JP2019159759A - Processing section determination device, processing section determination method, program, and diagnostic device - Google Patents

Abstract

To provide a processing section determination device, a processing section determination method, a program, and a diagnosis device that can accurately grasp an actual processing section.SOLUTION: A processing section determination device is comprised of: a first acquisition unit that acquires detection information output from a detection unit that detects a physical quantity that changes according to an operation of a target device; a second acquisition unit that acquires a signal indicating a section in which the target device is operated from the target device; a first identification unit that identifies a non-processing section and a processing section in the target device based on the detection information and the signal; and a second identification unit for identifying a non-processing section and a processing section in the target device based on the detection information, the signal, and the identification result by the first identification unit.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a processing section determination device, a processing section determination method, a program, and a diagnosis apparatus.

  As a method for predicting and estimating tool abnormalities and machining quality when machining a workpiece by a machine tool or the like, there is already a method of detecting and using a physical quantity such as information on the current value of the motor of the machine, vibration, or force. Are known.

  As a device that detects such physical quantities and determines tool abnormalities, etc., it detects the abnormality of machining operations by synchronizing the time series data of machine information such as spindle load and the event data of the device. Is known (Patent Document 1).

  However, with the technique described in Patent Document 1, it is difficult to clearly grasp the actual processing section from the event data of the apparatus.

  The present invention has been made in view of the above-described problems, and provides a processing section determination device, a processing section determination method, a program, and a diagnosis device that can accurately grasp an actual processing section. Objective.

  In order to solve the above-described problems and achieve the object, the present invention includes a first acquisition unit that acquires detection information output from a detection unit that detects a physical quantity that changes according to the operation of the target device; A second acquisition unit that acquires a signal indicating a section in which the target device is operated from the target device; and a first identification that identifies a non-processing section and a processing section in the target device based on the detection information and the signal And a second identification unit that identifies a non-processing section and a processing section in the target device based on the detection information, the signal, and the identification result by the first identification unit.

  According to the present invention, it is possible to accurately grasp an actual processing section.

FIG. 1 is a functional block diagram illustrating an example of the overall configuration of the diagnostic system according to the embodiment. FIG. 2 is a diagram illustrating an example of a hardware configuration of the processing machine according to the embodiment. FIG. 3 is a diagram illustrating an example of a hardware configuration of the diagnostic apparatus according to the embodiment. FIG. 4 is a diagram illustrating an example of detection information and ladder signals of the processing machine according to the embodiment. FIG. 5 is a diagram illustrating feature information extracted from the detection information by the diagnostic device according to the embodiment using frequency components. FIG. 6 is a diagram for explaining an example in which the diagnostic device according to the embodiment selects feature information of a non-machining section and a machining section. FIG. 7 is a diagram illustrating feature information selected from the non-processing section and the processing section by the diagnosis apparatus according to the embodiment as a frequency spectrum. FIG. 8 is a diagram illustrating an example in which the diagnostic device according to the embodiment calculates the likelihood of the machining section from the Euclidean distance. FIG. 9 is a diagram illustrating an operation of estimating the machining section from the machining section likelihood calculated by the diagnosis apparatus according to the embodiment. FIG. 10 is a flowchart illustrating an example of a procedure of abnormality determination processing at the initial stage of machining by the diagnostic device according to the embodiment. FIG. 11 is a flowchart illustrating an example of a processing section estimation process performed by the first identification unit according to the embodiment. FIG. 12 is a flowchart illustrating an example of a procedure of abnormality determination processing after a predetermined period has elapsed by the diagnostic device according to the embodiment. FIG. 13 is a flowchart illustrating an example of a processing section estimation process performed by the second identification unit according to the embodiment. FIG. 14 is a schematic diagram illustrating an example in which processing in the same process is continuously performed a plurality of times in the processing machine according to the embodiment.

  Hereinafter, embodiments of a processing section determination device, a processing section determination method, a program, and a diagnosis apparatus according to the present invention will be described in detail with reference to FIGS. Further, the present invention is not limited by the following embodiments, and components in the following embodiments can be easily conceived by those skilled in the art, are substantially the same, and have a so-called equivalent range. Is included. Furthermore, various omissions, substitutions, changes, and combinations of the components can be made without departing from the scope of the following embodiments.

(Overall configuration of diagnostic system)
Drawing 1 is a figure showing an example of the whole composition of diagnostic system 1 concerning an embodiment by a functional block. The diagnosis device 100 as a processing section determination device provided in the diagnosis system 1 acquires first detection information output from the detection unit 225 that detects a physical quantity that changes according to the operation of the processing machine 200 as a target device. Detection information receiving unit 112 as a unit, a processing information acquisition unit 101 as a second acquisition unit that acquires a ladder signal as a signal indicating a section in which the processing machine 200 is operated, from the processing machine 200, and detection information and signals Based on the first identification unit 104 for identifying the non-processing section and the processing section in the processing machine 200, the detection information, the signal, and the identification result by the first identification unit, the non-processing section and the processing in the processing machine 200 A second identification unit 105 that identifies a section. Below, the detail of the diagnostic system 1 of embodiment is demonstrated.

