Diagnostic device, diagnostic method, program and diagnostic system

Abstract

To provide a diagnostic device, diagnostic method, program and diagnostic system that can improve the accuracy of diagnosing an operation state.SOLUTION: A diagnostic device comprises: a reception unit that receives, from an object device, context information corresponding to a prescribed operation of multiple pieces of context information defining operation of operation objects forming the object device determined with respect to each type of operation performed by the object device, and detection information on a physical quantity changing in response to operation of the operation objects; and a determination unit that determines the state of an operation object by using the detection information when the operation object is during the prescribed operation and a model corresponding to the received context information of one or more models determined for each of the one or more pieces of context information.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present invention relates to a diagnostic device, a diagnostic method, a program, and a diagnostic system. [Background technology]

[0002] There is known a technique for detecting an abnormality in an image forming apparatus by comparing operation sound data collected by a sound collecting unit with operation sound data prepared in advance for each operation state based on context information.

[0003] For example, Patent Document 1 describes an image forming device that compares operating sound data of each unit (drum motor, paper feed motor, fixing motor, developing clutch, etc.) that has been collected and stored in advance with operating sound data collected by operating the image forming device, and when the difference is equal to or greater than a predetermined level, detects the sound as abnormal, and uses an operation sequence table for each unit to identify the unit that is generating the abnormal sound. Summary of the Invention [Problem to be solved by the invention]

[0004] However, there are cases where the method of Patent Document 1 cannot correctly diagnose an abnormality. For example, in Patent Document 1, the source of an abnormal sound is inferred using an operation timing chart (e.g., FIG. 20 of Patent Document 1), but since this is an operation timing chart for each unit, the operation information for inferring the source is rough, and there are cases where the source cannot be correctly inferred.

[0005] The present invention has been made in consideration of the above, and aims to provide a diagnostic device, a diagnostic method, a program, and a diagnostic system that can improve the accuracy of diagnosing operational conditions such as abnormalities. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems and achieve the object, the present invention comprises a receiving unit that receives from the target device context information corresponding to a predetermined operation among a plurality of context information that specifies the operation of an operating object that constitutes the target device, the context information being defined for each type of operation performed by the target device, and detection information of a physical quantity that changes according to the operation of the operating object, and a determining unit that determines a state of the operating object using the detection information when the operating object is performing the predetermined operation and a model corresponding to the received context information among one or more models defined for each of the one or more context information. Effect of the Invention

[0007] The present invention provides an advantage in that it is possible to improve the accuracy of diagnosing an abnormality. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a block diagram illustrating an example of the configuration of a diagnostic system according to the first embodiment. [Diagram 2] FIG. 2 is a block diagram illustrating an example of a hardware configuration of the processing machine according to the first embodiment. [Diagram 3] FIG. 3 is a block diagram illustrating an example of a hardware configuration of the diagnostic device according to the first embodiment. [Figure 4] FIG. 4 is a block diagram illustrating an example of a functional configuration of the diagnostic device according to the first embodiment. [Diagram 5] FIG. 5 is a diagram illustrating an example of the correspondence between context information and a model. [Figure 6] FIG. 6 is a flowchart showing an example of the diagnosis process in the first embodiment. [Figure 7] FIG. 7 is a flowchart illustrating an example of a model generation process in the first embodiment. [Figure 8] FIG. 8 is a diagram for explaining a specific example of the process according to the first embodiment. [Figure 9] FIG. 9 is a diagram illustrating an example in which a determination is made on a portion of context information. [Figure 10] FIG. 10 is a diagram showing an example in which a common model is used in another machining process. [Figure 11] FIG. 11 is a diagram showing an example in which the accumulated usage time is used as context information. [Figure 12] FIG. 12 is a diagram illustrating an example of generating a model. [Figure 13] FIG. 13 is a block diagram illustrating an example of a functional configuration of a diagnostic device according to the second embodiment. As shown in FIG. [Figure 14] FIG. 14 is a diagram for explaining a specific example of the process according to the second embodiment. [Figure 15] FIG. 15 is a diagram illustrating an example of a data structure of correspondence information used to determine detection information. [Figure 16] FIG. 16 is a diagram illustrating an example of the relationship between the context information and the processing section. [Figure 17] FIG. 17 is a diagram showing an example of a method for specifying a processing section. [Figure 18] FIG. 18 is a diagram showing another example of a method for specifying a processing section. [Figure 19] FIG. 19 is a graph showing an example of the relationship between the likelihood and the decision value r(k). [Figure 20] FIG. 20 is a diagram showing an example of a change in likelihood over time. [Figure 21] FIG. 21 is a diagram for explaining an example in which a plurality of threshold values ​​are used. [Figure 22] FIG. 22 is a diagram for explaining an example in which a plurality of threshold values ​​are used. [Figure 23-1] FIG. 23A is a diagram for explaining an example of the determination method according to the sixth modification. [Figure 23-2] FIG. 23-2 is a diagram for explaining an example of the determination method according to the sixth modification. [Figure 23-3] FIG. 23-3 is a diagram for explaining an example of the determination method according to the sixth modification. [Figure 24-1] FIG. 24A is a diagram for explaining an example of the determination method according to the seventh modification. [Figure 24-2]FIG. 24-2 is a diagram for explaining an example of the determination method according to the seventh modification. [Figure 24-3] FIG. 24-3 is a diagram for explaining an example of the determination method according to the seventh modification. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a diagnostic device, a diagnostic method, a program, and a diagnostic system according to the present invention will be described in detail below with reference to the accompanying drawings.

