JP7313001B2 - Tool life detection device and tool life detection method - Google Patents

Description

Application of Article 30, Paragraph 2 of the Patent Act Released in the 2017 (31st) Young Researcher Nurturing Society Research Presentation Meeting Research Paper Collection

The present invention relates to a tool life detection device and a tool life detection method attached to a machine tool for detecting the life of a machining tool during machining.

In metal machining, in order to maintain a certain level of machining quality, it is necessary to accurately grasp the working tool life and replace the working tool at an appropriate time.
Conventionally, the number of times of machining until the machined surface roughness of the object to be machined (hereinafter referred to as "work") becomes out of specification, and the number of times of machining with a safety factor taken into account has been used as a guideline for tool replacement.

The safety factor is empirically determined in consideration of chipping of machining tools and variations in machining tools. Chipping is the so-called edge chipping. Wear occurs gradually in proportion to the number of times of machining, but chipping occurs suddenly. Variation occurs due to the quality of working tools.
Therefore, the safety factor is expected to be about 10%.

I examined the machining tools that were exchanged at the number of times in anticipation of the safety factor. Then I found something that still works quite well. In other words, a machining tool whose machined surface roughness should be within the standard was replaced before the end of its service life.
If the working tools are used until the end of their life, productivity can be improved by increasing the number of times of working, and the cost of procuring the working tools can be reduced.
Therefore, there is a demand for a technology capable of detecting the life of a working tool with high accuracy.

For this reason, several techniques have been proposed for detecting the life of a working tool (see, for example, Patent Document 1 (FIG. 5)).

Patent Document 1 discloses that the Mahalanobis-Taguchi method (hereinafter referred to as the MT method) is used to determine the Mahalanobis distance (hereinafter referred to as the MD value) based on data obtained during machining, and the MD value is used to detect the life of a milling cutter (multi-blade tool).

The data obtained during machining are the spindle torque command value, X-axis direction vibration amplitude, Y-axis direction vibration amplitude, Z-axis direction vibration amplitude, X-axis direction strain, and Y-axis direction strain in the machine tool.
A conventional technique will be described with reference to FIG.
FIG. 15(a) is a graph for explaining identification of flank wear amount, and (b) is a graph for explaining identification of chipping. In both cases, the horizontal axis represents the number of times of processing (the number of passes), and the vertical axis represents the Mahalanobis distance.

In FIG. 15(a), for example, 8000 (vertical axis) is set as a criterion, and when the curve rising to the right reaches 8000, the tool is replaced (paragraph 0033 of Patent Document 1).
In FIG. 15(b), when a new multi-blade tool is artificially chipped, an MD value of about 2000 is obtained, and when a multi-blade tool after 830 passes is artificially chipped, an MD value of about 18000 is obtained.
Abnormality can be determined by separating the amount of wear and chipping (Patent Document 1, paragraph 0047).

Further, in Patent Document 1, when the machining Mahalanobis distance corresponding to the flank wear amount is displayed at a position separated by a predetermined distance above the machining Mahalanobis distance, the occurrence of chipping is detected (Patent Document 1, paragraph 0046).
However, the distance displayed for the isolating position varies. Determining with a changing distance is complicated and troublesome, so the effectiveness of Patent Document 1 for chipping remains questionable.
That is, the technique of Patent Document 1 is recognized as effective against wear, but its effectiveness against chipping remains questionable.
However, as the effective use of machining tools including multi-bladed tools is required, it is required to be able to accurately detect abnormalities including chipping.

JP 2017-226027 A

An object of the present invention is to provide a tool life detection device and a tool life detection method that can detect the life of a machining tool including chipping with high accuracy.

The inventors paid attention to the following two points in the technique of Patent Document 1.
For the first point, torque command value, vibration amplitude and strain are selected as data.
The second point is the method of processing the vibration amplitude.

