Abnormality detection system, abnormality detection device, abnormality detection method, and computer program

Abstract

The abnormality detection system is provided with: a sensor that is provided to the tool and measures a physical quantity of the tool that varies during a period in which the machine tool machines an object to be machined by the tool; and an abnormality detection device that detects an abnormality of the tool on the basis of measurement data of the sensor, the abnormality detection device including: a synchronization unit that acquires synchronization between target data, which is time-series measurement data of the sensor, and reference data, which is time-series measurement data of the sensor; the reference data is a measurement result of a time sequence of the physical quantity when the tool is normal; a distance calculation unit that calculates the distance between the reference data and the target data synchronized by the synchronization unit; and an abnormality detection unit that detects an abnormality of the tool by comparing the distance calculated by the distance calculation unit with a threshold value.

Description

Technical Field

[0001] The present disclosure relates to an anomaly detection system, an anomaly detection device, an anomaly detection method, and a computer program. Background Art

[0002] Patent document 1 discloses an abnormality detection device for detecting abnormalities of a tool of a machine tool. The abnormality detection device disclosed in Patent document 1 learns each measurement data of the tool's vibration information, cutting force information, sound information, spindle load, motor current, and power value through a one-class support vector machine (One Class SVM) method to create a normal model. During processing after the normal model is created, the measurement data is acquired while diagnosing whether the measurement data is normal or abnormal based on the normal model. The abnormality detection device further re-diagnoses the measurement data diagnosed as abnormal through methods different from the one-class support vector machine method, such as invariant analysis.

[0003] Prior art literature Patent Literature Patent Document 1: Japanese Patent Application Publication No. 2018-024055 Summary of the invention An abnormality detection system involved in one embodiment of the present disclosure detects an abnormality of a tool of a machine tool, wherein the abnormality detection system comprises: a sensor, which is arranged on the tool, and measures a physical quantity of the tool that changes during a period when the machine tool processes a processing object through the tool; and an abnormality detection device, which detects an abnormality of the tool based on the measurement data of the sensor, and the abnormality detection device includes: a synchronization unit, which synchronizes object data with reference data, wherein the object data is a time series of measurement data of the sensor, and the reference data is a time series of measurement results of the physical quantity when the tool is normal; a distance calculation unit, which calculates the distance between the reference data synchronized by the synchronization unit and the object data; and an abnormality detection unit, which detects an abnormality of the tool by comparing the distance calculated by the distance calculation unit with a threshold value.

[0004] The present disclosure can be implemented not only as an abnormality detection system having the characteristic configuration as described above, but also as an abnormality detection device included in the abnormality detection system, or as an abnormality detection method using characteristic processing in the abnormality detection system as steps. The present disclosure can be implemented as a computer program that enables a computer to function as an abnormality detection device, or a part or all of the abnormality detection device can be implemented as a semiconductor integrated circuit. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] Figure 1It is a diagram showing an example of the overall configuration of the abnormality detection system involved in the embodiment.

[0006] Figure 2A It is a diagram showing an example of the structure of the cutting tool involved in the embodiment.

[0007] Figure 2B It is a diagram showing another example of the structure of the cutting tool according to the embodiment.

[0008] Figure 2C It is a diagram showing another example of the structure of the cutting tool according to the embodiment.

[0009] Figure 2D It is a diagram showing another example of the structure of the cutting tool according to the embodiment.

[0010] Figure 3 It is a diagram showing an example of the configuration of a sensor module according to an embodiment.

[0011] Figure 4 This is a block diagram showing an example of the hardware configuration of the abnormality detection device according to the embodiment.

[0012] Figure 5 This is a functional block diagram showing an example of the function of the abnormality detection device involved in the embodiment.

[0013] Figure 6 This is a graph showing an example of a time series waveform of the strain of the cutting tool.

[0014] Figure 7 This is a graph showing an example of reference data.

[0015] Figure 8 This is a graph showing an example of the object data.

[0016] Fig. 9 This is a graph showing an example of measurement results of the strain sensor when a process under the same processing conditions is repeated a plurality of times.

[0017] Fig.10 This is a graph showing an example of the first moving average and the second moving average of the target data.

[0018] Fig.11 Yes Fig.10 The enlarged curve of the T part in .

[0019] Fig.12 This is a graph showing an example of the moving average difference.

[0020] Fig.13This is a graph for explaining an example of determining the cut-off period.

[0021] Fig.14A This is a graph showing an example of the clipped data before correction.

[0022] Fig. 14B This is a graph showing an example of corrected clipped data.

[0023] Fig.15 It is a graph for explaining DTW between reference data and cut data.

[0024] Fig.16A This is a graph showing an example of an abnormality degree graph.

[0025] Fig. 16B This is a graph showing another example of the abnormality degree graph.

[0026] Fig.17 This is a flowchart showing an example of an abnormality detection process of the abnormality detection device according to the embodiment.

[0027] Fig.18 This is a flowchart showing an example of synchronization processing. DETAILED DESCRIPTION

[0028] <Problems to be Solved by the Present Disclosure> The abnormality detection device disclosed in Patent Document 1 requires multiple sensors such as a vibration sensor for measuring the vibration of the tool, a cutting power meter for measuring the cutting force of the tool, and a sound sensor for measuring the sound of the tool. In addition, a large amount of measurement data needs to be acquired in machine learning, and even if the measurement data used for machine learning is acquired, the acquired measurement data needs to be input. Thus, machine learning requires a lot of effort.

[0029] <Effects of the Present Disclosure> According to the present disclosure, it is possible to detect abnormality of a tool of a machine tool without requiring a plurality of sensors and a large amount of data.

[0030] <Overview of Embodiments of the Present Disclosure> Hereinafter, the outline of the embodiments of the present disclosure will be described.

