JP2020104257A - Abnormality detection device and abnormality detection method of cutting tool - Google Patents

Abstract

To provide abnormality detection device and method, which can suppress extension of tool breakage.SOLUTION: Abnormality detection method or device for detecting breakage of a cutting tool with a plurality of blades (A) receives a measurement signal to be a measurement result of a current value or a power value of a spindle motor of a machine tool in which the cutting tool is mounted, (B) performs filter processing of the measurement signal and stores the signal in time series in a storage section, (C) detects an abnormality, and (D) outputs a command of machining machine control designated by a control parameter to the machine tool when determining the presence of tool breakage by the abnormality detection. In the (C) abnormality detection, (C1) the measurement signals of frequency analysis data number in a machining region in which simultaneous action blade number is predicted to be one are extracted from the stored measurement signals and a frequency spectrum is calculated, and (C2) the presence of the tool breakage is determined on the basis of amplitude of a determination frequency designated by the control parameter in the frequency spectrum (=rotation frequency×n, 1≤n<tool blade number, n is an integer) and a threshold designated by the control parameter.SELECTED DRAWING: Figure 1

Description

The present invention relates to an abnormality detection device and an abnormality detection method for a cutting tool in cutting.

As a background art of this technical field, there is International Publication No. 00/073018 (Patent Document 1). In this publication, "a vibration detection sensor that detects vibrations generated by machining of a machining center, and a peak detection unit that periodically counts the number of times a peak exceeding a predetermined value occurs in the vibrations detected by the sensor. And a comparison unit that compares the number of peak occurrences detected by the peak detector with a predetermined threshold value and outputs a comparison signal when the number of peak occurrences exceeds a predetermined threshold value, and a comparison signal from the comparison unit. According to the above, a work stop unit for outputting a stop signal to the control device of the machining center and a wireless calling unit for notifying the operator of the cutting tool replacement are provided."

International Publication No. 00/073018

In the configuration of Patent Document 1 described above, the vibration of the cutting edge of the cutting tool that is gradually chipped due to collision with the work material and irregularly generated is determined by the number of peak occurrences within a predetermined time, and the cutting tool is effective. Is what you use. However, when the cutting edge is suddenly largely chipped off, the number of peak occurrences does not increase and abnormality detection cannot be performed.

When the work material is a casting, foreign matters and cavities are present in the casting, and during cutting, these may cause a sudden chipping of the cutting edge. In an automatic production line that automatically carries in and out a work material as a product to a cutting machine, the production may be continued without noticing the chipping of the cutting edge. As a result, there is a problem that the tool expands in damage and leads to breakage, and the work material continuously suffers processing defects such as scratches and uncut parts.

Therefore, an object of the present invention is to provide an abnormality detection device and an abnormality detection method for a cutting tool capable of suppressing the expansion of tool damage and suppressing the occurrence of continuous machining defects.

An abnormality detection method or device for detecting a defect of a cutting tool having a plurality of blades of the present invention is (A) measurement of a current value or a power value of a spindle motor that rotates a spindle of a machine tool on which the cutting tool is mounted. The measurement signal as a result is received, (B) the measurement signal is filtered and stored in a storage unit in time series, (C) abnormality is detected, and (D) there is a tool defect by abnormality detection. When it is determined that the processing machine control command specified by the control parameter is output to the machine tool. Here, in (C) abnormality detection, (C1) the frequency spectrum is calculated by extracting the measurement signal of the number of frequency analysis data of the processing region in which the number of simultaneously acting blades is predicted to be 1 from the stored measurement signal. Then, (C2) the amplitude of the determination frequency (=rotation frequency×n, where 1≦n<the number of tool flutes, n is an integer) specified by the control parameter in the frequency spectrum, and the threshold value specified by the control parameter, Based on the above, it is determined that the tool is missing.

Other features of the present invention will be described in embodiments.

According to the present invention, it is possible to suppress the damage expansion of the tool and the occurrence of continuous machining defects.

It is a block diagram which shows the abnormality detection apparatus of the cutting tool which concerns on Example 1 connected to the processing machine. It is a perspective view showing the appearance of a cutting tool. It is a perspective view which shows the general view of the cutting tool with which the cutting edge was missing. It is a perspective view which shows the general view of the cutting tool with which the cutting edge and the tool body were damaged. It is a graph which shows the current value of a measurement signal. It is a graph which shows the electric current value of the measurement signal at the time of normal. It is a graph which shows the electric current value of the measurement signal at the time of abnormalities. It is a graph which shows the frequency spectrum distribution of the measurement signal at the time of normality. It is a graph which shows the frequency spectrum distribution of the measurement signal at the time of abnormality. It is a graph which shows the change of the n-fold component of a rotation frequency in the frequency analysis result which divided the whole processing time range. It is a cutting simulation result at the time of normality without a defect in a 4-flute tool. It is a graph which shows the frequency spectrum distribution of the cutting simulation result at the time of normality without a defect in a 4-flute tool. It is a cutting simulation result which assumed the state where one insert was missing in a 4-flute tool. It is a graph which shows the frequency spectrum distribution of the cutting simulation result which assumed the state where one insert was missing in a 4-flute tool. It is a waveform of the difference between the cutting simulation in the state where one insert is missing and the normal state in the four-blade tool. It is a graph which shows the frequency spectrum distribution of the waveform of the difference of the cutting simulation of the state where one insert of a tool was missing in a 4-flute tool, and a normal state. It is a result of showing a frequency with a large amount of change in amplitude by a magnification with respect to a rotation frequency by comparing a frequency spectrum distribution in a normal state and a state in which one insert is missing in a tool having 3 blades to 10 blades. It is an example of a screen for inputting control parameters on the display unit of the abnormality detection device. It is an example of a flowchart of the abnormality detection processing executed by the abnormality detection processing unit of the abnormality detection device. It is a block diagram which shows the abnormality detection apparatus of the cutting tool which concerns on Example 2 connected to the processing machine. It is a graph which shows the frequency spectrum distribution of the main spindle servomotor current in processing of a 3 blade tool. It is a graph which shows the change of the amplitude of the judgment frequency of a 3-flute tool with respect to the cumulative number of machining. It is a graph which shows the change of the level correction value of the amplitude of the decision frequency of the three-blade tool which carried out level correction on the basis of the amplitude of the decision frequency of the 1st processing with respect to the number of cumulative machining. It is a graph which shows the change of the level correction value of the amplitude of the decision frequency of the three-blade tool which carried out level correction on the basis of the amplitude of the decision frequency of the 1st processing with respect to the number of processes. The difference between the level correction value of the amplitude of the judgment frequency of each machining, which was level-corrected by the amplitude of the judgment frequency of the first machining of the 3-flute tool, and the level correction value of the machining one before is to the number of machining. It is a graph shown. It is the 1st example of a display in the method of making an abnormality determination using the level correction value of the amplitude of the determination frequency of each processing, which is level-corrected with the amplitude of the determination frequency of the first processing as a reference. It is the second display example in the method of making an abnormality determination using the level correction value of the amplitude of the determination frequency of each processing that is level-corrected with the amplitude of the determination frequency of the first processing as a reference. It is the 3rd example of a display in the method of making an abnormality judgment using the level correction value of the amplitude of the judgment frequency of each processing which level-corrected on the basis of the amplitude of the judgment frequency of the 1st processing. It is the 4th example of a display in the method of carrying out an abnormal judgment using the level correction value of the amplitude of the judgment frequency of each processing which level-corrected on the basis of the amplitude of the judgment frequency of the 1st processing. In the fourth display example of the method for determining an abnormality using the level correction value of the amplitude of the determination frequency of each process, which is level-corrected based on the amplitude of the determination frequency of the first process, an example in which the cumulative number of processes has increased Is. It is the 5th example of a display in the method of carrying out an abnormal judgment using the level correction value of the amplitude of the judgment frequency of each processing which level-corrected on the basis of the amplitude of the judgment frequency of the 1st processing. 9 is an example of a screen for inputting control parameters on the display unit of the abnormality detection device according to the second embodiment.

