Anomaly detection method and anomaly detection device

Abstract

We provide a new technology for detecting abnormalities in tools. [Solution] The abnormality detection method comprises an output step of outputting a periodic waveform of a signal corresponding to the distance between the cylindrical outer surface of the tool holder to which the tool is attached and the sensor, as the tool rotates during workpiece processing, and a generation step of generating information corresponding to the state of the tool based on the output waveform.

Description

[Technical Field]

[0001] The present invention relates to a technique for detecting the degree of abnormality of a tool. [Background Art]

[0002] In cutting performed by a machine tool such as a machining center, wear of a cutting tool is often determined based on the user's experience from visual observation, cutting sound, and processing marks. However, if the user has little experience, there is a risk that wear of the cutting tool may be overlooked, and in this case, the cutting tool may be damaged, or a workpiece that does not have desired dimensions may be produced. For this reason, various methods for detecting the state of a cutting tool without relying on the user's experience have been devised.

[0003] For example, an abnormality detection device and an abnormality detection method for a cutting tool have been devised, which are configured to measure a current value or a power value of a spindle motor that drives the cutting tool in a processing machine, detect chipping of a cutting edge of the cutting tool based on an increase in a specific frequency component in the frequency spectrum distribution thereof, and control the operation of the processing machine (see Patent Document 1). [Prior Art Document] [Patent Document]

[0004] [Patent Document 1] Japanese Unexamined Patent Application Publication No. 2020-104257 [Summary of the Invention] [Problem to be Solved by the Invention]

[0005] In the abnormality detection method described above, sudden chipping of a tool can be detected by continuously measuring signal waveforms of the current value and power value of the motor. However, in a case where wear of the tool gradually increases as the number of machining operations (total number of rotations) by the tool increases and the tool reaches the end of its service life, an abnormality does not necessarily appear in the signal waveform of the motor.

[0006] This invention has been made in view of these circumstances, and one of its objectives is to provide a new technology for detecting abnormalities in tools. [Means for solving the problem]

[0007] One aspect of the present invention is an anomaly detection method. This anomaly detection method comprises an output step of outputting a periodic waveform of a signal corresponding to the distance between the cylindrical outer surface of the tool holder to which the tool is attached and a sensor, as the tool rotates during workpiece machining, and a generation step of generating information corresponding to the state of the tool based on the output waveform.

[0008] Another aspect of the present invention is an anomaly detection device. This device comprises a sensor positioned opposite the cylindrical outer surface of a tool holder to which a tool is attached, a controller that acquires a signal from the sensor corresponding to the distance between the sensor and the outer surface during workpiece machining and outputs a periodic waveform of the signal associated with the rotation of the tool, and a generation unit that generates information corresponding to the state of the tool based on the output waveform. [Effects of the Invention]

[0009] According to the present invention, a new technology for detecting the lifespan of a tool can be provided. [Brief explanation of the drawing]

[0010] [Figure 1] This figure shows a schematic configuration of the machining center according to this embodiment. [Figure 2] This is a schematic diagram illustrating the cutting conditions of the workpiece. [Figure 3] This is a schematic diagram illustrating how to observe the cutting edge of a tool. [Figure 4] This figure shows a microscopic image of a worn blade edge. [Figure 5] This graph shows the relationship between the increase in the number of machining cycles and the wear width W1. [Figure 6] This graph shows the relationship between the increase in the number of machining cycles and the cutting resistance. [Figure 7]This figure shows waveform A1, which includes a DC component, output from an eddy current displacement sensor, and waveform A2, which is waveform A1 after BPF processing. [Figure 8] Figure 7 is an enlarged view of waveform A2 during the S region. [Figure 9] This figure shows the relationship between the total number of rotations of the tool over the entire machining cycle and the amplitude Vp-p value (waveform A3), and the moving average of the Vp-p value signal over 100 cycles (waveform A4). [Figure 10] This is a schematic block diagram of an anomaly detection device that performs the anomaly detection method according to this embodiment. [Figure 11] Figure 11(a) shows the waveform when the tool is not in contact with the workpiece, and Figure 11(b) shows the waveform when the tool is in contact with the workpiece. [Figure 12] This figure shows the intensity amplitude of the fundamental frequency and its integer multiples obtained by Fourier analysis of the waveform output from the sensor. [Figure 13] This figure shows the relationship between the number of processing steps and the nth-order frequency component. [Modes for carrying out the invention]

[0011] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In the following embodiment and its modified examples, substantially identical components are denoted by the same reference numerals, and their descriptions are omitted as appropriate.

