Recognition method, recognition device, and recognition program

Abstract

The recognition device acquires connection relationship information indicating a mechanical connection relationship of the plurality of servo motors and the connection mechanism, acquires an abnormality degree of each of the plurality of servo motors, recognizes an abnormality site from among each of the plurality of servo motors and the connection mechanism on the basis of the abnormality degree of each of the plurality of servo motors and the connection relationship information, and displays the abnormality site on the display.

Description

Technical Field

[0001] This disclosure relates to techniques for identifying abnormal parts of a servo system in which multiple servo motors are mechanically connected. Background Technology

[0002] Patent document 1 discloses the following technology: to estimate multiple causal models between variables based on the relationship between multiple mechanisms in the process implemented on the manufacturing production line, to obtain the anomaly detection results of the manufacturing production line, and to estimate the cause of the anomaly in multiple causal models based on the contribution rate of the variables that contribute to the anomaly in the anomaly detection results.

[0003] However, the technology described in Patent Document 1 does not identify areas where no sensors are installed as abnormal areas. Therefore, a problem arises where, if no sensors are installed in the connection mechanism that mechanically connects the servo motors, the connection mechanism cannot be identified as an abnormal area if an abnormality occurs in that mechanism.

[0004] Prior art literature

[0005] Patent documents

[0006] Patent Document 1: Japanese Patent Application Publication No. 2023-092184 Summary of the Invention

[0007] This disclosure was made to solve such a problem, and its purpose is to provide a technique that can identify abnormal parts of the connection mechanism even without placing sensors in the connection mechanism that mechanically connects the servo motors to each other.

[0008] One aspect of this disclosure involves an identification method for identifying abnormal parts of a servo system comprising multiple servo motors and a connecting mechanism mechanically connecting the multiple servo motors. A computer performs the following processing: acquiring connection relationship information representing the mechanical connection between the multiple servo motors and the connecting mechanism; acquiring the abnormality degree of each of the multiple servo motors; based on the abnormality degree of each of the multiple servo motors and the connection relationship information; identifying the abnormal part from each of the multiple servo motors and the connecting mechanism; and displaying the abnormal part on a display.

[0009] According to this disclosure, even without installing sensors in the connection mechanism that mechanically connects the servo motors to each other, it is possible to identify the connection mechanism as an abnormal part. Attached Figure Description

[0010] Figure 1 This is a diagram illustrating an example of the structure of a production system according to Embodiment 1 of this disclosure.

[0011] Figure 2 This is a diagram illustrating the action pattern.

[0012] Figure 3 It is a graph showing the relationship between the threshold and the degree of anomaly.

[0013] Figure 4 This is a flowchart illustrating an example of the processing of the identification device during the saving phase of motion data.

[0014] Figure 5 This is an illustration of the process where motion data is stored in the motion data storage unit.

[0015] Figure 6 This is a flowchart illustrating an example of abnormal part identification processing performed by the identification device according to Embodiment 1.

[0016] Figure 7 This is an example of an abnormal display screen.

[0017] Figure 8 This is another example of an abnormal display screen.

[0018] Figure 9 This is an example of a connection relationship display bar that shows the connection relationships of a collaborative servo system.

[0019] Figure 10 This is a diagram illustrating an example of the structure of a production system according to Embodiment 2 of this disclosure.

[0020] Figure 11 This is a flowchart illustrating an example of anomaly cause identification processing performed by the identification device according to Embodiment 2.

[0021] Figure 12 This is an example of an abnormal display screen. Detailed Implementation

[0022] (The understanding that forms the basis of this disclosure)

[0023] Production equipment using servo motors mostly consists of multiple servo motors working in tandem. Specifically, multiple servo motors are mechanically connected to each other via belts, or multiple servo motors are mechanically connected via arm elements, similar to a robotic arm.

[0024] In such production equipment, anomalies can be identified using measurement signals from built-in sensors for multiple servo motors. However, a problem exists: if an anomaly occurs in the connection mechanism that mechanically connects multiple servo motors, the connection mechanism cannot be identified as the abnormal part if no sensor is installed in that connection mechanism.

[0025] Here, if sensors are installed in all connection mechanisms that may produce abnormalities, the abnormal parts of the connection mechanisms can be identified based on the measurement signals from the sensors. However, due to cost limitations, it is difficult to install sensors in all connection mechanisms. In addition, there are also parts where it is difficult to install sensors due to physical limitations.

[0026] In the prior art, which presumes the cause of the abnormality, starting with Patent Document 1, as mentioned above, there is a problem that the connection mechanism without a sensor cannot be identified as an abnormal part.

[0027] This disclosure was made to solve such a problem, and its purpose is to provide a technique that can identify abnormal parts of the connection mechanism even without placing sensors in the connection mechanism that mechanically connects the servo motors to each other.

[0028] (1) One aspect of the present disclosure is a method for identifying abnormal parts of a servo system comprising multiple servo motors and a connecting mechanism that mechanically connects the multiple servo motors. The computer performs the following processing: acquiring connection relationship information representing the mechanical connection relationship between the multiple servo motors and the connecting mechanism; acquiring the abnormality degree of each of the multiple servo motors; based on the abnormality degree of each of the multiple servo motors and the connection relationship information, identifying the abnormal part from each of the multiple servo motors and the connecting mechanism; and displaying the abnormal part on a display.

[0029] According to this structure, since it acquires the anomaly degree of each of the multiple servo motors and the connection relationship information representing the mechanical connection relationship of the multiple servo motors, it is possible not only to identify multiple servo motors as abnormal parts, but also to identify the connection mechanism as an abnormal part based on this anomaly degree and connection relationship information. Therefore, even without installing sensors in the connection mechanism, the connection mechanism can be identified as an abnormal part.

[0030] (2) In the identification method described in (1) above, the connection relationship information may also be constituted by a chart data in which multiple blocks represent the multiple servo motors respectively and the connection relationship is represented by connecting lines that connect the multiple blocks to each other. The chart data is generated based on input operations performed by the operator.

[0031] In this case, since the connection relationship information is constructed by using a diagrammatic data that represents the servo motor with blocks and the connecting mechanism with connecting lines, the connection relationship information can be represented in an easy-to-understand way.

[0032] (3) In the identification method described in (2) above, the chart data can be further displayed on the display, and a mark indicating that it is the abnormal part can be displayed in the block or the connecting line corresponding to the abnormal part.