The present invention measures a current value or a power value of a motor (spindle motor) that drives a cutting tool in a processing machine, and in a frequency spectrum distribution thereof, a tool cutting edge that is a loss of the cutting tool due to an increase in a specific frequency component. It detects a defect and controls the operation of the processing machine.

Embodiments will be described below with reference to the drawings. In all the drawings for explaining the present embodiment, those having the same function are designated by the same reference numeral, and the repeated description thereof will be omitted in principle. However, the present invention should not be construed as being limited to the description of the embodiments below. It is easily understood by those skilled in the art that the specific configuration can be changed without departing from the idea or the spirit of the present invention.

In this embodiment, an example of an abnormality detecting device for detecting an abnormality of a tool based on a frequency analysis result of a current value or a power value will be described.

FIG. 1 is a block diagram showing a processing machine 4 including an abnormality detection device 1 for a cutting tool according to the first embodiment. This is a configuration in which the cutting tool abnormality detection device 1 is attached to the processing machine 4.

The processing machine 4 mounts a cutting tool 5 having a plurality of blades gripped by a tool holder 110 on a spindle 7, and rotates the cutting tool together with the spindle to machine a work material 6 such as a flat surface or a groove. It is an example of a milling machine tool.
In the processing machine 4, the cutting tool 5 is fixed to the main shaft 7 and the main shaft 7 is rotated by the main shaft motor 8 to process the work material 6 into a desired shape. At this time, the servo amplifier 9 controls the voltage value and the current value input to the spindle motor 8 by a command from the NC device 10 in which the NC program is stored. As a result, the spindle motor 8 is rotated at the rotation speed as instructed.

The abnormality detection device main body 15 can be configured on a general-purpose computer, and its hardware configuration includes an arithmetic unit 2 including a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), HDD (Hard Disk Drive), storage unit 3 configured by SSD (Solid State Drive) using flash memory and the like, input unit 11 configured by an input device such as a keyboard and mouse, LCD (Liquid Crystal display), a display device such as an organic EL display, a display unit 12 including various output devices, a NIC (Network Interface Card), a communication unit 13 including input/output interface devices, and the like.

The communication unit 13 measures a voltage value and a current value from between the servo amplifier 9 of the processing machine 4 and the spindle motor 8 via a wired network or a wireless network or an individual dedicated cable, and the NC device 10. Connected with.

The measuring unit 14 includes a voltage sensor, a current sensor, an AD converter, a communication unit, and the like. The abnormality detection device 1 includes an abnormality detection device body 15 and a measurement unit 14.

The arithmetic unit 2 realizes the following respective functional units by loading the cutting tool abnormality detection program 31 stored in the storage unit 3 into the RAM and executing it by the CPU. The calculation unit 2 includes an abnormality detection processing unit 21, a signal processing unit 22, an analysis determination unit 23, a processing machine command unit 24, and a control parameter setting unit 25.

The storage unit 3 has a storage area for storing each information of the cutting tool abnormality detection program 31, the control parameter 32, and the measurement signal 33.

The abnormality detection processing unit 21 is activated in response to a signal for the NC device 10 to start machining, and performs abnormality detection processing of the cutting tool until the end of machining in the machining area designated by the control parameter, or an abnormality of the cutting tool is detected. If it is, the control instruction of the processing machine 4 set in advance is executed until the issuing is completed.

The signal processing unit 22 is started by the abnormality detection processing unit 21 at the start of processing, and a measurement signal (digital signal) of a current value or a power value obtained between the servo amplifier 9 and the spindle motor 8 measured and output by the measuring unit 14 is output. The signal) is subjected to prior signal processing and recorded in the measurement signal 33 storage area of the storage unit 3. A noise removal filter such as a low pass filter, a high pass filter, a band pass filter, or a moving average filter is used as the prior signal processing method. This process continues until the end of processing.

The analysis determination unit 23 is activated by the abnormality detection processing unit 21 and determines that the cutting tool is abnormal with respect to the measurement signal 33 in a predetermined time range from a predetermined start time of the machining duration time stored in advance as the control parameter 32. Analyze for any perceptible changes. When it is determined that there is a change that is recognized as an abnormality, the abnormality detection processing unit 21 activates the processing machine command unit 24.

When it is determined that the cutting tool has a change that is recognized as abnormal, the processing machine command unit 24 issues a command for executing the operation of the processing machine 4 stored in advance as the control parameter 32 to the NC device 10 of the processing machine 4. Output to. The command at this time is tool replacement, tool withdrawal, or the like.

The control parameter setting unit 25 is activated by a user's instruction, receives a user input of a control parameter relating to a processing condition, a measurement condition, a processing machine control method, a signal processing method, etc. before the processing operation, and controls the storage unit 3. Register in parameter 32.

FIG. 2 shows an example of the cutting tool 5. The cutting tool 5 is a replaceable cutting edge tool having a structure in which an insert 41, which is a cutting edge of the tool, is fixed to a tool body 43 by a fixing screw 44. In the cutting tool 5, the work material 6 is processed in the vicinity of the corner portion 42 of the insert 41. The corner portion 42 becomes a cutting edge. The cutting tool 5 is a four-blade tool having four inserts 41, and the insert 41 is an example of four corners having four corner portions 42 serving as cutting edges.

FIG. 3 shows an example of the cutting tool 5 in which the cutting edge is missing. In the four inserts 41, the corner portion 42b of one insert 41b is missing. Since the tool body 43 is not damaged, the tool can be returned to a normal state by replacing only the insert 41b or all of the four inserts 41.

In addition, when the set tool life is reached, the insert is replaced by rotating the insert so that the unused cutting edge has a corner if it acts as a cutting edge. On the other hand, if all the cutting edges have been used up, replace the insert itself. With multi-blade tools, all inserts are typically replaced at the same time.

When the insert has a chip as shown in FIG. 3, when the chip is small, the insert is rotated so that the unused cutting edge works, and when the chip is large, the insert itself is replaced.

FIG. 4 shows an example of the cutting tool 5 in which the cutting edge and the tool body are damaged. This is an example in which the corner portions 42 of the four inserts 41 are all missing and the tool body 43 is also damaged 46. This is because the cutting was continued without noticing that one of the cutting edges shown in FIG. 3 was missing, and the damage was expanded to other inserts and the tool body. Since the work material machined with the tool in this state is scratched or left uncut, a large number of machining defects continuously occur until the tool damage is noticed.

As described above, when the work material is a cast product, it is difficult to completely prevent the tool loss because foreign matters and cavities existing in the cast product cause the cutting edge of the tool. If the machining can be stopped and the insert can be replaced when one insert is missing, it is possible to suppress damage to the tool body and to prevent machining defects that occur continuously.

FIG. 5 is a graph of the measurement signal 50 taking the current value measured by the measurement unit 14 as an example. This is an example of shoulder milling with a width of 6 mm and a depth of 5 mm under the conditions of a cutting speed of 45 m/min and a blade feed amount of 0.1 mm/tooth. The horizontal axis represents time [s], and the vertical axis represents the current value [A] in the spindle 7 of the processing machine measured between the spindle motor 8 and the servo amplifier 9.

In the signal waveform shown in FIG. 5, since the main shaft 7 is already measured from the state of rotation, the current value in the state of idling from 0 to about 2 seconds occurs, and from about 2 seconds. The cutting tool 5 comes into contact with the work material 6 to start machining. When the machining is started, a load is generated on the spindle 7, so that the control for rotating the spindle motor 8 at a constant number of rotations works from the servo amplifier 9 to increase the current value detected by the measuring unit 14.

FIG. 5 is an example of the measurement signal 50 of the current value when one of the four inserts 41 is missing during processing. A spike-shaped waveform 51 is seen around the processing time of about 12 seconds, and the load is momentarily increased. At this time, a large load is applied to one of the inserts 41 and a defect occurs.