[0012] First, a machining center capable of executing the anomaly detection method according to this embodiment will be described. Figure 1 is a diagram showing the schematic configuration of the machining center according to this embodiment. In the machining center 10 shown in Figure 1, a tool holder 14 for holding tools 12 such as cutting tools and drilling tools is mounted on the spindle head 16. The tools 12 process the workpiece material 17. The tools 12 include, for example, cutting tools such as end mills and grinding tools for drilling cemented carbide.

[0013] A machining center 10 is an apparatus that automatically selects various tools in accordance with machining processes, automatically mounts the selected tools to a spindle head 16, and performs multiple types of machining. In the machining center 10, tool change is performed by an unillustrated automatic tool change (ATC: Automatic Tool Holder Change) apparatus. The ATC apparatus automatically takes out a tool holder 14 with a tool attached thereto from a tool magazine, and automatically mounts the tool holder to the spindle head 16.

[0014] The tool holder 14 with a tool attached thereto has a conical fitting portion, and is mounted by fitting this fitting portion into a conical fitted portion formed on the spindle head 16. At this time, if chips or the like adhere between the fitting portion and the fitted portion, the tool will be mounted with an eccentric shaft. When machining is performed in this state, runout occurs in the tool, which significantly reduces the machining accuracy of the workpiece. Therefore, the machining center 10 according to the present embodiment includes an eddy current displacement sensor 26 for determining whether the tool holder 14 is normally mounted to the spindle head 16.

[0015] The eddy current displacement sensor 26 detects, as an electrical displacement signal, a distance d to a cylindrical outer peripheral surface 14a of the tool holder 14 mounted on the spindle head 16. This enables stable output of a periodic waveform of a signal corresponding to the distance d between the cylindrical outer peripheral surface 14a of the tool holder 14 and the eddy current displacement sensor 26 even in harsh environments where water, oil and other substances are scattered, such as when machining a workpiece in a machining center. Note that any sensor that can measure the distance d from a specific measurement point to the outer peripheral surface 14a of the tool holder 14 is not limited to the eddy current displacement sensor 26, and other sensors may be used. Furthermore, the sensor is not limited to non-contact sensors such as the eddy current displacement sensor 26, and contact-type sensors may also be used.

[0016] The electrical signal detected by the eddy current displacement sensor 26 is processed by a controller 28, logged by a data logger 22, transferred to a personal computer 24, and analyzed by the personal computer 24. The personal computer 24 is also connected to a cutting dynamometer 30 that measures the machining (cutting) resistance applied to a work material 17 during machining of the work material 17.

[0017] In the machining center 10 described above, continued use of tool 12 will lead to wear, and failure to replace the tool at the appropriate time may result in tool breakage or defective workpieces. Therefore, the change in wear was measured by machining a workpiece under predetermined cutting conditions. Figure 2 is a schematic diagram illustrating the cutting conditions of the workpiece. Tool 12 is an end mill with one cutting edge and a diameter of 12 mm. The cutting conditions are as follows: rotational speed during cutting is 8000 rpm, feed rate (speed of tool movement in the direction of arrow F) is 1200 [mm / min], depth of cut X is 0.3 mm, depth of cut Z is 6 mm, cutting distance Y per pass (per machining) is 60 mm, and workpiece depth L is 100 mm. The cutting method is down cutting, and it is dry cutting without coolant. The workpiece material is NAK55 (plastic mold steel).