[0031] (1) The abnormality detection system according to the present embodiment detects abnormality of a tool of a machine tool, wherein the abnormality detection system comprises: a sensor provided on the tool and measuring a physical quantity of the tool that changes while the machine tool processes a processing object with the tool; and an abnormality detection device that detects abnormality of the tool based on measurement data of the sensor, the abnormality detection device including: a synchronization unit that synchronizes object data with reference data, the object data being time-series measurement data of the sensor, the reference data being time-series measurement results of the physical quantity when the tool is normal; a distance calculation unit that calculates a distance between the reference data synchronized by the synchronization unit and the object data; and an abnormality detection unit that detects abnormality of the tool by comparing the distance calculated by the distance calculation unit with a threshold value. Thus, unlike abnormality determination using machine learning, a plurality of sensors are not required, and a large amount of measurement data for machine learning is not required. By synchronizing the reference data with the object data, the difference between the current state of the tool (at the time of abnormality detection) and the state at the time of normality can be accurately calculated as the distance, and the abnormality can be accurately detected.

[0032] (2) In the above (1), the abnormality detection device may further include a period setting unit, the period setting unit setting the object period of the same processing step in the reference data and the object data, and the distance calculation unit calculates the distance between the reference data and the object data within the object period set by the period setting unit. In this way, the reference data and the object data in the same processing step can be compared, and the distance that accurately represents the difference between the current state of the tool and the normal state can be calculated.

[0033] (3) In the above (2), the processing step may be an intermittent processing step in which cutting and non-cutting of the processing object are repeatedly performed. In the intermittent processing step, the state of the tool is easier to express as a physical quantity than in the continuous processing step in which cutting of the processing object is continuously performed. Therefore, it is possible to calculate a distance that accurately represents the difference between the current state of the tool and the normal state.

[0034] (4) In any of the above (1) to (3), the sensor may be a strain sensor that measures the strain of the tool as the physical quantity. The measurement value of the strain sensor shows a high value during the period when the tool is in contact with the object to be processed (cutting the object to be processed), and shows a low value during the period when the tool is not in contact with the object to be processed (not cutting the object to be processed). Using the measurement data of such a strain sensor, it is possible to accurately detect abnormalities in the tool.

[0035] (5) In any of the above (1) to (3), the sensor may be an acceleration sensor that measures the acceleration of the tool as the physical quantity. The measurement value of the acceleration sensor repeats high values ​​and low values ​​while the tool is vibrating, and shows a low value while the tool is not vibrating. Using such measurement data of the acceleration sensor, it is possible to accurately detect abnormality of the tool.

[0036] (6) In any one of the above (1) to (5), the abnormality detection device may further include a correction unit that corrects the baseline of the object data, and the distance calculation unit calculates the distance between the object data after the baseline is corrected by the correction unit and the reference data. Thus, even if the baseline differs for each object data, the abnormality of the tool can be stably detected.

[0037] (7) In any of the above (1) to (6), the distance between the reference data and the object data may be the distance between the time series waveform of the reference data and the time series waveform of the object data, or the difference between the distribution of the reference data and the distribution of the object data. Thus, a distance suitable for abnormality detection can be calculated based on the measurement data.

[0038] (8) In any of the above (1) to (7), the abnormality detection system may further include a display device that displays a time series graph of the distance between the reference data and the object data calculated by the distance calculation unit, so that a user can visually confirm the time change of the state of the tool.

[0039] (9) The abnormality detection device according to the present embodiment detects abnormality of a tool of a machine tool, wherein the abnormality detection device comprises: an acquisition unit that acquires time-series measurement data output from a sensor provided on the tool, the sensor measuring a physical quantity of the tool that changes while the machine tool processes a processing object with the tool; a synchronization unit that synchronizes object data with reference data, the object data being the time-series measurement data acquired by the acquisition unit, the reference data being a time-series measurement result of the physical quantity when the tool is normal; a distance calculation unit that calculates a distance between the reference data synchronized by the synchronization unit and the object data; and an abnormality detection unit that detects abnormality of the tool by comparing the distance calculated by the distance calculation unit with a threshold value. Thus, unlike abnormality determination using machine learning, a plurality of sensors are not required, and a large amount of measurement data for machine learning is not required. By synchronizing the reference data with the object data, the difference between the current state of the tool (at the time of abnormality detection) and the state at the time of normality can be accurately calculated as a distance, and the abnormality can be accurately detected.

[0040] (10) The abnormality detection method according to the present embodiment is performed by an abnormality detection device for detecting an abnormality of a tool of a machine tool, wherein the abnormality detection method includes the following steps: acquiring time-series measurement data outputted from a sensor provided on the tool, the sensor measuring a physical quantity of the tool that changes while the machine tool processes a processing object with the tool; synchronizing object data with reference data, the object data being the acquired time-series measurement data, the reference data being the time-series measurement result of the physical quantity when the tool is normal; calculating a distance between the synchronized reference data and the object data; and detecting an abnormality of the tool by comparing the calculated distance with a threshold value. Thus, unlike abnormality determination using machine learning, a plurality of sensors are not required, and a large amount of measurement data for machine learning is not required. By synchronizing the reference data with the object data, the difference between the current state of the tool (at the time of abnormality detection) and the normal state can be accurately calculated as a distance, and an abnormality can be accurately detected.

[0041] (11) The computer program according to the present embodiment is a computer program for detecting an abnormality of a tool of a machine tool, wherein the computer program causes a computer to execute the following steps: acquiring time-series measurement data output from a sensor provided on the tool, the sensor measuring a physical quantity of the tool that changes while the machine tool processes an object to be processed by the tool; synchronizing object data with reference data, the object data being the acquired time-series measurement data, the reference data being the time-series measurement result of the physical quantity when the tool is normal; calculating a distance between the synchronized reference data and the object data; and detecting an abnormality of the tool by comparing the calculated distance with a threshold value. Thus, unlike abnormality determination using machine learning, a plurality of sensors are not required, and a large amount of measurement data for machine learning is not required. By synchronizing the reference data with the object data, the difference between the current state of the tool (at the time of abnormality detection) and the normal state can be accurately calculated as a distance, and an abnormality can be accurately detected.

[0042] <Details of the embodiments of the present disclosure> Hereinafter, the details of the embodiments of the present invention will be described with reference to the drawings. It should be noted that at least a part of the embodiments described below may be arbitrarily combined.

