JPH0463662A - Tool abnormality detecting device - Google Patents

Abstract

PURPOSE:To detect a tool defect regardless of cutting conditions by comparing the average value of the frequency spectrum with a preset value, and detecting a tool abnormality based on the compared result. CONSTITUTION:The frequency spectrum in the preset frequency range at the low-frequency area of the detected value of the first detecting means is obtained by a frequency analyzing means. The frequency spectrum of the frequency component corresponding to the rotating speed S of a main spindle detected by the second detecting means of the frequency spectrum, its high-harmonic frequency component and near these frequency components is removed from the frequency spectrum, and the average value of the frequency spectrum after the removal is obtained by a calculating means 70. The calculated value of the calculating means 70 is compared with a preset value, and a tool abnormality is detected by a tool abnormality detecting means based on the compared result. A tool abnormality signal is inputted to an NC control device 20, and an NC machine tool 10 is forcefully stopped.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a tool abnormality detection device that automatically detects sudden tool breakage during machining.

[Conventional technology]

Conventional methods for detecting tool abnormalities include, for example, the following methods.

■If a tool breaks during machining, the cutting resistance increases, which puts a load on the machine's feed motor, causing the motor to lose power.
The motor current value to be supplied increases.

This method utilizes this phenomenon and monitors the feed motor current value, and if this current value exceeds a predetermined set value, it is determined that the tool is abnormal.

■It uses a high frequency signal (AE multiplied signal acoustic emission signal) generated when an object breaks, and monitors the AE multiplied signal generated when a tool breaks.
A method of determining a tool abnormality when this AE multiplied signal exceeds a predetermined threshold (Japanese Patent Laid-Open No. 59-227354).

■Set a vibration sensor on the tool post, analyze the output of the vibration sensor in a predetermined octave range, and normalize each band component at the component level of the band that gives the maximum band component of vibration during normal cutting. The band that gives the maximum band component during normal cutting and multiple bands before and after it are compared with those during normal cutting for each band, and if there is one or more bands where the difference in each band exceeds a certain value, the tool will be damaged. How to determine that has occurred (
(Japanese Unexamined Patent Publication No. 57-20625).

[What the invention tries to solve! lA8] However,
In method (2) above, depending on the degree of tool chipping, the feed motor current value may not increase, and in this case, tool abnormality cannot be detected. In addition, this method (2) has the problem that it cannot be applied to machining with low cutting resistance, such as finishing machining.

Furthermore, in the method (2) above, when chips or the like break, an AE multiplied signal is generated, which causes a problem of many false detections.

Furthermore, with method (■) above, when machining interrupted machining or when machining a workpiece with a change in thickness (black skin material), it is not possible to distinguish between the vibrations generated by these machining processes and the vibrations caused by tool breakage. First, there is the problem that there are many false positives.

This invention was made in view of the above circumstances, and an object of the present invention is to provide a tool abnormality detection device that can accurately detect tool abnormalities even during interrupted machining or machining with low cutting resistance. -ru.

[Means to solve the problem]

In this invention, a tool abnormality detection device for detecting a tool abnormality during machining of a machine tool includes a first detection means for detecting a drive signal of a feed motor of a machine tool or a predetermined physical signal during machining of a tool; a second detection means for detecting the rotational speed of the main shaft of the machine; a frequency analysis means for obtaining a frequency spectrum in a predetermined frequency range in the low frequency range of the detection value of the first detection means; Remove the frequency component corresponding to the spindle rotation speed detected by the second frequency spectrum detection means, its higher-order frequency components, and the frequency spectrum in the vicinity of these frequency components, and calculate the average value of the frequency spectrum after the removal. The tool includes a calculating means for calculating the value, and a tool abnormality detecting means for comparing the calculated value of the calculating means with a predetermined set value and detecting a plow tool abnormality based on the comparison result.

[Effect]

This invention samples a predetermined physical signal during machining of a tool, such as a drive signal of a feed motor of a machine tool or a load applied to a tool, and calculates its frequency spectrum.If a tool abnormality occurs, the frequency spectrum This method was developed by focusing on the fact that signals are generated in the low frequency region of the tool, and detects tool abnormalities using the following procedure.

