Abnormal classification device

The abnormality classification device efficiently classifies industrial machine anomalies by analyzing past data to identify known and unknown causes, enhancing maintenance efficiency.

JP7590461B2Active Publication Date: 2024-11-26FANUC LTD
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
JP2022571592
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Priority Date
2020-12-25
Filing Date
2021-12-22
Publication Date
2024-11-26
Estimated Expiration
2041-12-22

Smart Images

  • Figure 0007590461000001
    Figure 0007590461000001
  • Figure 0007590461000002
    Figure 0007590461000002
  • Figure 0007590461000003
    Figure 0007590461000003
Patent Text Reader

Abstract

This abnormality classification device determines whether a detected abnormality is a known abnormality or an unknown abnormality, and presents a user with a response not only in the case of a known abnormality, but also in the case of an unknown abnormality. This abnormality classification device acquires data related to a physical quantity detected when an abnormality has occurred in an industrial machine as abnormality data, creates a model for determining whether abnormality data is due to a known cause of abnormality and a model for classifying to which cause of abnormality abnormality data belongs, and then uses the created models to determine whether the abnormality data is due to a known cause of abnormality and to classify the abnormality data as being due to one cause of abnormality.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to an anomaly classification device that classifies anomalies that occur in industrial machinery. [Background technology]

[0002] In manufacturing sites such as factories, industrial machines such as machine tools and robots are installed to form production lines, and products are manufactured by controlling each industrial machine. Each industrial machine is equipped with sensors that measure physical quantities related to its operating state (current value, voltage value, temperature, vibration, sound, etc. related to each part), and based on the physical quantities detected by these sensors, it is possible to detect whether the industrial machine is operating within the normal range or is operating abnormally.

[0003] In order to detect abnormal operation of an industrial machine, a model for detecting normal or abnormal states is created based on data related to physical quantities detected when the industrial machine is operating, and the operation of the industrial machine is judged based on the model. Here, industrial machines often operate normally, but rarely operate abnormally. Therefore, it is difficult to collect data related to physical quantities detected when the industrial machine is operating abnormally. Therefore, in order to detect abnormal operation of an industrial machine, unsupervised learning is performed using data detected when the industrial machine is operating within a normal range, and a model created as a result is used to detect a state far from the normal operation of the industrial machine as abnormal operation (Patent Document 1, etc.). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2017-033470 A Summary of the Invention [Problem to be solved by the invention]

[0005] When an abnormality is detected, the user identifies the cause of the abnormality based on the data acquired at the time the abnormality was detected and determines what should be done to resolve the abnormality. This is because the severity of the abnormality and the action that the user should take vary depending on the location and type of failure that caused the abnormality.

[0006] On the other hand, it takes a lot of time and money to collect data on physical quantities detected when industrial machinery operates abnormally, and it is therefore difficult to prepare a model from the beginning that classifies all possible abnormalities that may occur in industrial machinery into their causes. Therefore, it is desirable to be able to determine whether a detected abnormality is a known abnormality or not, and to present to the user not only known abnormalities but also actions to be taken in the case of unknown abnormalities. [Means for solving the problem]

[0007] The anomaly classification device according to one aspect of the present invention solves the above-mentioned problems by classifying data at the time of an anomaly based on anomaly cases that occurred in the past and presenting the classification results to a user.

[0008] One aspect of the present invention is an anomaly classification device for classifying anomalies that occur in industrial machinery, the anomaly classification device including an anomaly data acquisition unit that acquires data relating to a physical quantity detected when an anomaly occurs in the industrial machinery as anomaly data, an anomaly data storage unit that stores the anomaly data, and a calculation unit that uses the anomaly data stored in the anomaly data storage unit to determine whether the anomaly data is based on a known anomaly cause. 1. A model for classifying which abnormal data belongs to which anomaly cause A second model different from the first model Model and Each individually A learning unit that creates the learning data and the learning unit creates the learning data. 1. a known anomaly determination unit that uses a model to determine whether or not the abnormal data is based on a known anomaly cause; Second and an abnormal data classifying unit that uses the model to classify which abnormal data is due to which abnormality cause. Effect of the Invention