Regarding the torque command value, vibration amplitude, and strain at the first point, the processing load increases in proportion to the number of data in data processing by a computer. If there are three types of data, data acquisition is also difficult.
As a result of experiments and studies, the inventors of the present invention have found that the number of acquired data can be reduced by placing more emphasis on chipping than wear and chipping. That is, chipping has a large effect on vibration, and has a small effect on the torque command value and strain. Based on this idea, I came up with the idea of limiting the acquired data to vibration.

Regarding the vibration amplitude at the second point, Patent Document 1 focuses on the magnitude of the vibration wave.
If there are a first wave and a second wave of the same magnitude, and the width of the first wave in the direction of the time axis (length of the time axis) is wide and the width of the second wave is narrow, the width of this wave will affect the life of the machining tool. Patent document 1 is missing this point.
The inventors have come to the conclusion that it is necessary to add the width of the wave to the amplitude of the oscillation.

The inventors of the present invention limited the acquired data to vibration, and proceeded with experiments and verifications based on analyzing the obtained vibration based on the amplitude and the width of the wave (length of the time axis), thereby completing the present invention.

The invention according to claim 1 is a tool life detection device that is attached to a machine tool that performs machining on the work by relative movement between the work tool and the work, and that detects the life of the work tool during machining work,
The tool life detecting device includes a vibration sensor attached to a component of the machine tool for detecting vibration, a calculation unit for calculating the Mahalanobis distance based on vibration information from the vibration sensor, a determination unit for determining whether the Mahalanobis distance obtained by the calculation unit is equal to or greater than a judgment value, and an abnormality display unit for displaying an abnormality when the judgment unit determines that the Mahalanobis distance is equal to or greater than the judgment value.
In the calculation unit, a sample line parallel to the processing time axis is drawn on the waveform curve obtained from the vibration information, the number of intersections of the waveform curve and the sample line is set as the amount of change, the sum of the line segments of the sample line divided by the waveform curve is set as the amount of change, and the amount of change and the amount of existence are used to calculate the Mahalanobis distance,
The calculation unit defines a frequency at which the blade of the processing tool hits the workpiece as a fundamental frequency,
Using this fundamental frequency as the primary frequency, determining secondary or higher frequencies based on this primary frequency,
Vibration information near at least the primary frequency is filtered from part of the vibration information, and the Mahalanobis distance is calculated using the filtered vibration information.

The invention according to claim 2 is the tool life detection device according to claim 1,
a component of the machine tool is a spindle of the machine tool to which the machining tool is attached;
The vibration sensor is attached to the spindle along the feed direction of the machining tool.

The invention according to claim 3 is the tool life detection device according to claim 1 or claim 2,
The calculation unit may calculate the Mahalanobis distance using part of the vibration information obtained when the machining tool makes one pass relative to the workpiece.

The invention according to claim 4 is a tool life detection method attached to a machine tool that performs machining on the work by relative motion between the work tool and the work, and detects the life of the work tool during machining work,
detecting vibrations occurring in components of the machine tool;
calculating a Mahalanobis distance based on the detected vibration information;
a step of detecting an abnormality when the calculated Mahalanobis distance is equal to or greater than the judgment value,
In the step of calculating the Mahalanobis distance, a sample line parallel to the processing time axis is drawn on the waveform curve obtained from the vibration information, the number of intersections of the waveform curve and the sample line is set as the amount of change, the sum of the line segments of the sample line divided by the waveform curve is set as the amount of change, and the amount of change and the amount of existence are used to calculate the Mahalanobis distance;
In the step of calculating the Mahalanobis distance, the frequency at which the blade of the processing tool hits the workpiece is defined as a fundamental frequency, this fundamental frequency is defined as a primary frequency, and secondary or higher frequencies are determined based on this primary frequency,
Vibration information near at least the primary frequency is filtered from part of the vibration information, and the Mahalanobis distance is calculated using the filtered vibration information.

The invention according to claim 5 is the tool life detection method according to claim 4 ,
In the step of calculating the Mahalanobis distance, part of the vibration information obtained when the machining tool makes one pass relative to the workpiece is used to calculate the Mahalanobis distance.

