signal processing means including a high-pass filter to attenuate lower frequency machinery noise, and a full wave energy detector to rectify and low-pass filter said signal;

means for sampling the unipolar output signal of said analog processing means and converting each sample to digital form; and

digital circuitry comprising means for comparing every sample with the running mean signal level of a given number of previous samples to detect the presence of a positive-going signal transient that may indicate a tool break event, means for comparing the mean signal level after and before said transient to detect a shift in mean level and a substantial change in background cutting noise, and means for generating a tool break alarm when the shift in mean level persists for a preselected period.

2. The monitor of claim 1 wherein said vibration sensor is an accelerometer most sensitive to frequencies around a resonant frequency which is above 30 KHz.

3. The monitor of claim 2 wherein the combination of said accelerometer and high-pass filter produces a band-pass filtering that favors frequencies in a band in the vicinity of said resonant frequency.

4. The monitor of claim 1 wherein the low-pass filter has a cutoff frequency less than 500 Hz to provide anti-aliasing filtering at the sampling rate of said sampling means.

5. A system for detecting cutting tool break events while machining a workpiece comprising:

an accelerometer which is most sensitive to frequencies around a resonant frequency and is positioned on a machine tool to sense vibrations at the tool-workpiece interface and convert these and other vibrations to an electrical signal;

an analog preprocessor including means for high-pass filtering said vibration signal to discriminate against lower frequency machinery noise, and means for rectifying and low-pass filtering said signal to detect the energy in a band around said resonant frequency;

means for sampling the unipolar output signal of said analog preprocessor and converting each sample to digital form;

digital circuitry to detect major tool break events capable of marring the workpiece and prevent false alarms on minor tool break events and spurious noise;

said digital circuitry comprising means for calculating the running mean signal level of a selected number of samples, a transient detector to compare each new sample with the running mean and detect a transient increase in signal level that may have its source in a major tool break event, a mean shift detector to compare the mean signal level after and before said transient to detect a shift in background cutting noise, and means for checking that said mean shift persists for a given period and hence generating a tool break alarm.

6. The system of claim 5 wherein said mean shift detector returns control to said transient detector to search for another transient after failure to detect said mean shift.

7. The system of claim 5 wherein said last-mentioned means returns control to said transient detector to search for another transient after failure to meet a mean shift persistence check criterion.

8. The system of claim 5 wherein the resonant frequency of said accelerometer is above 30 KHz.

9. The system of claim 8 wherein said low-pass filtering means has a cutoff frequency of less than 500 Hz to provide anti-aliasing filtering at the signal sampling rate of said sampling means.

10. The method of detecting tool breakage capable of marring the workpiece comprising:

sensing the vibrations of a cutting tool insert during a machining operation and converting these and other machine tool vibrations to an electrical signal;

preprocessing said vibration signal by high-pass filtering to discriminate against lower frequency machinery noise, and rectifying and low-pass filtering to detect the signal energy, the output signal containing a positive-going transient that may indicate a major tool break event;

sampling said output signal and converting each sample to digital form;

detecting said transient by comparing every sample with the running mean signal level of N previous samples, comparing the mean signal level after and before said transient to detect a mean shift and dismissing any transient not followed by a substantial change in background cutting noise, checking that said mean shift persists for a given period and dismissing any transient not meeting this second criterion, and generating a tool break alarm when both criteria are satisfied.