General Electric Company

General Elec. Co.

Charles E Thomas

Minyoung Lee

Steven Robert Hayashi

James F Bedard

Douglas Glenn Wildes

William S Yerazunis

Robert J Wojnarowski

Lawson P Harris

Charles W Eichelberger

Vijay C Jaswa

Henry A Rocha

Joseph W Erkes

A device for determining cutting tool wear and breakage has an accelerometer for sensing tool vibrations. Functions of the AC and DC power of the vibration signal, such as the powers themselves or the square root of the powers, are computed and compared. If the relationship between the AC and DC power changes beyond selected limits, an alarm is sounded or flashed. A method for determining cutting tool wear and breakage provides a tool vibration signal, computes functions of the AC and DC power of the vibration signal, and compares the AC and DC power signals.

Apparatus and method for tool break and wear detection

A system and method for monitoring machine tool operations provides a multi-level tool break alarm and uses information from multiple sensors of different types. Signal processing and pattern recognition logic techniques are applied to a cutting process high frequency vibration signal to detect major tool breaks requiring prompt stoppage of the cut. False alarm resistant detection of minor tool breaks, for which the response may not be an immediate stop of the cutting process, is obtained with information from a low frequency vibration sensor, an axis drive current sensor, or an axis velocity sensor. A minor tool break alarm is generated when signal transients in both a high frequency and low frequency channel signal are in close time coincidence, and a major alarm when the high frequency channel transient is followed by a persisting mean vibration level shift.

Multi-level tool break detection using multi-mode sensing

Gradual wear or breakage of a machine tool which occurs over a period of time is detected by sensing high frequency vibrations produced at the cutting tool/workpiece interface during a machining process. The high frequency vibrations are converted to a unipolar vibration signal which is processed to produce a tracking signal which tracks the minimum value of the vibration signal. The vibration signal is related to the effective cutting energy and it decreases due to gradual tool wear or breakage. The tracking minimum signal is compared to a predetermined threshold level to detect excessive wear or breakage of the tool, and an alarm is provided.

Tool break/wear detection using a tracking minimum of detected

Vibrations at the groove cutting tool-workpiece interface are sensed by a vibration sensor such as, for example, an accelerometer, and the vibration sensor output is processed to yield an electrical signal indicative of groove cutting tool cutting noise vibration. Digital analysis of the electrical signal after it is sampled is performed to detect various types of groove cutting tool break event signatures, and especially a type of signature in which a high amplitude positive going spike is followed by a significant shift in the mean amplitude signal level of the background cutting vibration. A machine tool monitor, system and method are disclsoed which detects such changes and interprets them to detect groove tool break events of sufficient magnitude to endanger the machined part.

Groove cutting tool break event detecting method and system

Substantial cutting condition changes which occur in intermittent bursts as rough surfaced workpieces are turned generate specific vibration signal signatures known as the runout condition. Either an upper or a lower threshold is set and all mean signal samples are compared with the threshold value. When the mean cutting noise exceeds the upper threshold or stays below the lower threshold for a preset number of signal samples, a tool break alarm is generated. Techniques are given to reduce false alarms on runout during initial rough surface cuts. The system comprises an accelerometer or other sensor whose signal is processed to attenuate lower frequency machinery noise and to detect the signal energy in a band below 100 kHz, then sampled, and the digitized signal samples analyzed by pattern recognition logic.

Tool break detecting method and system preventing false alarms due to

Sharp tools have different cutting characteristics from dull or worn tools. Among these differences is that a wear land develops on the cutting tool so that more of the cutting tool comes into contact with the workpiece during the cutting process. The increased contact area between the tool and workpiece forces more energy to be consumed by the cutting machine in making a cut because more energy is expended in non-productive work. Indications of an increase in non-productive work are the increased power or force necessary to operate a spindle in lathes, milling machines, etc., and the increased energy in cutting vibrations in a low frequency range emitted during the cutting process. Another indication of decreased efficiency of the cutting process is the decreased energy in cutting vibrations in a high frequency range emitted during the cutting process. A method and apparatus are described for continuously monitoring a ratio of spindle force or power or low frequency vibration energy to high frequency vibration energy during the cutting process, and generating an output warning signal when the radio reaches a preselected level correlated with excessive tool wear.

Cutting tool wear detection apparatus and method

A tool break detection system has an automatic gain control to adjust the analog signal channel gain and hold the cutting vibration signal at a desired average level. The AGC time constant is long enough that the detection of abrupt tool breaks and sudden large signal level transistions by the digital signature recognition logic is unaffected. However, the gradual decrease in signal level produced by a crumbly-type break would be removed by AGC action. In a system with a hardware AGC, crossing a high gain threshold resets gain to a low value and the resulting abrupt and persisting change in signal level is detected by the abrupt tool break logic. Another embodiment uses the gain command output by a software AGC and generates a break detected signal directly, without resetting the gain command, as gain rises above a high gain alarm level which is recalculated at the start of each cut.

