JPS61159353A - Detection of tool breakage - Google Patents

Abstract

PURPOSE:To detect breakage of various kinds of tools by sequentially detecting the peak value and effective value of the vibrational wave form in a plurality of pilot frequency bands set for each cutting process, and calculating and comparing the ratio between these values. CONSTITUTION:A piezoelectric acceleration sensor 13 is attached to the lower portion of a tool rest 12 of a lathe 11. The acceleration sensor 13 detects vibration of the tool rest 12 on cutting of a workpiece 8a. At this time, the peak value and effective value of the vibrational wave form in a plurality of pilot frequency bands set for each cutting process are sequentially detected and ratios between these values are calculated. Moving average process is performed on calculation of the ratios. Then, the ratios between the ratio obtained by the moving average process and said ratios before and after the time immediately after the data for the moving average process are calculated. Breakage of tool can be detected according to thus obtained ratio and the reference value set for each cutting process.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a method for detecting damage to tools such as cutting tools, and specifically, a method for detecting tool damage that can accurately detect damage such as wear and chipping of the tool. This is a proposal.

[Prior art]

It is important to accurately detect tool damage in order to improve productivity, reduce machining costs, and prevent deterioration of machining accuracy.

This type of tool damage detection method is described in "Mechanical Technology" No. 3.
There is a method using "Automatic detector for indexable tip for turning large objects" described in Vol. 2, No. 25, pp. 58-61.

That is, as shown in FIG. 5, a piezoelectric acceleration sensor 3 is attached to the tool rest 2 of the whirlpool 81.

The acceleration sensor 3 is for detecting vibrations of the lathe 1 during cutting, and the detected vibration waveform is inputted to a peak value detection circuit 5 and an average value detection circuit 6 via a filter 4. The detection results P and R of the peak value detection circuit 5 and the average value detection circuit 6 are input to the divider 9. Divider 9 is as below (L)
Executes the operation indicated by the formula and sends the result to comparator 7.
give to

Ctsu □ ... (11 The comparator 7 compares this C with a predetermined judgment reference value cb, and if c>cb, it activates a control relay (not shown) to stop the rotation ff1l.

[Problem that the invention seeks to solve]

By the way, as shown in Fig. 6, when the machining dimensions in the longitudinal direction of the workpiece 8 are different, or when cutting workpieces of different materials, the cutting conditions (cutting speed, tool feed rate, cutting speed, etc.) depth).

When the cutting conditions change, the frequency band in which vibration increases due to abnormality (wear, cutting damage) of the cutting tool differs.

However, in the conventional method as described above, the frequency band of the filter 4 is fixed, so when cutting the workpiece 8 where the cutting conditions change or due to external disturbances, the level fluctuation within this frequency band may occur. The problem was that it was impossible to deal with it if it occurred.

In addition, the frequency band of the detected vibration varies depending on the progress of tool abnormalities, resulting in the disadvantage that various tool abnormalities cannot be detected.

In addition, the material to be cut 8 has uneven thickness, bending, poor roundness, etc. as shown in FIG. 7(a), and these uneven thicknesses occur.

When cutting a curved or out-of-round part, the level of the ratio C varies greatly from the state shown in FIG. 7 (inside of FIG. 7) where these are not present in the workpiece 8, as shown in FIG. 7(e). death,
The values shown in the figure when a tool abnormality occurs and the dark value cb, which is the criterion for determining whether the tool is normal or abnormal, will occur without fail, and these will be regarded as tool abnormalities, and there is a risk of erroneous detection. There were also problems.

Means for Solving Problem C] The present invention has been made in view of the above circumstances, and by setting one or more monitoring frequency bands for each cutting process, the vibration waveform in this monitoring frequency band is determined. The peak value and the effective value are sequentially detected to find their ratio, and in the process of finding the ratio, moving average processing is performed, and the ratio of the ratio obtained by this to the ratio immediately after the data related to the moving average processing. By using this as a judgment index for tool abnormality detection, if the machining dimensions and materials differ for each cutting process, or even if the workpiece has uneven thickness, bending, or poor roundness, it can be detected that the same cutting conditions are used. Nevertheless, even when a disturbance occurs that causes the vibration waveform level to fluctuate, it is possible to detect various types of tool damage without the risk of false detection. By adopting itself,
It is an object of the present invention to provide a tool damage detection method that can eliminate the influence of noise entering a measurement system.