  As illustrated in FIG. 1, the diagnosis system 1 according to the embodiment includes a processing machine 200 and a diagnosis device 100 connected to the processing machine 200. The processing machine 200 is a machine tool that performs processing such as cutting, grinding, or polishing on a processing target using a tool. The processing machine 200 is an example of a target device to be diagnosed by the diagnostic device 100. The diagnosis apparatus 100 is an apparatus that is connected so as to be communicable with the processing machine 200 and diagnoses an abnormality in the operation of the processing machine 200.

  The processing machine 200 includes a numerical control unit 201, a communication control unit 221, a drive control unit 223, a drive unit 224, and a detection unit 225.

  The numerical control unit 201 is a functional unit that executes processing by the driving unit 224 by numerical control (NC: Numerical Control). For example, the numerical control unit 201 generates and outputs numerical control data for controlling the operation of the driving unit 224. Further, the numerical control unit 201 uses, as context information (machining information) indicating the operation state of the driving unit 224 that drives the tool, for example, a section from the feed operation to the machining target of the tool until the actual machining process ends (tool A ladder signal, which is an ON / OFF signal indicating a feed section), is output to the communication control unit 221. The context information is information determined in plural for each type of operation of the processing machine 200. In addition to the ladder signal, the context information includes, for example, identification information of the processing machine 200, identification information of the driving unit 224 (for example, tool identification information), a diameter of the tool driven by the driving unit 224, and the tool Configuration information such as material, etc., the operating state of the drive unit 224, the accumulated usage time from the start of use of the drive unit 224, the load applied to the drive unit 224, the rotational speed of the drive unit 224, the processing speed of the drive unit 224, etc. Information indicating processing condition information or the like may be included.

  For example, the numerical control unit 201 sequentially transmits context information corresponding to the current operation of the processing machine 200 to the diagnostic apparatus 100 via the communication control unit 221. The numerical control unit 201 changes the type of the driving unit 224 to be driven or the driving state (the rotational speed, the rotation speed, etc.) of the driving unit 224 according to the processing step when processing the processing target. Each time the numerical control unit 201 changes the operation type, the numerical control unit 201 sequentially transmits context information corresponding to the changed operation type to the diagnostic apparatus 100 via the communication control unit 221.

  The communication control unit 221 is a functional unit that controls communication with an external device such as the diagnostic device 100. For example, the communication control unit 221 transmits context information corresponding to the current operation to the diagnostic apparatus 100.

  The drive control unit 223 is a functional unit that drives and controls the drive unit 224 based on the numerical control data obtained by the numerical control unit 201.

  The drive unit 224 is a functional unit that is a target of drive control by the drive control unit 223. The drive unit 224 drives the tool under the control of the drive control unit 223. The drive unit 224 is an actuator (motor) that is driven and controlled by the drive control unit 223. The driving unit 224 may be any actuator as long as it is used for machining and is subject to numerical control. Two or more drive units 224 may be provided.

  The detection unit 225 is a functional unit that detects a physical quantity generated by the processing machine 200 and outputs information on the detected physical quantity to the diagnosis apparatus 100 as detection information (sensor data). The physical quantity generated by the processing machine 200 is vibration or sound generated by the processing machine 200. Such vibration or sound is generated, for example, when a tool installed in the processing machine 200 and a processing target come into contact during the processing operation. Alternatively, such vibration or sound is generated by the tool or the processing machine 200 itself. The number of detection units 225 is arbitrary. For example, a plurality of detection units 225 that detect the same physical quantity may be provided, or a plurality of detection units 225 that detect different physical quantities may be provided. For example, when a blade that is a tool used for machining breaks, a chipping of the blade, or the like occurs, vibration and sound during machining change. For this reason, it is possible to detect an abnormality in the operation of the processing machine 200 by detecting vibration data and acoustic data with the detection unit 225 and making a determination using a model that determines normal vibration and sound.

  The diagnostic apparatus 100 includes a communication control unit 111, a detection information reception unit 112, a processing information acquisition unit 101, a feature extraction unit 102, a generation unit 103, a first identification unit 104, and a second identification unit 105. An abnormality determination unit 106, a storage unit 113, an input unit 114, a display control unit 107, and a display unit 115.

  The communication control unit 111 is a functional unit that controls communication with the processing machine 200. For example, the communication control unit 111 receives context information from the numerical control unit 201 of the processing machine 200 via the communication control unit 221.

  The detection information receiving unit 112 as a first acquisition unit is a functional unit that receives detection information from the detection unit 225 installed in the processing machine 200.

  The processing information acquisition unit 101 as a second acquisition unit is a functional unit that acquires context information (processing information) received by the communication control unit 111 from the processing machine 200.

  The feature extraction unit 102 as an extraction unit is a functional unit that extracts feature information used in various determinations from detection information. For example, the feature information is used when identifying a non-machining section and a machining section from a section where a ladder signal included in the context information is ON, that is, a tool feed section. The feature information acquired when the processing machine 200 is normal is used when a model is generated by the generation unit 103 described later. Further, the feature information in the machining section acquired for determining the abnormality of the processing machine 200 is compared with the model generated by the generation unit 103 and used for abnormality determination.

  The generation unit 103 is a functional unit that generates models used for various determinations from the above-described feature information. For example, a model generated from the feature information of the non-processed section is used when identifying the non-processed section and the processed section. In addition, the model generated from the feature information of the machining section acquired when the processing machine 200 is normal is used for determining that the machining has been normally performed.