In the invention according to claim 1, a sample line parallel to the processing time axis is drawn on the waveform curve obtained from the vibration information, the number of intersections of the waveform curve and the sample line is used as the amount of change, and the sum of the line segments of the sample lines separated by the waveform curve is used as the amount of change, and the amount of change and the amount of existence are used to calculate the Mahalanobis distance.
A line segment of the sample line divided by the waveform curve corresponds to the width of the wave.

That is, in the present invention, only vibration information is selected as primary information from a large number of data (primary information). As a result, the load on the calculation unit can be greatly reduced.
The amount of change is defined as the number of intersections between the waveform curve and the sample line, the sum of the line segments of the sample lines separated by the waveform curve is defined as the amount of change, and the amount of change and the amount of existence are used to calculate the Mahalanobis distance.
Therefore, the present invention provides a tool life detection device capable of detecting the life of a working tool including chipping with high accuracy.
In addition, in the invention according to claim 1, the calculation is performed using vibration information filtered based on at least the fundamental frequency. Since filtering is performed based on at least the fundamental frequency, only vibration information when the blade hits the workpiece can be used for calculation, and the reliability of life determination can be further improved.

In the invention according to claim 2, the vibration sensor is attached to the spindle along the feed direction of the machining tool.
Assuming that the axial direction of the spindle is the z-axis direction, the feeding direction of the machining tool and perpendicular to the z-axis is the y-axis direction, and the direction perpendicular to both the y-axis and the z-axis is the x-axis direction, it was confirmed that the tool life can be detected regardless of whether the vibration sensor is mounted in the x-axis direction, the y-axis direction, or the z-axis direction.
However, it was found that the y-axis direction is more reliable in life determination than the x-axis direction and the z-axis direction. Therefore, it is recommended to install the vibration sensor along the feeding direction of the machining tool.

In the invention according to claim 3, part of the vibration information obtained when the machining tool makes one pass relative to the workpiece is used. By using a part of the vibration information for calculation, the load on the calculation unit can be reduced and the cost of the calculation unit can be reduced.

In the invention according to claim 4 , as in claim 1, a sample line parallel to the processing time axis is drawn on the waveform curve obtained from the vibration information, the number of intersections of the waveform curve and the sample line is used as the amount of change, the sum of the line segments of the sample line divided by the waveform curve is used as the amount of change, and the amount of change and the amount of existence are used to calculate the Mahalanobis distance.

That is, in the present invention, only vibration information is selected as primary information from a large number of data (primary information). As a result, the load on the calculation unit can be greatly reduced.
The amount of change is defined as the number of intersections between the waveform curve and the sample line, the sum of the line segments of the sample lines separated by the waveform curve is defined as the amount of change, and the amount of change and the amount of existence are used to calculate the Mahalanobis distance.
Therefore, the present invention provides a tool life detection method capable of detecting the life of a machining tool including chipping with high accuracy.
In addition, in the invention according to claim 4, calculation is performed using vibration information filtered based on at least the fundamental frequency. Since filtering is performed based on at least the fundamental frequency, only vibration information when the blade hits the workpiece can be used for calculation, and the reliability of life determination can be further improved.

In the invention according to claim 5 , as in claim 3, part of the vibration information obtained when the machining tool makes one pass relative to the workpiece is used. By using a part of the vibration information for calculation, the load on the calculation unit can be reduced and the cost of the calculation unit can be reduced.