FIG. 6 is a graph of the measured signal 50 before the tool is heavily loaded. This is obtained by extracting the measurement signal 50 in the range 52 having a processing time of 10 to 12 seconds in the graph of FIG. Since the number of simultaneously acting blades (the number of blades of the cutting tool acting on the processing of the work material at a certain time) is one, each amplitude of the waveform corresponds to the processing of each insert 41. Since the tool 5 is a four-blade tool, it means that the tool makes one rotation with four amplitudes. In FIG. 6, it can be seen that there are no extremely small amplitude variations, but each insert 41 of the tool 5 acts.

FIG. 7 is a graph of the measured signal 50 after the tool has been heavily loaded. This is obtained by extracting the measurement signal 50 of the processing time 13 to 15 seconds in the range 53 in the graph of FIG. In FIG. 7, a large amplitude waveform 55 is seen immediately after the small amplitude waveform 54 with large amplitude variations. This indicates that one of the inserts was missing and the cut amount was reduced, so the load was reduced, and the cut amount that was left uncut by the previous insert was added to the next insert, and the load increased. Further, since the amplitude 54 of the waveform 54 having a small amplitude is not completely zero, it can be estimated that the size of the insert defect is still smaller than the cut amount.

FIG. 8 is a frequency spectrum distribution of the measurement signal 50 for a machining time of 10 to 12 seconds before a heavy load is applied to the tool. In this figure, the amplitude is normalized to 1 on the basis of a frequency (hereinafter referred to as a fundamental frequency) obtained by multiplying the frequency of the number of revolutions of the main spindle 7 (hereinafter referred to as a revolution frequency) by the number of blades of the tool 5 (hereinafter referred to as a fundamental frequency) ( Hereinafter, the amplitude of the rotation frequency will be referred to as a rotation frequency component, the amplitude of the n-fold rotation frequency will be referred to as an n-fold component of the rotation frequency, and the amplitude of the fundamental frequency will be referred to as a fundamental frequency component).

Although there are peaks in the rotation frequency component 61, the rotation frequency double component 62, and the rotation frequency triple component 63, the peaks are smaller than the fundamental frequency component 60. This is the frequency spectrum distribution in the normal state where the insert 41 of the tool 5 is not missing.

FIG. 9 is a frequency spectrum distribution of the measurement signal 50 for a machining time of 13 to 15 seconds after a heavy load is applied to the tool. Similar to FIG. 8, the fundamental frequency component 65 is normalized to 1. The amplitudes of the rotation frequency component 66, the rotation frequency double component 67, and the rotation frequency triple component 68 are larger than those in FIG. In particular, the double component 67 of the rotation frequency has a larger increase amount than the double component 62 of the rotation frequency in the normal state. By setting the threshold value 64 so that the double component 67 of the rotation frequency exceeds, it is possible to detect an abnormality in which one insert 41 of the tool 5 is missing. Although the threshold 64 is shown in FIG. 8, when the tool is in a normal state, all the components of 1 to 3 times the rotation frequency are below the threshold.

FIG. 10 shows changes in each frequency component of the frequency analysis result in the entire processing time range. In the processing time of 1 second to about 30 seconds, frequency analysis is performed about every 2 seconds, the fundamental frequency component 70 is normalized to 1, and each frequency component is shown for each time. The plot time is the median value of the frequency analysis time, and the analysis result of, for example, 1 to 3 seconds is shown in 2 seconds.

Further, the frequency analysis includes a fast Fourier transform (FFT) and a discrete Fourier transform, but it is preferable to use the fast Fourier transform because analysis can be performed in a short time. In the case of Fast Fourier Transform, the number of data (the number of input samples of FFT collectively) must be a power of two. The current value shown in FIG. 5 is measured at a sampling cycle of 0.001 seconds, so when performing frequency analysis at intervals of about 2 seconds, 2048 (2^11) data is extracted and the extraction interval is 2.048 seconds. Becomes

In FIG. 10, the fundamental frequency component is 70, the rotation frequency component is 71, the rotation frequency double component is 72, and the rotation frequency triple component is 73, and the alternate long and short dash line 75 indicates the spike-like waveform 51 generation shown in FIG. It's time. Therefore, the tool is in a normal state before the alternate long and short dash line 75 and one insert of the tool is in a defective state after the alternate long and short dash line 75.

Before and after the alternate long and short dash line 75, the amount of increase in the double rotational frequency component is larger than the amount of change in the rotational frequency component 71 and triple rotational frequency component 73. Therefore, by focusing on the double component 72 of the rotation frequency and setting the threshold value 74, it is possible to determine the state in which one insert is missing. In particular, in the frequency analysis result, it is possible to prevent an erroneous determination that the abnormality is detected in the normal state by selecting a frequency component in which the increase amount is larger after the loss occurs than in the normal state.

The number of data for frequency analysis is preferably 1024 (2^10) or more because it makes it easy to discriminate the n-fold component of the rotation frequency in the frequency spectrum distribution. In addition, if frequency analysis is performed on all the processing time data, the change in the frequency analysis result will be extremely small when the timing of occurrence of the defect is just before the end of the process, making it difficult to determine the defect. It is preferable to divide the signal every few seconds and perform frequency analysis, because the change in the frequency analysis result becomes large when abnormality occurs in the defect determination.

FIG. 11 is a cutting simulation result in a normal state where the tool has no defect, and FIG. 12 is a frequency spectrum distribution of the result. The processing conditions are the same as in FIG. 5, and in FIG. 11, the maximum value of the cutting force is normalized to 1. FIG. 11 is a simulation result of the cutting force, but the result of the spindle servo motor current value of FIG. 6 and the tendency of acting on each blade are the same. Further, in FIG. 11, since each blade of the four-blade tool acts uniformly, there is no variation in the amplitude of the cutting force. Therefore, in the frequency spectrum distribution of FIG. 12, only the fundamental frequency component 80 has an amplitude, and the rotational frequency component and its n-fold component have no amplitude.

FIG. 13 is a cutting simulation result assuming a state where one insert of a tool is missing, and FIG. 14 is a frequency spectrum distribution of the result. In FIG. 13, the cutting force having the same magnitude as in FIG. 11 is normalized to 1. In FIG. 13, the waveform corresponding to the loss of the insert is 81, and the waveform of the insert that acts next to the defective insert and the load is increased is 82. The waveform 81 is zero because the insert is completely missing and does not work, and the amplitude of the next insert is doubled because the load is doubled.

Further, in the frequency spectrum distribution of FIG. 14, the double component 85 of the rotation frequency is the largest. Although the magnitude of the amplitude is different from that of FIG. 9 which is the frequency spectrum distribution of the measurement signal in the state where one insert is missing, the increasing tendency of each frequency component is similar.

FIG. 15 is a difference in waveform of the cutting simulation result between a state where one insert is missing and a normal state. When this waveform is added to the waveform in the normal state in FIG. 11, it becomes the same as the waveform in FIG. 13 in which one insert is missing. FIG. 16 is a frequency spectrum distribution of the difference waveform. No amplitude is seen in the fundamental frequency, and the double component of the rotation frequency is large. Therefore, when one insert is missing, the fundamental frequency component does not change in the frequency spectrum distribution, and the n-fold component of the rotation frequency increases. Therefore, the n-fold component of the rotation frequency is increased with reference to the fundamental frequency component. By monitoring, it can be seen that the insert defect can be determined.

FIG. 17 shows that, in a tool having 3 to 10 blades, the cutting force is calculated by cutting simulation, and the frequency spectrum distribution in the normal state and the state in which one insert is missing is compared to find a frequency with a large amount of change in the rotation frequency. It is shown by the magnification against. Note that the number of blades that act simultaneously is one. By selecting the n-fold component of the rotation frequency shown here as the target of defect detection, it becomes possible to detect the defect with high accuracy, especially when the number of blades that act simultaneously is one. In the table of FIG. 17, the magnification with the largest change amount is shown as “change amount 1st”, and the magnification with the largest change amount after 1st change is shown as “change amount 2nd”. When the number of blades is 4, as shown in FIG. 14, twice (n=2) is preferable, but when the number of blades is 8, for example, 2 to 5 times (n=2 to 5) is almost the same. It is possible to calculate by simulation that characteristics of the degree are shown, and 6 times is preferable as the next point.