[0009] One aspect of the present invention makes it possible to classify abnormality patterns based on data obtained during an abnormality, without prior knowledge required for abnormality pattern classification. It also makes it possible to determine whether or not the abnormality data is based on a known cause of the abnormality, separately from the classification of the abnormality pattern, thereby making it possible to determine unknown abnormalities with a high degree of accuracy. [Brief description of the drawings]

[0010] [Figure 1] FIG. 2 is a schematic hardware configuration diagram of an anomaly classification device according to an embodiment. [Diagram 2] FIG. 1 is a schematic functional block diagram of an anomaly classification device according to a first embodiment. [Diagram 3] 13 is a display example of the classification results of the anomaly causes. [Figure 4] This is an example of an unknown abnormality cause. [Diagram 5] 13 is another display example of the classification results of the anomaly causes. [Figure 6] FIG. 11 is a schematic functional block diagram of an anomaly classification device according to a second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a schematic hardware configuration diagram showing the main part of an anomaly classification device according to an embodiment of the present invention. The anomaly classification device 1 according to the present invention can be implemented, for example, as a control device that controls industrial machines including machine tools and robots based on a control program, and can also be implemented on a computer such as a personal computer attached to a control device that controls industrial machines including machine tools and robots based on a control program, or a personal computer, cell computer, fog computer 6, or cloud server 7 connected to the control device via a wired / wireless network. In this embodiment, an example is shown in which the anomaly classification device 1 is implemented on a personal computer connected to the control device via a network.

[0012] The CPU 11 provided in the anomaly classification device 1 according to this embodiment is a processor that controls the entire anomaly classification device 1. The CPU 11 reads a system program stored in the ROM 12 via the bus 22, and controls the entire anomaly classification device 1 in accordance with the system program. The RAM 13 temporarily stores temporary calculation data, display data, various data input from outside, and the like.

[0013] The non-volatile memory 14 is composed of, for example, a memory backed up by a battery (not shown) or an SSD (Solid State Drive), and the memory state is maintained even when the power supply of the anomaly classification device 1 is turned off. The non-volatile memory 14 stores data read from an external device 72 via an interface 15, data input via an input device 71, data detected by a sensor 4 acquired from an industrial machine 3 via a network 5, and the like. The data stored in the non-volatile memory 14 may be expanded in the RAM 13 at the time of execution / use. In addition, various system programs such as a known analysis program are written in advance in the ROM 12.

[0014] The industrial machine 3 is fitted with sensors 4 that detect physical quantities such as current, voltage, temperature, vibration, and sound of various parts during operation of the industrial machine 3. Examples of the industrial machine 3 include machine tools and robots.

[0015] The interface 15 is an interface for connecting the CPU 11 of the anomaly classification device 1 to an external device 72 such as a USB device. For example, data related to the operation of each industrial machine can be read from the external device 72. In addition, programs and setting data edited in the anomaly classification device 1 can be stored in an external storage means via the external device 72.

[0016] The interface 20 is an interface for connecting the CPU of the anomaly classification device 1 to a wired or wireless network 5. The network 5 is connected to an industrial machine 3, a fog computer 6, a cloud server 7, etc., and exchanges data with the anomaly classification device 1.

[0017] The display device 70 displays various data loaded into memory, data obtained as a result of executing programs, data output from a machine learning device 100 (described later), and the like, which are output via an interface 17. An input device 71 consisting of a keyboard, a pointing device, and the like passes commands and data based on operations by an operator to the CPU 11 via an interface 18.

[0018] The interface 21 is an interface for connecting the CPU 11 and the machine learning device 100. The machine learning device 100 includes a processor 101 that controls the entire machine learning device 100, a ROM 102 that stores a system program and the like, a RAM 103 that performs temporary storage in each process related to machine learning, and a non-volatile memory 104 that is used to store models and the like. The machine learning device 100 can observe each piece of information that can be acquired by the anomaly classification device 1 (for example, data indicating the operating state of the industrial machine 3) via the interface 21. In addition, the anomaly classification device 1 acquires the processing results output from the machine learning device 100 via the interface 21, and stores or displays the acquired results, or transmits them to another device via the network 5 or the like.