[0033] In this situation, the abnormal location can be easily identified.

[0034] (4) In any of the identification methods described in (1) to (3) above, if the connection relationship information is not obtained, the abnormal part is identified from each of the multiple servo motors based on the abnormality degree of each of the multiple servo motors, the abnormality degree corresponding to each of the multiple servo motors is displayed on the display, and the abnormality degree of the servo motor that is identified as the abnormal part is highlighted on the display.

[0035] In this situation, even if connection information cannot be obtained, it is possible to avoid situations where the servo motor does not indicate any abnormalities.

[0036] (5) In any of the identification methods described in (1) to (4) above, if the plurality of servo motors includes a first servo motor and a second servo motor connected to the first servo motor via the connecting mechanism and operated due to the operation of the first servo motor, the identification of the abnormal part includes: determining whether the abnormality degree of each of the first servo motor and the second servo motor indicates an abnormality; if it is determined that the abnormality degree of both the first servo motor and the second servo motor indicates an abnormality, identifying the connecting mechanism of the first servo motor and the second servo motor as the abnormal part; and if it is determined that the abnormality degree of only one of the first servo motor and the second servo motor indicates an abnormality, identifying that servo motor as the abnormal part.

[0037] In this case, among the first and second servo motors that have a causal relationship in their operation, it is possible to accurately identify which of the first servo motor, the second servo motor, and the connecting mechanism is the abnormal part.

[0038] (6) In any of the identification methods described in (1) to (7) above, when the servo system includes an action unit composed of multiple co-servo motors that perform coordinated actions among the multiple servo motors, the identification of the abnormal part includes: determining whether the abnormality degree of each of the multiple co-servo motors indicates an abnormality; if it is determined that all the abnormality degrees of the multiple co-servo motors indicate an abnormality, identifying the action unit as the abnormal part; if it is determined that not all the abnormality degrees of the multiple co-servo motors indicate an abnormality, identifying the co-servo motor whose abnormality degree indicates an abnormality as the abnormal part.

[0039] In this case, among multiple coordinated servo motors, it is possible to accurately identify which of the coordinated servo motors and motion units is the abnormal part.

[0040] (7) In any of the identification methods described in (1) to (6) above, it may be further to obtain object action data from the servo system, further to retrieve the registration action data with the highest similarity to the object action data from among the multiple registration action data that are pre-registered in the memory and associated with the abnormal cause information representing the abnormal cause, and to presume the abnormal cause represented by the abnormal cause information associated with the retrieved registration action data as the abnormal cause of the servo system, and further to display the presumed abnormal cause on the display.

[0041] In this case, the cause of the anomaly represented by the object's action data can be accurately determined, and the determined cause of the anomaly can be indicated.

[0042] (8) In the identification method described in (7) above, the object action data and the plurality of registered action data may respectively include: a command signal for driving the servo motor, a measurement signal of the servo motor when the servo motor moves based on the command signal, and the abnormality degree. The display of the display includes: displaying the command signal, the measurement signal and the abnormality degree contained in the retrieved registered action data, and the command signal, the measurement signal and the abnormality degree contained in the object action data.

[0043] In this case, since the command signals, measurement signals, and anomaly cause information contained in the retrieved registered action data, as well as the command signals, measurement signals, and anomaly degree corresponding to the object action data, are displayed, the operator can analyze the cause of the anomaly while comparing these signals.

[0044] (9) In the identification method described in (8) above, the plurality of registered action data may further include at least one of the following: a judgment result based on the operator’s normal or abnormal status; the measurement date and time of the retrieved registered action data; and a description related to the recovery operation for restoring the servo system to normal operation. The display of the display further includes: displaying at least one of the judgment result associated with the retrieved registered action data, the measurement date and time, and the description related to the recovery operation.

[0045] In this case, since at least one of the judgment result, measurement date and time, description related to the cause, and description related to the recovery operation contained in the retrieved registration action data is displayed, useful information for analyzing the cause of the anomaly can be provided.

[0046] (10) In the identification method described in (8) or (9) above, the display of the display may further include: displaying the instruction signal, the measurement signal and the abnormality contained in the retrieved registered action data and the instruction signal, the measurement signal and the abnormality contained in the object action data in a way that distinguishes the action mode.

[0047] In this case, since the command signals, measurement signals, and anomalies contained in the retrieved registered action data, as well as the command signals, measurement signals, and anomalies corresponding to the object action data, can be displayed differently for each action mode, it is possible to analyze the cause of anomalies for each action mode.

[0048] (11) In the identification method described in (8) or (9) above, the similarity may also be the similarity between the instruction signal contained in the plurality of registered action data and the instruction signal contained in the object action data.

[0049] In this case, the similarity can be calculated accurately.

[0050] (12) In any of the identification methods described in (1) to (11) above, the computer may be any of a servo amplifier, a motion controller, a personal computer, and a cloud server.

[0051] In this case, the abnormal location can be identified using any of the following: a servo amplifier, a motion controller, a personal computer, or a cloud server.

[0052] (13) In another aspect of this disclosure, the identification device identifies abnormal parts of a servo system comprising a plurality of servo motors and a connection mechanism mechanically connecting the plurality of servo motors. The processor of the identification device performs the following processing: acquiring connection relationship information representing the mechanical connection relationship between the plurality of servo motors and the connection mechanism; acquiring the abnormality degree of each of the plurality of servo motors; identifying the abnormal part from each of the plurality of servo motors and the connection mechanism based on the abnormality degree of each of the plurality of servo motors and the connection relationship information; and displaying the abnormal part on a display.

[0053] Based on this structure, an identification device can be provided that can identify the connection mechanism as an abnormal part even without installing a sensor in the connection mechanism.

[0054] (14) In another aspect of this disclosure, the identification procedure causes a computer to perform the following processing: acquiring connection relationship information representing the mechanical connection relationship between the plurality of servo motors and the connection mechanism; acquiring the degree of abnormality of each of the plurality of servo motors; identifying the abnormal part from each of the plurality of servo motors and the connection mechanism based on the degree of abnormality of each of the plurality of servo motors and the connection relationship information; and displaying the abnormal part on a display.

[0055] Based on this structure, an identification program can be provided that can identify the connection mechanism as an abnormal part even without placing a sensor in the connection mechanism.