However, in actual machining, with a tool having a plurality of blades, the inserts do not work completely evenly due to the mounting error (deviation) of the insert, the dimensional tolerance of the insert, the rotational runout of the spindle, and the like. Therefore, even in a normal state where the insert is not missing, the amplitude of each blade varies as shown in FIG. 6, and as shown in FIG. 8, the rotation frequency component 61 and its n-fold components 62 and 63 have amplitudes. You will be able to see it.

For this reason, the spindle servo motor currents in the state where one insert is removed and in the normal state are measured in actual machining, and the frequency spectrum distributions in normal state and in the state where one insert is removed are compared to show the largest change. The accuracy of the defect determination can be improved by selecting the frequency component. In particular, if the machining shape is not constant, the degree of change in frequency component varies depending on the machining area (variation in machining area and machining location), so the signal used for determination is the range in which the selected frequency component varies greatly. It is preferable to limit it to.

This is because when the number of simultaneously acting blades is large, the resultant force of the plurality of blades (inserts) appears in the measurement signal, and the influence of the missing blades (inserts) is reduced. For the determination of defects, it is preferable to select a processing condition in which the number of blades that act simultaneously is small, and it is desirable that only one blade acts, because each waveform of the measurement signal reflects the processing state of each insert, which is the easiest to determine a defect. ..

Further, it has been confirmed that when the main shaft to which the tool is attached has large rotational runout, the rotational frequency component is large and the variation is large even when the tool is in a normal state. Since the rotation frequency component increases due to factors other than the loss, the determination frequency of the loss is not the rotation frequency component, but other frequency components (for example, double the rotation frequency component to (blade number-1) times the component It is preferable to use the frequency component with the largest variation.

It is also possible to use the rotation frequency component for the judgment.In that case, in order to reduce the risk of judging as abnormal during normal operation, it is possible to suppress the runout of the spindle and the variation in the mounting of the insert on the tool. It is sometimes necessary to take measures to reduce the rotational frequency component.

≪Control parameter setting process≫
FIG. 18 shows an example of a screen on which the user inputs control parameters before executing the abnormality detection processing of the cutting tool by the abnormality detection device 1. The control parameter setting unit 25 activated by the user displays the control parameter input screen 100 on the display unit 12, receives the control parameter input by the user according to the guide, and registers it in the control parameter 32 of the storage unit 3.

In the processing condition input field 101 of the control parameter input screen 100, the number of rotations [min ^-1 ]: the number of rotations of the spindle 7 and the number of tool blades: the number of blades of the cutting tool 5 are input.

In the measurement condition input field 102, measurement time [s]: signal measurement time from the start to the end of processing (always input), sampling time [s]: sampling of current value and voltage value by the measuring unit 14 Enter the period.

In the processing machine control method input field 103, the processing machine control 105: the command content (processing machine operation) that the processing machine command unit 24 outputs to the processing machine 4 when the cutting tool 5 has an abnormality is displayed by pressing the pull-down 108. Select from the displayed menu (tool change, tool evacuation, etc.) and enter.

In the measurement signal processing method input field 104, a pre-processing method 106: a pre-signal processing for a measurement signal such as a voltage value and a current value executed by the signal processing unit 22 is displayed by pressing the pull-down 109 (low pass). Filter, high pass filter, band pass filter, moving average filter, etc.) and input.
Data extraction start time [s]: The signal processing unit 22 records the measurement signal during processing in the measurement signal 33 of the storage unit 3, and the target for executing the abnormality detection process from the measurement signal 33 is the number of simultaneously acting blades. The data extraction start time (designated by the elapsed time from the start time of the measurement time) for reading the measurement signal of the processing region on the processing path predicted to be one processing state is input as much as possible.
Number of frequency analysis data [pieces]: When performing the fast Fourier transform of frequency analysis, the digital signals (measurement signals) for each sampling time are grouped together by a power of 2 and spectrum analysis is performed. Enter the total number of data.
Frequency analysis frequency [number]: Input the number of times to continuously repeat the process of performing fast Fourier transform of the measurement signal of the frequency analysis data collectively in one time.
Judgment frequency (rotation frequency n times) 107: Input the judgment frequency (specify “n” value of n times the rotation frequency) in the frequency spectrum distribution adopted to judge the loss of the cutting tool. Alternatively, the actual frequency may be input.
Threshold: Amplitude (normalized value) of the judgment frequency used to judge the loss of the cutting tool is an appropriate threshold that is expected to exceed after the loss of the tool (remove one insert of the corresponding cutting tool on the processing machine. It is desirable to actually measure and determine in advance, but it may be determined with reference to empirical values under similar processing conditions in the past). In this embodiment, when the frequency spectrum distribution is calculated, the amplitude of the fundamental frequency is normalized to 1 and the amplitude of each frequency is normalized. Therefore, the threshold value specifies a value corresponding to the normalized amplitude. However, in the operation in which the amplitude of the fundamental frequency is not normalized to 1 when calculating the frequency spectrum distribution, the threshold value specifies the value corresponding to the unnormalized amplitude.

≪Abnormality detection process≫
FIG. 19 is an example of a flowchart of abnormality detection processing executed by the abnormality detection processing unit 21 of the abnormality detection device 1. Prior to this processing, the control parameter 32 is registered by the control parameter setting unit 25.

In step S10, the NC device 10 of the processing machine 4 issues a processing start signal at the start of processing of the work material 6, and the abnormality detection processing unit 21 is activated in response to the signal. The measuring unit 14 also starts measurement at the start of processing and outputs a measurement signal (digital signal) of a current value or power value obtained between the servo amplifier 9 of the processing machine 4 and the spindle motor 8.
The measurement of the measuring unit 14 is started based on the operation of the processing machine 4. A sensor may be attached to the holder of the target tool to detect the rotation of the tool, or information of the NC program may be monitored to read a specific character string or the like to be used as the measurement start trigger.

In step S11, the signal processing unit 22 is activated. The signal processing unit 22 continuously inputs the measurement signal from the measurement unit 14 thereafter, performs the pre-signal processing specified by the pre-processing method 106, and outputs the measurement signal to the measurement signal 33 storage area of the storage unit 3 in time series. To record.

In step S12, the data set in advance as the control parameter is the time to reach the corresponding machining path so that only the measurement signal of the machining path in which the number of simultaneous working blades is one is in the machining state. Read the extraction start time. From the measurement signal of the data extraction start time, it waits until the measurement signal of the frequency analysis data number is stored in the measurement signal 33 storage area of the storage unit 3. If the measurement signals of the number of frequency analysis data are already stored, the waiting is not necessary.

In step S13, the analysis determination unit 23 is activated, and the analysis determination unit 23 determines the number of data (frequency analysis data number) used to perform one fast Fourier transform from the measurement signal 33 storage area of the storage unit 3. The measurement signal is extracted and the frequency spectrum is calculated. Each frequency component is normalized by the standard for normalizing the fundamental frequency component to 1.

In step S14, the analysis determination unit 23 determines whether the amplitude of the determination frequency of the frequency spectrum exceeds the threshold value. If it exceeds, the process proceeds to S16. If not exceeded, the process proceeds to S15.

In step S15, if the number of times of frequency analysis to be continuously performed designated by the number of times of frequency analysis of the control parameter is plural and there is still frequency analysis to be performed, the process proceeds to step S12. When all the frequency analyzes performed continuously are completed, the process proceeds to S17. However, with the intention of performing frequency analysis again on the remaining machining path, if a plurality of data extraction start times are input in the control parameter setting process, and a plurality of frequency analysis times are also input correspondingly, If there is a next frequency analysis on the remaining machining path, the process proceeds to S12.