  The first identification unit 104 is a functional unit that identifies a machining section from a non-machining section at the initial stage of machining by the processing machine 200. For a certain period after the processing machine 200 is started up or after tool change, it cannot be said that there is sufficient data for identifying the machining section from the non-machining section, such as feature information and a model generated from the feature information. The first identification unit 104 estimates a machining section using a small amount of data. In order to realize such a function, the first identification unit 104 includes a selection unit 104a, a calculation unit 104b, and a section determination unit 104c.

  The selection unit 104 a is a functional unit that selects feature information for each predetermined frame from the feature information extracted by the feature extraction unit 102. The selection unit 104a selects feature information of a non-machining section in the tool feed section. The selection unit 104a also selects, as target feature information, feature information of a section that may be a machining section among tool feed sections in order to compare with feature information of a non-machining section. A model is generated by the generation unit 103 from the feature information of the selected non-processed section. Such a model is used for identifying a machining section as a model of a non-machining section as the first model.

  The calculation unit 104b compares the model of the non-machined section generated from the feature information of the non-machined section selected by the selection unit 104a with the target feature information, and calculates the likelihood of the machined section of each target feature information Part. As a comparison between the model and the target feature information, for example, a method for obtaining an outlier from the model of the target feature information is used.

  The section determination unit 104c is a functional unit that performs threshold determination on the likelihood of the machining section calculated by the calculation unit 104b. The feature information of the section determined as the processing section is determined as the feature information of the processing section. From this feature information, for example, a model is generated by the generation unit 103. Such a model is used for abnormality determination as a model of a machining section.

  The second identification unit 105 is a functional unit that identifies a non-machining section and a machining section after sufficient data for identifying the non-machining section and the machining section is obtained. In order to realize such a function, the second identification unit 105 includes a selection unit 105a, a calculation unit 105b, a section determination unit 105c, and a feature information collection unit 105d.

  The feature information collection unit 105d collects feature information of the non-processing section and the processing section extracted by the feature extraction unit 102 in the initial stage of processing by the processing machine 200. After sufficient feature information is collected, a non-machined section model as a second model and a machined section model as a third model are generated from the feature information by the generating unit 103, respectively. These models are used to identify non-machined sections and machined sections.

  The selection unit 105 a is a functional unit that selects feature information for each predetermined frame from the feature information extracted by the feature extraction unit 102. The selection unit 105a selects feature information of a non-machining section in the tool feed section. The selection unit 105a also selects, as target feature information, feature information of a section that may be a machining section among tool feed sections in order to compare with feature information of a non-machining section.

  The calculation unit 105b is a functional unit that compares the model generated from the feature information collected by the feature information collection unit 105d with the target feature information and calculates the processing section likelihood of each target feature information. As a comparison between the model generated from the feature information of the non-processed section and the target feature information, for example, a method for obtaining an outlier from the model of the target feature information is used. As a comparison between the model generated from the feature information of the machining section and the target feature information, for example, a method for obtaining an approximate value of the model of the target feature information is used.

  The section determination unit 105c is a functional unit that performs threshold determination on the likelihood of the machining section calculated by the calculation unit 105b. The feature information of the section determined as the processing section is determined as the feature information of the processing section.

  The abnormality determination unit 106 uses the feature information to be subjected to abnormality determination extracted by the feature extraction unit 102 and the model for each machining section generated by the generation unit 103, and the operation of the processing machine 200 is normal. It is a functional unit that determines whether or not there is.

  The storage unit 113 is a functional unit that stores the model generated by the generation unit 103 in association with the context information. In addition, the storage unit 113 stores the detection information acquired by the detection information reception unit 112 in association with the context information.

  The input unit 114 is a functional unit that performs operations such as inputting characters and numbers, selecting various instructions, and moving a cursor.

  The display control unit 107 is a functional unit that controls the display operation of the display unit 115. Specifically, the display control unit 107 causes the display unit 115 to display the result of abnormality determination by the abnormality determination unit 106, for example. The display unit 115 is a functional unit that displays various types of information according to control by the display control unit 107.

  In addition, each function part of the diagnostic apparatus 100 and the processing machine 200 has shown the function notionally, and is not limited to such a structure. For example, a plurality of functional units illustrated as independent functional units in FIG. 1 may be configured as one functional unit. On the other hand, the function of one function unit in FIG. 1 may be divided into a plurality of functions and configured as a plurality of function units.

  Moreover, the processing machine 200 and the diagnostic apparatus 100 may be connected in any connection form. For example, the processing machine 200 and the diagnostic apparatus 100 may be connected by a dedicated connection line, a wired network such as a wired LAN (Local Area Network), a wireless network, or the like.

  FIG. 1 shows an example in which one processing machine 200 is connected to the diagnostic apparatus 100. However, the present invention is not limited to this, and a plurality of processing machines 200 are connected to the diagnostic apparatus 100. And may be connected so that they can communicate with each other.

(Processing machine hardware configuration)
Next, a hardware configuration example of the processing machine 200 according to the embodiment will be described with reference to FIG. FIG. 2 is a diagram illustrating an example of a hardware configuration of the processing machine 200 according to the embodiment.

  As shown in FIG. 2, the processing machine 200 includes a CPU (Central Processing Unit) 20, a ROM (Read Only Memory) 20a, a RAM (Random Access Memory) 20b, a communication I / F (interface) 21, and a drive. The control circuit 23 is connected to be communicable with the bus 2B.