[0018] Under the above cutting conditions, machining was repeated, and the change in wear amount after a predetermined number of machining cycles was measured. Figure 3 is a schematic diagram illustrating the method of observing the cutting edge of the tool. As shown in Figure 3, the tool 12 was placed on the stage 32 such that the reference surface 12a (the surface opposite the rake face 12b) of the tool 12 was at a 75° angle to the stage 32, and the flank surface 12c was observed from directly above.

[0019] Figure 4 shows a micrograph of a worn cutting edge. As shown in Figure 4, as the machining time (number of machining cycles) increases, the wear widths W1 and W2 in the wear region R near the boundary between the rake face 12b and the flank face 12c also tend to increase. Figure 5 is a graph showing the relationship between the increase in the number of machining cycles and the wear width W2. Figure 6 is a graph showing the relationship between the increase in the number of machining cycles and the cutting resistance. As shown in Figure 5, the rate of increase in wear width accelerates once the number of machining cycles exceeds 1000. Also, as shown in Figure 6, the cutting resistance begins to increase once the number of machining cycles exceeds 1000, showing the same trend as the change in wear width.

[0020] When the rate of increase in wear width is not constant, or when the rate of increase in wear width differs depending on the type of tool, it is difficult to estimate the lifespan of a tool in advance. Furthermore, in mass production lines, it is not practical to remove tools and directly check the wear width at regular intervals. Therefore, determining the lifespan of a tool tends to rely on the user's experience. Accordingly, the inventors of this application have invented a new method that can indirectly estimate the wear state of a tool using a sensor without relying on the user's experience. In this embodiment, an eddy current displacement sensor 26 is used as this sensor.

[0021] Figure 7 shows waveform A1, which includes the DC component, output from the eddy current displacement sensor, and waveform A2, which is obtained by applying a bandpass filter (BPF) to waveform A1. Since the DC component of the sensor signal output from the eddy current displacement sensor 26 drifts due to the influence of temperature, this invention focuses on the AC component. To remove the DC component and noise and extract the AC component, a fourth-order Butterworth filter bandpass filter (BPF: fcl=500Hz, fch=0.01Hz) was applied. In Figure 7, period t1 represents the cutting time of 3s (cutting distance / travel speed = 60mm / 20 [mm / s]) during which the tool moves one cutting distance (60mm), and period t2 represents 2.5s excluding the start and end of one cutting time.

[0022] Figure 8 is an enlarged view of waveform A2 during the S region of Figure 7. As shown in Figure 8, the amplitude Vp-p value of the waveform voltage for each rotation of the tool (1 period T) was calculated in the waveform A2 data. Figure 9 shows the relationship between the number of rotations for the total number of machining operations of the tool and the amplitude Vp-p value (waveform A3), and the moving average of the Vp-p value signal over 100 operations (waveform A4). Note that the notation "n stages" in the figures refers to the portion that was repeatedly cut from the top surface of the workpiece 17 in Figure 2 with a cutting depth of 6 mm, the 1st stage being the portion with a cutting depth of 0-6 mm, the 2nd stage being the portion with a cutting depth of 6-12 mm, ..., and the 6th stage being the portion with a cutting depth of 30-36 mm.

[0023] As shown in Figure 9, the Vp-p value tends to increase with increasing total rotational speed (number of machining cycles). Since the Vp-p value correlates with the magnitude of tool runout, it can be inferred that the tool runout is increasing as the number of machining cycles increases.

[0024] As described above, the abnormality detection method according to this embodiment includes an output step of outputting a periodic waveform of a signal corresponding to the distance d between the cylindrical outer surface 14a of the tool holder 14 to which the tool 12 is attached and the eddy current displacement sensor 26, as the tool 12 rotates during machining of the workpiece 17, and a generation step of generating information corresponding to the state of the tool based on the output waveform. Figure 10 is a block diagram illustrating a schematic of an abnormality detection device that executes the abnormality detection method according to this embodiment.

[0025] The controller 28 includes an amplifier circuit 34, an A / D converter 36, a CPU 38, a memory 40, and an input / output circuit 42. The signal output from the eddy current displacement sensor 26 is processed through the various circuits inside the controller 28 and output from the controller 28 as a runout detection signal, which determines whether or not the tool holder 14 is properly mounted on the spindle head 16. The control device 25, which controls the entire machining center 10, controls the operation of the machining center 10 in accordance with the runout detection signal.

