Information processing device, information processing method, and information processing program

Abstract

To provide an information processor that can estimate a cause of an abnormality using a simple method.SOLUTION: An information processor comprises a causal model generation part that generates multiple causal models among variables based on a relationship between multiple mechanisms in a process carried out on a manufacturing line, dependence among multiple events related to multiple mechanisms, or control relationships of multiple mechanisms, an abnormality detection part that obtains an abnormality detection result of the manufacturing line, a contribution rate calculation part that calculates a contribution rate of the variable that contributes to abnormality among abnormality detection results, and an estimation part that estimates a cause of an abnormality for at least one of several causal models based on the calculated contribution rate of the variable contributing to the abnormality.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] The present disclosure relates to an information processing device, an information processing method, and an information processing program. [Background technology]

[0002] A production line in a factory, etc., is made up of multiple mechanisms such as conveyors and robot arms. If an abnormality occurs in any of the mechanisms in this production line, production of the product will stop, which could result in significant damage. For this reason, in factories, etc., maintenance personnel regularly patrol the production line to check for any abnormalities or signs of abnormalities.

[0003] When an abnormality or its precursor is detected in a production line, the true cause of the abnormality may lie in a mechanism preceding the mechanism where the abnormality was detected. Therefore, in order to identify the true cause of the abnormality, it is important to understand the causal relationships among each mechanism in the production line. However, as the number of mechanisms that make up a production line increases and the operating conditions of each mechanism can change daily, it is difficult to accurately understand the causal relationships among all mechanisms.

[0004] Therefore, in the past, experienced maintenance personnel relied on their own experience and intuition to understand the causal relationships between the multiple mechanisms that make up the production line and detect abnormalities or signs of abnormalities that occurred within the production line.To enable unskilled maintenance personnel to perform this type of maintenance work, there was a need for the development of technology that visualizes the causal relationships between the multiple mechanisms that make up the production line.

[0005] For example, Patent Document 1 discloses a device that provides a technology for visualizing the detection of abnormalities by easily modeling the relationships between devices in a manufacturing line. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Laid-Open No. 2009-64407 Summary of the Invention [Problem to be solved by the invention]

[0007] However, there are many different types of anomalies, and identifying the cause of an anomaly often requires a large amount of anomaly data, making it difficult to identify an anomaly easily.

[0008] The present disclosure has been made to solve the above-mentioned problems, and aims to provide an information processing device, an information processing method, and an information processing program that are capable of estimating the cause of an abnormality in a simple manner. [Means for solving the problem]

[0009] According to an example of the present disclosure, an information processing device includes a causal model generation unit that generates a plurality of causal models between variables based on the relationships between a plurality of mechanisms in processes performed on a production line, the dependencies between a plurality of events related to the plurality of mechanisms, or the control relationships between the plurality of mechanisms, an anomaly detection unit that acquires anomaly detection results for the production line, a contribution rate calculation unit that calculates the contribution rate of variables that contributed to the anomaly among the anomaly detection results, and an estimation unit that estimates the cause of the anomaly for at least one of the plurality of causal models based on the calculated contribution rate of the variables that contributed to the anomaly. With this configuration, it is possible to estimate the cause of the anomaly using a simple method.

[0010] The information processing device includes a display unit that displays an estimation result of the cause of the abnormality for at least one of the plurality of causal models based on the estimation result of the estimation unit. With this configuration, it is possible to easily estimate the cause of the abnormality from the estimation result displayed on the display unit.

[0011] The information processing device further includes a receiving unit that receives an input of a selection of multiple causal models from a user. The display unit displays an estimation result of the cause of the abnormality for a causal model selected from the multiple causal models based on the estimation result of the estimation unit in accordance with the reception of the input of the selection of the multiple causal models by the receiving unit. Since a causal model can be selected according to the user's selection, it is possible to estimate the cause of the abnormality using a causal model that suits the user's preferences.

[0012] The information processing device further includes an event relation registering unit that registers an event list corresponding to an abnormal event. By registering the event list, it is possible to easily estimate the cause of the abnormality from past data.

[0013] The event list includes an event summary, event data when the event occurred, the estimation result of the estimation unit, confirmed events related to multiple mechanisms, details of the event work, and information on the contribution rate of variables that contributed to the abnormality. By registering various data in the event list, it is possible to easily estimate the cause of the abnormality.

[0014] The information processing device further includes an event estimation unit that calculates a similarity between the event list registered by the event relation registration unit and the contribution rate of the variable that contributed to the abnormality calculated by the contribution rate calculation unit, and it is possible to easily estimate the cause of the abnormality from the similarity with the past event list.

[0015] According to an example of the present disclosure, an information processing method includes the steps of: generating a plurality of causal models between variables based on the relationships between a plurality of mechanisms in a process performed on a production line, the dependencies between a plurality of events related to the plurality of mechanisms, or the control relationships between the plurality of mechanisms; acquiring anomaly detection results for the production line; calculating a contribution rate of a variable that contributed to the anomaly among the anomaly detection results; and estimating the cause of the anomaly for at least one of the plurality of causal models based on the calculated contribution rate of the variable that contributed to the anomaly. With this configuration, it is possible to estimate the cause of the anomaly in a simple manner.

[0016] According to an example of the present disclosure, an information processing program causes a computer to execute the steps of: generating a plurality of causal models between variables based on relationships between a plurality of mechanisms in processes performed on a production line, dependencies between a plurality of events related to the plurality of mechanisms, or control relationships between the plurality of mechanisms; acquiring anomaly detection results for the production line; calculating contribution rates of variables that contributed to the anomaly among the anomaly detection results; and estimating the cause of the anomaly for at least one of the plurality of causal models based on the calculated contribution rates of the variables that contributed to the anomaly. With this configuration, it is possible to estimate the cause of the anomaly in a simple manner. [Effects of the Invention]

[0017] According to the present disclosure, it is possible to estimate the cause of an abnormality using a simple method. [Brief explanation of the drawings]

[0018] [Figure 1] 1 is a diagram schematically illustrating an example of a usage scene of a process analysis device 1 that is one form of an information processing device according to an embodiment. [Figure 2] FIG. 1 is a diagram illustrating an example of a hardware configuration of a process analysis device 1 according to an embodiment. [Figure 3] FIG. 2 is a diagram illustrating an example of a hardware configuration of a PLC 2 according to an embodiment. [Figure 4] FIG. 2 is a diagram schematically illustrating an example of a software configuration of the process analysis device 1 according to the embodiment. [Figure 5] 10 is a diagram illustrating a process of a first relationship identifying unit 114 of the process analysis device 1 according to the embodiment. FIG. [Figure 6] FIG. 10 is another diagram illustrating the processing of the first relationship identifying unit 114 of the process analysis device 1 according to the embodiment. [Figure 7] FIG. 10 is a flowchart illustrating processing of a second relationship identifying unit 115 of the process analysis device 1 according to the embodiment. [Figure 8] FIG. 2 is a conceptual diagram showing a structure of a program according to an embodiment expressed in the form of a tree structure. [Figure 9]FIG. 10 is a diagram illustrating an example of a control causal relationship according to an embodiment. [Figure 10] FIG. 10 is a diagram illustrating a process of a third relationship identifying unit 116 of the process analysis device 1 according to the embodiment. [Figure 11] FIG. 10 is a flow diagram illustrating processing in a data causal model generation unit 142 according to the embodiment. [Figure 12] 10A to 10C are diagrams illustrating the creation of undirected graph information based on state data according to an embodiment. [Figure 13] 10 is a diagram illustrating a control causal relationship based on status data 223 by a third relationship identifying unit 116 in a manufacturing process performed by a manufacturing line 3 according to an embodiment. FIG. [Figure 14] 10A and 10B are diagrams illustrating detection of an abnormality in an abnormality detection unit 117 according to the embodiment. [Figure 15] FIG. 10 is another diagram illustrating detection of anomaly in anomaly detection unit 117 according to the embodiment. [Figure 16] 10A to 10C are diagrams illustrating estimation of the cause of an abnormality by an estimation unit 119 according to the embodiment. [Figure 17] FIG. 10 is another diagram illustrating estimation of the cause of abnormality by estimation unit 119 according to the embodiment. [Figure 18] FIG. 10 is yet another diagram illustrating estimation of the cause of abnormality by estimation unit 119 according to the embodiment. [Figure 19] FIG. 10 is a diagram illustrating an event list in an event relation registration unit 120 according to the embodiment. [Figure 20] FIG. 10 is a diagram illustrating an event estimation process using an event list in an event estimation unit 121 according to an embodiment. [Figure 21] FIG. 10 is a flowchart of an event estimation process using an event list in an event estimation unit 121 according to an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0019] The embodiments will be described in detail with reference to the drawings. The same or corresponding parts in the drawings will be denoted by the same reference numerals, and the description thereof will not be repeated. The modifications described below may be combined selectively as appropriate.