[0019] FIG. 2 is a schematic block diagram showing functions of the anomaly classification device 1 according to the first embodiment of the present invention. The functions of the anomaly classification device 1 according to this embodiment are realized by the CPU 11 of the anomaly classification device 1 shown in FIG. 1 and the processor 101 of the machine learning device 100 executing a system program and controlling the operation of each part of the anomaly classification device 1 and the machine learning device 100.

[0020] The anomaly classification device 1 of this embodiment includes a data acquisition unit 110, an anomaly determination unit 120, an anomaly data acquisition unit 130, a label generation unit 140, and a classification result output unit 150. The machine learning device 100 included in the anomaly classification device 1 includes a learning unit 106, a known anomaly determination unit 107, and an anomaly data classification unit 108. In addition, the RAM 13 to the nonvolatile memory 14 of the anomaly classification device 1 are provided in advance with an acquired data storage unit 210 as an area for storing data acquired by the data acquisition unit 110 from the industrial machine 3 or the like, and an abnormality data storage unit 220 for storing data determined by the anomaly determination unit 120 to be data indicating an abnormal state as abnormal data, and the RAM 103 to the nonvolatile memory 104 of the machine learning device 100 are provided in advance with a model storage unit 109 as an area for storing a model created by machine learning by the learning unit 106.

[0021] The data acquisition unit 110 is realized by the CPU 11 included in the anomaly classification device 1 shown in FIG. 1 executing a system program read from the ROM 12, and mainly performing arithmetic processing by the CPU 11 using the RAM 13 and the non-volatile memory 14, and input control processing by the interface 15, 18, or 20. The data acquisition unit 110 acquires data related to physical quantities detected by the sensor 4 during operation of the industrial machine 3. The data acquisition unit 110 acquires data related to physical quantities such as values ​​of current and voltage flowing through each part during operation of the industrial machine 3, temperature (calorific value), vibration, and sound, which are detected by the sensor 4 attached to the industrial machine 3, for example. The data acquired by the data acquisition unit 110 may be instantaneous values ​​acquired at a predetermined timing, or may be time-series data acquired over a predetermined period of time. In addition, the data acquisition unit 110 may acquire data directly from the industrial machine 3 via the network 5, or may acquire data acquired and stored by an external device 72, the fog computer 6, the cloud server 7, or the like. The data acquired by the data acquisition unit 110 is stored in the acquired data storage unit 210.

[0022] The abnormality determination unit 120 is realized by the CPU 11 included in the abnormality classification device 1 shown in FIG. 1 executing a system program read from the ROM 12, and the CPU 11 mainly performing arithmetic processing using the RAM 13 and the non-volatile memory 14. The abnormality determination unit 120 determines the operation state of the industrial machine 3 based on data related to physical quantities detected during the operation of the industrial machine 3 acquired by the data acquisition unit 110. The abnormality determination unit 120 may determine that an abnormality has occurred when a value calculated based on data related to a predetermined physical quantity exceeds a predetermined threshold value, or may determine whether the operation of the industrial machine 3 is in a normal state or an abnormal state based on a predetermined model based on a result of statistical processing of the data related to the physical quantity, or may determine whether the operation of the industrial machine 3 is in a normal state or an abnormal state using a known machine learning method such as unsupervised learning or supervised learning. When using a machine learning technique, it is preferable to use a one-class classification technique such as One Class SVM, MT method, local outlier factor method, Auto Encoder, Variational Auto Encoder, etc. When using a machine learning technique, the abnormality determination unit 120 may be constructed on the machine learning device 100. The abnormality determination unit 120 outputs data related to a physical quantity determined to have been acquired when the operation of the industrial machine 3 was abnormal to the machine learning device 100.

[0023] The abnormality data acquisition unit 130 is realized by the CPU 11 included in the abnormality classification device 1 shown in FIG. 1 executing a system program read from the ROM 12 and performing calculation processing mainly by the CPU 11 using the RAM 13 and the non-volatile memory 14. The abnormality data acquisition unit 130 acquires, as abnormality data, data related to physical quantities determined by the abnormality determination unit 120 to have been acquired when the operation of the industrial machine 3 was abnormal, and stores the data in the abnormality data storage unit 220.

