Analysis monitoring device, analysis monitoring method, and analysis monitoring program

Abstract

A relationship analysis unit (120) uses at least one among feature amount data indicating a plurality of feature amounts of a target system in a target period, and state data indicating the state of each of a plurality of constituent elements of the target system at each time in the target period, to determine the relationship between elements of at least one analysis set among a plurality of analysis sets which is one of a plurality of sets having feature amounts and state groups, composed of one or more states, as elements, a plurality of sets having state groups as elements, and a plurality of sets having feature amounts as elements. A visualization unit (130) visualizes the relationship.

Description

Analysis monitoring device, analysis monitoring method, and analysis monitoring program

[0001] This disclosure relates to technology that can be used to understand the status of a system.

[0002] Manufacturing sites have a need to reduce waste and losses that hinder productivity and efficiency.

[0003] Patent Document 1 discloses the following analytical device. This analytical device identifies whether or not factors that reduce work efficiency have occurred based on the relationship between a data set of time periods during which the product is present and a data set of time periods during which the worker is present. Furthermore, the analytical device visualizes the time periods during which the product is present and the time periods during which the worker is present for each process on a timeline, and displays error marks superimposed at locations where it is determined that factors that reduce work efficiency have occurred. The analytical device also displays the distribution of work time in a histogram. Regarding factors that reduce work efficiency, for each process classification (automatic / semi-automatic / manual), conditions defining the relationship between the overlap of time periods during which the product is present (first time period) and time periods during which the worker is present (second time period) are pre-associated with factors that reduce work efficiency. For example, if the process classification is "automatic," two types of factors that reduce work efficiency are pre-associated: "none" and "setup changes, troubleshooting." If the process classification is "semi-automatic," three types of factors that reduce work efficiency are pre-associated: "none," "troubleshooting," and "feedback problems."

[0004] In this way, in conventional work analysis, work is analyzed from a specific perspective (e.g., staying time, waiting time, etc.), and the analysis results are presented by, for example, bar graphs or mark displays on a timeline. In the mark display, marks are superimposed together with factors at the time when a decrease in work efficiency is observed. However, the state of equipment, the state of people, and the environment (temperature, humidity, etc.) are not constant. That is, even for the same worker, the physical condition or skills change, and the state of equipment and the environment also change daily. Therefore, characteristic trends that were not seen at a certain analysis time point may emerge later, and the factors causing a decrease in work efficiency may also change. On the other hand, conventional work analysis does not consider the change of factors.

[0005] Japanese Patent Application Laid-Open No. 2023-105970

[0006] The present disclosure aims to visualize the change in relationships through daily analysis and updates by focusing on the relationships between states, the relationships between feature quantities, and the relationships between states and feature quantities.

[0007] The analysis monitoring device of the present disclosure uses at least one of feature quantity data indicating a plurality of feature quantities of a target system during a target period and state data indicating the state of each component of the target system at each time during the target period for each of the plurality of components of the target system, and determines the relationship between elements of at least any one of a plurality of analysis sets, which are any of a plurality of sets having as elements a state group composed of a feature quantity and one or more states, a plurality of sets having as elements state groups, and a plurality of sets having as elements feature quantities. It includes a relationship analysis unit.

[0008] According to the present disclosure, it is possible to determine the relationship between elements of an analysis set related to the target system and its change as information useful for grasping the situation of the target system.

[0009] A diagram showing the configuration of the analysis monitoring device 100 in Embodiment 1. A flowchart of the analysis monitoring method in Embodiment 1. A flowchart of step S110 in Embodiment 1. A diagram showing the state data 210 in Embodiment 1. A flowchart of step S120 in Embodiment 1. A diagram showing the correlation coefficient data 211 in Embodiment 1. A diagram showing the clustering data 212 in Embodiment 1. A flowchart of step S130 in Embodiment 1. A diagram showing an example of the clustering results in Embodiment 1. A diagram showing the visualization data 213 in Embodiment 1. A diagram showing an example of the clustering results in Embodiment 1. A diagram showing the visualization data 213 in Embodiment 1. A diagram showing an example of the clustering results in Embodiment 1. A diagram showing the visualization data 213 in Embodiment 1. A diagram showing the visualization data 214 in Embodiment 1. A diagram showing the configuration of the analysis monitoring device 100 in Embodiment 2. A flowchart of the analysis monitoring method in Embodiment 2. A diagram showing the relationship history data 220 in Embodiment 2. A diagram showing the relationship history data 221 in Embodiment 2. A flowchart of step S240 in Embodiment 2. A diagram showing visualization data 222(a) in Embodiment 2. A diagram showing visualization data 222(b) in Embodiment 2. A diagram showing visualization data 222(c) in Embodiment 2. A configuration diagram of the analysis monitoring device 100 in Embodiment 3. A flowchart of the analysis monitoring method in Embodiment 3. A flowchart of step S340 in Embodiment 3. A diagram showing relationship history data 230 and relationship list data 231 in Embodiment 3. A diagram showing relationship history data 232 in Embodiment 3. A diagram showing relationship history data 233 in Embodiment 3. A diagram showing relationship history data 234 in Embodiment 3. A diagram showing relationship list data 235 in Embodiment 3. A flowchart of step S350 in Embodiment 3. A diagram showing an example of displaying the relationship list (correlation coefficient) in Embodiment 3. A diagram showing an example of displaying the relationship list (relevant period) in Embodiment 3. A diagram showing an example of displaying the relationship list (relevant period) in Embodiment 3. A configuration diagram of the analysis monitoring device 100 in Embodiment 4.Flowchart of the analysis and monitoring method in Embodiment 4. Flowchart of step S440 in Embodiment 4. Diagram showing an example of displaying relationship changes in Embodiment 4. Diagram showing an example of displaying a relationship list (correlation coefficient) in Embodiment 4. Diagram showing an example of displaying relationship changes in Embodiment 4. Hardware configuration diagram of the analysis and monitoring device 100 in Embodiment.

[0010] In the embodiments and drawings, the same or corresponding elements are denoted by the same reference numeral. The descriptions of elements denoted by the same reference numeral as the described elements are omitted or simplified as appropriate. The arrows in the figures mainly indicate the flow of data or processing.

[0011] Embodiment 1. The analysis and monitoring device 100 will be described with reference to Figures 1 to 16.

[0012] ***Configuration Description*** The configuration of the analysis and monitoring device 100 will be described based on Figure 1. The analysis and monitoring device 100 is a computer equipped with hardware such as a processor 101, memory 102, auxiliary storage device 103, communication device 104, and input / output interface 105. These hardware components are connected to each other via signal lines.

[0013] The processor 101 is an integrated circuit (IC) that performs arithmetic processing and controls other hardware. For example, the processor 101 is a CPU, DSP, GPU, or a combination of these. IC is an abbreviation for Integrated Circuit. CPU is an abbreviation for Central Processing Unit. DSP is an abbreviation for Digital Signal Processor. GPU is an abbreviation for Graphics Processing Unit.

[0014] Memory 102 is a volatile or non-volatile storage device. Memory 102 is also called main memory. For example, memory 102 is RAM. Data stored in memory 102 is saved to auxiliary storage device 103 as needed. RAM is an abbreviation for Random Access Memory.