1 is a principle diagram of an experimental apparatus according to the present invention; FIG. 4 is a graph of machined surface roughness and vibration wave obtained by experiment. It is a figure explaining the variation|change_quantity and abundance which are used when calculating MD value. It is a figure explaining using a part of vibration information. 4 is a graph of MD values created based on work vibration information; It is a figure explaining the effectiveness of this invention. 1 is a principle diagram of an experimental apparatus having vibration sensors in the x-axis direction of a main axis; FIG. It is a graph of MD values created based on vibration information in the x-axis direction. It is a figure explaining filtering. It is a graph of MD value created based on the vibration information of the x-axis direction after filtering. It is a graph of MD value created based on the vibration information of the z-axis direction of a main axis. 5 is a graph of MD values created based on vibration information in the y-axis direction of the main axis; 1 is a basic configuration diagram of a tool life detection device according to the present invention; FIG. FIG. 4 is a flow chart explaining the operation of the tool life detection device according to the present invention; It is a figure explaining a conventional technique.

An embodiment of the present invention will be described below with reference to the accompanying drawings.

The tool life detecting device 31 according to the present invention will be described with reference to FIG. 13, and the basic principle of the present invention, which is the premise thereof, will be described with reference to FIGS. 1 to 12. FIG.
As shown in FIG. 1, the experimental apparatus 10 includes a table 11, a spindle 12, a vibration sensor 13, an amplifier 14 for amplifying vibration information from the vibration sensor 13, and a calculation unit 15 for calculating MD values based on the amplified vibration information.
In the experiment, the vibration sensor 13 used a small uniaxial acceleration pickup, but the type and format may be selected arbitrarily.

A lower plate 16 is placed on the table 11 , a work 17 is placed on the lower plate 16 , and an upper plate 18 is placed on the work 17 . Then, a plurality of (four in this example) bolts 19 are passed through the upper plate 18 , the workpiece 17 and the lower plate 16 and screwed into the table 11 . It is also possible to fix the lower plate 16 to the table 11 and fix the workpiece 17 and the upper plate 18 to the lower plate 16 with bolts 19 .

An end mill 21 as a cutting tool is attached to the spindle 12 . One side of the work 17 is cut by rotating the end mill 21 while relatively moving the end mill 21 and the work 17 in the x-axis, y-axis, and z-axis directions.

Experimental conditions:
・Work material: SKD11 hardened to the center
・Dimensions of workpiece: 5 mm x 100 mm x 100 mm

・Number of blades of end mill: 6 ・Spindle rotation speed: 3170 rpm ・Cutting depth: 0.1 mm
・Feeding speed: 5.3 mm/second ・1 pass time: about 20 seconds (time to cut 100 mm length once with an end mill)
・Replacement of workpieces: workpieces are replaced after about 200 passes. This is because the cutting allowance disappears after about 200 passes.

・Sampling frequency: 6.4kHz
・Measurement items: Machined surface roughness, cutting vibration, presence or absence of chipping:
For cutting vibration, the vibration data for each pass is taken as an absolute value, and the average value of the absolute values is taken.
Check for chipping with a microscope every 25 passes (2.5m).
・Calculation item: MD value (details will be explained in FIG. 3)

As shown in FIG. 2(a), when the cutting distance (m) is taken on the horizontal axis and the machined surface roughness (Ra (μm)) is taken on the vertical axis, the machined surface gradually becomes rougher. This is because the end mill gradually wears out.
When the horizontal axis exceeded 20 m and 40 m, the workpiece was changed, and when the third workpiece was cut, that is, when the cutting distance was around 70 m, the roughness jumped to 1.5 μm. When the blade of the end mill was observed with a microscope, chipping (chipping of the blade) was observed.

Thus, chipping can be detected by monitoring the machined surface roughness. However, it is difficult to measure the machined surface roughness during actual operation.
Vibration, on the other hand, can be measured during the cutting operation. Therefore, it is examined whether or not chipping can be detected by vibration.

As shown in FIG. 2(b), the vibration gradually increased. Note that there is a large turbulence in front of 40 m on the horizontal axis, but this turbulence is the boundary between the first day and the second day, and is the turbulence that occurred when the cutting work was stopped and started.
It is predicted from FIG. 2(a) that chipping occurs when the cutting distance is around 70 m. However, the change in vibration before and after the cutting distance of 70 m is not significant. Chipping is therefore difficult to detect.