In step S12, if there is the next frequency analysis in the frequency analysis performed continuously, the number of frequency analysis data used in the next frequency analysis is added to the measurement signal data used in the frequency analysis performed previously. It waits until the measurement signal is stored in the measurement signal 33 storage area of the storage unit 3. If the measurement signals of the number of frequency analysis data are already stored, the waiting is not necessary.

In step S16, the processing machine command unit 24 is activated, the processing machine command unit 24 creates a command to be output to the processing machine 4 designated by the processing machine control 105 of the control parameter, and outputs the command to the NC device 10.
Specific examples of control of the processing machine include replacement with another cutting tool 5, or withdrawal of the cutting tool 5 to stop processing. In the case of a processing machine equipped with an automatic tool changer, a plurality of cutting tools 5 are installed in the automatic tool changer, and after the abnormality is detected, the cutting tool 5 is controlled to be replaced with another cutting tool 5, thereby making The operation can be continued.

In step S17, the signal processing unit 22 is stopped.
In step S18, the abnormality detection processing unit 21 is stopped.

If it is determined that the cutting tool is missing, the abnormality detection processing outputs a command such as cutting tool replacement to the processing machine 4 and ends the processing. When restarting the machining after exchanging the cutting tool on the processing machine 4 side, a signal for resuming the machining is sent from the NC device 10 and the abnormality detection processing unit 21 of the abnormality detection device 1 is restarted. It will be. In that case, the measurement time is restarted from the time when the previous processing was interrupted.

Although the example in which the current value is used as the measurement signal has been described above, the same processing may be performed for the power value. The unit of the parameter is changed from the electric current value [A] to the electric power value [W], and otherwise the same.

According to this embodiment, since it is possible to detect the state where one insert of the tool is damaged, it is possible to suppress the damage expansion to the tool body and the occurrence of continuous machining defects. Therefore, it is possible to reduce the tool cost and the machining defects by suppressing the damage. It is possible to obtain the effect such as reduction of the cost of measures related to. For example, when a shoulder tool having a width of 6 mm, a depth of 5 mm, and a length of 200 mm is repeatedly cut with a stainless casting material with a cutting tool having a tool diameter of 40 mm and a number of blades of 4, a tool cost of about 50% is applied by applying this embodiment. It is possible to improve the number of products produced by 10% by reducing the number and reducing the setup time after tool breakage.

In this embodiment, the cutting tool 5 has been described as a type of cutting tool called a throw-away end mill capable of exchanging a plurality of cutting edges (inserts). The same thing can be explained with a cutting tool of a type called a solid end mill in which a plurality of cutting edges and a shank (handle portion of the end mill) are integrated. When this solid end mill is used to perform an abnormality detection process and detect a state in which one cutting edge is damaged, it can be immediately removed, re-polished, and re-coated for use.

In the present embodiment, level correction is performed using the amplitude of a specific frequency at a predetermined timing after the start of processing as a reference value, and a tool abnormality is detected by an increase in the level correction value of the amplitude of the determination frequency of each processing after the level correction. An example of the abnormality detection device will be described.

FIG. 20 is a block diagram illustrating the processing machine 4 including the abnormality detection device 111 for a cutting tool according to the second embodiment. This is a configuration in which a cutting tool abnormality detection device 111 is attached to the processing machine 4.

The abnormality detection device 111 is different from the abnormality detection device 1 described in the first embodiment in that the tool holder 160 having a tool information storage unit, the tool information acquisition sensor 112, the insert replacement detection unit 113, and a predetermined timing determination. The configuration is such that an amplitude storage unit 114 and a storage unit 115 for the level correction value of the determination amplitude of the immediately preceding machining are added.

The tool information acquisition sensor 112 acquires tool information from the tool holder 160 that has a tool information storage unit that holds the cutting tool 5. The tool holder 160 having the tool information storage unit is provided with an IC code chip or the like that stores the tool information. The insert replacement detection unit 113 detects the replacement of the insert from the information acquired by the tool information acquisition sensor 112.

In the IC code chip of the tool holder 160 provided with the tool information storage unit, by rewriting a specific code composed of numbers or the like at the time of insert replacement, the insert replacement detection unit 113 detects it and determines replacement of the insert. be able to.

The determination amplitude storage unit 114 of the predetermined timing provided in the storage unit 3 stores the amplitude of the determination frequency in the machining immediately after the insert replacement, based on the insert replacement signal. In the analysis determination unit 23, the level correction is performed on the basis of the stored amplitude of the determination frequency immediately after the insert replacement, which is stored, with respect to the amplitude of the determination frequency calculated in each processing that is continuously performed before the next insert replacement. To do. Then, the abnormality of the cutting tool 5 is determined by the increase of the level correction value of the amplitude of the determination frequency after the level correction. This level correction processing suppresses the influence of mounting variation that occurs in the amplitude of the determination frequency when the insert is replaced.

When performing test processing after insert replacement, the timing for storing the amplitude of the determination frequency may be set at a predetermined timing, such as when actually starting processing of the target product, not immediately after insert replacement. Good.

The storage unit 115 for storing the level correction value of the determination amplitude of the previous processing, which is included in the storage unit 3, stores the level correction value of the amplitude of the determination frequency after the level correction in the previous processing. The analysis determination unit 23 calculates the difference between the level correction value of the amplitude of the determination frequency after level correction in the next machining and the stored data, and determines the abnormality of the cutting tool 5 based on the increase in the difference.

FIG. 21 is a frequency spectrum distribution of a spindle servo motor current in machining a three-blade tool. This is an example of shoulder cutting with a width of 6 mm and a depth of 3 mm under the conditions of a cutting speed of 100 m/min and a blade feed amount of 0.15 mm/tooth. 121 is a rotation frequency component, 122 is a fundamental frequency component, and the fundamental frequency component 122 is normalized to 1. This result is immediately after the insert is replaced, but is the determination frequency of the three-flute tool (from the table of FIG. 17, the determination frequency is the rotation frequency) due to variations in the mounting of the insert, rotational runout of the spindle, and the like. The amplitude of the rotation frequency is often large from the beginning of using the insert.

FIG. 22 is a graph showing the change in the amplitude of the determination frequency of the three-blade tool with respect to the cumulative number of processes. This is an example in which a stainless cast material is used as a work material and is processed by a three-blade exchangeable end mill with a diameter of 20 mm. The width of 6 mm, the depth of 3 mm and the length of 200 mm is counted as one piece of shoulder milling. Further, the tool life was set to 30 and the inserts were replaced after every 30 machining (specifically, the replacement (switching) of the corner portion that becomes the cutting edge, which will be referred to as insert replacement hereinafter).

A range indicated by an arrow in the drawing represents processing in the same corner portion of the insert (hereinafter, each range indicated by an arrow is called processing of the insert 123, the insert 124, and the insert 125). "○" in the figure indicates the first machining immediately after the insert replacement.

In the processing of the insert 124, the amplitude of the determination frequency is larger than that of the other inserts from the first (124a) immediately after the replacement of the insert. In the result of this figure, no defect has occurred in any of the inserts, but in order to prevent the insert 124 from being detected as a tool abnormality, it is necessary to set the threshold value 126 to be larger than the amplitude of the insert 124. Therefore, when the amplitude of the determination frequency is small like the inserts 123 and 125, even if a defect occurs, there is a possibility that an abnormality cannot be detected without exceeding the threshold value.