[0010] (First embodiment) Fig. 1 is a block diagram showing an example of the configuration of a diagnostic system according to the first embodiment. As shown in Fig. 1, the diagnostic system includes a processing machine 200 and a diagnostic device 100. The processing machine 200 is an example of a target device to be diagnosed by the diagnostic device 100.

[0011] The processing machine 200 and the diagnostic device 100 may be connected in any form. For example, the processing machine 200 and the diagnostic device 100 are connected by a dedicated connection line, a wired network such as a wired LAN (local area network), a wireless network, or the like.

[0012] The processing machine 200 includes a numerical control unit 201, a communication control unit 202, and a machine tool 203. The machine tool 203 includes a sensor 211, a drive unit 212, and a tool 213.

[0013] Machine tool 203 is a machine that machines an object to be machined under the control of numerical control unit 201. Machine tool 203 includes a drive unit 212 that operates under the control of numerical control unit 201. Drive unit 212 is, for example, a motor. Tool 213 is an object that is actually driven by drive unit 212, and may be any tool that is used for machining and is subject to numerical control, such as a drill or end mill that performs machining on an object. Two or more drive units 212 may be provided.

[0014] The numerical control unit 201 executes machining by the machine tool 203 through numerical control. For example, the numerical control unit 201 generates and outputs numerical control data for controlling the operation of the drive unit 212. The numerical control unit 201 also outputs context information to the communication control unit 202. The context information is information that is determined for each type of operation of the processing machine 200. The context information includes, for example, information for identifying the tool 213 driven by the drive unit 212, the number of rotations of the drive unit 212, the rotation speed of the drive unit 212, movement information of the drive unit 212 and the tool 213, etc.

[0015] The numerical control unit 201 transmits, for example, context information indicating a current operation to the diagnosis device 100 via the communication control unit 202. When machining an object to be machined, the numerical control unit 201 changes the type of tool 213 driven by the driving unit 212 and the driving state (number of rotations, rotation speed, etc.) of the driving unit 212 according to the machining process. Every time the numerical control unit 201 changes the type of operation, it sequentially transmits context information corresponding to the changed type of operation to the diagnosis device 100 via the communication control unit 202.

[0016] The communication control unit 202 controls communication with an external device such as the diagnostic device 100. For example, the communication control unit 202 transmits to the diagnostic device 100 context information corresponding to a current operation.

[0017] The sensor 211 is a detection unit that detects a physical quantity that changes in response to the operation of the processing machine 200 and outputs detection information (sensor data). The type of the sensor 211 and the physical quantity that it detects may be any. For example, the sensor 211 may be a microphone, an acceleration sensor, or an AE (acoustic emission) sensor, and the detection information may be acoustic data, acceleration data, or data indicating an AE wave. The number of sensors 211 is arbitrary. A plurality of sensors 211 that detect the same physical quantity may be provided, or a plurality of sensors 211 that detect different physical quantities may be provided.

[0018] For example, when the blade of the tool 213 used in processing breaks or chips, the sound during processing changes. For this reason, by detecting acoustic data with the sensor 211 (microphone) and comparing it with a model indicating normal sounds, it becomes possible to detect abnormalities in the operation of the processing machine 200.

[0019] The diagnostic device 100 includes a communication control unit 101 and a determination unit 102. The communication control unit 101 controls communication with an external device such as a processing machine 200. For example, the communication control unit 101 receives context information and detection information from the processing machine 200. The determination unit 102 refers to the context information and the detection information to determine whether or not the operation of the processing machine 200 is normal. The functions of each unit will be described in detail later.

[0020] Fig. 2 is a block diagram showing an example of a hardware configuration of the processing machine 200. As shown in Fig. 2, the processing machine 200 includes a CPU (Central Processing Unit) 51, a ROM (Read Only Memory) 52, a RAM (Random Access Memory) 53, a communication I / F (Interface) 54, a drive control circuit 55, and a motor 56, which are connected via a bus 58.

[0021] The CPU 51 controls the entire processing machine 200. The CPU 51 executes a program stored in the ROM 52 or the like using, for example, the RAM 53 as a work area, thereby controlling the operation of the entire processing machine 200 and realizing a processing function.

[0022] The communication I / F 54 is an interface for communicating with an external device such as the diagnostic device 100. The drive control circuit 55 is a circuit for controlling the drive of a motor 56. The motor 56 drives a tool 213 used for machining, such as a drill, a cutter, and a table. The motor 56 corresponds to, for example, the drive unit 212 in FIG. 1. The sensor 57 is attached to the processing machine 200, detects a physical quantity that changes in response to the operation of the processing machine 200, and outputs detection information to the diagnostic device 100. The sensor 57 corresponds to, for example, the sensor 211 in FIG. 1.

[0023] The numerical control unit 201 and communication control unit 202 in FIG. 1 may be realized by having the CPU 51 in FIG. 2 execute a program, i.e., by software, or may be realized by hardware such as an IC (Integrated Circuit), or may be realized by a combination of software and hardware.

[0024] Fig. 3 is a block diagram showing an example of a hardware configuration of diagnostic device 100. As shown in Fig. 3, diagnostic device 100 has a configuration in which a CPU 61, a ROM 62, a RAM 63, a communication I / F 64, and an HDD (Hard Disk Drive) 65 are connected via a bus 66.