To overcome this difficulty, we use the MD value (Mahalanobis distance).
The MD value is obtained by calculating the amount of change and the existence value using the MT method calculation formula. The amount of change and the existence value will be explained with reference to FIG. Since the MT method calculation formula is well known, the description of the formula and calculation is omitted.

A comparative example is shown in FIG.
In FIG. 3(a), the vibration value 23 is plotted with cutting time on the horizontal axis and vibration on the vertical axis. This vibration value 23 is a digital value. In MT (Mahalanobis-Taguchi method), a suitable sample line 24 is drawn parallel to the horizontal axis (processing time axis). Then, in the standard method, the number of vibration values 23 existing above the sample line 24 is taken as the abundance. The number of points on the first peak 25 is six, the number of points on the second peak 26 is five, and the number of points on the third peak 27 is three. Calculating 6+5+3=14 yields an abundance of 14.

A sample line 24 drawn parallel to the processing time axis is delimited by a first peak 25 , a second peak 26 and a third peak 27 .
Let m1 be the line segment of the sample line 24 divided by the first mountain 25 . A line segment is a line of finite length, which is different from a line of infinite length. The unit of the line segment m1 is cutting time.

It is assumed that the length of the line segment m1 has a great influence on the damage given to the end mill. However, in the comparative example described with reference to FIG. 3A, only the magnitude of vibration is simply taken into account, and the line segment m1 is not considered at all. The inventors paid attention to this point.

Therefore, the inventors have found that the reliability of the MD value can be improved by considering the line segment m1.
FIG. 3(b) shows an example.
A waveform curve 28 shown in FIG. 3(b) was obtained by connecting the vibration values given by points (FIGS. 3(a) and 23) with a smooth curve. The number of intersections 29 (6 in this example) where the waveform curve 28 intersects with the sample line 24 was used as the amount of change.
In addition, the sum (m1+m2+m3) of the line segment m1 of the first peak 25, the line segment m2 of the second peak 26, and the line segment m3 of the third peak 27 was taken as the abundance.
In the present invention, the MD value is calculated based on the variation and abundance.

Although the sample line 27 and the waveform curve 28 have been described graphically to promote understanding, it goes without saying that they are virtual lines within the computer. The intersection point 29 and line segments m1 to m3 are also virtual points and lines.

Although the MD value is calculated for each pass, the following measures are preferably taken in order to reduce the load on the calculation section.
As shown in FIG. 4, of the vibration data for 20 seconds (for one pass), two seconds from the first second to the third second in the stable period are defined as the MT method calculation area. By performing calculations for only 2 seconds out of 20 seconds, the load on the calculation unit can be reduced to 1/10.
It should be noted that the MT method calculation area may be set at an arbitrary location, such as changing it to two seconds from the ninth second to the eleventh second.

MD values calculated based on FIGS. 3(b) and 4 are described in FIG. 5(b).
FIG. 5(a) is the same view as FIG. 2(b).
The horizontal axis of FIG. 5(a) between 0 and 67.5 m is notated in FIG. 5(b), and the horizontal axis of FIG. 5(a) between 67.5 and 69.9 m is expanded in FIG. 5(b).

The MD value between 0 and 67.5 m on the horizontal axis was 5 or less, as shown in FIG. 5(b). This interval is referred to as the normal range for convenience.
On the horizontal axis, MD values between 67.5 and 69.9 m varied between 0 and 249, as shown in FIG. 5(b). This interval is referred to as a boundary area for the sake of convenience.
On the horizontal axis, the MD value from 70 m exceeded 100, as shown in FIG. 5(b). This interval is referred to as an abnormal region for convenience.

For example, if the determination value is set to "50", the MD value becomes equal to or greater than the determination value at (68.6) in the middle of the boundary area, and it is possible to determine that there is an abnormality at this point.
This possibility was confirmed by another demonstration experiment. The results are described in FIG.