[0043] [1. Anomaly detection system] Figure 1 1 is a diagram showing an example of the overall structure of the abnormality detection system involved in the present embodiment. The abnormality detection system 10 detects an abnormality of a cutting tool 30 of a machine tool 20. The cutting tool 30 is mounted on the machine tool 20. The machine tool 20 uses the cutting tool 30 to perform cutting processing on a processing object. The "cutting processing" mentioned here includes "turning processing" in which a rotating processing object is cut by the cutting tool 30, and "milling processing" in which a fixed processing object is cut by the rotating cutting tool 30. The machine tool 20 can be a turning machine such as a lathe, or a milling machine such as a milling machine.

[0044] The abnormality detection system 10 includes a cutting tool 30, a wireless device 200, and an abnormality detection apparatus 300. The wireless device 200 is connected to the abnormality detection apparatus 300, for example, by wire. The wireless device 200 is, for example, an access point.

[0045] The cutting tool 30 includes a sensor module 100. As described later, the sensor module 100 includes a sensor.

[0046] It should be noted that the abnormality detection system 10 is not limited to the configuration including one cutting tool 30 , and may include a plurality of cutting tools 30 .

[0047] The cutting tool 30 transmits the measurement results of the sensors in the sensor module 100 to the abnormality detection device 300 in a time series.

[0048] More specifically, the cutting tool 30 wirelessly transmits a wireless signal including a data packet storing the measured value to the wireless device 200 .

[0049] The wireless device 200 acquires a data packet included in the wireless signal received from the cutting tool 30 and relays the data packet to the abnormality detection device 300 .

[0050] When the abnormality detection device 300 receives the sensor data packet from the cutting tool 30 via the wireless device 200 , it acquires the measurement information from the received sensor data packet and processes the acquired measurement information.

[0051] The cutting tool 30 and the wireless device 200 perform wireless communication using a communication protocol such as ZigBee in accordance with IEEE 802.15.4, Bluetooth (registered trademark) in accordance with IEEE 802.15.1, and UWB (Ultra Wide Band) in accordance with IEEE802.15.3a. It should be noted that a communication protocol other than the above may be used between the cutting tool 30 and the wireless device 200.

[0052] [2. Specific examples of cutting tools] Figure 2A This is a diagram showing an example of the structure of the cutting tool according to the present embodiment.

[0053] The turning tool 30A as an example of the cutting tool 30 is a turning tool for machining a rotating object, and is mounted on a machine tool such as a lathe. The turning tool 30A includes a cutting portion 31A and a sensor module 100 provided on the cutting portion 31A.

[0054] For example, a cutting insert 32 having a cutting edge can be mounted on the cutting portion 31A. Specifically, the cutting portion 31A is a tool holder that holds the cutting insert 32. That is, the turning tool 30A is a so-called throwaway turning tool.

[0055] More specifically, the cutting portion 31A includes a fixing member 33A and a fixing member 33B. The fixing members 33A and 33B hold the cutting insert 32 .

[0056] The cutting insert 32 is, for example, a polygonal shape such as a triangle, square, rhombus, or pentagon in a plan view. The cutting insert 32 has, for example, a through hole formed in the center of the upper surface and is fixed to the cutting portion 31A by fixing members 33A and 33B.

[0057] Figure 2B It is a diagram showing another example of the structure of the cutting tool according to the present embodiment.

[0058] The turning tool 30B as an example of the cutting tool 30 is a tool for turning, and is mounted on a machine tool such as a lathe. The turning tool 30B includes a cutting portion 31B and a sensor module 100 provided on the cutting portion 31B.

[0059] For example, the cutting portion 31B has a cutting edge 34. That is, the turning tool 30B is a solid turning tool or a brazed turning tool.

[0060] Figure 2C It is a diagram showing another example of the structure of the cutting tool according to the present embodiment. Figure 2C A cross-sectional view of a cutting tool is shown.

[0061] The milling tool 30C as an example of the cutting tool 30 is a tool for milling processing used for processing a fixed processing object, and is mounted on a machine tool such as a milling machine. The milling tool 30C includes a cutting portion 31C and a sensor module 100 provided on the cutting portion 31C.

[0062] For example, the cutting part 31C can be mounted with a cutting insert 32 having a cutting edge. Specifically, the cutting part 31C is a holder that holds the cutting insert 32. That is, the milling tool 30C is a so-called milling cutter.

[0063] More specifically, the cutting section 31C includes a plurality of fixing members 33C. The fixing members 33C hold the cutting insert 32 .

[0064] The cutting insert 32 is fixed to the cutting portion 31C via a fixing member 33C.

[0065] Figure 2D It is a diagram showing another example of the structure of the cutting tool according to the present embodiment.

[0066] The milling tool 30D as an example of the cutting tool 30 is a tool for milling processing and is mounted on a machine tool such as a milling machine. The milling tool 30D includes a cutting portion 31D and a sensor module 100 provided on the cutting portion 31D.

[0067] For example, the cutting portion 31D has a cutting edge 35. That is, the milling tool 30D is an end mill.

[0068] [3.Sensor module] Figure 3 It is a diagram showing an example of the configuration of a sensor module according to the present embodiment.

[0069] The sensor module 100 includes a processor 101 , a nonvolatile memory 102 , a volatile memory 103 , a communication interface (I / F) 104 , and a strain sensor 110A and a strain sensor 110B.

[0070] The volatile memory 103 is, for example, a volatile memory such as an SRAM (Static Random Access Memory) or a DRAM (Dynamic Random Access Memory). The nonvolatile memory 102 is, for example, a nonvolatile memory such as a flash memory or a ROM (Read Only Memory). In the nonvolatile memory 102, for example, a computer program not shown and data used in the execution of the computer program are stored. The computer program is a program for sending the measurement values ​​of the strain sensor 110A and the strain sensor 110B in a time series.

[0071] The processor 101 is, for example, a CPU (Central Processing Unit). However, the processor 101 is not limited to a CPU. The processor 101 may also be a GPU (Graphics Processing Unit). In a specific example, the processor 101 is a multi-core GPU. The processor 101 may also be, for example, an ASIC (Application Specific Integrated Circuit), or a programmable logic device such as a gate array or an FPGA (Field Programmable Gate Array).

