Facility operation status monitoring system and facility operation status monitoring method

Abstract

The system accurately detects signs of abnormality occurring in individual pieces of equipment 3 based on whether the equipment is maintained in a harmonious state, and further accurately detects signs of abnormality occurring in a group of equipment including a plurality of pieces of equipment 3 based on whether the group of equipment is maintained in a harmonious state. [Solution] The system comprises a first communication unit 11c that receives status information measured by at least one sensor 5 installed in a specified facility 3, and a first control unit 11a that derives reference status information that represents the status information during a specified period in the past, stores the reference status information in a memory unit 11b or a server memory unit 30b, and generates current status information that represents the status information after the specified period in the past, and the first control unit 11a determines the status of the facility 3 based on the generated current status information and the stored reference status information.

Description

[Technical Field]

[0001] The present invention relates to a facility operation state monitoring device, a facility operation state monitoring system, and a facility operation state monitoring method for monitoring the operation state of a facility or a group of facilities including a plurality of facilities in a factory, plant, or the like. [Background technology]

[0002] In recent years, there has been a demand for smart factories to address a variety of issues, such as responding to labor shortages caused by the declining birthrate and aging population, responding to shortened product lifecycles and high-mix low-volume production due to diversifying consumer needs, stabilizing quality and ensuring traceability in the manufacturing process, and further improving energy efficiency and reducing environmental impact in order to realize a decarbonized society.

[0003] Prerequisites for making factories smarter include the digitalization of manufacturing sites and stable equipment operation. The primary target of digitalization in manufacturing sites is manufacturing process information. Technologies for acquiring and managing manufacturing process information, such as operation control systems and MES (Manufacturing Execution Systems), have been introduced. The digital data used in these systems includes, for example, operating conditions, manufacturing information, and current / temperature information for manufacturing equipment (devices). To accurately detect signs of abnormalities in equipment or equipment groups consisting of multiple pieces of equipment, it is desirable to acquire diverse information from manufacturing equipment. However, incorporating a wide variety of sensors into manufacturing equipment in advance increases costs. Therefore, adopting sensors that can be retrofitted to manufacturing equipment and can wirelessly transmit acquired data is effective for digitalizing manufacturing sites.

[0004] To ensure stable operation of equipment or groups of equipment, anomaly detection often relies on the intuition and experience of skilled workers. Such anomaly detection involves checking on-site indicators, sounds, vibrations, odors, etc. during daily inspection patrols. Threshold management is also carried out for the output of current values ​​and temperature sensors in manufacturing equipment. Anomaly detection here refers to capturing (detecting) the occurrence of an "abnormal state." However, to ensure stable operation of equipment, it is necessary to perform appropriate maintenance on manufacturing equipment to prevent it from reaching an "abnormal state." In other words, to ensure stable operation of equipment, it is important to detect signs of an "abnormal state" before it reaches an "abnormal state."

[0005] As a technology for detecting signs of failure, for example, a failure sign detection system is known that includes a server that stores data indicating the state of equipment, and the server performs a short-time Fourier transform (STFT) on the data to generate a spectrogram, which is three-dimensional data of time x frequency x sound pressure, performs non-negative matrix factor analysis (NMF) on the spectrogram to extract time series clusters containing time series components and frequency clusters containing frequency components, and performs analysis processing to determine whether or not there is an abnormality in the equipment based on the time series clusters and the frequency clusters (Patent Document 1).

[0006] According to Patent Document 1, a system is provided that collects raw data indicating the status of small pumps, motors, machines, precision equipment, etc. from inexpensive sensors often used in IoT that are installed in each of them, automatically transmits the raw data to a server (cloud) via an IoT gateway device, stores the raw data in a storage device, and quickly and accurately analyzes the large amount of raw data stored in the storage device. Furthermore, Patent Document 1 states that a deterioration curve ranging from normal to abnormal is obtained based on time-series anomaly index data obtained by repeatedly executing a judgment process, and that the deterioration curve is predicted to be approximate if the evaluation object is the same, and that by collecting and evaluating data from the same type of evaluation object, it is possible to predict the time of failure.

[0007] Also known is a failure sign detection system that includes a data storage unit that stores data on normal operation of each manufacturing operation of a manufacturing device consisting of multiple operations as learning data, a sensor that measures the manufacturing operations of the manufacturing device, an operation detection unit that detects the start of each operation in the manufacturing operation, a divided data collection unit that divides the measurement data measured by the sensor into divided data for each operation and collects it, and a data analysis unit that analyzes abnormalities in each operation based on a comparison of the divided data for each operation with the learning data (Patent Document 2).

[0008] According to Patent Document 2, even when the manufacturing operation of a manufacturing device is made up of a plurality of operations, it is possible to detect abnormalities on an operation-by-operation basis with high accuracy. [Prior art documents] [Patent documents]

[0009] [Patent Document 1] Japanese Patent Application Publication No. 2019-067197 [Patent Document 2] Japanese Patent Application Publication No. 2020-154896 DISCLOSURE OF THE INVENTION [Problem to be solved by the invention]

[0010] However, the technologies disclosed in Patent Documents 1 and 2 only detect signs of failure by processing the output of a specific sensor, and do not mention detecting signs of failure (abnormality) in individual manufacturing equipment by focusing on the balance (harmony) of the outputs of multiple sensors. Furthermore, the conventional technologies only detect signs of failure in individual manufacturing equipment, and do not anticipate detecting signs of abnormality in an equipment group including multiple manufacturing equipment, a group of equipment groups including multiple equipment groups, or even in a higher-level group.

[0011] The present invention has been devised to solve the problems of the prior art, and its purpose is to accurately detect signs of abnormalities occurring in individual pieces of equipment based on whether the equipment is maintained in a harmonious state, and to further accurately detect signs of abnormalities occurring in a group of equipment including multiple pieces of equipment based on whether the group is maintained in a harmonious state. [Means for solving the problem]

[0012] The present invention, which has been made to solve the above problems, is an equipment operation status monitoring device comprising: a first communication unit that receives status information measured by at least one sensor provided in a predetermined facility; and a first control unit that derives reference status information that represents the status information for a predetermined period in the past, stores the reference status information in a predetermined storage unit, and generates current status information that represents the status information after the predetermined period in the past, wherein the first control unit determines the status of the facility based on the generated current status information and the stored reference status information. This makes it possible to accurately detect signs of abnormalities in individual facilities based on whether the facilities are maintaining a harmonious state.

[0013] In addition, in the present invention, the first control unit performs eigenvalue analysis on the state information and derives the reference state information and the current state information by reducing the dimension of the state information based on a first eigenvector and a second eigenvector extracted by the eigenvalue analysis, thereby making it possible to easily detect signs of abnormality occurring in individual pieces of equipment.

[0014] In addition, the present invention is a method for controlling the first control unit to calculate T for the reference state information and the current state information. 2 The degree of dispersion V is calculated based on the statistics, the degree of harmony H is calculated based on the Q statistics, and whether or not a harmonious state of the equipment is maintained is determined as the state of the equipment based on the degree of dispersion V and the degree of harmony H. This makes it possible to easily determine the harmony of the equipment 3.

[0015] In addition, in the present invention, the sensor measures at least one of acceleration information, geomagnetic information, vibration information, sound information, temperature information, humidity information, odor information, current information, and image information, thereby making it possible to appropriately select the sensor to be used for measurement and accurately detect the state of the equipment.

[0016] In addition, in the present invention, the sensor is retrofitted to the equipment and transmits the status information to the first communication unit wirelessly, thereby enabling the digitalization of manufacturing sites at low cost.

[0017] The present invention also provides an equipment operational status monitoring system comprising: a functional first layer including a plurality of first equipment operational status monitoring devices that are provided in each piece of equipment and monitor the status of the equipment based on the output of at least one sensor that measures status information of the equipment; and a functional second layer including at least one second equipment operational status monitoring device that is provided as a layer above the functional first layer and that monitors the status of a first equipment group formed by grouping a plurality of pieces of equipment based on the output of the plurality of first equipment operational status monitoring devices included in the functional first layer. This enables the second equipment operational status monitoring device to accurately detect signs of an abnormality in the equipment group, based on whether the equipment group including a plurality of pieces of equipment 3 is maintained in a harmonious state.

[0018] Furthermore, the present invention comprises a functional N-th layer, where N is an integer equal to or greater than 2, that is provided as a layer above the functional N-1-th layer and includes at least one N-th equipment operational status monitoring device that monitors the status of a group of equipment as a monitoring target based on outputs from multiple N-1-th equipment operational status monitoring devices included in the functional N-1-th layer. This makes it possible to accurately detect signs of abnormalities on the scale of a facility, a group of equipment including multiple facilities, a group including multiple equipment groups, or a group including a group of multiple equipment groups.

[0019] Furthermore, the present invention standardizes the dimensions of the outputs of the first equipment operational status monitoring device through the N-1 equipment operational status monitoring device, thereby enabling the configuration of the equipment operational status monitoring devices to be standardized across all layers, thereby reducing system costs.

[0020] In addition, in the present invention, the first equipment operation status monitoring device performs eigenvalue analysis on the input state information and outputs new state information by reducing the dimension of the state information using first and second eigenvectors extracted in descending order of eigenvalues, and each of the second to N-1 equipment operation status monitoring devices performs eigenvalue analysis on the state information output by the operation status monitoring device included in the layer immediately below and outputs new state information by reducing the dimension of the state information using first and second eigenvectors extracted in descending order of eigenvalues. This enables multi-scale state monitoring by individual equipment state monitoring and area state monitoring.

[0021] In addition, the present invention provides a first equipment operational state monitoring device that includes a first communication unit that receives the state information measured by the sensor, and a first control unit that derives reference state information that represents the state information for a predetermined period in the past and stores it in a predetermined storage unit, and generates current state information that represents the state information after the predetermined period in the past, and the first control unit determines the state of the equipment based on the generated current state information and the stored reference state information. This makes it possible to accurately detect signs of abnormality in individual equipment based on whether the equipment is maintained in a harmonious state.

[0022] Furthermore, in the present invention, the second equipment operational state monitoring device receives the reference state information and the current state information derived by the plurality of first equipment operational state monitoring devices. This enables the second monitoring device belonging to the upper layer to accurately detect signs of abnormality for each grouped piece of equipment (i.e., equipment group) based on the state quantity derived in the functional first layer (newly derived state information). It becomes Noh.

[0023] Furthermore, in the present invention, where N is an integer equal to or greater than 2, the N-th equipment operational state monitoring device included in the functional N-th layer comprises a second communication unit that receives the reference state information and the current state information output from the N-1 equipment operational state monitoring device included in the functional N-1 layer that is a layer below the functional N-th layer, and a second control unit that derives new reference state information that represents the reference state information and stores it in a predetermined storage unit, and generates new current state information that represents the current state information, where the second control unit determines the state of the equipment group as the monitoring target based on the generated new current state information and the stored new reference state information. This makes it possible to accurately detect signs of abnormalities on the scale of an equipment, an equipment group including multiple equipment, a group including multiple equipment groups, or a group including groups of multiple equipment groups.

[0024] Further, the present invention provides a method for monitoring the Nth equipment operation state, in which the Nth equipment operation state monitoring device performs T 2 The degree of dispersion V is calculated based on the statistics, the degree of harmony H is calculated based on the Q statistics, and whether or not a harmonious state of the equipment group is maintained is determined as the state of the equipment group based on the degree of dispersion V and the degree of harmony H. This makes it possible to easily determine the harmony of the equipment group.