  The CPU 20 is an arithmetic device that controls the entire processing machine 200. For example, the CPU 20 executes a program stored in the ROM 20a or the like using the RAM 20b as a work area (working area), thereby controlling the operation of the entire processing machine 200 and realizing a processing function. The numerical control unit 201 in FIG. 1 is realized by, for example, a program that runs on the CPU 20.

  The communication I / F 21 is an interface used for communication with an external device such as the diagnostic device 100. The communication I / F 21 is, for example, a NIC (Network Interface Card) corresponding to TCP (Transmission Control Protocol) / IP (Internet Protocol). The communication control unit 221 in FIG. 1 is realized by, for example, the communication I / F 21 and a program that operates on the CPU 20.

  The drive control circuit 23 is a circuit that controls driving of the motor 24. The motor 24 drives a tool 24a used for processing. The tool 24a includes a drill, an end mill, a bite tip, a grindstone, and a table on which a processing target is placed and moved according to processing. The drive control unit 223 in FIG. 1 is realized by the drive control circuit 23, for example. The drive unit 224 in FIG. 1 is realized by a motor 24, for example.

  The sensor 25 is composed of, for example, a microphone device, a vibration sensor, an acceleration sensor, an AE (Acoustic Emission) sensor, or the like, and is installed in the vicinity of a tool capable of detecting vibration or sound, for example. The sensor amplifier 25a to which the sensor 25 is connected is connected to the diagnostic device 100 so as to be communicable. The sensor 25 and the sensor amplifier 25a may be provided in the processing machine 200 in advance, or may be attached later to the processing machine 200 that is a completed machine. In addition, the sensor amplifier 25a is not limited to being installed on the processing machine 200, and may be installed on the diagnostic device 100 side. The detection unit 225 in FIG. 1 is realized by, for example, the sensor 25 and the sensor amplifier 25a.

  Note that the hardware configuration shown in FIG. 2 is an example, and the processing machine 200 does not have to include all the component devices, and may include other component devices. For example, the numerical control unit 201 and the communication control unit 221 illustrated in FIG. 1 may be implemented by causing the CPU 20 illustrated in FIG. 2 to execute a program, that is, by software, or by hardware such as an IC (Integrated Circuit). Alternatively, it may be realized by using a combination of software and hardware.

(Hardware configuration of diagnostic device)
Next, a hardware configuration example of the diagnostic apparatus 100 according to the embodiment will be described with reference to FIG. FIG. 3 is a diagram illustrating an example of a hardware configuration of the diagnostic apparatus 100 according to the embodiment.

  As illustrated in FIG. 3, the diagnostic device 100 includes a CPU 10, a ROM 10 a, a RAM 10 b, a communication I / F 11, a sensor I / F 12, an auxiliary storage device 13, an input device 14, and a display 15. 1B is configured to be communicable.

  The CPU 10 is an arithmetic device that controls the entire diagnostic device 100. For example, the CPU 10 executes a program stored in the ROM 10a or the like using the RAM 10b as a work area (working area), thereby controlling the operation of the entire diagnostic device 100 and realizing a diagnostic function. The processing information acquisition unit 101, the feature extraction unit 102, the generation unit 103, the first identification unit 104, the second identification unit 105, the abnormality determination unit 106, and the display control unit 107 in FIG. Realized by the program.

  The communication I / F 11 is an interface used for communication with an external device such as the processing machine 200. The communication I / F 11 is, for example, a NIC that supports TCP / IP. The communication control unit 111 in FIG. 1 is realized by, for example, the communication I / F 11 illustrated in FIG.

  The sensor I / F 12 is an interface that receives detection information from the sensor 25 installed in the processing machine 200 via the sensor amplifier 25a. The detection information receiving unit 112 in FIG. 1 is realized by, for example, a program that operates on the sensor I / F 12 and the CPU 10.

  The auxiliary storage device 13 is an HDD (Hard Disk Drive), SSD that stores various data such as setting information of the diagnostic device 100, detection information and context information received from the processing machine 200, OS (Operating System), and application programs. (Solid State Drive) or EEPROM (Electrically Erasable Programmable Read-Only Memory) or the like. The auxiliary storage device 13 is included in the diagnostic device 100, but is not limited thereto. For example, the auxiliary storage device 13 may be a storage device installed outside the diagnostic device 100, or the diagnostic device. 100 may be a storage device included in a server device capable of data communication with 100. The storage unit 113 in FIG. 1 is realized by, for example, the RAM 10b, the auxiliary storage device 13, and the like.

  The input device 14 is an input device such as a mouse or a keyboard that performs operations such as inputting characters and numbers, selecting various instructions, and moving a cursor. The input unit 114 in FIG. 1 is realized by the input device 14, for example.

  The display 15 is a display device such as a CRT (Cathode Ray Tube) display, an LCD (Liquid Crystal Display), or an organic EL (Electro-Luminescence) display that displays characters, numbers, various screens, operation icons, and the like. The display unit 115 in FIG. 1 is realized by, for example, the display 15.