[0015] The auxiliary storage device 103 is a non-volatile storage device. For example, the auxiliary storage device 103 is a ROM, HDD, flash memory, SSD, or a combination thereof. Data stored in the auxiliary storage device 103 is loaded into memory 102 as needed. ROM is an abbreviation for Read Only Memory. HDD is an abbreviation for Hard Disk Drive. SS is an abbreviation for Solid State Drive.

[0016] The communication device 104 is a receiver and transmitter. For example, the communication device 104 is a communication chip or NIC. For example, the communication device 104 communicates using a wired LAN, wireless LAN, or Bluetooth®. Communication of the analysis and monitoring device 100 is performed using the communication device 104. NIC is an abbreviation for Network Interface Card. LAN is an abbreviation for Local Area Network.

[0017] The input / output interface 105 is a port to which input and output devices are connected. For example, the input / output interface 105 is a USB terminal. For example, input devices include keyboards, mice, touch panels, and microphones. For example, output devices include displays and speakers. An example of a display is a Liquid Crystal Display (LCD). Input and output of the analysis and monitoring device 100 are performed via the input / output interface 105. USB is an abbreviation for Universal Serial Bus.

[0018] The analysis and monitoring device 100 comprises elements such as a state group definition unit 110, a relationship analysis unit 120, and a visualization unit 130. These elements are implemented using software.

[0019] The auxiliary storage device 103 stores analysis and monitoring programs that enable the computer to function as a state group definition unit 110, a relationship analysis unit 120, and a visualization unit 130. The analysis and monitoring programs are loaded into memory 102 and executed by the processor 101. The auxiliary storage device 103 also stores the operating system (OS). At least a portion of the OS is loaded into memory 102 and executed by the processor 101. The processor 101 executes the analysis and monitoring programs while executing the OS. OS is an abbreviation for Operating System.

[0020] The data of the analysis monitoring program (input data, output data, etc.) is stored in the storage unit 190. Memory 102 functions as the storage unit 190. However, storage devices such as auxiliary storage device 103, registers in the processor 101, and cache memory in the processor 101 may function as the storage unit 190 instead of memory 102, or together with memory 102.

[0021] The analysis and monitoring program can be recorded (stored) in a computer-readable format on a non-volatile recording medium such as an optical disc or flash memory.

[0022] ***Explanation of Operation*** The operation procedure of the analysis monitoring device 100 corresponds to the analysis monitoring method. Furthermore, the operation procedure of the analysis monitoring device 100 corresponds to the processing procedure of the analysis monitoring program.

[0023] The analysis and monitoring method will be explained based on Figure 2. Steps S110 to S130 are performed for each target period and each data unit. The target period is the period to be analyzed. The length of the target period is, for example, one week or one month. The data unit is the unit of data to be analyzed. Examples of data units are machines, workers, processes, tasks, etc.

[0024] In step S110, the state group definition unit 110 defines multiple state groups using state data.

[0025] The status data indicates the state of each component of the target system at each time point during the target period. The status data is stored in the storage unit 190.

[0026] The target system is the system being analyzed. An example of a target system is a system in a manufacturing plant. Examples of components of the target system include "machinery," "equipment," "people (workers)," "environment," and "sensing data."

[0027] A state group consists of one or more states of one or more components. When a component is a "machine" or "equipment," the types of states include operating states and the states of ancillary equipment (e.g., doors). Examples of operating states include "start," "stop," and "warning." When a component is a "person (worker)," the types of states include working states and presence / absence states. Examples of working states include "working," "on break," and "moving." When a component is an "environment," the types of states include "temperature" and "humidity." The component "sensing data" is data that shows the measured values ​​(sensor values) of various sensors corresponding to the work. Examples of types of sensor values ​​include "pressure" and "current."

[0028] Multiple sets of states differ from each other in that one or more states of one or more of their constituent elements.

[0029] Step S110 will be explained in detail based on Figure 3. In step S111, the state group definition unit 110 accepts the specification of one or more state columns in the state data.

[0030] A specific example of step S111 will be explained based on Figure 4. The state data 210 is an example of state data in which the operating status of the machine and the working status of a person are recorded. The state data 210 has state columns such as "operating status of machine (1)", "working status of person", and "door A of machine (1)" for each "date and time". The state group definition unit 110 displays the state data 210 on the display, the user specifies one or more state columns that they want to analyze, and the state group definition unit 110 accepts the user's specification.

[0031] Returning to Figure 3, we will continue the explanation from step S112. In step S112, the state group definition unit 110 defines multiple state groups by combining the specified state column with the possible values ​​for each specified state column.

[0032] A specific example of step S112 will be explained based on Figure 4. Here, we assume that the specified state columns are "operating state of machine (1)" and "working state of person". We also assume that "operating state of machine (1)" takes three values ​​corresponding to three states: "started", "stopped", and "warning", and that "working state of person" takes two values ​​corresponding to two states: "working" and "absent". The state group definition unit 110 defines eleven state groups. The defined state groups are as follows: The first state group consists of one state: "machine (1) is started". The second state group consists of one state: "machine (1) is stopped". The third state group consists of one state: "machine (1) is warning". The fourth state group consists of one state: "person is working". The fifth state group consists of one state: "person is absent". The sixth state group consists of two states: "machine (1) is started" and "person is working". The seventh state group consists of two states: "Machine (1) is running" and "No person is present." The eighth state group consists of two states: "Machine (1) is stopped" and "A person is working." The ninth state group consists of two states: "Machine (1) is stopped" and "No person is present." The tenth state group consists of two states: "Machine (1) is warning" and "A person is working." The eleventh state group consists of two states: "Machine (1) is warning" and "No person is present."

[0033] Returning to Figure 2, we will continue the explanation from step S120. In step S120, the relationship analysis unit 120 uses at least one of the feature data and the state data to determine the relationships between the elements of at least one of the multiple analysis sets.

[0034] Feature data represents multiple features of the target system during the specified period. An example of a feature is the number of times each of several programs executed on the target system's machine is run.

[0035] Step S120 will be explained in detail based on Figure 5. In step S121, the relationship analysis unit 120 calculates multiple feature quantities of the target system for the target period using feature information data.

[0036] Feature information data indicates information that identifies multiple features. Feature information data is stored in the storage unit 190. If the features are the execution counts of multiple programs, the feature information data indicates, for example, the programs executed at each time. The relationship analysis unit 120 then calculates the execution count for each program. The multiple execution counts calculated become multiple features.

[0037] The data representing the multiple calculated features becomes the feature data.

[0038] In step S122, the relationship analysis unit 120 defines multiple analysis sets.

[0039] An analysis set is a set of elements to be analyzed. Specifically, an analysis set is a set of elements consisting of a feature and a set of states. However, an analysis set may also be a set of sets of states as elements, or a set of features as elements.

[0040] If the analysis set is a set of features and a set of states, the relationship analysis unit 120 defines multiple analysis sets by combining multiple features and multiple set of states. If the analysis set is a set of set of states, the relationship analysis unit 120 defines multiple analysis sets by combining multiple set of states. In this case, step S121 is unnecessary. If the analysis set is a set of features, the relationship analysis unit 120 defines multiple analysis sets by combining multiple features. In this case, step S110 is unnecessary.

