WO2021053782A1 - Analysis device for event that can occur in production facility - Google Patents

Analysis device for event that can occur in production facility Download PDF

Info

Publication number
WO2021053782A1
WO2021053782A1 PCT/JP2019/036697 JP2019036697W WO2021053782A1 WO 2021053782 A1 WO2021053782 A1 WO 2021053782A1 JP 2019036697 W JP2019036697 W JP 2019036697W WO 2021053782 A1 WO2021053782 A1 WO 2021053782A1
Authority
WO
WIPO (PCT)
Prior art keywords
causal relationship
feature
production facility
feature amount
adopted
Prior art date
Application number
PCT/JP2019/036697
Other languages
French (fr)
Japanese (ja)
Inventor
一貴 西田
Original Assignee
オムロン株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by オムロン株式会社 filed Critical オムロン株式会社
Priority to PCT/JP2019/036697 priority Critical patent/WO2021053782A1/en
Publication of WO2021053782A1 publication Critical patent/WO2021053782A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B19/00—Programme-control systems
    • G05B19/02—Programme-control systems electric
    • G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS], computer integrated manufacturing [CIM]
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B23/00—Testing or monitoring of control systems or parts thereof
    • G05B23/02—Electric testing or monitoring
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Definitions

  • the present invention relates to an analysis device, an analysis method, and an analysis program for events that may occur in production equipment.
  • the causal element corresponds to, for example, a driving means such as a servomotor in a manufacturing facility and a monitoring means such as a sensor.
  • a driving means such as a servomotor in a manufacturing facility
  • a monitoring means such as a sensor.
  • a causal relationship model is constructed by representing it with an arrow. As a result, it is possible to quickly identify the cause when an event occurs.
  • the present invention has been made to solve the above problems, and is capable of constructing a causal relationship model capable of identifying the true cause of an event such as an abnormality even if there is a problem of multicollinearity. It is an object of the present invention to provide an analysis method and an analysis program.
  • the first analyzer is a production facility for producing a product, and has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production.
  • a driving means and a monitoring means are provided in a production facility having at least one controllable feature amount, and are analysis devices for events that can occur in the production facility, and include a control unit and a storage unit.
  • the storage unit stores a plurality of the feature amounts, and the control unit has a step of calculating a correlation between the plurality of feature amounts with respect to the event, and any one of the correlations.
  • the step of constructing the causal relationship that is, the step of applying the causal relationship related to the one feature amount to the causal relationship of the other feature amount, is executed.
  • the causal relationship model was constructed after removing highly correlated features. If the mechanism corresponding to the feature amount is the root cause of the abnormality, a causal relationship model that does not include this may be generated, and there is a problem that the root cause of the abnormality cannot be identified.
  • the causal relationship model is constructed.
  • the causal relationship model is reconstructed by incorporating the mechanism related to the features that were not used. Therefore, it is possible to construct a causal relationship model including a mechanism corresponding to all related features. As a result, the true cause of the abnormality can be reliably identified.
  • the correlation is calculated as 1 or more. For example, there is one correlation between two features, three correlations between three features, and six correlations between four features. When any one of the one or more correlations calculated in this way shows a strong correlation satisfying a predetermined requirement, the above processing is performed.
  • the second analyzer is a production facility for producing a product, and has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production.
  • a driving means and a monitoring means are provided in a production facility having at least one controllable feature amount, and are analysis devices for events that can occur in the production facility, and include a control unit and a storage unit.
  • the storage unit stores a plurality of the feature quantities
  • the control unit includes a step of calculating a correlation between the plurality of feature quantities for the event and any one of the correlations.
  • a step of generating a group of feature quantities so as to adopt one feature quantity having the strong correlation and not adopt the other feature quantity, and the group A step of constructing a causal relationship model showing a causal relationship between a plurality of the included feature amounts and a step of writing the other feature amount together with the one feature amount in the causal relationship model are executed. ..
  • the same effect as that of the first analyzer can be obtained. That is, in the second analysis apparatus according to the present invention, one of the highly correlated features is not adopted, and then the causal relationship model is constructed. After that, the causal relationship model is adopted for the constructed causal relationship model. The causal relationship model is reconstructed by adding the mechanism related to the features that were not used. Therefore, it is possible to construct a causal relationship model including a mechanism corresponding to all related features. As a result, the true cause of the abnormality can be reliably identified.
  • the third analyzer is a production facility for producing a product, and has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production.
  • a driving means and a monitoring means are provided in a production facility having at least one controllable feature amount, and are analysis devices for events that can occur in the production facility, and include a control unit and a storage unit.
  • the storage unit stores a plurality of the feature amounts, and the control unit has a step of calculating a correlation between the plurality of feature amounts with respect to the event, and any one of the correlations.
  • the two feature quantities having the strong correlation are combined to form a composite feature quantity, which is a causal relationship between the step and the plurality of the feature quantities and the composite feature quantity.
  • the causal relationship related to the synthetic feature amount are applied to the two said feature amounts constituting the synthetic feature amount, and the causal relationship between all the feature amounts obtained by deleting the synthetic feature amount.
  • the same effect as that of the first analyzer can be obtained. That is, in the second analyzer according to the present invention, a synthetic feature amount is generated by combining highly correlated features, and a causal relationship model including the synthetic feature amount is constructed. After that, the causal relationship constructed is constructed. After deleting the synthetic features from the model, the causal relationship model is reconstructed using all the related features including the features used for the synthetic features. Therefore, it is possible to construct a causal relationship model including a mechanism corresponding to all related features. As a result, the true cause of the abnormality can be reliably identified.
  • the first analysis device can further include a step of forming a causal relationship between the one feature amount and the other feature amount.
  • the third analysis device can further include a step of forming a causal relationship between the two feature quantities constituting the synthetic feature quantity.
  • the first analysis method is a production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production.
  • a method of analyzing an event that can occur in a production facility and the driving means and the monitoring means are provided in the production facility and have at least one controllable feature amount. If the step of calculating the correlation between a plurality of feature quantities and any one of the above correlations show a strong correlation that satisfies a predetermined requirement, one of the said feature quantities having the strong correlation is adopted. Then, a causal relationship model showing a causal relationship between the step of generating a group including the plurality of feature amounts and the plurality of the feature amounts included in the group is constructed so as not to adopt the other feature amount. This step is a step of constructing a causal relationship between all the feature quantities including the other feature quantity that has not been adopted, and the causal relationship related to the one feature quantity is set to the other feature quantity. It has steps to apply to the causal relationship of.
  • the second analysis method is a production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production.
  • the driving means and the monitoring means are provided in the production facility and have at least one controllable feature amount, and are an analysis method of an event that can occur in the production facility, wherein the plurality of feature amounts for the event are If the step of calculating the correlation between the above and any one of the above correlations shows a strong correlation that satisfies a predetermined requirement, one feature quantity having the strong correlation is adopted and the other feature is adopted.
  • the third analysis method is a production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production.
  • the step includes a step of constructing a causal relationship model showing a causal relationship between all the feature amounts in which the synthetic feature amount is deleted.
  • the first analysis program according to the present invention is a production facility that produces a product, and has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production.
  • a driving means and a monitoring means are provided in a production facility having at least one controllable feature quantity, and are an analysis program of an event that can occur in the production facility, and can occur in the computer and the production facility.
  • the step of constructing a model and the step of constructing a causal relationship between all the feature quantities including the other feature quantity that has not been adopted, and the causal relationship relating to the one feature quantity is the other.
  • the step of applying to the causal relationship of the feature amount of is executed.
  • the second analysis program according to the present invention is a production facility that produces a product, and has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production.
  • a drive means and a monitoring means are provided in a production facility having at least one controllable feature amount, and are an analysis program of an event that can occur in the production facility. If one of the steps of calculating the correlation between the feature quantities of the above and one of the above correlations shows a strong correlation satisfying a predetermined requirement, one of the feature quantities having the strong correlation is adopted.
  • a step of generating a group of feature quantities so as not to adopt the other feature quantity a step of constructing a causal relationship model showing a causal relationship between a plurality of the feature quantities included in the group, and the causal relationship. In the model, the step of writing the other feature amount together with the one feature amount is executed.
  • the third analysis program according to the present invention is a production facility that produces a product, and has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production.
  • a drive means and a monitoring means are provided in a production facility having at least one controllable feature amount, and are an analysis program of an event that can occur in the production facility. If one of the steps of calculating the correlation between the feature quantities of the above and one of the above correlations shows a strong correlation satisfying a predetermined requirement, the two feature quantities having the strong correlation are combined and synthesized.
  • a step of constructing a causal relationship model showing a causal relationship between all the said feature quantities in which the synthetic feature quantity is deleted while applying the quantity is executed.
  • FIG. 1 An example of a situation in which the present invention is applied is schematically illustrated. It is a block diagram which shows the hardware structure of the analysis apparatus which concerns on one Embodiment of this invention. It is the schematic of the production equipment which concerns on one Embodiment of this invention. It is the figure which superposed the node of the causal relation model on the schematic diagram of a packaging machine. It is a block diagram which shows the functional structure of an analyzer. It is a flowchart which shows an example of construction of a causal relational model. It is a figure which shows the selection method of a feature amount after calculating the correlation. It is a flowchart which shows an example of construction of a causal relational model. This is an example of a causal relationship model. This is an example of a causal relationship model.
  • the present embodiment an embodiment according to one aspect of the present invention (hereinafter, also referred to as “the present embodiment”) will be described with reference to the drawings.
  • the embodiments described below are merely examples of the present invention in all respects. Needless to say, various improvements and modifications can be made without departing from the scope of the present invention. That is, in carrying out the present invention, a specific configuration according to the embodiment may be appropriately adopted.
  • the data appearing in the present embodiment is described in natural language, more specifically, it is specified in a pseudo language, a command, a parameter, a machine language, etc. that can be recognized by a computer.
  • FIG. 1 schematically illustrates an example of an application scene of the production system according to the present embodiment.
  • the production system according to the present embodiment includes a packaging machine 3 which is an example of production equipment, and an analysis device 1.
  • the analyzer 1 is configured to build a causal relationship model between a servomotor (driving means, monitoring means (for example, encoder of a servomotor, etc.)) and various sensors (monitoring means) provided in the packaging machine 3. It is a computer.
  • the driving means such as a servomotor, the monitoring means, and the monitoring means such as various sensors are collectively referred to as the mechanism 21.
  • the analysis device 1 constructs a causal relationship model between the mechanisms 21 for various events occurring in the packaging machine 3, for example, an abnormality. Therefore, first, a plurality of measurement data regarding the states of the plurality of mechanisms 21 constituting the packaging machine 3 are acquired.
  • the analysis device 1 calculates the feature amount from the acquired plurality of measurement data 124.
  • the feature amount is a numerical value such as an amplitude, a maximum value, a minimum value, and an average value of the output value (torque, speed, etc.) of each mechanism 21.
  • the correlation of the calculated features is calculated. In the example of FIG. 1, it is assumed that the feature amounts A to C are calculated and it is determined that the feature amounts A and B have a high correlation among the feature amounts A to C.
  • the causal relationship model between the mechanisms 21 corresponding to the features A and C is constructed.
  • a causal relationship model having an edge facing the feature amount A from the feature amount C is constructed.
  • the causal relationship model shows the relationship between the mechanisms 21 corresponding to the feature amounts, but here, for convenience of explanation, the feature amounts A to C are shown as the mechanisms corresponding to these. .. This point is the same in the description of the present specification.
  • the causal relationship model incorporating the feature quantity B that has not been adopted is reconstructed. That is, since the features A and B have a high correlation, they are regarded as equivalent and incorporated into the causal relationship model as nodes to which the same edges are connected.
  • the causal relational model incorporating the feature amount C is reconstructed together with the edge from the feature amount C to the feature amount B. From the above, it is possible to construct a causal relationship model that includes all features (mechanisms) even if the causal relationship model has a problem of multicollinearity.
  • each mechanism may be, for example, a conveyor, a robot arm, a servomotor, a cylinder (molding machine or the like), a suction pad, a cutter device, a sealing device, or the like.
  • each mechanism may be a composite device such as a printing machine, a mounting machine, a reflow furnace, or a substrate inspection device.
  • each mechanism includes, for example, a device that detects some information by various sensors, a device that acquires data from various sensors, and some device from the acquired data, in addition to the device that involves some physical operation as described above. It may include a device that performs internal processing such as a device that detects information and a device that processes acquired data.
  • One mechanism may be composed of one or a plurality of devices, or may be composed of a part of the devices.
  • One device may be configured by a plurality of mechanisms.
  • each may be regarded as a different mechanism. For example, when the same device executes the first process and the second process, the device that executes the first process is regarded as the first mechanism, and the device that executes the second process is the first. It may be regarded as the mechanism of 2.
  • FIG. 2 is a block diagram showing an example of the hardware configuration of the analysis device 1 according to the present embodiment
  • FIG. 3 is a diagram showing a schematic configuration of the packaging machine.
  • FIG. 2 schematically illustrates an example of the hardware configuration of the analysis device 1 according to the present embodiment.
  • the analysis device 1 is a computer to which the control unit 11, the storage unit 12, the communication interface 13, the input device 14, the display device 15, and the drive 16 are electrically connected. ..
  • the communication interface is described as "communication I / F".
  • the control unit 11 includes a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), and the like, which are hardware processors, and is configured to execute information processing based on a program and various data.
  • the storage unit 12 is an example of a memory, and is composed of, for example, an auxiliary storage device such as a hard disk drive or a solid state drive. In the present embodiment, the storage unit 12 stores various information such as the analysis program 121, the variable setting data 122, the causal relationship model data 123, and the measurement data 124.
  • the analysis program 81 is a program for causing the analysis device 1 to execute information processing (FIGS. 6 and 8 described later) relating to the derivation of the causal relationship between the plurality of mechanisms 21.
  • the analysis program 121 includes a series of instructions for this information processing.
  • the variable setting data 122 is data related to the combination of highly correlated feature quantities
  • the causal relationship model data 123 is data related to the generated causal relationship model, which is a causal relationship model before reconstruction, which will be described later.
  • the data includes both the causal relationship model after reconstruction and the reconstruction.
  • the measurement data 124 is data relating to the states of the plurality of mechanisms 21 constituting the packaging machine 3.
  • the communication interface 13 is, for example, a wired LAN (Local Area Network) module, a wireless LAN module, or the like, and is an interface for performing wired or wireless communication via a network.
  • the analysis device 1 uses this communication interface 13 to perform data communication via a network with another information processing device such as a control device (not shown) configured to control the operation of the packaging machine 3, for example. This can be performed to acquire a plurality of measurement data 124.
  • the type of network may be appropriately selected from, for example, the Internet, a wireless communication network, a mobile communication network, a telephone network, a dedicated network, and the like. However, the route for acquiring the measurement data 124 does not have to be limited to such an example.
  • the input device 14 is, for example, a device for inputting a mouse, a keyboard, or the like.
  • the display device 15 is an example of an output device, for example, a display. The operator can operate the analysis device 1 via the input device 14 and the display device 15.
  • the display device 15 may be a touch panel display. In this case, the input device 14 may be omitted.
  • the drive 16 is, for example, a CD drive, a DVD drive, or the like, and is a drive device for reading a program stored in the storage medium 91.
  • the type of the drive 16 may be appropriately selected according to the type of the storage medium 91.
  • At least one of the analysis program 121, the variable setting data 122, the causal relationship model data 123, and the plurality of measurement data 124 may be stored in the storage medium 91.
  • the storage medium 91 stores the information of the program or the like by electrical, magnetic, optical, mechanical or chemical action so that the information of the program or the like recorded by the computer or other device, the machine or the like can be read. It is a medium to do.
  • the analysis device 1 may acquire at least one of the analysis program 81 and a plurality of measurement data 124 from the storage medium 91.
  • FIG. 2 illustrates a disc-type storage medium such as a CD or DVD as an example of the storage medium 91.
  • the type of the storage medium 91 is not limited to the disk type, and may be other than the disk type.
  • Examples of storage media other than the disk type include semiconductor memories such as flash memories.
  • the control unit 11 may include a plurality of hardware processors.
  • the hardware processor may be composed of a microprocessor, an FPGA (field-programmable gate array), a DSP (digital signal processor), or the like.
  • the storage unit 12 may be composed of a RAM and a ROM included in the control unit 11. At least one of the communication interface 13, the input device 14, the display device 15, and the drive 16 may be omitted.
  • the analysis device 1 may further include an output device other than the display device 15 such as a speaker.
  • the analysis device 1 may be composed of a plurality of computers. In this case, the hardware configurations of the computers may or may not match.
  • the analysis device 1 may be a general-purpose information processing device such as a desktop PC (Personal Computer) or a tablet PC, a general-purpose server device, or the like, in addition to an information processing device designed exclusively for the provided service. Further, the analysis device 1 may be configured to be able to control the operation of the packaging machine 3. In this case, the analysis device 1 may be a PLC (programmable logic controller). Further, the analysis device 1 may be provided with an input / output interface for connecting to the packaging machine 3, and measurement data 124 may be acquired via this input / output interface.