FIG. 23 shows a coordinate axis whose vertical axis has the same unit as the amplitude unit of the determination frequency represented by the vertical axis of FIG. 22 and whose range is expanded to a negative region, and the first machining determination of each insert is performed. It is a graph which shows the change of the level correction value of the amplitude of the judgment frequency of the 3-flute tool which level-corrected so that the amplitude of frequency may be matched with 0 of the vertical axis with respect to the cumulative number of machining. This is the following two correction amounts that are level-corrected so that the amplitude of the first judgment frequency of machining in each insert becomes zero with respect to the amplitude of the respective judgment frequency of machining of each insert shown in FIG. The same amount of level correction (the above correction amount is subtracted) is applied to the amplitude of the determination frequency of each process after the eye. The level correction values of the amplitude of the determination frequency are gathered near zero, and the threshold value 127 (hereinafter, referred to as the level-corrected threshold value) is set closer to the data of the insert 123 and the insert 125 as compared with the threshold value 126 of FIG. be able to.

As a result, even if the increase in the amplitude of the determination frequency is small when the insert defect occurs, the detection can be performed. This tool abnormality determination method is particularly effective when the rotation frequency is used as the determination frequency. For level correction, the amplitude of the determination frequency of the first machining of the insert is stored as a reference value for level correction, and the level correction value is obtained by subtracting the reference value for level correction from the subsequent amplitude. Is an example. Furthermore, the "first" is an example of the first processing. In the work defined by the processor, the "second" processing or the first processing may be performed as long as it is considered to be the first processing after the insert replacement. For example, the "first" processing is a trial processing for confirming the state of the tool, and the "second" processing is a case where an actual target product is processed. The reference value may use the amplitude at another time point.

FIG. 24 is a graph showing changes in the level correction value of the amplitude of the determination frequency of the three-blade tool, which is level-corrected on the basis of matching the amplitude of the determination frequency of the first processing with 0 on the vertical axis with respect to the number of processes. is there. This is a result of an insert different from the insert shown in FIG. 23 (specifically, another corner portion). Although no data exceeding the threshold 127 after the level correction is found, it is an example in which the cutting tool is confirmed to be defective after noticing that the machined surface is scratched after the tenth machining (128). The scratches on the machined surface are generated from the sixth machining (129), and it is highly probable that the cutting tool has a defect at this point.

In the case of FIG. 24, the level correction value of the amplitude of the determination frequency greatly increases in the sixth processing as compared with the fifth processing, but does not exceed the threshold 127 after level correction. It is presumed that this is because the level correction value of the amplitude of the determination frequency tends to decrease as the number of machining increases from the first machining to the fifth machining.

As this phenomenon, there is one that was particularly protruded and fixed among the three inserts at the time of insert replacement, and as the number of processing increases, the wear of the protruding insert progresses, and It is considered that the action was equalized and the rotation frequency component (level correction value of the amplitude of the determination frequency) decreased. Then, it is estimated that the wear was increased in the sixth machining, and the insert was chipped.

FIG. 25 shows the level correction value of the amplitude of the determination frequency for the first machining of the three-blade tool shown in FIG. 24, the difference value between the current and previous values, and the difference value as the coordinate on the vertical axis. It is a graph plotted for each processing. Here, the level correction value of the first processing is taken as a difference from its own level correction value, and the coordinate on the vertical axis is set to 0. In calculating the difference value, a past level correction value other than the “previous” level correction value may be used. This is because there may be a case where it is not preferable to use the previous level correction value for some reason such as elimination of an abnormal value or accuracy of numerical calculation.

In the case of FIG. 25, the difference value is significantly increased by the sixth (130) processing in which the occurrence of insert defect is estimated. By detecting the abnormality by exceeding this with respect to the threshold 131 (hereinafter referred to as the difference threshold), even if the abnormality occurs after the amplitude of the rotation frequency decreases as the machining progresses, the abnormality of the tool can be detected. Is possible. In addition, here, the difference from the level correction value of the immediately preceding processing is shown by a real number including a negative value, but evaluation by an absolute value is also effective.

The tool abnormality determination methods described so far can be used not only individually but also in combination. The usage examples and characteristics of each judgment method are shown.

The method of determining the abnormality of the tool by directly determining the amplitude of the determination frequency shown in the first embodiment (hereinafter, referred to as the method of directly determining the abnormality of the tool by the amplitude of the determination frequency) is the rotation frequency. It is useful for cutting tools with 4 or more blades that are not used for. Further, it is characterized in that data acquisition and processing (storage and calculation) are easier than other determination methods. This technique is used alone.

The correction amount level-corrected so that the amplitude of the determination frequency of the first machining is zero, which is shown in the second embodiment, is also applied to the amplitudes of the determination frequencies of the subsequent second and subsequent processes. A method of determining a tool abnormality when a threshold value is exceeded using a level correction value that has undergone level correction (hereinafter referred to as a tool abnormality determination method based on level-corrected data) is a three-piece method that uses a rotation frequency as a determination frequency. It is particularly effective for cutting tools with blades or less, and there is no problem when applied to cutting tools with four or more blades. Further, this method is effective for detecting the abnormality of the tool in the case where the abnormality of the tool gradually progresses with the progress of machining and finally becomes large. For example, it is possible to detect the case where the wear of the insert gradually progresses and finally the chip becomes large, or the screw for fixing the insert is gradually loosened and becomes large.

The difference between the level correction value of the amplitude of the determination frequency of each processing, which is level-corrected based on the amplitude of the determination frequency of the first processing, and the difference between the level correction value of the immediately previous processing, which will be described later in the second embodiment. The method of calculating and using the difference to determine a tool abnormality when the threshold value is exceeded (hereinafter referred to as the difference data determination method) is most likely to reflect changes in the cutting edge, and can detect relatively small defects. Is. Therefore, it is effective for all tools, including tools that use the rotation frequency as the determination frequency.

In addition, in the case where the cutting edge gradually wears (changes) and its accumulation increases and becomes abnormal, the tool abnormality determination method based on the level-corrected data is more suitable than the determination method based on the difference data. Therefore, by using these two determination methods together, it is possible to detect tool abnormalities in more various cases.

≪Judgment result display≫
FIG. 26 is a first display example of the level correction value of the amplitude of the determination frequency after the level correction on the display unit 12. In the level correction value display area 139 of the amplitude of the determination frequency after level correction provided in the display unit 12, the level correction value of the amplitude of the determination frequency of each machining is level-corrected based on the amplitude of the determination frequency of the first machining. In this example, 141 and the level-corrected threshold value 140 are displayed for cumulative processing. The level correction values 142a, 142b, 142c of the amplitude of the first processing determination frequency immediately after the insert replacement are displayed as white circles “◯” to distinguish them from other data. It is possible to confirm the transition of the level correction value of the amplitude of the determination frequency of the insert (142c) that is currently being machined and the inserts (142a, 142b) that have been replaced up to two inserts before that.

FIG. 27 is a second display example of the level correction value of the amplitude of the determination frequency after the level correction on the display unit 12. This is an example in which the cumulative machining number of the first machining immediately after the insert replacement is displayed by arrows 143a, 143b, 143c.

FIG. 28 is a third display example of the level correction value of the amplitude of the determination frequency after the level correction on the display unit 12. This is an example in which the cumulative machining number of the first machining immediately after the insert replacement is displayed by alternate long and short dashed lines 144a, 144b, 144c.

FIG. 29 is a fourth display example of the level correction value of the amplitude of the determination frequency after the level correction on the display unit 12. It is an example in which the marker is switched and displayed every time the insert is replaced. 145, 146, and 147 are the results of different inserts, and the insert can be discriminated by the type of marker.

FIG. 30 is an example in which the cumulative number of processes is increased in the fourth display example of the level correction value of the amplitude of the determination frequency after the level correction on the display unit 12. In the example of setting the data for 3 times of the set tool life of 3 times to be displayed at one time, when the data of 3 times is exceeded, the cumulative machining number on the horizontal axis is shifted by the set tool life (30 pieces) and displayed. is there. The data of the insert 146 displayed in the center of FIG. 29 is shifted to the left end in FIG. 30, and the data of the insert 148 is displayed with the same marker as the insert 146.

