JPH057140B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of the Invention] The present invention is a method for monitoring tool breakage and abnormalities in machine tools using acoustic emission (hereinafter referred to as AE) that occurs during cutting and tool breakage.
This invention relates to a tool breakage detection device that automatically detects tool breakage.

[Summary of the invention]

The tool breakage detection device according to the present invention was developed based on the knowledge that it is rare for a tool without wear to break suddenly, and that in many cases, a tool with advanced wear breaks. A.E.
To facilitate signal processing, the AE signal is pre-processed in advance to detect wear that is a sign of tool breakage. By transmitting the AE sensor signal to the signal processing section only when the tool wear exceeds a predetermined level, the processing of the signal processing section is made easier, the load is reduced, and the reliability of breakage detection is improved. This is what I did.

[Prior art and its problems]

When a machine tool uses a tool to cut an object to be machined (hereinafter referred to as a workpiece), the tool may break or become clogged with chips for some reason, resulting in abnormal cutting. With the recent progress in factory automation, there is a strong demand for automatic detection of tool breakage and abnormal cutting. As a method for detecting tool breakage in machine tools,
Conventionally, devices have been proposed that install an AE sensor near the tool or workpiece of a machine tool and detect tool breakage based on the AE signal obtained from the sensor.

However, in conventional tool breakage detection devices, the AE signal obtained from the AE sensor is sent to the signal processing section, regardless of the wear state of the tool, and the signal processing section detects tool breakage based on the average value of the amplitude or a specific frequency. was. However, the signal obtained from the AE sensor contains the signal generated when the tool breaks.
Various other signals are included in the AE signal, including AE signals obtained from other causes, such as signals generated from chips on the workpiece, electrical noise associated with the opening and closing of a solenoid, and shock when an object comes into contact with the workpiece. There was a problem in that the reliability of breakage detection was extremely low because it could not be sufficiently separated from sound and the like.

[Purpose of the invention]

The present invention has been made to solve the problems of the conventional tool breakage detection device, and includes:
Considering that tool breakage rarely occurs suddenly, signs of tool breakage can be obtained from tool wear.
It is an object of the present invention to provide a tool breakage detection device that can reduce the burden on an AE signal processing unit and improve the reliability of tool breakage detection.

[Structure and effects of the invention]

The present invention is an AE installed near the tool of a machine tool.
A tool breakage detection device that includes a sensor and detects breakage based on an AE signal obtained when the tool breaks, a sampling means that samples the output signal of the AE sensor at predetermined time intervals; maximum density detection means for detecting the maximum density of a probability density function of the data obtained at each predetermined time; a comparison means for providing an output when the maximum density obtained by the maximum density detection means exceeds a predetermined level; and an output of the comparison means. a signal preprocessing section having a gate means for opening the gate of the AE sensor based on the signal preprocessing section;
The present invention is characterized by comprising a signal processing section that identifies breakage of the tool when an output signal from the sensor is given.

According to the present invention having such characteristics, AE
The signal obtained from the sensor is first given to the signal pre-processing unit to obtain the probability density function of the AE sensor output, and tool wear is detected based on the maximum density. In this way, the wear state of the tool can be detected without malfunction regardless of the shape of the chips. When the wear of the tool exceeds a predetermined value and breakage is likely to occur, the AE signal is transmitted to the signal processing section and signal processing is performed. This eliminates the possibility of erroneously detecting various signals from the workpiece or workbench as tool breakage signals, reducing the burden on the signal processing unit and greatly improving the reliability of tool breakage detection. Become.