In this demonstration experiment, a workpiece was cut using a turret lathe.
As shown in FIG. 6(a), the MD value was small and stable while grinding up to 160 workpieces. At point A, the skilled worker in charge of the lathe noticed an anomaly. The production number at point A is 184 pieces.

As shown in FIG. 6B, when the judgment value is given, the MD value at point B becomes equal to or greater than the judgment value. The production number at point B is 171 pieces.

Assume that the 171st to 184th workpieces are "defective".
In FIG. 6(b) to which the present invention is applied, the number of defective products is limited to one.
On the other hand, in the case of FIG. 6A to which the present invention is not applied, the calculation of 187-170=17 results in 17 defective products.

By applying the present invention, it is possible to minimize the occurrence of defective products, which is a remarkable effect. That is, according to the present invention, it is possible to continue using the cutting tool to the limit while minimizing the occurrence of defective products.

By the way, the vibration sensor 13 is attached to the upper plate 18 in FIG. In this attachment, it is necessary to detach the vibration sensor 13 when exchanging the work 17, which increases the man-hours for detachment.
Since the number of man-hours is required to be reduced, mounting the vibration sensor 13 on the spindle 12 is considered.

As shown in FIG. 7, the vibration sensor 13 was attached to the main shaft 12 along the x-axis direction, an experiment was conducted, and the MD value was calculated. The results are shown in FIG.
As shown in FIG. 8, the MD values of the borderline and abnormal regions were not significantly different from those of the normal region. This makes it difficult to give a judgment value, and even if it is given, there is a risk that an erroneous judgment will be made, and countermeasures are required.

1, the vibration sensor 13 was attached to the work 17 side, so that the vibration could be directly detected by the vibration sensor 13. However, in FIG. A bearing was interposed between the spindle 12 and the end mill 21, and it was considered that the vibration was damped by the clearance of this bearing. Measures must be taken to compensate for this attenuation.

The end mill 21 has six blades per turn at a pitch of 60°, and attention was paid to the difference in vibration between when the blades are in contact with the work 17 and when they are not in contact with the work 17 . In other words, the inventors focused on the fact that the reliability of the vibration data when the blade is not in contact is reduced.

FIG. 9A is the same diagram as FIG.
The vibration information in the MT method calculation area of FIG. 9(a) was subjected to frequency analysis as shown in FIG. 9(b). Since the 6-flute end mill is rotated at 3170 revolutions per minute, the calculation of 6*3170÷60=317 gives a fundamental cutting frequency of 317 Hz.
In FIG. 9B, horizontal lines are added to 317 Hz (first order), 634 Hz (second order), 951 Hz (third order), 1268 Hz (fourth order), and 1585 Hz (fifth order).
Then, as shown in FIG. 9C, each frequency is filtered with a width of about 10 Hz, and the filtered waveform is used as a new MT method calculation region.

FIG. 10 shows the result of recalculating the MD value based on this new MT method calculation area.
Compared with FIG. 8, in FIG. 10, the MD value in the boundary region is significantly different from the MD value in the normal region. Therefore, with FIG. 10, abnormality detection becomes easy and accurate.

In FIG. 9(c), only 317 Hz (fundamental frequency) was set as a new MT method calculation area, and a difference was observed between the normal area, the boundary area, and the abnormal area, though not as much as in FIG.
Next, when two frequencies consisting of 317 Hz (primary frequency) and 634 Hz (secondary frequency) were set as a new MT method calculation area and calculation was performed, the result was a little closer to that shown in FIG.
That is, when calculating based on the 1st to n-order frequencies (where n is an integer of 2 or more), the larger the value of n, the clearer the difference between the normal region, boundary region, and abnormal region.

Although n is set to 5 in the embodiment, n may be appropriately set depending on the situation. In some situations, the nth order frequency may be absent.
Therefore, the fundamental frequency is assumed to be the primary frequency, the secondary or higher frequency is determined based on this primary frequency, at least the vibration information near the primary frequency is filtered from a part of the vibration information, and the filtered vibration information is used to calculate the Mahalanobis distance.