[0072] The communication I / F 104 is a communication interface for a communication protocol such as ZigBee in accordance with IEEE 802.15.4, Bluetooth in accordance with IEEE 802.15.1, and UWB in accordance with IEEE 802.15.3a. The communication I / F 104 is implemented by a communication circuit such as a communication IC (Integrated Circuit).

[0073] The sensor module 100 includes a battery 105. The battery 105 supplies power to the processor 101, the nonvolatile memory 102, the volatile memory 103, the communication I / F 104, and the strain sensors 110A and 110B.

[0074] The strain sensor 110A and the strain sensor 110B are provided, for example, near the cutting edge of the cutting tool 30. The strain sensors 110A and 110B are mounted on the cutting tool 30 in such a manner as to measure strains in different directions. For example, the strain sensor 110A measures the strain in the length direction of the cutting tool 30, and the strain sensor 110B measures the strain in the width direction of the cutting tool 30. Hereinafter, the strain sensor 110A and the strain sensor 110B are also collectively referred to as the strain sensor 110. The strain sensor 110 is an example of a sensor. The sensor is driven by the power supplied from the battery 105.

[0075] It should be noted that the sensor module 100 is not limited to the configuration including two strain sensors 110, and may also include one or more than three strain sensors 110. In addition, the sensor module 100 may also replace the strain sensor 110, or on the basis of the strain sensor 110, further include other sensors such as an acceleration sensor, a pressure sensor, a sound sensor, and a temperature sensor.

[0076] [4. Hardware configuration of abnormality detection device] Figure 4 This is a block diagram showing an example of the hardware configuration of the abnormality detection device according to the present embodiment.

[0077] The abnormality detection device 300 includes a processor 301 , a nonvolatile memory 302 , a volatile memory 303 , an input / output interface (I / O) 304 , a graphic controller 305 , and a display device 306 .

[0078] The volatile memory 303 is, for example, a volatile memory such as SRAM or DRAM. The nonvolatile memory 302 is, for example, a nonvolatile memory such as a flash memory, a hard disk, or a ROM. The nonvolatile memory 302 stores an abnormality detection program 307 as a computer program and data used in the execution of the abnormality detection program 307. Each function of the abnormality detection device 300 is realized by executing the abnormality detection program 307 by the processor 301. The abnormality detection program 307 can be stored in a recording medium such as a flash memory, a ROM, or a CD-ROM.

[0079] The nonvolatile memory 302 stores reference data 308. The reference data 308 is data used by the abnormality detection program 307, and is a time-series measurement result of strain when the cutting tool 30 is normal.

[0080] A measurement result database (DB) 309 is provided in the nonvolatile memory 302. The measurement result DB 309 stores the measurement results of the strain sensor 110. More specifically, in the measurement result DB 309, the measurement values ​​of the strain output from the strain sensor 110 at every certain sampling period are accumulated in a time series. The reference data 308 is, for example, data created based on past measurement results stored in the measurement result DB 309.

[0081] The processor 301 is, for example, a CPU. The processor 301 may be one or more CPUs. However, the processor 301 is not limited to a CPU. The processor 301 may also be a GPU. In a specific example, the processor 301 is a multi-core GPU. The processor 301 may be, for example, an ASIC, or a programmable logic device such as a gate array or an FPGA. In this case, the ASIC or the programmable logic device is configured to be able to perform the same processing as the abnormality detection program 307.

[0082] The I / O 304 is connected to the wireless device 200. The I / O 304 is, for example, a communication I / F, and can communicate based on a specific communication protocol. The I / O 304 includes, for example, an Ethernet interface (“Ethernet” is a registered trademark). The I / O 304 can receive the measurement result of the strain sensor 110 from the cutting tool 30 via the wireless device 200.

[0083] The graphics controller 305 is connected to the display device 306 and controls the display on the display device 306. The graphics controller 305 includes, for example, a GPU and a VRAM (Video RAM), and stores data displayed on the display device 306 in the VRAM, and periodically reads out one frame of image data from the VRAM to generate an image signal. The generated image signal is output to the display device 306, and the image is displayed on the display device 306. The function of the graphics controller 305 may also be included in the processor 301. A part of the area of ​​the volatile memory 303 may also be used as a VRAM.

[0084] The display device 306 includes, for example, a liquid crystal panel or an OEL (organic electroluminescence) panel, and can display text or graphic information.

[0085] [5. Function of abnormality detection device] Figure 5 The abnormality detection device 300 has the functions of an acquisition unit 311 , a synchronization unit 312 , a period setting unit 313 , a correction unit 314 , a distance calculation unit 315 , an abnormality detection unit 316 , and a display control unit 317 .

[0086] The acquisition unit 311 acquires measurement data of the strain of the cutting tool 30 during the period when the machine tool 20 processes the object to be processed by the cutting tool 30. The measurement data is time series data of the measurement value of the strain sensor 110. The strain of the cutting tool 30 is an example of a physical quantity that changes during the period when the machine tool 20 processes the object to be processed by the cutting tool 30.

[0087] Figure 6 This is a graph showing an example of a time series waveform of the strain of the cutting tool. Figure 6 In the figure, the vertical axis represents strain and the horizontal axis represents time. In the cutting process of the object to be processed, multiple processes with different processing conditions are included. For example, in the turning process of a processing object by a lathe, multiple outer diameter processing processes and multiple inner diameter processing processes are included. In a series of cutting processes, the waveform pattern of the measured value of the strain of the cutting tool 30 changes in each process. Figure 6 In the example of FIG. 1 , a cutting process of a processing object includes process A, process B, process C, process D, process E, and process F. Process A is performed during period P1, process B is performed during period P2, process C is performed during period P3, process D is performed during period P4, process E is performed during period P5, and process F is performed during period P6. Processing conditions of process A, process B, process C, process D, process E, and process F are different from each other. Therefore, waveform patterns of strains in process A, process B, process C, process D, process E, and process F are different from each other.