[0025] Furthermore, in the present invention, the equipment operation status monitoring device included in a layer above the first functional layer further receives the status information measured by the sensor, making it possible to determine the status of the equipment group by referring to the temperature, humidity, odor, vibration, etc. of the entire environment in which the equipment group consisting of multiple pieces of equipment is installed.

[0026] The present invention also provides an equipment operational status monitoring system including at least one sensor provided in each piece of equipment for measuring status information of the equipment, a first equipment operational status monitoring device that derives primary reference status information representative of the status information during a predetermined period in the past and primary current status information representative of the status information during a monitoring period after the predetermined period in the past, and determines the status of each piece of equipment based on the primary reference status information and the primary current status information, and a monitoring unit that derives secondary reference status information representative of the primary reference status information for an equipment group formed by grouping multiple pieces of equipment based on the primary reference status information output by multiple first equipment operational status monitoring devices, derives secondary current status information representative of the primary current status information for the equipment group based on the primary current status information output by the multiple first equipment operational status monitoring devices, and determines the status of the equipment group based on the secondary reference status information and the secondary current status information. This makes it possible to accurately detect signs of abnormality in each piece of equipment based on whether the equipment is maintained in a harmonized state, and further accurately detect signs of abnormality in each equipment group based on whether the equipment group including multiple pieces of equipment is in harmony.

[0027] The present invention also provides an equipment operational status monitoring system including: at least one sensor provided in each piece of equipment for measuring status information of the equipment; a monitoring unit that receives the status information over a predetermined period in the past and derives primary reference status information representative of the status information and secondary reference status information representative of the primary reference status information; a first equipment operational status monitoring device that acquires the primary reference status information from the monitoring unit, derives primary current status information representative of the status information during a monitoring period after the predetermined period in the past, and determines the status of the equipment based on the primary reference status information and the primary current status information; and a second equipment operational status monitoring device that acquires the secondary reference status information from the monitoring unit, derives secondary current status information representative of the primary current status information output by the first equipment operational status monitoring device during the monitoring period, and determines the status of an equipment group formed by grouping multiple pieces of equipment based on the secondary reference status information and the secondary current status information. This allows each monitoring device 1 to function as an edge device and perform monitoring processing on an equipment 3 or a group of equipment in real time.

[0028] Further, the present invention provides a method for monitoring the operation state of the first equipment, the method comprising: 2 The degree of dispersion V is calculated based on the statistics, the degree of harmony H is calculated based on the Q statistics, and whether or not a harmonious state of the equipment is maintained is determined as the state of the equipment based on the degree of dispersion V and the degree of harmony H. This makes it possible to easily determine the harmony of the equipment 3.

[0029] Further, the present invention provides a method for monitoring the operation state of the second equipment, the method comprising: 2 The degree of dispersion V is calculated based on the statistics, the degree of harmony H is calculated based on the Q statistics, and whether or not a harmonious state of the equipment group is maintained is determined as the state of the equipment group based on the degree of dispersion V and the degree of harmony H. This makes it possible to easily determine the harmony of the equipment group.

[0030] The present invention also provides an equipment operation status monitoring method that measures status information of individual equipment, derives primary reference status information representative of the status information during a predetermined period in the past based on the status information, derives primary current status information representative of the status information during a monitoring period after the predetermined period in the past based on the status information, determines the status of each equipment based on the primary reference status information and the primary current status information, derives secondary reference status information representative of the primary reference status information for an equipment group formed by grouping multiple pieces of equipment based on multiple pieces of primary reference status information, derives secondary current status information representative of the primary current status information for the equipment group based on the multiple pieces of primary current status information, and determines the status of the equipment group based on the secondary reference status information and the secondary current status information. This makes it possible to accurately detect signs of abnormality in each equipment based on whether the equipment is maintained in a harmonious state, and further accurately detect signs of abnormality in each equipment group based on whether an equipment group including multiple pieces of equipment is harmonious.

[0031] The present invention also provides an equipment operation status monitoring method that measures status information of individual equipment, receives the status information over a predetermined period in the past, derives primary reference status information representative of the status information and secondary reference status information representative of the primary reference status information, acquires the primary reference status information, derives primary current status information representative of the status information during a monitoring period after the predetermined period in the past, determines the status of the equipment based on the primary reference status information and the primary current status information, acquires the secondary reference status information, derives secondary current status information representative of the primary current status information during the monitoring period, and determines the status of an equipment group formed by grouping a plurality of the equipment based on the secondary reference status information and the secondary current status information. This makes it possible to perform monitoring processing on equipment 3 or an equipment group in real time. [Effects of the Invention]

[0032] According to the present invention, it is possible to accurately detect signs of an abnormality occurring in individual pieces of equipment based on whether the equipment is maintained in a harmonious state.Furthermore, it is possible to accurately detect signs of an abnormality occurring in a group of equipment including a plurality of pieces of equipment based on whether the group of equipment is maintained in a harmonious state. [Brief explanation of the drawings]

[0033] [Figure 1] FIG. 1 is an explanatory diagram showing an example of a manufacturing line 90 to which a facility operation state monitoring system Sy1 according to a first embodiment of the present invention is applied. [Figure 2] Diagram showing the monitoring levels in EBSM [Figure 3] Diagram showing the hierarchical structure of EBSM [Figure 4] EBSM configuration diagram focusing on the first monitoring device 1a included in the first functional layer L1 [Figure 5] EBSM configuration diagram focusing on the second monitoring device 1b included in the functional second layer L2 [Figure 6] Block diagram showing the functions of the monitoring device 1 [Figure 7] An explanatory diagram showing an example of feature extraction [Figure 8] EBSM state map illustration [Figure 9] EBSM-VH map [Figure 10] FIG. 10 is an explanatory diagram illustrating an example of individual device status monitoring for each robot and area status monitoring for a group of robots 60. [Figure 11] FIG. 10 is an explanatory diagram illustrating an example of individual device status monitoring for each conveyor and area status monitoring for a group of conveyors 70. [Figure 12] FIG. 10 is an explanatory diagram illustrating an example of individual equipment status monitoring for each AGV and area status monitoring for an AGV group 80. [Figure 13] FIG. 1 is an explanatory diagram illustrating an example of monitoring the state of an area in a production line 90. [Figure 14] FIG. 10 is an explanatory diagram showing an example of harmony determination according to a second embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0034] (First embodiment) FIG. 1 is an explanatory diagram showing an example of a production line 90 to which an equipment balance state monitoring system Sy1 according to a first embodiment of the present invention is applied. Hereinafter, the equipment balance state monitoring system Sy1 may be referred to as EBSM (Equipment Balance State Monitoring). The EBSM monitors the state of each piece of equipment 3 (such as manufacturing equipment) included in the production line 90, and also performs monitoring processing to monitor the state of an equipment group consisting of multiple pieces of equipment 3 (as will be described later, groups of equipment groups are also subject to monitoring processing). When the EBSM detects a sign of an abnormality occurring in a piece of equipment 3 or a group of equipment, it notifies a manager of the production line 90 or the like that a sign has been detected.

[0035] 1, a production line 90 includes, as equipment 3, a first robot 61 to a fourth robot 64, which are so-called industrial robots, a first conveyor 71 to a third conveyor 73 that transport work-in-progress and the like along the production process, and a first AGV 81 (AGV: Automated Guided Vehicle) and a second AGV 82, which are unmanned guided vehicles that transport raw materials, parts, finished products, and the like. Hereinafter, the first robot 61 to the fourth robot 64 may be collectively referred to as a robot group 60, the first conveyor 71 to the third conveyor 73 may be collectively referred to as a conveyor group 70, and the first AGV 81 and the second AGV 82 may be collectively referred to as an AGV group 80. The robot group 60, the conveyor group 70, and the AGV group 80 each constitute an equipment group, and further, the robot group 60, the conveyor group 70, and the AGV group 80 collectively constitute a higher-level equipment group (i.e., in this case, the production line 90).

[0036] Of course, other equipment 3 may be introduced into production line 90. Examples of other equipment 3 include machine tools, presses, injection molding machines, cutting machines, polishing machines, press-fitting machines, welding machines, surface mounting machines, visual inspection machines, dimensional inspection machines, electrical characteristic inspection machines, X-ray inspection machines, image processing devices, AMRs (Autonomous Mobile Robots), cranes, lifters, motors, inverters, cleaning machines, dryers, cooling machines, heating machines, packaging machines, and labelers. There are no particular limitations on the combination of equipment 3 introduced into production line 90, the number of pieces of equipment 3, or the number of equipment groups.

[0037] Each piece of equipment 3 is provided with a corresponding equipment operational state monitoring device (hereinafter, sometimes simply referred to as a "monitoring device 1"). In the following description, the monitoring device 1 that monitors the individual state of equipment 3 included in the production line 90 may be referred to as a first equipment operational state monitoring device (first monitoring device 1a). As shown in FIG. 1, each piece of equipment 3 is provided with a corresponding first monitoring device 1a. Furthermore, a second equipment operational state monitoring device (second monitoring device 1b) is provided as a monitoring device 1 that belongs to a higher level (upper layer) than the multiple first monitoring devices 1a. The second monitoring devices 1b (three in this case) correspond to the robot group 60, the conveyor group 70, and the AGV group 80, which are the first equipment group 3a, respectively, and monitor the state of each equipment group.

[0038] Figure 2 is an explanatory diagram showing the monitoring levels in EBSM. In Figure 2, the state of a facility 3 or a group of facilities is shown as layers, with the degree of abnormality or urgency increasing toward the outside. EBSM determines whether a facility 3 or a group of facilities consisting of multiple facilities 3 is "maintaining a harmonious state" (whether it is within the harmonious region HA described below), "whether it is in a state outside the harmonious region HA," or "whether it is within the range where an unprecedented state occurs." Here, "harmonious state" refers to a state in which the relative relationship between the outputs of the multiple sensors 5 constituting the sensor group 5S (see Figures 3 and 4) for a facility 3 is within a predetermined range, or a state in which the relative relationship between the states of each facility 3 constituting the facility group is within a predetermined range for a group of facilities.

[0039] Furthermore, a "state outside the harmonic region HA" refers to a state that deviates from the harmonic region HA but is still considered to be within the normal range (i.e., a state that has been measured before), and an "unexperienced state" refers to a state that deviates even further from the "state outside the harmonic region HA" (a state that has never been measured before), but is not necessarily an abnormality (i.e., a state in which a process alarm or the like is not issued). On the other hand, states in which an obvious abnormality is detected in equipment 3 or a group of equipment and an alarm is issued from the equipment 3 or the like (process alarm), and states in which the operation of the equipment 3 or a group of equipment is stopped due to the abnormality (interlock) are excluded from the states monitored by EBSM. Whether to issue a process alarm or execute an interlock is determined individually by each equipment 3, for example, based on the output of the sensor 5, etc.

[0040] FIG. 3 is an explanatory diagram showing the hierarchical structure of the EBSM. As shown in the figure, the EBSM has a hierarchical structure consisting of N layers (N is an integer equal to or greater than 1), including a functional first layer L1 as the lowest layer to a functional Nth layer LN as the highest layer. Each layer includes one or more monitoring devices 1. The relationship between the monitoring devices 1 between layers (relationships focused on the transmission and reception of information) may be stored as branch tree structure information in the server storage unit 30b (see FIGS. 4 and 5) or the information terminal 40. In this case, a monitoring device 1 can obtain monitoring-related information (reference state information and current state information, which will be described later; these may be collectively referred to as "monitoring-related information") derived by a monitoring device 1 included in the layer immediately below it by inquiring of the server 30 or the information terminal 40 about its own monitoring device ID.