[0025] The CPU 61 controls the entire diagnostic device 100. The CPU 61 executes a program stored in the ROM 62 or the like using the RAM 63 as a work area, for example, to control the operation of the entire diagnostic device 100 and realize a diagnostic function. The communication I / F 64 is an interface for communicating with an external device such as the processing machine 200. The HDD 65 stores information such as setting information of the diagnostic device 100, context information transmitted by the numerical control unit 201 of the processing machine 200 and received from the communication control unit 202, and detection information transmitted by the sensor 211. Instead of the HDD 65, or together with the HDD 65, a non-volatile storage means such as an EEPROM (Electrically Erasable Programmable Read-Only Memory) or an SSD (Solid State Drive) may be provided.

[0026] Fig. 4 is a block diagram showing an example of a functional configuration of the diagnostic device 100. As shown in Fig. 4, the diagnostic device 100 includes a reception unit 103, a feature extraction unit 104, a generation unit 105, and a storage unit 111 in addition to the above-mentioned communication control unit 101 and determination unit 102.

[0027] Storage unit 111 stores various information necessary for the diagnostic function of diagnostic device 100. Storage unit 111 can be realized by, for example, RAM 63 and HDD 65 in Fig. 3. For example, storage unit 111 stores one or more models used to determine an abnormality.

[0028] The model is generated by learning, for example, using detection information detected when the processing machine 200 is operating normally. Alternatively, depending on the purpose of detection, the model is generated by attaching a new tool 213, performing a predetermined processing, and acquiring data until the tool 213 becomes abnormal (such as a broken blade or chipping). The learning method and the form of the model to be learned may be any method. For example, models such as GMM (Gaussian mixture model) and HMM (Hidden Markov model) and corresponding model learning methods can be applied.

[0029] Also, for example, a predetermined machining period after a new tool 213 is attached may be set as a learning period, and normal and abnormal states may be made into rules and stored as models in the storage unit 111. For example, the first 10 machining operations after a new tool 213 is attached and machining is started may be set as a learning period for determining diagnostic rules. The diagnostic rules may be determined in advance separately from the actual machining, and the determined rules may be stored in the storage unit 111 as models.

[0030] In this embodiment, a model is generated for each piece of context information. The storage unit 111 stores, for example, context information and a model corresponding to the context information in association with each other. FIG. 5 is a diagram showing an example of the association between context information and a model. As shown in FIG. 5, a model is generated for each tool 213 attached to the processing machine 200, for example. If the tool 213 is a drill, a model is generated for each diameter of the drill. If there are three diameters of φ1 mm, φ3 mm, and φ5 mm as the diameter of the drill, a model is generated for each diameter. Then, when the determination unit 102 recognizes the drill diameter used for processing based on the context data received from the processing machine 200, for example, the determination unit 102 selects a learning model that matches the drill diameter and uses it for determination.

[0031] The communication control unit 101 includes a receiving unit 101a and a transmitting unit 101b. The receiving unit 101a receives various information transmitted from an external device such as the processing machine 200. For example, the receiving unit 101a receives context information corresponding to the current operation of the processing machine 200 and detection information transmitted by the sensor 211. The transmitting unit 101b transmits various information to the external device.

[0032] The feature extraction unit 104 extracts feature information (feature amount) from the detection information to be used for generating a model and for judgment by the judgment unit 102. The feature information may be any information that indicates the feature of the detection information. For example, when the detection information is acoustic data collected by a microphone, the feature extraction unit 104 may extract feature amounts such as energy, frequency spectrum, and MFCC (Mel-Frequency Cepstrum Coefficients).

[0033] The generation unit 105 creates a model for determining normal operation by learning using feature information extracted from detection information during normal operation. When the model is generated by an external device, the generation unit 105 may not be provided. When context information for which a model is not defined and detection information corresponding to the context information are input, the generation unit 105 may generate a model corresponding to the context information by using feature information extracted from the detection information.

[0034] Although the model is generated by the generation unit 105 in the above embodiment, the model may be generated externally, and the externally generated model may be received by the communication I / F 64 and stored in the HDD 65 for use.

[0035] The determination unit 102 determines the operation state of the processing machine 200 by using the detection information received by the receiving unit 101a and a model corresponding to the received context information. The operation state of the processing machine 200 indicates, for example, whether the operation of the processing machine 200 is normal or not. For example, the determination unit 102 requests the feature extraction unit 104 to extract feature information from the detection information. The determination unit 102 calculates a likelihood indicating the likelihood that the feature information extracted from the detection information is normal by using the corresponding model. The determination unit 102 compares the likelihood with a predetermined threshold, and determines that the operation of the processing machine 200 is normal when the likelihood is equal to or greater than the threshold. Also, the determination unit 102 determines that the operation of the processing machine 200 is abnormal when the likelihood is less than the threshold. Since the feature amount changes depending on the state of the tool 213, the determination of the operation state of the processing machine 200 also includes determining an index such as the wear level of the tool 213.

[0036] The method of determining whether or not the detection information is normal is not limited to this, and any method that can determine whether or not the detection information is normal using the detection information and the model may be used. For example, instead of directly comparing the likelihood value with a threshold value, a value indicating the fluctuation of the likelihood may be compared with a threshold value.