[0020] Hereinafter, an embodiment of a process analysis device, which is one form of an information processing device, will be described with reference to the drawings.

[0021] <1. Application Examples> FIG. 1 is a diagram illustrating a typical example of a usage scenario of a process analysis device 1, which is an information processing device according to an embodiment. As illustrated in FIG. 1, the process analysis device 1 according to an embodiment acquires status data 223 related to the status of multiple mechanisms 31 constituting a production line 3. The production line 3 may be configured with multiple devices or a single device, such as a packaging machine, as long as it is capable of manufacturing some kind of product. Each mechanism 31 may be configured with one or more devices, or may be configured as part of a device as long as it is capable of performing some process in the production process. When one mechanism 31 is configured as part of a device, the multiple mechanisms 31 may be configured as a single device. When the same device performs multiple processes, each may be considered as a separate mechanism 31. For example, when the same device performs a first process and a second process, the device performing the first process may be considered as a first mechanism 31, and the device performing the second process may be considered as a second mechanism 31. The status data 223 may include any type of data related to the status of each mechanism 31 constituting the production line 3. Each mechanism 31 may be composed of a device or part of a device such as a conveyor, a robot arm, a servo motor, a cylinder (such as a molding machine), a suction pad, a cutting device, or a sealing device. Each mechanism 31 may also be a composite device such as a printer, a mounting machine, a reflow oven, or a circuit board inspection device. In addition to devices that perform physical operations such as those described above, each mechanism 31 may also include devices that perform internal processing, such as devices that detect information using various sensors, devices that acquire data from various sensors, devices that detect information from the acquired data, and devices that process the acquired data. As a specific example, in a production line equipped with an optical sensor that detects marks on objects moving along a conveyor, the optical sensor and devices that use the information detected by the optical sensor may be treated as each mechanism 31. Furthermore, the status data 223 for each item may be data indicating at least one of torque, speed, acceleration, temperature, current, voltage, air pressure, pressure, flow rate, position, dimensions (height, length, width), and area. Such status data 223 can be obtained using known measuring devices such as sensors and cameras. For example, the flow rate can be obtained by a float sensor.Additionally, the position, size and area can be obtained by an image sensor.

[0022] The status data 223 may be composed of data obtained from one or more measuring devices. The status data 223 may be the data obtained from the measuring device itself, or may be data obtainable by applying some processing to data obtained from the measuring device, such as position data obtained from image data. The status data 223 for each item is obtained corresponding to each mechanism 31. Each measuring device is appropriately positioned so as to be able to monitor each mechanism 31 of the production line 3. The PLC 2 operates the production line 3 and collects the status data 223 for each item from each measuring device. The control unit 11 obtains the status data 223 from the PLC 2, relating to the state of each mechanism 31 when the production line 3 is operating normally.

[0023] The process analysis device 1 according to the embodiment also acquires a control program 222 for controlling the operation of the production line 3. The control program 222 may include any type of program that controls the operation of each mechanism 31 constituting the production line 3. The control program 222 may be composed of one program or multiple programs. The control program 222 may be written in at least one of a ladder diagram language, a function block diagram language, a structured text language, an instruction list language, a sequential function chart language, and the C language so that it can be executed by the PLC 2. In the embodiment, the operation of the production line 3 is controlled by a programmable logic controller (PLC) 2. The process analysis device 1 acquires multiple pieces of status data 223 and the control program 222 from the PLC 2.

[0024] The process analysis device 1 identifies a causal relationship (first causal relationship) between a plurality of mechanisms 31 in a process executed in the manufacturing line 3. As an example, the first causal relationship is also referred to as a design causal model.

[0025] The process analysis device 1 analyzes the control program 222 and identifies a causal relationship (second causal relationship) between the multiple mechanisms 31 based on the control program 222. As an example, the second causal relationship is also referred to as a control causal model.

[0026] The process analysis device 1 analyzes the acquired state data 223 to identify a causal relationship (third causal relationship) between the multiple mechanisms 31. As an example, the third causal relationship is also referred to as a data causal model.

[0027] The process analysis device 1 according to the embodiment estimates the cause of the abnormality n based on each causal model and the abnormality detection result.

[0028] <2. Hardware configuration of the process analysis device> Fig. 2 is a diagram schematically illustrating an example of a hardware configuration of a process analysis apparatus 1 according to an embodiment. As shown in Fig. 2, the process analysis apparatus 1 according to the embodiment is a computer to which a control unit 11, a storage unit 12, a communication interface 13, an input device 14, an output device 15, and a drive 16 are electrically connected. In Fig. 2, the communication interface is referred to as a "communication I / F."

[0029] The control unit 11 includes a hardware processor such as a CPU (Central Processing Unit), RAM (Random Access Memory), and ROM (Read Only Memory), and controls each component in accordance with information processing. The storage unit 12 is an auxiliary storage device such as a hard disk drive or solid state drive, and stores the process analysis program 17 executed by the control unit 11, etc.

[0030] The process analysis program 17 is a program for causing the process analysis device 1 to execute a process of analyzing the causal relationships between multiple mechanisms 31 in the manufacturing process carried out by the manufacturing line 3, using information about the process executed on the manufacturing line 3, status data 223, control program 222, etc.

[0031] 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 process analysis device 1 can perform data communication via the network with the PLC 2 using this communication interface 13. 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, or the like.

[0032] The input device 14 is a device for inputting data, such as a mouse or a keyboard. The output device 15 is a device for outputting data, such as a display or a speaker. An operator can operate the process analysis apparatus 1 via the input device 14 and the output device 15.

[0033] 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 selected appropriately depending on the type of the storage medium 91. The process analysis program 17 may be stored in this storage medium 91.

[0034] The storage medium 91 is a medium that stores information such as a program by electrical, magnetic, optical, mechanical, or chemical action so that the information can be read by a computer or other device, machine, etc. The process analysis device 1 may acquire the process analysis program 17 from the storage medium 91.

[0035] 2 illustrates a disk-type storage medium such as a CD or DVD as an example of the storage medium 91. However, the type of 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.