[0088] Return to Figure 5 The synchronization unit 312 synchronizes the object data with the reference data 308, wherein the object data is the time series measurement data acquired by the acquisition unit 311, and the reference data 308 is the time series measurement result of the strain of the cutting tool 30 in a normal state. The magnitude of the strain varies depending on the cutting process, so if the measured value of the strain in a specific process is not evaluated, it is impossible to accurately detect the abnormality of the cutting tool 30. The synchronization unit 312 synchronizes the object data with the reference data 308 in order to compare the object data with the reference data 308 in the same process.

[0089] Figure 7 is a graph showing an example of benchmark data. Figure 8 is a graph showing an example of object data. Figure 7 as well as Figure 8, the vertical axis represents strain, and the horizontal axis represents time. The abnormality detection device 300 compares the object data 400 with the reference data 308 to detect the abnormality of the cutting tool 30. Therefore, it is necessary to synchronize the object data 400 with the reference data 308. It should be noted that synchronizing the object data 400 with the reference data 308 means making the start time of the process in the object data 400 consistent with the start time of the process in the reference data 308, and making the end time of the process in the object data 400 consistent with the end time of the process in the reference data 308. Figure 8 Shown with Figure 7 The reference data 308 shown is synchronized with the object data 400. By synchronizing the object data 400 and the reference data 308 in this way, the waveform of the reference data and the waveform of the object data 400 in the same process can be compared, and abnormality of the cutting tool 30 can be detected with high accuracy.

[0090] Fig. 9 This is a graph showing an example of the measurement results of the strain sensor when the process under the same processing conditions is repeated multiple times. Fig. 9 In the example of FIG. 1 , the measurement results of the strain sensor in three processes, process A1, process A2, and process A3, are shown. Process A1, process A2, and process A3 are processes with the same processing conditions. Fig. 9 In the process A2 and A3, the waveforms are offset due to thermal drift. The waveform offset will be described later. Fig. 9 As shown in FIG. 1 , when the steps under the same processing conditions are repeated a plurality of times, waveform patterns of strains in each of the steps A1 , A2 , and A3 are acquired.

[0091] The synchronization unit 312 determines a cutout period including one process in the object data 400, and extracts (cuts out) a time series waveform pattern of the determined cutout period from the object data 400. The cutout period is a period corresponding to a period indicating a waveform pattern in the reference data 308. Fig. 9 In the example, the synchronization unit 312 determines the cutout periods P_A1, P_A2, and P_A3 of the steps A1, A2, and A3, respectively, and cuts out the time series waveform patterns of the steps A1, A2, and A3 from the object data 400.

[0092] The following describes determination of the cutout period by the synchronization unit 312. The synchronization unit 312 calculates the first moving average and the second moving average of the target data 400 in order to determine the cutout period. Fig.10 is a graph showing an example of the first moving average and the second moving average of the object data 400. Fig.10 In the figure, the vertical axis represents the moving average of strain and the horizontal axis represents time. Fig.10 In the figure, the upper curve graph represents the first moving average, and the lower curve graph represents the second moving average. The number of data points used to calculate the average values ​​of the first moving average and the second moving average is different. For example, the number of data points of the first moving average is 100 points, and the number of data points of the second moving average is 200 points.

[0093] Fig.11 Yes Fig.10 The enlarged curve of the T section in Fig.11 In the figure, the vertical axis represents the moving average of strain and the horizontal axis represents time. Fig.11 In the figure, the black line graph represents the first moving average, and the gray line graph represents the second moving average. Fig.11 As shown in FIG. 1 , the first moving average and the second moving average have different values ​​due to the different number of data points. That is, the first moving average and the second moving average show substantially the same value in the period P_S where the value is approximately constant, and show different values ​​in the period P_C where the value changes. The synchronization unit 312 uses such a change in value to determine the change point of the moving average, that is, the boundary point between the period P_S and the period P_T.

[0094] In a specific example, the synchronization unit 312 calculates the absolute value of the difference between the first moving average and the second moving average (hereinafter referred to as “moving average difference”). Fig.11 The moving average difference in the period P_S is small, and the moving average difference in the period P_T includes a large value. Fig.12 is a graph showing an example of a moving average difference. Fig.12 , the vertical axis represents the moving average difference, and the horizontal axis represents time. The synchronization unit 312 compares the moving average difference with a threshold value (first threshold value), and determines the time point ts0 when the moving average difference exceeds the first threshold value. That is, the synchronization unit 312 compares the moving average difference with the first threshold value in chronological order from the start time of the moving average difference, and determines the time point ts0 when the moving average difference first exceeds the first threshold value.

[0095] The synchronization unit 312 determines a temporary period starting from the determined time point ts0. The temporary period is determined using a registration period pre-registered based on the reference data 308. For example, a user registers a period of a specific process included in the reference data 308 as the registration period. Information on the registration period is stored in the non-volatile memory 302, for example. The synchronization unit 312 determines a period of the same length as the registration period as the temporary period starting from the time point ts0.

[0096] The synchronization unit 312 compares the moving average difference with the first threshold in reverse order of time from the time point te01 as the end point of the temporary period, and determines the time point te02 when the moving average difference exceeds the first threshold for the first time. The synchronization unit 312 determines the period from the time point ts0 to the time point te02 as the process period.

[0097] The synchronization unit 312 determines the cutout period based on the process period. The cutout period is a period for cutting out waveform data from the object data 400. Fig.13 is a graph for explaining an example of determining the cut-off period. Fig.13 In the figure, the vertical axis represents the moving average difference, and the horizontal axis represents time. The synchronization unit 312 determines the time point ts1 that is advanced by the margin from the time point ts0, and determines the time point te1 that is postponed by the margin from the time point te02. The synchronization unit 312 determines the period from the time point ts1 to the time point te1 as the interception period. The margin is determined based on the process period. For example, the margin is 1 / 2 of the difference between the registration period and the process period. However, the margin may also be a predetermined period.