A tool damage detection method according to the present invention detects a vibration waveform during cutting of the machine tool with a vibration detector attached to the machine tool, and detects tool damage based on the detected vibration waveform. The peak value and effective value of the vibration waveform in one or more monitoring frequency bands set for each frequency band are sequentially detected and their ratio is determined, and in the process of determining the ratio, moving average processing is performed, and the ratio determined by the moving average processing is and,
The present invention is characterized in that a ratio is calculated with the ratio immediately after the data related to the moving average processing, and damage to the tool is detected based on the calculated ratio and a reference value set for each cutting process.

[First Embodiment] The present invention will be described in detail below based on drawings showing embodiments thereof. FIG. 1 is a schematic diagram showing the implementation state of the tool damage detection method according to the present invention together with its signal processing system.

A piezoelectric acceleration sensor 13 is attached to the lower part of the tool rest 12 of the lathe 11. The acceleration sensor 13 detects vibrations of the tool post 12 during cutting of the material 8 to be cut. Acceleration sensor 1
The detection signal No. 3 is amplified by a preamplifier 31, and the output of the preamplifier 31 is given to three channels of signal processing systems A, B, and C, and can also be taken out from a terminal for monitoring. Signal processing system A, B.

C is a disconnection switch 70a, 70b which has a similar circuit configuration and is opened and closed by a switch control signal given from a control device 30 which will be described later regarding the output of the preamplifier 31.

70c, and receives the output of the preamplifier 31 through the attenuators 32a, 32b, 32c, and the amplifier 10a.

Filter 4a, 4b+4c via 10b, 10c
, the peak value of the input signal within the pass frequency band is sequentially detected, the effective value is detected at a predetermined sampling period, the ratio of the detection results is determined each time, and then a predetermined number of pieces of this time series data are detected. Calculate the moving average value for each spectrum, and calculate the ratio between the moving average value and the ratio immediately after the data related to the moving average processing, in other words, the ratio between the peak value and the effective value at the time of detection, and this ratio. By calculating the ratio of the moving average value of the previous data salt and comparing this calculated ratio with a reference value, it is possible to detect whether or not a byte is damaged. Attenuators 32a, 32b, 32c and amplifier 1
0a, 10b.

10c is for adjusting the level of each channel signal. 33 is a calibration oscillator, and a preamplifier 31
Attenuators 32a and 32b convert the output into outputs.

32c, and performs adjustment of the attenuators 32a, 32b, 32c, amplifiers 10al, 10b, and 110c.

Since the circuit configurations of signal processing systems A, B, and C are similar, signal processing system A will be described below.

The output of the amplifier 10a is applied to a bypass filter 40a via a bypass switch 42a, the output of the bypass filter 40a is applied to a low-pass filter 41a via a bypass switch 43a, and the output of the low-pass filter 41a is applied to a peak value detection circuit 15a.

It is applied to the effective value detection circuit 16a. When the bypass switches 42a and 43a are switched to the bypass side, the circuits of the bypass filter 40a and the low-pass filter 41a are closed, respectively. Bypass filter 40a.

The cut-off frequency of the low-pass filter 41a, that is, the passband frequency of the band-pass filter 4a is variable.

The peak value detection circuit 15a sequentially detects a peak of 4 diPa in the pass frequency band, and provides the detection result to the divider 50a. Similarly, the effective value detection circuit 16a detects the effective value Ra at a predetermined sampling period, and sends the detection result to the divider 50a.
give to The divider 50a executes the operation shown in equation (2) below.

a C=□ ...(2) a Divider 5 (the calculation result of la is the moving average processing bag fi52a
and none as given to divider 51a, respectively.

The moving average processing device 52a sequentially reads only a predetermined number of the output C of the divider 50a, and calculates this time series data C+ (
i = n- to +Ln- to +2-n) moving average value Cn
The time required for the moving average processing is set to be sufficiently shorter than the cutting time required for each of the cutting steps α to δ of the workpiece 8. This moving average value Cn is given to the divider 51a.

The divider 51a waits for the data Cn of the output C of the divider 50a at the time of the detection, and executes the calculation shown in the following equation (3) based on this Cn+1 and the moving average value τn up to the point immediately before, and calculates the result. is given to the comparator 17a.

It is also provided to the moving average processing device 52a for calculation.

The output of the divider 51a when detecting the C leaf 2 is expressed by the following equation (4).