[Explanation of Examples]

(Overall Configuration of Embodiment) FIG. 1 is a block diagram showing an embodiment of a tool breakage detection device according to the present invention. This embodiment shows a tool breakage detection device attached to a drilling machine controlled using a numerical control device. A workpiece 1 is fixed on the base of the drilling machine, and a drill 2 is inserted from the top of the workpiece 1. The workpiece 1 is opened by rotating and pushing down at a predetermined speed. The operation of the drill 2 is controlled by a numerical control device 3. It is assumed that the drill 2 used here is automatically replaced by an automatic tool changer (not shown). Now, before cutting the workpiece 1, a pseudo-breakage signal generator 4 made of PZT or the like, like the AE sensor, is installed in advance at the position where the drill blade contacts the upper part of the workpiece 1. The drive circuit 5 drives the pseudo-breakage signal generator 4, and is configured in advance to oscillate a drive waveform that is similar to and has the same power spectrum distribution as the AE output waveform when the tool breaks. Its amplitude level is given externally. An AE sensor 6 for detecting an AE signal is provided near the tool on which the workpiece 1 is placed, for example, on the base as shown in FIG. AE sensor 6
is a wideband AE that detects AE signals from tools such as drill 2 and AE signals from pseudo-breakage signal generator 4.
It is a sensor, and its output is given to the signal preprocessing section 7. The signal preprocessing section 7 detects a sign of tool breakage based on the degree of wear of the tool, as will be described later, and if a sign is detected, it provides the output from the AE sensor 6 to the AE signal processing section 8. The AE signal processing unit 8 amplifies the signal from the AE sensor 6 to a predetermined level, detects a signal indicating tool breakage or abnormal cutting, and provides the signal to a central processing unit (hereinafter referred to as CPU) 10 through an input/output interface 9. It is. The CPU 10 is a read-only memory (hereinafter referred to as ROM) 11 that stores a system control program and a communication control program with the numerical control device 3.
A storage means consisting of a random access memory (hereinafter referred to as RAM) 12 containing AE sensor sensitivity information corresponding to the tool used by the numerical control device 3 is connected to the memory. The CPU 10 is also connected to the input/output interface 13 for information about the cutting process.
An indicator 14 that displays the AE signal level, abnormal cutting or breakage of the tool, and the tool number, type, and preparation information.
An input key 15 for setting the sensitivity of the AE sensor is connected. Further, a numerical control device 3 is connected via a signal transmission line 16. Based on these inputs, the CPU 10 applies a predetermined drive level of the pseudo-breakage signal generator 4 to the drive circuit 5, detects the optimal sensitivity of the AE sensor, and sequentially stores it in the RAM 12, and also outputs the AE sensor when the tool is used. The sensitivity is set and the tool is controlled to detect tool breakage based on the output of the AE signal processing section.

(Configuration of Signal Preprocessing Unit) FIG. 2 is a block diagram showing the detailed configuration of the signal preprocessing unit 7. As shown in FIG. The signal preprocessing section 7 is the AE sensor 6
Since the AE signal is directly processed, it may be stored in the AE sensor 6. In this figure, the output of the AE sensor 6 is given to a preamplifier 20 in a signal preprocessing section 7. Preamplifier 20
The output of the AE sensor 6 is amplified to an appropriate level and the output is transmitted to the band filter 21. The bandpass filter 21 is a filter that passes the AE signal with a frequency between 100 and 300 KHz among the AE signals obtained from the AE sensor 6, and removes noise components associated with mechanical vibration etc. and performs full-wave rectification of the output. circuit 22
tell to. The A/D converter 23 is a full-wave rectifier circuit 22
The output is converted into digital data at a predetermined minute sampling period, and the output is provided to the microcomputer 24. The microcomputer 24 has an arithmetic processing section and a storage means, and processes the applied AE signal according to an arithmetic processing procedure described later, compares it with a predetermined threshold level, and outputs a gate signal. be. The gate circuit 25 receives the signal from the AE sensor 6 when the gate signal is given from the microcomputer 24.
The AE signal is passed through the signal pre-processing unit 7 to the signal processing unit 8
It is something that can be conveyed to people.