[0056] This disclosure can also be implemented as an abnormal part identification system that operates based on such an abnormal part identification program. Furthermore, it goes without saying that such a computer program can be distributed via computer-readable non-transitory recording media such as CD-ROMs or communication networks such as the Internet.

[0057] Furthermore, the embodiments described below are all specific examples of this disclosure. The numerical values, shapes, constituent elements, steps, and order of steps shown in the following embodiments are examples and are not intended to limit this disclosure. In addition, any constituent elements in the following embodiments that are not described in the independent claims representing the highest-level concept are described as arbitrary constituent elements. Furthermore, in all embodiments, the various contents can be combined.

[0058] (Implementation Method 1)

[0059] Figure 1This is a diagram illustrating an example of the structure of a production system 1 according to Embodiment 1 of this disclosure. The production system 1 includes: an identification device 10, a motion controller 20, a servo amplifier 30, and a servo system 40. The motion controller 20, the servo amplifier 30, and the servo system 40 constitute a production equipment 60. The production equipment 60 is an example of a device.

[0060] The identification device 10 is used to identify abnormal parts of the servo system 40. The servo system 40 includes multiple servo motors M and a connection mechanism K that mechanically connects the multiple servo motors M. The connection mechanism K consists of a mechanical component that transmits power from one servo motor M to another. Examples of the connection mechanism K include a belt, an arm mechanism that connects multiple joints of a robot arm to each other, a plate, etc. Details of the structure of the servo system 40 will be described later.

[0061] The identification device 10 can be a personal computer installed on-site where the motion controller 20, servo amplifier 30, and servo system 40 are located, or it can be a cloud server. The motion controller 20, servo amplifier 30, and identification device 10 are connected via a network NT. An example of a network NT is a local area network or the Internet. The motion controller 20 and servo amplifier 30 are connected to each other, and the servo amplifier 30 and servo system 40 are connected via LAN cables or the like.

[0062] The servo system 40 includes, for example, a production device used for manufacturing products. An example of a production device is an industrial robot that performs tasks such as machine installation, processing, operation, or handling. The servo system 40 includes a work arm that grips or processes parts. The work arm includes multiple arm elements and one or more joints connecting the multiple arm elements. The production device is, for example, installed on a production line in a factory. Although the production device has been described as being composed of an industrial robot, this is just an example; any device driven by a servo motor can be configured in any way.

[0063] The servo motor M is a motor that precisely operates the production device according to the drive signal output from the servo amplifier 30. For example, the servo motor M is installed in the joint of the work arm, causing the arm element to rotate a given angle in either the forward or reverse direction. The servo motor M includes a sensor (not shown) that detects the state of the servo motor M, and the measurement signal detected by the sensor is input to the servo amplifier 30. One example of the sensor is a torque sensor that detects the torque of the servo motor M. Therefore, the measurement signal is a torque signal. However, this is just one example; the measurement signal can also be an acceleration signal or a speed signal, and is not limited to a torque signal.

[0064] The servo amplifier 30 controls the servo motor M so that the servo motor M moves according to the command signal output from the motion controller 20. The servo amplifier 30 generates a drive signal corresponding to the command signal output from the motion controller 20 and inputs it to the servo motor M. Based on the measurement signal output from the servo motor M, the servo amplifier 30 performs feedback control on the servo motor M so that the servo motor M performs the action corresponding to the command signal.

[0065] The motion controller 20 generates a command signal to move the servo motor M in a predetermined action mode, and inputs the generated command signal to the servo amplifier 30. The command signal is a time-series data specifying the position of the servo motor M from the beginning to the end of a certain action mode.

[0066] Furthermore, while the command signal has been described as a position command signal specifying the position of the servo motor M, this is only one example. The command signal can also be a speed command signal, acceleration command signal, or torque command signal specifying the speed, acceleration, or torque of the servo motor M. Alternatively, the command signal can include at least two of the following: position command signal, speed command signal, acceleration command signal, and torque command signal.

[0067] The identification device 10 includes a processor 11, a communication device 12, an input device 13, a display 14, and a memory 15. The processor 11 is composed of a central processing unit (CPU) and includes an acquisition unit 111, a first identification unit 112, and a display control unit 113. In addition, the acquisition unit 111 to the display control unit 113 may also be composed of a dedicated integrated circuit such as an ASIC.

[0068] The acquisition unit 111 uses the communication device 12 to acquire command signals and measurement signals sent from the servo amplifier 30. Based on the acquired command signals and measurement signals, the acquisition unit 111 generates motion data and stores it in the motion data storage unit 151.

[0069] The acquisition unit 111 acquires connection relationship information representing the mechanical connection relationships between multiple servo motors M and the connection mechanism K from the connection relationship information storage unit 154. The connection relationship information is composed of graphical data where multiple servo motors M are represented by multiple blocks, and the connection relationships are represented by connecting lines linking the blocks together. This graphical data is generated based on input operations performed by the operator. The operator generates the graphical data by performing an input operation based on the actual structure of the servo system 40, arranging blocks corresponding to the servo motors M on the screen of the drawing application, and connecting the arranged blocks with connecting lines according to the connection relationships between the servo motors M. The acquisition unit 111 generates connection relationship information containing the input graphical data and saves the generated connection relationship information in the connection relationship information storage unit 154. Therefore, the connection relationship information becomes pre-created information. An operator is someone who manages the production equipment 60.

[0070] The acquisition unit 111 acquires the anomaly degree representing the time-varying anomaly degree of each of the multiple servo motors M. Specifically, the acquisition unit 111 acquires the anomaly degree by inputting the command signal sent from the servo amplifier 30 and the measurement signal into the learning model stored in the learning model storage unit 152.

[0071] The learning model is generated through unsupervised machine learning on the normal motion data (hereinafter referred to as normal data) of the servo system 40. The learning model is a model that takes input command signals and measurement signals and outputs the degree of abnormality of the servo system 40 based on the input command signals and measurement signals. Therefore, the normal command signals and measurement signals contained in the motion data storage unit 151 are used as the learning data for the learning model.

[0072] As an algorithm for learning models, methods such as k-nearest neighbors or k-means can be used. Anomaly score is an indicator representing the degree of deviation between action data and normal data; the greater the deviation, the larger the score, and vice versa.