[0026] At least a portion of the output process according to this embodiment is performed by the controller 28. The periodic waveform amplified by the amplifier circuit 34 is input to the personal computer 24 via an external data logger 22, which also functions as an A / D converter. The personal computer 24 then generates information corresponding to the tool's state based on the output waveform. Specifically, as shown in Figure 8, the personal computer 24 generates information corresponding to the tool's state from the amplitude Vp-p values ​​of the output waveform. This allows the personal computer 24 to estimate the tool's wear state with simple signal processing without having to calculate the nth-order frequency components as described later. Thus, the abnormality detection device according to this embodiment comprises an eddy current displacement sensor 26, a controller 28, a data logger 22, and a personal computer 24.

[0027] Here, information corresponding to the state of the tool refers to, for example, information that estimates the tool's lifespan or wear, and specifically, information that estimates the amount of tool wear. The personal computer 24 may use the estimated information to display the current amount of tool wear on the screen, or it may notify the operator with sound or a message when the amount of wear exceeds a predetermined threshold. In this way, even if the state of the tool 12 changes as the machining time by the tool 12 increases, the abnormality detection method according to this embodiment can generate information corresponding to the state of the tool 12 based on the periodic waveform of the signal corresponding to the distance d between the cylindrical outer surface 14a of the tool holder 14 to which the tool 12 is attached and the eddy current displacement sensor 26, as the tool 12 rotates while machining the workpiece 17. As a result, abnormalities in the tool 12 can be detected using the information generated while the machining center 10 is running, thus reducing the downtime of the machining center 10. Furthermore, since tool abnormalities can be detected in advance even by operators who are not experienced, tool damage and a decrease in workpiece yield can be suppressed.

[0028] In this embodiment, both a so-called runout detection signal, which determines whether the tool holder 14 is properly mounted on the spindle head 16 based on the signal obtained from the eddy current displacement sensor 26, and a signal corresponding to the amount of tool wear can be used. This is because the former is a signal obtained when the tool 12 is not in contact with the workpiece 17, and the latter is a signal obtained when the tool 12 is in contact with the workpiece 17 and machining is in progress, and the waveforms of the two signals are different. Figure 11(a) shows the waveform when the tool is not in contact with the workpiece, and Figure 11(b) shows the waveform when the tool is in contact with the workpiece.

[0029] Next, we will describe other processing of the signal output from the eddy current displacement sensor 26. Figure 12 shows the intensity amplitude of the fundamental frequency and its integer multiples obtained by Fourier analysis of the waveform output from the sensor. In Figure 12, since an end mill tool with one cutting edge is rotated at 8000 rpm, the fundamental frequency is 8000 / 60 ≈ 133 Hz, and peaks appear at its integer multiples. In Figure 12, the dotted line represents the initial amplitude, and the dotted line represents the amplitude after machining (2003 machining cycles). Note that when using the eddy current displacement sensor 26, the distance between the sensor and the cutting tool is measured, making it less susceptible to external noise. Therefore, it is thought that the waveform shows peaks only at integer multiples of the fundamental frequency.

[0030] Figure 13 shows the relationship between the number of machining cycles and the nth-order frequency component. Here, n is an integer, and Figure 13 shows the case where n is 1 to 5, i.e., the change in the 1st to 5th-order frequency components with increasing number of machining cycles. As shown in Figure 13, the integrated value of the signal amplitude corresponding to the 1st and 2nd-order frequency components increases with increasing slope from around 1000 machining cycles. Using this change, the personal computer 24 processes the output waveform to calculate the nth-order (n is a natural number) frequency component and generates information corresponding to the tool state from the amplitude of the nth-order frequency component signal. This makes it possible to estimate the tool state by processing the periodic waveform associated with the rotation of the tool 12 during workpiece machining.