[0037] The accepting unit 103 accepts input of context information different from the context information received by the receiving unit 101a from the processing machine 200. For example, the accumulated usage time can be configured to be acquired from the processing machine 200. In this case, the processing machine 200 may have a function of resetting (initializing) the accumulated usage time when, for example, the tool 213 is replaced.

[0038] The receiving unit 103 may be configured to receive the accumulated usage time instead of acquiring it from the processing machine 200. The receiving unit 103 receives context information input from an operation unit such as a keyboard or a touch panel. The context information received by the receiving unit 103 is not limited to the accumulated usage time, and may be, for example, information on the specifications of the tool 213 to be used (such as the diameter of the blade, the number of blades, the material, whether the tool 213 is coated, etc.) or information on the material to be processed (such as the material). The receiving unit 103 may be configured to receive the context information from an external device such as a server device or a personal computer. If there is no need to receive context information from sources other than the processing machine 200, the receiving unit 103 may not be provided.

[0039] Each unit in FIG. 4 (communication control unit 101, judgment unit 102, reception unit 103, feature extraction unit 104, and generation unit 105) may be realized by causing CPU 61 in FIG. 3 to execute a program, that is, by software, or by hardware such as an IC (Integrated Circuit), or by a combination of software and hardware.

[0040] Next, the diagnostic process performed by the diagnostic device 100 according to the first embodiment configured as described above will be described with reference to Fig. 6. Fig. 6 is a flowchart showing an example of the diagnostic process in the first embodiment.

[0041] As described above, the numerical control unit 201 of the processing machine 200 sequentially transmits context information indicating the current operation to the diagnosis device 100. The receiving unit 101a receives the context information thus transmitted from the processing machine 200 (step S101). The sensor 211 of the processing machine 200 sequentially outputs detection information during processing. The receiving unit 101a receives the detection information (sensor data) thus transmitted from the processing machine 200 (step S102).

[0042] The feature extraction unit 104 extracts feature information from the received detection information (step S103). The judgment unit 102 judges whether the processing machine 200 is operating normally or not by using the extracted feature information and a model corresponding to the received context information (step S104). The judgment unit 102 outputs the judgment result (step S105). Any method may be used to output the judgment result. For example, if the diagnostic device 100 is equipped with a display device such as a display, the judgment unit 102 may display the judgment result on the display device. The judgment unit 102 may output the judgment result to an external device such as a server device or a personal computer.

[0043] Next, the model generation process by the diagnosis device 100 according to the first embodiment configured as described above will be described with reference to FIG. 7. FIG. 7 is a flowchart showing an example of the model generation process in the first embodiment. Note that the model generation process is executed in advance, for example, before the diagnosis process. As described above, the model generation process may be configured to be executed when context information for which a model is not defined is input. Also, when the model is generated externally as described above, the model generation process may not be executed.

[0044] The receiving unit 101a receives context information transmitted from the processing machine 200 (step S201). The receiving unit 101a receives detection information (sensor data) transmitted from the processing machine 200 (step S202).

[0045] The context information and detection information received in this manner are used to generate a model. Since a model is generated for each piece of context information, the detection information needs to be associated with the corresponding context information. For this reason, for example, the receiving unit 101a stores the received detection information in the storage unit 111 or the like in association with the context information received at the same timing. Each piece of information may be temporarily stored in the storage unit 111 or the like, and it may be confirmed that the information is normal, and a model may be generated using only the information in the normal state. In other words, a model may be generated using detection information labeled as normal.

[0046] The confirmation of whether the information is normal (labeling) may be performed at any timing after the information is stored in the storage unit 111, etc., or may be performed in real time while the processing machine 200 is operating. A model may be generated assuming that the information is normal without performing labeling. If the information assumed to be normal is actually abnormal, the generated model will not correctly execute the judgment process. Whether such a situation exists can be determined based on the frequency of abnormal judgments, and measures can be taken such as deleting the erroneously generated model. A model generated from abnormal information may also be used as a model for judging abnormality.

[0047] The feature extraction unit 104 extracts feature information from the collected detection information (step S203). The generation unit 105 generates a model for the context information by using the feature information extracted from the detection information associated with the same context information (step S204). The generation unit 105 stores the generated model in, for example, the storage unit 111 (step S205).

[0048] Next, a specific example of the model generation process and the diagnosis process according to this embodiment will be described. Fig. 8 is a diagram for explaining a specific example of the process according to this embodiment.

[0049] 8 shows, for example, a model generation process and a diagnosis process for a part of a process for machining a certain part. In the model generation process, a plurality of pieces of detection information (detection information 711a to 711c in FIG. 8) received together with the context information 701 are used. Note that the number of pieces of detection information is not limited to three, and can be any number.

[0050] The context information 701 indicates that the machining process includes an operation of driving four motors (motor A, motor B, motor C, and motor D). The feature extraction unit 104 extracts feature information from the received detection information. The generation unit 105 generates a model for each piece of context information corresponding to each motor using the feature information extracted from the corresponding detection information. The generated model is stored in the memory unit 111 or the like. FIG. 8 shows an example in which a model ("motor B") generated for context information when motor B is driven is stored in the memory unit 111. The stored model is referred to in the subsequent diagnosis process.

[0051] In the diagnosis process, similarly to the model generation process, detection information 721 is received together with context information 701. When the context information indicates that "motor B is being driven", the determination unit 102 determines whether or not the operation of the processing machine 200 is normal, for example, using detection information received during the period in which this context information was received and the model "motor B" stored in the storage unit 111.