[0041] An example of the configuration of EBSM will be described below with reference to FIG. 3. In EBSM, the functional first layer L1 includes first monitoring devices 1a (1a_1) to 1a (1a_4). Note that the number of first monitoring devices 1a is not limited. The robots, conveyors, and AGVs described above each correspond to a facility 3. The robot group 60 and conveyor group 70 correspond to a first facility group 3a, respectively. The first facility group 3a (3a_1) may include more facilities 3. The first monitoring device 1a belonging to the lowest functional first layer L1 receives measurement results (status information) from sensors 5 (sensor group 5S) installed in the facility 3 to be monitored. Based on the received status information, the first monitoring device 1a detects whether or not there is a sign of an abnormality occurring in the facility 3. That is, in the functional first layer L1, one first monitoring device 1a corresponds to one facility 3, and each first monitoring device 1a determines whether or not there is a sign of an abnormality occurring in the corresponding facility 3. Specifically, each first monitoring device 1a determines whether there are any signs of an abnormality occurring for each of the first robot 61, second robot 62, etc., first conveyor 71, second conveyor 72, etc., first AGV 81, and second AGV 82 (all see Figure 1) associated with that first monitoring device 1a.

[0042] A functional second layer L2 is provided above the functional first layer L1. A second monitoring device 1b belonging to the functional second layer L2 receives monitoring-related information from a first monitoring device 1a (here, the first monitoring device 1a (1a_1) and the first monitoring device 1a (1a_2)) belonging to the lower layer. Then, based on the monitoring-related information, the second monitoring device 1b determines whether or not there are signs of an abnormality in a first equipment group 3a consisting of multiple pieces of equipment 3 (here, equipment 3 (3_1) and equipment 3 (3_2)). That is, in the functional second layer L2, one monitoring device 1 corresponds to an equipment group consisting of multiple pieces of equipment 3, and each monitoring device 1 determines whether or not there are signs of an abnormality in each equipment group.

[0043] Specifically, in the second functional layer L2, the presence or absence of signs of an abnormality is determined for each equipment group (first equipment group 3a) such as the robot group 60 consisting of the first robot 61, the second robot 62, etc., the conveyor group 70 consisting of the first conveyor 71, the second conveyor 72, etc., and the AGV group 80 consisting of the first AGV 81, the second AGV 82, etc. In other words, the robot group 60, the conveyor group 70, and the AGV group 80 are each part of the first equipment group 3a. There is no particular limit to the number of equipment groups that make up the first equipment group 3a.

[0044] A functional third layer L3 is provided above the functional second layer L2. The third monitoring device 1c belonging to the functional third layer L3 receives monitoring-related information from the second monitoring device 1b (here, the second monitoring device 1b (1b_1) and the second monitoring device 1b (1b_2)) belonging to the layer below. Based on the monitoring-related information, the third monitoring device 1c determines whether or not there are signs of an abnormality for the equipment group including the first equipment group 3a (3a_1) and the first equipment group 3a (3a_2). That is, in the functional third layer L3, one monitoring device 1 (third monitoring device 1c) corresponds to the first equipment group 3a (3a_1) obtained by grouping the equipment 3 and the second equipment group 3b obtained by grouping the other first equipment group 3a (3a_2), and each third monitoring device 1c determines whether or not there are signs of an abnormality for the corresponding second equipment group 3b. Specifically, the third functional layer L3 determines whether there are any signs of an abnormality occurring in units of a group of equipment (i.e., production line 90 in this case) that includes multiple first equipment groups 3a, such as robot groups 60, conveyor groups 70, and AGV groups 80. There is no particular limit to the number of first equipment groups 3a that make up the second equipment group 3b.

[0045] Further layers above the functional third layer L3 may be provided, such as a fourth layer, an N-1th layer, and an Nth layer. In the upper layers, the presence or absence of signs of an abnormality is determined for each group of equipment groups grouped in the layer immediately below, and then for each group of those groups, and so on. The unit (grouping unit) for determining the presence or absence of signs of an abnormality in each layer can be determined, for example, as follows: The scale of grouping of equipment 3 increases as the layer goes higher, and the presence or absence of signs of an abnormality is determined by the monitoring device 1 belonging to each layer.

[0046] Functional first layer L1 (bottom layer): each equipment 3 unit such as the first robot 61, the first conveyor 71, the first AGV 81, etc. Functional second layer L2: Units of the first equipment group 3a, such as a robot group 60 including a first robot 61 and a second robot 62, a conveyor group 70 including a first conveyor 71 and a second conveyor 72, and an AGV group 80 including a first AGV 81 and a second AGV. Functional third layer L3: Units of the second equipment group 3b (production line 90) including the robot group 60, the conveyor group 70, and the AGV group 80 Functional fourth layer L4: A group unit of multiple production lines 90 that manufacture the same product Functional 5th layer L5: A factory unit including a group of multiple production lines 90 that manufacture different products. Functional N-1 layer LN-1: A group of factories located in a specific country Functional Nth layer LN (top layer): Unit of factories located worldwide

[0047] In this way, the higher the layer, the more information is aggregated, and the number of monitoring devices 1 included in each layer decreases. The functional Nth layer LN (top layer) contains one monitoring device 1. The grouping units in each layer described above are merely examples, and the grouping units can be set arbitrarily. For example, robots with the same task content may be treated as a group of equipment, and a new layer may be added by defining units such as a first robot group performing a first task A and a second robot group performing a second task B. Hereinafter, monitoring of individual pieces of equipment 3 in the functional first layer L1 may be referred to as individual equipment status monitoring, and monitoring in layers above the functional first layer L1 (i.e., monitoring of equipment groups grouped on a predetermined scale) may be referred to as area status monitoring.

[0048] EBSM uses the concept of "harmony" to determine whether there are signs of an abnormality. EBSM detects signs of an abnormality in individual pieces of equipment 3 or in groups of equipment 3 based on harmony, i.e., whether the equipment 3 maintains a harmonious state. In conventional technology, for example, when measurement results are output from multiple sensors 5 (sensor group 5S) in a single piece of equipment 3, if all of the measurement results are within the normal range, it is determined that there are no signs of an abnormality. However, even if all of the outputs from the sensor group 5S are within the normal range, if the balance of the outputs from each sensor 5 deviates from a predetermined range, this could be a sign of an abnormality. Furthermore, in an equipment group including multiple pieces of equipment 3, if the balance of the status of each piece of equipment 3 deviates from a predetermined range, this could be a sign of an abnormality. This is similar to the process by which an experienced manager, for example, grasps the situation at a site from a macro perspective and intuitively recognizes that "something is different from usual."

[0049] In other words, EBSM monitors the harmony of the equipment group in the same way that a person listens to a symphony played by an orchestra.Then, using the "degree of balance" of the equipment group when the entire group is in harmony as a benchmark, it derives the "degree of imbalance" during the monitoring period.An experienced manager can also be interpreted as sensing this "degree of imbalance" as the "atmosphere of Facility 3 and the equipment group."

[0050] Understanding the individual harmony of equipment 3 (individual equipment status monitoring) is a fundamental requirement in factories and other workplaces. However, on-site, multiple pieces of equipment 3 often operate in coordination rather than independently. In such cases, the balance of each piece of equipment 3 within a group of equipment must be considered. In other words, in addition to individual equipment status monitoring, monitoring the group of equipment as a whole (area status monitoring) is also required. Furthermore, as the number of monitored equipment (number of equipment 3) increases, the size of the data management and application servers (see server 30 in Figure 4) also increases. It is therefore difficult to plan a system construction plan that takes into account scale-up in advance. Area status monitoring using a hierarchical structure is also an effective solution to this problem. In other words, even if a new piece of equipment 3 is introduced into the manufacturing line 90, the new piece of equipment 3 is simply added to the first functional layer L1, and the hierarchical structure remains unchanged. In other words, the configuration shown in Figure 3 allows for easy addition and deletion of equipment 3 or groups of equipment.

[0051] FIG. 4 is a configuration diagram of the EBSM focusing on the first monitoring device 1a included in the first functional layer L1. FIG. 4 shows the configuration of the first monitoring device 1a and the functions of the first monitoring device 1a in the EBSM. As shown in the figure, the EBSM includes the first monitoring device 1a to the Nth monitoring device 1z (monitoring devices 1), equipment 3, sensors 5 that measure status information of the equipment 3, a server 30, and an information terminal 40. There may be multiple sensors 5, in which case each sensor 5 constitutes a sensor group 5S. The first monitoring device 1a, the second monitoring device 1b to the Nth monitoring device 1z, the server 30, and the information terminal 40 transmit and receive information to and from each other via a network 50.

[0052] Here, "status information" refers to measurement data (vector data) measured in time series by the sensors 5, and when a sensor group 5S is made up of multiple sensors 5, it refers to a group of data measured in time series by the sensor group 5S. This data group is obtained by concatenating the vector data. When the number of sensors 5 included in the sensor group 5S is M1 (M1 is a positive integer) and each sensor 5 outputs one piece of measurement data per measurement, M1-dimensional vector data is obtained by one measurement. Note that "one measurement" includes not only measurements taken by each sensor 5 at synchronized timing, but also asynchronous measurements taken by each sensor 5 at different timings. In the latter case, the latest measurement values ​​of each sensor 5 at a specific point in time constitute M1-dimensional vector data.

[0053] Furthermore, even if the sensor group 5S is composed of only one sensor 5, when one sensor 5 outputs M2 (M2 is a positive integer) pieces of measurement data in one measurement (for example, M2 = 6 in a 6-axis inertial sensor), M2-dimensional vector data is obtained in one measurement. Furthermore, when the number of sensors 5 constituting the sensor group 5S is M3 (M3 is a positive integer), M4 sensors among them output one piece of measurement data in one measurement, and M3 - M4 sensors output M2 pieces of measurement data in one measurement, M4 + (M3 - M4) × M2-dimensional vector data is obtained in one measurement. In this way, the state information is composed of multi-dimensional vectors. Of course, there is no particular limit to the number of dimensions of the state information.

[0054] To simplify the following explanation, it is assumed that the sensor group 5S outputs an M-dimensional vector (M is an integer equal to or greater than 3) in one measurement. That is, each element of the status information constitutes an M-dimensional vector. If the number of samples in the sensor group 5S is SN1, the status information is made up of SN1 M-dimensional vectors. Note that the types and numbers of sensors 5 installed in each facility 3 may be different, and the number of dimensions of each element of the status information output by each facility 3 may be different.

[0055] The monitoring device 1 is configured, for example, by a small industrial computer. The first monitoring device 1a includes a first control unit 11a, a storage unit 11b, a first communication unit 11c, and a network communication unit 11d. The first control unit 11a includes a processor such as a CPU (Central Processing Unit) and operates according to a control program stored in a storage unit 11b composed of a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. The storage unit 11b is also used as a work memory. The storage unit 11b is further provided with a non-volatile memory such as an EEPROM, in which predetermined data is stored. The predetermined data includes unique identification information (monitoring device ID) corresponding to each first monitoring device 1a. The first control unit 11a and the other components are connected by a bus 11e or the like, and the first control unit 11a controls the other components via the bus 11e or the like.