[0098] The interception period as described above is set as a period during which a specific process is performed using the cutting tool 30. In a specific example, the interception period is set as an intermittent processing process in which cutting and non-cutting of the processing object are repeatedly performed. In a continuous processing process in which cutting of the processing object is continuously performed, even if an abnormality occurs in the cutting tool 30, the cutting tool 30 stably generates a certain strain. In contrast, in an intermittent processing process, the cutting tool 30 does not generate strain when not cutting, and generates strain when cutting. That is, the state in which the cutting tool 30 generates strain and the state in which the cutting tool 30 does not generate strain are repeated. Therefore, the abnormality generated in the cutting tool 30 is manifested, for example, as a change in the strain in the cutting tool 30. Therefore, the abnormality of the cutting tool 30 can be accurately detected using the measurement results of the strain in the intermittent processing process.

[0099] Return to Figure 5 The synchronization unit 312 extracts the waveform data of the determined extraction period from the target data 400 . The extracted waveform data (hereinafter referred to as “extracted data”) is synchronized with the reference data 308 .

[0100] The period setting unit 313 sets a target period (hereinafter referred to as a “window”) for the same processing step in the reference data 308 and the cutout data. Figure 7 as well as Figure 8 The window W is a part of a period in a process. The window W is a period common to the reference data 308 and the cut data. For example, the period setting unit 313 sets a part of the period of the reference data 308 as the window W.

[0101] Return to Figure 5 The correction unit 314 corrects the baseline of the clipped data. As described above, the waveform data of the measured value of the strain is offset due to thermal drift. Fig. 9 During cutting, friction heat is generated in the cutting tool 30. The measurement result of the strain sensor 110 changes due to the influence of heat. This is called thermal drift. Fig. 9 As shown, the strain waveform is shifted due to thermal drift. In order to accurately detect abnormality of the cutting tool 30, the correction unit 314 corrects the shift of the strain waveform caused by the thermal drift.

[0102] exist Fig. 9 In the example, the same process is repeatedly performed in one cutting process, and due to thermal drift, the waveform pattern of the strain in the subsequent process A2 and process A3 is offset to the positive side relative to the waveform pattern of the strain in the previous process A1. Similarly, due to thermal drift, the waveform pattern of the strain in the subsequent process A3 is offset to the positive side relative to the waveform pattern of the strain in the previous process A2.

[0103] Fig.14A is a graph showing an example of the clipped data before correction. Fig. 14B is a graph showing an example of corrected clipped data. Fig.14A as well as Fig. 14B In the figure, the vertical axis represents strain and the horizontal axis represents time. Fig.14A as well as Fig. 14B In FIG. 4 , the black line graph represents the reference data 308, and the gray line graph represents the clipped data 410. For example, the reference data 308 is corrected data. That is, the reference data 308 is corrected in advance for the waveform deviation caused by thermal drift.

[0104] In a specific example, the correction unit 314 calculates the average value of the strain measurement values ​​of the intercepted data 410 of the window W, and corrects the baseline of the intercepted data 410 by subtracting the average value from each measurement value of the intercepted data 410 of the window W. Fig. 14B As shown, the baseline of the corrected clipped data 410A matches the baseline of the reference data 308. It should be noted that the baseline of the reference data 308 may not be corrected in advance. In this case, the correction unit 314 corrects the baseline of the reference data 308 by subtracting the average value from each measurement value of the reference data 308 of the window W.

[0105] Return to Figure 5The distance calculation unit 315 calculates the distance between the reference data 308 and the object data 400 (the clipped data 410A) synchronized by the synchronization unit 312. In a specific example, the distance calculation unit 315 calculates the distance between the reference data 308 and the clipped data 410A in the window W set by the period setting unit 313. The distance calculation unit 315 calculates the distance between the clipped data 410A and the reference data 308 after the baseline is corrected by the correction unit 314.

[0106] An example of the distance between the reference data 308 and the clipped data 410A is the distance between the time series waveform of the reference data 308 and the time series waveform of the clipped data 410A. In a more specific example, the distance between the reference data 308 and the clipped data 410A is the DTW (dynamic time warping) between the reference data 308 and the clipped data 410A.

[0107] Fig.15 410A is a graph for explaining the DTW between the reference data 308 and the clipped data 410A. Fig.15 In the figure, the vertical axis represents strain and the horizontal axis represents time. Fig.15 In FIG. 4 , the upper graph represents the reference data 308, and the lower graph represents the cut data 410A. Fig.15 In the figure, for ease of understanding, the reference data 308 and the clipped data 410A are shown in different coordinate systems, but in fact, the DTW between the waveform of the reference data 308 and the waveform of the clipped data 410A on a common coordinate axis (the axis of strain and the time axis) is calculated. In the calculation of DTW, the distance calculation unit 315 cyclically calculates the distance (absolute value of error) between each point of the reference data 308 and each point of the clipped data 410A, and determines the minimum value of the calculated distances as the distance between the reference data 308 and the clipped data 410A. The determined distance is set as the abnormality of the cutting tool 30.

[0108] Return to Figure 5 The abnormality detection unit 316 detects the abnormality of the cutting tool 30 by comparing the distance (abnormality) calculated by the distance calculation unit 315 with the threshold value (second threshold value). Specifically, the abnormality detection unit 316 determines that the cutting tool 30 has an abnormality when the abnormality exceeds the second threshold value, and determines that the cutting tool 30 has no abnormality (normal) when the abnormality is below the second threshold value. The second threshold value is determined based on the position of the window W. However, the second threshold value can also be preset. For example, the second threshold value is determined by Hotelling's T 2 However, the second threshold value may also be determined using other methods.

[0109] When the abnormality degree of a window W is compared with the second threshold value, the period setting unit 313 sets a period slightly offset from the original window W as a new window W. In the new window W, the cutout data 410 is cut out from the object data 400, the correction unit 314 corrects the baseline of the cutout data 410, the distance calculation unit 315 calculates the distance between the reference data 308 and the cutout data 410A, and the abnormality detection unit 316 compares the abnormality degree with the second threshold value. For example, a period for abnormality detection (hereinafter referred to as "detection object period") is pre-set in the object data 400, and the period setting unit 313 can sequentially set the window W from the start point to the end point of the detection object period, each time offset by a predetermined time. Thus, the abnormality degree is calculated sequentially throughout the detection object period, and the abnormality detection of the cutting tool 30 is performed. The detection object period is set to a part or all of the cutout period, for example.