FIG. 31 is a fifth display example of the level correction value of the amplitude of the determination frequency after the level correction on the display unit 12. The horizontal axis indicates the number of machining, and is displayed for each insert (specifically, for each corner part that becomes a cutting edge). In this example, the same data as the insert 145 of FIG. 29 is displayed.

26 to 31, in each case, the reference value for level correction when level correction is performed so that the amplitude of the first determination frequency for machining is adjusted to 0 on the vertical axis for each insert replacement It is possible to confirm the transition of the level correction value calculated by applying it to the amplitude of the judgment frequency of each processing. Although not shown here, the result of the determination method based on the difference data is also displayed so that it can be confirmed each time the insert is replaced.

It should be noted that the display example shown here is merely an example, and there is no limitation on the markers to be displayed, broken lines, etc., and the number of data to be displayed at one time. Further, in the present embodiment, the unit of processing is indicated by the number of processing (pieces), but if it is increased by the progress of processing, it may be replaced with the cutting volume or the cutting distance.

≪Control parameter setting process≫
FIG. 32 shows a screen example in which the user inputs control parameters before executing the abnormality detection processing of the cutting tool by the abnormality detection device 111 of the second embodiment. FIG. 32 shows a measurement signal processing method input field 150 obtained by expanding the measurement signal processing method input field 104 in the control parameter input screen 100 of the first embodiment shown in FIG. 18 for the second embodiment.

In the measurement signal processing method input field 150, in addition to the threshold value input field 151 of the method of directly determining the tool abnormality by the amplitude of the determination frequency of the first embodiment, there is a threshold value input field 152 after level correction of the second embodiment. A difference threshold input field 153 is added. The threshold value of the tool abnormality determination method based on the level-corrected data is input to the threshold value input field 152 after level correction. Further, the threshold value of the determination method based on the difference data is input in the difference threshold value input field 153.

Note that it is better to limit the input fields of each threshold to the thresholds of the determination method used, but even if an input field of thresholds that are not used is provided, if a specific numerical value is input, the determination will be made with that threshold. If you do not use this method, there is no problem even if there is an input field.

This is an example of the setting in which, when 0.0 is input in the threshold value input field of the measurement signal processing method input field 150 shown in FIG. 32, the method of determining with the threshold value is not used. The threshold value input field 151 is a threshold value for the method of directly determining the tool abnormality by the amplitude of the determination frequency. Since 0.0 is input, this method is set not to be used.

In the first embodiment, an example in which the servo motor current or power is measured by the measuring unit 14 is shown. However, depending on the processing machine, a PC (personal computer) may be connected via an Ethernet (registered trademark) line to perform servo control by a measurement application. The motor current can be acquired.

That is, the measurement unit 14 of FIG. 1 is configured to include the measurement unit in the calculation unit 2 of the abnormality detection device body 15. There is no problem in using this measurement result for the measurement signal. Also, if the measurement application has a signal processing function such as a filter, there is no problem in using that function.

Further, in the second embodiment, an example in which the tool information acquisition sensor 112 and the insert replacement detection unit 113 detect the timing of insert replacement is shown. However, depending on the processing machine, this may also be a PC (personal computer) via an Ethernet line. It is possible to connect and acquire the machining number of the tool directly from the machining machine, and to determine the insert replacement from the information.

That is, the insert replacement detection unit 113 of FIG. 20 is configured to include the insert replacement detection unit in the calculation unit 2 of the abnormality detection device body 15. There is no problem even if the result of insert replacement judgment based on the acquired number of processes is used for the insert replacement signal.

Further, in the first and second embodiments, the fundamental frequency component is normalized to 1 in the frequency spectrum distribution, and the n-fold component of the rotation frequency designated by the control parameter is threshold-determined, but it is not always normalized to 1. There is no need to do so, and the determination may be made based on the ratio or the like, and it is not necessarily limited to the one having all the configurations described.

As described above, in the first to third embodiments, the following has been described.

Anomaly detection method to detect the loss of cutting tools equipped with multiple blades,
(A) receiving a measurement signal that is a measurement result of a current value or a power value of a spindle motor that rotates a spindle of a machine tool on which the cutting tool is mounted,
(B) The measurement signal is filtered and stored in time series in the storage unit,
(C) Abnormality detected,
(D) When it is determined by the abnormality detection of (C) that there is a tool loss, a processing machine control command designated by a control parameter is output to the machine tool,
The abnormality detection in (C)
(C1) From the stored measurement signal, a frequency spectrum is calculated by extracting a measurement signal of the number of frequency analysis data of the machining region in which the number of simultaneously acting blades is predicted to be 1.
(C2) Based on the amplitude of the determination frequency (=rotational frequency×n, where 1≦n<number of tool flutes, n is an integer) specified by the control parameter in the frequency spectrum, and the threshold value specified by the control parameter It is determined that the tool is missing.

The determination frequency may be a value that specifies the frequency of the next or maximum change in the frequency spectrum when the cutting tool changes from the normal state to the state where one blade is missing.

In the abnormality detection of (C), the data extraction start time is specified by the control parameter,
The data extraction start time may be used as a control parameter for frequency-analyzing the measurement signal in the processing region where the number of simultaneously acting blades is one in the measurement time based on the processing start signal emitted from the NC device.

In addition, the determination of (C2) that the tool is missing is
(C2a) The amplitude of the determination frequency at a predetermined processing timing is stored as a level correction reference value,
(C2b) Level correction is performed by subtracting the level correction reference value from the amplitude calculated in each subsequent process to obtain a level correction value,
(C2c) It may be determined that the tool is missing based on the level correction value and the threshold value.

The level correction reference value is the amplitude of the determination frequency in the first machining after replacing the blade of the cutting tool,
In each subsequent machining, the determination of (C2c) that the tool is missing is
(C2c1) A difference value between the level correction value calculated based on the amplitude and a past value is calculated,
(C2c2) It may be determined that there is a tool defect based on the calculated difference value and the threshold value.

It should be noted that the change in the level correction value according to the cumulative machining number is displayed on the (E) display unit so that the data for each blade replacement can be discriminated, or the (E) display unit displays the machining. The change in the level correction value and the change in the difference value according to the number may be displayed so that the data for each blade replacement can be discriminated.

In the above description, the above-described abnormality detection method is realized by hardware (for example, an abnormality detection device or a computer), a program for executing the abnormality detection method by a computer, the hardware and the machine tool, and the like. The system configuration including is also described. The program may be distributed by being stored in a computer-readable storage medium (a medium of the same type as the storage unit 3 may be used). The program may be distributed by a program distribution computer. The program distribution computer reads a program from the storage medium storing the program (a medium of the same type as the storage unit 3 may be used) and the program for distributing the program to other computers via a network interface. It may have a unit (may be the same type as the arithmetic unit 2) and a network interface.