[0041] In step S123, the relationship analysis unit 120 calculates the correlation coefficient between the elements of each analysis group.

[0042] When the analysis set is a pair of a feature quantity and a state group, the relationship analysis unit 120 calculates the number of occurrences of the state group in the target period for each state group using the state data, and calculates the correlation coefficient between the feature quantity and the number of occurrences for each analysis set using the feature quantity data. When the analysis set is a pair of state groups, the relationship analysis unit 120 calculates the number of occurrences of the state group in the target period for each state group using the state data, and calculates the correlation coefficient between the numbers of occurrences for each analysis set. When the analysis set is a pair of feature quantities, the relationship analysis unit 120 calculates the correlation coefficient between the feature quantities for each analysis set using the feature quantity data.

[0043] The data indicating the calculated plurality of correlation coefficients is referred to as correlation coefficient data.

[0044] FIG. 6 shows the correlation coefficient data 211. The correlation coefficient data 211 is an example of correlation coefficient data indicating a plurality of correlation coefficients when, in a certain process A, the analysis set is a pair of the number of executions of a program (feature quantity) and the state of a machine. In the correlation coefficient data 211, the correlation coefficient of the analysis set that is a pair of the feature quantity of the number of executions of the program (1) and the state of the startup of the machine (1) is 0.71.

[0045] Returning to FIG. 5, the description will be continued from step S124. In step S124, the relationship analysis unit 120 clusters a plurality of reference elements of the plurality of analysis sets based on the respective correlation coefficients of the plurality of analysis sets.

[0046] The reference element is one element of the analysis set. When the analysis set is a pair of a feature quantity and a state group, the reference element is a feature quantity or a state group. When the analysis set is a pair of state groups, the reference element is one state group. When the analysis set is a pair of feature quantities, the reference element is one feature quantity.

[0047] By clustering, the reference elements whose correlation coefficients of the analysis sets are similar to each other are grouped together. Clustering is also called grouping. One or more reference elements grouped by clustering are referred to as a reference element cluster. A cluster is also called a group.

[0048] The data indicating the clustering result is referred to as clustering data.

[0049] Figure 7 shows the clustering data 212. Clustering data 212 is an example of clustering data where the analysis set is a set of features and a set of states, and the reference elements are features. The "Cluster" column shows the reference element cluster. The "Classified Features" column shows one or more reference elements belonging to the reference element cluster. In the correlation coefficient data 211 (see Figure 6), the number of executions of program (1), program (4), and program (5) are similar. Therefore, in the clustering data 212 (see Figure 7), the number of executions of program (1), program (4), and program (5) are classified into cluster (1). In the correlation coefficient data 211 (see Figure 6), the number of executions of program (2), program (3), and program (6) are similar. Therefore, in the clustering data 212 (see Figure 7), the number of executions of program (2), program (3), and program (6) are classified into cluster (2).

[0050] Returning to Figure 5, step S125 will be explained. In step S125, the relationship analysis unit 120 determines the relationship of each matching element to each reference element cluster based on the clustering results.

[0051] The matching element is the other element (not the reference element) of the analysis pair. If the analysis pair is a pair of a feature and a state group, and the reference element is the feature, then the matching element is the state group. If the analysis pair is a pair of state groups, then the matching element is the other state group. If the analysis pair is a pair of features, then the matching element is the other feature.

[0052] The relationship is defined using a format. The format is stored in the memory unit 190. For example, the following format is used: "In FFF, state S is a combination of feature i and CCC."

[0053] "FFF" is a string indicating the data unit for which the correlation coefficient is calculated. "CCC" is a string indicating the relationship. "CCC" is defined by the user in combination with the correlation coefficient X. The correlation coefficient X is the condition for the value of the correlation coefficient. "State S" is the state that forms a pair with feature i in the analysis set. "Feature i" is the state that forms a pair with state S in the analysis set. "i" is a number that identifies the feature.

[0054] Assume that "FFF" is "Process A" and "CCC" is "strongly correlated". Assume that the correlation coefficient X is greater than or equal to X1 and less than X2, with X1 being 0.7 and X2 being 1.1. In this case, the relationship analysis unit 120 extracts analysis pairs with correlation coefficients greater than or equal to X1 and less than X2 from the correlation coefficient data 211 (see Figure 6). Then, based on the clustering data 212 (see Figure 7), the relationship analysis unit 120 creates the following statements to indicate the relationships between the extracted analysis pairs. The first relationship is shown as "In Process A, the state of "Machine (1) is started" is strongly correlated with the number of times Program (1) is executed, the number of times Program (4) is executed, and the number of times Program (5) is executed." The second relationship is shown as "In Process A, the state of "Machine (1) is stopped" is strongly correlated with the number of times Program (1) is executed, the number of times Program (4) is executed, and the number of times Program (5) is executed." The third relationship is shown to be that "In process A, the state of 'Machine (1) is warning' is strongly correlated with the number of times program (1) is executed, the number of times program (4) is executed, and the number of times program (5) is executed."

[0055] Multiple combinations of correlation coefficients X(X1, X2) and "CCC" can be defined. For example, for X1 = 0.4 and X2 = 0.7, "CCC" can be defined as "moderate correlation".

[0056] Even when the analysis set is a set of state groups or a set of features, the relationship can be determined using the format and clustering results, just as when the analysis set is a set of features and state groups.

[0057] An example of the format when the analysis set is a set of states is shown below: "In FFF, state S is state i and CCC."

[0058] An example of the format when the analysis pair is a pair of features is shown below: "In FFF, feature k is a combination of feature i and CCC."

[0059] Returning to Figure 2, step S130 will be explained. In step S130, the visualization unit 130 visualizes the determined relationship.

[0060] Visualization is performed as follows: The visualization unit 130 determines the order of multiple reference element clusters and displays the multiple reference element clusters in a line. For each reference element cluster, the visualization unit 130 displays the multiple reference elements belonging to the reference element cluster in a line. The visualization unit 130 displays the multiple reference element clusters for each target period in the order of the target periods.

[0061] Step S130 will be explained in detail based on Figure 8. In step S131, the visualization unit 130 selects the display target.

[0062] The display target is the type of information to be displayed. Examples of display targets include the reference period and feature quantities.

[0063] For example, the display targets are selected by the user as follows: If the user specifies only features, the visualization unit 130 selects only the specified features to be displayed. If the user specifies only a reference period, the visualization unit 130 selects only the specified reference period to be displayed. If the user specifies both a reference period and features, the visualization unit 130 selects both the specified reference period and the specified features to be displayed.

[0064] For example, the display targets are selected based on at least one of the reference period and the correlation coefficient as follows: The visualization unit 130 selects the feature quantities of analysis sets for which the correlation coefficient has been within a predetermined range as the display targets. The visualization unit 130 selects the feature quantities of analysis sets for which the correlation coefficient during the reference period is within a predetermined range as the display targets.

[0065] In step S132, the visualization unit 130 acquires analysis data related to the display target.