[0031] The personal computer 24 generates notification information that notifies the tool of its lifespan when the amplitude of the nth-order frequency component signal exceeds a predetermined value. In this embodiment, when the amplitude of the first-order frequency component signal exceeds a predetermined value, the personal computer 24 determines that the amount of tool wear has exceeded a predetermined value and generates notification information that the tool has reached the end of its lifespan. Alternatively, when the amplitudes of both the first-order and second-order (or nth-order) frequency component signals exceed a predetermined value, the personal computer 24 may determine that the amount of tool wear has exceeded a predetermined value and generate notification information that the tool has reached the end of its lifespan. This allows for accurate detection of the tool's lifespan (abnormality). Here, the first-order frequency corresponds to the rotation period of the tool. For example, the first-order frequency is 100 Hz when an end mill with one tooth rotates at 100 rpm. Similarly, the first-order frequency is 200 Hz when an end mill with two tooth rotates at 100 rpm. The nth-order frequency is n times the frequency of the first-order frequency.

[0032] Although the tool rotation speed during workpiece machining was 8000 rpm as described above, the abnormality detection method according to this embodiment can be implemented even if the tool diameter, workpiece material, machining speed, etc. are not necessarily at this rotation speed. For example, the tool rotation speed during workpiece machining may be 100 rpm or higher, or 500 rpm or higher. This makes it possible to detect tool abnormalities even when using a large-diameter tool at a low rotation speed.

[0033] [Differentiation] In the abnormality detection device according to the above embodiment, a data logger 22 and a personal computer 24 are provided separately from the controller 28. However, a signal processing section such as Fourier transform may be provided inside the controller 28, and the generated wear detection signal may be output to the control device 25, which may then display information corresponding to the amount of tool wear. Furthermore, the Fourier transform may be, for example, a Fast Fourier Transform (FFT).

[0034] It should be noted that the present invention is not limited to the embodiments and modifications described above, and the components can be modified and implemented without departing from the spirit of the invention. Various inventions may be formed by appropriately combining the multiple components disclosed in the embodiments and modifications described above. In addition, some components may be deleted from all the components shown in the embodiments and modifications described above. [Explanation of symbols]

[0035] 10 Machining center, 12 Tool, 12b Rake face, 12c Relief face, 14 Tool holder, 14a Outer surface, 16 Spindle head, 17 Workpiece, 22 Data logger, 24 Personal computer, 25 Control unit, 26 Eddy current displacement sensor, 28 Controller.

Claims

1. An output process that outputs a periodic waveform of a signal corresponding to the distance between the cylindrical outer surface of the tool holder to which the tool is attached and the sensor, as the tool rotates during workpiece machining. A generation step that generates information corresponding to the state of the tool based on the output waveform, An anomaly detection method comprising the following features.

2. The abnormality detection method according to claim 1, wherein the generation step involves processing the output waveform to calculate the nth-order (n is a natural number) frequency component, and generating information corresponding to the state of the tool from the amplitude of the signal of the nth-order frequency component.

3. The abnormality detection method according to claim 2, wherein the generation step generates notification information that notifies the tool of its lifespan when the amplitude of the signal of the nth-order frequency component exceeds a predetermined value.

4. The abnormality detection method according to claim 2, wherein the generation step generates information corresponding to the state of the tool from the amplitude of a signal of at least the first-order frequency component.

5. The abnormality detection method according to claim 1, wherein the generation step generates information corresponding to the state of the tool from the amplitude of the output waveform.

6. The abnormality detection method according to any one of claims 1 to 5, wherein the sensor is an eddy current displacement sensor.

7. An abnormality detection method according to any one of claims 1 to 5, wherein the rotational speed of the tool during workpiece processing is 100 rpm or more.

8. The abnormality detection method according to any one of claims 1 to 5, wherein the tool is an end mill.

9. A sensor is provided at a position opposite the cylindrical outer surface of the tool holder to which the tool is attached, During workpiece machining, a controller acquires a signal from the sensor corresponding to the distance between the sensor and the outer surface, and outputs a periodic waveform of the signal in accordance with the rotation of the tool. A generation unit that generates information corresponding to the state of the tool based on the output waveform, An anomaly detection device equipped with the following features.

Images