[0036] It should be noted that, with regard to the specific hardware configuration of the process analysis device 1, components can be omitted, replaced, or added as appropriate depending on the embodiment. For example, the control unit 11 may include multiple processors. The process analysis device 1 may be configured with multiple information processing devices. Furthermore, the process analysis device 1 may be an information processing device designed specifically for the services provided, as well as a general-purpose server device, a PC (Personal Computer), etc.

[0037] <3.PLC> Fig. 3 is a diagram illustrating an example of a hardware configuration of a PLC 2 according to an embodiment. As shown in Fig. 3, the PLC 2 is a computer to which a control unit 21, a storage unit 22, an input / output interface 23, and a communication interface 24 are electrically connected. As a result, the PLC 2 is configured to control the operation of each mechanism 31 of the production line 3. In Fig. 3, the input / output interface and the communication interface are referred to as an "input / output I / F" and a "communication I / F," respectively.

[0038] The control unit 21 includes a CPU, RAM, ROM, etc., and is configured to execute various information processes based on programs and data. The storage unit 22 is configured, for example, with RAM, ROM, etc., and stores a control program 222, status data 223, etc. The control program 222 is a program for controlling the operation of the production line 3. The status data 223 is data related to the status of each mechanism 31. The input / output interface 23 is an interface for connecting to an external device and is configured appropriately depending on the external device to be connected. In the embodiment, the PLC 2 is connected to the production line 3 via the input / output interface 23. Note that, when different status data can be acquired for a single device, the target single device may be regarded as multiple mechanisms 31 or as a single mechanism 31. Therefore, the number of input / output interfaces 23 may be the same as the number of mechanisms 31 constituting the production line 3, or may be different from the number of mechanisms 31 constituting the production line 3.

[0039] The communication interface 24 is, for example, a wired LAN module, a wireless LAN module, etc., and is an interface for performing wired or wireless communication. The PLC 2 can perform data communication with the process analysis device 1 via the communication interface 24.

[0040] It should be noted that the specific hardware configuration of PLC 2 can be omitted, replaced, or added as appropriate depending on the embodiment. For example, control unit 21 may include multiple processors. Storage unit 22 may be configured with RAM and ROM included in control unit 21. Storage unit 22 may be configured with an auxiliary storage device such as a hard disk drive or solid state drive. Furthermore, PLC 2 may be replaced with an information processing device designed specifically for the services provided, or with a general-purpose desktop PC, tablet PC, or the like depending on the object to be controlled.

[0041] <4. Software configuration of the process analysis device> 4 is a diagram schematically illustrating an example of the software configuration of the process analysis apparatus 1 according to the embodiment. As shown in FIG. 4, the control unit 11 of the process analysis apparatus 1 loads the process analysis program 17 stored in the storage unit 12 into the RAM. The control unit 11 then interprets and executes the process analysis program 17 loaded into the RAM using the CPU to control each component. As a result, as shown in FIG. 4, the process analysis apparatus 1 according to the embodiment includes, as software modules, a first acquisition unit 111, a second acquisition unit 112, a third acquisition unit 113, a first relationship identification unit 114, a second relationship identification unit 115, a third relationship identification unit 116, an anomaly detection unit 117, a contribution rate calculation unit 118, an estimation unit 119, an event relationship registration unit 120, an estimation unit 119, and a display unit 122.

[0042] The first acquisition unit 111 acquires process data 220 and device installation position data 221 as information relating to processes executed on the production line 3. The second acquisition unit 112 acquires a control program 222 for controlling the operation of the production line 3. The third acquisition unit 113 acquires status data 223. The process data 220 and device installation position data 221 as information relating to processes executed on the production line 3 acquired by the first acquisition unit 111 may be acquired from the PLC 2. Alternatively, the data may be input by the user, or may be acquired from information previously stored in a storage unit.

[0043] The first relationship identifying unit 114 identifies a causal relationship (first causal relationship) between multiple mechanisms 31 in the processes executed in the manufacturing line 3 based on the process data 220 and device installation position data 221 acquired by the first acquiring unit 111.

[0044] The second relationship identifying unit 115 analyzes the control program 222 acquired by the second acquiring unit 112, and identifies a causal relationship (second causal relationship) between the plurality of mechanisms 31 based on the control program 222.

[0045] The third relationship identifying unit 116 analyzes the state data 223 acquired by the third acquiring unit 113 to identify a causal relationship between the plurality of mechanisms 31 (third causal relationship).

[0046] The anomaly detection unit 117 detects anomalies in the processes executed on the production line 3. Specifically, it calculates an anomaly score and determines whether the anomaly score exceeds a predetermined threshold. The anomaly detection unit 117 determines an anomaly when the anomaly score exceeds the predetermined threshold. In this example, the anomaly score indicates the degree of deviation from the normal range.

[0047] The contribution rate calculation unit 118 calculates the contribution rate of the variable that contributed to the abnormality score based on the abnormality score.

[0048] The estimation unit 119 acquires a plurality of causal models from each of the first relationship identification unit 114 to the third relationship identification unit 116. The estimation unit 119 acquires the contribution rates of the variables that contributed to the anomaly score calculated by the contribution rate calculation unit 118. Based on the acquired information on the contribution rates of the variables, the estimation unit 119 estimates the cause of the anomaly in at least one of the plurality of causal models, and outputs the result to the display unit 122. When an anomaly is detected, the user can easily identify candidate causes and quickly perform recovery work.

[0049] The event relation registration unit 120 registers an event list corresponding to an event. The event list includes an event summary, event data when the event occurred, the estimation result of the estimation unit 119, confirmed events related to multiple mechanisms, the event work content, and information on the contribution rate of variables that contributed to the abnormality.

[0050] The event estimation unit 121 acquires the event list registered by the event relation registration unit 120. The event estimation unit 121 acquires the contribution rates of the variables that contributed to the anomaly score calculated by the contribution rate calculation unit 118, extracts an event list with a high similarity in contribution rate, and outputs the extracted event list to the display unit 122. When an anomaly is detected, the user can easily identify a candidate event list and quickly perform recovery work.

[0051] Each software module of the process analysis apparatus 1 will be described in detail in an operation example below. In the embodiment, an example is described in which each software module of the process analysis apparatus 1 is implemented by a general-purpose CPU. However, some or all of the above software modules may be implemented by one or more dedicated hardware processors. Furthermore, with regard to the software configuration of the process analysis apparatus 1, software modules may be omitted, replaced, or added as appropriate depending on the embodiment.

[0052] <5.1 First causal relationship> 5 is a diagram illustrating the processing of the first relationship identifying unit 114 of the process analysis apparatus 1 according to the embodiment. However, the processing method described below is merely an example, and various methods are possible.

[0053] Referring to FIG. 5(A), process data for materials MA and MB is shown. As the process data, the process sequence of the material MA and the variable information of the mechanism for the material MA are set.

[0054] In this example, as an example, the conveyor axis (ProductFeed), film transport sub-axis (FilmFeedSub), and top seal axis (TopSeal) are set as the process order for the material MA.

[0055] As the process data, the process sequence of the material MB and the variable information of the mechanism for the material MB are set.

[0056] In this example, as an example, the process order of the material MB is set as follows: film feed main shaft (FilmFeedMain), film feed sub shaft (FilmFeedSub), and top seal shaft (TopSeal).

[0057] FIG. 5(B) is a diagram for explaining the process causal relationships based on the process data of materials MA and MB.