[0066] The analysis data shows various information (analysis results) obtained in step S120. For example, the analysis data shows the correlation coefficients of multiple analysis sets, relationship information between elements of each analysis set, and multiple reference element clusters. The analysis data is stored in the storage unit 190 in association with the target period.

[0067] For example, let's assume the analysis data shows the correlation coefficients for multiple analysis sets for 50 programs. Let's also assume the display target is the execution count (feature) of 18 programs from program (1) to program (18). In this case, the visualization unit 130 extracts the correlation coefficients from the analysis data for each analysis set that includes any of the 18 programs from program (1) to program (18) as an element.

[0068] In step S133, the visualization unit 130 determines the arrangement order of the display targets.

[0069] For example, the arrangement order of the display targets (reference elements) is determined using correlation coefficient data and clustering data as follows: <First policy> The visualization unit 130 determines the arrangement order of the reference elements belonging to each reference element cluster, in which the reference elements belonging to the reference element cluster are adjacent to each other. <Second policy> The visualization unit 130 calculates the average correlation coefficient for each reference element cluster and determines the arrangement order of the reference element clusters in order of the magnitude of the average correlation coefficients. For example, multiple reference element clusters are arranged from left to right (or from top to bottom) in descending order of the average correlation coefficient. <Third policy> The visualization unit 130 re-determines the arrangement order of the display targets according to the clustering results for each target period within the reference period.

[0070] Let me explain the third policy. Each reference element cluster in the current target period will be called the current reference element cluster. Each reference element cluster in the previous target period (previous target period) will be called the previous reference element cluster. Among the multiple reference element clusters in the current period, any reference element cluster that is different from any of the previous reference element clusters will be called a new reference element cluster. Examples of new reference element clusters are as follows: An example of a new reference element cluster is a reference element cluster formed by merging two or more previous reference element clusters. An example of a new reference element cluster is a reference element cluster that corresponds to a previous reference element cluster with one or more reference elements added. An example of a new reference element cluster is a reference element cluster that corresponds to a previous reference element cluster with one or more reference elements deleted. If the multiple reference element clusters in the current period include a new reference element cluster, the visualization unit 130 determines the order in which the multiple reference element clusters in the current period are arranged. The visualization unit 130 also determines the order in which the reference elements belonging to each new reference element cluster are arranged. In this process, the visualization unit 130 determines the order in which two or more reference elements belonging to the new reference element cluster are adjacent to each other, compared to two or more reference elements that belonged to the same reference element cluster in the previous process. Among the multiple reference element clusters in the current process, any reference element cluster that is the same as any of the previous reference element clusters is referred to as a common reference element cluster. The order in which one or more reference elements belonging to a common reference element cluster are arranged is the same as the previous order.

[0071] In step S134, the visualization unit 130 displays the analysis data to be displayed in the determined arrangement order.

[0072] The displayed analytical data is referred to as visualized data.

[0073] Figures 9 to 15 show examples of clustering results and visualization data 213. Visualization data 213 is an example of visualization data. Visualization data 213 represents a heatmap of correlation coefficients. The displayed features are those specified by the user. The relationship analysis was performed every seven days.

[0074] Figure 9 shows the clustering results for the period from April 1 to April 21, 2024.

[0075] Figure 10 shows the visualization of data 213 for the period from April 1 to April 21, 2024. (A) For the period from April 1 to April 7, 2024, features (1), (4), and (5) constitute the first reference element cluster, and features (2), (3), and (6) through (18) constitute the second reference element cluster (see Figure 9). Therefore, features (1), (4), and (5) are arranged consecutively, and features (2), (3), and (6) through (18) are arranged consecutively. Also, the average correlation coefficient of the first reference element cluster is higher than the average correlation coefficient of the second reference element cluster. Therefore, the first and second reference element clusters are arranged from left to right.

[0076] (B) For the period from April 8 to April 14, 2024, the clustering results were the same as for period (A). Therefore, the order of the features remains unchanged.

[0077] (C) For the period from April 15 to April 21, 2024, the clustering results were the same as for period (B). Therefore, the order of the features remains unchanged.

[0078] Figure 11 shows the clustering results for the period from April 22 to May 5, 2024.

[0079] Figure 12 shows the visualization data 213 for the period from April 1 to May 5, 2024. (D) For the period from April 22 to April 28, 2024, the second reference element cluster was divided into the (2-1) reference element cluster and the (2-2) reference element cluster (see Figure 11). The (2-1) reference element cluster consists of features (2), (16), and (18). The (2-2) reference element cluster consists of features (3), (6), and (7) through (17). Therefore, features (2), (16), and (18) are arranged consecutively, and features (3), (6), and (7) through (17) are arranged consecutively. The order of the features in the second reference element cluster from April 1 to April 21, 2024 has also been changed accordingly.

[0080] (E) For the period from April 29 to May 5, 2024, the clustering results were the same as for period (D). Therefore, the order of the features remains unchanged.

[0081] Figure 13 shows the clustering results for the period from May 6 to May 26, 2024.

[0082] Figure 14 shows the visualization data 213 for the period from April 1 to May 12, 2024. (F) For the period from May 6 to May 12, 2024, the first reference element cluster and the (2-2) reference element cluster were merged into the third reference element cluster (see Figure 13). The third reference element cluster consists of features (1) and (3) through (17). The average correlation coefficient of the (2-1) reference element cluster is higher than the average correlation coefficient of the third reference element cluster. Therefore, the (2-1) reference element cluster and the third reference element cluster are arranged from left to right. In addition, in the third reference element cluster, the features (1, 4, 5) that belonged to the former first reference element cluster are arranged consecutively, and the features (3, 6-17) that belonged to the former (2-2) reference element cluster are arranged consecutively. In other words, the (2-1) reference element cluster, the former 1st reference element cluster, and the former (2-2) reference element cluster are arranged from left to right.

[0083] Figure 15 shows the visualization of data 213 for the period from April 1 to May 26, 2024. (G) For the periods from May 13 to May 19, 2024 and from May 20 to May 26, 2024, the clustering results were the same as for period (F). Therefore, the order of the features remains unchanged.

[0084] From the visualization data 213 shown in Figures 10, 12, 14, and 15, the status of the target system can be understood as follows. Assume that the analysis pair consisted of a feature called "number of program executions" and a state called "machine (1) is warning".

[0085] In Figure 10, it can be seen that as of April 8, 2024, programs (1), (4), and (5) are programs related to "warnings." Furthermore, as of April 22, 2024, it can be seen that programs (1), (4), and (5) continue to be programs related to "warnings." It can be seen that the factors causing machine (1) to issue a warning have not changed. In Figure 12, it can be seen that as of April 29, 2024, programs (2), (16), and (18) have newly become programs related to "warnings." It can be seen that the factors causing machine (1) to issue a warning have changed. In Figure 15, as of May 27, 2024, programs (2), (16), and (18) are programs related to "warnings." It can be easily seen that the influence of programs (1), (4), and (5) has decreased.