First, a signal such as the drive signal of the feed motor is sampled, and the frequency spectrum of the sampling signal is obtained by Fourier transforming the sampling signal in a predetermined frequency range. The total sum S1 in the frequency range) and the number N1 of data in the low frequency region are determined.

Next, the frequency components of the spindle rotation signal, its higher-order frequency components, and the power spectrum in the vicinity of these frequency components are extracted from the frequency spectrum in the predetermined low frequency region, and the sum S2 of the extracted spectra and its data are extracted. Find the number N2.

Then, using these obtained data, for example, G-(S
By executing the calculation 1-52)/(N+-N2), the average value G of the frequency spectrum obtained by removing the harmonic components of the spindle rotation signal from the frequency spectrum of the sampling signal is calculated. Then, this calculated value G is compared with a predetermined threshold value Gr, and when G≧Gr, it is determined that the tool is abnormal.

〔Example〕

Hereinafter, the present invention will be described in detail according to embodiments shown in the accompanying drawings.

First, the present invention will be explained in detail.

FIG. 2 shows the experimental equipment according to the present invention, and the workpiece 1
Two defective parts (grained parts) 2.3 are created in advance. These defective parts 2.3 are made by drilling holes in the surface of the workpiece 1 and filling them with sand hardened with an adhesive. This work 1 is rotationally driven around an axis G by rotation of a main shaft with both ends supported by a chuck and a center.

The table 5 with the tool 4 (bite in this case) attached is
As the feed motor 6 rotates, the workpiece 1 is moved in the feed direction along the guide line <-7, thereby cutting the workpiece 1. The table 5 is also movable in the cutting direction.

When processing was performed using this configuration, the results shown in Table 1 below were obtained.

Table 1 Cutting area Workpiece A Normal B Sand eating C Sand eating D Normal tool Normal Normal Defective Defective At this time, the feed motor current value I supplied to the feed motor 6 is sampled, and the motor current value l and its frequency spectrum are When calculated, the results shown in Fig. 3 were obtained for each cutting area. That is, in this case, the tool 4 is
As shown in the table, defects occur in cutting areas C and D, but the sent motor current value I (in Figure 3, the value converted from the sent motor current value I to voltage V is used) There are no major changes in the cutting areas A, B, C, and D. but,
In the frequency spectrum, a power spectrum occurs in the low frequency range in the cutting region C1D, which is significantly different from the cutting region ASB in which no defects occur.

This is because when a tool breaks during cutting, the force balance at the cutting point is disrupted, which has a large effect on the frequency spectrum in the low frequency range of the signal that correlates to cutting resistance (in this case, the sent motor current value l). By exerting

If such a phenomenon is used to detect tool abnormalities, 1
Since tool abnormalities can be detected regardless of cutting conditions, tool defects can be reliably detected. Also, in this case, the frequency to be monitored may be in the low frequency range of about 0 to IClOH2, preferably about O to 50H2, so the frequency range of 10KH2 to 2M)IZ, such as the AE multiplied signal due to the destruction described above, is This is completely different, and it is possible to reliably prevent false detection due to confusion between chip cutting and tool loss. Also,
If the rotational frequency of the spindle and its higher-order frequencies are removed from the above frequency spectrum, and tool abnormalities are determined using the frequency spectrum from which the rotational frequency of the spindle and its higher-order frequencies are removed, vibrations caused by intermittent machining can be reduced. The influence of frequency spectrum fluctuations that occur when machining a workpiece (black skin material) that has been damaged or has changed thickness is eliminated, making it possible to reliably detect only tool damage.

FIG. 1 shows an embodiment of this invention, in which N
An NC lathe machine 10 is used as a C machine tool. N
The C lathe machine 10 has an N'C control device 2 equipped with a powerful electric panel.
The drive is controlled by 0.

In this case, in order to handle up to two-dimensional machining, the Z-axis current 1z and the X-axis current lx are sampled as the feed motor current value I, and these current values 1z and Ix are manually input to the amplifier 30. In addition, as mentioned above, the spindle rotation speed S is sampled in order to remove the spindle rotation speed S and its higher-order frequency n5 (n-1, 2, ...) from the frequency spectrum of the feed motor current value I. The spindle rotation speed S is also manually input to the amplifier 30.