[0056] The first communication unit 11c includes a communication module (not shown) that complies with wireless communication standards such as BLE (Bluetooth (registered trademark) Low Energy), UWB (Ultra-Wideband), ZigBee (registered trademark), and LTE-M (Long Term Evolution for Machines). The first communication unit 11c receives status information of the equipment 3 transmitted from the sensor group 5S. The first control unit 11a monitors the status of the equipment 3 based on the status information and determines whether there is a sign of an abnormality occurring in the equipment 3.

[0057] The network communication unit 11d includes a communication module (not shown) that complies with a wireless communication standard such as LTE, 4G, or 5G. The network communication unit 11d transmits status information acquired from the sensor group 5S (or information obtained after processing the status information) to the server 30 via the network 50. Of course, the network communication unit 11d may be configured to comply with the WiFi (Wireless Fidelity) standard, and the first monitoring device 1a may be connected to the network 50 via a wireless router or the like.

[0058] The server 30 is a known computer system and is composed of a server control unit 30a and a server storage unit 30b. The server control unit 30a is composed of a CPU and memory (not shown) and controls the components of the server 30. The server storage unit 30b may include a large-capacity storage configured with ROM, RAM, and a RAID (Redundant Array of Independent Disks) or the like. This large-capacity storage is provided with an equipment monitoring database (equipment monitoring DB). The equipment monitoring DB stores, in association with each other, a unique identifier (equipment ID) corresponding to each of the multiple pieces of equipment 3, a unique identifier (sensor ID) corresponding to the sensor 5 installed in the piece of equipment 3, and a unique identifier (monitoring device ID) corresponding to each monitoring device 1.

[0059] The information terminal 40 is, for example, a mobile information terminal such as a smartphone or tablet terminal, or a PC (Personal Computer), and includes a user interface such as a display unit and an input unit, a calculation unit, a memory, etc. (not shown). The information terminal 40 acquires the monitoring results (including an EBSM state map, an EBSM-VH map, etc., which will be described later) of each facility 3 or a facility group from the monitoring device 1 (or the server 30), and presents them to a manager, etc. Furthermore, when a sign of an abnormality is detected in each facility 3 or a facility group, the monitoring device 1 may notify the information terminal 40 in the form of an email.

[0060] Each sensor 5 constituting the sensor group 5S includes a sensor module 5a and a sensor communication unit 5b. Examples of the sensor module 5a include an acceleration sensor, a geomagnetic sensor, a microphone as a sound detection sensor, a piezoelectric element as a vibration detection unit, a temperature sensor, a humidity sensor, an odor sensor, and an ammeter that measures the drive current of an actuator such as a motor provided in the facility 3. Note that a so-called six-axis inertial sensor including an angular velocity sensor (in this case, six-dimensional state information is output from one sensor 5) may also be used as the acceleration sensor.

[0061] Here, the acceleration sensor measures, for example, the displacement speed and displacement distance (or, if it is a six-axis inertial sensor, the displacement angular velocity and displacement angle) of the operating part of the equipment 3 (including the entire equipment 3 if the equipment 3 is an AGV). The geomagnetic sensor measures, for example, the direction in which the entire equipment 3 is facing. The microphone detects sound information emitted from the equipment 3. The piezoelectric element detects vibrations of the equipment 3. The temperature sensor measures, for example, the temperature of a part attached to the equipment 3. The humidity sensor measures, for example, the amount of steam emitted from the equipment 3 that uses heated steam for processing. The odor sensor measures, for example, the degree of deterioration of lubricating oil or the status of the equipment 3 that uses volatile solvents for processing. The ammeter measures, for example, the load on the drive source of the actuator, motor, etc. used in the equipment 3. The combination of sensors 5 provided in each equipment 3 is arbitrary, and a single equipment 3 may be provided with multiple sensors 5 (for example, acceleration sensors) that measure the same event, or sensors 5 that measure different events (for example, an acceleration sensor and a temperature sensor).

[0062] The sensor 5 is preferably provided with an attraction portion such as a permanent magnet, a suction cup, an adhesive sheet, or an easily dismantled adhesive, or an engagement portion (none of which are shown) that mechanically engages with the equipment 3, and is attachable to or detachable from the equipment 3 after installation. The sensor communication unit 5b preferably includes a communication module (not shown) that complies with wireless communication standards such as BLE, UWB, ZigBee, or LTE-M. The sensor communication unit 5b transmits the measured status information to the first monitoring device 1a (first communication unit 11c) that corresponds to the equipment 3 to which the sensor 5 is attached.

[0063] As described above, in the monitoring device 1 (first monitoring device 1a) of the first embodiment, the sensor 5 is retrofitted to the equipment 3, and the sensor 5 wirelessly transmits status information to the first communication unit 11c of the first monitoring device 1a. This makes it possible to digitize the manufacturing site at low cost.

[0064] The sensor 5 may include an imaging device equipped with an image sensor. In this case, the imaging device may be installed at a position separated from the corresponding equipment 3 and may capture an image of the entire equipment 3 or a predetermined specific part (for example, a joint or a displacing part if the equipment 3 is a robot). The captured image is transmitted to the first monitoring device 1a. The imaging device may be, for example, a visible light camera. Of course, an NIR (Near Infrared) camera, a thermal camera, or a stereo camera may also be used as the imaging device.

[0065] In this way, the sensors 5 included in the sensor group 5S measure at least one of acceleration information, geomagnetic information, vibration information, sound information, temperature information, humidity information, odor information, current information, and image information, making it possible to appropriately select the sensors 5 to be used for measurement and accurately detect the state of the equipment 3.

[0066] As will be described in detail later, the first control unit 11a of the first monitoring device 1a derives reference state information (information obtained by reducing the dimension of the state information based on the results of eigenvalue analysis; details will be described later) representing state information over a predetermined period in the past based on state information measured by the sensor 5 (sensor group 5S), stores the information in a predetermined storage unit (storage unit 11b or server storage unit 30b), and generates current state information (information obtained by reducing the dimension of the state information based on eigenvectors generated when deriving the reference state information; details will be described later) representing state information after the predetermined period in the past (for example, a time point at which the presence or absence of signs of an abnormality is determined; hereinafter, this may be referred to as the "monitoring period"). Then, the first control unit 11a determines the state of the equipment 3 (presence or absence of signs of an abnormality) based on the generated current state information and the stored reference state information. This makes it possible to accurately detect signs of an abnormality in each piece of equipment 3 based on whether the equipment 3 is maintaining a harmonious state.

[0067] FIG. 5 is a configuration diagram of the EBSM focusing on the second monitoring device 1b included in the functional second layer L2. The second monitoring device 1b includes a second control unit 12a, a storage unit 11b, a first communication unit 11c, and a network communication unit 11d. The second control unit 12a includes a processor such as a CPU, similar to the first control unit 11a already described, and operates according to a control program stored in the storage unit 11b, which is composed of ROM, RAM, etc. Note that a description of components common to the first monitoring device 1a described using FIG. 4 will be omitted. The above-mentioned third monitoring device 1c to Nth monitoring device 1z (see FIG. 2) are provided above the second monitoring device 1b. The configurations and functions of the third monitoring device 1c to Nth monitoring device 1z are equivalent to those of the second monitoring device 1b.

[0068] As will be described in detail later, the second control unit 12a of the second monitoring device 1b (second equipment operational state monitoring device) receives the reference state information and current state information output from the first monitoring device 1a (first equipment operational state monitoring device). This enables the second monitoring device 1b belonging to the upper layer to accurately detect signs of abnormality for each grouped piece of equipment 3 (i.e., equipment group; see the first equipment group 3a in FIG. 2) based on the state quantity (newly derived state information) derived in the first functional layer L1.

[0069] More specifically, the second equipment operational state monitoring device (second monitoring device 1b) included in the functional second layer L2 includes a second communication unit (network communication unit 11d) that receives reference state information and current state information output from the first equipment operational state monitoring device (first monitoring device 1a) included in the functional first layer L1, and a second control unit 12a that derives new reference state information representative of the reference state information and stores it in a predetermined storage unit (storage unit 11b or server control unit 30a) and generates new current state information representative of the current state information.The second control unit 12a determines the state of a first equipment group 3a consisting of multiple pieces of equipment 3 based on the generated new current state information and the stored new reference state information.This makes it possible to accurately detect signs of an abnormality in the equipment group based on whether the equipment group including the multiple pieces of equipment 3 (here, the first equipment group 3a) maintains a harmonious state.

[0070] As described above, the configuration of the monitoring device 1 included in the functional third layer L3 to the functional N-th layer LN is equivalent to that of the second monitoring device 1b. Therefore, the above description can be generalized as follows: when N is an integer equal to or greater than 2, the N-th equipment operational state monitoring device (N-th monitoring device 1z) included in the functional N-th layer LN includes a second communication unit (network communication unit 11d) that receives reference state information and current state information output from the N-1st equipment operational state monitoring device (N-1st monitoring device 1y) included in the functional N-1st layer LN-1, which is a layer below the functional N-th layer, and a second control unit 12a that derives new reference state information representative of the reference state information and stores it in a predetermined storage unit (storage unit 11b or server control unit 30a) and generates new current state information representative of the current state information, and the second control unit 12a determines the state of the equipment group to be monitored (the N-2th equipment group, but if N-2=0, each equipment 3) based on the generated new current state information and the stored new reference state information.

[0071] In this configuration, the monitoring device 1 belonging to the upper layer receives the reference state information and current state information output from the monitoring device 1 belonging to the layer immediately below, derives new reference state information representative of the reference state information and stores it in a predetermined storage unit (storage unit 11b or server control unit 30a), generates new current state information representative of the current state information, and determines the state of the equipment group to be monitored based on the generated new current state information and the stored new reference state information. This makes it possible to accurately detect signs of abnormalities on a scale such as a equipment 3, an equipment group including multiple equipment 3, a group including multiple equipment groups, a group including groups of multiple equipment groups, and so on.

[0072] The second monitoring device 1b may further acquire status information from the sensor 5. This also applies to the monitoring device 1 on a higher layer. That is, when N is an integer equal to or greater than 2, the Nth equipment operational status monitoring device (Nth monitoring device 1z) included in the functional Nth layer LN may receive reference status information and current status information output from the N-1th equipment operational status monitoring device () included in the functional N-1th layer LN-1, which is a layer below the functional Nth layer, and may also acquire status information from the sensor 5. This makes it possible to determine the status of an equipment group, for example, by referring to the temperature, humidity, odor, vibration, etc. of the entire environment in which an equipment group made up of multiple pieces of equipment 3 is installed.

[0073] FIG. 6 is a block diagram showing the functions of the monitoring device 1. FIG. 6 shows functional blocks of the monitoring device 1 included in the first functional layer L1 to the third functional layer L3. The first monitoring device 1a included in the first functional layer L1 acquires state information from a group of sensors 5S. The second monitoring device 1b included in the second functional layer L2 acquires reference state information and current state information output from a plurality of first monitoring devices 1a. Furthermore, the third monitoring device 1c included in the third functional layer L3 acquires reference state information and current state information output from a plurality of second monitoring devices 1b.