[0086] The visualization unit 130 may display multiple visualization data corresponding to multiple data units side by side on the same screen. Figure 16 shows the display of visualization data 214. Visualization data 214 includes visualization data for process A and visualization data for process B. For each of process A and process B, a heat map representing the relationship between state (1) and each feature is displayed. The state for process A may be different from the state for process B. For example, the state for process A may be state (1) and the state for process B may be state (3). The data unit is not limited to processes. Multiple data units do not have to be of the same type. For example, a heat map representing the relationship between state (1) and each feature may be displayed for both worker (1) and process A. The number of data units is not limited to two, but may be three or more. The analysis pairs related to relationships may be pairs of state groups or pairs of feature pairs.

[0087] ***Effects of Embodiment 1*** Embodiment 1 aims to enable understanding of the state of a system (a system in natural science). Understanding the state can be used to improve the system. Therefore, in Embodiment 1, the relationships between states or the relationship between states and feature quantities obtained from various observation data are analyzed daily for various states in the field, and changes in these relationships are visualized.

[0088] Embodiment 1 is characterized by enabling easy understanding of changes in the relationship between a state and a set of features through visualization.

[0089] Conventionally, it has been difficult to easily grasp changes in the relationship between states and features. On the other hand, the analysis and monitoring device 100 analyzes and visualizes the relationships between states or the relationships between states and features obtained from various observation data on a daily basis. This allows users to easily grasp changes in relationships. By checking the visualization results daily, users can easily understand whether the situation on site is the same as before or has changed, and use this information to make improvements. When there are many features and visualization is performed without considering the order of the features, it is difficult to grasp the similarity and changes in relationships. In contrast, Embodiment 1 makes it possible to easily grasp the similarity and changes in relationships even when there are many features.

[0090] ***Supplement to Embodiment 1*** Embodiment 1 shows an example of operation in which multiple columns representing the state to be analyzed are specified, and a state group is defined by calculating combinations of states from the multiple columns based on the types of values ​​that each specified column can take. However, the method of defining a state group is not limited to this method. For example, combinations of columns may be automatically created for all columns, and those to be used as a state group may be defined from the created combinations. Alternatively, all automatically created combinations may be used as the entire state group.

[0091] Embodiment 2. This embodiment involves accumulating the relationship analysis results as a history and visualizing the relationships according to conditions specified by the user. The main differences from Embodiment 1 will be explained based on Figures 17 to 24.

[0092] ***Configuration Description*** The configuration of the analysis monitoring device 100 will be described based on Figure 17. The analysis monitoring device 100 further includes a relationship storage unit 140. The analysis monitoring program further causes a computer to function as the relationship storage unit 140.

[0093] ***Explanation of Operation*** The analysis and monitoring method will be explained based on Figure 18. Steps S210 to S240 are executed for each target period and each data unit.

[0094] In step S210, the state group definition unit 110 defines a plurality of state groups using state data. Step S210 is the same as step S110 in Embodiment 1.

[0095] In step S220, the relationship analysis unit 120 uses at least one of the feature data and the state data to determine the relationships between the elements of at least one of the multiple analysis sets. Step S220 is the same as step S120 in Embodiment 1.

[0096] In step S230, the relationship storage unit 140 records the determined relationship in the relationship history data.

[0097] Relationship history data shows the relationships determined for each period.

[0098] Figure 19 shows the relationship history data 220. The relationship history data 220 is an example of relationship history data when analysis is performed every 7 days. The relationship history data 220 shows information for each column: "Period", "Process", "State", and "Relationship between State and Feature".

[0099] Figure 20 shows the relationship history data 221. Relationship history data 221 is an example of relationship history data when analysis is performed every 7 days. In addition to the information in the columns for "Period," "Process," "State," and "Relationship between State and Feature," relationship history data 221 also shows the information in the column for "Worker."

[0100] Relationship history data can represent any information. The information shown in relationship history data is not limited to the information in the columns shown in Figures 19 and 20. For example, if the analysis is performed by machine, the relationship history data will include the machine number and the relationship information corresponding to the machine number. For example, if the analysis is performed by element work, the relationship history data will include the relationship information corresponding to each element work.

[0101] Returning to Figure 18, step S240 will be explained. In step S240, the visualization unit 130 visualizes the determined relationship.

[0102] Step S240 will be explained in detail based on Figure 21. In step S241, the visualization unit 130 receives a specification of the data target and sets the specified data target.

[0103] The data target is specified by the user using a combination of columns in the relationship history data. For example, a user can specify the data target for relationship history data 220 (see Figure 19) as "Process column is Process A, AND Status column is 'Machine (1) is in Warning'". If the relationship history for Process B is recorded in the relationship history data, the user can specify the data target as, for example, "Process column is Process B, AND Status column is 'Machine (1) is in Warning'". For example, a user can specify the data target for relationship history data 221 (see Figure 20) as "Worker column is Worker (1), AND Status column is 'Machine (1) is in Warning'". However, the data targets are not limited to these examples.

[0104] In step S242, the visualization unit 130 selects the display target. Step S242 is the same as step S131 in Embodiment 1.

[0105] In step S243, the visualization unit 130 acquires analysis data related to the data target and the display target.

[0106] In step S244, the visualization unit 130 determines the arrangement order of the display targets. Step S244 corresponds to step S133 in Embodiment 1.

[0107] If the difference in the mean correlation coefficients between clusters is below a threshold, the visualization unit 130 does not need to sort the reference element clusters by the magnitude of the mean correlation coefficients between them. The threshold may be entered interactively by the user, predefined, or set by other means.

[0108] In step S245, the visualization unit 130 displays the analysis data to be displayed in the determined arrangement order.

[0109] Figures 22 to 24 show the visualization of the data 222. The data 222 is an example of visualization data. The data displayed are the features of analysis sets that have had a correlation coefficient of 0.7 or higher.

[0110] In Figure 22, visualization data 222(a) shows the features of analysis pairs that had a correlation coefficient of 0.7 or higher during the period from July 1, 2024 to September 15, 2024. During the period from July 1, 2024 to July 21, 2024, the correlation coefficients of each analysis pair of features (1), (7), and (8) were not 0.7 or higher. However, during the period from September 2, 2024 to September 15, 2024, the correlation coefficients of each analysis pair of features (1), (7), and (8) were 0.7 or higher. Therefore, features (1), (7), and (8) are displayed. Note that there were no features to display during the period from July 22 to September 1, so a blank space is displayed between July 21 and September 2.

[0111] Visualization data 222(b) in Figure 22 and visualization data 222(c) in Figure 23 are different visualization data 222 from visualization data 222(a). For the period from July 1 to July 21, 2024, features (3), (5), and (6) constituted the first reference element cluster. For the period from September 2 to September 15, 2024, features (1), (7), and (8) constituted the second reference element cluster. For the period from July 1 to July 21, the average correlation coefficient of the first reference element cluster is 0.91. For the period from September 2 to September 15, the average correlation coefficient of the second reference element cluster is 0.90. The difference (0.1) between the average correlation coefficient of the first reference element cluster and the average correlation coefficient of the second reference element cluster is below the threshold. In this case, the order of placement of the first and second reference element clusters is not restricted. In visualization data 222(b), the first reference element cluster is located to the left of the second reference element cluster. In visualization data 222(c), the second reference element cluster is located to the left of the first reference element cluster.