  • a PLC programmable logic controller
  • FIG. 3 schematically illustrates an example of the hardware configuration of the packaging machine 3 according to the present embodiment.
  • the wrapping machine 3 is a so-called horizontal pillow wrapping machine, which is a device for wrapping contents WA such as food (dried noodles, etc.) and stationery (eraser, etc.).
  • the type of the content WA can be appropriately selected according to the embodiment, and is not particularly limited.
  • the wrapping machine 3 broadcasts the film roll 30 on which the wrapping film is wound, the film transport unit 31 for transporting the wrapping film, the content transport unit 32 for transporting the content WA, and the content on the wrapping film. It is provided with a bag making portion 33.
  • the packaging film can be, for example, a resin film such as a polyethylene film.
  • the film roll 30 includes a winding core, and the packaging film is wound around the winding core.
  • the winding core is rotatably supported around the axis, whereby the film roll 30 is configured so that the packaging film can be unwound while rotating.
  • the film transport unit 31 includes a drive roller driven by a servomotor (servo 1) 311, a passive roller 312 to which a rotational force is applied from the drive roller, and a plurality of pulleys 313 that guide the packaging film while applying tension. , Is equipped.
  • the film transport section 31 is configured to feed the packaging film from the film roll 30 and transport the delivered packaging film to the bag making section 33 without loosening it.
  • the content transfer unit 32 includes a conveyor 321 that conveys the content WA to be packaged, and a servomotor (servo 2) 322 that drives the conveyor 321.
  • the content transporting section 32 is connected to the bag making section 33 via the lower part of the film transporting section 31.
  • the content WA transported by the content transport unit 32 is supplied to the bag making unit 33 and packaged by the packaging film supplied from the film transport unit 31.
  • a fiber sensor (sensor 1) 324 for detecting the position of the content WA is provided in the information downstream of the conveyor 321.
  • a fiber sensor (sensor 2) 325 for detecting the riding of the contents WA and the like.
  • the bag making section 33 cuts the conveyor 331, the servo motor (servo 3) 332 that drives the conveyor 331, the center seal section 333 that seals the packaging film in the transport direction, and the packaging film on both ends in the transport direction. It includes an end seal portion 334 that seals at each end portion.
  • the conveyor 331 conveys the content WA conveyed from the content transfer unit 32 and the packaging film supplied from the film transfer unit 31.
  • the packaging film supplied from the film transport unit 31 is supplied to the center seal unit 333 while being appropriately bent so that both side edge portions in the width direction overlap each other.
  • the center seal portion 333 is composed of, for example, a pair of left and right heating rollers (heaters 1 and 2), and seals both side edge portions of the bent packaging film along the transport direction by heating.
  • the packaging film is formed in a tubular shape.
  • the content WA is put into the packaging film formed in this tubular shape.
  • a fiber sensor (sensor 3) 336 that detects the position of the content WA is provided above the conveyor 331.
  • the end seal portion 334 includes, for example, a roller driven by a servomotor (servo 4) 335, a pair of cutters that open and close by rotation of the rollers, and heaters (heaters 3) provided on both sides of each cutter.
  • a servomotor servo 4
  • the end seal portion 334 is configured so that the tubular packaging film can be cut in the direction orthogonal to the transport direction and can be sealed by heating at the cut portion.
  • the tip portion of the packaging film formed in a tubular shape is sealed on both sides in the transport direction and separated from the subsequent portion to become a packaging body WB containing the contents WA.
  • the above-mentioned packaging machine 3 can wrap the contents WA in the following steps. That is, the film transport unit 31 feeds out the packaging film from the film roll 30. In addition, the content transport unit 32 transports the content WA to be packaged. Next, the center seal portion 333 of the bag making portion 33 forms the unwound packaging film into a tubular shape. Then, after the content WA is put into the formed tubular packaging film, the tubular packaging film is cut in the direction orthogonal to the transport direction by the end seal portion 334, and both sides of the cut portion in the transport direction are cut. Seal by heating with. As a result, a horizontal pillow type package WB containing the content WA is formed. That is, the packaging of the contents WA is completed.
  • the drive of the packaging machine 3 can be controlled by a PLC or the like provided separately from the packaging machine 3.
  • the above-mentioned measurement data 124 can be acquired from the PLC.
  • 10 mechanisms are set in order to establish a causal relationship of abnormalities. That is, the above-mentioned servos 1 to 4, heaters 1 to 3, and sensors 1 to 3 are set as mechanisms, and a causal relationship between these mechanisms when an abnormality occurs is constructed as a causal relationship model. Details will be described later.
  • FIG. 5 schematically illustrates an example of the software configuration of the analysis device 1 according to the present embodiment.
  • the control unit 11 of the analysis device 1 expands the analysis program 81 stored in the storage unit 12 into the RAM. Then, the control unit 11 interprets and executes the analysis program 81 expanded in the RAM by the CPU to control each component.
  • the analysis device 1 includes a data acquisition unit 111, a variable setting unit 112, a model construction unit 113, a model reconstruction unit 114, and an output unit 115 as software modules. Operates as a computer. That is, in the present embodiment, each software module of the analysis device 1 is realized by the control unit 11 (CPU).
  • the data acquisition unit 111 acquires a plurality of measurement data 124 relating to the states of the plurality of mechanisms 21 constituting the packaging machine 3, and calculates the feature amount of each mechanism 21 from the measurement data 124.
  • the variable setting unit 112 calculates the correlation between the calculated feature amounts, and stores the combination of the feature amounts with high correlation in the storage unit 12. Further, the variable setting unit 112 determines the feature amount to be adopted in the construction of the causal relationship model from the combination of the feature amount having high correlation. Details will be described later.
  • the model building unit 113 builds a causal relationship model between a plurality of mechanisms 2 by statistically analyzing the above-mentioned features.
  • the model reconstruction unit 114 reconstructs the causal relationship model by incorporating the features that were not adopted in the construction of the causal relationship model among the combinations of the above-mentioned highly correlated features into the generated causal relationship model. ..
  • the output unit 115 outputs data related to the reconstructed causal relationship model to a display device or the like.
  • each software module of the analysis device 1 will be described in detail in an operation example described later.
  • an example in which each software module of the analysis device 1 is realized by a general-purpose CPU is described.
  • some or all of the above software modules may be implemented by one or more dedicated hardware processors.
  • software modules may be omitted, replaced, or added as appropriate according to the embodiment.
  • FIG. 6 illustrates an example of the processing procedure of the analysis device 1 according to the present embodiment.
  • the processing procedure of the analysis device 1 described below is an example of the "analysis method" of the present invention.
  • the processing procedure described below is only an example, and each processing may be changed as much as possible. Further, with respect to the processing procedure described below, steps can be omitted, replaced, and added as appropriate according to the embodiment.
  • step S101 the control unit 11 operates as the data acquisition unit 111 and acquires a plurality of measurement data 124 relating to the states of the plurality of mechanisms 21 constituting the packaging machine 3.
  • the control unit 11 acquires a plurality of measurement data 124 via a network from a control device (not shown) configured to control the operation of the packaging machine 3 by using the communication interface 13.
  • the route for acquiring the measurement data 124 does not have to be limited to such an example.
  • a plurality of measurement data 124s may be stored in an external storage device such as NAS (Network Attached Storage) or another information processing device.
  • the control unit 11 may acquire a plurality of measurement data 124s from the external storage device or another information processing device via the network, the storage medium 91, or the like.
  • the analysis device 1 may be configured to directly control the operation of the packaging machine 3.
  • the control unit 11 may acquire the measurement data 124 of each case directly from the sensor that observes the state of each mechanism 21 constituting the packaging machine 3.
  • the measurement data 124 of each case may include all kinds of data regarding the state of each mechanism 21 constituting the packaging machine 3.
  • the measurement data 124 of each case is, for example, data indicating at least one of torque, velocity, acceleration, temperature, current, voltage, pneumatic pressure, pressure, flow rate, position, dimension (height, length, width) and area. It may be.
  • Such measurement data 124 can be obtained by a measuring device such as a known sensor or camera.
  • the flow rate can be obtained by a float sensor.
  • the position, dimensions, and area can be obtained by an image sensor.
  • the measurement data 124 of each case may be composed of data obtained from one or a plurality of measurement devices. Further, the measurement data 124 of each case may be the data obtained from the measurement device as it is, or some information processing is applied to the data obtained from the measurement device such as the position data calculated from the image data. It may be the data acquired by the above. The measurement data 124 of each case is acquired corresponding to each mechanism 21.
  • the control unit 11 calculates one or a plurality of feature quantities from the acquired measurement data 124. In the present embodiment, the control unit 11 calculates a plurality of feature quantities from the measurement data 124 of each case.
  • the type of the feature amount does not have to be particularly limited and may be appropriately selected according to the embodiment.
  • the calculated feature amount is, for example, the amplitude in the frame, the maximum value, the minimum value, the average value, the variance value, the standard deviation, the instantaneous value (one-point sample), or the like. You can.
  • the calculated feature amount is, for example, the "on” time, the “off” time, the duty ratio, the "on” number of times, and the “off” number of times in each frame. And so on.
  • the number of feature quantities to be calculated may not be particularly limited, and may be appropriately selected depending on the embodiment.
  • the number of feature quantities calculated from the measurement data 124 of each case may be the same or different.
  • the control unit 11 proceeds to the next step S102.
  • step S102 the control unit 11 operates as the variable setting unit 112, and first calculates the correlation between the feature quantities. That is, the control unit 11 calculates the correlation between each feature amount calculated from the measurement data of one of the plurality of measurement data 124 and each feature amount calculated from the other measurement data.
  • a correlation coefficient can be used. For example, when there are five feature quantities A to E, all the correlation coefficients between the five feature quantities A to E (10 correlation coefficients in total) are obtained, and it is determined that the correlation is high.
  • the combination of feature quantities is stored in the storage unit 12 as variable setting data 122.
  • the determination that the correlation is high can be, for example, a correlation coefficient of 0.6 or more, preferably 0.7 or more, more preferably 0.8 or more, and particularly preferably 0.9 or more. The determination is not limited, and can be determined by various known methods.
  • step S103 the control unit 11 operates as the model construction unit 113, and constructs a causal relationship model between the plurality of mechanisms 21 by statistically analyzing the above-mentioned features.
  • the method for statistically analyzing a plurality of feature quantities does not have to be particularly limited, and may be appropriately selected according to the embodiment.
  • Examples of the analysis method include GLASSO (Graphical LASSO), covariance selection method, SGS (Spirtes, Glymour, and Scenes), GM (Graphical Modeling), PC (Peter & Clark), GES (Greedy Equivalent Search), and FCI ( FastCausalInference), LiNGAM (LinerNon-GaussianAcyclicModel), Bayesian network, etc. may be used.
  • the control unit 11 statistically analyzes a plurality of feature quantities by the following processing procedure.
  • FIG. 8 illustrates an example of the processing procedure of the causal relationship analysis by the analysis device 1 according to the present embodiment.
  • the process of step S103 according to the present embodiment includes the following processes of steps S201 to S202.
  • the processing procedure described below is only an example, and each processing may be changed as much as possible. Further, with respect to the processing procedure described below, steps can be omitted, replaced, and added as appropriate according to the embodiment.
  • step S201 the control unit 11 calculates the conditional independence between the calculated features.
  • the control unit 11 has conditional independence between each feature amount calculated from the measurement data of one of the plurality of measurement data 124 and each feature amount calculated from the other measurement data. Is calculated.
  • the type of conditional independence does not have to be particularly limited and may be appropriately selected depending on the embodiment.
  • the calculated conditional independence may be, for example, a partial correlation coefficient, a correlation coefficient, a covariance, a conditional probability, an accuracy matrix, or the like.
  • the control unit 11 may calculate one type of conditional independence or may calculate a plurality of types of conditional independence. After calculating the conditional independence between the features, the control unit 11 proceeds to the next step S202.
  • step S202 the control unit 11 determines whether or not there is a causal relationship between the mechanisms 21 based on the conditional independence calculated for each. As an example, the control unit 11 determines the presence or absence of a causal relationship by comparing the calculated conditional independence with the threshold value. When the value of conditional independence is proportional to the degree of having a causal relationship, the control unit 11 determines whether or not the calculated value of conditional independence is equal to or greater than the threshold value. Then, the control unit 11 determines that there is a causal relationship between the corresponding mechanisms 21 when the calculated conditional independence value is equal to or greater than the threshold value, and when it is not, the control unit 11 determines that there is a causal relationship between the corresponding mechanisms 21. Is judged to have no causal relationship.
  • the threshold value may be set as appropriate.
  • the method of determining the presence or absence of a causal relationship based on conditional independence does not have to be limited to such an example, and may be appropriately determined according to the type of conditional independence and the like.
  • the correspondence between the degree of causality and conditional independence may be opposite.
  • the control unit 11 determines that there is a causal relationship between the corresponding mechanisms 21 when the value of the conditional independence is equal to or less than the threshold value, and when not, the control unit 11 determines that there is a causal relationship between the corresponding mechanisms 21. It may be determined that there is no causal relationship. Further, the control unit 11 may determine the presence or absence of a causal relationship between the mechanisms 21 based on a plurality of types of conditional independence.
  • the control unit 11 can build a causal relationship model between each mechanism 21. For example, in the case of the above-mentioned packaging machine 3, graphs of a causal relationship model as shown in FIGS. 9A to 9C are constructed.
  • FIG. 9A shows a causal relationship model when the leather belt is worn as an abnormality. That is, the torque mean value and the standard deviation of the position, which are the feature amounts of the servo 1 (servomotor 311), affect the speed mean value and the torque maximum value, which are the feature amounts of the servo 2 (servomotor 322).
  • a causal relationship model is constructed in which the torque of the servo 4 (servo motor 335) is affected by the torque average value.
  • FIG. 9B shows a causal relationship model when the chain of the conveyor 321 of the content transporting unit 32 becomes loose as an abnormality. That is, the ON time, which is the feature amount of the sensor 2 (fiber sensor 325), affects the turn ON time, which is the feature amount of the sensor 3 (fiber sensor 336), which further affects the torque average value of the servo 4. , A causal relationship model is constructed.
  • FIG. 9C shows a causal relationship model when a sealing defect of the packaging film occurs as an abnormality.
  • a causal relationship model is constructed in which only the average torque value of the servo 4 is the cause of this abnormality, and the causal relationship model data 123 is stored in the storage unit 12.
  • step S104 the control unit 11 operates as the model reconstruction unit 114, incorporates the features not adopted in the construction of the causal relational model into the generated causal relational model, and incorporates the causal relational model into the generated causal relational model. Rebuild.
  • the feature quantity B is adopted.
  • a causal relationship model is constructed for the four feature quantities A, B, D, and E (A, B, D, and E correspond to the group according to the present invention).
  • the feature quantity C is incorporated into the causal relational model. That is, since the feature amount C has a high correlation with the feature amount B, it is considered to be equivalent to the feature amount B, and the feature amount C having an edge similar to that of the feature amount B is incorporated into the causal relationship model.
  • the causal relationship model is reconstructed by incorporating the feature amount C and providing an edge from the feature amount C toward the feature amount A in the same manner as the feature amount B.
  • the causal relationship model reconstructed in this way is stored in the storage unit 12 as the causal relationship model data 123.
  • FIG. 11A when a causal relationship such as A ⁇ B ⁇ D ⁇ E is established and the feature amount C that has been rejected is incorporated, the same as the feature amount B. Is provided with an edge from the feature amount A to the feature amount C, and from the feature amount C to the feature amount D.
  • FIG. 11B As a modification of FIG. 10, an edge connecting the feature amounts B and C having high correlation can be provided.
  • step S105 the control unit 11 operates as the output unit 115 and outputs causal relationship information indicating the reconstructed causal relationship model.
  • the output destination and expression format of the causal relationship information may not be particularly limited, and may be appropriately selected according to the embodiment.
  • the display device 15 is adopted as the output destination of the causal relationship information.
  • the control unit 11 causes the display device 15 to display the causal relationship information as an output process.
  • as the expression format of the causal relationship information for example, graphs as shown in FIGS. 9A to 9C are adopted. That is, in the present embodiment, the output process of step S104 includes generating a graph representing the constructed causal relationship model, and outputting the generated graph as causal relationship information.
  • the output process includes switching the display form of the causal relationship information between the two forms. That is, in the present embodiment, the control unit 11 is specified by using each mechanism 21 as an item and using the first form expressing the specified causal relationship model and each feature amount as an item.
  • the causal relationship information is output by switching the second form that expresses the causal relationship model.
  • both the causal relationship model before reconstruction and the causal relationship model after reconstruction, or these can be switched and displayed.
  • the causal relationship model was constructed after removing highly correlated features. If the mechanism 21 corresponding to the feature amount is the root cause of the abnormality, a causal relationship model that does not include the mechanism 21 may be generated, and there is a problem that the root cause of the abnormality cannot be identified.
  • one of the highly correlated features is removed and then the causal relationship model is constructed. After that, the constructed causal relationship model is removed.
  • the causal relationship model is being reconstructed by incorporating the mechanism related to the features. Therefore, it is possible to construct a causal relationship model including a mechanism corresponding to all related features. As a result, the true cause of the abnormality can be reliably identified.
  • variable setting unit 112 constructs a causal relationship model by adopting one of the highly correlated features and not adopting the other, but the variable setting unit 112 performs other processing. It can be performed.
  • feature quantities B and C with high correlation are combined to generate feature quantities X (corresponding to the synthetic feature quantities according to the present invention), and then feature quantities A, X, D, and E are generated. It can be used to build a causal relationship model.
  • the feature amount X can be generated, for example, as follows. Since the feature quantities A to E are time-series data, the feature quantities X can be generated by extracting the data in the same time zone for the feature quantities B and C and simply adding them. In addition, for example, it can be generated by linearly combining feature quantities B and C.
  • the feature amount C is also written together with the feature amount B in the causal relationship model. Keep it.
  • the feature amount C can be displayed adjacent to the feature amount B of the output causal relational model. That is, if the causal relationship model is drawn or stored as data so that it is clearly shown that the feature amount B and the feature amount C are equivalent, it is "combined writing".
  • a causal relationship model is built when an abnormality occurs, but it is also possible to build a causal relationship model for fluctuations and displacements that occur in the normal range when it is not abnormal. Such fluctuations, displacements, and anomalies correspond to the events of the present invention.