[0058] As shown in the figure, nodes corresponding to the conveyor axis (ProductFeed), film transport sub-axis (FilmFeedSub), top seal axis (TopSeal), and film transport main axis (FilmFeedMain) are provided.

[0059] Then, a design causal relationship (design causal model) is generated in which a vector from the film transport main axis (FilmFeedMain) to the film transport sub axis (FilmFeedSub), a vector from the film transport sub axis (FilmFeedSub) to the top seal axis (TopSeal), and a vector from the conveyor axis (ProductFeed) to the film transport sub axis (FilmFeedSub) are set.

[0060] 6 is another diagram illustrating the processing of the first relationship identifying unit 114 of the process analysis apparatus 1 according to the embodiment. However, the processing method described below is merely an example, and various methods are possible.

[0061] Referring to FIG. 6(A), device installation location data is shown. The device installation location data is set in association with the device name and information about the location where the device is installed and the direction indicating the relationship.

[0062] In this example, as an example, a work detection sensor (ProductDetectionSensorSignal) is set for the conveyor axis (ProductFeed), and vector information relating to the right direction is set.

[0063] A workpiece overhang sensor (ProductOverMountDetectionSensorSignal) is set for the conveyor axis (ProductFeed), and vector information relating to the right direction is set.

[0064] A workpiece misplacement sensor (ProductMisplaceDetectionSensorSignal) is set for the film feed sub-axis (FilmFeedSub), and vector information relating to the right direction is set.

[0065] A center seal heater (CenterHeaterDPC_MV) is set for the film feed sub-axis (FilmFeedSub), and vector information relating to the left direction is set.

[0066] The top seal heater (TopSealHeaterDPU_MV) is set for the top seal axis (TopSeal), and vector information relating to the left direction is set.

[0067] The top seal heater lower (TopSealLowerHeaterDPC_MC) is set for the top seal axis (TopSeal), and vector information relating to the left direction is set.

[0068] FIG. 6B is a diagram illustrating a design causal relationship based on device installation position data. With reference to Fig. 6(B), a design causal relationship (design causal model) based on device installation position data is generated for the design causal relationship in Fig. 5(B). A node is added based on the device installation position data in Fig. 5(B), and a design causal model to which the causal relationship of the node is added is generated.

[0069] Specifically, a vector is set from the conveyor axis (ProductFeed) to the work detection sensor (ProductDetectionSensorSignal). Also, a vector is set from the conveyor axis (ProductFeed) to the work over-mount sensor (ProductOverMountDetectionSensorSignal). A vector is set from the film transport sub-axis (FilmFeedSub) to the work misplacement sensor (ProductMisplaceDetectionSensorSignal). A vector is set from the film transport sub-axis (FilmFeedSub) to the work misplacement sensor (ProductMisplaceDetectionSensorSignal). A vector is set from the center seal heater (CenterHeaterDPC_MV) to the film transport sub-axis (FilmFeedSub). A vector is set from the top seal heater upper (TopSealHeaterDPU_MV) to the top seal axis (TopSeal). A vector is set from the top seal heater lower (TopSealLowerHeaterDPC_MC) to the top seal axis (TopSeal).

[0070] The first relationship identifying unit 114 generates a design causal relationship (design causal model) of the above configuration based on the process data 220 and the device installation position data 221.

[0071] <5.2 Second causal relationship> 7 is a flow diagram illustrating processing by the second relationship identifying unit 115 of the process analysis apparatus 1 according to the embodiment. However, the processing method described below is merely an example, and various methods are possible. In the following explanation, for convenience of explanation, it is assumed that the production line 3 includes four mechanisms F1 to F4 as the multiple mechanisms 31, and that in step S102, the second relationship identifying unit 115 has acquired the control program 222 that uses variables v1 to v4 corresponding to the four mechanisms F1 to F4.

[0072] First, the second relationship identification unit 115 performs a syntax analysis of the acquired control program 222 and constructs an abstract syntax tree from the control program 222 (step S1401). A known syntax analysis method using top-down syntax analysis or bottom-up syntax analysis may be used to construct the abstract syntax tree. For example, a syntax analyzer that handles character strings that comply with a specific formal grammar may be used to construct the abstract syntax tree. Once the construction of the abstract syntax tree is complete, the control unit 11 proceeds to the next step S1402.

[0073] FIG. 8 is a conceptual diagram in which the structure of a program according to the embodiment is expressed in a tree structure. 8 shows an abstract syntax tree 2211 obtained from the syntax "if(a>0)[v1=a;]else[v2=-a;]" in the control program 222. The abstract syntax tree is a data structure that represents the structure of a program in a tree structure in order to interpret the meaning of the program.

[0074] Specifically, the second relationship identification unit 115 omits tokens (phrases) used in the program, such as parentheses, that are unnecessary for interpreting the meaning of the program, and extracts tokens that are relevant to interpreting the meaning of the program. The second relationship identification unit 115 then associates operators, such as conditional branches, with nodes, and operands, such as variables, with leaves. By parsing the control program 222 in this manner, the second relationship identification unit 115 can construct an abstract syntax tree, such as the one shown in FIG. 8. The abstract syntax tree constructed in this manner represents the relationships between variables, operators, and nodes (such as the relationship between operations and operands). The structure of this abstract syntax tree may be modified, changed, omitted, or otherwise altered as appropriate, as long as the content it represents remains unchanged.

[0075] 7, second relationship identification unit 115 extracts variables (v1 to v4) related to each mechanism 31 and operators including conditional branching and assignment operations from the abstract syntax tree constructed in step S1401 (step S1402). For example, when abstract syntax tree 2211 shown in FIG. 8 is obtained, second relationship identification unit 115 extracts ("if", "v1", "=", "a", "v2", "=", and "-a") from abstract syntax tree 2211. When this extraction is complete, second relationship identification unit 115 proceeds to the next process.

[0076] The second relationship identifying unit 115 arranges the variables and operators extracted in step S1402 in order, and limits the targets for monitoring the execution order to parts related to the variables (v1 to v4) corresponding to each mechanism 31 for which a causal relationship is to be identified (step S1403). In the above example, the second relationship identifying unit 115 further extracts ("if", "v1", and "v2") from ("if", "v1", "=", "a", "v2", "=", and "-a").

[0077] Then, in the next step S1404, the second relationship identifying unit 115 randomly selects a conditional branch and attempts to execute the control program 222, thereby initializing the ordering of the limited variables (step S1403). Specifically, the second relationship identifying unit 115 can initialize the ordering of the limited variables by monitoring the order in which the limited variables are used in the attempt to execute the control program 222.

[0078] FIG. 9 is a diagram illustrating an example of a control causal relationship according to an embodiment. FIG. 9(A) illustrates an example of initialization of the control program 222 in which variables v1 to v4 corresponding to four mechanisms F1 to F4 are used.

[0079] Graph 2212 in FIG. 9(A) shows the order relationships of (1) to (5) below as a result of initialization. (1) Of the variables “v1” to “v4,” the variable “v1” is used first. (2) After variable “v1”, variable “v2” is used with a probability of 0.5, and variable “v3” is used with a probability of 0.5. "v3" is used. (3) After variable “v2”, variable “v3” is used with a probability of 1. (4) After variable “v3”, variable “v4” is used with a probability of 1. (5) Of the variables “v1” to “v4”, the variable “v4” is used last.

[0080] By such initialization, the second relationship identifying unit 115 can generate a control causal relationship based on the control program.