[0112] ***Effects of Embodiment 2*** The analysis and monitoring device 100 analyzes the relationships between states or the relationships between states and feature quantities obtained from various observation data on a daily basis, accumulates the analysis results, and extracts and visualizes the analysis results that correspond to the conditions specified by the user in an easy-to-understand manner. This allows the user to easily grasp the situation on site.

[0113] Embodiment 3. This embodiment, which detects new relationships and visualizes them so that a list of relationships can be easily grasped, will be explained with reference to Figures 25 to 36, mainly showing the differences from Embodiments 1 and 2.

[0114] ***Configuration Description*** The configuration of the analysis monitoring device 100 will be described based on Figure 25. The analysis monitoring device 100 further includes a detection unit 150. The analysis monitoring program further causes a computer to function as the detection unit 150.

[0115] ***Explanation of Operation*** The analysis and monitoring method will be explained based on Figure 26. Steps S310 to S350 are performed for each target period and each data unit.

[0116] Steps S310 to S330 are the same as steps S210 to S230 in Embodiment 2.

[0117] In step S340, the detection unit 150 detects a new relationship based on the history of the determined relationship.

[0118] Step S340 will be explained in detail based on Figure 27. In step S341, the detection unit 150 refers to the relationship list data and determines whether the relationship list data is empty.

[0119] The relationship list data shows a list of determined relationships. Relationships in the relationship list data do not overlap. The relationship list data is stored in the storage unit 190.

[0120] If the relationship list data is empty (i.e., no relationships are listed in the relationship list data), the process proceeds to step S342. If the relationship list data is not empty (i.e., relationships are listed in the relationship list data), the process proceeds to step S343.

[0121] In step S342, the detection unit 150 refers to the relationship history data and records the relationships described in the relationship history data in the relationship list data.

[0122] Figure 28 shows the relationship history data 230 and the relationship list data 231. The relationship history data 230 is an example of relationship history data showing relationships for the period from April 1, 2024 to April 7, 2024. The relationship list data 231 is an example of relationship list data. If the relationship list data 231 is empty, the detection unit 150 writes the three relationships described in the relationship history data 230 into the relationship list data 231.

[0123] Returning to Figure 27, the explanation continues. After step S342, the process ends.

[0124] In step S343, the detection unit 150 refers to the relationship history data and determines whether the relationship determined in the current analysis (current relationship) has changed from the relationship determined in the previous analysis (previous relationship).

[0125] Figure 29 shows the relationship history data 232. Relationship history data 232 is an example of relationship history data showing relationships for the period from April 1, 2024 to April 21, 2024. In relationship history data 232, the "relationship between state and feature quantity" has not changed from relationship history data 230 (see Figure 28). In this case, the detection unit 150 determines that the current relationship has not changed from the previous relationship as of April 22.

[0126] Returning to Figure 27, let's continue the explanation. If the current relationship has not changed from the previous relationship (step S343), the process ends.

[0127] Figure 30 shows relationship history data 233. Relationship history data 233 is an example of relationship history data showing relationships for the period from April 1, 2024 to April 28, 2024. In relationship history data 233, the "relationship between state and feature" has changed from relationship history data 232 (see Figure 29). Specifically, the relationship "has a moderate correlation with the number of executions of program (2, 16, 18)" has increased. In this case, the detection unit 150 determines that the current relationship has changed from the previous relationship as of April 29.

[0128] Returning to Figure 27, let's continue the explanation. If the current relationship has changed from the previous relationship (step S343), the process proceeds to step S344.

[0129] In step S344, the detection unit 150 refers to the relationship history data and determines whether the current relationship is a new relationship. A new relationship is one that has been determined for the first time.

[0130] In the relationship history data 233 (see Figure 30), the relationships in this case are "strong correlation with the number of executions of programs (1, 4, 5)" and "moderate correlation with the number of executions of programs (2, 16, 18)". Of the relationships in this case, "moderate correlation with the number of executions of programs (2, 16, 18)" is a new relationship. In this case, the detection unit 150 determines that the relationship in this case is a new relationship as of April 29.

[0131] Returning to Figure 27, let's continue the explanation. If the relationship in question is a new relationship (step S344), the process proceeds to step S345.

[0132] In step S345, the detection unit 150 records the current relationship (new relationship) in the relationship list data.

[0133] Figure 32 shows the relationship list data 235. Relationship list data 235 is an example of relationship list data. Compared to relationship list data 231, relationship list data 235 has an increase in relationships with No. 2.

[0134] Returning to Figure 27, the explanation continues. After step S345, the process ends.

[0135] Figure 31 shows relationship history data 234. Relationship history data 234 is an example of relationship history data showing relationships for the period from April 1, 2024 to May 5, 2024. In relationship history data 234, the "relationship between state and feature" has changed from relationship history data 233 (see Figure 30). Specifically, the relationship "has a moderate correlation with the number of executions of program (2, 16, 18)" has decreased. In this case, the detection unit 150 determines that the current relationship has changed from the previous relationship as of May 6 (step S343). In relationship history data 234, the current relationship is "has a strong correlation with the number of executions of program (1, 4, 5)" and is not a new relationship. In this case, the detection unit 150 determines that the current relationship is not a new relationship as of May 6 (step S344).

[0136] Returning to Figure 27, let's continue the explanation. If the relationship in question is not a new relationship (step S344), the process ends.

[0137] Returning to Figure 26, step S350 will be explained. In step S350, the visualization unit 130 visualizes the determined relationship.

[0138] Step S350 will be explained in detail based on Figure 33. In step S351, the visualization unit 130 acquires analysis data for each relationship shown in the relationship list data.

[0139] In the relationship list data 235 (see Figure 32), for process A, there are three states: "Machine (1) started," "Machine (1) stopped," and "Machine (1) issued a warning," and two types of relationships are described for each of these states. Relationship No. 1 indicates a "strong correlation," so for analysis sets where the correlation coefficient is 0.7 or higher, analysis data for each feature (1, 4, 5) during the period in which the correlation coefficient was 0.7 or higher is obtained. Referring to the relationship history data 234 (Figure 31), the period in which features (1), (4), and (5) are relevant is all the weeks from April 1st to May 5th. Relationship No. 2 indicates a "moderate correlation," so for analysis sets where the correlation coefficient is between 0.4 and 0.7, analysis data for each feature (2, 16, 18) during the period in which the correlation coefficient was between 0.4 and 0.7 is obtained. Referring to the relationship history data 234 (Figure 31), the relevant period for features (2), (16), and (18) is the week from April 22 to April 28.

[0140] In step S352, the visualization unit 130 calculates the average of the correlation coefficients for the relevant period for each relationship shown in the relationship list data.

[0141] In the example described in step S351, the average of the correlation coefficients for all weeks from April 1st to May 5th is calculated for feature (1), feature (4), and feature (5). Additionally, the average of the correlation coefficients for the weeks from April 22nd to April 28th is calculated for feature (2), feature (16), and feature (18).

[0142] In step S353, the visualization unit 130 displays the average of the correlation coefficients for the relevant period for each relationship shown in the relationship list data.

[0143] Figure 34 shows an example of a relationship list (correlation coefficient) display. The relationship between process A and process B with state (1) and state (2) is visualized. The average of the correlation coefficients for each relationship is shown by a mark on the axis. State (1) is the state where "machine (1) is started". State (2) is the state where "machine (1) is stopped".