Abstract

A first analysis device according to the present invention is provided to a production facility that is for production of products and that has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production, the driving means and the monitoring means having one or more controllable feature amounts, the first analysis device being for an event that can occur in the production facility. The first analysis device is provided with a control unit and a storage unit, wherein the storage unit stores a plurality of the feature amounts and the control unit causes execution of: a step for calculating correlations with respect to the event between the plurality of feature amounts; a step for, when one of the correlations indicates a strong correlation satisfying a predetermined requirement, generating a group including the plurality of feature amounts such that one of the feature amounts having a strong correlation is employed and the other feature amounts are not employed; a step for constructing a causal relation model that indicates causal relations between the plurality of feature amounts included in the group; a step for constructing causal relations between all the feature amounts including the other feature amounts that are unemployed; and a step for applying causal relations pertaining to the one feature amount also to causal relations of the other feature amounts.

Description

生産設備に生じ得る事象の解析装置Analysis device for events that may occur in production equipment
　本発明は、生産設備に生じ得る事象の解析装置、解析方法、及び解析プログラムに関する。 The present invention relates to an analysis device, an analysis method, and an analysis program for events that may occur in production equipment.
　生産設備に異常等、所定の事象が発生したとき、その事象の原因要素を特定し、さらに原因要素間の関係を特定した因果関係モデルを構築することが提案されている（例えば、特許文献１）。原因要素とは、例えば、製造設備におけるサーボモータ等の駆動手段、センサ等の監視手段が該当し、例えば、サーボモータの事象がセンサの事象の原因になっている場合には、この因果関係を矢印で表すことで、因果関係モデルを構築する。これにより、事象が生じたときの原因の特定を迅速に行うことができる。 When a predetermined event such as an abnormality occurs in a production facility, it has been proposed to identify the causal element of the event and to construct a causal relationship model that further specifies the relationship between the causal elements (for example, Patent Document 1). ). The causal element corresponds to, for example, a driving means such as a servomotor in a manufacturing facility and a monitoring means such as a sensor. For example, when an event of a servomotor is a cause of an event of a sensor, this causal relationship is considered. A causal relationship model is constructed by representing it with an arrow. As a result, it is possible to quickly identify the cause when an event occurs.
特開２００７－２０７１０１号公報JP-A-2007-207101
　ところで、従来、因果関係モデルの生成時に多重共線性の問題がある場合には、相関の高い特徴量を除去した上で、因果関係モデルの構築が行われることが提案されていた。しかしながら、このようにすると、除去した特徴量に対応する機構が、異常の真因である場合には、これを含まない因果関係モデルが生成されるおそれがあり、異常の真因を特定できないという問題があった。 By the way, conventionally, when there is a problem of multicollinearity when generating a causal relationship model, it has been proposed that a causal relationship model is constructed after removing highly correlated features. However, in this way, if the mechanism corresponding to the removed features is the root cause of the abnormality, a causal relationship model that does not include this may be generated, and the root cause of the abnormality cannot be identified. There was a problem.
　本発明は、上記問題を解決するためになされたものであり、多重共線性の問題があっても、異常などの事象の真因を特定できる因果関係モデルを構築することができる、解析装置、解析方法、及び解析プログラムを提供することを目的とする。 The present invention has been made to solve the above problems, and is capable of constructing a causal relationship model capable of identifying the true cause of an event such as an abnormality even if there is a problem of multicollinearity. It is an object of the present invention to provide an analysis method and an analysis program.
　本発明に係る第１の解析装置は、製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析装置であって、制御部と、記憶部と、を備え、前記記憶部は、複数の前記特徴量を記憶し、前記制御部は、前記事象に対する前記複数の特徴量の間の相関を算出するステップと、前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある一方の前記特徴量を採用し、他方の前記特徴量を採用しないように、前記複数の特徴量を含むグループを生成するステップと、前記グループに含まれる複数の前記特徴量の間の因果関係を示した因果関係モデルを構築するステップと、採用されなかった前記他方の特徴量を含む、全ての前記特徴量の間の因果関係を構築するステップであって、前記一方の特徴量に係る因果関係を、前記他方の特徴量の因果関係にも適用するステップと、を実行させる。 The first analyzer according to the present invention is a production facility for producing a product, and has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production. A driving means and a monitoring means are provided in a production facility having at least one controllable feature amount, and are analysis devices for events that can occur in the production facility, and include a control unit and a storage unit. The storage unit stores a plurality of the feature amounts, and the control unit has a step of calculating a correlation between the plurality of feature amounts with respect to the event, and any one of the correlations. When a strong correlation that satisfies a predetermined requirement is shown, one of the feature quantities having a strong correlation is adopted, and a group including the plurality of feature quantities is generated so as not to adopt the other feature quantity. Between the step, the step of constructing a causal relationship model showing the causal relationship between the plurality of features included in the group, and all the features including the other feature that was not adopted. The step of constructing the causal relationship, that is, the step of applying the causal relationship related to the one feature amount to the causal relationship of the other feature amount, is executed.
　この構成によれば、次の効果を得ることができる。従来、因果関係モデルの生成時に多重共線性の問題がある場合には、相関の高い特徴量を除去した上で、因果関係モデルの構築が行われることがあったが、このようにすると、除去した特徴量に対応する機構が、異常の真因である場合には、これを含まない因果関係モデルが生成されるおそれがあり、異常の真因を特定できないという問題があった。 According to this configuration, the following effects can be obtained. In the past, when there was a problem of multicollinearity when generating a causal relationship model, the causal relationship model was constructed after removing highly correlated features. If the mechanism corresponding to the feature amount is the root cause of the abnormality, a causal relationship model that does not include this may be generated, and there is a problem that the root cause of the abnormality cannot be identified.
　これに対して、本発明に係る解析装置では、相関の高い特徴量のうちの一方を非採用とした上で、因果関係モデルを構築するが、その後、構築された因果関係モデルに対し、採用しなかった特徴量に係る機構を組み込んで、因果関係モデルを再構築している。そのため、関連のある全ての特徴量に対応する機構を含んだ因果関係モデルを構築することができる。その結果、異常の真因を確実に特定することができる。 On the other hand, in the analysis apparatus according to the present invention, one of the highly correlated features is not adopted, and then the causal relationship model is constructed. The causal relationship model is reconstructed by incorporating the mechanism related to the features that were not used. Therefore, it is possible to construct a causal relationship model including a mechanism corresponding to all related features. As a result, the true cause of the abnormality can be reliably identified.
　なお、複数の特徴量の間の相関を算出すると、相関は１以上算出される。例えば、２つの特徴量の間の相関は１つであり、３つの特徴量の間の相関は３つであり、４つの特徴量の間の相関は６つである。こうして算出された１以上の相関のうちのいずれか１つが、所定の要件を充足する強い相関を示す場合、上記のような処理を行う。 When the correlation between a plurality of feature quantities is calculated, the correlation is calculated as 1 or more. For example, there is one correlation between two features, three correlations between three features, and six correlations between four features. When any one of the one or more correlations calculated in this way shows a strong correlation satisfying a predetermined requirement, the above processing is performed.
　本発明に係る第２の解析装置は、製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析装置であって、制御部と、記憶部と、を備え、前記記憶部は、複数の前記特徴量を記憶し、前記制御部は、前記事象に対する前記複数の特徴量の間の相関を算出するステップと、前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある一方の特徴量を採用し、他方の特徴量を採用しないように、特徴量のグループを生成するステップと、前記グループに含まれる複数の前記特徴量の間の因果関係を示した因果関係モデルを構築するステップと、前記因果関係モデルにおいて、前記一方の特徴量に前記他方の特徴量を併記するステップと、を実行させる。 The second analyzer according to the present invention is a production facility for producing a product, and has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production. A driving means and a monitoring means are provided in a production facility having at least one controllable feature amount, and are analysis devices for events that can occur in the production facility, and include a control unit and a storage unit. The storage unit stores a plurality of the feature quantities, and the control unit includes a step of calculating a correlation between the plurality of feature quantities for the event and any one of the correlations. When a strong correlation that satisfies a predetermined requirement is shown, a step of generating a group of feature quantities so as to adopt one feature quantity having the strong correlation and not adopt the other feature quantity, and the group A step of constructing a causal relationship model showing a causal relationship between a plurality of the included feature amounts and a step of writing the other feature amount together with the one feature amount in the causal relationship model are executed. ..
　この構成によれば、上記第１の解析装置と同様の効果を得ることができる。すなわち、本発明に係る第２の解析装置では、相関の高い特徴量のうちの一方を非採用とした上で、因果関係モデルを構築するが、その後、構築された因果関係モデルに対し、採用しなかった特徴量に係る機構を併記して、因果関係モデルを再構築している。そのため、関連のある全ての特徴量に対応する機構を含んだ因果関係モデルを構築することができる。その結果、異常の真因を確実に特定することができる。 According to this configuration, the same effect as that of the first analyzer can be obtained. That is, in the second analysis apparatus according to the present invention, one of the highly correlated features is not adopted, and then the causal relationship model is constructed. After that, the causal relationship model is adopted for the constructed causal relationship model. The causal relationship model is reconstructed by adding the mechanism related to the features that were not used. Therefore, it is possible to construct a causal relationship model including a mechanism corresponding to all related features. As a result, the true cause of the abnormality can be reliably identified.
　本発明に係る第３の解析装置は、製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析装置であって、制御部と、記憶部と、を備え、前記記憶部は、複数の前記特徴量を記憶し、前記制御部は、前記事象に対する前記複数の特徴量の間の相関を算出するステップと、前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある２つの特徴量を組合せ、合成特徴量とする、ステップと、複数の前記特徴量及び前記合成特徴量の間の因果関係を構築するステップと、前記合成特徴量に係る因果関係を、当該合成特徴量を構成する２つの前記特徴量に適用するとともに、前記合成特徴量を削除した全ての前記特徴量の間の因果関係を示す因果関係モデルを構築するステップと、を実行させる。 The third analyzer according to the present invention is a production facility for producing a product, and has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production. A driving means and a monitoring means are provided in a production facility having at least one controllable feature amount, and are analysis devices for events that can occur in the production facility, and include a control unit and a storage unit. The storage unit stores a plurality of the feature amounts, and the control unit has a step of calculating a correlation between the plurality of feature amounts with respect to the event, and any one of the correlations. When a strong correlation that satisfies a predetermined requirement is shown, the two feature quantities having the strong correlation are combined to form a composite feature quantity, which is a causal relationship between the step and the plurality of the feature quantities and the composite feature quantity. And the causal relationship related to the synthetic feature amount are applied to the two said feature amounts constituting the synthetic feature amount, and the causal relationship between all the feature amounts obtained by deleting the synthetic feature amount. To execute the step of building a causal relationship model showing.
　この構成によれば、上記第１の解析装置と同様の効果を得ることができる。すなわち、本発明に係る第２の解析装置では、相関の高い特徴量を組み合わせて合成特徴量を生成し、この合成特徴量を含んだ因果関係モデルを構築するが、その後、構築された因果関係モデルに対し、合成特徴量を削除した上で、合成特徴量に用いられた特徴量を含む、関連のある全て特徴量を用いて、因果関係モデルを再構築している。そのため、関連のある全ての特徴量に対応する機構を含んだ因果関係モデルを構築することができる。その結果、異常の真因を確実に特定することができる。 According to this configuration, the same effect as that of the first analyzer can be obtained. That is, in the second analyzer according to the present invention, a synthetic feature amount is generated by combining highly correlated features, and a causal relationship model including the synthetic feature amount is constructed. After that, the causal relationship constructed is constructed. After deleting the synthetic features from the model, the causal relationship model is reconstructed using all the related features including the features used for the synthetic features. Therefore, it is possible to construct a causal relationship model including a mechanism corresponding to all related features. As a result, the true cause of the abnormality can be reliably identified.
　上記第１の解析装置においては、前記一方の特徴量と前記他方の特徴量との間に、因果関係を形成するステップと、をさらに備えることができる。 The first analysis device can further include a step of forming a causal relationship between the one feature amount and the other feature amount.
　上記第３の解析装置においては、前記合成特徴量を構成する２つの特徴量の間に、因果関係を形成するステップと、をさらに備えることができる。 The third analysis device can further include a step of forming a causal relationship between the two feature quantities constituting the synthetic feature quantity.
　本発明に係る第１の解析方法は、製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析方法であって、前記生産設備に生じ得る事象に対する前記複数の特徴量の間の相関を算出するステップと、前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある一方の前記特徴量を採用し、他方の前記特徴量を採用しないように、前記複数の特徴量を含むグループを生成するステップと、前記グループに含まれる複数の前記特徴量の間の因果関係を示した因果関係モデルを構築するステップと、採用されなかった前記他方の特徴量を含む、全ての前記特徴量の間の因果関係を構築するステップであって、前記一方の特徴量に係る因果関係を、前記他方の特徴量の因果関係にも適用するステップと、を備えている。 The first analysis method according to the present invention is a production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production. A method of analyzing an event that can occur in a production facility and the driving means and the monitoring means are provided in the production facility and have at least one controllable feature amount. If the step of calculating the correlation between a plurality of feature quantities and any one of the above correlations show a strong correlation that satisfies a predetermined requirement, one of the said feature quantities having the strong correlation is adopted. Then, a causal relationship model showing a causal relationship between the step of generating a group including the plurality of feature amounts and the plurality of the feature amounts included in the group is constructed so as not to adopt the other feature amount. This step is a step of constructing a causal relationship between all the feature quantities including the other feature quantity that has not been adopted, and the causal relationship related to the one feature quantity is set to the other feature quantity. It has steps to apply to the causal relationship of.
　本発明に係る第２の解析方法は、製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析方法であって、前記事象に対する前記複数の特徴量の間の相関を算出するステップと、前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある一方の特徴量を採用し、他方の特徴量を採用しないように、特徴量のグループを生成するステップと、前記グループに含まれる複数の前記特徴量の間の因果関係を示した因果関係モデルを構築するステップと、前記因果関係モデルにおいて、前記一方の特徴量に前記他方の特徴量を併記するステップと、を備えている。 The second analysis method according to the present invention is a production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production. The driving means and the monitoring means are provided in the production facility and have at least one controllable feature amount, and are an analysis method of an event that can occur in the production facility, wherein the plurality of feature amounts for the event are If the step of calculating the correlation between the above and any one of the above correlations shows a strong correlation that satisfies a predetermined requirement, one feature quantity having the strong correlation is adopted and the other feature is adopted. In the step of generating a group of feature quantities so as not to adopt a quantity, the step of constructing a causal relationship model showing a causal relationship between a plurality of the feature quantities included in the group, and the causal relationship model. It includes a step of writing the other feature amount together with the one feature amount.
　本発明に係る第３の解析方法は、製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析方法であって、前記事象に対する前記複数の特徴量の間の相関を算出するステップと、前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある２つの特徴量を組合せ、合成特徴量とする、ステップと、複数の前記特徴量及び前記合成特徴量の間の因果関係を構築するステップと、前記合成特徴量に係る因果関係を、当該合成特徴量を構成する２つの前記特徴量に適用するとともに、前記合成特徴量を削除した全ての前記特徴量の間の因果関係を示す因果関係モデルを構築するステップと、を備えている。 The third analysis method according to the present invention is a production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production. A method of analyzing an event that is provided in a production facility and can occur in the production facility, wherein the driving means and the monitoring means have at least one controllable feature amount, and the plurality of feature quantities for the event. When the step of calculating the correlation between the above and any one of the above correlations shows a strong correlation satisfying a predetermined requirement, the two feature quantities having the strong correlation are combined and combined with the composite feature quantity. The step, the step of constructing a causal relationship between the plurality of the feature amounts and the synthetic feature amount, and the causal relationship related to the synthetic feature amount are applied to the two said feature amounts constituting the synthetic feature amount. At the same time, it includes a step of constructing a causal relationship model showing a causal relationship between all the feature amounts in which the synthetic feature amount is deleted.
　本発明に係る第１の解析プログラムは、製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析プログラムであって、コンピュータに、前記生産設備に生じ得る事象に対する前記複数の特徴量の間の相関を算出するステップと、前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある一方の前記特徴量を採用し、他方の前記特徴量を採用しないように、前記複数の特徴量を含むグループを生成するステップと、前記グループに含まれる複数の前記特徴量の間の因果関係を示した因果関係モデルを構築するステップと、採用されなかった前記他方の特徴量を含む、全ての前記特徴量の間の因果関係を構築するステップであって、前記一方の特徴量に係る因果関係を、前記他方の特徴量の因果関係にも適用するステップと、を実行させる。 The first analysis program according to the present invention is a production facility that produces a product, and has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production. A driving means and a monitoring means are provided in a production facility having at least one controllable feature quantity, and are an analysis program of an event that can occur in the production facility, and can occur in the computer and the production facility. When the step of calculating the correlation between the plurality of feature quantities for an event and any one of the correlations show a strong correlation satisfying a predetermined requirement, one of the features having the strong correlation. A causal relationship showing a causal relationship between a step of generating a group including the plurality of feature quantities and a plurality of the feature quantities included in the group so as to adopt the quantity and not adopt the other feature quantity. The step of constructing a model and the step of constructing a causal relationship between all the feature quantities including the other feature quantity that has not been adopted, and the causal relationship relating to the one feature quantity is the other. The step of applying to the causal relationship of the feature amount of is executed.