[0081] FIG. 9B is a diagram illustrating the control causal relationships based on the control program 222 by the second relationship identifying unit 115 in the manufacturing process carried out by the manufacturing line 3.

[0082] Referring to FIG. 9(B), as a control causal model based on the control program, a vector is set from the workpiece overhang sensor (ProductOverMountDetectionSensorSignal) to the conveyor axis (ProductFeed). A vector is set from the workpiece detection sensor (ProductDetectionSensorSignal) to the conveyor axis (ProductFeed). A vector is set from the film conveyance main axis (FilmFeedMain) to the film conveyance sub axis (FilmFeedSub). A vector is set from the film conveyance sub axis (FilmFeedSub) to the top seal axis (TopSeal). A vector is set from the conveyor axis (ProductFeed) to the film conveyance sub axis (FilmFeedSub).

[0083] <5.3 Third causal relationship> 10 is a diagram illustrating the processing of the third relationship identifying unit 116 of the process analysis device 1 according to the embodiment. However, the processing method described below is merely an example, and various methods are possible.

[0084] Referring to FIG. 10, the third relationship identifying unit 116 includes a constraint model generating unit 140 and a data causal model generating unit 142.

[0085] The constraint model generating unit 140 generates a constraint model based on the design causal model and the control causal model, by combining them.

[0086] The data causal model generation unit 142 generates a data causal model based on the constraint model and the state data.

[0087] The generation of the data causal model is described below. FIG. 11 is a flow diagram illustrating processing in the data causal model generation unit 142 according to the embodiment.

[0088] In the following description, for the sake of convenience, it is assumed that the production line 3 includes four mechanisms F1 to F4 as the plurality of mechanisms 31.

[0089] Referring to FIG. 11(A), the data causal model generating unit 142 acquires the state data 223 of each of the mechanisms F1 to F4 (step S101).

[0090] Next, the data causal model generation unit 142 analyzes the state data (step S103).

[0091] FIG. 11B is a subroutine flow diagram illustrating the analysis process of the status data.

[0092] 11(B), the data causal model generation unit 142 calculates a feature quantity 2221 from the state data 223 of each case acquired in step S101 (step S1301). The type of the feature quantity 2221 is not particularly limited and may be selected appropriately depending on the embodiment. Furthermore, a method for calculating the feature quantity 2221 can be determined appropriately depending on the embodiment. As a specific example, in this embodiment, the data causal model generation unit 142 calculates the feature quantity 2221 from the state data 223 by the following method. First, the data causal model generation unit 142 divides the acquired state data 223 of each case into frames to define a processing range for calculating the feature quantity 2221. The length of each frame may be set appropriately depending on the embodiment.

[0093] For example, the data causal model generation unit 142 may divide the status data 223 of each case into frames of a certain time length. However, the production line 3 does not necessarily operate at regular time intervals. Therefore, if the status data 223 of each case is divided into frames of a certain time length, there is a possibility that the operation of each mechanism 31 reflected in each frame will be out of sync.

[0094] Therefore, the data causal model generation unit 142 may divide the state data 223 into frames for each takt time. The takt time is the time it takes for the production line 3 to produce a predetermined number of products. This takt time can be determined based on a signal that controls the production line 3, for example, a control signal that the PLC 2 uses to control the operation of each mechanism 31 on the production line 3.

[0095] Next, the data causal model generation unit 142 calculates the value of the feature amount 2221 from each frame of the state data 223. When the state data 223 is continuous value data such as measurement data, the data causal model generation unit 142 may calculate, for example, the amplitude, maximum value, minimum value, average value, variance value, standard deviation, instantaneous value (single-point sample), etc. within the frame as the feature amount 2221. When the state data 223 is discrete value data such as detection data, the control unit 11 may calculate, for example, the "on" time, "off" time, duty ratio, number of "on" times, number of "off" times, etc. within each frame as the feature amount 2221. When the calculation of each feature amount 2221 is thereby completed, the data causal model generation unit 142 proceeds to the next step S1302.

[0096] Next, the data causal model generation unit 142 calculates the correlation coefficient or partial correlation coefficient between the feature quantities 2221 (step S1302). The correlation coefficient can be calculated using the following formula 1. Furthermore, the partial correlation coefficient can be calculated using the following formula 2.

[0097]

number

[0098] In addition, r ij indicates the element in the i-th row and j-th column of the matrix 2222. i and x j corresponds to data indicating the feature amount 2221 calculated from the state data 223 of each case. i and X j are respectively, x i and x j n indicates the number of features 2221 used to calculate the correlation.

[0099]

number

[0100] Note that the matrix R(r ij ) inverse matrix in R-1 (r ij ) and r ij indicates the element in the i-th row and j-th column of the inverse matrix of matrix 2222.

[0101] As a result, the data causal model generation unit 142 can obtain a matrix 2222 having correlation coefficients or partial correlation coefficients as each element. The correlation coefficients and partial correlation coefficients between the feature quantities 2221 indicate the strength of the relationship between the corresponding mechanisms 31. In other words, the strength of the relationship between the corresponding mechanisms 31 is specified by each element of the matrix 2222. Upon completing the calculation of the correlation coefficients or partial correlation coefficients between the feature quantities 2221, the data causal model generation unit 142 proceeds to the next step S1303.

[0102] Next, the data causal model generation unit 142 constructs undirected graph information 2223 indicating the strength of the relationship between the corresponding mechanisms 31 based on the correlation coefficients or partial correlation coefficients between the feature quantities 2221 (step S1303).

[0103] For example, the data causal model generation unit 142 creates a node corresponding to each mechanism 31. Then, when the value of the correlation coefficient or partial correlation coefficient calculated between two mechanisms 31 is equal to or greater than a threshold, the control unit 11 connects the two corresponding nodes with an edge. On the other hand, when the value of the correlation coefficient or partial correlation coefficient calculated between two mechanisms 31 is less than the threshold, the control unit 11 does not connect the two corresponding nodes with an edge. Note that the threshold may be a fixed value defined in the process analysis program 121 or a set value that can be changed by an operator or the like. Furthermore, the thickness of the edge may be determined according to the magnitude of the value of the corresponding correlation coefficient or partial correlation coefficient.

[0104] FIG. 12 is a diagram illustrating the creation of undirected graph information based on state data according to an embodiment. As shown in FIG. 12, undirected graph information 2223 can be created. As an example, four nodes corresponding to four mechanisms F1 to F4 are created. Edges are then formed between the nodes of mechanisms F1 and F2, between the nodes of mechanisms F1 and F3, between the nodes of mechanisms F2 and F3, and between the nodes of mechanisms F3 and F4. Furthermore, since the correlation between mechanisms F1 and F3 and between mechanisms F3 and F4 is greater than the correlation between other mechanisms, the edges between the nodes of mechanisms F1 and F3 and between the nodes of mechanisms F3 and F4 are formed thicker than the other edges. Thus, once the construction of undirected graph information 2223 indicating the strength of the relationships between multiple mechanisms 31 is completed, the analysis process of the state data 223 according to this embodiment is completed.

[0105] In the example of FIG. 12 , the undirected graph information 2223 represents the formed undirected graph using an image. However, the output format of the undirected graph information 2223 is not limited to an image and may be represented by text, etc. Also, in the above example, edges are not formed between nodes (mechanisms 31) with weak relationships by comparing the correlation coefficient or partial correlation coefficient with a threshold. However, the method of removing edges between nodes with weak relationships is not limited to this example. For example, after forming a graph in which all nodes are connected by edges, the data causal model generation unit 142 may delete edges from the formed graph in order of decreasing correlation coefficient or partial correlation coefficient so that the goodness of fit index (GFI, SRMR, etc.) representing the degree of deviation does not exceed a threshold.