[0144] Returning to Figure 33, step S354 will be explained. In step S354, the visualization unit 130 displays the corresponding period for each relationship shown in the relationship list data, for the feature quantities related to the relationship.

[0145] Figure 35 shows an example of the display of the relationship list (relevant period). When the "Check Period" button for the relationship with state (1) related to process A is pressed in the relationship list (correlation coefficient) display in Figure 34, the relationship list (relevant period) in Figure 35 is displayed. The horizontal axis represents the period, and the vertical axis represents the average of the correlation coefficients.

[0146] Figure 36 shows an example of displaying a relationship list (for the relevant period). The period may be represented on the vertical axis and the average of the correlation coefficients on the horizontal axis.

[0147] Similarly, relationships other than those relating to state (1) of process A are also visualized for the relevant period.

[0148] ***Effects of Embodiment 3*** Embodiment 3 solves the problem that the relationship between the state and the features is unclear, and it is not easy to grasp a list of the relationships that have appeared. With Embodiment 3, the user can easily grasp a list of the relationships that have appeared. Furthermore, since the user can grasp the timing of the appearance of each relationship, they can easily grasp how the state of the system is changing or not changing.

[0149] ***Supplement to Embodiment 3*** Embodiment 3 shows an example of visualizing a list of relationships by marking the average value of the correlation coefficient on the axis and visualizing the classification of features for each relationship number. However, the method of visualizing a list of relationships is not limited to this method.

[0150] Embodiment 3 shows an example of accumulating a history of the relationship between states and features for a process. However, the unit of analysis for the relationship between states and features is not limited to processes. For example, the history of the relationship between states and features, the relationship between states, or the relationship between features may be accumulated for units such as machines, workers, or elemental tasks.

[0151] Embodiment 4. This embodiment compares the relationships obtained in the previous analysis with the relationships obtained in the current analysis and notifies the user if a change occurs in the relationships. The main differences from Embodiments 1 to 3 are explained based on Figures 37 to 42.

[0152] ***Configuration Description*** The configuration of the analysis monitoring device 100 will be described based on Figure 37. The analysis monitoring device 100 further includes a notification unit 160. The analysis monitoring program further causes a computer to function as the notification unit 160.

[0153] ***Explanation of Operation*** The analysis and monitoring method will be explained based on Figure 38. Steps S410 to S450 are performed for each target period and each data unit.

[0154] Steps S410 to S430 are the same as steps S210 to S230 in Embodiment 2.

[0155] In step S440, the detection unit 150 detects a new relationship based on the history of the determined relationship.

[0156] Step S440 will be described in detail based on Figure 39. Steps S441 to S443 are the same as steps S341 to S343 in Embodiment 3.

[0157] In step S444, the detection unit 150 instructs the notification unit 160 to notify it of the change in relationship. The notification unit 160 then notifies the user of the change in relationship. The change in relationship is notified, for example, by sound.

[0158] In the relationship history data 233 (see Figure 30), the "relationship between state and feature" has changed from the relationship history data 232 (see Figure 29). Specifically, the relationship "has a moderate correlation with the number of executions of program (2, 16, 18)" has increased. In this case, the notification unit 160 notifies the user that the relationship has changed and provides information about the changed relationship (process A, state (1), state (2), state (3)).

[0159] Let's continue the explanation of step S444. The notification unit 160 also instructs the visualization unit 130 to display the change in the relationship.

[0160] Steps S445 and S446 are the same as steps S344 and S345 in Embodiment 3.

[0161] Returning to Figure 38, step S450 will be explained. In step S450, the visualization unit 130 visualizes the determined relationship. Step S450 is the same as step S350 in Embodiment 3. However, if the display of a change in the relationship is instructed, the visualization unit 130 displays the change in the relationship.

[0162] Figure 40 shows an example of displaying relationship changes. For process A, there are changes in the relationships with state (1), state (2), and state (3), and warnings indicating these changes are displayed superimposed on a heat map as shown in Figure 16. The heat map is omitted. For process B, there are no changes in relationships, and no warnings are displayed for process B.

[0163] Figure 41 shows an example of the display of the relationship list (correlation coefficient). Warnings indicating changes in relationships may be superimposed on the display of the relationship list (correlation coefficient).

[0164] Figure 42 shows an example of displaying relationship changes. Relationship changes are displayed for each worker. The relationship display is omitted. The data unit is not limited to processes, but may be a worker or other unit.

[0165] ***Effects of Embodiment 4*** Embodiment 4 solves the problem of not noticing changes in relationships. The analysis monitoring device 100 analyzes the relationships between states or the relationships between each state and feature quantities obtained from various observation data on a daily basis, and notifies the user of any changes in relationships through warning displays and sounds. This allows the user to know whether the system is continuing in the same state or not, to notice new relationships when new factors that have not appeared as negative factors for the system before appear, and to respond quickly to improve the system. Examples of negative factors include factors that reduce the work efficiency of the manufacturing site. Also, when there are many relationship monitoring targets, the user can immediately notice when relationships have changed. Examples of monitoring targets include analysis results for each worker, analysis results for each process, and analysis results for each state.

[0166] ***Supplement to Embodiment 4*** Embodiment 4 shows an example of superimposing a warning to indicate that a change has occurred. However, the indication of a change is not limited to this example. For example, a flashing display or other method may be used to indicate a change.

[0167] Notifications to the user may be made using sounds such as beeps, or they may not be made using sounds.

[0168] ***Supplement to the Embodiment*** The heatmap may have the time axis (period) on the vertical axis and the feature quantities on the horizontal axis, or the feature quantities on the vertical axis and the time axis (period) on the horizontal axis. Visualization of relationships is not limited to heatmaps. For example, changes in relationships may be visualized using a method that shows the classification of feature quantities or by other methods.

[0169] In this embodiment, an example was shown in which the number of times each program number executed by the machine is used as a feature, that is, an example in which the "number of times program i (i=1 to n) is executed" is used as feature i (i=1 to n). Then, an example was shown in which the correlation coefficient between the number of occurrences of each state and each feature (the number of executions of each program) is calculated for each process. However, the items for which the correlation coefficient is calculated are not limited to the number of occurrences of each state and the number of executions of each program.

[0170] The metrics for a state are not limited to the number of occurrences. For example, the metrics for a state may be the duration of the state or the frequency of the state. The features are not limited to the number of times the program has been executed. For example, the features may be the execution time or the execution frequency. The features are not limited to information about the program. For example, the features may be statistical values ​​calculated from the values ​​of various observed sensor data.

[0171] The relationships that can be visualized are not limited to the relationships between state groups and features. For example, relationships between state groups or between features may also be visualized. It is possible to visualize changes in these relationships.

[0172] The length of the period to be analyzed is not limited to one week. For example, the length of the period may be two weeks or one month. The period may be entered by the user using an input device, or it may be predefined in a definition file stored in the storage unit 190.

[0173] The method for selecting the display targets (period and features) is not limited to selection based on user specifications or selection based on a combination of period and correlation coefficient.