The amplifier 30 amplifies these input values and generates a Z-axis current IzSX
The amplified value of the axis current Ix is input to the A/D converter 50, and the two-axis current 1z and the X-axis current lx are input.

and the amplified value of the spindle rotation speed S are input to the low-pass filter 40. The low-pass filter 40 removes noise from the input signal and then inputs the signal to the A/D converter 50 . A/D converter 5
0, the Z axis current 1z and the Z-axis current IzSX
The shaft current lx and the spindle rotation speed S are A/D converted and input into a personal computer 70 (hereinafter referred to as a personal computer).

In this case, the FFT 60 has two channels, and performs Fourier transform on each of the input two-axis current 1z and the X-axis current I in parallel. FFT6
0 and the personal computer 70 perform data communication via a GPIB bus interface. PC 70 is l10
It is connected to the NC control device 20 via 80.

The FFT60 mainly executes the following processes.

In this process, as shown in Fig. 4, the frequency spectrum P(f) obtained by fast Fourier transform (frequency range; e.g. 0 to 500 H2) is sliced by a predetermined threshold Th, and the portion smaller than the threshold Th is sliced. By cutting the noise that is often included in low-level signals. The calculation formula is as follows.

IF P(f')(Th THEN P(f)-
ThNote that from now on, the frequency spectrum obtained by cutting the portion smaller than the threshold Th will be referred to as P(r)-.

In this process, as shown in FIGS. 5 and 6, the threshold T
Frequency spectrum P(
f) - frequency spectrum within the predetermined calculation target range (a to d) (in the case of Fig. 6, a-5H2, d-30
Hz), and the total sum (area) Sl of the frequency spectrum of the extracted portion and the number N1 of data in the calculation target range a to d are determined.

That is, s 1-dpm-dr Number of data in N1-a-d It is.

However, the area S1 is digitally calculated as the sum of the power spectra in the calculation target range a to d.

The calculation for this process (2) will be referred to as calculation I. In addition,
The sum S1 in this case is an additive calculation value because the logarithmic scale power spectrum (dB value) is used and these dB values are added.

As shown in FIGS. 5 and 6, this process includes frequency components of the spindle rotation signal S, its higher-order frequency components, and the vicinity of these frequency components from the frequency spectrum P(f) within the calculation target range. The power spectrum is extracted, and the total sum S2 of the extracted spectra (the total sum of the hatched portion in FIG. 5) and the number of data N2 are determined.

That is, bi-x 1-h ci-i-x1+h S 2- ,? 5pm -ar N2-Number of extracted data i-1, 2,...L XI h; Setting value. The calculation for this process (2) is called operation (2).

Further, the personal computer 70 mainly executes the following processes.

■Calculation of various variables necessary for Fourier transform The list of variables to be calculated is as follows.

Th: Threshold a for slicing; Lower limit d of the range to be calculated; Upper limit Xl of the range to be calculated; Fundamental wave of the main axis frequency h; Value for determining neighboring frequencies. Note that when determining the ranges a and d to be calculated is set to a low frequency range of about 0 to 100H2, preferably about 0 to 50H2, which is affected only by tool abnormalities. For example, a-5H2 and B-30H2. Further, the fundamental wave x1 of the spindle frequency is calculated from the input spindle rotation speed signal S.

Each of these variables is input to the FFT 60. ? FFT6
0, the sum S1 of the frequency spectrum of the feed motor current calculated in the above process of ■■, and the calculation target range a - d
The average value G is calculated according to the following formula using the number N1 of data in the middle, the sum S2 of the frequency components of the spindle rotation signal S, its higher-order frequency components, and the power spectrum in the vicinity of these frequency components, and the number N2 of data. .

G-(S+-32)/(N1-N2) This process ■
The operation for is called operation (■).

(2) Output of tool abnormality signal The above average value G is compared with a predetermined threshold value Gr, and when G≧Gr, a tool abnormality signal is output to the NC control device 20. In the NC control device 20 to which the tool abnormality signal is input, for example,
The NC machine tool 10 is forcibly stopped.

Hereinafter, the operation processing of the FFT 60 and the personal computer 7o will be explained in order according to the flowchart of FIG. 7 and the like.

With the start of cutting, Z-axis current Iz, X-axis current lx,
and the spindle rotation speed f are sampled, and after passing through the amplifier 30, the low-pass filter 40, and the A/D converter 50, these signals are input to the FFT 60 and the personal computer 70 (steps 100 to 130 in FIG. 7).