[0024] The learning unit 106 included in the machine learning device 100 is realized by the processor 101 included in the machine learning device 100 shown in FIG. 1 executing a system program read from the ROM 102, and the processor 101 mainly performing arithmetic processing using the RAM 103 and the non-volatile memory 104. The learning unit 106 creates a model to be used in a determination process related to abnormal data based on the abnormal data stored in the abnormal data storage unit 220, and stores the model in the model storage unit 109. When creating a model, the learning unit 106 uses abnormal data stored in the abnormal data storage unit 220 to which a label is attached regarding the cause of the abnormality. The model created by the learning unit 106 includes at least a model that can be used to determine whether or not the abnormal data is a known abnormal cause using the abnormal data as input, and a model that can be used to classify the abnormal data into which abnormal cause the data is based if the data is a known abnormal cause. In the case where a model for determining whether or not an anomaly cause is known and a model for classifying anomaly causes are created separately, for example, the model for determining whether or not an anomaly cause is known may be one class SVM, MT method, local outlier factor method, Auto Encoder, Variational Auto Encoder, etc. In addition, the model for classifying anomaly causes may be k-nearest neighbor method, linear discriminant analysis, neural network, etc. The parameters (hyperparameters, thresholds, etc.) of each model may be set by the user. The model used to determine whether the abnormal data is based on a known abnormality cause and the model used to classify the abnormal data based on which abnormality cause may be created as a common classification model. In this case, the classification model may be a classification model that receives abnormal data as input and outputs a confidence score indicating which class the abnormal data belongs to. For example, the value of the Softmax function of the output layer of the neural network may be output as the confidence score for the class classification result. When using such a model, if the confidence score output from the model is equal to or lower than a predetermined threshold value for any class (label of the abnormality cause), it can be determined that the abnormality is not based on a known abnormality cause. Also, if there is a class for which the confidence score output from the model is equal to or higher than a certain value, it can be determined that the abnormal data is classified into that class. The learning unit 106 stores the created model in the model storage unit 109 .

[0025] The known anomaly determination unit 107 included in the machine learning device 100 is realized by the processor 101 included in the machine learning device 100 shown in FIG. 1 executing a system program read from the ROM 102, and the processor 101 mainly performing arithmetic processing using the RAM 103 and the non-volatile memory 104. The known anomaly determination unit 107 determines whether or not an anomaly that has occurred in the industrial machine 3 is due to a known anomaly cause, based on data related to a physical quantity that the anomaly determination unit 120 has determined to have been acquired when the operation of the industrial machine 3 is abnormal. The known anomaly determination unit 107 uses a model stored in the model storage unit 109 to determine whether or not data determined to be abnormal by the anomaly determination unit 120 is due to a known anomaly cause. When the known anomaly determination unit 107 determines that an anomaly that has occurred in the industrial machine 3 is due to a known anomaly cause, it instructs the anomaly data classification unit 108 to classify the anomaly cause. Furthermore, if the known anomaly determination unit 107 determines that the anomaly is not due to a known anomaly cause, that is, if it determines that an unknown anomaly cause has occurred, it instructs the label generation unit 140 to assign a label.

[0026] The abnormal data classification unit 108 included in the machine learning device 100 is realized by the processor 101 included in the machine learning device 100 shown in FIG. 1 executing a system program read from the ROM 102, and the processor 101 mainly performing arithmetic processing using the RAM 103 and the non-volatile memory 104. The abnormal data classification unit 108 classifies and outputs the cause of abnormality for abnormal data in which the known abnormality determination unit 107 has determined that an abnormality occurring in the industrial machine 3 is due to a known abnormality cause. For example, when the model stored in the model storage unit 109 is a k-nearest neighbor method or the like, the abnormal data classification unit 108 classifies the abnormality cause based on which cluster the abnormality data is in the vicinity of. Also, for example, when the model stored in the model storage unit 109 is a neural network that outputs a score of the abnormality cause, the abnormality cause is classified based on the score calculated based on the abnormality data. The abnormal data classification unit 108 outputs the result of the classification of the abnormality cause to the classification result output unit 150.