Next, FIG. 3 shows a memory map of the storage means within the microcomputer 24. As shown in this figure, the storage means is provided with sample data areas X(o), X(1), ... It has probability density tables A(o), A(1), . . . A(k) indicating probability densities of sample data. Each area of the probability density table holds the number of data that takes the value of each sample data. In addition, the storage means includes a sample data pointer i used when signal processing the sample data, and a maximum density function that stores the sample data number corresponding to the sample data X(i) with the highest probability of occurrence as a result of the calculation processing described later. A maximum value register m and a threshold level Th for detecting the maximum density when tool wear reaches a predetermined limit value are provided.

(Configuration of AE Signal Processing Section) Next, FIG. 4 is a block diagram showing the detailed configuration of the AE signal processing section 8. As shown in FIG. In this figure, the output of the AE sensor 6 is first given to an analog switch 30. The analog switch 30 is a switch that cuts off the analog signal based on the control signal from the CPU 10, and its output terminal is connected to the variable gain amplifier 31.
It is connected to the. The amplifier 31 is a variable amplification factor amplifier whose amplification factor can be set based on the control input from the CPU 10, and provides its output to the CPU 10 via two bandpass filters 32, 33 and the input/output interface 9. It is something. The bandpass filter 32 has a center frequency
300KHz, bandpass filter 33 is the center frequency
It is a 50 kHz filter and transmits only signals near the respective center frequencies to the next-stage detectors 34 and 35. The detectors 34 and 35 each detect the input signal and obtain an output according to the amplitude.The output of the detector 34 is sent to a differentiating circuit 36 and
The outputs of are given to comparators 37, respectively. These bandpass filters 32, 33, detectors 34, 3
5 and the comparator 37 form a frequency identification means for identifying the AE signal at the time of breakage. Differential circuit 3
Reference numeral 6 transmits only steep changes in the input signal to the next stage level judger 38. The level determiner 38 compares the input signal with a predetermined reference level,
If the input signal is large, the output is transmitted to the breakage detection circuit 39 and the abnormal cutting detection circuit 40. Also, a comparator 37 compares the outputs of the detectors 34 and 35, and transmits the output to the breakage detection circuit 39 only when the output of the detector 34 is large.
tell to. The breakage detection circuit 39 is a logic circuit that performs the logical product of these inputs to detect tool breakage, and transmits a detection signal to the CPU 10 via the input/output interface 9. Also, abnormal cutting detection circuit 4
0 detects abnormal cutting based on the output of the level judger 38 and outputs the signal via the input/output interface 9.
This is to be transmitted to the CPU 10.

(AE Sensor Sensitivity Setting Operation) Next, the operation when installing the AE sensor of this tool breakage detection device in a machine tool, in this case a drilling machine, will be explained in this embodiment. First, the AE sensor 6 is attached to a predetermined position, and then the drive circuit 5 is operated in accordance with the level given from the input key 15 that corresponds to the tool being used. Then, the pseudo-breakage signal from the pseudo-breakage signal generator 4 will be transmitted to the AE via the workpiece 1.
The information is transmitted to the sensor 6. At this time, the pseudo-breakage signal obtained from the AE sensor 6 has the same power spectrum as when the tool breaks, and also has a waveform similar to the waveform when the tool breaks in the time domain. this
The AE signal passes through the gate 25 of the signal preprocessing section 7.
The signal is transmitted to the AE signal processing unit 8 and further transmitted from the input/output interface 9 to the CPU 10 via the analog switch 30 and variable gain amplifier 31. Based on the level, the CPU 10 changes the amplification factor of the variable amplification factor amplifier 31 so that the output level becomes appropriate, and sets the amplification factor to the optimum amplification factor.