[0073] The following describes an example of the learning model's processing using the k-nearest neighbors method. The learning model calculates the distances between the command signal and measurement signal that become the judgment target and multiple normal command signals and measurement signals. Next, the learning model extracts the top k normal command signals and measurement signals from among the multiple normal command signals and measurement signals, ranked from closest to furthest from the command signal and measurement signal that became the judgment target. The learning model then calculates the average distance between the extracted top k command signals and measurement signals and the command signal and measurement signal that became the judgment target as the anomaly score. Furthermore, the distance is calculated using the Euclidean distance between a vector defined by the values ​​of the normal command signals and measurement signals at a certain time point t and a vector defined by the values ​​of the command signals and measurement signals that became the judgment target at time point t. Thus, the anomaly score is composed of time series data calculated based on time point t.

[0074] The first identification unit 112 identifies abnormal parts from each of the multiple servo motors M and the connection mechanism K based on the abnormality degree and connection relationship information of each of the multiple servo motors M. The details of the identification of abnormal parts will be described later.

[0075] The display control unit 113 displays the abnormal parts identified by the first identification unit 112 on the display screen 14.

[0076] The display control unit 113 displays the graphical data contained in the connection relationship information on the display 14, and displays a mark indicating that it is an abnormal part on the block or connecting line corresponding to the abnormal part.

[0077] The communication device 12 connects the identification device 10 to the network NT. The communication device 12 receives command signals and corresponding measurement signals from the servo amplifier 30. The communication device 12 may also receive command signals and corresponding measurement signals from the motion controller 20.

[0078] The input device 13 consists of a mouse, keyboard, or touch panel, etc., and accepts instructions from the user.

[0079] Display 14 is composed of a display device such as a liquid crystal display or an organic EL display, and displays... Figure 7 The error message shown is 700.

[0080] The memory 15 is composed of a hard disk drive (HDD) or a solid-state drive (SSD). The memory 15 includes: an action data storage unit 151, a learning model storage unit 152, a threshold storage unit 153, and a connection relationship information storage unit 154.

[0081] The motion data storage unit 151 stores the motion data generated by the acquisition unit 111.

[0082] The learning model storage unit 152 stores a learning model that receives input command signals and measurement signals and outputs anomaly scores. Furthermore, the learning model can also be composed of a learning model that has undergone machine learning pre-processed for each action mode. For example, the learning model can also be composed of learning models corresponding to acceleration mode, deceleration mode, transition mode, and stable mode. For example, the learning model for acceleration mode is generated by performing unsupervised machine learning on the command signals and measurement signals during normal operation in acceleration mode. In this case, the acquisition unit 111 only needs to input the command signals and measurement signals into the learning model for the corresponding action mode, such as inputting the command signals and measurement signals for acceleration mode into the learning model for acceleration mode, and inputting the command signals and measurement signals for deceleration mode into the learning model for deceleration mode, thereby calculating the anomaly score.

[0083] The threshold storage unit 153 stores the threshold for comparing the anomaly level calculated by the learning model. In this embodiment, the threshold is stored for the acceleration mode, deceleration mode, transition mode, and stable mode, respectively.

[0084] Figure 2 This is a diagram illustrating the action patterns. In Figure 2 The waveforms of the position command signal, velocity command signal, acceleration command signal, and measurement signal are shown from top to bottom.

[0085] The operation period of the servo motor M is divided into an acceleration period P1, a deceleration period P2, and a constant speed period P3. Furthermore, the constant speed period P3 is further divided into a transition period P31, which includes the initial phase of the constant speed period P3, and a stabilization period P32, which includes the final phase of the constant speed period P3. The transition period P31 is the period when the speed command signal value is zero, but the measured signal value is greater than the given value due to inertia. The stabilization period P32 is the period when the measured signal value is below the given value.

[0086] The acquisition unit 111 acquires a position command signal from the servo amplifier 30, and differentiates the acquired position command signal to generate a speed command signal. The acquisition unit 111 differentiates the speed command signal to generate an acceleration command signal. The acquisition unit 111 sets the period during which the absolute value of the acceleration command signal is above a certain value and has a positive sign as the acceleration period P1. The acquisition unit 111 sets the period during which the absolute value of the acceleration command signal is above a certain value and has a negative sign as the deceleration period P2. The acquisition unit 111 sets the period during which the absolute value of the acceleration command signal is below a certain value as the constant speed period P3. The acquisition unit 111 sets the period from the start time of the constant speed period P3 to the elapsed time of a given period as the transition period P31. During the constant speed period P3, the acquisition unit 111 sets the remaining period from the end time of the transition period P31 as the stabilization period P32.

[0087] The acquisition unit 111 determines the operation mode of the servo motor M during the acceleration period P1 as the acceleration mode, the operation mode of the servo motor M during the deceleration period P2 as the deceleration mode, the operation mode of the servo motor M during the constant speed period P3 as the constant speed mode, the operation mode of the servo motor M during the transition period P31 as the transition mode, and the operation mode of the servo motor M during the stabilization period P32 as the stabilization mode.

[0088] Figure 3 This is chart 300 showing the relationship between the threshold and the degree of anomaly. In Figure 3 In the diagram, the vertical axis represents the degree of abnormality, and the horizontal axis represents time. The degree of abnormality 303 is divided into an acceleration period P1, a deceleration period P2, and a constant speed period P3. The constant speed period P3 is further divided into a transition period P31 and a stabilization period P32. The threshold 302 is set to a value that varies for each action period, i.e., a value that varies for each action mode. In this example, the value of the threshold 302 decreases in the order of acceleration period P1, deceleration period P2, transition period P31, and stabilization period P32. When the degree of abnormality 303 is above the threshold 302, the acquisition unit 111 determines that the servo system 40 is abnormal; when the degree of abnormality 303 is below the threshold, it determines that the servo system 40 is normal. Therefore, for example, the stabilization period P32 is more likely to be determined as abnormal than the acceleration period P1 and the deceleration period P2. In this way, by setting different values ​​for the threshold according to the action mode, appropriate abnormality determination can be made according to the action mode.

[0089] Figure 4This is a flowchart illustrating an example of the processing of the identification device 10 during the data storage phase. First, in step S1, the acquisition unit 111 uses the communication device 12 to acquire the command signal and measurement signal sent from the servo amplifier 30. When the production equipment 60 is operating, the servo amplifier 30 sends the command signal and measurement signal at a given sampling rate; therefore, the acquisition unit 111 acquires the command signal and measurement signal at the given sampling rate.