[0027] 1 executes a system program read from ROM 12, and is realized by the CPU 11 mainly performing calculation processing using the RAM 13 and non-volatile memory 14 by the CPU 11, and input / output processing using the interfaces 17, 18, etc. The label generation unit 140 generates a label relating to the abnormality cause (a label that gives a meaning related to the operation and maintenance of the machine) for abnormality data that is determined by the known abnormality determination unit 107 to be not based on a known abnormality cause. For example, the label generating unit 140 may display on the display device 70 abnormality data that is determined to be not caused by a known abnormality cause, and generate a label related to the abnormality cause based on information related to the abnormality cause input by the user via the input device 71 for the displayed abnormality data, or may obtain alarm information that occurs after the abnormality data is obtained from the industrial machine 3, and generate a label related to the abnormality cause based on the obtained alarm information, or may comprehensively identify the abnormality cause and generate a label based on information obtained from other machines, information obtained from a higher-level computer such as the fog computer 6 or the cloud server 7, information related to the environment (environmental temperature and humidity, visual information and audio information obtained from an external sensor, etc.). The label generating unit 140 assigns the generated label to the abnormality data and stores it in the abnormality data storage unit 220.

[0028] In addition, the learning unit 106 may execute the re-learning process when the label generating unit 140 stores the abnormal data to which a new label related to the abnormality cause has been added in the abnormality data storage unit 220. For example, the learning unit 106 may execute the re-learning process when a predetermined number of abnormal data to which a label related to the abnormality cause has been added to the abnormality data storage unit 220 after the previous learning process has been executed to create a model. In addition, the learning unit 106 may execute the re-learning process when a predetermined number of abnormal data to which a label related to the same abnormality cause has been added to the abnormality data storage unit 220. With this configuration, the abnormality classification device 1 will eventually be able to classify the abnormality cause of abnormal data that was generated based on an unknown abnormality cause at the time of installation. Therefore, by continuing to use the anomaly classification device 1, it becomes possible to more appropriately support the user in dealing with failures.

[0029] The classification result output unit 150 is realized by the CPU 11 included in the anomaly classification device 1 shown in Fig. 1 executing a system program read from the ROM 12, and mainly performing calculation processing by the CPU 11 using the RAM 13 and the non-volatile memory 14, and output processing using the interfaces 17, 20, etc. The classification result output unit 150 outputs the classification result of the abnormal data by the abnormal data classification unit 108 to the display device 70 or a machine or device connected to the network 5. In addition, the classification result output unit 150 outputs information related to the abnormal data determined to be an unknown abnormality by the known anomaly determination unit 107 to the display device 70 or a machine or device connected to the network 5.

[0030] 3 is an example of the display of the classification results of abnormal data by the classification result output unit 150. In the example of FIG. 3, the degree of abnormality indicated by abnormal data detected during the operation of an industrial machine is displayed in a graph with the date and time on the horizontal axis. In this example, when specific abnormal data is selected, the classification results of the selected data are displayed. The classification results can be displayed in various ways, such as displaying only the single class with the highest certainty (301 in FIG. 3), displaying a list of the certainty for each class (302 in FIG. 3), or displaying a graph of the certainty (bar graph, pie chart, radar chart), etc.

[0031] FIG. 4 shows an example of a display when anomaly data that is determined by the known anomaly determination unit 107 to be not based on a known anomaly cause is selected. In this case, too, only the unknown anomaly with the highest certainty is displayed (303 in FIG. 4), and the certainty for each class is also displayed in a list ( Figure 4 304) may be used.

[0032] Fig. 5 is another display example of the classification result of abnormal data by the classification result output unit 150. As illustrated in Fig. 5, the classification result output unit 150 may display a list of the history of abnormality causes related to abnormalities that have occurred in a plurality of industrial machines 3.

[0033] The anomaly classification device 1 having the above configuration can classify anomaly patterns based on data at the time an anomaly occurs, even if there is no prior knowledge for classifying the cause of an anomaly based on the pattern at which the anomaly occurs. In addition, by determining whether an anomaly that has occurred in the industrial machinery 3 is due to a known anomaly cause (known anomaly determination) separately from the classification of anomaly patterns, it becomes possible to determine unknown anomalies with a high degree of accuracy.