[0074] In FIG. 6 , the reference state information and current state information are directly transmitted from the lower-level monitoring device 1 to the higher-level monitoring device 1. However, this information may be stored in the server storage unit 30b of the server 30. In this case, the higher-level monitoring device 1 accesses the server 30 to acquire the reference state information and current state information output by the lower-level monitoring device 1. In the first embodiment, the functional blocks of the monitoring devices 1 included in each functional layer are identical. Also, while the first control unit 11a and the second control unit 12a are distinguished in the above-described FIGS. 5 and 6 , they are essentially identical. That is, in the first embodiment, all monitoring devices 1 have the same physical configuration and may share common functional blocks. This facilitates the addition of new equipment 3, the removal of existing equipment 3, or the modification of the layer structure in the EBSM, improving maintainability. Furthermore, the standardization of the monitoring devices 1 reduces the cost of building and maintaining the EBSM.

[0075] 6, the monitoring device 1 includes functional blocks such as a monitoring method identification processing unit 20a, a data collection processing unit 20b, an operation format identification processing unit 20c, a feature quantity generation processing unit 20d, a feature quantity accumulation processing unit 20e, a base information import processing unit 20f, a state quantity calculation processing unit 20g, a state quantity accumulation processing unit 20h, a state identification processing unit 20i, an identification result accumulation processing unit 20j, a harmony determination processing unit 20k, an attachment file creation processing unit 20m, an email transmission processing unit 20n, a state visualization processing unit 20p, and an image file accumulation processing unit 20r. The processing of these functional blocks is executed by hardware such as the first control unit 11a (second control unit 12a), the storage unit 11b, the first communication unit 11c, and the network communication unit 11d described with reference to FIGS.

[0076] In the EBSM, base information is acquired prior to executing a monitoring process (i.e., during a predetermined period in the past). Here, the base information includes the output (status information) of the sensor group 5S installed in the facility 3 and reference status information (primary, secondary, tertiary, etc. reference status information, described later) used by each monitoring device 1 when executing the monitoring process. The process of acquiring the base information based on the output of the sensor group 5S will be described below with reference to FIGS. 3, 4, and 6. [Get base information] <Data collection process> First, the processing of the data collection processing unit 20b will be described. The first monitoring device 1a receives the outputs (status information) of the sensor group 5S provided in the facility 3.

[0077] <Feature generation processing> The processing of the feature generation processing unit 20d will be described below. The first control unit 11a may extract features from the status information (individual M-dimensional vectors) as preprocessing. Specifically, the status information may be normalized, for example, or a deviation value may be derived if the overall variance of the status information is known in advance. In this case, the extracted features are used as the status information for subsequent processing. Of course, the status information may be data obtained by digitizing the output of the sensor 5. The first control unit 11a attaches to the status information an equipment ID corresponding to each facility 3, a sensor ID corresponding to the sensor 5, a monitoring device ID corresponding to each monitoring device 1, and a timestamp when the status information was acquired, and transmits the attached information to the server 30. The server control unit 30a associates the received status information with the monitoring device ID, facility ID, and sensor ID, and stores the associated information in chronological order in the equipment monitoring DB.

[0078] FIG. 7 is an explanatory diagram showing an example of feature extraction. In FIG. 7, five sensors 5 are installed in the facility 3. Measurement results from each sensor 5 are received by the first control unit 11a of the first monitoring device 1a. The first control unit 11a pre-derives the output variation range of each sensor 5, derives the output deviation value of each sensor 5 by referring to the variation range, or ranks the deviation values ​​based on a predetermined threshold. In this example, the deviation values ​​are ranked as high, low, and average, and a predetermined numerical value is assigned to each. When the number of sensors 5 is five as in this example, five-dimensional vector data is obtained as the feature (status information). In the processing of the feature accumulation processing unit 20e, the first control unit 11a stores the feature (status information) in the storage unit 11b or the server storage unit 30b.

[0079] <Operational type identification process> Furthermore, the first control unit 11a may determine the operation mode of each piece of equipment 3 based on the status information. Here, the operation mode refers to, for example, whether the target equipment 3 operates continuously or intermittently, whether the load during operation is constant or fluctuating, whether there is a change in operation due to a change in the materials handled by the equipment 3, and whether there is dependency of the equipment 3 on environmental changes (season, outside temperature, atmospheric pressure, etc.). The first control unit 11a generates operation mode labels corresponding to various operation modes. Then, the first control unit 11a attaches the operation mode labels to the status information and stores them in the memory unit 11b or the server memory unit 30b.

[0080] Through the above process, the status information is stored as base information in the server storage unit 30b. As described above, a timestamp indicating the year, month, day, hour, minute, and second when the status information was acquired, and an operating format label may be added to the status information.

[0081] [Monitoring process in the first monitoring device 1a (monitoring the status of individual devices)] Hereinafter, the monitoring process of the monitoring device 1 (first monitoring device 1a) included in the first functional layer L1 will be described with reference to Figures 3, 4, and 6. Here, individual device status monitoring refers to a mode of monitoring the status of each piece of equipment 3 based on the balance of outputs from multiple sensors 5 provided in each piece of equipment (each piece of equipment 3). <Individual equipment status monitoring - data collection and processing> The monitoring method identification processing unit 20a specifies the data source to be input to the data collection processing unit 20b. For the first monitoring device 1a, the status information output by the sensor 5 is selected as the data source. In the processing of the data collection processing unit 20b, the first monitoring device 1a receives the output (status information) of the sensor group 5S provided in the facility 3. The first control unit 11a stores the received status information in the memory unit 11b. Note that the status information may be transmitted to the server 30 in the same manner as described in "Acquisition of base information" <Data collection processing>.

[0082] <Individual equipment status monitoring - feature generation processing> The feature generation process is the same as the process described in "Acquisition of base information," so a description thereof will be omitted. <Operational type identification process> The operation mode identification process is the same as the process described in “Acquisition of base information.” The first control unit 11a may determine the operation mode of each piece of equipment 3 based on the feature amount generated by the feature amount generation process.

[0083] <Individual equipment status monitoring - base information import processing> In the processing of the base information import processing unit 20f, the first monitoring device 1a accesses the equipment monitoring DB by transmitting its own monitoring device ID, the equipment ID and the sensor ID that it manages to the server 30, and acquires the status information (i.e., the above-mentioned base information) output by each sensor 5 during a predetermined period in the past. At this time, by specifying the above-mentioned timestamp and operation format label, the period during which the base information was accumulated and the operation format for performing the monitoring process are selected.

[0084] <Individual equipment status monitoring - status quantity calculation processing> In the processing of the state quantity calculation processing unit 20g, the first monitoring device 1a performs eigenvalue analysis on the acquired base information to calculate the coordinates of the first eigenvector axis and the second eigenvector axis. The coordinates of the first eigenvector axis and the second eigenvector axis, which are extracted in descending order of eigenvalues ​​by the eigenvalue analysis, reflect the balance of the outputs of the sensors 5. In other words, if the balance (harmony) of the outputs of the sensors 5 provided in the equipment 3 is disrupted, the values ​​of the coordinates change.

[0085] Hereinafter, the base information to be subjected to eigenvalue analysis is assumed to be a plurality of M-dimensional vectors (SN1) as described above. Here, the "predetermined period in the past" may be any period during which all of the equipment 3 constituting the production line 90 is assumed to be operating normally. Specifically, for example, a predetermined period from the time of installation of the production line 90 (see FIG. 1) (such as several months to several years from the initial installation) or a period during which the yield of manufactured products was higher than a predetermined value can be appropriately selected. In individual equipment status monitoring, the "predetermined period in the past" may be any period during which the equipment 3 to be monitored is assumed to be operating normally. As described above, if a timestamp is stored in the equipment management DB along with the base information, the period can be appropriately set by an administrator or the like.

[0086] The first control unit 11a performs eigenvalue analysis on the entire base information and derives the first and second eigenvectors in descending order of eigenvalue. Of course, the third eigenvector or an eigenvector with a smaller eigenvalue may also be derived. Note that when visualization is performed as a two-dimensional map in the <state visualization process> described below, it is sufficient to derive the first and second eigenvectors, and when visualization is performed as a three-dimensional map, it is sufficient to derive up to the third eigenvector. The derived eigenvectors are stored in the storage unit 11b. Of course, the eigenvectors may be transmitted to the server 30 and stored in the server storage unit 30b. Hereinafter, for simplicity of explanation, it is assumed that the first control unit 11a derives the first and second eigenvectors.

[0087] The first control unit 11a individually projects the SN1 pieces of base information onto a two-dimensional space (a Cartesian coordinate system NAx1) defined by the derived first and second eigenvectors, and determines coordinate values ​​in the coordinate system. The state information converted by this process may be referred to as reference state information hereinafter. As described above, the state information is an M-dimensional vector of SN1 pieces, and the reference state information is a two-dimensional vector of SN1 pieces. The reference state information may not be a so-called instantaneous vector, but may be a vector obtained by taking a time average over a predetermined period (a vector representing the overall state over a certain period). In this way, the reference state information is information obtained by reducing the dimensions of the base information (state information) and is information representative of the state information over a predetermined period in the past. In the following description, the reference state information used in the monitoring process for the functional first layer L1 may be referred to as primary reference state information. The order, such as primary, secondary, etc., corresponds to the hierarchy shown in FIG. 3 (primary corresponds to the lowest hierarchy).

[0088] In this way, the first monitoring device 1a included in the first functional layer L1 generates primary reference state information as a state quantity of the facility 3 in which the sensor group 5S is installed, based on the output of the sensor group 5S installed in each facility 3. The primary reference state information corresponds to a state vector of the facility 3 during a predetermined period in the past.

[0089] During the monitoring period, the first control unit 11a acquires status information from the sensor group 5S in the same manner as described in [Acquisition of Base Information] <Data Collection Processing>. The status information may be obtained by extracting features in the same manner as described in [Acquisition of Base Information] <Feature Generation Processing>, and the features may be used as the status information. The status information may be acquired at a timing to determine whether or not there are signs of an abnormality. If the status information is stored in the equipment monitoring DB described above, information corresponding to a specific period or specific operation type may be extracted based on a timestamp or an operation type label. The number of samples (SN2) of status information acquired during the monitoring period is arbitrary, but it is preferable that SN2 be multiple when comparing with the distribution of primary reference status information.

[0090] Specifically, the period for acquiring status information during the monitoring period should be set to, for example, 5 to 20 minutes. This is based on the EBSM technical concept of "grasping the 'atmosphere' of the equipment's operating status over a relatively long period." In other words, the status information used in the monitoring process consists of SN2 M-dimensional vectors.

[0091] The first control unit 11a projects the state information onto an orthogonal coordinate system NAx1 defined by the first and second eigenvectors derived when calculating the reference state information, and determines coordinate values ​​in the coordinate system. The state information converted by this process may hereinafter be referred to as current state information.

[0092] The state information acquired during the monitoring period is an M-dimensional vector of SN2 elements. In other words, the current state information is a two-dimensional vector of SN2 elements. Thus, the current state information is information obtained by reducing the dimension of the state information and is representative of the state information acquired during the monitoring period (monitoring period) after the predetermined period in the past when the reference state information was derived. As described above, the state information may be stored in the server storage unit 30b along with a timestamp and an operation format label. By specifying the timestamp and operation format label, current state information for any period and any operation format can be calculated. Of course, the current state information may be generated for each sampling by the sensor group 5S. Furthermore, the current state information may not be a so-called instantaneous vector, but may be a vector obtained by taking a time average over a predetermined period. In the following description, the current state information used in the monitoring process for the first functional layer L1 may be referred to as primary current state information. Note that the order, such as primary, secondary, etc., corresponds to the hierarchy shown in FIG. 3 (primary corresponds to the lowest hierarchy).