  Note that the hardware configuration illustrated in FIG. 3 is an example, and the diagnosis apparatus 100 does not have to include all the component devices, and may include other component devices. For example, each functional unit (processing information acquisition unit 101, feature extraction unit 102, generation unit 103, first identification unit 104, second identification unit 105, abnormality determination unit 106, and diagnosis unit 100 shown in FIG. The display control unit 107) causes the CPU 10 to execute a program, that is, may be realized by software, may be realized by hardware such as an IC, or may be realized by using software and hardware together. Also good. Further, when the diagnostic device 100 specializes in the diagnostic operation of the processing machine 200 and transmits the diagnostic result to an external server device or the like, the input device 14 and the display 15 may not be provided.

(Example of operation of first identification unit)
Next, an operation example of the first identification unit 104 will be described with reference to FIGS. The 1st identification part 104 identifies a process area from a non-process area. When the first identification unit 104 identifies the machining section from the non-machining section, the feature extraction unit 102 extracts feature information from the detection information of the non-machining section.

  FIG. 4 is a diagram illustrating an example of detection information and a ladder signal of the processing machine 200 according to the embodiment. As shown in FIG. 4, the sensor data (detection information) includes a waveform portion indicating a non-processing section and a waveform portion indicating a processing section. The non-machining section is a section before and after the tool contacts the machining target. The machining section is a section in which the tool is in contact with the machining target and performing an actual machining process.

  On the other hand, the processing machine 200, for example, turns on the ladder signal at the start of the machining operation by the tool, feeds the tool to the machining target, and turns off the ladder signal when the actual machining processing is completed. In this case, the processing machine 200 does not turn on the ladder signal after the tool contacts the processing target. For this reason, the tool feed section in which the ladder signal is in the ON state includes a non-machining section in which the tool is not in contact with the machining target and a machining process in which the tool is in contact with the machining target. Section.

  Since the distance between the object to be machined and the tool is not known depending on the machining conditions and the like, it may be difficult to accurately determine the machining section based on such a ladder signal. However, if the detection information of the non-machining section is included in the abnormality determination of the processing machine 200, the accuracy of the determination is lowered. Therefore, it is necessary to accurately estimate the section where the processing is actually performed.

  For example, when the detection information is frequency data collected by a vibration sensor or a microphone device, the feature extraction unit 102 extracts energy, frequency spectrum, MFCC (mel frequency cepstrum coefficient), and the like as feature information. In the present embodiment, the extracted feature information is assumed to be a frequency spectrum.

  FIG. 5 is a diagram illustrating feature information extracted from detection information by the diagnostic apparatus 100 according to the embodiment as frequency components. The section shown in FIG. 5 includes one tool feed section. That is, the section of FIG. 5 includes one machining section. The feature extraction unit 102 extracts feature information from detection information of a section including one tool feed section among detection information received by the detection information reception unit 112 in order to estimate one machining section.

  Specifically, the feature extraction unit 102 extracts feature information, for example, by performing Fourier transform on the detection information for each frame. Here, the frame indicates the data amount of the detection information for a predetermined time (for example, 20 [ms], 40 [ms], etc.). For example, the frame is obtained by performing Fourier transform on the detection information. This corresponds to the data amount of the window length in the case of a frequency spectrum. The feature information shown in FIG. 5 is associated with the frame time of the corresponding detection information. The frequency spectrum of the processing section has a feature that is different from the frequency spectrum of the non-processing section. However, in FIG. 5, in order to schematically show the difference between the feature information of the non-machining section and the feature information of the machining section, the frequency component is not shown in the non-machining section. This does not indicate that there is no frequency component in the non-processed section.

  The selection unit 104a of the first identification unit 104 selects feature information of the non-machined section from the feature information extracted as described above in order to generate a model of the non-machined section. In addition, the selection unit 104a selects target feature information from the feature information extracted as described above in order to compare with the model of the non-processed section.

  FIG. 6 is a diagram illustrating an example in which the diagnosis apparatus 100 according to the embodiment selects feature information of a non-processing section and a processing section. In FIG. 6, feature information corresponding to the frame of the selected portion 401 is selected in the feature information of the non-processed section, and feature information corresponding to the frame of the selected portion 402 is selected in the feature information of the processed section.

  When selecting the feature information of the non-processed section, the selection unit 104a selects, for example, the feature information of the section where the ladder signal is in the OFF state, as in the selection portion 401. However, the selection of the feature information of the non-processed section by the selection unit 104a is not limited to this. For example, since there is a high possibility that the section immediately after the ladder signal is turned ON is also a non-processed section, the selection unit 104a may select the feature information of the section. Or the selection part 104a may select the information produced | generated by the machine learning with respect to the feature information contained in a non-processed area as the feature information of a non-processed area. A model is generated by the generation unit 103 from the feature information of the non-processed section selected in this way.

  On the other hand, when selecting the feature information of the machining section, the selection unit 104a may select the feature information of the latter half part of the tool feed section in which the ladder signal is in the ON state, as in the selection part 402, for example. This is because the first part of the tool feed section may include a non-machining section. The selection unit 104a sequentially selects feature information in units of frames in order to compare with a model generated from feature information of a non-machining section in the first half of the tool feed section. Note that when the selection unit 104a sequentially selects feature information in units of frames, it does not prevent each frame from overlapping.

  FIG. 7 is a diagram illustrating feature information selected from the non-processing section and the processing section by the diagnosis apparatus 100 according to the embodiment using a frequency spectrum. FIG. 7 shows a frequency spectrum as feature information corresponding to the frame of the selected portion 401 and a frequency spectrum as feature information corresponding to the frame of the selected portion 402. Since the frequency spectrum of the selected portion 402 is a frequency spectrum in a machining section where the tool is in contact with the machining target and performing machining, the frequency spectrum has a large amplitude in a specific frequency component. On the other hand, the frequency spectrum of the selection portion 401 is a frequency spectrum in a non-machining section where the tool is not in contact with the machining target, and thus has a small amplitude overall.