The personal computer 70 calculates the fundamental wave x1 of the spindle frequency from the input spindle rotation speed S, and outputs this calculated value x1 to the FFT 60 along with other set values TJ a s d s and h.

In FFT60, these input values Th5a Sd 1x1
The above-mentioned process ■■■ is executed using and h.

That is, in the FFT60, the input Z-axis current Iz,
These Z-axis currents 1z are obtained by performing a fast Fourier transform on each of the axis currents Ix.

After determining each frequency spectrum of the X-axis current Ix (step 140), each frequency spectrum P(r) is sliced by a threshold Th as shown in FIG. 4 to obtain a slice waveform P(r)- (step 140). 150). Next FF
At T60, the determined Z-axis current 1z.

Slice waveform p of each frequency spectrum of X-axis current 1x
The total sum S1 of the power spectra within a predetermined calculation target range a - d of (r) - and the number N of data in this range a - d
1 and executes operation I (step 160).

In addition, in the FFT60, the fundamental wave x1 of the spindle frequency input from the personal computer 70 and the value are used to calculate the calculation target range a-
Frequency spectrum within d (pH) - Extract the frequency components of the spindle rotation signal S, its higher-order frequency components, and the power spectrum in the vicinity of these frequency components, and calculate the sum S2 of the extracted spectra and the number N2 of data thereof. Operation (2) is executed (step 170).

Then, the FFT 60 sequentially sends the sums S1 and N2 and the numbers of data N1 and N2 determined by the above operations I and I to the personal computer 70.

The personal computer 70 that has received these data S1, N2, N1, and N2 uses these data to calculate the average value G (=
(SIN2)/(N1-N2)) is executed (step 180).

Furthermore, the personal computer 70 compares the average value G of the power spectrum obtained by the above calculation with a predetermined set value G (step 190), and outputs a tool abnormality signal to the NC control device 20 when G≧Gr. (Step 2
00). The NC control device 20 that receives the tool abnormality signal executes abnormality processing to forcibly stop the NC machine tool 10, for example. When G<Gr, processing is continued as it is.

In this way, this device detects tool defects by analyzing the frequency of the feeding motor current, so it can accurately detect tool abnormalities that could not be detected by conventional feeding motor current level comparison alone. You will be able to do this. In other words, Figures 8(a) and (b) both show experimental results showing the fluctuations in the sent motor current and the frequency spectrum when a tool defect occurs, but compared to the level comparison of the conventional sent motor current. Figure 8(b), which could not be detected
Even under such conditions, fluctuations corresponding to tool defects occur in the frequency spectrum obtained by this device, and these tool defects can be reliably detected.

Further, FIG. 9 shows experimental results showing the results of the calculation I and the calculation (2) when intermittent machining is performed in both the case where no tool defect occurs and the case where a tool defect occurs.

As can be seen from the comparison between Figures 9(a) and (b), the results of calculation I show that the frequency spectra of the two do not change significantly due to the influence of vibrations from interrupted machining, regardless of whether a tool defect occurs or not. However, as can be seen from the comparison of Figures 9(c) and (d), if the rotational frequency of the spindle and its higher-order frequencies are excluded by calculation A significant difference appears in the frequency spectrum depending on when a tool defect occurs, making it possible to reliably detect tool defects.

In the above embodiment, assuming that two-dimensional machining is performed, both the Z-axis current Iz and the X-axis current lx of the motor current to be sent are sampled, but in the case of one-dimensional machining, What is necessary is to sample the motor current sent by one of the motors corresponding to the processing and perform fast Fourier transformation.

Further, in the above embodiment, the feed motor current is detected and tool defects are detected from its frequency spectrum, but the command value of this feed motor current may also be detected, or the sending motor voltage may be detected. It may also be detected. In addition, vibration of the tool rest may be detected. Furthermore, for example, a table 5 on which a tool is attached
A three-dimensional dynamic strain meter may be attached to the tool (see FIG. 2), and the load applied to the tool may be detected by this dynamic strain meter. Further, other physical signals may be detected during cutting by the machine tool.