[0093] As described above, in the first embodiment, the first control unit 11a performs eigenvalue analysis on the state information, and derives the reference state information and the current state information by reducing the dimension of the state information using the first eigenvector and the second eigenvector extracted in descending order of eigenvalue. This makes it possible to easily detect signs of abnormality occurring in each piece of equipment 3.

[0094] The first monitoring device 1a included in the lowest layer (functional first layer L1) generates primary current state information as a state quantity of the facility 3 in which the sensor group 5S is installed, based on the output of the sensor group 5S installed in each facility 3. The primary current state information corresponds to a state vector during the monitoring period. The primary reference state information and primary current state information derived by the state quantity calculation processing unit 20g are stored in the memory unit 11b or the server memory unit 30b by the state quantity accumulation processing unit 20h.

[0095] As described above, in the first embodiment, the first equipment operating state monitoring device (first monitoring device 1a) includes a first communication unit 11c that receives state information measured by the sensor 5, and a first control unit 11a that derives reference state information that represents state information during a predetermined period in the past and stores it in a predetermined storage unit 11b or a server storage unit 30b, and generates current state information that represents state information after the predetermined period in the past (monitoring period), and the first control unit 11a determines the state of the equipment 3 based on the generated current state information and the stored reference state information. This makes it possible to accurately detect signs of abnormality in each piece of equipment 3 based on whether the equipment 3 is maintaining a harmonious state.

[0096] The primary reference state information and primary current state information stored in the server storage unit 30b are input to the second control unit 12a of the monitoring device 1 (here, the second monitoring device 1b) in the upper layer (here, the functional second layer L2) and processed by the data collection processing unit 20b. Specifically, the upper layer monitoring device 1 (second monitoring device 1b) accesses the equipment monitoring DB in the server storage unit 30b and acquires the reference state information (primary reference state information) and current state information (primary current state information) derived by the lower layer monitoring device 1 (first monitoring device 1a).

[0097] <Individual equipment status monitoring - status identification processing> The processing of the state identification processing unit 20i will be described below. As described above, the primary reference state information as a vector is (dimensionally reduced) information that represents the state information acquired during the period when the equipment 3 is operating normally. The primary reference state information is a vector corresponding to the harmonious state of the equipment 3, and is derived as many times as the number of sampled state information (SN1). Although the primary reference state information has variations, in the above-mentioned Cartesian coordinate system NAx1, the coordinates of the primary reference state information, i.e., the end point of the vector, are projected into a specific region. This region is called the harmonic region HA.

[0098] On the other hand, the primary current state information as a vector is (dimensionally reduced) information representing the state information of the equipment 3 in a period (period during which monitoring is performed) after the predetermined period in the past when the reference state information was derived. The primary current state information is also derived by the number of sampled state information (SN2). The coordinates of the primary current state information, i.e., the end points of the vectors, are projected onto the above-mentioned Cartesian coordinate system NAx1. That is, in the state identification process, the first control unit 11a projects the primary reference state information and the primary current state information onto the Cartesian coordinate system NAx1. Hereinafter, the projected result may be referred to as the EBSM state map.

[0099] The first control unit 11a compares the distribution of the primary reference state information with the distribution of the primary current state information to determine whether or not there is a sign of an abnormality for each piece of equipment 3 included in the first functional layer L1. The range of the area (harmony area HA) containing the primary reference state information may be adjusted, for example, by expanding or contracting. The distribution of the primary reference state information (i.e., the harmonic area HA) and the distribution of the primary current state information may be separated and analyzed, for example, using a known clustering tool. If, for example, a group of clustered primary current state information exists outside the harmonic area HA, the first control unit 11a determines that the harmony of the manufacturing equipment has deteriorated (it has not yet broken down, but has deviated from a normal state (harmonious state). In other words, it is in a "pre-illness" state, like the human body). The first control unit 11a, as processing by the identification result accumulation processing unit 20j, stores information regarding the presence or absence of a sign of an abnormality in the storage unit 11b or the server storage unit 30b.

[0100] Furthermore, the first control unit 11a may determine whether there is a sign of an abnormality based on whether the end point of each piece of generated primary current state information is included in the harmonic region HA. If the end point of the primary current state information is located outside the harmonic region HA, it may be determined that there is a sign of an abnormality. Of course, the degree of discrepancy (Euclidean distance) of the primary current state information from the center of the harmonic region HA (for example, the average of the primary reference state information) may be used to rank or quantify the harmonicity. The degree of discrepancy or rank may then be used to determine whether the state is a "harmonious state," a "state outside the harmonic region HA," or an "unexperienced state," as shown in FIG. 2.

[0101] <Individual equipment status monitoring - status visualization processing> The processing of the state visualization processing unit 20p will be described below. FIG. 8 is an explanatory diagram of an EBSM state map. The first control unit 11a generates an EBSM state map by plotting the above-mentioned primary reference state information for SN1 and primary current state information for SN2 on two axes based on the first eigenvector and second eigenvector obtained by eigenvalue analysis. The first control unit 11a then transmits the generated EBSM state map to, for example, the information terminal 40. The information terminal 40 displays the EBSM state map on a display unit (not shown). This visualizes the EBSM state map. Then, in the processing of the image file storage processing unit 20r, the first control unit 11a stores the EBSM state map as image information in the storage unit 11b or the server storage unit 30b.

[0102] <Individual equipment status monitoring - Harmony determination processing> The processing of the harmony determination processing unit 20k will be described below. The first control unit 11a derives "dispersion V" and "harmony H" as harmony evaluation indices based on the result of the above-mentioned eigenvalue analysis. Here, the dispersion V is T 2 Based on the statistics, the harmony degree H is derived based on the Q statistic as follows: Note that the standard deviation value in the following explanation refers to the square root of the unbiased variance.

[0103] <Dispersion degree V> First, using [Equation 1], the scaling value x of factor i and analysis lot j is calculated. ij In individual equipment status monitoring, the analysis lot j refers to the output of the sensor group 5S obtained in one sampling, and the factor i refers to the output (status information) of each sensor 5 included in the output of the sensor group 5S. In area status monitoring, which will be described later, the analysis lot j refers to information corresponding to the output of multiple sensor groups 5S obtained in one sampling (i.e., linked reference status information or current status information), and the factor i refers to each individual element included in the linked reference status information or current status information.

number

[0104] Next, using [Equation 2], the state value (r-th eigenvector score value) τ in the r-th eigenvector analysis lot j rj Calculate.

number

[0105] Next, using [Equation 3], T in analysis lot j 2 Calculate statistics.

number

[0106] Next, using [Equation 4], the dispersion V in the analysis lot j is calculated. j Calculate.

number

[0107] where: JPEG0007809306000007.jpg45: T of the rth eigenvector in the base information 2 Statistical mean s r : T of the rth eigenvector in the base information 2 Standard deviation of a statistic

[0108] <Harmony level H> First, using [Equation 5], the restored value from the state value (score) of the analysis lot j in factor i is Calculate JPEG0007809306000008.jpg55.

number

[0109] In addition, the restored value Since JPEG0007809306000010.jpg55 is restored using up to the second eigenvector (composite vector in two-dimensional space), information on the third to mth eigenvectors is missing. Therefore, using [Equation 6], the information loss amount q of the two-dimensional space composite vector from the first to second eigenvectors in the analysis lot j including factor i is calculated as ij Calculate.

number

[0110] Here, if the sum of the amount of missing information of all factors is taken as the Q statistic, the Q statistic for the analysis lot j is expressed by [Equation 7].

number

[0111] Based on [Equation 7], the harmony degree H for the analysis lot j is calculated using the following [Equation 8].

number

[0112] FIG. 9 is an explanatory diagram of the EBSM-VH map. FIG. 9 is a graph in which the harmony H and the dispersion V are calculated for the reference state information of SN1 items and the current state information of SN2 items described above, and the dispersion V is plotted on the horizontal axis and the harmony H on the vertical axis. Here, the harmony region HA may be defined as a range in which the harmony H≦SH and the dispersion V≦SH, for example, with a threshold SH of 3. That is, if the harmony H and the dispersion V are within the threshold range, it is determined that the harmonious state of the equipment 3 is maintained. If at least one of the values ​​of the harmony H or the dispersion V falls outside the harmony region HA, it may be determined that the harmonious state of the equipment 3 is not being maintained, i.e., there is a sign of an abnormality occurring in the equipment 3. Of course, the value of the threshold SH is arbitrary. Furthermore, based on the distribution of the reference state information and the current state information in the EBSM-VH map, the "harmonious state," "state outside the harmony region HA," and "unexperienced state" shown in FIG. 2 may be determined.

[0113] In this way, the first control unit 11a calculates T 2The degree of dispersion V is calculated based on the statistics, the degree of harmony H is calculated based on the Q statistics, and whether the state of the equipment 3 is maintained in harmony is determined based on the degree of dispersion V and the degree of harmony H. This makes it possible to easily determine the harmony of the equipment 3.

[0114] The processing of the attachment file creation processing unit 20m and the email sending processing unit 20n will be described below. The first control unit 11a generates text information or image information based on the results of the harmony determination processing described above, and creates a file in a predetermined format. Then, the email with the file attached is sent to, for example, an information terminal 40 assigned a predetermined email address.

[0115] In this way, individual equipment status monitoring is performed for each piece of equipment 3. In individual equipment status monitoring, monitoring processing may be performed using newly obtained status information for a specific period or specific operation type based on the timestamp or operation type label described above. In other words, in individual equipment status monitoring, if monitoring the status of each piece of equipment 3 is considered macroscopic status monitoring, microscopic status monitoring can be performed for the purpose of evaluating harmony focused on any period or operating conditions, thereby realizing two-stage parallel monitoring of balanced status (hereinafter sometimes referred to as "two-stage parallel status monitoring").

[0116] [Monitoring process in upper layer monitoring device 1 (area status monitoring)] Hereinafter, the monitoring process (area state monitoring) of the monitoring device 1 included in the upper layers (functional second layer L2 to functional Nth layer LN) of the functional first layer L1 will be explained using Figures 3, 5, and 6. Here, area state monitoring refers to collecting state quantities (reference state information and current state information) derived by multiple monitoring devices 1 belonging to the lower layers, calculating new state quantities based on the collected state quantities, and monitoring the balance state between equipment groups made up of multiple pieces of equipment 3. Furthermore, area state monitoring includes monitoring on a scale such as a group of equipment groups, or even a group of groups of equipment groups, etc.

[0117] For simplicity of explanation, the following description will be mainly focused on the functions and operations of the second monitoring device 1b included in the functional second layer L2. However, the operations and functions of the monitoring devices 1 (the third monitoring device 1c to the Nth monitoring device 1z) in layers above the functional second layer L2 are equivalent to those of the second monitoring device 1b. This enables multi-scale status monitoring by the above-mentioned individual device status monitoring and area status monitoring in the first embodiment.

[0118] <Area status monitoring - data collection and processing> The second control unit 12a of the second monitoring device 1b specifies the type of data to be input by the data collection processing unit 20b as processing by the monitoring method identification processing unit 20a. For the second monitoring device 1b included in the functional second layer L2, the reference state information (primary reference state information) and current state information (primary current state information) output by the first monitoring device 1a belonging to the lower layer are selected as the type of data to be input. As described above, in the first monitoring device 1a, the primary reference state information and primary current state information derived by processing by the state quantity calculation processing unit 20g are stored in the server storage unit 30b, and the second monitoring device 1b in the upper layer accesses the equipment monitoring DB in the server storage unit 30b to acquire the reference state information (primary reference state information) and current state information (primary current state information) derived by the multiple first monitoring devices 1a belonging to the layer immediately below.