[0090] Next, in step S2, the acquisition unit 111 assigns identifiers to the command signals and measurement signals acquired in step S1, generating motion data. Here, the motion data is managed on a per-action basis. Therefore, when the acquisition unit 111 acquires all time-series data of the command signals and measurement signals from the start to the end of one action mode, it assigns identifiers to the time-series data of the command signals and measurement signals, as well as the measurement date and time of the motion data, generating motion data.

[0091] Next, in step S3, the acquisition unit 111 saves the motion data generated in step S2 to the motion data storage unit 151. Furthermore, the operator can input, from the reason display screen 900 described in Embodiment 2, a judgment result indicating normal or abnormal operation, a description related to the cause of the abnormality, and a description related to the operation used to restore the production equipment 60 to normal operation, based on the saved motion data. When this information is input, the acquisition unit 111 simply includes this information in the content of the motion data. When the processing in step S3 ends, the processing returns to step S1. That is, Figure 4 The processing is repeatedly performed while the production equipment is operating at 60%.

[0092] Figure 5 This is an explanatory diagram of the processing of motion data stored in the motion data storage unit 151. When the acquisition unit 111 acquires a command signal and a measurement signal corresponding to a motion mode, it inputs the acquired command signal and measurement signal into the learning model, thereby calculating an anomaly level. The acquisition unit 111 compares the anomaly level with a threshold to determine whether the servo system 40 is abnormal or normal. If it is determined to be abnormal, the acquisition unit 111 establishes a correspondence between the anomaly level and the command signal and the measurement signal. On the other hand, if it is determined to be normal, the acquisition unit 111 establishes a correspondence between the normality level and the command signal and the measurement signal.

[0093] If an operator, having confirmed the absence of any abnormality in the servo system 40, inputs an indication indicating an abnormality using the input device 13, the acquisition unit 111 establishes a correspondence between the abnormality level and the command signal and the measurement signal. Conversely, if an operator inputs an indication indicating normal operation using the input device 13, the acquisition unit 111 establishes a correspondence between the normality level and the command signal and the measurement signal.

[0094] When an operator inputs a comment using input device 13, acquisition unit 111 establishes a correspondence between the comment and command signals and measurement signals. The comment includes the date and time of the anomaly, a memo describing the confirmation content for servo system 40, and details of the recovery operation.

[0095] The acquisition unit 111 generates motion data by establishing a correspondence between the level and comment and the command signal and measurement signal, and saves the generated motion data in the motion data storage unit 151. At this time, the acquisition unit 111 saves motion data containing normal levels in the normal data storage unit 161, and saves motion data containing abnormal levels in the abnormal data storage unit 162.

[0096] Thus, the acquisition unit 111 distinguishes between motion data containing normal levels and motion data containing abnormal levels, and stores them in the motion data storage unit 151.

[0097] Figure 6 This is a flowchart illustrating an example of abnormal part identification processing performed by the identification device 10 according to Embodiment 1. In step S11, the acquisition unit 111 acquires connection relationship information from the connection relationship information storage unit 154.

[0098] Next, in step S12, the acquisition unit 111 acquires the command signal and the measurement signal from the servo amplifier 30.

[0099] Next, in step S14, the acquisition unit 111 acquires the anomaly degree by inputting the instruction signal and measurement signal acquired in step S12 into the learning model.

[0100] Next, the first identification unit 112 identifies abnormal parts based on the connection relationship information obtained in step S11 and the degree of abnormality obtained in step S13.

[0101] Figure 7 This is an example of an error display screen 700. The following uses... Figure 7The detailed process for identifying and handling abnormal parts is explained. The abnormality display screen 700 includes a connection relationship display bar 710 and an abnormality degree display bar 720. The connection relationship display bar 710 displays graphical data included in the connection relationship information. In this example, a servo system 40 containing three servo motors M1, M2, and M3 and two connecting mechanisms K1 and K2 is displayed. Servo motors M1, M2, and M3 are displayed as blocks. Connecting mechanisms K1 and K2 are displayed as connecting lines linking the blocks together.

[0102] Servo motors M1 and M2 are connected via a connecting mechanism K1. The connecting mechanism K1 transmits the motion of servo motor M1 to servo motor M2. Servo motor M2 operates in response to the motion of servo motor M1. Servo motors M2 and M3 are connected via a connecting mechanism K2. The connecting mechanism K2 transmits the motion of servo motor M2 to servo motor M3. Servo motor M3 operates in response to the motion of servo motor M2.

[0103] The arrows on the connecting lines indicate the direction of power transmission from the servo motor M. In this example, power is transmitted from servo motor M1 to servo motor M2, and from servo motor M2 to servo motor M3. Therefore, the arrow on the connecting line corresponding to connecting mechanism K1 points from servo motor M1 to servo motor M2, and the arrow on the connecting line corresponding to connecting mechanism K2 points from servo motor M2 to servo motor M3. Such a connection relationship in the servo system 40, where one servo motor M operates due to the action of another servo motor M, is called a causal relationship.

[0104] The anomaly display bar 720 displays a graph showing the time progression of the anomalies of servo motors M1 to M3, which are displayed in the connection relationship display bar 710. The vertical axis of the graph represents the anomaly degree, and the horizontal axis represents time. Anomalies 721, 722, and 723 represent the anomalies of servo motors M1, M2, and M3, respectively.

[0105] In this example, the anomaly level 721 of servo motor M1 did not exceed the threshold, but the anomaly level 722 of servo motor M2 exceeded the threshold. Subsequently, the anomaly level 723 of servo motor M3 exceeded the threshold. Servo motors M1 and M2 are components that are not prone to failure. Therefore, the probability of an anomaly occurring in servo motors M2 and M3 is low, while the probability of an anomaly occurring in the connecting mechanism K2 is high. Therefore, the first identification unit 112 does not identify servo motors M2 and M3 as abnormal parts, but identifies the connecting mechanism K2 as an abnormal part.

[0106] That is, the first identification unit 112 determines whether the abnormality levels 721, 722, and 723 of each of the servo motors M1 (an example of the first servo motor), M2 (an example of both the first and second servo motors), and M3 (an example of the second servo motor) indicate an abnormality. Here, the first identification unit 112 only needs to determine that the abnormality levels 721 to 723 indicate an abnormality if the period during which the abnormality levels 721 to 723 exceed the threshold is longer than a given time. If the first identification unit 112 determines that the abnormality levels of both servo motors M2 and M3 indicate an abnormality, it identifies the connecting mechanism K2 provided between servo motors M2 and M3 as an abnormal part.