[0052] Similarly, when other context information is received, the determination unit 102 performs a determination using the corresponding detection information and the corresponding model. It is not necessary to perform a determination for all the context information. Fig. 9 is a diagram showing an example in which a determination is performed for a part of the context information.

[0053] In the example of FIG. 9, a model is generated only when the context information indicates that "motor B is being driven." Furthermore, the diagnostic process is executed when context information 701-2 indicating that "motor B is being driven" is received. This allows the diagnostic process to be executed using only detection information that is effective for determining an abnormality. For example, when using acoustic data as detection information, there are cases where a processing step includes a section that does not need to be determined, such as a silent section. By excluding such unnecessary sections from the determination target, it is possible to reduce erroneous determinations and reduce calculation costs. In other words, it is possible to achieve high accuracy and efficiency in the diagnostic process.

[0054] Even if the machining processes are different from each other, for example, when the same motor is used, the corresponding models may be commonly used to execute the diagnosis process. Fig. 10 is a diagram showing an example in which a common model is used in another machining process.

[0055] Context information 901 in Fig. 10 indicates that the machining process includes an operation of driving four motors (motor X, motor Y, motor Z, and motor B). The use of motor B is common to the machining process shown in Fig. 8, for example. Therefore, in the machining process in Fig. 10, the determination unit 102 can execute a diagnosis process using the same model "motor B" and detection information 921.

[0056] FIG. 11 is a diagram showing an example in which the accumulated usage time is further used as context information. The context information 1001 indicates the accumulated usage time of each motor. In the example of FIG. 11, a model is generated for each range of the accumulated usage time (such as 0 to 6 months, 6 to 12 months, etc.) and for each type of motor. For example, the storage unit 111a stores a model of each motor whose accumulated usage time is 0 to 6 months. The storage unit 111b stores a model of each motor whose accumulated usage time is 6 to 12 months. The storage unit 111c stores a model of each motor whose accumulated usage time is, for example, 12 months or more. The storage units 111a to 111c may be realized by physically different storage media or may be realized by the same storage medium. The determination unit 102 identifies a corresponding model using the context information 1001 together with the context information 701, and executes a diagnosis process using the identified model.

[0057] Fig. 12 is a diagram showing an example of generating a model when context information for which a model has not been defined is input. Context information 1101 in Fig. 12 indicates that a machining process includes an operation for driving four motors (motor X, motor C, motor D, and motor B). It is assumed that a model has not been generated for the context information indicating the driving of motor X.

[0058] In this case, the feature extraction unit 104 extracts feature information from each of the multiple pieces of detection information (detection information 1111a, 1111b, and 1111c in FIG. 12) during a period corresponding to context information indicating that "motor X is being driven". The generation unit 105 uses the extracted feature information to generate a model "motor X" corresponding to the context information indicating that "motor X is being driven", and stores the model in the storage unit 111. This makes it possible to determine whether there is an abnormality during the period in which motor X will be driven in the future.

[0059] In this way, in the first embodiment, context information indicating the current operation is received from the processing machine 200, and an abnormality can be determined using a model corresponding to the received context information. Therefore, it is possible to identify the operating drive unit with high accuracy and diagnose the abnormality with high accuracy.

[0060] Second Embodiment In the first embodiment, one type of detection information is used to determine whether or not the processing machine 200 is normal. The number of pieces of detection information used for the determination is not limited to one, and may be two or more. The diagnosis system of the second embodiment determines whether or not the processing machine 200 is abnormal by using a plurality of pieces of detection information.

[0061] The configuration of the diagnostic system of the second embodiment is the same as that of the first embodiment shown in FIG. 1, and therefore the description will be omitted. FIG. 13 is a block diagram showing an example of the functional configuration of a diagnostic device 100-2 of the second embodiment. As shown in FIG. 13, the diagnostic device 100-2 includes a communication control unit 101, a determination unit 102-2, a reception unit 103, a feature extraction unit 104, a generation unit 105, and a storage unit 111.

[0062] In the second embodiment, the function of the judgment unit 102-2 is different from that in the first embodiment. Other configurations and functions are the same as those in the block diagram of the diagnostic device 100 according to the first embodiment in FIG. 4, so the same reference numerals are used and the description thereof is omitted here.

[0063] The determination unit 102-2 uses a plurality of pieces of detection information to determine whether or not the operation of the processing machine 200 is normal. For example, the determination unit 102 switches the detection information used for the determination in accordance with the context information.

[0064] Fig. 14 is a diagram for explaining a specific example of the process according to this embodiment. In the example of Fig. 14, context information 1201 and multiple types of detection information 1221a and 1221b are received. The detection information 1221a is, for example, acoustic data. The detection information 1221b is, for example, acceleration data.

[0065] When the context information indicates that "motor A is being driven", the determination unit 102-2 performs a determination using feature information extracted from the detection information 1222b of the period corresponding to the context information, among the detection information 1221b. When the context information indicates that "motor B is being driven", the determination unit 102-2 performs a determination using feature information extracted from the detection information 1222a of the period corresponding to the context information, among the detection information 1221a. The determination unit 102-2 also reads out a model corresponding to the context information from the storage unit 111a-2 or the storage unit 111b-2 and uses it.

[0066] The detection information corresponding to the context information may be determined, for example, by using the correspondence information stored in the storage unit 111. Fig. 15 is a diagram showing an example of a data structure of the correspondence information used to determine the detection information. The correspondence information includes, for example, sensor data and context information. The determination unit 102-2 can determine the detection information corresponding to the context information by referring to such correspondence information.