As shown in FIG. 11(a), the vibration sensor 13 was attached to the main shaft 12 along the z-axis direction.
MD values as shown in FIG. 11(b) were obtained. This MD value is based on the new MT method computational domain.

As shown in FIG. 12(a), the vibration sensor 13 was attached to the main shaft 12 along the y-axis direction. "Along the y-axis direction" means that the axis of the vibration sensor 13 is arranged parallel to the y-axis. The same applies to the x-axis direction and the z-axis direction.
MD values as shown in FIG. 12(b) were obtained. This MD value is based on the new MT method computational domain.

FIG. 10 (x-axis direction), FIG. 11(b) (z-axis direction), and FIG. 12(b) (y-axis direction) are compared.
Focusing on the boundary area, FIG. 10 (x-axis direction) and FIG. 12(b) (y-axis direction) are superior in terms of abnormality detection.
Also, when focusing on the abnormal region, FIG. 12B (y-axis direction) is superior.

Therefore, although the vibration sensor 13 can be mounted anywhere on the main shaft 12 to detect an abnormality, it is preferably arranged in the x-axis direction or the y-axis direction, and most preferably in the y-axis direction.

Next, the tool life detection device 31 according to the present invention will be described with reference to FIGS. 13 and 14. FIG.
As shown in FIG. 13, a machine tool 30 has a table 11 for supporting a workpiece 17 and a main spindle 12 to which a machining tool 21 is detachably attached, and includes a tool life detection device 31 .

The tool life detection device 31 includes a vibration sensor 13 attached to a component of the machine tool 30 (for example, the workpiece 17 and the spindle 12), an amplifier 14 that amplifies vibration information obtained by the vibration sensor 13, a calculation unit 15 that calculates the MD value based on the amplified vibration information, a determination unit 32 that compares the calculated MD value with a judgment value, and an abnormality display unit 33 that notifies the surroundings of the abnormality when the judgment unit 32 determines that there is an abnormality. Abnormalities are indicated by any one or a combination of these, such as a buzzer or bell sound, a lamp or rotary warning light light, a graphical or textual display.

The machine tool 30 may be any one of a milling machine, a drilling machine, a boring machine, and a lathe as long as it has a table 11 for supporting the workpiece 17 . If the processing tool 21 is a cutting tool, the cutting tool may be a milling cutter, an end mill, a drill, or a cutting tool.

At ST (step number) 01 in FIG. 14, a vibration sensor is attached to the machine tool.
In ST02, the judgment value MDb is read.
Next, machining is started (ST03), the MD value is calculated for each pass (ST04), and it is determined whether or not the calculated MDcal is greater than or equal to the determination value MDb (ST05).

If MDcal is less than the judgment value MDb, the process returns to ST04 to continue machining. When the work is interrupted due to a break or the like, or when the work is finished manually, this flow is finished in ST06.

When it is determined in ST05 that MDcal is greater than or equal to the determination value MDb, an abnormality is displayed (ST07).
Although it is possible to stop the machining along with the error display, it is preferable to wait until the pass is completed (ST08).
When the pass is completed, the machining is stopped (ST09), the machining tool is replaced (ST10), and the process returns to ST03 when continuing the operation (ST11).

That is, the tool life detection method according to the present invention consists of the following steps.
The process consists of the steps of detecting vibration generated in the components of the machine tool (ST01 to ST03), the step of calculating the Mahalanobis distance based on the detected vibration information (ST04), and the step of detecting an abnormality when the calculated Mahalanobis distance is equal to or greater than the judgment value (ST05).

Then, in the step of calculating the Mahalanobis distance, a sample line parallel to the processing time axis is drawn on the waveform curve obtained from the vibration information, the number of intersections of the waveform curve and the sample line is used as the amount of change, and the sum of the line segments of the sample lines separated by the waveform curve is used as the amount of change, and the amount of change and the amount of existence are used to calculate the Mahalanobis distance.

It should be noted that this flow describes a preferred example, and may be changed as appropriate.