(Operation of Signal Preprocessing Unit) Next, the operation of the preprocessing unit when monitoring tool breakage will be described. During machining, an opening is made using a drill 2 by removing the pseudo-breakage signal generator 4 from the workpiece 1 placed on the base. In this way, an AE signal is obtained from the AE sensor 6 at the same time as cutting, and is amplified by the preamplifier 20 of the signal preprocessing section 7. The AE signal is rectified through a bandpass filter 21 and given to an A/D conversion circuit 23, where it is converted into digital data at a predetermined sampling period. Now, when the operation starts, the microcomputer 24, as shown in the flowchart of FIG.
n) and the sample data pointer i, and in step 42, the A/D-converted sampling data at every predetermined period T is temporarily stored in the sample data areas X(o), X(1), ...X(n) in the memory. are stored sequentially.
FIG. 6 is a graph showing the relationship between the contents of the sample data pointer i and each sample data X(i) for an arbitrary rectified waveform. Next, the process proceeds to step 43, where the sample data pointer i is compared with the last value n of the sample data. Initially, the sample data pointer i is o, so proceed to step 44,
The first sample data X(o) is read, and then the process proceeds to step 45, where the content of the probability data area A{X(o)} corresponding to the value of the read sample data X(o) is incremented. The process then proceeds to step 46, where the value of the sample data pointer i is incremented, and the process returns to step 43. In this way, by repeating the loop from step 43 to step 46 until the sample data pointer i reaches the last value n of the sample data, all sample data values are distributed into probability data areas according to their sizes. Complete the probability density function table for {X(i)}. The process then proceeds to step 47 where the sample data pointer i and the maximum value register m are cleared, and the process proceeds to step 48 where the sample data pointer i and the final value n are compared, and if these values are not equal, the process proceeds to step 49. In step 49, since the first sample data pointer i is o, it is checked whether the probability density data A{X(m)} exceeds A{X(o)}. If A{X(m)} is large, step 50
In step 51, the value of the sample data pointer i at that time is stored in the maximum value register m, and in step 51, the sample data pointer i is incremented. If A{X(m)} does not exceed the probability value A{X(i)} of the sample data in step 49 as in the initial operation, i is incremented in step 51 without going through step 50. and return to step 48. In this way, by repeating the loop of steps 48 to 51, all probability density data A
The maximum value A{X(m)} is found for A(k) from (o),
At the same time, the sample data number i corresponding to this maximum value is held in the maximum value register m.
The program then proceeds to step 52 to check whether the maximum value A{X(m)} exceeds the threshold level Th. If this level Th has not been exceeded, there has not been much wear, so the step
Return to step 41 and repeat the same process, and if the threshold level Th is exceeded, step
At step 53, a gate open signal is given to the gate circuit 25, and the process ends.

Now, since the maximum density (m) is proportional to the wear as shown in Figure 7, the threshold level
By appropriately selecting Th, it is possible to open the gate circuit 25 when wear exceeds a predetermined value. Furthermore, as already shown in Japanese Patent Application No. 59-110564, etc., when arc-shaped chips or coil-shaped chips are obtained, the average value of the AE signal is
(v) and maximum density (m) are at almost the same level. Furthermore, when spiral chips are generated, the average value (v) of the AE signal becomes a considerably large value, but the level of maximum density (m) does not increase. Therefore, if the threshold level is set to a predetermined level using the maximum density (v) of the AE signal, the maximum density (m) will not increase due to changes in the chip shape and will only be proportional to wear, preventing malfunctions. will not occur.