[0067] Next, modified examples that can be applied to each of the above embodiments will be described.

[0068] (Variation 1) The context information only indicates the section in which a certain driving unit 212 is driven, and there are cases where it is not possible to precisely extract the actual machining section in which, for example, the tool 213 is rotated by this driving unit 212 and is contacting and machining the material. In other words, there are cases where the accuracy of the abnormality determination is deteriorated.

[0069] 16 is a diagram showing an example of the relationship between context information and a machining section. Context information 1501 indicates that motor B is being driven. Based only on context information 1501, for example, detection information corresponding to waveform section 1512 among the detection information is identified. However, the section in which the material is actually being machined is waveform section 1511. For example, when sound data is used as the detection information, waveform section 1511 corresponds to a section of sound data detected when tool 213 comes into contact with the material and generates a sound.

[0070] Therefore, in the first modification, the context information and the detection information are combined to identify the actual processing section. That is, the determination unit 102 determines a period to be used for determination among the periods identified by the received context information based on the received detection information, and performs determination using the detection information and the model for the determined period.

[0071] For example, the determination unit 102 determines the processing section by identifying the time when the feature of the detection information is switched. FIG. 17 is a diagram showing an example of a method for identifying the processing section. In the example of FIG. 17, the determination unit 102 identifies a time 1601 when the amplitude of the detection information is switched to exceed a predetermined threshold (for example, "a" and "-a"), and a time 1602 when the amplitude is subsequently switched to fall below the threshold. The determination unit 102 extracts a section 1611 between the time 1601 and the time 1602. The determination unit 102 then executes the determination using the feature information extracted from the detection information of the section 1611. This makes it possible to execute the determination with higher accuracy using the detection information of the section corresponding to the actual processing section.

[0072] (Variation 2) In the second modification, a section identification signal is used to extract feature information (identify a processing section). The section identification signal is generated by the diagnostic device 100 based on the context information transmitted from the processing machine 200, but it may also be generated by the processing machine 200 and transmitted to the diagnostic device 100.

[0073] Fig. 18 is a diagram showing an example of a method for identifying a processing section in this modified example. When a section identification signal 1801 indicates that the section is a processing section (e.g., high level), the determination unit 102 uses feature information extracted from detection information of the corresponding section for the determination. In the example of Fig. 18, the determination unit 102 does not determine sections 1811, 1813, and 1814 as processing sections, but determines a section in section 1812 where the signal 1801 indicates that the section is a processing section (e.g., a high level section) as the processing section.

[0074] For example, when generating a section identification signal when a drill is used as the tool 213 to drill a hole in a workpiece, the diagnostic device 100 uses context information on the feed rate of the drill to the workpiece, detects that the feed rate has reached the feed rate during cutting, which indicates that the drill has started to drill a hole in the workpiece, and sets the section identification signal to high level (during machining). The diagnostic device 100 also detects a change in the context information when the feed rate has changed from the feed rate during cutting, which indicates the time point when the drill has finished drilling a hole in the workpiece, to the feed rate during non-cutting, and sets the section identification signal from high level to low level (not during machining).

[0075] As another method, the coordinate information of the machining position may be used as context information, and the section identification signal may be generated based on the movement of the drill or the machining object to the machining coordinate position and the movement of the drill or the machining object from the machining coordinate position. In this way, the diagnostic device 100 may use any context information as long as it is capable of identifying the machining section in which the tool 213 is actually machining the machining object.

[0076] When the processing machine 200 generates the section identification signal, it is programmed in advance to generate a signal that makes the section identification signal high under the same conditions as those used by the diagnosis device 100 .

[0077] In this way, the section identification signal is used to create the next model using only the detection information during the period when the section identification signal is at a high level, rather than using the detection information for the entire period when the motor B is driven. Then, the state of the tool 213 is diagnosed based only on the detection information when the tool 213 is actually machining the workpiece. This makes it possible to determine the state of the tool 213 (for example, wear and chipping of the drill bit) based only on the detection information when the tool 213 and the workpiece actually come into contact with each other, improving the diagnosis result.

[0078] (Variation 3) As described above, when determining whether or not something is normal, the value of the likelihood itself may be used, or a value indicating the fluctuation of the likelihood may be used. In Modification 3, an example of the value indicating the fluctuation of the likelihood will be described. For example, the variance of the likelihood may be used as the value indicating the fluctuation of the likelihood. For example, the variance of the likelihood in the section X is calculated by the following formula (1). n X is the number of frames in the interval X, k is the frame index, and x i is the likelihood at frame i (frame likelihood), μ X represents the average frame likelihood in the section X. Note that a frame corresponds to a unit section for calculating the likelihood.

number

[0079] Furthermore, a value obtained by scoring based on the variance as in formula (1) may be used for the judgment. For example, r(k) shown in the following formula (2) may be used for the judgment. r(k)=V S (k) / V L (k) ...(2)

[0080] S and L represent the type of section X. S represents a short section, and L represents a long section. S (k) indicates the value calculated by equation (1) for a short section. V L(k) indicates the value calculated by equation (1) for the long section. For each of S and L, n X , μ X The value corresponding to n S , μ S , and n L , μ L is calculated.