  The calculation unit 104b of the first identification unit 104 compares the model corresponding to the non-machining section with target feature information of a section that may be a processing section, and calculates the processing section likelihood of each target feature information. To do. As a method for comparing the model with the target feature information, machine learning such as a cross-correlation coefficient or GMM (Gaussian Mixture Model) is used in addition to a method of obtaining and comparing the Euclidean distance.

  FIG. 8 is a diagram illustrating an example in which the diagnostic device 100 according to the embodiment calculates the likelihood of the machining section from the Euclidean distance. In FIG. 8, the closer the machining section likelihood is to “1”, the higher the likelihood of the processing section. Since the processing section likelihood is calculated from the feature information extracted from the detection information including noise, as shown in FIG.

  The section determination unit 104c of the first identification unit 104 determines whether or not the processing section likelihood of each feature information calculated by the calculation unit 104b is larger than a predetermined threshold.

  FIG. 9 is a diagram for explaining an operation of estimating the machining section from the machining section likelihood calculated by the diagnosis apparatus 100 according to the embodiment. As illustrated in FIG. 9, the section determination unit 104c estimates an “estimated processing section” whose processing section likelihood is larger than a predetermined threshold as a processing section. Conversely, the section determination unit 104c estimates a section whose processing section likelihood is smaller than the threshold as a non-processing section. In order to estimate the machining section, the section determination unit 104c performs threshold determination on the likelihood of the machining section, but the present invention is not limited to this, and for example, using a statistical method, You may perform determination with respect to a section likeness.

  The first identification unit 104 estimates a section in which the section determination unit 104c determines that the processing section likelihood is larger than a predetermined threshold as the processing section.

(Example of operation of second identification unit)
Next, an operation example of the second identification unit 105 will be described with reference to FIGS. The second identifying unit 105 identifies the non-processed section and the processed section using the identification result obtained by the first identifying unit 104.

  The feature information collection unit 105d of the second identification unit 105 collects the feature information of the non-processing section and the processing section specified from the identification result by the first identification unit 104. When a certain period of time elapses after the processing machine 200 is started up or after tool replacement, a predetermined number of feature information shown in FIG. 7 is collected. The feature information collection unit 105d sets a flag when a sufficient number of data is generated to generate a model from each feature information. Using this flag as a trigger, the generation unit 103 generates a non-machined section model and a machined section model from a predetermined number of collected feature information.

  The selection unit 105a of the second identification unit 105 selects feature information extracted based on the detection information of the non-processed section. Further, the selection unit 105a selects feature information extracted based on detection information of a section that may be a processing section. That is, the selection unit 105a selects the feature information corresponding to the frame of the selection portion 401 and the feature information corresponding to the frame of the selection portion 402 in FIG.

  The calculation unit 105b of the second identification unit 105 compares the model of the non-processed section with the plurality of feature information selected by the selection unit 105a, and calculates the processing section likelihood of each feature information. In addition, the calculation unit 105b compares the model of the machining section with the feature information described above, and calculates the likelihood of the machining section of each feature information. That is, as shown in FIG. 8, each feature information is compared with both the non-machined section model and the machined section model, and the machining section likelihood is calculated.

  As shown in FIG. 9, the section determination unit 105c of the second identification unit 105 determines whether or not the processing section likelihood of each feature information calculated by the calculation unit 105b is larger than a predetermined threshold.

  The second identification unit 105 estimates the section determined by the section determination unit 105c that the likelihood of the processing section is larger than a predetermined threshold as the processing section.

(Example of abnormality judgment processing at the initial stage of machining)
Next, an example of abnormality determination processing of the processing machine 200 at the initial stage of processing will be described with reference to FIGS. 10 and 11. FIG. 10 is a flowchart illustrating an example of a procedure of abnormality determination processing at an initial stage of machining by the diagnostic device 100 according to the embodiment.

  Prior to the abnormality determination, first, a model when the processing machine 200 is normal is generated.

  As shown in FIG. 10, in step S <b> 110, the processing information acquisition unit 101 acquires context information from the processing machine 200 via the communication control unit 111. In step S120, the detection information receiving unit 112 acquires the detection information from the detection unit 225. The context information and detection information acquired at this time are information when the processing machine 200 is operating normally.

  In step S130, the feature extraction unit 102 extracts feature information of a section including a non-processed section and a processed section from a predetermined section based on a ladder signal.

  In step S140, the first identification unit 104 estimates a machining section.

  In step S150, the generation unit 103 generates a model from the estimated feature amount of the machining section.

  In step S <b> 160, the first identification unit 104 determines to switch to the second identification unit 105. Specifically, the first identification unit 104 determines whether or not a flag indicating that sufficient data has been collected by the feature information collection unit 105d of the second identification unit 105 is set. Since sufficient data has not been collected at this stage (No), the first identification unit 104 does not switch to the second identification unit 105 and returns to step S101.

  Next, the diagnostic apparatus 100 performs abnormality determination on the processing machine 200.