[0034] FIG. 6 is a schematic block diagram showing functions of an anomaly classification device 1 according to the second embodiment of the present invention. The functions of the anomaly classification device 1 according to this embodiment are realized by the CPU 11 of the anomaly classification device 1 shown in FIG. 1 and the processor 101 of the machine learning device 100 executing a system program and controlling the operation of each part of the anomaly classification device 1 and the machine learning device 100.

[0035] The anomaly classification device 1 according to this embodiment has the same functions as the respective functions of the anomaly classification device 1 according to the first embodiment, except that the anomaly data acquisition unit 130 acquires the anomaly data acquired when an anomaly is detected in the industrial machine 3. In this way, the anomaly classification device 1 can be utilized to externally determine the occurrence of an anomaly and classify the anomaly data detected when an anomaly occurs. The anomaly classification device 1 has a function of classifying the cause of an anomaly for a known anomaly, and creating a label and learning when the anomaly is determined to be an unknown anomaly, thereby being able to fully provide the effects of the present invention.

[0036] Although one embodiment of the present invention has been described above, the present invention is not limited to the above-described embodiment, and can be embodied in various forms by making appropriate modifications. [Explanation of symbols]

[0037] 1 Anomaly classification device 3. Industrial machinery 4 Sensors 5. Network 6. Fog Computer 7. Cloud Server 11 CPU 12 ROM 13 RAM 14 Non-volatile memory 15,17,18,20,21 Interface 22 Bus 70 Display device 71 Input Devices 72 External equipment 110 Data Acquisition Section 120 Abnormality determination section 130 Abnormal data acquisition unit 140 Label Generation Unit 150 Classification result output section 210 Acquired data storage unit 220 Abnormal data storage unit 100 Machine Learning Machines 101 Processor 102 ROM 103 RAM 104 Non-volatile memory 106 Learning Department 107 Known Anomaly Judgment Unit 108 Abnormal Data Classification Section 109 Model Memory Section

Claims

1. An anomaly classification device that classifies anomalies occurring in an industrial machine, an abnormality data acquisition unit that acquires data relating to a physical quantity detected when an abnormality occurs in the industrial machine as abnormality data; an abnormality data storage unit that stores the abnormality data; a learning unit that uses the abnormal data stored in the abnormal data storage unit to individually create a first model for determining whether the abnormal data is based on a known abnormality cause, and a second model different from the first model for classifying the abnormal data as to which abnormality cause it belongs; a known anomaly determination unit that determines whether or not abnormality data is based on a known cause of anomaly using the first model created by the learning unit; an abnormal data classification unit that classifies abnormal data into which abnormality cause the abnormal data is based by using the second model created by the learning unit; An anomaly classification device comprising:

2. moreover, a data acquisition unit that acquires data related to a physical quantity detected in the industrial machine; an abnormality determination unit that determines whether an operation of the industrial machine is normal or abnormal based on the data acquired by the data acquisition unit, The abnormality data acquisition unit acquires data determined to be abnormal by the abnormality determination unit as abnormal data. The anomaly classification device according to claim 1 .

3. The method further includes a label generating unit that assigns a label related to an anomaly cause to the anomaly data that is determined by the known anomaly determining unit to be not based on a known anomaly cause. The anomaly classification device according to claim 1 .

4. a classification result output unit that combines the label added by the label generation unit with the classification result by the abnormal data classification unit and outputs the combined result, The anomaly classification device according to claim 3 .

5. the label generation unit acquires a label to be assigned to the abnormal data via a user interface; The anomaly classification device according to claim 3 .

6. the label generation unit generates a label to be assigned to the abnormality data based on any one of information on the machine to be diagnosed, information on other machines, and information on environmental conditions; The anomaly classification device according to claim 3 .

7. the learning unit re-learns the first model and the second model using abnormality data to which the label generating unit has assigned a label related to an abnormality cause; The anomaly classification device according to claim 3 .

Citation Information

Patent Citations

  • Method and apparatus for monitoring device state

    CN108199795A

  • Abnormality sign diagnostic device and abnormality sign diagnostic method

    JP2017033470A

  • Abnormality detection system, model generation device, abnormality detection device, abnormality detection method, model generation program, and abnormality detection program

    JP2017207904A

  • Failure prediction method, failure prediction device, and failure prediction program

    JP2019185422A

  • Asset Condition Monitoring

    US20150346066A1