[0110] The display control unit 317 outputs the video signal to the display device 306 and controls the display of the display device 306. When the abnormality detection unit 316 detects an abnormality in the cutting tool 30, the display control unit 317 causes the display device 306 to display that the abnormality has been detected. Thus, the user is notified of the abnormality detection.

[0111] The display control unit 317 can display a time series graph of the abnormality degree calculated by the distance calculation unit 315 (hereinafter referred to as “abnormality degree graph”) on the display device 306 . Fig.16A is a graph showing an example of an abnormality curve graph. Fig. 16B is a graph showing another example of an abnormality curve graph. Fig.16A as well as Fig. 16B In the figure, the vertical axis represents the abnormality and the horizontal axis represents the time. Fig.16A is an abnormality curve diagram when the cutting tool 30 is normal. Fig. 16B is an abnormality curve diagram when the cutting tool 30 is abnormal. Fig.16A as well as Fig. 16B The abnormality degree graph shown here allows the user to check the temporal transition of the abnormality degree. By displaying the time series graph of the abnormality degree, when the cutting tool 30 wears and changes from a normal state to an abnormal state, the user can check how the state changes.

[0112] [6. Operation of abnormality detection device] When the processor 301 of the abnormality detection device 300 executes the abnormality detection program 307 , the processor 301 performs the abnormality detection process described below. Fig.17 This is a flowchart showing an example of an abnormality detection process performed by the abnormality detection device according to the present embodiment.

[0113] While the cutting tool 30 is used and the object to be processed is cut by the machine tool 20, the strain sensor 110 measures the strain of the cutting tool 30. The sensor module 100 wirelessly transmits the measurement result of the strain sensor 110 to the abnormality detection device 300. The abnormality detection device 300 receives the measurement result of the strain and stores the measurement result in the measurement result DB 309. In this way, the time series measurement data is accumulated in the measurement result DB 309.

[0114] The processor 301 acquires the object data 400 which is the measurement data to be detected as abnormality from the measurement result DB (step S101 ).

[0115] The processor 301 executes a synchronization process of synchronizing the acquired object data 400 with the reference data 308 (step S102 ).

[0116] Fig.18 This is a flowchart showing an example of synchronization processing.

[0117] In the synchronization process, the processor 301 calculates the first moving average and the second moving average of the object data 400, and calculates the difference between the moving averages (step S201). Next, the processor 301 compares the difference between the moving averages with a first threshold value, and determines the time point ts0 when the difference between the moving averages exceeds the first threshold value (step S202).

[0118] The processor 301 uses the registered period to determine the temporary period with the determined time point ts0 as the starting point (step S203). The processor 301 compares the moving average difference with the first threshold in reverse order of time from the time point te01 as the end point of the temporary period, and determines the time point te02 when the moving average difference exceeds the first threshold for the first time. The processor 301 determines the period from the time point ts0 to the time point te02 as the process period (step S204).

[0119] The processor 301 adds margins before and after the process period and determines the cutout period (step S205 ). The processor 301 cuts out waveform data (cutout data) during the cutout period from the object data 400 (step S206 ). The synchronization process ends as described above.

[0120] Return to Fig.17 The processor 301 sets a window W in the synchronized cut-out data 410 and the reference data 308 (step S103). The first window W is set to the beginning of the detection target period.

[0121] The processor 301 calculates the average value of the measured values ​​of the clipped data 410 of the window W, and subtracts the average value from each measured value. In this way, the baseline of the clipped data 410 is corrected (step S104 ).

[0122] The processor 301 calculates the distance between the corrected clipped data 410A and the reference data 308 as the abnormality (step S105 ).

[0123] Processor 301 compares the abnormality level with the second threshold value (step S106 ). If the abnormality level is less than or equal to the second threshold value (No in step S106 ), processor 301 proceeds to step S109 .

[0124] When the abnormality degree exceeds the second threshold value (YES in step S106 ), the processor 301 determines that an abnormality has occurred in the cutting tool 30. That is, the processor 301 detects an abnormality in the cutting tool 30 (step S107 ).

[0125] The processor 301 causes the display device 306 to display that an abnormality of the cutting tool 30 is detected (step S108 ). Thus, the user is notified that an abnormality is detected.

[0126] The processor 301 determines whether the window W reaches the end of the detection target period (step S109). If the window W does not reach the end of the detection target period (No in step S109), the processor 301 returns to step S103 and sets the period after a predetermined time from the window W as a new window W. The processor 301 executes the steps after step S104. By sequentially updating the window W, the abnormality is calculated throughout the detection target period, and it is determined whether the cutting tool 30 is abnormal.

[0127] When window W reaches the end of the detection target period (YES in step S109 ), processor 301 creates an abnormality curve graph of the detection target period and displays the abnormality curve graph on display device 306 (step S110 ). The abnormality detection process ends as described above.

[0128] [7. Modifications] In the above-mentioned embodiment, the interception period is set in the intermittent processing step, and the abnormality of the cutting tool in the intermittent processing step is detected, but it is not limited to this. The interception period can also be set in the continuous processing step, and the abnormality of the cutting tool in the continuous processing step is detected. In addition, the period including one or more intermittent processing steps and one or more continuous processing steps can also be set as the interception period, and the abnormality of the cutting tool in the interception period can be detected.

[0129] In the above-mentioned embodiment, as the degree of abnormality, the distance between the time series waveform of the reference data and the time series waveform of the intercepted data (object data) is calculated, but it is not limited to this. As the degree of abnormality, the difference between the distribution of the reference data and the distribution of the intercepted data (object data) can also be calculated. The distance between the time series waveform of the reference data and the time series waveform of the intercepted data is not limited to DTW. For example, it can also be the Euclidean distance. In the case where the Euclidean distance is set as the degree of abnormality, the phase of the time series waveform of the intercepted data can also be made consistent with the phase of the time series waveform of the reference data. As the difference between the distribution of the reference data and the distribution of the intercepted data, the cross entropy can be calculated, and the density ratio can also be estimated.