　本発明に係る第２の解析プログラムは、製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析プログラムであって、コンピュータに、前記事象に対する前記複数の特徴量の間の相関を算出するステップと、前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある一方の特徴量を採用し、他方の特徴量を採用しないように、特徴量のグループを生成するステップと、前記グループに含まれる複数の前記特徴量の間の因果関係を示した因果関係モデルを構築するステップと、前記因果関係モデルにおいて、前記一方の特徴量に前記他方の特徴量を併記するステップと、を実行させる。 The second analysis program according to the present invention is a production facility that produces a product, and has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production. A drive means and a monitoring means are provided in a production facility having at least one controllable feature amount, and are an analysis program of an event that can occur in the production facility. If one of the steps of calculating the correlation between the feature quantities of the above and one of the above correlations shows a strong correlation satisfying a predetermined requirement, one of the feature quantities having the strong correlation is adopted. A step of generating a group of feature quantities so as not to adopt the other feature quantity, a step of constructing a causal relationship model showing a causal relationship between a plurality of the feature quantities included in the group, and the causal relationship. In the model, the step of writing the other feature amount together with the one feature amount is executed.
　本発明に係る第３の解析プログラムは、製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析プログラムであって、コンピュータに、前記事象に対する前記複数の特徴量の間の相関を算出するステップと、前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある２つの特徴量を組合せ、合成特徴量とする、ステップと、複数の前記特徴量及び前記合成特徴量の間の因果関係を構築するステップと、前記合成特徴量に係る因果関係を、当該合成特徴量を構成する２つの前記特徴量に適用するとともに、前記合成特徴量を削除した全ての前記特徴量の間の因果関係を示す因果関係モデルを構築するステップと、を実行させる。 The third analysis program according to the present invention is a production facility that produces a product, and has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production. A drive means and a monitoring means are provided in a production facility having at least one controllable feature amount, and are an analysis program of an event that can occur in the production facility. If one of the steps of calculating the correlation between the feature quantities of the above and one of the above correlations shows a strong correlation satisfying a predetermined requirement, the two feature quantities having the strong correlation are combined and synthesized. The two said features constituting the synthetic feature amount, the step of making the feature amount, the step of constructing the causal relationship between the plurality of the feature amount and the synthetic feature amount, and the causal relationship related to the synthetic feature amount. A step of constructing a causal relationship model showing a causal relationship between all the said feature quantities in which the synthetic feature quantity is deleted while applying the quantity is executed.
　本発明によれば、多重共線性の問題があっても、異常などの事象の真因を特定できる因果関係モデルを構築することができる。 According to the present invention, it is possible to construct a causal relationship model that can identify the true cause of an event such as an abnormality even if there is a problem of multicollinearity.
本発明が適用される場面の一例を模式的に例示する。An example of a situation in which the present invention is applied is schematically illustrated. 本発明の一実施形態に係る解析装置のハードウェア構成を示すブロック図である。It is a block diagram which shows the hardware structure of the analysis apparatus which concerns on one Embodiment of this invention. 本発明の一実施形態に係る生産設備の概略図である。It is the schematic of the production equipment which concerns on one Embodiment of this invention. 包装機の概略図に因果関係モデルのノードを重ねた図である。It is the figure which superposed the node of the causal relation model on the schematic diagram of a packaging machine. 解析装置の機能構成を示すブロック図である。It is a block diagram which shows the functional structure of an analyzer. 因果関係モデルの構築の例を示すフローチャートである。It is a flowchart which shows an example of construction of a causal relational model. 相関を算出した後の、特徴量の選択方法を示す図である。It is a figure which shows the selection method of a feature amount after calculating the correlation. 因果関係モデルの構築の例を示すフローチャートである。It is a flowchart which shows an example of construction of a causal relational model. 因果関係モデルの例である。This is an example of a causal relationship model. 因果関係モデルの例である。This is an example of a causal relationship model. 因果関係モデルの例である。This is an example of a causal relationship model. 因果関係モデルの再構築の例である。This is an example of reconstructing a causal relationship model. 因果関係モデルの再構築の例である。This is an example of reconstructing a causal relationship model. 因果関係モデルの再構築の例である。This is an example of reconstructing a causal relationship model. 因果関係モデルの再構築の例である。This is an example of reconstructing a causal relationship model. 因果関係モデルの再構築の例である。This is an example of reconstructing a causal relationship model.
　以下、本発明の一側面に係る実施の形態（以下、「本実施形態」とも表記する）を、図面に基づいて説明する。ただし、以下で説明する本実施形態は、あらゆる点において本発明の例示に過ぎない。本発明の範囲を逸脱することなく種々の改良や変形を行うことができることは言うまでもない。つまり、本発明の実施にあたって、実施形態に応じた具体的構成が適宜採用されてもよい。なお、本実施形態において登場するデータを自然言語により説明しているが、より具体的には、コンピュータが認識可能な疑似言語、コマンド、パラメータ、マシン語等で指定される。 Hereinafter, an embodiment according to one aspect of the present invention (hereinafter, also referred to as “the present embodiment”) will be described with reference to the drawings. However, the embodiments described below are merely examples of the present invention in all respects. Needless to say, various improvements and modifications can be made without departing from the scope of the present invention. That is, in carrying out the present invention, a specific configuration according to the embodiment may be appropriately adopted. Although the data appearing in the present embodiment is described in natural language, more specifically, it is specified in a pseudo language, a command, a parameter, a machine language, etc. that can be recognized by a computer.
　＜１．適用例＞
　まず、図１を用いて、本発明が適用される場面の一例について説明する。図１は、本実施形態に係る生産システムの適用場面の一例を模式的に例示する。本実施形態に係る生産システムは、生産設備の一例である包装機３と、解析装置１と、を備えている。解析装置１は、包装機３に設けられたサーボモータ（駆動手段、監視手段（例えば、サーボモータのエンコーダ等））や各種センサ（監視手段）間の因果関係モデルを構築するように構成されたコンピュータである。なお、以下では、サーボモータ等の駆動手段、監視手段や各種センサ等の監視手段を合わせて機構２１と称することとする。 <1. Application example>
First, an example of a situation in which the present invention is applied will be described with reference to FIG. FIG. 1 schematically illustrates an example of an application scene of the production system according to the present embodiment. The production system according to the present embodiment includes a packaging machine 3 which is an example of production equipment, and an analysis device 1. The analyzer 1 is configured to build a causal relationship model between a servomotor (driving means, monitoring means (for example, encoder of a servomotor, etc.)) and various sensors (monitoring means) provided in the packaging machine 3. It is a computer. In the following, the driving means such as a servomotor, the monitoring means, and the monitoring means such as various sensors are collectively referred to as the mechanism 21.
　解析装置１は、包装機３に生じる種々の事象、例えば、異常について、機構２１間の因果関係モデルを構築する。そこで、まず、包装機３を構成する複数の機構２１の状態に関する複数件の計測データを取得する。 The analysis device 1 constructs a causal relationship model between the mechanisms 21 for various events occurring in the packaging machine 3, for example, an abnormality. Therefore, first, a plurality of measurement data regarding the states of the plurality of mechanisms 21 constituting the packaging machine 3 are acquired.
　続いて、解析装置１は、取得された複数件の計測データ１２４から、特徴量を算出する。特徴量は、各機構２１の出力値（トルク、速度など）の振幅、最大値、最小値、平均値などの数値である。次に、算出された特徴量の相関を算出する。図１の例では、特徴量Ａ～Ｃが算出され、特徴量Ａ～Ｃのうち、特徴量Ａ，Ｂの相関が高いと判定されたとする。 Subsequently, the analysis device 1 calculates the feature amount from the acquired plurality of measurement data 124. The feature amount is a numerical value such as an amplitude, a maximum value, a minimum value, and an average value of the output value (torque, speed, etc.) of each mechanism 21. Next, the correlation of the calculated features is calculated. In the example of FIG. 1, it is assumed that the feature amounts A to C are calculated and it is determined that the feature amounts A and B have a high correlation among the feature amounts A to C.
　この場合、特徴量Ａ，Ｂのうち、特徴量Ａを採用し、特徴量Ｂを採用しないようにする。したがって、この例では、特徴量Ａ，Ｃを用いて、これらに対応する機構２１間の因果関係モデルを構築する。その結果、例えば、特徴量Ｃから特徴量Ａに向くエッジを有する因果関係モデルが構築されたとする。なお、因果関係モデルは、特徴量に対応する機構２１間の関係を示すものであるが、ここでは、説明の便宜のため、特徴量Ａ～Ｃを、これらに対応する機構として示すこととする。なお、この点は、本明細書の記載において同じである。 In this case, of the feature quantities A and B, the feature quantity A is adopted and the feature quantity B is not adopted. Therefore, in this example, the causal relationship model between the mechanisms 21 corresponding to the features A and C is constructed. As a result, for example, it is assumed that a causal relationship model having an edge facing the feature amount A from the feature amount C is constructed. The causal relationship model shows the relationship between the mechanisms 21 corresponding to the feature amounts, but here, for convenience of explanation, the feature amounts A to C are shown as the mechanisms corresponding to these. .. This point is the same in the description of the present specification.
　こうして、因果関係モデルが構築された後、採用されなかった特徴量Ｂを組み込んだ因果関係モデルを再構築する。すなわち、特徴量Ａ，Ｂは相関が高いため、等価とみなし、同じエッジが接続されたノードとして因果関係モデルに組み込む。図１の例では、特徴量Ｃから特徴量Ｂに向かうエッジとともに、特徴量Ｃを組み込んだ因果関係モデルを再構築する。以上より、多重共線性の問題があるような因果関係モデルであっても、全ての特徴量（機構）が含まれた因果関係モデルを構築することができる。 In this way, after the causal relationship model is constructed, the causal relationship model incorporating the feature quantity B that has not been adopted is reconstructed. That is, since the features A and B have a high correlation, they are regarded as equivalent and incorporated into the causal relationship model as nodes to which the same edges are connected. In the example of FIG. 1, the causal relational model incorporating the feature amount C is reconstructed together with the edge from the feature amount C to the feature amount B. From the above, it is possible to construct a causal relationship model that includes all features (mechanisms) even if the causal relationship model has a problem of multicollinearity.
　なお、上記の説明では、生産設備の例として包装機３を示しているが、何らかの物を生産可能であればよく、その種類は、特に限定されなくてもよい。各機構の種類は、特に限定されなくてもよく、実施の形態に応じて適宜選択されてよい。各機構は、例えば、コンベア、ロボットアーム、サーボモータ、シリンダ（成形機等）、吸着パッド、カッター装置、シール装置等であってよい。また、各機構は、上述した包装機３のほか、例えば、印刷機、実装機、リフロー炉、基板検査装置等の複合装置であってもよい。更に、各機構は、例えば、上記のような何らかの物理的な動作を伴う装置の他に、例えば、各種センサにより何らかの情報を検知する装置、各種センサからデータを取得する装置、取得したデータから何らかの情報を検知する装置、取得したデータを情報処理する装置等の内部処理を行う装置を含んでもよい。１つの機構は、１又は複数の装置で構成されてもよいし、装置の一部で構成されてもよい。複数の機構により１つの装置が構成されてもよい。また、同一の装置が複数の処理を実行する場合には、それぞれを別の機構とみなしてもよい。例えば、同一の装置が第１の処理と第２の処理とを実行する場合に、第１の処理を実行する当該装置を第１の機構とみなし、第２の処理を実行する当該装置を第２の機構とみなしてもよい。 In the above description, the packaging machine 3 is shown as an example of the production equipment, but the type may not be particularly limited as long as it can produce something. The type of each mechanism does not have to be particularly limited and may be appropriately selected according to the embodiment. Each mechanism may be, for example, a conveyor, a robot arm, a servomotor, a cylinder (molding machine or the like), a suction pad, a cutter device, a sealing device, or the like. In addition to the packaging machine 3 described above, each mechanism may be a composite device such as a printing machine, a mounting machine, a reflow furnace, or a substrate inspection device. Further, each mechanism includes, for example, a device that detects some information by various sensors, a device that acquires data from various sensors, and some device from the acquired data, in addition to the device that involves some physical operation as described above. It may include a device that performs internal processing such as a device that detects information and a device that processes acquired data. One mechanism may be composed of one or a plurality of devices, or may be composed of a part of the devices. One device may be configured by a plurality of mechanisms. Further, when the same device executes a plurality of processes, each may be regarded as a different mechanism. For example, when the same device executes the first process and the second process, the device that executes the first process is regarded as the first mechanism, and the device that executes the second process is the first. It may be regarded as the mechanism of 2.
　＜２．構成例＞
　＜２－１．ハードウェア構成＞
　次に、本実施形態に係る解析装置１及び包装機３のハードウェア構成の一例について説明する。図２は、本実施形態に係る解析装置１のハードウェア構成の一例を示すブロック図であり、図３は包装機の概略構成を示す図である。 <2. Configuration example>
<2-1. Hardware configuration>
Next, an example of the hardware configuration of the analysis device 1 and the packaging machine 3 according to the present embodiment will be described. FIG. 2 is a block diagram showing an example of the hardware configuration of the analysis device 1 according to the present embodiment, and FIG. 3 is a diagram showing a schematic configuration of the packaging machine.
　＜２－１－１．解析装置＞
　次に、図２を用いて、本実施形態に係る解析装置１のハードウェア構成の一例について説明する。図２は、本実施形態に係る解析装置１のハードウェア構成の一例を模式的に例示する。 <2-1-1. Analyst>
Next, an example of the hardware configuration of the analysis device 1 according to the present embodiment will be described with reference to FIG. FIG. 2 schematically illustrates an example of the hardware configuration of the analysis device 1 according to the present embodiment.
　図２に示されるとおり、本実施形態に係る解析装置１は、制御部１１、記憶部１２、通信インタフェース１３、入力装置１４、表示装置１５、及びドライブ１６が電気的に接続されたコンピュータである。なお、図２では、通信インタフェースを「通信Ｉ／Ｆ」と記載している。 As shown in FIG. 2, the analysis device 1 according to the present embodiment is a computer to which the control unit 11, the storage unit 12, the communication interface 13, the input device 14, the display device 15, and the drive 16 are electrically connected. .. In FIG. 2, the communication interface is described as "communication I / F".
　制御部１１は、ハードウェアプロセッサであるＣＰＵ（Central Processing Unit）、ＲＡＭ（Random Access Memory）、ＲＯＭ（Read Only Memory）等を含み、プログラム及び各種データに基づいて情報処理を実行するように構成される。記憶部１２は、メモリの一例であり、例えば、ハードディスクドライブ、ソリッドステートドライブ等の補助記憶装置により構成される。本実施形態では、記憶部１２は、解析プログラム１２１、変数設定データ１２２、因果関係モデルデータ１２３、計測データ１２４等の各種情報を記憶する。 The control unit 11 includes a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), and the like, which are hardware processors, and is configured to execute information processing based on a program and various data. To. The storage unit 12 is an example of a memory, and is composed of, for example, an auxiliary storage device such as a hard disk drive or a solid state drive. In the present embodiment, the storage unit 12 stores various information such as the analysis program 121, the variable setting data 122, the causal relationship model data 123, and the measurement data 124.
　解析プログラム８１は、複数の機構２１間の因果関係の導出に関する情報処理（後述する図６及び図８）を解析装置１に実行させるためのプログラムである。解析プログラム１２１は、この情報処理の一連の命令群を含む。変数設定データ１２２は、後述するように、相関の高い特徴量の組合せに関するデータであり、因果関係モデルデータ１２３は、生成された因果関係モデルに関するデータであり、後述する再構築前の因果関係モデルと再構築後の因果関係モデルの両方を含むデータである。また、計測データ１２４は、上述したように、包装機３を構成する複数の機構２１の状態に関するデータである。 The analysis program 81 is a program for causing the analysis device 1 to execute information processing (FIGS. 6 and 8 described later) relating to the derivation of the causal relationship between the plurality of mechanisms 21. The analysis program 121 includes a series of instructions for this information processing. As will be described later, the variable setting data 122 is data related to the combination of highly correlated feature quantities, and the causal relationship model data 123 is data related to the generated causal relationship model, which is a causal relationship model before reconstruction, which will be described later. The data includes both the causal relationship model after reconstruction and the reconstruction. Further, as described above, the measurement data 124 is data relating to the states of the plurality of mechanisms 21 constituting the packaging machine 3.
　通信インタフェース１３は、例えば、有線ＬＡＮ（Local Area Network）モジュール、無線ＬＡＮモジュール等であり、ネットワークを介した有線又は無線通信を行うためのインタフェースである。解析装置１は、この通信インタフェース１３により、例えば、包装機３の動作を制御するように構成された制御装置（不図示）等の他の情報処理装置との間でネットワークを介したデータ通信を行い、複数件の計測データ１２４を取得することができる。ネットワークの種類は、例えば、インターネット、無線通信網、移動通信網、電話網、専用網等から適宜選択されてよい。ただし、計測データ１２４を取得する経路は、このような例に限定されなくてもよい。 The communication interface 13 is, for example, a wired LAN (Local Area Network) module, a wireless LAN module, or the like, and is an interface for performing wired or wireless communication via a network. The analysis device 1 uses this communication interface 13 to perform data communication via a network with another information processing device such as a control device (not shown) configured to control the operation of the packaging machine 3, for example. This can be performed to acquire a plurality of measurement data 124. The type of network may be appropriately selected from, for example, the Internet, a wireless communication network, a mobile communication network, a telephone network, a dedicated network, and the like. However, the route for acquiring the measurement data 124 does not have to be limited to such an example.
　入力装置１４は、例えば、マウス、キーボード等の入力を行うための装置である。また、表示装置１５は、出力装置の一例であり、例えば、ディスプレイである。オペレータは、入力装置１４及び表示装置１５を介して、解析装置１を操作することができる。なお、表示装置１５は、タッチパネルディスプレイであってもよい。この場合、入力装置１４は省略されてもよい。 The input device 14 is, for example, a device for inputting a mouse, a keyboard, or the like. The display device 15 is an example of an output device, for example, a display. The operator can operate the analysis device 1 via the input device 14 and the display device 15. The display device 15 may be a touch panel display. In this case, the input device 14 may be omitted.
　ドライブ１６は、例えば、ＣＤドライブ、ＤＶＤドライブ等であり、記憶媒体９１に記憶されたプログラムを読み込むためのドライブ装置である。ドライブ１６の種類は、記憶媒体９１の種類に応じて適宜選択されてよい。上記解析プログラム１２１、変数設定データ１２２、因果関係モデルデータ１２３、及び複数件の計測データ１２４の少なくともいずれかは、この記憶媒体９１に記憶されていてもよい。 The drive 16 is, for example, a CD drive, a DVD drive, or the like, and is a drive device for reading a program stored in the storage medium 91. The type of the drive 16 may be appropriately selected according to the type of the storage medium 91. At least one of the analysis program 121, the variable setting data 122, the causal relationship model data 123, and the plurality of measurement data 124 may be stored in the storage medium 91.
　記憶媒体９１は、コンピュータその他装置、機械等が記録されたプログラム等の情報を読み取り可能なように、当該プログラム等の情報を、電気的、磁気的、光学的、機械的又は化学的作用によって蓄積する媒体である。解析装置１は、この記憶媒体９１から、上記解析プログラム８１、及び複数件の計測データ１２４の少なくともいずれかを取得してもよい。 The storage medium 91 stores the information of the program or the like by electrical, magnetic, optical, mechanical or chemical action so that the information of the program or the like recorded by the computer or other device, the machine or the like can be read. It is a medium to do. The analysis device 1 may acquire at least one of the analysis program 81 and a plurality of measurement data 124 from the storage medium 91.
　ここで、図２では、記憶媒体９１の一例として、ＣＤ、ＤＶＤ等のディスク型の記憶媒体を例示している。しかしながら、記憶媒体９１の種類は、ディスク型に限定される訳ではなく、ディスク型以外であってもよい。ディスク型以外の記憶媒体として、例えば、フラッシュメモリ等の半導体メモリを挙げることができる。 Here, FIG. 2 illustrates a disc-type storage medium such as a CD or DVD as an example of the storage medium 91. However, the type of the storage medium 91 is not limited to the disk type, and may be other than the disk type. Examples of storage media other than the disk type include semiconductor memories such as flash memories.
　なお、解析装置１の具体的なハードウェア構成に関して、実施形態に応じて、適宜、構成要素の省略、置換及び追加が可能である。例えば、制御部１１は、複数のハードウェアプロセッサを含んでもよい。ハードウェアプロセッサは、マイクロプロセッサ、ＦＰＧＡ（field-programmable gate array）、ＤＳＰ（digital signal processor）等で構成されてよい。記憶部１２は、制御部１１に含まれるＲＡＭ及びＲＯＭにより構成されてもよい。通信インタフェース１３、入力装置１４、表示装置１５及びドライブ１６の少なくともいずれかは省略されてもよい。解析装置１は、例えば、スピーカ等の表示装置１５以外の出力装置を更に備えてもよい。解析装置１は、複数台のコンピュータで構成されてもよい。この場合、各コンピュータのハードウェア構成は、一致していてもよいし、一致していなくてもよい。また、解析装置１は、提供されるサービス専用に設計された情報処理装置の他、デスクトップＰＣ（Personal Computer）、タブレットＰＣ等の汎用の情報処理装置、汎用のサーバ装置等であってもよい。更に、解析装置１は、包装機３の動作を制御可能に構成されてよい。この場合、解析装置１は、ＰＬＣ（programmable logic controller）であってもよい。また、解析装置１は、包装機３と接続するための入出力インタフェースを備えてもよく、この入出力インタフェースを介して計測データ１２４を取得してもよい。 Regarding the specific hardware configuration of the analysis device 1, components can be omitted, replaced, or added as appropriate according to the embodiment. For example, the control unit 11 may include a plurality of hardware processors. The hardware processor may be composed of a microprocessor, an FPGA (field-programmable gate array), a DSP (digital signal processor), or the like. The storage unit 12 may be composed of a RAM and a ROM included in the control unit 11. At least one of the communication interface 13, the input device 14, the display device 15, and the drive 16 may be omitted. The analysis device 1 may further include an output device other than the display device 15 such as a speaker. The analysis device 1 may be composed of a plurality of computers. In this case, the hardware configurations of the computers may or may not match. Further, the analysis device 1 may be a general-purpose information processing device such as a desktop PC (Personal Computer) or a tablet PC, a general-purpose server device, or the like, in addition to an information processing device designed exclusively for the provided service. Further, the analysis device 1 may be configured to be able to control the operation of the packaging machine 3. In this case, the analysis device 1 may be a PLC (programmable logic controller). Further, the analysis device 1 may be provided with an input / output interface for connecting to the packaging machine 3, and measurement data 124 may be acquired via this input / output interface.