[0119] <Area status monitoring - feature generation processing> As described above, the primary reference state information output by each first monitoring device 1a is a two-dimensional vector. The second monitoring device 1b concatenates the primary reference state information output by multiple first monitoring devices 1a, and generates 2×P1-dimensional state information if, for example, the number of first monitoring devices 1a = P1. Here, it is preferable that the number of samples of primary reference state information input to the second monitoring device 1b be the same (SN3) for each first monitoring device 1a.

[0120] <Area status monitoring - state quantity calculation processing> The second control unit 12a performs eigenvalue analysis on the SN3 pieces of 2×P1-dimensional state information. Through the eigenvalue analysis, the SN3 pieces of 2×P1-dimensional state information are reduced in dimension to SN3 two-dimensional vectors. The coordinates of the first eigenvector axis and the second eigenvector axis, which are extracted in descending order of eigenvalues ​​extracted by the eigenvalue analysis, reflect the balance between the primary reference state information derived for each piece of equipment 3. In other words, if the balance (harmony) of the outputs of the multiple first monitoring devices 1a is disrupted, the value of each coordinate changes.

[0121] Then, the second control unit 12a of the second monitoring device 1b projects the primary reference state information onto a two-dimensional space (an orthogonal coordinate system NAx2) defined by the derived first and second eigenvectors, and determines coordinate values ​​in the coordinate system. Through this process, the second monitoring device 1b derives two-dimensional secondary reference state information. The secondary reference state information is stored in the storage unit 11b or the server storage unit 30b. The secondary reference state information is then input to a higher-level monitoring device 1, and the monitoring device 1 derives tertiary or higher-level reference state information.

[0122] Furthermore, the primary current state information output by each first monitoring device 1a is a two-dimensional vector of SN2 pieces as described above. Similar to the processing for the reference state information, the second monitoring device 1b concatenates the primary current state information output by the multiple first monitoring devices 1a and projects the concatenated primary current state information onto the orthogonal coordinate system NAx2 to derive secondary current state information. The secondary current state information is stored in the storage unit 11b or the server storage unit 30b.

[0123] As shown in FIG. 5, the output of the sensor 5 may be input to the second monitoring device 1b as status information. However, it is preferable that the sensor 5 here does not measure the status of individual pieces of equipment 3, but rather measures the status of at least a group of pieces of equipment collectively. In this case, the output of the sensor 5 is processed in the same manner as described above in the section <Individual Equipment Status Monitoring - Data Collection Processing> and further linked to 2×P1-dimensional status information. That is, if the output of the sensor 5 is P2-dimensional, the linked status information will be 2×P1+P2-dimensional, and the second control unit 12a performs eigenvalue analysis on the SN3 pieces of status information in the 2×P1+P2 dimension. Then, taking the output of the sensor 5 into account, secondary reference status information and secondary current status information are derived as SN3 two-dimensional vectors.

[0124] The status information acquired by the sensor 5 may be input not only to the second monitoring device 1b included in the functional second layer L2, but also to a monitoring device 1 belonging to a layer above the functional first layer L1. In this way, in the first embodiment, the monitoring device 1 included in a layer above the functional first layer L1 may further receive status information measured by the sensor 5. This makes it possible to perform monitoring taking into account information about the environment in which the equipment group including the plurality of pieces of equipment 3 is located. Note that "information about the environment in which the equipment group including the plurality of pieces of equipment 3 is located" may include, for example, temperature, humidity, and odor.

[0125] <Area status monitoring - status identification processing, harmony determination processing, status visualization processing, etc.> Here, the secondary reference state information is (dimensionally reduced) information representing state information acquired during a period when an equipment group (the first equipment group 3a in FIG. 3 ) including multiple pieces of equipment 3 is operating normally. The secondary current state information is (dimensionally reduced) information representing state information of the equipment group (the first equipment group 3a) during a period (a period during which monitoring is performed) after a predetermined period in the past when the secondary reference state information was derived. The harmonic area HA in the area state monitoring is defined by projecting the secondary reference state information onto the Cartesian coordinate system NAx2 derived by eigenvalue analysis. Then, based on whether the position of the secondary current state information projected onto the Cartesian coordinate system NAx2 is within the harmonic area HA, the presence or absence of a sign of an abnormality occurring in the first equipment group 3a, i.e., the harmonicity, is determined. Of course, as described in <Individual Equipment State Monitoring—State Identification Processing>, the harmonicity may also be determined based on the distribution of the clusters constituted by the secondary current state information and the harmonic area HA.

[0126] Furthermore, the second control unit 12a projects the secondary reference state information and the secondary current state information onto the Cartesian coordinate system NAx2 and outputs the result as an EBSM state map. In the area state monitoring, the EBSM state map is also transmitted to, for example, the information terminal 40 and displayed. Furthermore, the harmony H and dispersion V described above may be calculated to determine whether the harmonized state of the equipment group is maintained. The relationship between the harmony H and dispersion V is visualized as an EBSM-VH map. In the area state monitoring, if monitoring the state of the equipment group is considered macroscopic state monitoring, microscopic state monitoring can be performed to evaluate harmony focusing on any period or operating conditions. This enables two-stage parallel state monitoring, in which two-stage balance states are monitored in parallel.

[0127] The secondary reference state information is input to a higher-level monitoring device 1, which derives tertiary or higher reference state information. The secondary current state information is also input to a higher-level monitoring device 1, which derives tertiary or higher current state information. The higher-level monitoring device 1 performs the above-mentioned <area state monitoring-feature quantity generation process> and <area state monitoring-state quantity calculation process> based on the state information input from the lower-level monitoring device 1, and sequentially generates tertiary reference state information to Nth-order reference state information and tertiary current state information to Nth-order current state information.

[0128] Thus, the equipment operational status monitoring system Sy1 (EBSM) of the first embodiment comprises a functional first layer L1 including multiple first equipment operational status monitoring devices (first monitoring devices 1a) that monitor the status of equipment 3 based on the output of at least one sensor 5 installed in each equipment 3 and measuring the status information of the equipment 3, and a functional second layer L2 including at least one second equipment operational status monitoring device (second monitoring device 1b) installed as a layer above the functional first layer L1 and that monitors the status of a first equipment group 3a, which is a grouping of multiple equipment 3, based on the output of the multiple first equipment operational status monitoring devices (first monitoring devices 1a) included in the functional first layer L1. This enables the second equipment operational status monitoring device to accurately detect signs of an abnormality in the equipment group (first equipment group 3a) including multiple equipment 3, based on whether the equipment group (first equipment group 3a) is maintaining a harmonious state.

[0129] To generalize this, the equipment operational status monitoring system Sy1 (EBSM) of the first embodiment, where N is an integer equal to or greater than 2, is provided as an upper layer of the functional N-1th layer LN-1, and includes at least one Nth equipment operational status monitoring device (monitoring device 1) that monitors the status of an equipment group (an N-1th equipment group formed by grouping multiple N-2th equipment groups; however, when N-2=0, the N-2th equipment group means equipment 3) as a monitoring target based on outputs of multiple N-1th equipment operational status monitoring devices (N-1th monitoring device 1y) included in the functional N-1th layer LN-1. This makes it possible to accurately detect signs of abnormalities on a scale such as equipment 3, an equipment group (first equipment group 3a) including multiple equipment 3, a group including multiple equipment groups, a group including groups of multiple equipment groups, and so on.

[0130] The output of the first equipment operational state monitoring device (first monitoring device 1a) to the output of the (N-1)th equipment operational state monitoring device (one, i.e., reference state information and current state information per sampling) have a common dimension (here, two dimensions). This makes it possible to standardize the configuration of the monitoring device 1 across all layers, thereby reducing system costs.

[0131] The first equipment operational state monitoring device (first monitoring device 1a) performs eigenvalue analysis on the input state information (state information output by sensor 5 or state information generated by the feature generation process) and outputs new state information (reference state information, current state information) by reducing the dimension of the state information based on the first and second eigenvectors extracted by the eigenvalue analysis. Each of the second equipment operational state monitoring device (second monitoring device 1b) through the N-1st equipment operational state monitoring device (N-1st monitoring device 1y) performs eigenvalue analysis on the state information (reference state information) output by the operational state monitoring device (monitoring device 1) included in the layer immediately below and reduces the dimension of the state information based on the first and second eigenvectors extracted by the eigenvalue analysis, thereby outputting new state information (new reference state information, new current state information). This new state information is input to the monitoring device 1 in the layer immediately above. This enables multi-scale state monitoring through individual equipment status monitoring and area status monitoring.

[0132] In addition, when N is an integer of 2 or more, the Nth equipment operation status monitoring device (Nth monitoring device 1z) calculates T 2 The dispersion V may be calculated based on the statistics, the harmony H may be calculated based on the Q statistics, and whether the harmony of the equipment group is maintained may be determined as the state of the equipment group based on the dispersion V and the harmony H. This makes it possible to easily determine the harmony of the equipment group.

[0133] Fig. 10 is an explanatory diagram illustrating an example of individual device status monitoring for each robot and area status monitoring for the robot group 60. Fig. 10 shows the process of individual device status monitoring executed for the first robot 61, second robot 62, third robot 63, and fourth robot 64 that perform work on the production line 90 (see Fig. 1), and area status monitoring executed for the robot group 60 made up of these multiple robots. Hereinafter, the explanation will continue using Fig. 10 in conjunction with Figs. 4 and 5.

[0134] <Individual equipment status monitoring for each robot> Here, the sensor group 5S (see Figure 4) mounted on each robot outputs nine-dimensional state information per sampling (one lot). The number of samplings when creating the base information is SN4. As explained in "Acquisition of Base Information" and "Monitoring Process in the First Monitoring Device 1a (Individual Device Status Monitoring)," the first control unit 11a of the first monitoring device 1a corresponding to each robot performs eigenvalue analysis on the state information acquired over a predetermined past period, i.e., the base information (9 dimensions × SN4 lots), to generate two-dimensional primary reference state information (SN4). Furthermore, the state information acquired over the monitoring period (9 dimensions per lot because the output of the same sensor group 5S is used) is projected onto coordinate axes defined by the eigenvectors obtained by eigenvalue analysis to generate primary current state information (SN4 two-dimensional vectors). Based on the primary reference state information and the primary current state information, the presence or absence of signs of an abnormality is determined for each robot. Furthermore, an EBSM state map and an EBSM-VH map corresponding to each robot are generated and visualized.

[0135] <Area status monitoring for 60 robots> The two-dimensional primary reference state information and primary current state information generated by each robot are input to the upper-level monitoring device 1 (here, the second monitoring device 1b (see Figure 5)). The second monitoring device 1b then concatenates the two-dimensional primary reference state information (4 SN units each) generated by each robot (four robots in this case) to generate SN4 units of 2 x 4 = 8-dimensional "state information of the robot group 60." The second monitoring device 1b then performs eigenvalue analysis on the state information of the robot group 60 (8 dimensions x 4 SN units) to derive secondary reference state information (2 dimensions x 4 SN units).

[0136] Furthermore, the second monitoring device 1b concatenates the primary current state information output by the four first monitoring devices 1a. Then, a weighting operation is performed on the concatenated eight-dimensional primary current state information using the first eigenvector and second eigenvector derived by eigenvalue analysis to derive secondary current state information (two-dimensional). Based on the secondary reference state information and the secondary current state information, it is determined whether there is a sign of an abnormality occurring in the robot group 60. Furthermore, an EBSM state map and an EBSM-VH map are generated and visualized for the robot group 60.