  In steps S110 and S120, context information and detection information are acquired. These context information and detection information are acquired when abnormality determination of the processing machine 200 becomes necessary.

  In step S130, the feature extraction unit 102 extracts feature information of a section including a non-processed section and a processed section.

  In step S140, the first identification unit 104 estimates a machining section.

  In step S <b> 150, the abnormality determination unit compares the previously generated normal machining section model with the feature information based on the currently acquired machining section detection information and performs abnormality determination on the processing machine 200. .

  In step S <b> 160, the first identification unit 104 determines to switch to the second identification unit 105. If the flag indicating that sufficient data has been collected by the feature information collection unit 105d of the second identification unit 105 is set (Yes), the first identification unit 104 switches to the second identification unit 105. Do.

  Thus, the abnormality determination process for the processing machine 200 in the initial stage of processing is completed. Thereafter, the diagnosis apparatus 100 identifies the non-processing section and the processing section by the function of the second identification unit 105, and performs abnormality determination.

  Next, details of step S140 will be described with reference to FIG.

  FIG. 11 is a flowchart illustrating an example of a processing section estimation process performed by the first identification unit 104 according to the embodiment. As shown in FIG. 11, in step S141, the selection unit 104a selects feature information of a non-processed section. In step S142, the generation unit 103 generates a model of the non-machined section from the feature information of the non-machined section.

  In step S143, the selection unit 104a sequentially selects target feature information from sections that may be processing sections. The calculation unit 104b compares the target feature information with the model of the non-machined section, and obtains an outlier. In step S144, the calculation unit 104b further calculates the likelihood of the processing section for each target feature information.

  In step S145, the section determination unit 104c estimates, as a processing section, a section determined that the processing section likelihood is greater than a predetermined threshold.

  Thus, the processing section estimation process by the first identification unit 104 ends.

(Example of abnormality determination processing after a predetermined period of time)
After a predetermined period has elapsed since the beginning of machining, the second identification unit 105 identifies the non-machining section and the machining section, and the diagnostic device 100 generates a model and determines abnormality.

  FIG. 12 is a flowchart illustrating an example of a procedure of abnormality determination processing after a predetermined period has elapsed by the diagnostic apparatus 100 according to the embodiment. In step S200, models of the non-machining section and the machining section are generated. Such a model is generated from a predetermined number of feature information collected by the feature information collection unit 105d.

  In steps S210 and S220, the context information and detection information of the processing machine 200 that requires abnormality determination are acquired.

  In step S230, the feature extraction unit 102 extracts feature information of a section including a non-processed section and a processed section.

  In step S240, the second identification unit 105 estimates a machining section.

  In step S250, the abnormality determination unit compares the model of the machining section generated in step S200 with the feature information based on the detection information acquired in step S220, and determines abnormality of the processing machine 200.

  Thus, the abnormality determination process for the processing machine 200 after the predetermined period has elapsed.

  Next, details of step S240 will be described with reference to FIG.

  FIG. 13 is a flowchart illustrating an example of a processing section estimation process performed by the second identification unit 105 according to the embodiment. As shown in FIG. 13, in step S421, the calculation unit 105b compares the model of the non-processed section generated in step S200 of FIG. 12 with each target feature amount. Further, the model of the processing section in step S200 is compared with each target feature amount, and an approximate value for the model is obtained for each feature information. In step S242, the calculation unit 105b further calculates the likelihood of the processing section for each target feature information.

  In step S243, the section determination unit 105c estimates, as a processing section, a section that has been determined that the processing section likelihood is greater than a predetermined threshold.

  Thus, the processing section estimation processing by the second identification unit 105 ends.

  For example, in the information acquisition device of Patent Document 1, time-series data of machine information such as spindle load and the event data of the device such as a program name, a tool number, and an override value by an operator operation are output in association with each other. Thereby, an abnormality in the machining operation is detected. However, it is difficult to accurately grasp the actual machining section by separating the non-machining section and the machining section from the event data of the apparatus.

  In the diagnosis apparatus 100 of the embodiment, even in the initial stage of machining, a model of a non-machined section is generated from a small amount of feature information, and the machined section is identified from the non-machined section. Thereby, predetermined accuracy is obtained in the section estimation.

  Moreover, in the diagnostic apparatus 100 of the embodiment, after a predetermined period has elapsed from the initial stage of machining, models are generated for both the non-machined section and the machined section based on the abundant feature information obtained so far, Used for identification. This makes it possible to grasp the actual machining section with higher accuracy.

  As described above, in the diagnosis apparatus 100 according to the embodiment, since appropriate algorithms are used at the initial stage of machining and after a predetermined period, a more efficient machining section can be estimated.

  Moreover, in the processing machine 200, as shown in FIG. 14, the process using the same tool may be performed continuously several times. In FIG. 14, for example, steps 1-1 and 1-2 are the same, steps 1-4 to 1-6 are the same, and steps 1-7 to 1-9 are the same. Such a process is processing based on the same context information. Even in such a case, the first identification unit 104 and the second identification unit 105 of the embodiment can identify the non-processing section and the processing section and accurately estimate the processing section.

(Modification)
In the above-described embodiment, when generating the model of the non-machining section in order to distinguish the non-machining section and the machining section by the first identification unit 104, the feature information of the actual non-machining section is used. It is not limited to this. Before starting the processing, the feature information may be acquired by operating the processing machine 200 in advance in a state where there is no workpiece. Using this as feature information in the non-machined section, a model of the non-machined section can be generated.