　＜２－１－２．包装機＞
　次に、図３を用いて、本実施形態に係る包装機３のハードウェア構成の一例を説明する。図３は、本実施形態に係る包装機３のハードウェア構成の一例を模式的に例示する。包装機３は、いわゆる横ピロー包装機であり、食品（乾燥麺等）、文房具（消しゴム等）等の内容物ＷＡを包装する装置である。但し、内容物ＷＡの種類は、実施の形態に応じて適宜選択可能であり、特には限定されない。この包装機３は、包装フィルムが巻き取られたフィルムロール３０と、包装フィルムを搬送するフィルム搬送部３１と、内容物ＷＡを搬送する内容物搬送部３２と、内容物を包装フィルムで放送する製袋部３３と、を備えている。 <2-1-2. Packaging machine ＞
Next, an example of the hardware configuration of the packaging machine 3 according to the present embodiment will be described with reference to FIG. FIG. 3 schematically illustrates an example of the hardware configuration of the packaging machine 3 according to the present embodiment. The wrapping machine 3 is a so-called horizontal pillow wrapping machine, which is a device for wrapping contents WA such as food (dried noodles, etc.) and stationery (eraser, etc.). However, the type of the content WA can be appropriately selected according to the embodiment, and is not particularly limited. The wrapping machine 3 broadcasts the film roll 30 on which the wrapping film is wound, the film transport unit 31 for transporting the wrapping film, the content transport unit 32 for transporting the content WA, and the content on the wrapping film. It is provided with a bag making portion 33.
　包装フィルムは、例えば、ポリエチレンフィルム等の樹脂フィルムとすることができる。フィルムロール３０は巻き芯を備えており、包装フィルムはその巻き芯に巻き取られている。巻き芯は軸周りに回転可能に支持されており、これにより、フィルムロール３０は、回転しながら包装フィルムを繰り出すことができるように構成されている。 The packaging film can be, for example, a resin film such as a polyethylene film. The film roll 30 includes a winding core, and the packaging film is wound around the winding core. The winding core is rotatably supported around the axis, whereby the film roll 30 is configured so that the packaging film can be unwound while rotating.
　フィルム搬送部３１は、サーボモータ（サーボ１）３１１により駆動される駆動ローラと、この駆動ローラから回転力を付与される受動ローラ３１２と、包装フィルムにテンションをかけながらガイドする複数のプーリ３１３と、を備えている。これにより、フィルム搬送部３１は、フィルムロール３０から包装フィルムを繰り出し、繰り出した包装フィルムを弛ませることなく製袋部３３に搬送するように構成されている。 The film transport unit 31 includes a drive roller driven by a servomotor (servo 1) 311, a passive roller 312 to which a rotational force is applied from the drive roller, and a plurality of pulleys 313 that guide the packaging film while applying tension. , Is equipped. As a result, the film transport section 31 is configured to feed the packaging film from the film roll 30 and transport the delivered packaging film to the bag making section 33 without loosening it.
　内容物搬送部３２は、包装対象となる内容物ＷＡを搬送するコンベア３２１と、コンベア３２１を駆動するサーボモータ（サーボ２）３２２と、を備えている。図３に例示されるように、内容物搬送部３２は、フィルム搬送部３１の下方を経て、製袋部３３に連結している。これにより、内容物搬送部３２により搬送される内容物ＷＡは、製袋部３３に供給されるとともに、フィルム搬送部３１から供給された包装フィルムにより包装される。また、コンベア３２１の下流の情報には、内容物ＷＡの位置を検知するファイバセンサ（センサ１）３２４が設けられている。さらに、コンベア３２１の下方には、内容物ＷＡの乗り上げ等を検知するファイバセンサ（センサ２）３２５が設けられている。これらセンサ１，２により、内容物ＷＡが正しく包装されるために、正しい位置で搬送されているか否かを検知する。 The content transfer unit 32 includes a conveyor 321 that conveys the content WA to be packaged, and a servomotor (servo 2) 322 that drives the conveyor 321. As illustrated in FIG. 3, the content transporting section 32 is connected to the bag making section 33 via the lower part of the film transporting section 31. As a result, the content WA transported by the content transport unit 32 is supplied to the bag making unit 33 and packaged by the packaging film supplied from the film transport unit 31. Further, in the information downstream of the conveyor 321, a fiber sensor (sensor 1) 324 for detecting the position of the content WA is provided. Further, below the conveyor 321 is provided a fiber sensor (sensor 2) 325 for detecting the riding of the contents WA and the like. These sensors 1 and 2 detect whether or not the contents WA are transported in the correct position in order to be correctly packaged.
　製袋部３３は、コンベア３３１と、コンベア３３１を駆動するサーボモータ（サーボ３）３３２と、包装フィルムを搬送方向にシールするセンターシール部３３３と、包装フィルムを搬送方向の両端側で切断し、各端部でシールするエンドシール部３３４と、を備えている。 The bag making section 33 cuts the conveyor 331, the servo motor (servo 3) 332 that drives the conveyor 331, the center seal section 333 that seals the packaging film in the transport direction, and the packaging film on both ends in the transport direction. It includes an end seal portion 334 that seals at each end portion.
　コンベア３３１は、内容物搬送部３２から搬送された内容物ＷＡとフィルム搬送部３１から供給された包装フィルムとを搬送する。フィルム搬送部３１から供給された包装フィルムは、幅方向の両側端縁部同士が重なるように適宜折り曲げられつつ、センターシール部３３３に供給される。センターシール部３３３は、例えば、左右一対の加熱ローラ（ヒータ１，２）により構成されており、折り曲げられた包装フィルムの両側端縁部を加熱により搬送方向に沿ってシールする。これにより、包装フィルムは、筒状に形成される。内容物ＷＡは、この筒状に形成された包装フィルム内に投入される。また、エンドシール部３３４の上流側には、コンベア３３１の上方に、内容物ＷＡの位置を検知するファイバセンサ（センサ３）３３６が設けられている。 The conveyor 331 conveys the content WA conveyed from the content transfer unit 32 and the packaging film supplied from the film transfer unit 31. The packaging film supplied from the film transport unit 31 is supplied to the center seal unit 333 while being appropriately bent so that both side edge portions in the width direction overlap each other. The center seal portion 333 is composed of, for example, a pair of left and right heating rollers (heaters 1 and 2), and seals both side edge portions of the bent packaging film along the transport direction by heating. As a result, the packaging film is formed in a tubular shape. The content WA is put into the packaging film formed in this tubular shape. Further, on the upstream side of the end seal portion 334, a fiber sensor (sensor 3) 336 that detects the position of the content WA is provided above the conveyor 331.
　一方、エンドシール部３３４は、例えば、サーボモータ（サーボ４）３３５により駆動されるローラと、ローラの回転によって開閉する一対のカッタと、各カッタの両側に設けられるヒータ（ヒータ３）と、を有している。これらにより、エンドシール部３３４は、搬送方向に直交する方向に筒状の包装フィルムをカットするとともに、カットした部分で加熱によりシールすることができるように構成されている。このエンドシール部３３４を通過すると、筒状に形成された包装フィルムの先端部分は、搬送方向の両側でシールされ、後続から分離されて、内容物ＷＡを内包する包装体ＷＢとなる。 On the other hand, the end seal portion 334 includes, for example, a roller driven by a servomotor (servo 4) 335, a pair of cutters that open and close by rotation of the rollers, and heaters (heaters 3) provided on both sides of each cutter. Have. As a result, the end seal portion 334 is configured so that the tubular packaging film can be cut in the direction orthogonal to the transport direction and can be sealed by heating at the cut portion. After passing through the end seal portion 334, the tip portion of the packaging film formed in a tubular shape is sealed on both sides in the transport direction and separated from the subsequent portion to become a packaging body WB containing the contents WA.
　＜２－１－３．包装工程＞
　以上の包装機３は、次のような工程で、内容物ＷＡの包装を行うことができる。すなわち、フィルム搬送部３１によって、フィルムロール３０から包装フィルムを繰り出す。また、内容物搬送部３２によって、包装対象となる内容物ＷＡを搬送する。次に、製袋部３３のセンターシール部３３３によって、繰り出された包装フィルムを筒状に形成する。そして、形成した筒状の包装フィルムに内容物ＷＡを投入した上で、エンドシール部３３４によって、搬送方向に直交する方向に筒状の包装フィルムをカットすると共に、カットした部分の搬送方向の両側で加熱によりシールする。これにより、内容物ＷＡを内包する横ピロー型の包装体ＷＢが形成される。すなわち、内容物ＷＡの包装が完了する。 <2-1-3. Packaging process>
The above-mentioned packaging machine 3 can wrap the contents WA in the following steps. That is, the film transport unit 31 feeds out the packaging film from the film roll 30. In addition, the content transport unit 32 transports the content WA to be packaged. Next, the center seal portion 333 of the bag making portion 33 forms the unwound packaging film into a tubular shape. Then, after the content WA is put into the formed tubular packaging film, the tubular packaging film is cut in the direction orthogonal to the transport direction by the end seal portion 334, and both sides of the cut portion in the transport direction are cut. Seal by heating with. As a result, a horizontal pillow type package WB containing the content WA is formed. That is, the packaging of the contents WA is completed.
　なお、包装機３の駆動の制御は、包装機３とは別個に設けたＰＬＣなどで行うこともできる。この場合、上述した計測データ１２４は、ＰＬＣから取得することができる。また、上記のように構成された包装機３では、一例として、図４に示すように、異常の因果関係を構築するために、１０個の機構が設定されている。すなわち、上述したサーボ１～４，ヒータ１～３，及びセンサ１～３が機構として設定され、異常が発生するときの、これら機構間の因果関係が因果関係モデルとして構築される。詳細は後述する。 Note that the drive of the packaging machine 3 can be controlled by a PLC or the like provided separately from the packaging machine 3. In this case, the above-mentioned measurement data 124 can be acquired from the PLC. Further, in the packaging machine 3 configured as described above, as an example, as shown in FIG. 4, 10 mechanisms are set in order to establish a causal relationship of abnormalities. That is, the above-mentioned servos 1 to 4, heaters 1 to 3, and sensors 1 to 3 are set as mechanisms, and a causal relationship between these mechanisms when an abnormality occurs is constructed as a causal relationship model. Details will be described later.
　＜２－２．ソフトウエア構成＞
　次に、図５を用いて、本実施形態に係る解析装置１のソフトウェア構成の一例を説明する。図５は、本実施形態に係る解析装置１のソフトウェア構成の一例を模式的に例示する。 <2-2. Software configuration>
Next, an example of the software configuration of the analysis device 1 according to the present embodiment will be described with reference to FIG. FIG. 5 schematically illustrates an example of the software configuration of the analysis device 1 according to the present embodiment.
　解析装置１の制御部１１は、記憶部１２に記憶された解析プログラム８１をＲＡＭに展開する。そして、制御部１１は、ＲＡＭに展開された解析プログラム８１をＣＰＵにより解釈及び実行して、各構成要素を制御する。これによって、図３に示されるとおり、本実施形態に係る解析装置１は、データ取得部１１１、変数設定部１１２、モデル構築部１１３、モデル再構築部１１４、及び出力部１１５をソフトウェアモジュールとして備えるコンピュータとして動作する。すなわち、本実施形態では、解析装置１の各ソフトウェアモジュールは、制御部１１（ＣＰＵ）により実現される。 The control unit 11 of the analysis device 1 expands the analysis program 81 stored in the storage unit 12 into the RAM. Then, the control unit 11 interprets and executes the analysis program 81 expanded in the RAM by the CPU to control each component. As a result, as shown in FIG. 3, the analysis device 1 according to the present embodiment includes a data acquisition unit 111, a variable setting unit 112, a model construction unit 113, a model reconstruction unit 114, and an output unit 115 as software modules. Operates as a computer. That is, in the present embodiment, each software module of the analysis device 1 is realized by the control unit 11 (CPU).
　データ取得部１１１は、包装機３を構成する複数の機構２１の状態に関する複数件の計測データ１２４を取得するとともに、計測データ１２４から各機構２１の特徴量を算出する。変数設定部１１２は、算出された特徴量間の相関を算出、相関の高い特徴量の組合せを記憶部１２に記憶する。また、変数設定部１１２は、相関の高い特徴量の組合せの中から、因果関係モデルの構築において採用する特徴量を決定する。詳細は、後述する。 The data acquisition unit 111 acquires a plurality of measurement data 124 relating to the states of the plurality of mechanisms 21 constituting the packaging machine 3, and calculates the feature amount of each mechanism 21 from the measurement data 124. The variable setting unit 112 calculates the correlation between the calculated feature amounts, and stores the combination of the feature amounts with high correlation in the storage unit 12. Further, the variable setting unit 112 determines the feature amount to be adopted in the construction of the causal relationship model from the combination of the feature amount having high correlation. Details will be described later.
　モデル構築部１１３は、上述した特徴量を統計的に解析することで、複数の機構２間の因果関係モデルを構築する。モデル再構築部１１４は、上述した相関が高い特徴量の組合せの中で、因果関係モデルの構築に採用されなかった特徴量を、生成された因果関係モデルに組み込み、因果関係モデルを再構築する。出力部１１５は、再構築された因果関係モデルに関するデータを表示装置等に出力する。 The model building unit 113 builds a causal relationship model between a plurality of mechanisms 2 by statistically analyzing the above-mentioned features. The model reconstruction unit 114 reconstructs the causal relationship model by incorporating the features that were not adopted in the construction of the causal relationship model among the combinations of the above-mentioned highly correlated features into the generated causal relationship model. .. The output unit 115 outputs data related to the reconstructed causal relationship model to a display device or the like.
　解析装置１の各ソフトウェアモジュールに関しては後述する動作例で詳細に説明する。なお、本実施形態では、解析装置１の各ソフトウェアモジュールがいずれも汎用のＣＰＵにより実現される例について説明している。しかしながら、以上のソフトウェアモジュールの一部又は全部が、１又は複数の専用のハードウェアプロセッサにより実現されてもよい。また、解析装置１のソフトウェア構成に関して、実施形態に応じて、適宜、ソフトウェアモジュールの省略、置換及び追加が行われてもよい。 Each software module of the analysis device 1 will be described in detail in an operation example described later. In this embodiment, an example in which each software module of the analysis device 1 is realized by a general-purpose CPU is described. However, some or all of the above software modules may be implemented by one or more dedicated hardware processors. Further, regarding the software configuration of the analysis device 1, software modules may be omitted, replaced, or added as appropriate according to the embodiment.
　＜３．動作例＞
　次に、図６を用いて、解析装置１の動作例について説明する。図６は、本実施形態に係る解析装置１の処理手順の一例を例示する。以下で説明する解析装置１の処理手順は、本発明の「解析方法」の一例である。ただし、以下で説明する処理手順は一例に過ぎず、各処理は可能な限り変更されてもよい。また、以下で説明する処理手順について、実施の形態に応じて、適宜、ステップの省略、置換、及び追加が可能である。 <3. Operation example>
Next, an operation example of the analysis device 1 will be described with reference to FIG. FIG. 6 illustrates an example of the processing procedure of the analysis device 1 according to the present embodiment. The processing procedure of the analysis device 1 described below is an example of the "analysis method" of the present invention. However, the processing procedure described below is only an example, and each processing may be changed as much as possible. Further, with respect to the processing procedure described below, steps can be omitted, replaced, and added as appropriate according to the embodiment.
　［ステップＳ１０１］
　ステップＳ１０１では、制御部１１は、データ取得部１１１として動作し、包装機３を構成する複数の機構２１の状態に関する複数件の計測データ１２４を取得する。 [Step S101]
In step S101, the control unit 11 operates as the data acquisition unit 111 and acquires a plurality of measurement data 124 relating to the states of the plurality of mechanisms 21 constituting the packaging machine 3.
　本実施形態では、制御部１１は、通信インタフェース１３を利用して、包装機３の動作を制御するように構成された制御装置（不図示）からネットワークを介して複数件の計測データ１２４を取得する。ただし、計測データ１２４を取得する経路は、このような例に限定されなくてもよい。例えば、複数件の計測データ１２４は、ＮＡＳ（Network Attached Storage）等の外部記憶装置又は他の情報処理装置に保持されていてもよい。この場合、制御部１１は、ネットワーク、記憶媒体９１等を介して、当該外部記憶装置又は他の情報処理装置から複数件の計測データ１２４を取得してもよい。また、例えば、解析装置１は、包装機３の動作を直接的に制御するように構成されていてもよい。この場合、制御部１１は、包装機３を構成する各機構２１の状態を観測するセンサから直接的に各件の計測データ１２４を取得してもよい。 In the present embodiment, the control unit 11 acquires a plurality of measurement data 124 via a network from a control device (not shown) configured to control the operation of the packaging machine 3 by using the communication interface 13. To do. However, the route for acquiring the measurement data 124 does not have to be limited to such an example. For example, a plurality of measurement data 124s may be stored in an external storage device such as NAS (Network Attached Storage) or another information processing device. In this case, the control unit 11 may acquire a plurality of measurement data 124s from the external storage device or another information processing device via the network, the storage medium 91, or the like. Further, for example, the analysis device 1 may be configured to directly control the operation of the packaging machine 3. In this case, the control unit 11 may acquire the measurement data 124 of each case directly from the sensor that observes the state of each mechanism 21 constituting the packaging machine 3.
　各件の計測データ１２４は、包装機３を構成する各機構２１の状態に関するあらゆる種類のデータを含んでよい。各件の計測データ１２４は、例えば、トルク、速度、加速度、温度、電流、電圧、空圧、圧力、流量、位置、寸法（高さ、長さ、幅）及び面積の少なくともいずれかを示すデータであってよい。このような計測データ１２４は、公知のセンサ、カメラ等の計測装置によって得ることができる。例えば、流量は、フロートセンサにより得ることができる。また、位置、寸法、及び面積は、画像センサにより得ることができる。 The measurement data 124 of each case may include all kinds of data regarding the state of each mechanism 21 constituting the packaging machine 3. The measurement data 124 of each case is, for example, data indicating at least one of torque, velocity, acceleration, temperature, current, voltage, pneumatic pressure, pressure, flow rate, position, dimension (height, length, width) and area. It may be. Such measurement data 124 can be obtained by a measuring device such as a known sensor or camera. For example, the flow rate can be obtained by a float sensor. In addition, the position, dimensions, and area can be obtained by an image sensor.
　各件の計測データ１２４は、１又は複数の計測装置から得られるデータにより構成されてもよい。また、各件の計測データ１２４は、計測装置から得られるデータそのままであってもよいし、画像データから算出される位置データ等のように計測装置から得られたデータに何らかの情報処理を適用することで取得されたデータであってもよい。各件の計測データ１２４は、各機構２１に対応して取得される。 The measurement data 124 of each case may be composed of data obtained from one or a plurality of measurement devices. Further, the measurement data 124 of each case may be the data obtained from the measurement device as it is, or some information processing is applied to the data obtained from the measurement device such as the position data calculated from the image data. It may be the data acquired by the above. The measurement data 124 of each case is acquired corresponding to each mechanism 21.
　複数件の計測データ１２４の取得が完了すると、制御部１１は、取得された各件の計測データ１２４から１又は複数の特徴量を算出する。本実施形態では、制御部１１は、各件の計測データ１２４から複数の特徴量を算出する。特徴量の種類は、特に限定されなくてもよく、実施の形態に応じて適宜選択されてよい。計測データ１２４が連続値データである場合、算出される特徴量は、例えば、フレーム内の振幅、最大値、最小値、平均値、分散値、標準偏差、瞬時値（１点サンプル）等であってよい。また、計測データ１２４が離散値データである場合には、算出される特徴量は、例えば、各フレーム内の「on」時間、「off」時間、Ｄｕｔｙ比、「on」回数、「off」回数等であってよい。算出する特徴量の数も、特に限定されなくてもよく、実施の形態に応じて適宜選択されてよい。各件の計測データ１２４から算出される特徴量の数は同じであってもよいし、異なっていてもよい。各件の計測データ１２４から複数の特徴量を算出すると、制御部１１は、次のステップＳ１０２に処理を進める。 When the acquisition of the plurality of measurement data 124 is completed, the control unit 11 calculates one or a plurality of feature quantities from the acquired measurement data 124. In the present embodiment, the control unit 11 calculates a plurality of feature quantities from the measurement data 124 of each case. The type of the feature amount does not have to be particularly limited and may be appropriately selected according to the embodiment. When the measurement data 124 is continuous value data, the calculated feature amount is, for example, the amplitude in the frame, the maximum value, the minimum value, the average value, the variance value, the standard deviation, the instantaneous value (one-point sample), or the like. You can. When the measurement data 124 is discrete value data, the calculated feature amount is, for example, the "on" time, the "off" time, the duty ratio, the "on" number of times, and the "off" number of times in each frame. And so on. The number of feature quantities to be calculated may not be particularly limited, and may be appropriately selected depending on the embodiment. The number of feature quantities calculated from the measurement data 124 of each case may be the same or different. When a plurality of feature quantities are calculated from the measurement data 124 of each case, the control unit 11 proceeds to the next step S102.
　［ステップＳ１０２］
　ステップＳ１０２では、制御部１１は、変数設定部１１２として動作し、まず、各特徴量間の相関を算出する。すなわち、制御部１１は、複数件の計測データ１２４のうちの一の計測データから算出された各特徴量と他の計測データから算出された各特徴量との間の相関を算出する。相関としては、例えば、相関係数を用いることができる。例えば、５つの特徴量Ａ～Ｅがある場合には、その５つの特徴量Ａ～Ｅの間の全ての相関係数（合計１０個の相関係数）を求め、相関が高いと判断された特徴量の組合せを、変数設定データ１２２として記憶部１２に記憶する。相関が高いとの判定は、例えば、相関係数が０．６以上、好ましくは０．７以上、さらに好ましくは０．８以上、特に好ましくは０．９以上とすることができるが、特には限定されるものではなく、種々の公知の方法で判定することができる。 [Step S102]