The output of the divider 51a is given to the comparator 17a.

A required reference value sb is set in the comparator 17a, and when Sn>Sb, the timer 23a is activated.
If this condition continues for the time set in the timer 23a, the alarm lamp 24a is turned on and the Wl device 27 is also made to sound. The output of the timer 23a can be taken out from the terminal 26a. If the output of the comparator 17a is shorter than the set time of the timer 23a, the timer 23a returns to zero. The reference value sb of the comparator 17a is set to an appropriate value by the control device 30 in accordance with changes in cutting conditions and the like.

Signal processing systems B and C are given the same numbers as signal processing system A, with suffixes ``c'' and ``c'' added thereto, and the explanation thereof will be omitted.

It goes without saying that the signal processing system may have four or more channels.

Therefore, when implementing the method of the present invention, the disconnection switch 7
Opening/closing of 0a, 70b, 70c, each filter 4a
.

4b, 4c passband frequency settings (that is, bypass switches 42a, 43a, 42b, 4
3b, 42c, 43c switching position settings, bypass filters 40a, 40b.

40c, O-bus filters 41a, 41b,
41c (cutoff frequency G setting) and reference value setting of the comparator 17a are sequentially controlled by the control device 30 as the cut-off-1 process progresses.

[Effect]

Next, the control device 30 based on the flowchart shown in FIG.
The following describes the control details. Now, each cutting process α, β, and T is performed for the workpiece 8 shown in FIG.

Monitoring conditions such as the used channel, cutoff frequency band, and comparator reference value are set in the control device 30 for each of δ.

When cutting starts, the control device 30 first determines the process,
Set monitoring conditions such as channels used in the α process. The number of channels used may be single or multiple, and
The lathe 11 is stopped when the terminal 26a output is monitored and an output is obtained from these, that is, when the input signal of the comparator continues to be larger than its reference value for the time set on the timer in any of the channels used. let At this time, the alarm lamp 24a, 24b, or 24c lights up, and the alarm 27 can determine which channel the abnormality has been detected by the 0 lighting lamp that sounds.

Next, when entering the β step, the switch 70a, 70b or 70c is closed depending on the channel used in that case.

In this case, when the channel used in the α process is used again, the cutoff frequency band.

If it is necessary to change the reference value of the comparator, the time limit of the timer, etc., it is necessary to change these settings when moving from the α process to the β process. Then, the β process is monitored, and if no abnormality has occurred, the δ process name is monitored.

According to the present invention, the number of channels of signal processing systems A, B, and C and their monitoring frequency bands are independently determined according to the cutting process, etc., and the reference value of the comparator and the time limit of the timer are determined. Therefore, if these monitoring conditions are set appropriately in advance based on actual results, etc., it is possible to detect a variety of tool breakages, and also to detect the damage of the workpiece 8 whose cutting conditions change.
It is also possible to accurately detect tool damage.

Furthermore, n is calculated using the moving average value of the ratio Ci between the peak value Pa and the effective value Ra, and the ratio Sn between this τn and CnY1 is used as a judgment index for tool abnormality detection. 8
Even if there is a deviation in thickness or a defect in roundness, the abnormality of the tool can be accurately detected without the risk of false detection, unlike the above-mentioned conventional method.

The reason for this will be explained below based on FIG. FIG. 3 is a graph in which the vertical axis shows the ratio c1 [FIG. 3 (Jl)) and the ratio Sn (35th peak)], and the horizontal axis shows time.

Now, as described above, if the workpiece 8 has an uneven thickness, a bend, or a defect in roundness, the level of the ratio C2 changes greatly. However, when looking at the moving average value τn, the ratio C
, due to its own property of smoothing, the 3rd viIJ (
As shown by the broken line in a), the time course is generally constant.

However, the ratio Sn is determined when an abnormality occurs in the tool in which the value of cl rises and falls sharply [in Fig. 3 (a), the data Cn
+ 1), the molecule Cn+ 1 increases significantly more than the data before and after it, so its value Sn increases significantly and exceeds its reference value Sb, as shown in FIGS. Therefore, when selecting the standard value sb1 and the time required for the moving average processing to an appropriate value, even when the workpiece 8 has uneven thickness 2, bends, poor roundness, etc., these and tool abnormalities will occur. It is possible to discriminate between the two and perform accurate detection.