[0081] 19 is a graph showing an example of the relationship between the likelihood and the judgment value r(k). In a judgment method in which the likelihood is compared with a threshold, for example, if the threshold is set to 0.97, then the range between 200 and 300 frames is judged as normal. On the other hand, in a judgment method in which the judgment value r(k) is compared with a threshold (for example, 1.0), it is possible to detect the fluctuation in the likelihood between 200 and 300 frames and judge it as abnormal.

[0082] (Variation 4) The determination unit 102 may be configured to not only determine whether the likelihood is normal or abnormal by comparing the likelihood with a threshold value, but also determine and output a change in the likelihood over time. For example, when one normal model is used, the determination unit 102 may store the calculated likelihood in the storage unit 111 or the like, and obtain a change in the likelihood (whether it is decreasing over time, etc.).

[0083] 20 is a diagram showing an example of the change in likelihood over time. The determination unit 102 can determine the change in likelihood over time by, for example, storing the determined time and the likelihood. The determination unit 102 is similar to the above embodiment in that it determines that an abnormality has occurred when the likelihood falls below a threshold. The determination unit 102 may be further configured to determine and output information indicating the change in likelihood over time, and the time when the likelihood is predicted to fall below the threshold.

[0084] (Variation 5) The determination unit 102 may be configured to determine the degree (rank, level) of quality using multiple thresholds. Figures 21 and 22 are diagrams for explaining an example in which multiple thresholds are used. Each diagram shows an example in which four thresholds (threshold 1 to threshold 4) are used, but the number of thresholds is not limited to four.

[0085] 21 shows an example in which quality ranks are divided into five levels, R1 to R5, based on the magnitude relationship with four thresholds. The determination unit 102 determines and outputs the quality rank by comparing each threshold with the likelihood. The quality rank can be used as information (quality information) indicating the quality of an object (workpiece) machined by the processing machine 200, for example.

[0086] 22 shows an example in which it is determined whether the quality is normal (no problem), requires caution, or is abnormal (below the quality standard) based on the magnitude relationship with four thresholds. If it is normal, it is further divided into two levels, L11 and L12. Similarly, if it requires caution, it is further divided into two levels, L21 and L22. It may be further divided into three or more levels.

[0087] (Variation 6) The determination unit 102 may calculate a plurality of likelihoods using a plurality of models determined according to the quality rank, and determine whether or not the quality is normal, as well as the quality rank if the quality is normal. Figs. 23-1 to 23-3 are diagrams for explaining an example of the determination method according to the sixth modification.

[0088] 23-1, 23-2, and 23-3 show examples of likelihoods calculated by models for determining quality ranks R1, R2, and R3, respectively. Quality rank R1 represents, for example, quality determined to be normal in period 1, which is the early stage of the change over time. Quality rank R2 represents, for example, quality determined to be normal in period 2, which is the middle stage of the change over time. Quality rank R3 represents, for example, quality determined to be normal in period 3, which is the final stage of the change over time.

[0089] The determination unit 102 calculates multiple likelihoods using multiple models corresponding to such multiple quality ranks. Usually, the likelihood when any one of the models is used exceeds a threshold. The determination unit 102 determines the quality rank indicated by the model that calculated the likelihood when the threshold is exceeded, and outputs it as a determination result. If the likelihood does not exceed the threshold when any of the models is used, the determination unit 102 determines that an abnormality has occurred.

[0090] A plurality of models may be generated in advance for each quality rank. According to this modification, it is possible to determine not only whether or not something is normal, but also the level of quality. It is also possible to determine which period of time change the current situation corresponds to.

[0091] (Variation 7) The determination unit 102 may execute the determination by switching between a plurality of models depending on the elapsed time. Fig. 24-1 to Fig. 24-3 are diagrams for explaining an example of the determination method according to the seventh modification.

[0092] FIG. 24-1, FIG. 24-2, and FIG. 24-3 show examples of likelihoods calculated by models generated for Period 1, Period 2, and Period 3, respectively. Period 1, Period 2, and Period 3 correspond to the initial, middle, and final periods of the time-dependent change, respectively. Each model may be generated in advance using detection information for the corresponding period.

[0093] The determination unit 102 determines which model period to apply by using context information for identifying a period such as cumulative usage time. The determination unit 102 determines whether or not the state is normal using the identified model. According to this modification, it is possible to determine whether the current state is a valid change over time (normal state) or a state that deviates from the valid change over time (abnormal state).

[0094] (Variation 8) The determination unit 102 may perform the determination using a specified model among a plurality of models determined according to the quality rank. For example, when a workpiece of higher quality is requested, the determination unit 102 uses a model determined for a high quality rank. When a workpiece of normal quality is requested, the determination unit 102 uses a model determined for a quality rank lower than the high quality rank. Which model to use may be determined using, for example, context information received by the receiving unit 101a and information received via the accepting unit 103 or the like.

[0095] The program executed by the diagnostic device of this embodiment is provided in a state that it is pre-installed in a ROM or the like.

[0096] The program executed by the diagnostic device of this embodiment may be configured to be provided as a computer program product by being recorded on a computer-readable recording medium such as a CD-ROM, a flexible disk (FD), a CD-R, or a digital versatile disk (DVD) in an installable or executable format file.

[0097] Furthermore, the program executed by the diagnostic device of the present embodiment may be stored on a computer connected to a network such as the Internet and provided by being downloaded via the network. Also, the program executed by the diagnostic device of the present embodiment may be provided or distributed via a network such as the Internet.