In step S102, the control unit 11 operates as the variable setting unit 112, and first calculates the correlation between the feature quantities. That is, the control unit 11 calculates the correlation between each feature amount calculated from the measurement data of one of the plurality of measurement data 124 and each feature amount calculated from the other measurement data. As the correlation, for example, a correlation coefficient can be used. For example, when there are five feature quantities A to E, all the correlation coefficients between the five feature quantities A to E (10 correlation coefficients in total) are obtained, and it is determined that the correlation is high. The combination of feature quantities is stored in the storage unit 12 as variable setting data 122. The determination that the correlation is high can be, for example, a correlation coefficient of 0.6 or more, preferably 0.7 or more, more preferably 0.8 or more, and particularly preferably 0.9 or more. The determination is not limited, and can be determined by various known methods.
　そして、相関係数が高い特徴量の組合せが算出されると、そのうちの一方を因果関係モデルの構築に採用し、他方を採用しないこととする。例えば、図７に示すように、５つの特徴量Ａ～Ｅのうち、特徴量Ｂ，Ｃの相関が高い場合には、特徴量Ｂのみを採用し、特徴量Ｃは採用しないこととする。したがって、因果関係モデルの構築には、特徴量Ａ，Ｂ，Ｄ，Ｅの４つを用いることとする。なお、特徴量Ｂ，Ｃのいずれを採用するかは、特には限定されず、いずれであってもよい。こうして、因果関係モデルの構築に用いる特徴量が選択されると、これを変数設定データして記憶部１２に記憶しておく。 Then, when a combination of features with a high correlation coefficient is calculated, one of them is adopted for the construction of the causal relationship model, and the other is not adopted. For example, as shown in FIG. 7, when the correlation between the feature amounts B and C is high among the five feature amounts A to E, only the feature amount B is adopted and the feature amount C is not adopted. Therefore, four feature quantities A, B, D, and E are used to construct the causal relationship model. It should be noted that which of the feature amounts B and C is adopted is not particularly limited, and may be any of them. In this way, when the feature amount used for constructing the causal relationship model is selected, it is stored in the storage unit 12 as variable setting data.
　［ステップＳ１０３］
　ステップＳ１０３では、制御部１１は、モデル構築部１１３として動作し、上述した特徴量を統計的に解析することで、複数の機構２１間の因果関係モデルを構築する。 [Step S103]
In step S103, the control unit 11 operates as the model construction unit 113, and constructs a causal relationship model between the plurality of mechanisms 21 by statistically analyzing the above-mentioned features.
　複数件の特徴量を統計的に解析する方法は、特に限定されなくてもよく、実施の形態に応じて適宜選択されてよい。当該解析方法として、例えば、ＧＬＡＳＳＯ（Graphical LASSO）、共分散選択法、ＳＧＳ（Spirtes,Glymour,and Scheines）、ＧＭ（Graphical Modeling）、ＰＣ（Peter & Clark）、ＧＥＳ（Greedy Equivalent Search）、ＦＣＩ（Fast Causal Inference）、ＬｉＮＧＡＭ（Liner Non-Gaussian Acyclic Model）、ベイジアンネットワーク等が利用されてよい。本実施形態では、制御部１１は、以下の処理手順により、複数の特徴量を統計的に解析する。 The method for statistically analyzing a plurality of feature quantities does not have to be particularly limited, and may be appropriately selected according to the embodiment. Examples of the analysis method include GLASSO (Graphical LASSO), covariance selection method, SGS (Spirtes, Glymour, and Scenes), GM (Graphical Modeling), PC (Peter & Clark), GES (Greedy Equivalent Search), and FCI ( FastCausalInference), LiNGAM (LinerNon-GaussianAcyclicModel), Bayesian network, etc. may be used. In the present embodiment, the control unit 11 statistically analyzes a plurality of feature quantities by the following processing procedure.
　＜因果関係の解析＞
　図８を更に用いて、ステップＳ１０３の処理の一例を詳細に説明する。図８は、本実施形態に係る解析装置１による因果関係解析の処理手順の一例を例示する。本実施形態に係るステップＳ１０３の処理は、以下のステップＳ２０１～ステップＳ２０２の処理を含む。ただし、以下で説明する処理手順は一例に過ぎず、各処理は可能な限り変更されてよい。また、以下で説明する処理手順について、実施の形態に応じて、適宜、ステップの省略、置換、及び追加が可能である。 <Analysis of causal relationship>
An example of the process of step S103 will be described in detail with reference to FIG. FIG. 8 illustrates an example of the processing procedure of the causal relationship analysis by the analysis device 1 according to the present embodiment. The process of step S103 according to the present embodiment includes the following processes of steps S201 to S202. However, the processing procedure described below is only an example, and each processing may be changed as much as possible. Further, with respect to the processing procedure described below, steps can be omitted, replaced, and added as appropriate according to the embodiment.
　［ステップＳ２０１］
　ステップＳ２０１では、制御部１１は、算出された各特徴量間の条件付き独立性を算出する。本実施形態では、制御部１１は、複数件の計測データ１２４のうちの一の計測データから算出された各特徴量と他の計測データから算出された各特徴量との間の条件付き独立性を算出する。条件付き独立性の種類は、特に限定されなくてもよく、実施の形態に応じて適宜選択されてよい。算出される条件付き独立性は、例えば、偏相関係数、相関係数、共分散、条件付き確率、精度行列等であってよい。本ステップＳ２０１では、制御部１１は、一種類の条件付き独立性を算出してもよいし、複数種類の条件付き独立性を算出してもよい。各特徴量間の条件付き独立性を算出すると、制御部１１は、次のステップＳ２０２に処理を進める。 [Step S201]
In step S201, the control unit 11 calculates the conditional independence between the calculated features. In the present embodiment, the control unit 11 has conditional independence between each feature amount calculated from the measurement data of one of the plurality of measurement data 124 and each feature amount calculated from the other measurement data. Is calculated. The type of conditional independence does not have to be particularly limited and may be appropriately selected depending on the embodiment. The calculated conditional independence may be, for example, a partial correlation coefficient, a correlation coefficient, a covariance, a conditional probability, an accuracy matrix, or the like. In this step S201, the control unit 11 may calculate one type of conditional independence or may calculate a plurality of types of conditional independence. After calculating the conditional independence between the features, the control unit 11 proceeds to the next step S202.
　［ステップＳ２０２］
　ステップＳ２０２では、制御部１１は、それぞれ算出された条件付き独立性に基づいて、各機構２１間の因果関係の有無を判定する。一例として、制御部１１は、算出された条件付き独立性と閾値とを比較により、因果関係の有無を判定する。因果関係を有することの程度に条件付き独立性の値が比例している場合には、制御部１１は、算出された条件付き独立性の値が閾値以上であるか否かを判定する。そして、制御部１１は、算出された条件付き独立性の値が閾値以上であるときに、該当する各機構２１間に因果関係が有ると判定し、そうではないときには、該当する各機構２１間に因果関係が無いと判定する。閾値は、適宜設定されてよい。 [Step S202]
In step S202, the control unit 11 determines whether or not there is a causal relationship between the mechanisms 21 based on the conditional independence calculated for each. As an example, the control unit 11 determines the presence or absence of a causal relationship by comparing the calculated conditional independence with the threshold value. When the value of conditional independence is proportional to the degree of having a causal relationship, the control unit 11 determines whether or not the calculated value of conditional independence is equal to or greater than the threshold value. Then, the control unit 11 determines that there is a causal relationship between the corresponding mechanisms 21 when the calculated conditional independence value is equal to or greater than the threshold value, and when it is not, the control unit 11 determines that there is a causal relationship between the corresponding mechanisms 21. Is judged to have no causal relationship. The threshold value may be set as appropriate.
　なお、条件付き独立性に基づいて因果関係の有無を判定する方法は、このような例に限定されなくてもよく、条件付き独立性の種類等に応じて適宜決定されてよい。例えば、因果関係を有することの程度と条件付き独立性との対応関係は反対であってもよい。この場合、制御部１１は、条件付き独立性の値が閾値以下である場合に、該当する各機構２１間に因果関係が有ると判定し、そうではない場合に、該当する各機構２１間に因果関係が無いと判定してもよい。また、制御部１１は、複数種類の条件付き独立性に基づいて、各機構２１間の因果関係の有無を判定してもよい。 Note that the method of determining the presence or absence of a causal relationship based on conditional independence does not have to be limited to such an example, and may be appropriately determined according to the type of conditional independence and the like. For example, the correspondence between the degree of causality and conditional independence may be opposite. In this case, the control unit 11 determines that there is a causal relationship between the corresponding mechanisms 21 when the value of the conditional independence is equal to or less than the threshold value, and when not, the control unit 11 determines that there is a causal relationship between the corresponding mechanisms 21. It may be determined that there is no causal relationship. Further, the control unit 11 may determine the presence or absence of a causal relationship between the mechanisms 21 based on a plurality of types of conditional independence.
　これらの判定の結果、制御部１１は、各機構２１間の因果関係モデルを構築することができる。例えば、上述した包装機３の場合、図９Ａ～図９Ｃに示すような因果関係モデルのグラフが構築される。 As a result of these determinations, the control unit 11 can build a causal relationship model between each mechanism 21. For example, in the case of the above-mentioned packaging machine 3, graphs of a causal relationship model as shown in FIGS. 9A to 9C are constructed.
　例えば、図９Ａは、異常として、革ベルトの摩耗が発生したときの因果関係モデルを示している。すなわち、サーボ１（サーボモータ３１１）の特徴量であるトルク平均値と位置の標準偏差が、サーボ２（サーボモータ３２２）の特徴量である速度平均値とトルク最大値に影響を与え、さらにこれらがサーボ４（サーボモータ３３５）のトルク平均値に影響を与える、という因果関係モデルが構築される。 For example, FIG. 9A shows a causal relationship model when the leather belt is worn as an abnormality. That is, the torque mean value and the standard deviation of the position, which are the feature amounts of the servo 1 (servomotor 311), affect the speed mean value and the torque maximum value, which are the feature amounts of the servo 2 (servomotor 322). A causal relationship model is constructed in which the torque of the servo 4 (servo motor 335) is affected by the torque average value.
　図９Ｂは、異常として、内容物搬送部３２のコンベア３２１のチェーンの緩みが発生したときの因果関係モデルを示している。すなわち、センサ２（ファイバセンサ３２５）の特徴量であるＯＮ時間が、センサ３（ファイバセンサ３３６）の特徴量であるターンＯＮ時間に影響を与え、さらにこれがサーボ４のトルク平均値に影響を与える、という因果関係モデルが構築される。 FIG. 9B shows a causal relationship model when the chain of the conveyor 321 of the content transporting unit 32 becomes loose as an abnormality. That is, the ON time, which is the feature amount of the sensor 2 (fiber sensor 325), affects the turn ON time, which is the feature amount of the sensor 3 (fiber sensor 336), which further affects the torque average value of the servo 4. , A causal relationship model is constructed.
　図９Ｃは、異常として、包装フィルムのシール不良が発生したときの因果関係モデルを示している。この異常については、サーボ４のトルク平均値のみが原因であるという因果関係モデルが構築され、因果関係モデルデータ１２３として、記憶部１２に記憶される。 FIG. 9C shows a causal relationship model when a sealing defect of the packaging film occurs as an abnormality. A causal relationship model is constructed in which only the average torque value of the servo 4 is the cause of this abnormality, and the causal relationship model data 123 is stored in the storage unit 12.
　こうして、各機構２１間の因果関係モデルを構築すると、制御部１１は、次のステップＳ１０４に処理を進める。 When the causal relationship model between each mechanism 21 is constructed in this way, the control unit 11 proceeds to the next step S104.
　［ステップＳ１０４］
　ステップＳ１０４では、制御部１１は、モデル再構築部１１４として動作し、上述したように、因果関係モデルの構築に採用されなかった特徴量を、生成された因果関係モデルに組み込み、因果関係モデルを再構築する。 [Step S104]
In step S104, the control unit 11 operates as the model reconstruction unit 114, incorporates the features not adopted in the construction of the causal relational model into the generated causal relational model, and incorporates the causal relational model into the generated causal relational model. Rebuild.
　例えば、図１０に示すように、５つの特徴量Ａ～Ｅについて、特徴量Ｂ，Ｃの相関が高い場合、特徴量Ｂを採用したとする。そして、４つの特徴量Ａ，Ｂ，Ｄ，Ｅについて因果関係モデルを構築する（Ａ，Ｂ，Ｄ，Ｅが、本発明に係るグループに相当する）。これに続いて、特徴量Ｃを因果関係モデルに組み込む。すなわち、特徴量Ｃは、特徴量Ｂと相関が高いため、特徴量Ｂと等価であると見なし、特徴量Ｂと同様のエッジを有する特徴量Ｃを因果関係モデルに組み込む。すなわち、特徴量Ｃを組み込み、特徴量Ｂと同様に、特徴量Ｃから特徴量Ａに向かうエッジを設けた、因果関係モデルを再構築する。こうして、再構築された因果関係モデルは、因果関係モデルデータ１２３として、記憶部１２に記憶される。 For example, as shown in FIG. 10, when the correlation between the feature quantities B and C is high for the five feature quantities A to E, it is assumed that the feature quantity B is adopted. Then, a causal relationship model is constructed for the four feature quantities A, B, D, and E (A, B, D, and E correspond to the group according to the present invention). Following this, the feature quantity C is incorporated into the causal relational model. That is, since the feature amount C has a high correlation with the feature amount B, it is considered to be equivalent to the feature amount B, and the feature amount C having an edge similar to that of the feature amount B is incorporated into the causal relationship model. That is, the causal relationship model is reconstructed by incorporating the feature amount C and providing an edge from the feature amount C toward the feature amount A in the same manner as the feature amount B. The causal relationship model reconstructed in this way is stored in the storage unit 12 as the causal relationship model data 123.
　また、例えば、図１１Ａに示すように、Ａ→Ｂ→Ｄ→Ｅのような因果関係が構築されている場合に、非採用となった特徴量Ｃを組み込む場合には、特徴量Ｂと同様に、特徴量Ａから特徴量Ｃ、及び特徴量Ｃから特徴量Ｄに向かうエッジを設ける。なお、エッジの接続方法として、例えば、図１１Ｂに示すように、図１０の変形例として、相関の高い特徴量Ｂ，Ｃを接続するエッジを設けることもできる。 Further, for example, as shown in FIG. 11A, when a causal relationship such as A → B → D → E is established and the feature amount C that has been rejected is incorporated, the same as the feature amount B. Is provided with an edge from the feature amount A to the feature amount C, and from the feature amount C to the feature amount D. As a method of connecting the edges, for example, as shown in FIG. 11B, as a modification of FIG. 10, an edge connecting the feature amounts B and C having high correlation can be provided.
　［ステップＳ１０５］
　ステップＳ１０５では、制御部１１は、出力部１１５として動作し、再構築された因果関係モデルを示す因果関係情報を出力する。 [Step S105]
In step S105, the control unit 11 operates as the output unit 115 and outputs causal relationship information indicating the reconstructed causal relationship model.
　因果関係情報の出力先及び表現形式はそれぞれ、特に限定されなくてもよく、実施の形態に応じて適宜選択されてよい。本実施形態では、因果関係情報の出力先として、表示装置１５が採用される。制御部１１は、出力処理として、因果関係情報を表示装置１５に表示させる。また、本実施形態では、因果関係情報の表現形式として、例えば、図９Ａ～図９Ｃに示すようなグラフが採用される。つまり、本実施形態では、ステップＳ１０４の出力処理は、構築された因果関係モデルを表現するグラフを生成すること、及び生成されたグラフを因果関係情報として出力することを含む。 The output destination and expression format of the causal relationship information may not be particularly limited, and may be appropriately selected according to the embodiment. In the present embodiment, the display device 15 is adopted as the output destination of the causal relationship information. The control unit 11 causes the display device 15 to display the causal relationship information as an output process. Further, in the present embodiment, as the expression format of the causal relationship information, for example, graphs as shown in FIGS. 9A to 9C are adopted. That is, in the present embodiment, the output process of step S104 includes generating a graph representing the constructed causal relationship model, and outputting the generated graph as causal relationship information.
　更に、本実施形態では、出力処理は、因果関係情報の表示形態を２つの形態の間で切り替えることを含んでいる。すなわち、本実施形態では、制御部１１は、各機構２１を項目として利用して、特定された因果関係モデルを表現する第１形態、及びそれぞれの特徴量を項目として利用して、特定された因果関係モデルを表現する第２形態を切り替えて因果関係情報を出力する。また、再構築前の因果関係モデルと、再構築後の因果関係モデルの両方、またはこれらを切り替えて表示することもできる。 Further, in the present embodiment, the output process includes switching the display form of the causal relationship information between the two forms. That is, in the present embodiment, the control unit 11 is specified by using each mechanism 21 as an item and using the first form expressing the specified causal relationship model and each feature amount as an item. The causal relationship information is output by switching the second form that expresses the causal relationship model. In addition, both the causal relationship model before reconstruction and the causal relationship model after reconstruction, or these can be switched and displayed.
　＜４．特徴＞
　以上のとおり、本実施形態に係る解析装置１によれば、次の効果を得ることができる。従来、因果関係モデルの生成時に多重共線性の問題がある場合には、相関の高い特徴量を除去した上で、因果関係モデルの構築が行われることがあったが、このようにすると、除去した特徴量に対応する機構２１が、異常の真因である場合には、これを含まない因果関係モデルが生成されるおそれがあり、異常の真因を特定できないという問題があった。 <4. Features>
As described above, according to the analysis device 1 according to the present embodiment, the following effects can be obtained. In the past, when there was a problem of multicollinearity when generating a causal relationship model, the causal relationship model was constructed after removing highly correlated features. If the mechanism 21 corresponding to the feature amount is the root cause of the abnormality, a causal relationship model that does not include the mechanism 21 may be generated, and there is a problem that the root cause of the abnormality cannot be identified.
　これに対して、本実施形態に係る解析装置１では、相関の高い特徴量のうちの一方を除去した上で、因果関係モデルを構築するが、その後、構築された因果関係モデルに対し、除去した特徴量に係る機構を組み込んで、因果関係モデルを再構築している。そのため、関連のある全ての特徴量に対応する機構を含んだ因果関係モデルを構築することができる。その結果、異常の真因を確実に特定することができる。 On the other hand, in the analysis device 1 according to the present embodiment, one of the highly correlated features is removed and then the causal relationship model is constructed. After that, the constructed causal relationship model is removed. The causal relationship model is being reconstructed by incorporating the mechanism related to the features. Therefore, it is possible to construct a causal relationship model including a mechanism corresponding to all related features. As a result, the true cause of the abnormality can be reliably identified.
　＜５．変形例＞
　以上、本発明の実施の形態を詳細に説明してきたが、前述までの説明はあらゆる点において本発明の例示に過ぎない。本発明の範囲を逸脱することなく種々の改良や変形を行うことができることは言うまでもない。例えば、以下のような変更が可能である。なお、以下では、上記実施形態と同様の構成要素に関しては同様の符号を用い、上記実施形態と同様の点については、適宜説明を省略した。以下の変形例は適宜組み合わせ可能である。 <5. Modification example>
Although the embodiments of the present invention have been described in detail above, the above description is merely an example of the present invention in all respects. Needless to say, various improvements and modifications can be made without departing from the scope of the present invention. For example, the following changes can be made. In the following, the same reference numerals will be used for the same components as those in the above embodiment, and the same points as in the above embodiment will be omitted as appropriate. The following modifications can be combined as appropriate.
　＜５－１＞
　上記実施形態では、変数設定部１１２において、相関の高い特徴量のうちの一方を採用し、他方を採用しないことで、因果関係モデルを構築しているが、変数設定部１１２では、他の処理を行うことができる。 <5-1>
In the above embodiment, the variable setting unit 112 constructs a causal relationship model by adopting one of the highly correlated features and not adopting the other, but the variable setting unit 112 performs other processing. It can be performed.
　例えば、図１２に示すように、相関の高い特徴量Ｂ，Ｃを組合せ、特徴量Ｘ（本発明に係る合成特徴量の相当）を生成した上で、特徴量Ａ，Ｘ，Ｄ，Ｅを用いて因果関係モデルを構築することができる。この場合、特徴量Ｘは、例えば、次のように生成することができる。特徴量Ａ～Ｅは、時系列のデータであるため、特徴量Ｂ，Ｃについて、同じ時間帯のデータを抽出し、これを単純に加算することで、特徴量Ｘを生成することができる。その他、例えば、特徴量Ｂ，Ｃを線形的に組み合わせることで、生成することができる。 For example, as shown in FIG. 12, feature quantities B and C with high correlation are combined to generate feature quantities X (corresponding to the synthetic feature quantities according to the present invention), and then feature quantities A, X, D, and E are generated. It can be used to build a causal relationship model. In this case, the feature amount X can be generated, for example, as follows. Since the feature quantities A to E are time-series data, the feature quantities X can be generated by extracting the data in the same time zone for the feature quantities B and C and simply adding them. In addition, for example, it can be generated by linearly combining feature quantities B and C.
　あるいは、図１３に示すように、相関の高い特徴量Ｂ，Ｃのうち、特徴量Ｂを用いて因果関係モデルを構築した後、因果関係モデルには、特徴量Ｂとともに特徴量Ｃを併記しておく。この場合、出力された因果関係モデルの特徴量Ｂに隣接して特徴量Ｃを表示することができる。すなわち、特徴量Ｂと特徴量Ｃとが等価であることが明示されるように、因果関係モデルが描画されたり、データとして記憶されていれば、「併記」となる。 Alternatively, as shown in FIG. 13, after constructing a causal relationship model using the feature amounts B among the highly correlated feature amounts B and C, the feature amount C is also written together with the feature amount B in the causal relationship model. Keep it. In this case, the feature amount C can be displayed adjacent to the feature amount B of the output causal relational model. That is, if the causal relationship model is drawn or stored as data so that it is clearly shown that the feature amount B and the feature amount C are equivalent, it is "combined writing".
　＜５－２＞
　異常が発生したときに因果関係モデルを構築しているが、異常ではない場合、つまり正常範囲で生じた変動や変位に対する因果関係モデルを構築することもできる。このような変動、変位、及び異常が、本発明の事象に相当する。 <5-2>
A causal relationship model is built when an abnormality occurs, but it is also possible to build a causal relationship model for fluctuations and displacements that occur in the normal range when it is not abnormal. Such fluctuations, displacements, and anomalies correspond to the events of the present invention.
　１…解析装置、
　１１…制御部
　１２…記憶部、
　３…包装機（生産設備） 1 ... Analytical device,
11 ... Control unit 12 ... Storage unit,
3 ... Packaging machine (production equipment)