In addition, in the above embodiment, the sums S1 and N2 are obtained as addition values of the log scale, but when a linear scale is used for the frequency spectrum, each power spectrum is A synergistic operation may be performed in which the values are multiplied, or the above-mentioned sum may be obtained by adding linear scale values.

Further, in the above embodiment, only tool defects are detected, but the present invention may be applied to detecting the presence or absence of defects on the surface of a workpiece. That is, by appropriately combining tool defect detection and workpiece defect detection, appropriate instructions can be given to the machine, and machine monitoring can be automated.

〔Effect of the invention〕

As explained above, according to the present invention, an appropriate physical quantity for detecting tool defects such as the feed motor current value is detected, the frequency spectrum in the low frequency range of this detection signal is determined, and the spindle rotation speed is determined from this frequency spectrum. Corresponding frequency components, their higher-order frequency components, and frequency spectra in the vicinity of these frequency components are removed, the average value of the frequency spectrum after the removal is determined, this average value is compared with a predetermined setting value, and this comparison is performed. Since tool abnormalities are detected based on the results, tool defects can be appropriately detected regardless of cutting conditions, and false detections due to AE multiplication signal confusion caused by broken chips etc. can be prevented. Furthermore, tool defects can be reliably detected even when machining a workpiece with a slight change in thickness during interrupted machining.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a diagram showing experimental equipment according to the invention, FIG. 3 is a diagram showing experimental results, and FIG. 4 explains a slice waveform of a frequency spectrum. Graphs, FIGS. 5 and 6 are diagrams explaining the fast Fourier transform processing of the embodiment, FIG. 7 is a flowchart showing the operation of the embodiment, and FIGS. 8 and 9 are diagrams showing experimental results. 1... Workpiece 2.3... Sand eating defect 4... Tool 6... Feed motor 10... NC machine tool
20... NC control device 60... Fast Fourier transformer 70... Personal computer Fig. 2 Fig. 1 Fig. 3 Zhou Wanhuan (f) (b) Fig. 4 Fig. 7 Fig. 5 f @ Recording (Hz) Figure 6 Figure 8

Claims (6)

[Claims] (1) A tool abnormality detection device that detects a tool abnormality during machining of a machine tool, which includes a first detection means that detects a drive signal of a feed motor of the machine tool, and a second detection means that detects the rotation speed of the main shaft of the machine tool. a detection means, a frequency analysis means for obtaining a frequency spectrum in a predetermined frequency range in a low frequency range of the detected value of the first detection means, and a frequency spectrum detected by the second detection means from this frequency spectrum. a calculation means for removing a frequency component corresponding to the spindle rotation speed, its higher-order frequency component, and a frequency spectrum in the vicinity of these frequency components, and obtaining an average value of the frequency spectrum after the removal; and a calculation value of the calculation means. A tool abnormality detecting device comprising: a tool abnormality detecting means for comparing the value with a predetermined setting value and detecting a tool abnormality based on the comparison result. (2) The tool abnormality detection device according to (1), wherein the first detection means detects a current value or a voltage value of the feed motor. (3) The tool abnormality detection device according to (1), wherein the first detection means detects a command value of a drive signal for a feed motor. (4) The frequency analysis means has means for obtaining a frequency spectrum at or above the predetermined level by slicing the frequency spectrum at a predetermined level, and the calculation means uses the frequency spectrum at or above the predetermined level. The tool abnormality detecting device according to claim 1, wherein the calculation is performed using the following steps. (5) A tool abnormality detection device for detecting tool abnormality during machining of a machine tool, comprising a first detection means for detecting a predetermined physical signal during machining of the tool, and a first detection means for detecting the rotation speed of the main axis of the machine tool. a frequency analysis means for obtaining a frequency spectrum in a predetermined frequency range in a low frequency range of the detection value of the first detection means; and a frequency analysis means for determining the frequency spectrum from this frequency spectrum by the second detection means. a calculation means for removing a frequency component corresponding to the detected spindle rotation speed, its higher-order frequency component, and a frequency spectrum in the vicinity of these frequency components, and obtaining an average value of the frequency spectrum after the removal; and calculation of the calculation means. A tool abnormality detection device comprising: tool abnormality detection means for comparing a value with a predetermined set value and detecting a tool abnormality based on the comparison result. (6) The tool abnormality detection device according to claim (5), wherein the first detection means detects a load applied to the tool.