(Monitoring Operation) Next, a description will be given of a monitoring operation for monitoring tool breakage using the optimal sensitivity data corresponding to each tool set in this way. FIG. 8 is a flowchart showing this monitoring operation. When the monitoring operation starts, first, in step 60, based on the input of the tool used, the optimum value of the amplification factor corresponding to the tool is stored in the RAM 1.
2 to set the amplification factor of the variable amplification factor amplifier 31 of the AE signal processing section 8. If the machining continues, the tool, in this case the drill 2, will wear out, so after a certain period of time, the AE signal from the AE sensor 6 will be transmitted from the signal preprocessing section 7 via the gate circuit 25 to the AE signal.
The signal is transmitted to the signal processing section 8. Then, in response to cutting of the workpiece 1, an AE signal is applied to an analog switch 30, amplified by an optimum amplification factor, and applied to two bandpass filters 32 and 33. Now, the power spectrum distribution of the AE signal given by the AE sensor 6 during normal cutting is concentrated around a frequency of 50 KHz, as shown by curve b in Figure 9, and monotonically attenuates in the higher frequency range. It is becoming. Furthermore, as is known from many experiments, the power spectrum distribution when a tool breaks is represented by curve a in FIG. 9, and it has been revealed that it has a peak at a frequency of around 300 KHz. This is considered to be because the signal source is not caused by mechanical vibration, but a phenomenon peculiar to ultrasonic waves that occurs during non-plastic fracture of a tool occurs. Therefore, only the AE signals near the respective frequency components are extracted by the two bandpass filters 32 and 33 and sent to the detectors 34 and 33.
35, and by comparing the output levels, it is possible to clearly distinguish between normal conditions and tool breakage conditions. The comparator 37 compares these outputs and provides a signal to the breakage detection circuit 39 only when the tool breaks.

On the other hand, a signal similar to the power spectrum distribution shown by curve a in FIG. 9 may be generated due to contact or friction between chips and the tool workpiece during cutting. In this case, bandpass filters 32, 3
Even if the center frequency and Q value of 3, the threshold level of comparator 37, etc. are set appropriately, a signal caused by contact or friction between chips and a tool or workpiece may be mistakenly judged as a tool breakage signal. . Therefore, in this embodiment, attention is also paid to the time domain waveform of the AE signal seen when the tool breaks, and these signals are separated. In other words, the AE obtained when the tool breaks
As shown in Figure 10a, the signal waveform is a signal with a sharp rise at the time of breakage, while the AE caused by contact between chips and the tool or workpiece and wear.
As shown in FIG. 10b, the signal has a waveform that does not show a sharp rise and continues for a predetermined period of time. Therefore, in the AE signal processing unit 8, the output of the wave detector 34 is applied to a differentiating circuit 36, and only steep signals such as those caused by breakage are separated and applied to the level determiner 38. The level determiner 38 provides an output to the breakage detection circuit 39 and the abnormal cutting detection circuit 40 when the input signal is large. The abnormal cutting detection circuit 40 is connected to the CPU 1 from the input/output interface 9 based on the output of the level judger 38.
Inform 0 of abnormal cutting. In the flowchart shown in FIG. 8, the CPU 10 is the abnormal cutting detection circuit 40.
It is checked whether or not an abnormal cutting signal is transmitted from the machine (step 61). If there is no abnormal cutting signal, the cutting operation is being performed normally, so the process proceeds to step 62 and the cutting level is displayed on the display 14. Then, the process returns to step 61 and the same process is repeated to monitor cutting abnormalities. Now, if an abnormal cutting signal is transmitted from the abnormal cutting detection circuit 40, step 63
Then, it is checked whether a breakage signal is given from the breakage detection circuit 39. Breakage detection circuit 39
detects tool breakage by the logical product of the comparator 37 and the level determiner 38, and transmits the breakage output to the CPU 10 from the input/output interface 9 when the tool breaks. Therefore, if no breakage signal is given in step 63, abnormal cutting has been performed, and therefore, in step 64, the display 14 displays abnormal cutting, and the process returns to step 60. If a breakage detection signal is given in step 63, tool breakage has been detected, so the process proceeds to step 65, where the display 14 displays the occurrence of abnormal vibration and transmits the data to the numerical control device 3 to control its operation. Stop. Then, in step 66, the analog switch 30 of the AE signal processing section 8 is turned off to end the process. After the tool breaks, the display 1 displays a large AE signal that is generated due to abnormal contact or friction between the broken tool and the workpiece.
This is so that the signal level at the time of breakage can be recognized by not displaying it by 4.

If the wear exceeds a certain level, there is a high possibility that the tool will break, so after the wear has become considerably large due to the pre-processing in the signal pre-processing section 7, the gate is opened and the AE signal is transferred to the AE signal processing section 8. is giving to The AE signal processing unit 8 improves the reliability of breakage detection by combining both frequency domain breakage detection and time domain breakage detection.