Claims (11)

  1. 　製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析装置であって、
    　制御部と、
    　記憶部と、
    を備え、
    　前記記憶部は、複数の前記特徴量を記憶し、
    　前記制御部は、
    　前記事象に対する前記複数の特徴量の間の相関を算出するステップと、
    　前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある一方の前記特徴量を採用し、他方の前記特徴量を採用しないように、前記複数の特徴量を含むグループを生成するステップと、
    　前記グループに含まれる複数の前記特徴量の間の因果関係を示した因果関係モデルを構築するステップと、
    　採用されなかった前記他方の特徴量を含む、全ての前記特徴量の間の因果関係を構築するステップであって、前記一方の特徴量に係る因果関係を、前記他方の特徴量の因果関係にも適用するステップと、
    を実行させる、解析装置。     A production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production, and the driving means and the monitoring means can be controlled at least. It is an analysis device for events that can occur in a production facility and that has one feature quantity.
    Control unit and
    Memory and
    With
    The storage unit stores a plurality of the feature amounts,
    The control unit
    A step of calculating the correlation between the plurality of features for the event, and
    When any one of the above correlations shows a strong correlation satisfying a predetermined requirement, the feature amount of one having the strong correlation is adopted, and the feature amount of the other is not adopted. Steps to generate a group containing multiple features,
    A step of constructing a causal relationship model showing a causal relationship between a plurality of the feature quantities included in the group, and a step of constructing a causal relationship model.
    It is a step of constructing a causal relationship among all the feature amounts including the other feature amount that has not been adopted, and the causal relationship related to the one feature amount is changed to the causal relationship of the other feature amount. And the steps to apply
    An analyzer that runs.
  2. 　製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析装置であって、
    　制御部と、
    　記憶部と、
    を備え、
    　前記記憶部は、複数の前記特徴量を記憶し、
    　前記制御部は、
    　前記事象に対する前記複数の特徴量の間の相関を算出するステップと、
    　前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある一方の特徴量を採用し、他方の特徴量を採用しないように、特徴量のグループを生成するステップと、
    　前記グループに含まれる複数の前記特徴量の間の因果関係を示した因果関係モデルを構築するステップと、
    　前記因果関係モデルにおいて、前記一方の特徴量に前記他方の特徴量を併記するステップと、
    を実行させる、解析装置。     A production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production, and the driving means and the monitoring means can be controlled at least. It is an analysis device for events that can occur in a production facility and that has one feature quantity.
    Control unit and
    Memory and
    With
    The storage unit stores a plurality of the feature amounts,
    The control unit
    A step of calculating the correlation between the plurality of features for the event, and
    When any one of the above correlations shows a strong correlation that satisfies a predetermined requirement, the feature quantity of the feature quantity is adopted so that one feature quantity having the strong correlation is adopted and the other feature quantity is not adopted. Steps to generate a group and
    A step of constructing a causal relationship model showing a causal relationship between a plurality of the feature quantities included in the group, and a step of constructing a causal relationship model.
    In the causal relationship model, the step of writing the other feature amount together with the one feature amount, and
    An analyzer that runs.
  3. 　製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析装置であって、
    　制御部と、
    　記憶部と、
    を備え、
    　前記記憶部は、複数の前記特徴量を記憶し、
    　前記制御部は、
    　前記事象に対する前記複数の特徴量の間の相関を算出するステップと、
    　前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある２つの特徴量を組合せ、合成特徴量とする、ステップと、
    　複数の前記特徴量及び前記合成特徴量の間の因果関係を構築するステップと、
    　前記合成特徴量に係る因果関係を、当該合成特徴量を構成する２つの前記特徴量に適用するとともに、前記合成特徴量を削除した全ての前記特徴量の間の因果関係を示す因果関係モデルを構築するステップと、
    を実行させる、解析装置。     A production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production, and the driving means and the monitoring means can be controlled at least. It is an analysis device for events that can occur in a production facility and that has one feature quantity.
    Control unit and
    Memory and
    With
    The storage unit stores a plurality of the feature amounts,
    The control unit
    A step of calculating the correlation between the plurality of features for the event, and
    When any one of the above correlations shows a strong correlation that satisfies a predetermined requirement, the two feature quantities having the strong correlation are combined to form a composite feature quantity.
    Steps to build a causal relationship between the plurality of features and the synthetic features,
    A causal relationship model showing a causal relationship between all the feature amounts in which the synthetic feature amount is deleted while applying the causal relationship related to the synthetic feature amount to the two feature amounts constituting the synthetic feature amount is obtained. Steps to build and
    An analyzer that runs.
  4. 　前記一方の特徴量と前記他方の特徴量との間に、因果関係を形成するステップと、をさらに備えている、請求項１に記載の解析装置。 The analysis device according to claim 1, further comprising a step of forming a causal relationship between the one feature amount and the other feature amount.
  5. 　前記合成特徴量を構成する２つの特徴量の間に、因果関係を形成するステップと、をさらに備えている、請求項３に記載の解析装置。 The analysis device according to claim 3, further comprising a step of forming a causal relationship between the two feature quantities constituting the synthetic feature quantity.
  6. 　製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析方法であって、
    　前記生産設備に生じ得る事象に対する前記複数の特徴量の間の相関を算出するステップと、
    　前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある一方の前記特徴量を採用し、他方の前記特徴量を採用しないように、前記複数の特徴量を含むグループを生成するステップと、
    　前記グループに含まれる複数の前記特徴量の間の因果関係を示した因果関係モデルを構築するステップと、
    　採用されなかった前記他方の特徴量を含む、全ての前記特徴量の間の因果関係を構築するステップであって、前記一方の特徴量に係る因果関係を、前記他方の特徴量の因果関係にも適用するステップと、
    を備えている、解析方法。     A production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production, and the driving means and the monitoring means can be controlled at least. It is a method of analyzing an event that is provided in a production facility and can occur in the production facility, which has one characteristic quantity.
    A step of calculating the correlation between the plurality of features with respect to an event that may occur in the production facility, and
    When any one of the above correlations shows a strong correlation satisfying a predetermined requirement, the feature amount of one having the strong correlation is adopted, and the feature amount of the other is not adopted. Steps to generate a group containing multiple features,
    A step of constructing a causal relationship model showing a causal relationship between a plurality of the feature quantities included in the group, and a step of constructing a causal relationship model.
    It is a step of constructing a causal relationship among all the feature amounts including the other feature amount that has not been adopted, and the causal relationship related to the one feature amount is changed to the causal relationship of the other feature amount. And the steps to apply
    The analysis method is equipped with.
  7. 　製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析方法であって、
    　前記事象に対する前記複数の特徴量の間の相関を算出するステップと、
    　前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある一方の特徴量を採用し、他方の特徴量を採用しないように、特徴量のグループを生成するステップと、
    　前記グループに含まれる複数の前記特徴量の間の因果関係を示した因果関係モデルを構築するステップと、
    　前記因果関係モデルにおいて、前記一方の特徴量に前記他方の特徴量を併記するステップと、
    を備えている、解析方法。     A production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production, and the driving means and the monitoring means can be controlled at least. It is a method of analyzing an event that is provided in a production facility and can occur in the production facility, which has one characteristic quantity.
    A step of calculating the correlation between the plurality of features for the event, and
    When any one of the above correlations shows a strong correlation that satisfies a predetermined requirement, the feature quantity of the feature quantity is adopted so that one feature quantity having the strong correlation is adopted and the other feature quantity is not adopted. Steps to generate a group and
    A step of constructing a causal relationship model showing a causal relationship between a plurality of the feature quantities included in the group, and a step of constructing a causal relationship model.
    In the causal relationship model, the step of writing the other feature amount together with the one feature amount, and
    The analysis method is equipped with.
  8. 　製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析方法であって、
    　前記事象に対する前記複数の特徴量の間の相関を算出するステップと、
    　前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある２つの特徴量を組合せ、合成特徴量とする、ステップと、
    　複数の前記特徴量及び前記合成特徴量の間の因果関係を構築するステップと、
    　前記合成特徴量に係る因果関係を、当該合成特徴量を構成する２つの前記特徴量に適用するとともに、前記合成特徴量を削除した全ての前記特徴量の間の因果関係を示す因果関係モデルを構築するステップと、
    を備えている、解析方法。     A production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production, and the driving means and the monitoring means can be controlled at least. It is a method of analyzing an event that is provided in a production facility and can occur in the production facility, which has one characteristic quantity.
    A step of calculating the correlation between the plurality of features for the event, and
    When any one of the above correlations shows a strong correlation that satisfies a predetermined requirement, the two feature quantities having the strong correlation are combined to form a composite feature quantity.
    Steps to build a causal relationship between the plurality of features and the synthetic features,
    A causal relationship model showing a causal relationship between all the feature amounts in which the synthetic feature amount is deleted while applying the causal relationship related to the synthetic feature amount to the two feature amounts constituting the synthetic feature amount is obtained. Steps to build and
    The analysis method is equipped with.
  9. 　製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析プログラムであって、
    　コンピュータに、
    　前記生産設備に生じ得る事象に対する前記複数の特徴量の間の相関を算出するステップと、
    　前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある一方の前記特徴量を採用し、他方の前記特徴量を採用しないように、前記複数の特徴量を含むグループを生成するステップと、
    　前記グループに含まれる複数の前記特徴量の間の因果関係を示した因果関係モデルを構築するステップと、
    　採用されなかった前記他方の特徴量を含む、全ての前記特徴量の間の因果関係を構築するステップであって、前記一方の特徴量に係る因果関係を、前記他方の特徴量の因果関係にも適用するステップと、
    を実行させる、解析プログラム。     A production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production, and the driving means and the monitoring means can be controlled at least. It is an analysis program of an event that is installed in a production facility and can occur in the production facility, and has one feature quantity.
    On the computer
    A step of calculating the correlation between the plurality of features with respect to an event that may occur in the production facility, and
    When any one of the above correlations shows a strong correlation satisfying a predetermined requirement, the feature amount of one having the strong correlation is adopted, and the feature amount of the other is not adopted. Steps to generate a group containing multiple features,
    A step of constructing a causal relationship model showing a causal relationship between a plurality of the feature quantities included in the group, and a step of constructing a causal relationship model.
    It is a step of constructing a causal relationship among all the feature amounts including the other feature amount that has not been adopted, and the causal relationship related to the one feature amount is changed to the causal relationship of the other feature amount. And the steps to apply
    An analysis program that executes.
  10. 　製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析プログラムであって、
    　コンピュータに、
    　前記事象に対する前記複数の特徴量の間の相関を算出するステップと、
    　前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある一方の特徴量を採用し、他方の特徴量を採用しないように、特徴量のグループを生成するステップと、
    　前記グループに含まれる複数の前記特徴量の間の因果関係を示した因果関係モデルを構築するステップと、
    　前記因果関係モデルにおいて、前記一方の特徴量に前記他方の特徴量を併記するステップと、
    を実行させる、解析プログラム。     A production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production, and the driving means and the monitoring means can be controlled at least. It is an analysis program of an event that is installed in a production facility and can occur in the production facility, and has one feature quantity.
    On the computer
    A step of calculating the correlation between the plurality of features for the event, and
    When any one of the above correlations shows a strong correlation that satisfies a predetermined requirement, the feature quantity of the feature quantity is adopted so that one feature quantity having the strong correlation is adopted and the other feature quantity is not adopted. Steps to generate a group and
    A step of constructing a causal relationship model showing a causal relationship between a plurality of the feature quantities included in the group, and a step of constructing a causal relationship model.
    In the causal relationship model, the step of writing the other feature amount together with the one feature amount, and
    An analysis program that executes.
  11. 　製品を生産する生産設備であって、当該生産設備の駆動を行う少なくとも１つの駆動手段及び前記生産の監視を行う少なくとも１つの監視手段を有し、前記駆動手段及び監視手段が、制御可能な少なくとも１つの特徴量を有している、生産設備に設けられ、当該生産設備で生じ得る事象の解析プログラムであって、
    　コンピュータに、
    　前記事象に対する前記複数の特徴量の間の相関を算出するステップと、
    　前記相関のうちのいずれか１つが、所定の要件を充足する強い相関を示した場合、当該強い相関関係にある２つの特徴量を組合せ、合成特徴量とする、ステップと、
    　複数の前記特徴量及び前記合成特徴量の間の因果関係を構築するステップと、
    　前記合成特徴量に係る因果関係を、当該合成特徴量を構成する２つの前記特徴量に適用するとともに、前記合成特徴量を削除した全ての前記特徴量の間の因果関係を示す因果関係モデルを構築するステップと、
    を実行させる、解析プログラム。     A production facility for producing a product, which has at least one driving means for driving the production facility and at least one monitoring means for monitoring the production, and the driving means and the monitoring means can be controlled at least. It is an analysis program of an event that is installed in a production facility and can occur in the production facility, and has one feature quantity.
    On the computer
    A step of calculating the correlation between the plurality of features for the event, and
    When any one of the above correlations shows a strong correlation that satisfies a predetermined requirement, the two feature quantities having the strong correlation are combined to form a composite feature quantity.
    Steps to build a causal relationship between the plurality of features and the synthetic features,
    A causal relationship model showing a causal relationship between all the feature amounts in which the synthetic feature amount is deleted while applying the causal relationship related to the synthetic feature amount to the two feature amounts constituting the synthetic feature amount is obtained. Steps to build and
    An analysis program that executes.
PCT/JP2019/036697 2019-09-19 2019-09-19 Analysis device for event that can occur in production facility WO2021053782A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/JP2019/036697 WO2021053782A1 (en) 2019-09-19 2019-09-19 Analysis device for event that can occur in production facility