[0107] Figure 8 This is another example of an abnormal display screen 700. In Figure 8 In the middle, the structure of the servo system 40 and Figure 7 The same applies. In this example, only the abnormality degree 722 of servo motor M2 indicates an abnormality. Therefore, the first identification unit 112 identifies servo motor M2 as an abnormal part. That is, when the first identification unit 112 determines that the abnormality degree of one of the servo motors M2 and M3 located on both sides of the connecting mechanism is abnormal, it does not identify the connecting mechanism K2 as an abnormal part, but instead identifies the servo motor M2 as an abnormal part.

[0108] Figure 9 This is a diagram illustrating an example of a connection relationship display bar 710 that shows the connection relationships of a collaborative servo system 40. Hereinafter, using... Figure 9 The anomaly identification method for the cooperative servo system 40 is explained. A cooperative servo system 40 is one in which multiple servo motors M work together to move a connecting mechanism K. For example, a structure in which multiple servo motors M work together to move the axle of a vehicle conforms to the cooperative servo system 40.

[0109] Figure 9 The servo system 40 shown includes two motion units 901 and 902. Motion unit 901 includes four servo motors M1 to M4. Servo motors M1 to M4 are examples of coordinated servo motors that cause the connecting mechanism K1 to move in tandem. An example of the connecting mechanism K1 is an axle. Motion unit 902 has the same structure as the motion unit.

[0110] The first identification unit 112 determines that only the abnormality level of the servo motor M1 is abnormal in the action unit 901. In this case, the first identification unit 112 identifies the servo motor M1 as an abnormal part.

[0111] On the other hand, the first identification unit 112 determines that the abnormality level of each of the servo motors M5 to M8 in the motion unit 902 is abnormal. Generally, since the servo motors M are difficult to fail, it is difficult to assume that all the servo motors M5 to M8 constituting the motion unit 902 fail simultaneously, and the connection mechanism K2 is more likely to fail. Therefore, the first identification unit 112 does not identify the servo motors M5 to M8 as abnormal parts, but identifies the connection mechanism K2 as an abnormal part.

[0112] Furthermore, it is assumed that the abnormality level of servo motors M1 and M2 in the action unit 901 indicates an abnormality. Although it is rare for multiple servo motors M1 and M2 to fail simultaneously, if the abnormality level of servo motors M3 and M4 connected to the common connection mechanism K1 does not indicate an abnormality, the possibility of servo motors M1 and M2 failing is relatively high. In this case, the first identification unit 112 identifies servo motors M1 and M2 as abnormal parts.

[0113] Return to reference Figure 6 In step S15, the display control unit 113 displays the abnormal location. Figure 7 In the example, since the connecting mechanism K2 is identified as an abnormal part, the display control unit 113 displays a mark R1 indicating the abnormal part on the servo motor M2. Figure 8 In the example, because servo motor M2 is identified as an abnormal part, a marker R1 is displayed for servo motor M2. Figure 9 In the example, since the servo motor M1 and the motion unit 902 are identified as abnormal parts, the display control unit 113 displays the mark R1 on the servo motor M1 and the motion unit 902.

[0114] Thus, according to this embodiment, since the anomaly degree of each of the multiple servo motors M and the connection relationship information representing the mechanical connection relationship of the multiple servo motors M are obtained, based on this anomaly degree and connection relationship information, not only can the multiple servo motors be identified as abnormal parts, but the connection mechanism K can also be identified as an abnormal part. Therefore, even without installing a sensor in the connection mechanism K, the connection mechanism K can be identified as an abnormal part.

[0115] (Implementation Method 2)

[0116] The identification device 10A in Embodiment 2 identifies the cause of the malfunction in the servo system 40. Figure 10 This is a diagram illustrating an example of the structure of the production system 1A according to Embodiment 2 of this disclosure. Furthermore, in this embodiment, the same reference numerals are used for the same components as in Embodiment 1, and descriptions are omitted. Relative to the identification device 10, the processor 11 further includes a second identification unit 114.

[0117] The acquisition unit 111 acquires command signals and measurement signals from the servo amplifier 30 and stores them in the motion data storage unit 151. Here, the acquired command signals and measurement signals are examples of object motion data. Object motion data refers to the motion data of the identified object that has become the cause of the anomaly. In this embodiment, the motion data stored in the motion data storage unit 151 is referred to as registered motion data.

[0118] The object action data and the registered action data respectively include: the command signal for driving the servo motor M, the measurement signal of the servo motor M when the servo motor moves based on the command signal, and the anomaly degree obtained by inputting the command signal and the measurement signal into the learning model.

[0119] The registered action data includes: the operator's judgment result (normal or abnormal), the measurement date and time of the retrieved registered action data, the cause of the abnormality, and a description related to the recovery operation used to restore the servo system 40 to normal operation. The cause of the abnormality information is a description of the cause of the abnormality entered by the operator.

[0120] The second identification unit 114 retrieves the registration action data with the highest similarity to the target action data from among multiple registration action data pre-registered in the action data storage unit 151, i.e., multiple registration action data associated with anomaly cause information indicating the cause of the anomaly. The second identification unit 114 presumes the anomaly cause represented by the anomaly cause information associated with the retrieved registration action data as the anomaly cause of the servo system 40. The operator inputs the cause display screen 900 based on the operator's judgment result of normal or abnormal, the operator's description related to the cause of the anomaly, and the description related to the operation used to restore the servo system 40 to normal operation. The measurement date and time of the action data refers to the year, month, day, and time when the measurement signal contained in the action data was measured. Furthermore, the action data may also contain the judgment result of normal or abnormal determined by the acquisition unit 111 based on the anomaly degree output by the learning model.

[0121] Similarity is the similarity between the command signals contained in multiple registered action data and the command signals contained in the object action data. Here, the command signals contained in the multiple registered action data are called registered command signals, and the command signals contained in the object command signals are called object command signals. Similarity is defined by the difference between the registered command signal and the object command signal. This difference is obtained by summing the differences between the values ​​of the registered command signal and the object command signal calculated for each of the multiple sampling points, from the start time point to the end time point of the object command signal. The smaller the difference between the registered command signal and the object command signal, the higher the similarity.