  In the above-described embodiment, the detection information is, for example, vibration data or acoustic data, but other data such as a motor current value, load, torque, and the like can be used as detection information.

  In the above-described embodiment, the apparatus to be diagnosed is the processing machine 200, for example, but a machine such as an assembly machine, a measuring machine, an inspection machine, or a washing machine may be the target apparatus.

  Note that the program executed by the above-described embodiment and the diagnostic system of each modification may be configured to be provided by being incorporated in advance in a ROM or the like.

  In addition, the program executed in the above-described embodiment and the diagnostic system of each modified example is a file in an installable format or an executable format, and is a CD-ROM (Compact Disc Read Only Memory), a flexible disk (FD), a CD. -R (Compact Disk-Recordable), DVD (Digital Versatile Disk), and the like may be recorded on a computer-readable recording medium and provided as a computer program product.

  Further, the program executed by the diagnostic system of the above-described embodiment and each modification may be stored on a computer connected to a network such as the Internet and provided by being downloaded via the network. . Moreover, you may comprise so that the program run with the diagnostic system of the above-mentioned embodiment and each modification may be provided or distributed via networks, such as the internet.

  In addition, the program executed in the above-described embodiment and the diagnostic system of each modification has a module configuration including the above-described functional units, and as actual hardware, a CPU (processor) reads the program from the ROM. By executing the above, each function unit described above is loaded onto the main storage device, and each function unit is generated on the main storage device.

1 diagnostic system 1B bus 2B bus 10 CPU
10a ROM
10b RAM
11 Communication I / F
12 Sensor I / F
13 Auxiliary storage device 14 Input device 15 Display 20 CPU
20a ROM
20b RAM
21 Communication I / F
DESCRIPTION OF SYMBOLS 23 Drive control circuit 24 Motor 24a Tool 25 Sensor 25a Sensor amplifier 100 Diagnosis apparatus 101 Processing information acquisition part 102 Feature extraction part 103 Generation part 104 1st identification part 105 2nd identification part 106 Abnormality determination part 107 Display control part 111 Communication Control unit 112 Detection information receiving unit 113 Storage unit 114 Input unit 115 Display unit 200 Processing machine 201 Numerical control unit 221 Communication control unit 223 Drive control unit 224 Drive unit 225 Detection unit

JP 2017-033346 A

Claims (7)

A first acquisition unit that acquires detection information output from a detection unit that detects a physical quantity that changes according to the operation of the target device;
A second acquisition unit that acquires a signal indicating a section in which the target device is operated from the target device;
A first identification unit for identifying a non-processing section and a processing section in the target device based on the detection information and the signal;
A second identification unit that identifies a non-processing section and a processing section in the target device based on the detection information, the signal, and the identification result by the first identification unit;
Processing section determination device. An extraction unit that extracts feature information from detection information of the non-processing section identified by the first identification unit;
Generating a first model using the feature information of the non-processing section for learning, and using the first model for identifying the processing section from the non-processing section,
The processing section determination apparatus according to claim 1. The first acquisition unit includes:
Obtaining detection information corresponding to the detection information of the non-processing section, obtained by operating the target device in advance in the absence of an object to be processed,
An extraction unit for extracting feature information from the detection information in the preliminary operation;
Using the feature information for learning, generating a first model in advance, and using the first model for identifying the processing section from the non-processing section,
The processing section determination apparatus according to claim 1. An extraction unit that extracts feature information from detection information of the non-processing section and the processing section identified by the second identification unit;
A second model is generated using the feature information of the non-processing section for learning, a third model is generated using the feature information of the processing section for learning, and the second and third models are used as the non-processing section. Used to identify the processing section and the processing section,
The processing section determination device according to any one of claims 1 to 3. A processing section determination method executed by a processing section determination apparatus that determines a processing section of a target device,
A first acquisition unit in which a first acquisition unit acquires detection information output from a detection unit that detects a physical quantity that changes according to the operation of the target device;
A second acquisition unit in which a second acquisition unit acquires a signal indicating a section in which the target device is operated from the target device;
A first identification unit for identifying a non-processing section and a processing section in the target device based on the detection information and the signal;
A second identification unit including a second identification step for identifying a non-processing section and a processing section in the target device based on the detection information, the signal, and the identification result of the first identification step; ,
Processing section judgment method. On the computer,
A first acquisition step of acquiring detection information output from a detection unit that detects a physical quantity that changes according to the operation of the target device;
A second acquisition step of acquiring a signal indicating a section in which the target device is operated from the target device;
A first identification step for identifying a non-processing section and a processing section in the target device based on the detection information and the signal;
Based on the detection information, the signal, and the identification result of the first identification step, a second identification step for identifying a non-processing section and a processing section in the target device is performed.
program. A first acquisition unit that acquires detection information output from a detection unit that detects a physical quantity that changes according to the operation of the target device;
A second acquisition unit that acquires a signal indicating a section in which the target device is operated from the target device;
A first identification unit for identifying a non-processing section and a processing section in the target device based on the detection information and the signal;
A second identification unit that identifies a non-processing section and a processing section in the target device based on the detection information, the signal, and the identification result by the first identification unit;
An abnormality determination unit that determines the presence or absence of an abnormality of the target device using detection information of the identified processing section,
Diagnostic device.

Images