In this embodiment, breakage is detected by analog processing in the AE signal processing section 8;
It is also possible to sample the sensor output after A/D conversion and perform all subsequent processing by digital signal processing. In this case, it is conceivable to replace the bandpass filter with a digital filter, replace the differentiation circuit with a difference operation, etc., and perform signal processing using the CPU. In the present invention, since the AE signal is not transmitted to the signal processing unit until the tool is not worn to a considerable extent, there is no need to perform complex signal processing, making it possible to significantly reduce the load on the signal processing unit, and reducing processing capacity. It becomes possible to detect tool breakage even with the use of an inexpensive microcomputer or the like.

Furthermore, in this embodiment, the signal processing section detects breakage by the AND of breakage detection in the frequency domain and breakage detection in the time domain. It is also possible to use or combine other breakage detection methods.

Furthermore, although this embodiment describes a breakage detection device applied to a drilling machine using a numerical control device, the present invention can also be applied to other machine tools, such as various machine tools such as lathes and milling machines, and even large-scale machining centers. It is also possible to apply

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the tool breakage detection device according to the present invention, FIG. 2 is a block diagram showing the detailed configuration of the signal preprocessing section, and FIG. 3 is a block diagram showing the detailed configuration of the signal preprocessing section. A memory map showing the memory contents, Fig. 4 is a block diagram showing the detailed configuration of the AE signal processing section, Fig. 5 is a flowchart showing the operation of the signal preprocessing section for identifying tool wear, and Fig. 6 is a block diagram showing the detailed configuration of the AE signal processing section. A graph showing the relationship between the sample data pointer i and each sample data X(i) for the rectified waveform, FIG. 7 is a graph showing the relationship between maximum density and degree of wear, and FIG. 8 is a flowchart showing the tool breakage monitoring operation. Chart, Figure 9 is a diagram showing the power spectrum of the AE signal obtained from the AE sensor 6,
Figure 10a shows the AE signal waveform obtained when the tool breaks.
FIG. 10b is a diagram showing the AE signal waveform obtained when chips are generated. DESCRIPTION OF SYMBOLS 1... Workpiece, 2... Drill, 3... Numerical control device, 4... Pseudo-breakage signal generator, 5... Drive circuit, 6... AE sensor, 7... Signal preprocessing unit, 8
...AE signal processing unit, 10...CPU, 11...
ROM, 12...RAM, 14...Display, 15
...Input key, 20...Preamplifier, 21,3
2, 33...Band pass filter, 22...Full wave rectifier circuit, 23...A/D converter, 24...Microcomputer, 25...Gate circuit, 34,
35... Detector, 36... Differential circuit, 37... Comparator, 38... Level determiner, 39... Breakage detection circuit, 40... Abnormal cutting detection circuit.

Claims (1)

[Claims] 1. A tool breakage detection device that has an AE sensor installed near a tool of a machine tool and detects breakage based on an AE signal obtained when the tool breaks, comprising: sampling means for sampling at predetermined time intervals; maximum density detection means for detecting the maximum density of a probability density function of data sampled by the sampling means at predetermined time intervals; a signal preprocessing unit having a comparison unit that provides an output when the level is exceeded, and a gate unit that opens the gate of the AE sensor based on the output of the comparison unit; and an output of the AE sensor from the gate unit of the signal preprocessing unit. A tool breakage detection device comprising: a signal processing unit that identifies tool breakage when a signal is applied. 2 The signal processing section is obtained when the tool breaks.
Frequency identification means that outputs when an AE signal with a frequency component that has a strong correlation with the frequency component of the AE signal is given, and a rising signal detection means that outputs when a signal that rises suddenly from the AE sensor is given. and logic output means for identifying tool breakage based on the AND output of the frequency identification means and the rising signal detection means. Detection device.