[0122] The display control unit 113 displays the cause of the abnormality deduced by the second identification unit 114 on the display screen 14.

[0123] The display control unit 113 displays the command signals, measurement signals, and anomalies contained in the registered action data retrieved by the second identification unit 114, and the command signals, measurement signals, and anomalies contained in the object action data.

[0124] The display control unit 113 displays the judgment result associated with the retrieved registration action data, the measurement date and time, the description related to the cause, and the description related to the recovery operation on the display 14.

[0125] The display control unit 113 can distinguish the operation modes of the instruction signals, measurement signals, and abnormality levels contained in the registered action data retrieved by the second identification unit 114, and the instruction signals, measurement signals, and abnormality signals contained in the object action data.

[0126] Figure 11 This is a flowchart illustrating an example of the anomaly cause identification processing performed by the identification device 10A according to Embodiment 2. First, in step S21, the acquisition unit 111 acquires object motion data from the servo amplifier 30. The object motion data includes command signals and measurement signals.

[0127] The following steps S22 and S23 are performed in parallel with the following step S24. In step S22, the second identification unit 114 retrieves the registered action data with the highest similarity to the object action data from among the multiple registered action data stored in the action data storage unit 151. Specifically, the second identification unit 114 retrieves the registered action data with the highest similarity by comparing the command signals contained in the multiple registered action data with the command signals contained in the object action data. Hereinafter, the registered action data with the highest similarity will be referred to as the retrieved registered action data.

[0128] In step S23, the second identification unit 114 calculates the anomaly degree of the retrieved registration action data. Specifically, the second identification unit 114 calculates the anomaly degree of the retrieved registration action data by inputting the instruction signal and measurement signal contained in the retrieved registration action data into the learning model.

[0129] In step S24, the second recognition unit 114 calculates the anomaly degree of the object motion data. Specifically, the second recognition unit 114 calculates the anomaly degree of the object motion data by inputting the command signal and measurement signal contained in the object motion data into the learning model.

[0130] Next, in step S25, the second identification unit 114 displays a display screen 900 on the monitor 14 showing the abnormal cause indicated by the abnormal cause information contained in the retrieved registration action data, and the cause display screen 900 showing the abnormality degree of each of the object action data and the retrieved registration action data.

[0131] Figure 12 This diagram illustrates an example of a cause display screen 900. In the right column of the cause display screen 900, display columns 1, 910, 920, and 930 are displayed. Display column 910 displays the command signal 911 of the object motion data. Display column 920 overlays the measurement signal 921 contained in the object motion data and the measurement signal 922 contained in the retrieved registration motion data. Display column 930 overlays the anomaly degree 931 of the object motion data and the anomaly degree 932 of the retrieved motion data. In display columns 910 to 930, the vertical axis represents the value of each signal, and the horizontal axis represents time. That is, display columns 910 to 930 display the command signal 911, measurement signals 921 and 922, and anomaly degrees 931 and 932 in a time sequence.

[0132] Display bars 910 to 930 distinguish between acceleration, deceleration, transition, and stabilization periods. For example, display bars 910 to 930 differentiate between each action period by using different background colors, displaying boundary lines, or displaying text indicating the action period. Thus, display bars 910 to 930 can distinguish between multiple action modes.

[0133] Display panes 910 to 930 display emphasized targets 990 indicating changes in anomaly levels. Here, emphasized targets 990 are displayed during periods when anomaly level 931 is above the threshold. The emphasized target 990 is a target whose rectangular area representing periods when anomaly level 931 is above the threshold is displayed in a semi-transparent color.

[0134] The left column of the cause display screen 900 displays: event summary display 940, level display 950, abnormal cause display 960, and restoration content display 970.

[0135] The event summary display bar 940 displays the event corresponding to the command signal 911 displayed in the first display bar 910. An event is information indicating a phenomenon that occurs in the servo system 40. In this example, since an anomaly occurred during acceleration in operation mode A, a message indicating this is displayed as an event.

[0136] The rating display bar 950 is where the operator inputs confirmation results to verify whether the servo system 40 has any abnormalities. Operators who determine the system is normal input a comment indicating "normal" in the rating display bar 950, while operators who determine the system is abnormal input a comment indicating "abnormal." For example, even if the learning model determines the system is normal but an actual abnormality occurs in the servo system 40, the operator can still input a comment indicating this in the rating display bar 950. Furthermore, the rating display bar 950 can also display a normal rating or an abnormal rating generated based on the acquisition unit 111's determination of the abnormality level 931.

[0137] The error cause display field 960 displays the error cause indicated by the error cause information contained in the registered action data. In this example, since the error cause is slippage of the belt connected to the servo system 40, a comment indicating this situation is displayed. Furthermore, if the error cause information does not correspond to the retrieved registration data, the error cause display field 960 becomes blank. In this case, the operator only needs to input the error cause in the error cause display field 960. Thus, the error cause information indicated by the error cause input into the error cause display field 960 is stored in the action data storage unit 151 in correspondence with the object action data. Therefore, later, when the object action data is retrieved as retrieved registration action data, the error cause corresponding to that retrieved registration action data is displayed in the error cause display field 960.

[0138] The recovery content display field 970 is where the operator inputs the job content for the recovery operation of the servo system 40 that has malfunctioned. Here, belt cleaning is entered as the job content.

[0139] The acquisition unit 111 updates the object action data by including the information input into the event summary display bar 940, the level display bar 950, and the restoration content display bar 970 into the object action data. Additionally, the acquisition unit 111 updates the object action data by including the exception cause information displayed in the exception cause display bar 960 into the object action data. The acquisition unit 111 then saves the updated object action data in the action data storage unit 151. This object action data is then used as registration action data.

[0140] Thus, according to this embodiment, the cause of the anomaly represented by the object's motion data can be accurately determined, and the determined cause of the anomaly can be indicated.

[0141] (Variation example)

[0142] The following variations are possible with respect to this disclosure.

[0143] (1) The identification device 10 can also be installed in the servo amplifier 30 or the motion controller 20, or in a personal computer or cloud server.

[0144] (2) The memory 15 can also be installed in the servo amplifier 30 or the motion controller 20, or in a personal computer or cloud server.