[0106] The data causal model generation unit 142 adds a direction to the created undirected graph information by utilizing the information of the constraint model, and identifies the order relationship of the multiple mechanisms 31 in the production line 3.

[0107] FIG. 13 is a diagram illustrating a control causal relationship based on status data 223 by third relationship identifying unit 116 in the manufacturing process performed by manufacturing line 3 according to the embodiment.

[0108] Referring to FIG. 13, in this example, an example of a generated data causal model is shown.

[0109] A vector is set that goes from the work misplacement sensor (ProductMisplaceDetectionSensorSignal) to the conveyor axis (ProductFeed). A vector is set that goes from the conveyor axis (ProductFeed) to the film transport sub-axis (FilmFeedSub). A vector is set that goes from the film transport main axis (FilmFeedMain) to the work detection sensor (ProductDetectionSensorSignal). A vector is set that goes from the center seal heater (CenterHeaterDPC_MV) to the top seal lower heater (TopSealLowerHeaterDPC_MC).

[0110] <6. Anomaly detection> FIG. 14 is a diagram illustrating detection of an abnormality in abnormality detection unit 117 according to the embodiment.

[0111] Referring to FIG. 14, abnormality detection unit 117 detects an abnormality based on the status data. In this example, data of three variables P1 to P3 will be described as an example of the state data.

[0112] A three-dimensional space with three axes is set for variables P1 to P3, and a data group indicating the normal range and data indicating an abnormal state are shown as a learned model. The black point cloud data indicates a data group indicating the normal range. On the other hand, hatched data indicates data that is far from the normal range. The distance from the center point of the data group indicating the normal range is the abnormality score. In other words, the greater the distance, the higher the abnormality score value.

[0113] In this example, a three-dimensional space of three variables P1 to P3 is described, but it can of course be expanded to a multidimensional space according to the number of variables. It is possible to calculate the contribution rate of a variable according to the direction of the data from the center of the normal range as well as the distance. For example, if the amount of displacement with respect to the axis corresponding to variable P3 is large, the contribution rate of variable P3 will be large. In addition to the above method, the contribution rate of a variable to an abnormality may also be calculated using the importance of a decision tree (or random forest).

[0114] FIG. 15 is another diagram illustrating detection of anomaly in anomaly detection unit 117 according to the embodiment.

[0115] Referring to Figure 15(A), data changes from point to point are shown. In this example, a two-dimensional case is shown, with a change of Δx on the x-axis and a change of Δy on the y-axis. In the two-dimensional case, the change in distance Δx and Δy is calculated as the anomaly score.

[0116] 15(B), the data change of a point from the distribution is shown. In the three-dimensional case, the change in distance from the center of the distribution to the point is calculated as the anomaly score.

[0117] As an example of an algorithm for calculating the anomaly score, an anomaly detection algorithm such as an anomaly detection based on Mahalanobis distance, LOF (Local Outlier Factor), or Isolation Forest may be used.

[0118] <7. Estimation of the cause of the abnormality> FIG. 16 is a diagram illustrating estimation of the cause of abnormality by estimation unit 119 according to the embodiment.

[0119] Referring to FIG. 16, in this example, an estimation result screen 300 estimated by estimation unit 119 and output to display unit 122 is shown.

[0120] The estimation result screen 300 includes score data 310 , variables that contributed to the anomaly score and contribution rate data 320 , and a causal model 301 .

[0121] The estimation result screen 300 also has a button 330 for selecting a design causal model that can accept input from the user, a button 340 for selecting a control causal model, and a button 350 for selecting a data causal model.

[0122] The estimation unit 119 can switch and output estimation results according to input from the user.

[0123] As an example, a case is shown in which the user has selected button 340 for selecting a control causal model.

[0124] The score data 310 indicates a score value that changes along the time axis. In this example, the score value exceeds a threshold value TH1 at time T1.

[0125] The anomaly detection unit 117 detects an anomaly based on the score value exceeding the threshold value TH1 at time T1. The contribution rate calculation unit 118 calculates the contribution rate of the variable based on the score value at time T1.

[0126] As an example, the contribution rate calculation unit 118 outputs contribution rate data 320. In this example, the contribution rate data 320 shows a case where the film transport sub-axis (FilmFeedSub) has a contribution rate of 45%, the top seal axis (TopSeal) has a contribution rate of 23%, the conveyor axis (ProductFeed) has a contribution rate of 14%, the film transport main axis (FilmFeedMain) has a contribution rate of 4%, and the work detection sensor (ProductDetectionSensorSignal) has a contribution rate of 2%.

[0127] The estimation unit 119 extracts the variables with the first and second highest contribution rates from the calculated contribution rates of the variables. In this example, the film feed sub-axis (FilmFeedSub) and the top seal axis (TopSeal) are extracted.

[0128] In this example, the case where the control causal model button 340 is selected is shown. The estimation unit 119 identifies the nodes of the film feed sub-axis (FilmFeedSub) and the top seal axis (TopSeal) in the control causal model.

[0129] The estimation unit 119 extracts nodes connected to the node in the selected control causal model.

[0130] The estimation unit 119 extracts a node of the conveyor axis (ProductFeed) connected to the film transport sub-axis (FilmFeedSub). Then, it extracts a node of the conveyor workpiece overhang sensor (ProductOverMountDetectionSensorSignal) and a node of the workpiece misplacement sensor (ProductMisplaceDetectionSensorSignal) as nodes connected to the conveyor axis (ProductFeed).

[0131] Furthermore, the estimation unit 119 extracts the nodes of the work detection sensor (ProductDetectionSensorSignal) and the film transport main axis (FilmFeedMain) as nodes connected to the film transport sub axis (FilmFeedSub).

[0132] The estimation unit 119 extracts the film feed sub axis (FilmFeedSub) as a node connected to the top seal axis (TopSeal).

[0133] This makes it possible to easily check the relationships between nodes in the control causal model, making it easier to infer the cause of an abnormality.

[0134] In the above, a case has been described in which an abnormality is detected based on the score value exceeding the threshold value TH1 at time T1 in the score data 310. Then, a description has been given of the calculation of the contribution rate of the variable at time T1.

[0135] Note that the time T1 may be configured to allow the user to specify any time. For example, a pointing device may be used to specify a time, and a score value at the specified time may be obtained. Based on the obtained score value, contribution rate data 320 may be output as described above, and the contribution rates of the variables may be displayed. Then, the variables with the first and second highest contribution rates may be extracted from the contribution rate data 320, and an estimate of the selected causal model may be displayed in the same manner as described above.

[0136] FIG. 17 is another diagram illustrating estimation of the cause of abnormality by estimation unit 119 according to the embodiment.

[0137] Referring to FIG. 17, in this example, an estimation result screen 302 estimated by estimation unit 119 and output to display unit 122 is shown.

[0138] Compared to the estimation result screen 300, the estimation result screen 302 shows the case where button 330 for selecting a design causal model that can accept input from the user, button 340 for selecting a control causal model, and button 350 for selecting a data causal model are all selected.

[0139] A causal model 312 according to the selection is displayed on the estimation result screen 302. The other configurations are similar, so detailed description thereof will not be repeated.