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/JP2019/036697 WO2021053782A1 (en) 2019-09-19 2019-09-19 Analysis device for event that can occur in production facility

Publications (1)

Publication Number Publication Date
WO2021053782A1 true WO2021053782A1 (en) 2021-03-25

Family

ID=74884433

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2019/036697 WO2021053782A1 (en) 2019-09-19 2019-09-19 Analysis device for event that can occur in production facility

Country Status (1)

Country Link
WO (1) WO2021053782A1 (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007207101A (en) * 2006-02-03 2007-08-16 Infocom Corp Graph generation method, graph generation program, and data mining system
JP2017054432A (en) * 2015-09-11 2017-03-16 日本電信電話株式会社 Data analyzer, data analysis method, and data analysis processing program
WO2017122340A1 (en) * 2016-01-15 2017-07-20 三菱電機株式会社 Plan generating device, plan generating method, and plan generating program
JP2018116545A (en) * 2017-01-19 2018-07-26 オムロン株式会社 Prediction model creating device, production facility monitoring system, and production facility monitoring method
JP2018151883A (en) * 2017-03-13 2018-09-27 株式会社東芝 Analysis device, analysis method, and program

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007207101A (en) * 2006-02-03 2007-08-16 Infocom Corp Graph generation method, graph generation program, and data mining system
JP2017054432A (en) * 2015-09-11 2017-03-16 日本電信電話株式会社 Data analyzer, data analysis method, and data analysis processing program
WO2017122340A1 (en) * 2016-01-15 2017-07-20 三菱電機株式会社 Plan generating device, plan generating method, and plan generating program
JP2018116545A (en) * 2017-01-19 2018-07-26 オムロン株式会社 Prediction model creating device, production facility monitoring system, and production facility monitoring method
JP2018151883A (en) * 2017-03-13 2018-09-27 株式会社東芝 Analysis device, analysis method, and program

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
KIKUCHI MINORU ET AL.: "A moving target detection system from infrared image using genetic algorithms", THE TRANSACTIONS OF THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS, vol. J84-D-II, no. 10, October 2001 (2001-10-01), pages 2224 - 2233 *

Similar Documents

Publication Publication Date Title
US11106197B2 (en) Prediction model creation apparatus, production facility monitoring system, and production facility monitoring method
WO2020183973A1 (en) Display system, display method, and display program
EP3023338B1 (en) Program library and control apparatus
WO2020183975A1 (en) Display system
WO2021053782A1 (en) Analysis device for event that can occur in production facility
WO2022030567A1 (en) Display system, display method, and display program
JP7006640B2 (en) Control devices, control methods, and control programs
CN108027263A (en) Ultrasound Instrument for the gas for measuring reduced size
WO2020183974A1 (en) Display system
JP2022114330A (en) Anomaly detection system, anomaly detection method and anomaly detection program
WO2021011397A1 (en) Machine learning for splice improvement
CN112513761B (en) Display system, display method, and recording medium for display program
WO2022176760A1 (en) Model generation device, model generation method, model generation program, estimation device, estimation method, and estimation program
WO2021044500A1 (en) Abnormality processing support device
WO2021048783A1 (en) A procedure to solve an unknown fault of at least part of an automatic machine for manufacturing or packing consumer products
JP2021196999A (en) Control device, host controller, control method, and control program
EP4286822A1 (en) Virtual sensing system for condition monitoring of a container packaging machine
WO2022209057A1 (en) Information processing device, information processing method, and information processing program
WO2022190552A1 (en) Information processing device, information processing method, and information processing program
JP2016084176A (en) Control device, system and library program
CN116457798A (en) Process analysis device, process analysis method, and process analysis program
Shui et al. Roll-to-Roll Manufacturing System Modeling and Analysis by Stream of Variation Theory

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19945712

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19945712

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: JP