[0098] The program executed by the diagnostic device of this embodiment has a modular structure including each of the above-mentioned units (communication control unit, judgment unit, etc.), and in terms of actual hardware, the CPU (processor) reads the program from the ROM and executes it, thereby loading each of the above-mentioned units onto the main memory device, and each unit is generated on the main memory device. [Explanation of symbols]

[0099] 51 CPU 52 ROM 53 RAM 54 Communication I / F 55 Drive control circuit 56 Motor 57 Sensors 58 Bus 61 CPU 62 ROM 63 RAM 64 Communication I / F 65 HDD 66 Bus 100 Diagnostic Equipment 101 Communication control unit 101a Receiver 101b Transmitter 102 Judgment section 103 Reception 104 Feature Extraction Unit 105 Generation part 111 Storage section 200 processing machines 201 Numerical Control Unit 202 Communication control section 203 Machine tools 211 Sensors 212 Drive unit [Prior art documents] [Patent documents]

[0100] [Patent Document 1] JP 2006-184722 A

Claims

1. a receiving unit that receives, from the target device, context information corresponding to a predetermined operation among a plurality of context information that specifies the operation of an operation target constituting the target device, the context information being determined for each type of operation performed by the target device, and detection information of a physical quantity that changes in response to the operation of the operation target; a determination unit that determines a state of the motion object using detection information when the motion object is performing a predetermined motion and a model corresponding to the received context information among one or more models defined for each of the one or more pieces of context information; A diagnostic device comprising:

2. the determination unit identifies detection information when the motion object is performing a predetermined motion based on the context information, and determines a state of the motion object using the identified detection information and the model. The diagnostic device of claim 1 .

3. The determination unit determines a state of the motion object during a period in which the motion object is performing a predetermined motion, the period being identified by the received context information, from the received detection information, and using a model corresponding to the received context information. The diagnostic device of claim 1 .

4. The receiving unit receives a plurality of pieces of detection information each corresponding to a different physical quantity, the determination unit determines a state of the object by using the detection information determined according to the context information among the plurality of detection information and a model corresponding to the received context information; The diagnostic device of claim 1 .

5. A generation unit that, when context information for which a model is not defined is received, generates a model for the context information for which a model is not defined by using the detection information corresponding to the context information. The diagnostic device of claim 1 .

6. a receiving unit that receives input of context information different from the context information received by the receiving unit, the determination unit determines a state of the object of operation by using the received detection information and a model corresponding to at least one of the context information received by the receiving unit and the context information accepted by the accepting unit. The diagnostic device of claim 1 .

7. the determination unit determines a period to be used for determination from among the periods specified by the received context information, based on the received detection information, and determines a state of the motion target using the detection information for the determined period and the model. The diagnostic device of claim 1 .

8. the determination unit determines a likelihood of the received detection information for the model, and determines a state of the motion object by comparing at least one of the likelihood and a value indicating a fluctuation of the likelihood with a threshold value; The diagnostic device of claim 1 .

9. The model is generated using detection information of the physical quantity when the target device is operating normally; The determination unit determines a likelihood of the received detection information with respect to the model, and outputs quality information determined in accordance with the likelihood. The diagnostic device of claim 1 .

10. the determining unit determines and outputs any one of a plurality of quality information by comparing at least one of the likelihood and a value indicating a fluctuation of the likelihood with a plurality of thresholds; The diagnostic device of claim 9.

11. A plurality of the models are determined according to the context information and quality; the determination unit determines a state of the operation target for each of a plurality of models, and outputs information indicating quality corresponding to a model determined to be normal. The diagnostic device of claim 1 .

12. The context information is a usage time of a drive unit included in the target device. The diagnostic device of claim 1 .

13. A plurality of the models are determined according to the context information and quality; The determination unit determines a state of the object of operation using a model corresponding to a specified quality and the received context information. The diagnostic device of claim 1 .

14. A receiving step of receiving, from the target device, context information corresponding to a predetermined operation among a plurality of context information defining an operation of an operation object constituting the target device, the context information being determined for each type of operation performed by the target device, and detection information of a physical quantity that changes in response to the operation of the operation object; a determination step of determining a state of the motion object using detection information when the motion object is performing a predetermined motion and a model corresponding to the received context information among one or more models defined for each of the one or more pieces of context information; A diagnostic method comprising:

15. On the computer, a receiving step of receiving, from the target device, context information corresponding to a predetermined operation among a plurality of context information defining an operation of an operation target constituting the target device, the context information being determined for each type of operation performed by the target device, and detection information of a physical quantity that changes in response to the operation of the target device; a determination step of determining a state of the motion object using detection information when the motion object is performing a predetermined motion and a model corresponding to the received context information among one or more models defined for each of the one or more pieces of context information; A program for executing the above.

16. A diagnostic system including a diagnostic device and a target device to be diagnosed by the diagnostic device, The diagnostic device comprises: a receiving unit that receives, from the target device, context information corresponding to a predetermined operation among a plurality of context information that specifies the operation of an operation target constituting the target device, the context information being determined for each type of operation performed by the target device, and detection information of a physical quantity that changes in response to the operation of the target device; a determination unit that determines a state of the motion object using detection information when the motion object is performing a predetermined motion and a model corresponding to the received context information among one or more models defined for each of the one or more pieces of context information; The target device is A detection unit that detects the physical quantity; A transmitter that transmits the detected physical quantity to the diagnostic device. Diagnostic system.

Images