[0140] The estimation unit 119 identifies the nodes of the film feed sub-axis (FilmFeedSub) and the top seal axis (TopSeal) in the design causal model, the control causal model, and the data causal model.

[0141] The estimation unit 119 extracts nodes connected to the identified nodes in the selected design causal model.

[0142] The estimation unit 119 extracts the nodes of the conveyor axis (ProductFeed), the center seal heater (CenterHeaterDPC_MV), and the film transport main axis (FilmFeedMain) connected to the film transport sub axis (FilmFeedSub).

[0143] The estimation unit 119 extracts the film conveyance sub-axis (FilmFeedSub), the top seal heater lower axis (TopSealLowerHeaterDPC_MC), and the top seal heater upper axis (TopSealHeaterDPU_MV) as nodes connected to the top seal axis (TopSeal).

[0144] The estimation unit 119 extracts nodes connected to the identified nodes in the selected control causal model.

[0145] The estimation unit 119 extracts a node of the conveyor axis (ProductFeed) connected to the film transport sub-axis (FilmFeedSub). Then, it extracts a node of the conveyor workpiece overhang sensor (ProductOverMountDetectionSensorSignal) and a node of the workpiece misplacement sensor (ProductMisplaceDetectionSensorSignal) as nodes connected to the conveyor axis (ProductFeed).

[0146] Furthermore, the estimation unit 119 extracts the nodes of the work detection sensor (ProductDetectionSensorSignal) and the film transport main axis (FilmFeedMain) as nodes connected to the film transport sub axis (FilmFeedSub).

[0147] The estimation unit 119 extracts the film feed sub axis (FilmFeedSub) as a node connected to the top seal axis (TopSeal).

[0148] The estimation unit 119 extracts nodes connected to the identified nodes in the selected data causal model.

[0149] The estimation unit 119 extracts a node of the conveyor axis (ProductFeed) connected to the film transport sub-axis (FilmFeedSub), and then extracts a node of the work misplacement sensor (ProductMisplaceDetectionSensorSignal) as a node connected to the conveyor axis (ProductFeed).

[0150] This makes it possible to easily check the relationships between nodes in the design causal model, control causal model, and data causal model, making it easier to infer the cause of an abnormality.

[0151] In this example, the format in which the respective causal models are combined and displayed has been described, but these combinations are arbitrary.

[0152] FIG. 18 is yet another diagram illustrating estimation of the cause of an abnormality by estimation unit 119 according to the embodiment.

[0153] Figure 18(A) shows the score data 310 described in Figure 14. In this example, the contribution rate of a variable when the score value at time T1 exceeds the threshold value TH1 is calculated.

[0154] Referring to FIG. 18(B), a transition relationship diagram for estimating the cause of an abnormality is shown. Referring to Figure 18 (C), there are shown a button 360 for setting a selection index for layer R2 and a button 370 for setting a selection index for layer R3, which can be provided on the estimation result screen and accept input from the user.

[0155] As an example, in this example, the hierarchy selection index is set so that any one of design causality, control causality, and data causality can be selected.

[0156] In this example, a case is shown in which control causality is selected as the selection index for hierarchy R2, and data causality is selected as the selection index for hierarchy R3.

[0157] 18(B) shows an example in which variables PA and PB with high contribution rates corresponding to the score value are identified in layer R1. Then, in layer R2, in response to an input of button 360 for setting the selection index for layer R2, the control causal model is referenced to extract variables connected to variables PA and PB.

[0158] In this example, the variable PA is connected to the variables PC and PD with reference to the control causal model.

[0159] Also, the variable PB is shown to be connected to the variables PE and PF with reference to the control causal model as the control causal relationship.

[0160] Then, in layer R3, in response to input of button 370 for setting the selection index for layer R3, the data causal model is referenced and variables connected to variables PC to PF are extracted.

[0161] In this example, variables PG and PH are extracted as variables connected to variable PD, and variables PI and PJ are extracted as variables connected to variable PF.

[0162] This allows the user to change the transition relationship diagram by referring to any model of the transition relationship, making it easy to check the relationships between nodes and facilitating the estimation of the cause of an abnormality.

[0163] In this example, we have explained the case where control causality is selected as the selection index for layer R2 and data causality is selected as the selection index for layer R3, but it is of course possible to change the causality to a different causality.

[0164] <8.Event List> 19 is a diagram illustrating an event list in event relation registration unit 120 according to an embodiment. Referring to FIG. 19, an event list in which a plurality of events are registered by a user is shown.

[0165] Specifically, the event relation registration unit 120 stores an event list including an event summary, event data, a causal model, actual physical phenomena, and details of recovery work. The event list also includes data on variables and contribution rates calculated by the contribution rate calculation unit 118.

[0166] The event summary shows that the event data and causal model corresponding to "poor seal adhesive strength" are registered, along with "press roller wear" as the actual physical phenomenon and "press roller replacement" as the recovery work. Also shown is the case where "39.1", "1.8", "9.3", and "12.3" are registered as the contribution rates of variables PA to PD, respectively.

[0167] Additionally, the event summary shows that event data and a causal model are registered for "poor seal adhesive strength," with "leather belt wear" registered as the actual physical phenomenon and "leather belt replacement" registered as the restoration work. Also, the contribution rates of variables PA through PD are shown to be "39.1," "1.8," "9.3," and "12.3," respectively. Also, the contribution rates of variables PA through PD are shown to be "19.1," "21.8," "39.3," and "12.3," respectively.

[0168] The event summary shows that the event data and causal model corresponding to "poor seal adhesive strength" are registered, along with "poor contact of current collecting ring" as the actual physical phenomenon and "replacement of current collecting ring" as the restoration work. Also shown is the case where "29.1", "31.8", "29.3", and "2.3" are registered as the contribution rates of variables PA to PD, respectively.

[0169] Also, as an event summary, the event data and causal model corresponding to "poor seal adhesive strength" are registered, and "film meandering" is registered as the actual physical phenomenon, and "correction of distortion of film transport axis" is registered as the recovery work content. Also, the contribution rates of variables PA to PD are registered as "9.1", "11.8", "19.3", and "12.3", respectively.

[0170] 20 is a diagram illustrating an event estimation process using an event list in event estimation unit 121 according to an embodiment. Referring to Fig. 20, event estimation unit 121 performs event estimation using the event list registered in event relation registration unit 120.

[0171] Specifically, the event estimation unit 121 acquires event data and calculates the similarity of the contribution rate.

[0172] As an example, the cosine similarity S ij It is also possible to use

[0173]

number

[0174] x i : Contribution rate of each variable in the event data list, x j : Contribution rate of each variable calculated by the contribution rate calculation unit 118 In this example, the event estimation unit 121 extracts an event list with a high degree of similarity in contribution rate. As an example, the case where the topmost event list with a high degree of similarity is extracted is shown.

[0175] Specifically, this is an event list in which the event summary "poor seal adhesive strength", the causal model, the actual physical phenomenon "press roller wear", and the recovery work content "press roller replacement" are registered.

[0176] Therefore, based on this information, the user is presented with "press roller wear" as the suspected location of the abnormality, and the recovery work to be done, which is to replace the press roller, is presented, thereby enabling recovery work to be carried out for the event at an early stage.

[0177] The event estimation unit 121 may present all event data to the user if the (similarity) is equal to or greater than a predetermined value.