[0174] The target systems are not limited to systems in manufacturing sites. The analysis and monitoring method of the embodiment is a technology that analyzes the characteristics and changes (transitions) of a system whose characteristics are unknown, and visualizes the characteristics of the system so that they can be easily understood, and can be applied to any system.

[0175] Based on Figure 43, the hardware configuration of the analysis and monitoring device 100 will be described. The analysis and monitoring device 100 includes a processing circuit 109. The processing circuit 109 is hardware that implements a state group definition unit 110, a relationship analysis unit 120, a visualization unit 130, a relationship storage unit 140, a detection unit 150, and a notification unit 160. The processing circuit 109 may be dedicated hardware, or it may be a processor 101 that executes a program stored in memory 102.

[0176] If the processing circuit 109 is dedicated hardware, the processing circuit 109 may be, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC, an FPGA, or a combination thereof. ASIC is an abbreviation for Application Specific Integrated Circuit. FPGA is an abbreviation for Field Programmable Gate Array.

[0177] The analysis and monitoring device 100 may include multiple processing circuits that replace the processing circuit 109.

[0178] In the processing circuit 109, some functions may be implemented by dedicated hardware, while the remaining functions may be implemented by software or firmware.

[0179] Thus, the functions of the analysis and monitoring device 100 can be realized by hardware, software, firmware, or a combination thereof.

[0180] Each embodiment is an example of a preferred form and is not intended to limit the technical scope of this disclosure. Each embodiment may be implemented in part or in combination with other embodiments. Procedures described using flowcharts, etc., may be modified as appropriate.

[0181] The "part" in each element of the analysis and monitoring device 100 may be read as "processing," "process," "circuit," or "circuit."

[0182] 100 Analysis and monitoring device, 101 Processor, 102 Memory, 103 Auxiliary storage device, 104 Communication device, 105 Input / output interface, 109 Processing circuit, 110 State group definition unit, 120 Relationship analysis unit, 130 Visualization unit, 140 Relationship storage unit, 150 Detection unit, 160 Notification unit, 190 Storage unit, 210 State data, 211 Correlation coefficient data, 212 Clustering data, 213 Visualization data, 214 Visualization data, 220 Relationship history data, 221 Relationship history data, 222 Visualization data, 230 Relationship history data, 231 Relationship list data, 232 Relationship history data, 233 Relationship history data, 234 Relationship history data, 235 Relationship list data.

Claims

1. An analysis monitoring device comprising a relationship analysis unit that uses feature data showing multiple features of a target system for a target period and state data showing the state of each component of the target system at each time point in the target period to determine the relationships between elements of at least one of a plurality of analysis sets, which are either a plurality of sets consisting of a feature and a state group consisting of one or more states, a plurality of sets consisting of state groups as elements, or a plurality of sets consisting of feature items as elements.

2. The analysis and monitoring device according to claim 1, further comprising a state group definition unit that defines a plurality of state groups in which one or more states of one or more of the plurality of components are different from each other, using the state data.

3. The analysis monitoring device according to claim 1 or 2, wherein the relationship analysis unit calculates an index value for each state group during the target period using the state data, calculates a correlation coefficient between the feature and the index value for each analysis set in which the feature and the state group are elements using the feature data, and determines the relationship based on the correlation coefficient of each of the plurality of analysis sets.

4. The analysis monitoring device according to claim 1 or 2, wherein the relationship analysis unit calculates an index value for each state group during the target period using the state data, calculates a correlation coefficient between the index values ​​for each analysis set in which the state groups are elements, and determines the relationship based on the correlation coefficients of each of the plurality of analysis sets.

5. The analysis monitoring device according to claim 1 or 2, wherein the relationship analysis unit calculates the correlation coefficient between features for each analysis set in which the features are elements, using the feature data, and determines the relationship based on the correlation coefficients of each of the plurality of analysis sets.

6. The analysis monitoring device according to any one of claims 3 to 5, wherein the relationship analysis unit clusters the multiple reference elements of the multiple analysis sets with respect to a reference element which is one element of each analysis set based on the correlation coefficient of each of the multiple analysis sets, and determines the relationship of each matching element to the reference element cluster for each reference element cluster with respect to the other element of each analysis set based on the cluster link result.

7. The analysis and monitoring device according to claim 6, further comprising a visualization unit that displays a plurality of reference elements belonging to each reference element cluster in a continuous sequence.

8. The analysis and monitoring device according to claim 7, wherein the visualization unit determines the order in which the multiple reference element clusters are arranged and displays the multiple reference element clusters in order.

9. The analysis monitoring device according to claim 7 or claim 8, wherein the visualization unit displays a plurality of reference element clusters for each of the plurality of target periods in order of the duration of the plurality of target periods.

10. The analysis monitoring device according to claim 9, comprising a relationship storage unit that records the relationships in relationship history data for each target period, wherein the visualization unit determines the order of the plurality of reference element clusters for each target period by referring to the relationship history data.

11. The analysis monitoring device according to claim 6, comprising: a relationship storage unit that records the relationships in relationship history data for each target period; a detection unit that detects new relationships from the relationship history data for each target period and records the new relationships in relationship list data; and a visualization unit that displays one or more reference element clusters related to each relationship using the relationship list data.

12. The analysis and monitoring device according to claim 11, wherein the visualization unit displays the degree of the relationship together with the reference element cluster for each reference element cluster.

13. The analysis monitoring device according to claim 11 or claim 12, wherein the visualization unit displays the period during which the relationship conditions were met for each of the reference element clusters.

14. The analysis monitoring device according to any one of claims 11 to 13, wherein the analysis monitoring device comprises a notification unit, the detection unit determines whether the current relationship has changed from the previous relationship by referring to the relationship history data for each target period, and the notification unit notifies the user of the change in the relationship when it is determined that the current relationship has changed from the previous relationship.

15. The analysis monitoring device according to claim 14, wherein the visualization unit determines that the current relationship has changed from the previous relationship, and displays that there has been a change in the relationship along with one or more reference element clusters relating to the changed relationship.

16. The analysis and monitoring device according to any one of claims 7 to 15, wherein the visualization unit displays information relating to the reference element cluster limited to the information of the specified display target.

17. The analysis monitoring device according to any one of claims 7 to 15, wherein the visualization unit selects a display target based on at least one of the reference period and the correlation coefficient of each analysis group, and displays information relating to the reference element cluster limited to the information of the selected display target.

18. An analysis monitoring method for determining the relationships between elements of at least one of a plurality of analysis sets, which are either a plurality of sets consisting of a feature quantity and a state group consisting of one or more states, a plurality of sets consisting of state groups as elements, or a plurality of sets consisting of feature quantities as elements. This method uses feature quantity data that shows a plurality of features of a target system for a target period and state data that shows the state of each component of the plurality of components of the target system at each time point in the target period.

19. An analysis monitoring program that causes a computer to perform a relationship analysis process to determine the relationships between elements of at least one of a plurality of analysis sets, which are either a plurality of sets consisting of a feature quantity and a state group consisting of one or more states, a plurality of sets consisting of state groups as elements, or a plurality of sets consisting of feature quantities as elements. This analysis set uses feature quantity data that shows a plurality of features of the target system for the target period and state data that shows the state of each component of the plurality of components of the target system for each time point in the target period.

Images