In the embodiment, the processing tool 21 is a cutting tool (such as an end mill), but the processing tool 21 may be a grinding tool, a polishing tool, or a boring tool.
Further, in the embodiments, the component of the machine tool to which the vibration sensor is attached is the workpiece or the spindle, but the component of the machine tool may be any of the machining tool, the base of the turret, and the stage for fixing the workpiece, and may be any element as long as the vibration generated during machining is transmitted.

INDUSTRIAL APPLICABILITY The present invention is suitable for a tool life detection device and tool life detection method attached to a machine tool.

12... Spindle, 13... Vibration sensor, 15... Calculation part, 17... Work, 21... Processing tool (cutting tool, end mill), 24... Sample line, 28... Waveform curve, 29... Intersection, 30... Machine tool, 31... Tool life detection device, 32... Judgment part, 33... Abnormal display part, m1 to m3... Line segment.

Claims (5)

A tool life detection device that is attached to a machine tool that performs machining on the work by relative movement between the machining tool and the work, and that detects the life of the machining tool during machining,
The tool life detecting device includes a vibration sensor attached to a component of the machine tool for detecting vibration, a calculation unit for calculating the Mahalanobis distance based on vibration information from the vibration sensor, a determination unit for determining whether the Mahalanobis distance obtained by the calculation unit is equal to or greater than a judgment value, and an abnormality display unit for displaying an abnormality when the judgment unit determines that the Mahalanobis distance is equal to or greater than the judgment value.
In the calculation unit, a sample line parallel to the processing time axis is drawn on the waveform curve obtained from the vibration information, the number of intersections of the waveform curve and the sample line is set as the amount of change, the sum of the line segments of the sample line divided by the waveform curve is set as the amount of change, and the amount of change and the amount of existence are used to calculate the Mahalanobis distance,
The calculation unit defines a frequency at which the blade of the processing tool hits the workpiece as a fundamental frequency,
Using this fundamental frequency as the primary frequency, determining secondary or higher frequencies based on this primary frequency,
A tool life detecting device characterized by filtering vibration information in the vicinity of at least the primary frequency from part of the vibration information, and calculating the Mahalanobis distance using the filtered vibration information. The tool life detection device according to claim 1,
a component of the machine tool is a spindle of the machine tool to which the machining tool is attached;
A tool life detecting device, wherein the vibration sensor is attached to the spindle along the feed direction of the machining tool. The tool life detection device according to claim 1 or claim 2,
The tool life detection device, wherein the calculation unit calculates the Mahalanobis distance using part of the vibration information obtained when the machining tool makes one pass relative to the workpiece. A tool life detection method that is attached to a machine tool that performs machining on the work by relative motion between the machining tool and the work, and detects the life of the machining tool during machining,
detecting vibrations occurring in components of the machine tool;
calculating a Mahalanobis distance based on the detected vibration information;
a step of detecting an abnormality when the calculated Mahalanobis distance is equal to or greater than the judgment value,
In the step of calculating the Mahalanobis distance, a sample line parallel to the processing time axis is drawn on the waveform curve obtained from the vibration information, the number of intersections of the waveform curve and the sample line is set as the amount of change, the sum of the line segments of the sample line divided by the waveform curve is set as the amount of change, and the amount of change and the amount of existence are used to calculate the Mahalanobis distance;
In the step of calculating the Mahalanobis distance, the frequency at which the blade of the processing tool hits the workpiece is defined as a fundamental frequency, this fundamental frequency is defined as a primary frequency, and secondary or higher frequencies are determined based on this primary frequency,
A tool life detection method, wherein vibration information near at least the primary frequency is filtered from a part of the vibration information, and the Mahalanobis distance is calculated using the filtered vibration information. The tool life detection method according to claim 4 ,
In the step of calculating the Mahalanobis distance, part of the vibration information obtained when the machining tool makes one pass relative to the workpiece is used to calculate the Mahalanobis distance. A method for detecting a tool life.

Images