DESCRIPTION OF SYMBOLS 1... Abnormality detection device, 2... Calculation part, 3... Storage part, 4... Processing machine, 5... Cutting tool, 6... Work material, 7... Spindle, 8... Spindle motor, 9... Servo amplifier, 10... NC device , 11... Input unit, 12... Display unit, 13... Communication unit, 14... Measuring unit, 15... Abnormality detection device body, 21... Abnormality detection processing unit, 22... Signal processing unit, 23... Analysis determination unit, 24... Processing Machine command unit, 25... Control parameter setting unit, 31... Cutting tool abnormality detection program, 32... Control parameter, 33... Measurement signal,
41... Insert, 42... Corner, 43... Tool body, 44... Fixing screw, 46... Damage, 50... Measurement signal, 51... Spike-shaped waveform, 52... Range before tool breakage, 53... Range after tool breakage , 54... Waveform with small amplitude, 55... Waveform with large amplitude, 60... Basic frequency component, 61... Rotation frequency component, 62... Double frequency component, 63... Triple frequency component, 64... Threshold value, 65 ... fundamental frequency component, 66... rotation frequency component, 67... rotation frequency double component, 68... rotation frequency triple component, 70... fundamental frequency component, 71... rotation frequency component, 72... rotation frequency double component, 73... Triple frequency component of rotation frequency, 74... Threshold value, 75... Dash-dotted line, 80... Fundamental frequency component, 81... Waveform corresponding to loss of insert, 82... Waveform of insert that bears machining of defective insert, 83... Fundamental frequency component, 84... Rotation frequency component, 85... Rotation frequency double component, 86... Rotation frequency triple component, 91... Rotation frequency component, 92... Rotation frequency double component, 93... Rotation frequency triple Component, 100... Control parameter input screen, 101... Machining condition input field, 102... Measurement condition input field, 103... Machine control method input field, 104... Measurement signal processing method input field, 105... Machine control (control parameter) , 106... Pre-processing method (control parameter), 107... Judgment frequency (rotation frequency n times) (control parameter), 108... Processing machine control method item selection pull-down, 109... Pre-processing method item selection pull-down, 110... Tool Holder, 111... Abnormality detection device, 112... Tool information acquisition sensor, 113... Insert replacement detection unit, 114... Storage unit for determination amplitude of predetermined timing, 115... Storage unit for level correction value of determination amplitude of previous machining , 121... Rotation frequency component, 122... Basic frequency component, 123... First insert corner machining range, 124... Second insert corner machining range, 125... Third insert corner machining range, 126... Threshold, 127... Threshold value after level correction, 128... Level correction value for 10th processing, 129... Level correction value for 6th processing, 130... Difference from previous processing data for 6th processing, 131 ... Threshold of difference, 139... Amplitude display area of determination frequency after level correction, 140... Threshold after level correction, 142... First processing data immediately after insert replacement, 143... First processing immediately after insert replacement Arrow mark indicating 144... One point indicating the first machining immediately after insert replacement Chain line, 145... First insert machining range, 146... Second insert machining range, 147... Third insert machining range, 148... Fourth insert machining range, 150... Measurement signal processing method input field , 151... Threshold input field for direct determination of amplitude of determination frequency, 152... Threshold input field after level correction, 153... Difference threshold input field, 160... Tool holder provided with tool information storage section

Claims (17)

An abnormality detection device for detecting a loss of a cutting tool having a plurality of blades,
A measuring unit that measures a current value or a power value of a spindle motor that rotates a spindle of a machine tool equipped with the cutting tool,
A signal processing unit that filters the measurement signal from the measurement unit and stores it in time series in the storage unit, and
An abnormality detection processing unit,
When it is determined that there is a tool loss in the abnormality detection processing unit, a processing machine command unit that outputs a processing machine control command specified by a control parameter to the machine tool,
Equipped with
The abnormality detection processing unit,
From the stored measurement signal, the frequency spectrum is calculated by extracting the measurement signal of the machining region in which the number of simultaneously acting blades is predicted to be 1,
In the frequency spectrum, based on the amplitude of the determination frequency (=rotation frequency×n, where 1≦n<the number of tool flutes, n is an integer) specified by the control parameter, and the threshold value specified by the control parameter, It is judged that there is a tool loss,
An abnormality detection device for a cutting tool, which is characterized in that The number of blades of the cutting tool is 4,
The determination frequency is twice the rotation frequency of the cutting tool,
The abnormality detection device for a cutting tool according to claim 1, wherein. The number of edges of the cutting tool is 8,
The determination frequency is at least one of 2 to 5 times the rotation frequency,
The abnormality detection device for a cutting tool according to claim 1. The determination frequency is a value that specifies the maximum or next maximum frequency change in the frequency spectrum when the cutting tool changes from a normal state to a state where one blade is missing,
The abnormality detection device for a cutting tool according to claim 1, wherein. In the abnormality detection processing unit, the data extraction start time is specified by the control parameter,
The data extraction start time is a control parameter for frequency-analyzing the measurement signal in the processing region where the number of simultaneously acting blades is one in the measurement time based on the processing start signal emitted from the NC device.
The abnormality detection device for a cutting tool according to claim 1, wherein. The abnormality detection processing unit,
Stores the amplitude of the judgment frequency at the predetermined processing timing as the level correction reference value,
From the amplitude calculated in each subsequent process, a level correction value is obtained by performing a level correction by subtracting the level correction reference value,
Based on the level correction value and the threshold value, it is determined that the tool is missing.
The abnormality detection device for a cutting tool according to claim 1, wherein. The level correction reference value is the amplitude of the determination frequency in the first machining after replacing the blade of the cutting tool,
In each subsequent processing, the abnormality detection processing unit,
Calculate the difference value with the past value of the level correction value calculated based on the amplitude,
Based on the calculated difference and the threshold value, it is determined that the tool is missing.
The abnormality detection device for a cutting tool according to claim 6. The change in the level correction value according to the cumulative number of machining is displayed on the display unit so that the data for each blade replacement can be discriminated.
The abnormality detection device for a cutting tool according to claim 6, wherein. A change in the level correction value and a change in the difference value according to the number of machining are displayed on the display unit so that the data for each blade replacement can be discriminated.
The abnormality detection device for a cutting tool according to claim 7, wherein. An abnormality detection method for detecting a loss of a cutting tool having a plurality of blades,
(A) receiving a measurement signal that is a measurement result of a current value or a power value of a spindle motor that rotates a spindle of a machine tool on which the cutting tool is mounted,
(B) The measurement signal is filtered and stored in time series in the storage unit,
(C) Abnormality detected,
(D) When it is determined by the abnormality detection of (C) that there is a tool loss, a processing machine control command designated by a control parameter is output to the machine tool,
The abnormality detection in (C)
(C1) From the stored measurement signal, a frequency spectrum is calculated by extracting a measurement signal of the number of frequency analysis data of the machining region in which the number of simultaneously acting blades is predicted to be 1.
(C2) Based on the amplitude of the determination frequency (=rotational frequency×n, where 1≦n<number of tool flutes, n is an integer) specified by the control parameter in the frequency spectrum, and the threshold value specified by the control parameter Determine that the tool is missing,
A method for detecting an abnormality in a cutting tool, which is characterized in that The determination frequency is a value that specifies the maximum or next maximum frequency change in the frequency spectrum when the cutting tool changes from a normal state to a state where one blade is missing,
The method for detecting abnormality of a cutting tool according to claim 10, wherein. In the abnormality detection of (C), the data extraction start time is specified by the control parameter,
The data extraction start time is a control parameter for frequency-analyzing the measurement signal in the processing region where the number of simultaneously acting blades is one in the measurement time based on the processing start signal emitted from the NC device.
The method for detecting abnormality of a cutting tool according to claim 10, wherein. The determination of the tool loss in (C2) is
(C2a) The amplitude of the determination frequency at a predetermined processing timing is stored as a level correction reference value,
(C2b) Level correction is performed by subtracting the level correction reference value from the amplitude calculated in each subsequent process to obtain a level correction value,
(C2c) Based on the level correction value and the threshold value, it is determined that the tool is missing.
The method for detecting abnormality of a cutting tool according to claim 10, wherein. The level correction reference value is the amplitude of the determination frequency in the first machining after replacing the blade of the cutting tool,
In each subsequent machining, the determination of (C2c) that the tool is missing is
(C2c1) A difference value between the level correction value calculated based on the amplitude and a past value is calculated,
(C2c2) Based on the calculated difference value and the threshold value, it is determined that a tool is missing.
The method for detecting an abnormality of a cutting tool according to claim 13. (E) A change in the level correction value according to the cumulative number of processes is displayed on the display unit so that data for each blade replacement can be discriminated.
The method for detecting an abnormality of a cutting tool according to claim 13, wherein. (E) The display unit displays the change in the level correction value and the change in the difference value according to the number of machining operations so that the data for each blade replacement can be discriminated.
The method for detecting abnormality of a cutting tool according to claim 14, wherein. A program that causes a computer to execute the method according to any one of claims 10 to 16.

Images