[0178] 21 is a flow diagram of an event estimation process using an event list in event estimation unit 121 according to an embodiment. Referring to FIG. 21, event estimation unit 121 acquires an event list (step S101). Specifically, the event estimation unit 121 acquires the event list registered in event relation registration unit 120 described with reference to FIG. 20.

[0179] Next, the event estimation unit 121 calculates the similarity between the contribution rate of the variables included in the acquired event list and the contribution rate of the variables calculated by the contribution rate calculation unit 118 (step S102). Specifically, the event estimation unit 121 calculates the cosine similarity S ij Calculate.

[0180] Next, the event estimation unit 121 extracts a list of candidate events based on the calculated similarity (step S103).

[0181] Then, the event estimation unit 121 ends the process (END). This process makes it possible to extract an event list and perform recovery work for the event promptly.

[0182] Although the embodiments of the present disclosure have been described, the embodiments disclosed herein should be considered to be illustrative and not restrictive in all respects. The scope of the present disclosure is defined by the claims, and it is intended to include all modifications within the meaning and scope of the claims. [Explanation of symbols]

[0183] 1 Process analysis device, 3 Manufacturing line, 11, 21 Control unit, 12, 22 Memory unit, 13, 24 Communication interface, 14 Input device, 15 Output device, 16 Drive, 17 Process analysis program, 23 Input / output interface, 31 Mechanism, 91 Storage medium, 111 First acquisition unit, 112 Second acquisition unit, 113 Third acquisition unit, 114 First relationship identification unit, 115 Second relationship identification unit, 116 Third relationship identification unit, 117 Anomaly detection unit, 118 Contribution rate calculation unit, 119 Estimation unit, 120 Event relationship registration unit, 121 Event estimation unit, 122 Display unit, 130 Syntax analysis unit, 132 Program dependency analysis unit, 134 Variable tracking processing unit, 136 Model generation unit, 140 Constraint model generation unit, 142 Data causal model generation unit, 220 Process data, 221 Device installation position data, 222 Control programs, 223 state data.

Claims

1. a causal model generation unit that generates a plurality of causal models between variables based on the relationships between a plurality of mechanisms in a process performed on a production line, the dependencies between a plurality of events related to the plurality of mechanisms, or the control relationships between the plurality of mechanisms; an abnormality detection unit that acquires an abnormality detection result of the production line; a contribution rate calculation unit that calculates a contribution rate of a variable that contributed to the anomaly in the anomaly detection result; an estimation unit that estimates a cause of the abnormality for at least one of the plurality of causal models based on the calculated contribution rate of the variable that contributed to the abnormality; a display unit that displays an estimation result of the cause of an abnormality for at least one of the plurality of causal models based on the estimation result of the estimation unit; a reception unit that receives an input of a selection of the plurality of causal models from a user, The display unit displays an estimation result of the cause of an abnormality for a causal model selected from the plurality of causal models based on an estimation result from the estimation unit in accordance with reception of an input of a selection of the plurality of causal models by the reception unit.

2. A causal model generation unit that generates a plurality of causal models between variables based on the relationships between multiple mechanisms in a process performed on a manufacturing line, the dependencies between multiple events related to the multiple mechanisms, or the control relationships between the multiple mechanisms; an abnormality detection unit that acquires an abnormality detection result of the production line; a contribution rate calculation unit that calculates a contribution rate of a variable that contributed to the anomaly in the anomaly detection result; an estimation unit that estimates a cause of the abnormality for at least one of the plurality of causal models based on the calculated contribution rate of the variable that contributed to the abnormality; an event-related registration unit that registers an event list corresponding to an abnormal event; the event list includes an event summary, event data when the event occurred, an estimation result by the estimation unit, confirmed events related to the plurality of mechanisms, details of the event work, and information on the contribution rate of variables that contributed to the abnormality; The information processing apparatus further comprises an event estimation unit that calculates a similarity with an event list registered by the event relation registration unit based on the contribution rate of the variable that contributed to the abnormality calculated by the contribution rate calculation unit.

3. generating a plurality of causal models between variables based on the relationships between a plurality of mechanisms in a process performed on a manufacturing line, the dependencies between a plurality of events related to the plurality of mechanisms, or the control relationships between the plurality of mechanisms; acquiring an abnormality detection result of the production line; calculating a contribution rate of a variable that contributed to the anomaly in the anomaly detection result; and estimating a cause of the anomaly for at least one of the plurality of causal models based on the calculated contribution rate of the variable that contributed to the anomaly, a step of displaying an estimation result of the cause of the abnormality for at least one of the plurality of causal models based on the estimation result obtained by the estimating step; receiving an input of a selection of the plurality of causal models from a user; an information processing method, wherein the displaying step includes a step of displaying an estimation result of the cause of an abnormality for a causal model selected from the plurality of causal models based on the estimation result in accordance with receiving an input of a selection of the plurality of causal models.

4. A step of generating multiple causal models between variables based on the relationships between multiple mechanisms in processes performed on a manufacturing line, the dependencies between multiple events related to the multiple mechanisms, or the control relationships between the multiple mechanisms; acquiring an abnormality detection result of the production line; calculating a contribution rate of a variable that contributed to the anomaly in the anomaly detection result; a step of estimating a cause of the anomaly for at least one of the plurality of causal models based on the calculated contribution rate of the variable that contributed to the anomaly; registering an event list corresponding to the abnormal event; the event list includes an event summary, event data when the event occurred, an estimation result of the estimating step, confirmed events related to the plurality of mechanisms, event work content, and information on the contribution rate of variables that contributed to the abnormality; The information processing method further comprises a step of calculating a similarity with the event list registered in the registering step based on the contribution rate of the variable that contributed to the anomaly calculated in the calculating step.

5. An information processing program for causing a computer to execute a control method, The control method includes: generating a plurality of causal models between variables based on the relationships between a plurality of mechanisms in a process performed on a manufacturing line, the dependencies between a plurality of events related to the plurality of mechanisms, or the control relationships between the plurality of mechanisms; acquiring an abnormality detection result of the production line; calculating a contribution rate of a variable that contributed to the anomaly in the anomaly detection result; a step of estimating a cause of the anomaly for at least one of the plurality of causal models based on the calculated contribution rate of the variable that contributed to the anomaly; a step of displaying an estimation result of the cause of the abnormality for at least one of the plurality of causal models based on the estimation result obtained by the estimating step; receiving an input of a selection of the plurality of causal models from a user; an information processing program, wherein the displaying step includes a step of displaying an estimation result of the cause of an abnormality for a causal model selected from the plurality of causal models based on the estimation result in accordance with receiving an input of a selection of the plurality of causal models.

6. An information processing program for causing a computer to execute a control method, The control method includes: generating a plurality of causal models between variables based on the relationships between a plurality of mechanisms in a process performed on a manufacturing line, the dependencies between a plurality of events related to the plurality of mechanisms, or the control relationships between the plurality of mechanisms; acquiring an abnormality detection result of the production line; calculating a contribution rate of a variable that contributed to the anomaly in the anomaly detection result; a step of estimating a cause of the anomaly for at least one of the plurality of causal models based on the calculated contribution rate of the variable that contributed to the anomaly; registering an event list corresponding to the abnormal event; the event list includes an event summary, event data when the event occurred, an estimation result of the estimating step, confirmed events related to the plurality of mechanisms, event work content, and information on the contribution rate of variables that contributed to the abnormality; The control method further includes a step of calculating a similarity with the event list registered in the registering step based on the contribution rate of the variable that contributed to the abnormality calculated in the calculating step.

Images