Detection by automatic gain control features of gradual cutting tool

The initial contact of a rotating milling tool to a workpiece is detected in the presence of spurious noise spikes by digitally processing samples of a preprocessed vibration signal concurrently through basic tool touch detection logic and milling tool touch detection enhancement logic, and generating a touch indication if either logic set is satisfied. The basic logic detects a continuously increasing signal level higher than a threshold set above background noise. The milling logic detects an above-threshold signal sample and sets up an acceptance window about the expected time of contact of the next cutter. Tool touch is declared when an above-threshold signal occurs during the window; the milling logic is reset by a noise spike in the interval before the window starts. The basic logic detects tool touch in a dense spiky noise situation.

Acoustic detection of milling tool touch to a workpiece

A Machine Tool Monitor detects significant cutting tool breakage and the first contact of an advancing tool to a workpiece, and does this by monitoring vibration signals produced by the machining of parts and interpreting patterns in these signals. Information from the part program improves the performance of the detector and optimizes it for the cutting conditions called for by the machine tool control. The analog channel gain of the monitor is adjusted and parameters controlling the digital pattern recognition logic are selected using part program information on machining parameters. The tool touch or tool break detection mode is selected by the part program.

Vibration-sensing tool break and touch detector optimized for machining

Vibrations at the cutting tool-workpiece interface are sensed and the accelerometer output is preprocessed to yield a cutting noise vibration signal. Digital analysis of the sampled signal is performed to detect a gradual cutting noise signal level decrease tool break signature, either continuously decreasing or decreasing in a series of small abrupt steps. A tool break alarm is generated, without false alarming on metal-to-air tool path transitions, as the mean cutting noise signal level falls below an upper check limit and, after a preset time has elapsed, crosses a lower check limit.

Detection of tool breaks that cause slowly decreasing cutting noise

In some machining conditions the background noise in an acoustic cutting tool break detection system is a low mean level with fairly dense high amplitude noise spikes. A common tool break vibration signature is the sudden appearance of a dense spiky noise. Digital signal pattern recognition logic uses an up/down counter to reject the noise on the basis of its lower spike density while alarming on such a tool break signature. Preprocessed vibration signal samples are tested against a detection threshold and an alarm generated if those whose amplitude is above the threshold amplitude exceeds those below by a preset count.

Detector to discriminate between tool break acoustic signatures and

A Machine Tool Monitor detects the sudden increase in vibration signal level when a slowly advancing cutting tool first touches the workpiece. The delay to a tool touch decision is reduced while still avoiding false alarms on high amplitude single-peaked and multi-peaked noise spikes. Samples of the preprocessed analog vibration signal that are above a low threshold level are analyzed by digital tool touch detection logic. The signal slope polarity is determined by comparison of a new sample to the tracking mean, or by sample-to-sample amplitude comparisons, and presented to an up/down counter. A tool touch alarm is generated when the positive slope counts exceed the negative slope counts by a preset number slightly greater than occur in a time period from the start of a noise spike until its peak amplitude is passed.

Acoustic tool touch detector with minimized detection delay

Substantial cutting condition changes that occur gradually, as opposed to the more usual sudden large change, are detected by setting upper and lower cutting noise mean level thresholds. When the mean cutting noise exceeds the upper threshold or stays below the lower threshold for a preset number of signal samples, a tool break alarm is generated. Techniques are given to reduce false alarms at the start and end of the cut and on runout on initial rough surface cuts. The system comprises an accelerometer or other sensor whose signal is preprocessed to attenuate lower frequency machinery noise and detect the signal energy in a band below 100 KHz, then sampled, and the digitized signal samples analyzed by pattern recognition logic.

Acoustic tool break detection system and method

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.

Acoustic detection of tool break events in machine tool operations

This tool break detection system relies on monitoring changes in the cutting noise itself, rather than detecting the tool fracture acoustic signal. A broken tool capable of damaging the workpiece is detected, and tool break events that do not affect cutting conditions are ignored. The signal from a sensor such as an accelerometer is preprocessed to attenuate low frequency machinery noise and detect the signal energy in a band below 100 KHz, then sampled, and the digitized signal samples analyzed by pattern recognition logic. Runout false alarms during rough surface cutting are prevented; after detection of an abrupt increase or decrease in signal level, the confirmation period to test for a persistent shift in mean level is set longer than the workpiece revolution period.

Growing community of inventors

Scotia, NY, United States of America

Charles E Thomas