[Second Embodiment] In the first embodiment described above, the ratio sn was used as a judgment index for tool abnormality detection, but in the second embodiment, the moving average value τn of the ratio C1 itself is used as a judgment index.

That is, the moving average processing device 52a output Cn is given to the comparator 17a, 100 is compared with b using its reference value, and τn
>τb, the configuration is such that the output is similarly output to the timer 23a to detect an abnormality in the tool.

In the case of such a configuration, as described above, the moving average value τn remains approximately constant even when the workpiece 8 has uneven thickness 9 bending or poor roundness, but an abnormality occurs in the tool. Since the level increases when
Not only can the two be discriminated and accurate detection can be performed, but the following effects can also be achieved.

In other words, when noise whose frequency is in the monitoring frequency band enters the measurement system and is taken in as the peak value Pa, the peak value increases rapidly, resulting in an increase in C1, as shown in Figure 4 (al). As shown, there is a possibility that the threshold value cb may be exceeded.In such a case, in the conventional method described above, this will be erroneously detected as a tool abnormality.However, if the moving average value cn is used as a judgment index, , the number of data to be processed k
In addition, when the reference value τb is selected as an appropriate value, the value of Cn is smoothed despite the sudden increase in the peak value Pa, as shown in Fig. 4). Since the temporal change of n in the moving average value is a smooth curve, there is no risk of exceeding the reference value cb.

On the other hand, when an abnormality occurs in the tool, the moving average process 〇 increases significantly and exceeds the reference value τb, so that it is possible to distinguish between the two. Therefore, even if noise enters the measurement system, tool abnormalities can be accurately detected.

〔effect〕

As detailed above, the tool damage detection method according to the present invention sets one or more optimal monitoring frequency bands depending on the tool damage and its progress, and sets a peak value and an effective value in the monitoring frequency band. is detected to determine these ratios, and in the process of determining the ratio, moving average processing is performed, and the ratio itself or the ratio between this and the ratio immediately after the moving average processing is used for tool abnormality detection. Since it is used as a judgment index, it can be used to detect various types of tool damage, even when the cutting conditions of the workpiece material change, or when the workpiece material has uneven thickness, bending, or poor roundness. The present invention has excellent effects, such as being able to accurately detect tool abnormalities without the risk of erroneous detection even when noise enters the measurement system.

Note that in the above embodiment, the moving average value of Ci is calculated, but it is also possible to calculate the moving average values of each of Pa and Ra, and then calculate the ratio thereof.

Further, in the above-described embodiments, the present invention was applied to detecting breakage of a cutting tool, but it is of course applicable to detecting breakage and wear of other tools such as a milling cutter. Further, in the above-described embodiment, the monitoring conditions for each process are automatically changed, but some or all of the monitoring conditions may be changed manually.

[Brief explanation of drawings]

FIG. 1 is a schematic diagram showing the implementation state of the method of the present invention together with its signal processing system, FIG. 2 is a flowchart of the control device, and FIG.
Figure 4 is a graph for explaining the effects of the present invention, Figure 5 is a schematic diagram showing the implementation state of the conventional method, Figure 6 is a schematic diagram showing the material to be cut, and Figure 7 is a problem with the conventional method. It is a graph showing points.

Claims (1)

[Scope of Claims] 1. A method for detecting vibration waveforms during cutting of the machine tool using a vibration detector attached to the machine tool, and detecting damage to the tool based on the detected vibration waveforms, comprising: setting for each cutting process. The peak value and effective value of the vibration waveform in one or more monitored frequency bands are sequentially detected and their ratio is determined. In the process of determining the ratio, moving average processing is performed, and the ratio determined by the moving average processing and the relevant Tool damage is characterized in that a ratio of the data related to moving average processing to the ratio immediately after the data is calculated, and damage to the tool is detected based on the calculated ratio and a reference value set for each cutting process. Detection method. 2. In a method of detecting vibration waveforms during cutting of the machine tool with a vibration detector attached to the machine tool and detecting damage to the tool based on the detected vibration waveforms, one or more monitors set for each cutting process are used. The peak value and effective value of the vibration waveform in the frequency band are sequentially detected and their ratio is determined. In the process of determining the ratio, moving average processing is performed, and the ratio determined by the moving average processing is compared with the ratio set for each cutting process. A method for detecting damage to a tool, characterized in that damage to the tool is detected based on a reference value.