[0145] (3) In Embodiment 1, if the first identification unit 112 does not acquire connection relationship information from the acquisition unit 111, it only needs to identify the abnormal part from each of the multiple servo motors M based on the abnormality degree of each of the multiple servo motors M. The situation where connection relationship information is not acquired is, for example, the situation where the operator has not created the chart data of the servo system 40. In this case, the first identification unit 112 only needs to determine whether the abnormality degree of each of the multiple servo motors M exceeds a threshold, and determine the servo motor M whose abnormality degree exceeds the threshold as abnormal. In this case, the display control unit 113 can also highlight the servo motor M determined to be abnormal on the display. In this case, the display control unit 113 only needs to... Figure 7 In the connection relationship display bar 710, the block that indicates the servo motor M that has obtained the abnormality degree is displayed. The mark R1 can be displayed for the block that indicates the servo motor M that is abnormal.

[0146] Industrial availability

[0147] According to this disclosure, it is useful in production equipment that includes servo motors.

Claims

1. A method for identifying abnormal parts of a servo system comprising multiple servo motors and a connecting mechanism mechanically connecting the multiple servo motors. The computer performs the following processing: Obtain connection relationship information representing the mechanical connection relationship between the plurality of servo motors and the connection mechanism; Obtain the anomaly degree of each of the multiple servo motors; Based on the anomaly degree and connection relationship information of each of the plurality of servo motors, the abnormal part is identified from each of the plurality of servo motors and the connection mechanism; The abnormal area is displayed on the monitor.

2. The identification method according to claim 1, wherein, The connection relationship information is constituted by a graphical representation of the multiple servo motors, using multiple blocks to represent the multiple blocks respectively, and connecting lines connecting the multiple blocks to represent the connection relationship. The chart data is generated based on input operations performed by the operator.

3. The identification method according to claim 2, wherein, Furthermore, the chart data is displayed on the monitor. Furthermore, a mark indicating that it is the abnormal part is displayed in the block or the connecting line corresponding to the abnormal part.

4. The identification method according to claim 1 or 2, wherein, In the absence of the connection relationship information, based on the anomaly degree of each of the plurality of servo motors, the abnormal part is identified from each of the plurality of servo motors. The abnormality degree corresponding to each of the plurality of servo motors is displayed on the display, and the abnormality degree of the servo motors that are identified as abnormal parts is highlighted on the display.

5. The identification method according to claim 1 or 2, wherein, In the case where the plurality of servo motors includes a first servo motor and a second servo motor connected to the first servo motor via the connecting mechanism and operated due to the operation of the first servo motor, The identification of the abnormal parts includes: Determine whether the abnormality degree of each of the first servo motor and the second servo motor indicates an abnormality; If the abnormality degree of both the first servo motor and the second servo motor is determined to be abnormal, the connection mechanism of the first servo motor and the second servo motor is identified as the abnormal part. and If it is determined that the abnormality degree of one of the first servo motors and the second servo motor is abnormal, the servo motor of that individual is identified as the abnormal part.

6. The identification method according to claim 1 or 2, wherein, In the case where the servo system includes an action unit consisting of multiple coordinated servo motors that perform coordinated actions among the plurality of servo motors, The identification of the abnormal parts includes: Determine whether the abnormality degree of each of the plurality of cooperative servo motors indicates an abnormality; If it is determined that all of the abnormality degrees of the multiple collaborative servo motors are abnormal, the motion unit is identified as the abnormal part. If it is determined that not all of the multiple collaborative servo motors show an abnormality in their abnormality degree, the collaborative servo motors that show an abnormality in their abnormality degree are identified as the abnormal parts.

7. The identification method according to claim 1, wherein, Furthermore, object motion data is obtained from the servo system. Furthermore, from among the multiple registered action data that are pre-registered in the memory and associated with exception cause information representing exception reasons, the registered action data with the highest similarity to the object action data is retrieved, and the exception cause represented by the exception cause information associated with the retrieved registered action data is presumed as the exception cause of the servo system. Furthermore, the presumed cause of the anomaly is displayed on the monitor.

8. The identification method according to claim 7, wherein, The object action data and the plurality of registered action data respectively include: the command signal for driving the servo motor, the measurement signal of the servo motor when the servo motor moves based on the command signal, and the anomaly degree. The display includes: displaying the instruction signal, the measurement signal, and the anomaly degree contained in the retrieved registered action data, and the instruction signal, the measurement signal, and the anomaly degree contained in the object action data.

9. The identification method according to claim 8, wherein, The plurality of registered action data further includes at least one of the following: Based on the operator's judgment result of normal or abnormal; The date and time of measurement of the retrieved registration action data; and Description relating to the recovery operation used to restore the servo system to normal operation. The display further includes at least one of the following: displaying the judgment result associated with the retrieved registration action data, the measurement date and time, and a description related to the recovery operation.

10. The identification method according to claim 8 or 9, wherein, The display further includes: displaying the instruction signal, the measurement signal, and the anomaly degree contained in the retrieved registered action data, and the instruction signal, the measurement signal, and the anomaly degree contained in the object action data, respectively, in a manner that distinguishes the action patterns.

11. The identification method according to claim 8 or 9, wherein, The similarity is the similarity between the instruction signal contained in the plurality of registered action data and the instruction signal contained in the object action data.

12. The identification method according to claim 1 or 2, wherein, The computer can be any of a servo amplifier, motion controller, personal computer, or cloud server.

13. An identification device for identifying abnormal parts of a servo system comprising a plurality of servo motors and a connection mechanism mechanically connecting the plurality of servo motors. The processor of the identification device performs the following processing: Obtain connection relationship information representing the mechanical connection relationship between the plurality of servo motors and the connection mechanism; Obtain the anomaly degree of each of the multiple servo motors; Based on the anomaly degree and connection relationship information of each of the plurality of servo motors, the abnormal part is identified from each of the plurality of servo motors and the connection mechanism; The abnormal area is displayed on the monitor.

14. An identification program that enables a computer to perform an identification method for identifying abnormal parts of a servo system comprising multiple servo motors and a connecting mechanism mechanically connecting the multiple servo motors. The identification program causes the computer to perform the following processing: Obtain connection relationship information representing the mechanical connection relationship between the plurality of servo motors and the connection mechanism; Obtain the anomaly degree of each of the multiple servo motors; Based on the anomaly degree of each of the multiple servo motors and the connection relationship information, the abnormal part is identified from each of the multiple servo motors and the connection mechanism; The abnormal area is displayed on the monitor.

Images