[0130] In the above-mentioned embodiment, the abnormality of the cutting tool 30 is detected using the measurement data of the strain of the cutting tool 30 by the strain sensor 110, but the present invention is not limited thereto. An acceleration sensor may be provided on the cutting tool 30, and the abnormality of the cutting tool 30 may be detected using the time series acceleration measured by the acceleration sensor. When the acceleration sensor is used, the difference between the distribution of the reference data of the acceleration and the distribution of the target data of the acceleration can be calculated as the abnormality degree.

[0131] [8. Additional explanation] The embodiments disclosed this time are illustrative in all aspects and are not restrictive. The scope of rights of the present invention is indicated by the claims rather than the embodiments described above, and includes all modifications within the meaning and scope equivalent to the claims.

[0132] Description of Reference Numerals 10: Anomaly detection system; 20: Machine tools; 30, 30A, 30B, 30C, 30D: turning tools; 31A, 31B, 31C, 31D: cutting part; 32: cutting blade; 33A, 33B, 33C: fixing components; 34, 35: cutting edge; 100: sensor module; 101: Processor; 102: non-volatile memory; 103: Volatile memory; 104: communication interface (communication I / F); 105: Battery; 110, 110A, 110B: strain sensors; 200: wireless devices; 300: anomaly detection device; 301: processor; 302: non-volatile memory; 303: Volatile memory; 304: Input and output interface (I / O); 305: Graphics controller; 306: display device; 307: Anomaly detection program; 308: Benchmark data; 309: measurement result DB; 311: Acquisition Department; 312: Synchronization Department; 313: Period Setting Department; 314: Correction Department; 315: distance calculation unit; 316: Abnormality detection unit; 317: display control unit; 400: object data; 410, 410A: intercept data; A, B, C, D, E, F, A1, A2, A3: process; P1, P2, P3, P4, P5, P6, P_S, P_C: period; P_A1, P_A2, P_A3: interception period; ts0, ts1, te01, te02, te1: time points; W: Window.

Claims

1. An abnormality detection system for detecting abnormalities of a tool of a machine tool, wherein: The anomaly detection system comprises: a sensor provided on the tool for measuring a physical quantity that changes while the tool is processing an object to be processed by the machine tool; and an abnormality detection device for detecting an abnormality of the tool based on the measurement data of the sensor, The abnormality detection device comprises: a synchronization unit that synchronizes object data with reference data, wherein the object data is time-series measurement data of the sensor, and the reference data is time-series measurement results of the physical quantity when the tool is normal; a distance calculation unit that calculates a distance between the reference data synchronized by the synchronization unit and the target data; as well as The abnormality detecting unit detects abnormality of the tool by comparing the distance calculated by the distance calculating unit with a threshold value.

2. The abnormality detection system according to claim 1, wherein: The abnormality detection device further includes a period setting unit that sets a target period of the same processing step in the reference data and the target data. The distance calculation unit calculates the distance between the reference data and the target data within the target period set by the period setting unit.

3. The abnormality detection system according to claim 2, wherein: The processing step is an intermittent processing step of repeatedly performing cutting and non-cutting of the processing object.

4. The abnormality detection system according to any one of claims 1 to 3, wherein: The sensor is a strain sensor that measures strain of the tool as the physical quantity.

5. The abnormality detection system according to any one of claims 1 to 3, wherein: The sensor is an acceleration sensor that measures acceleration of the tool as the physical quantity.

6. The abnormality detection system according to any one of claims 1 to 5, wherein: The abnormality detection device further includes a correction unit, which corrects the baseline of the object data. The distance calculation unit calculates a distance between the object data after the baseline has been corrected by the correction unit and the reference data.

7. The abnormality detection system according to any one of claims 1 to 6, wherein: The distance between the reference data and the object data is the distance between the time series waveform of the reference data and the time series waveform of the object data, or the difference between the distribution of the reference data and the distribution of the object data.

8. The abnormality detection system according to any one of claims 1 to 7, wherein: The abnormality detection system further includes a display device that displays a time series graph of the distance between the reference data and the target data calculated by the distance calculation unit.

9. An abnormality detection device for detecting abnormalities of a tool of a machine tool, wherein: The abnormality detection device comprises: an acquisition unit that acquires time-series measurement data output from a sensor provided on the tool, the sensor measuring a physical quantity that changes while the tool is processing an object to be processed by the machine tool; a synchronization unit that synchronizes object data with reference data, wherein the object data is the time-series measurement data acquired by the acquisition unit, and the reference data is the time-series measurement result of the physical quantity when the tool is normal; a distance calculation unit that calculates a distance between the reference data synchronized by the synchronization unit and the target data; as well as The abnormality detecting unit detects abnormality of the tool by comparing the distance calculated by the distance calculating unit with a threshold value.

10. An abnormality detection method, performed by an abnormality detection device for detecting an abnormality of a tool of a machine tool, wherein: The anomaly detection method comprises the following steps: acquiring time-series measurement data output from a sensor provided on the tool, the sensor measuring a physical quantity that changes while the tool is processing an object to be processed by the machine tool; Synchronizing object data and reference data, wherein the object data is the acquired time series measurement data, and the reference data is the time series measurement result of the physical quantity when the tool is normal; calculating a distance between the synchronized reference data and the object data; and Anomalies of the tool are detected by comparing the calculated distance with a threshold value.

11. A computer program for detecting an abnormality of a tool of a machine tool, wherein: The computer program is used to make the computer execute the following steps: acquiring time-series measurement data output from a sensor provided on the tool, the sensor measuring a physical quantity that changes while the tool is processing an object to be processed by the machine tool; Synchronizing object data and reference data, wherein the object data is the acquired time series measurement data, and the reference data is the time series measurement result of the physical quantity when the tool is normal; calculating a distance between the synchronized reference data and the object data; and Anomalies of the tool are detected by comparing the calculated distance with a threshold value.

Images