[0137] Fig. 11 is an explanatory diagram illustrating an example of individual device status monitoring for each conveyor and area status monitoring for conveyor group 70. Fig. 11 shows the process of individual device status monitoring performed for first conveyor 71, second conveyor 72, and third conveyor 73 installed on production line 90 (see Fig. 1), and area status monitoring performed for conveyor group 70 consisting of these multiple conveyors. Hereinafter, explanation will continue using Fig. 11 in conjunction with Figs. 4 and 5.

[0138] <Individual equipment status monitoring for each conveyor> Here, the sensor group 5S mounted on the first conveyor 71 outputs 12-dimensional status information, the sensor group 5S mounted on the second conveyor 72 outputs 6-dimensional status information, and the sensor group 5S mounted on the third conveyor 73 outputs 12-dimensional status information. Similar to the individual device status monitoring for each robot described above, the first control unit 11a determines whether there are signs of an abnormality occurring for each conveyor. Furthermore, an EBSM status map and an EBSM-VH map corresponding to each conveyor are generated and visualized.

[0139] <Area status monitoring for conveyor group 70> The two-dimensional primary reference state information and primary current state information generated by each conveyor are input into the upper-layer monitoring device 1 (here, the second monitoring device 1b (Fig. 5)). Then, the second monitoring device 1b concatenates the two-dimensional primary reference state information generated by each conveyor (here, three conveyors) to generate "state information of the conveyor group 70" in 2×3 = 6 dimensions. Similar to the processing of <area state monitoring for the robot group 60> described above, the second monitoring device 1b performs eigenvalue analysis on the state information of the conveyor group 70, that is, the base information, to derive secondary reference state information (two-dimensional). Then, the primary current state information acquired during the monitoring period is projected onto the coordinate axes defined by the eigenvectors obtained by eigenvalue analysis to derive secondary current state information. Based on the secondary reference state information and the secondary current state information, it is determined whether there is a sign of an abnormality occurring in the conveyor group 70. Furthermore, for the conveyor group 70, an EBSM state map and an EBSM-VH map are generated and visualized.

[0140] Fig. 12 is an explanatory diagram for explaining an example of individual device state monitoring for each AGV and area state monitoring for the AGV group 80. In Fig. 12, it shows the process of individual device state monitoring executed for the first AGV 81 and the second AGV 82 that perform operations on the manufacturing line 90 (see Fig. 1), and area state monitoring executed for the AGV group 80 composed of these multiple AGVs. Hereinafter, the explanation will continue by using Figs. 4 and 5 in combination with Fig. 12.

[0141] <Individual device state monitoring for each conveyor> Here, the sensor group 5S mounted on the first AGV 81 outputs state information in 21 dimensions, and the sensor group 5S mounted on the second AGV 82 outputs state information in 15 dimensions. Similar to <individual device state monitoring for each robot> and <individual device state monitoring for each conveyor> described above, the first control unit 11a determines whether there is a sign of an abnormality occurring for each AGV. Furthermore, an EBSM state map and an EBSM-VH map corresponding to each AGV are generated and visualized.

[0142] <Area state monitoring for the AGV group 80> The two-dimensional primary reference state information and primary current state information generated by each AGV are input to the upper-level monitoring device 1 (here, the second monitoring device 1b (Figure 5)). The second monitoring device 1b then concatenates the two-dimensional primary reference state information generated by each AGV (here, two AGVs) to generate 2 x 2 = 4-dimensional "state information of the AGV group 80." Similar to the above-mentioned process of <area state monitoring for the robot group 60>, the second monitoring device 1b performs eigenvalue analysis on the state information of the AGV group 80, i.e., base information, to derive secondary reference state information (two-dimensional). The primary current state information acquired during the monitoring period is then projected onto coordinate axes defined by the eigenvectors obtained by the eigenvalue analysis to derive secondary current state information. Based on the secondary reference state information and the secondary current state information, the presence or absence of signs of an abnormality in the AGV group 80 is determined. Furthermore, an EBSM state map and an EBSM-VH map are generated and visualized for the AGV group 80.

[0143] <Area status monitoring for production line 90> FIG. 13 is an explanatory diagram illustrating an example of area state monitoring in a production line 90. A second monitoring device 1b (e.g., second monitoring device 1b (1b_1) shown in FIG. 3) monitoring the group of robots 60, a second monitoring device 1b (e.g., second monitoring device 1b (1b_2) shown in FIG. 3) monitoring the group of conveyors 70, and a second monitoring device 1b (not shown in FIG. 3) monitoring the group of AGVs 80 each output the above-mentioned secondary reference state information and secondary current state information. The secondary reference state information and secondary current state information are then input to a monitoring device 1 in the immediately higher layer (here, a third monitoring device 1c included in the functional third layer L3).

[0144] Here, the secondary reference state information output from the second monitoring device 1b, which monitors the robot group 60, the conveyor group 70, and the AGV group 80, is two-dimensional as described above. The third monitoring device 1c performs eigenvalue analysis on a six-dimensional vector (SN4 in number) concatenating these secondary reference state information to derive third-order reference state information (SN4 two-dimensional vectors). Then, the second current state information (six dimensions) acquired during the monitoring period is projected onto a coordinate system defined by the eigenvectors obtained by the eigenvalue analysis to derive third-order current state information (two-dimensional). Based on the third reference state information and the third current state information, the presence or absence of signs of an abnormality is determined for the production line 90, which includes the robot group 60, the conveyor group 70, and the AGV group 80. Furthermore, an EBSM state map and an EBSM-VH map are generated and visualized for the production line 90.

[0145] The tertiary reference state information and the tertiary current state information are input to a higher-level monitoring device 1, which derives quaternary, quintic, etc. reference state information and current state information, and determines whether there are any signs of an abnormality occurring on a scale such as a factory including multiple production lines 90, or a group of such factories, etc.

[0146] In the first embodiment, when creating base information, the first monitoring device 1a belonging to the first functional layer L1 acquires status information from the sensor 5, and derives primary reference status information. Then, by deriving the primary reference status information, the second monitoring device 1b belonging to the second functional layer L2 derives secondary reference status information. This chain of processes for generating reference status information is repeated up to the third functional layer L3 to the Nth functional layer LN. Furthermore, in the monitoring process stage, the first monitoring device 1a acquires status information from the sensor 5, and derives primary current status information. Then, by deriving the primary current status information, the second monitoring device 1b derives secondary current status information. This chain of processes for generating current status information is also repeated up to the third functional layer L3 to the Nth functional layer LN. The equipment operational status monitoring system Sy1 is constructed by this chain of information processing.

[0147] As described above, the status information acquired from the sensor 5 can be stored in the server storage unit 30b. Therefore, in the above-described <Individual Device Status Monitoring—Status Quantity Calculation Process> and <Region Status Monitoring—Status Quantity Calculation Process>, each monitoring device 1 can acquire base information collected over a predetermined period in the past from the server 30, perform eigenvalue analysis offline, and generate reference status information to be referenced during the monitoring period. Alternatively, the server control unit 30a can perform eigenvalue analysis based on the stored base information to derive reference status information to be used by each monitoring device 1 and transmit this information to each monitoring device 1. In this case, each monitoring device 1 stores the reference status information in the storage unit 11b. During the monitoring period, each time status information is output online from the sensor 5, current status information is transmitted from the lower layer to the upper layer, and all monitoring devices 1 calculate new current status information in real time and perform monitoring processing. In this way, in EBSM, each monitoring device 1 can function as an edge device.

[0148] In this way, the equipment operation status monitoring system Sy1 of the first embodiment may include at least one sensor 5 provided in each piece of equipment 3 to measure status information of the equipment 3; a monitoring unit (server 30 or information terminal 40) that receives the status information over a predetermined period in the past and derives primary reference status information representative of the status information and secondary reference status information representative of the primary reference status information; a first equipment operation status monitoring device (first monitoring device 1a) that acquires the primary reference status information from the monitoring unit, derives primary current status information representative of the status information over a monitoring period after the predetermined period in the past, and determines the status of the equipment 3 based on the primary reference status information and the primary current status information; and a second equipment operation status monitoring device (second monitoring device 1b) that acquires secondary reference status information from the monitoring unit, derives secondary current status information representative of the primary current status information output by the first equipment operation status monitoring device over the monitoring period, and determines the status of an equipment group (first equipment group 3a) formed by grouping multiple pieces of equipment 3 based on the secondary reference status information and the secondary current status information. This allows each monitoring device 1 to function as an edge device and to perform monitoring processing on the facility 3 or a group of facilities in real time.

[0149] The first equipment operational state monitoring device (first monitoring device 1a) calculates T 2 The degree of dispersion V may be calculated based on the statistics, the degree of harmony H may be calculated based on the Q statistics, and whether the state of the equipment 3 is maintained in harmony may be determined based on the degree of dispersion V and the degree of harmony H. This makes it possible to easily determine the harmony of the equipment 3.

[0150] Furthermore, the second equipment operation state monitoring device (second monitoring device 1b) calculates T 2 The dispersion V may be calculated based on the statistics, the harmony H may be calculated based on the Q statistics, and whether the harmony of the equipment group is maintained may be determined as the state of the equipment group based on the dispersion V and the harmony H. This makes it possible to easily determine the harmony of the equipment group.

[0151] Furthermore, the equipment operating state monitoring method of the first embodiment may measure state information of each equipment 3, receive the state information over a predetermined period in the past, derive primary reference state information representative of the state information and secondary reference state information representative of the primary reference state information, acquire the primary reference state information, derive primary current state information representative of the state information over a monitoring period after the predetermined period in the past, determine the state of the equipment 3 based on the primary reference state information and the primary current state information, acquire the secondary reference state information, derive secondary current state information representative of the primary current state information over the monitoring period, and determine the state of a group of equipment formed by grouping multiple equipment 3 based on the secondary reference state information and the secondary current state information. This makes it possible to perform monitoring processing on the equipment 3 or the equipment group in real time.

[0152] On the other hand, the hierarchical structure of the monitoring device 1 shown in Fig. 3 may be virtual or conceptual. That is, if the tree structure of the monitoring device 1 shown in Fig. 3 is defined in the server 30 shown in Fig. 4 and Fig. 5, the monitoring devices 1 in higher hierarchical levels, except for the monitoring device 1 in the functional first layer L1 that acquires the output of the sensor 5, may be virtual. In this case, the functions and processes of the monitoring device 1 in higher hierarchical levels than the functional first layer L1 may be consolidated in, for example, the server 30 or information terminal 40 as a monitoring unit.

[0153] As described above, the equipment operational state monitoring system Sy1 of the first embodiment includes at least one sensor 5 provided in each equipment 3 to measure state information of the equipment 3; a first equipment operational state monitoring device (first monitoring device 1a) that derives primary reference state information representative of the state information during a predetermined period in the past, derives primary current state information representative of the state information during a monitoring period after the predetermined period in the past, and determines the state of each equipment 3 based on the primary reference state information and the primary current state information; and a monitoring unit that derives secondary reference state information representative of the primary reference state information for an equipment group formed by grouping multiple equipment 3 based on the primary reference state information output by the multiple first equipment operational state monitoring devices (first monitoring devices 1a), derives secondary current state information representative of the primary current state information for the equipment group based on the primary current state information output by the multiple first equipment operational state monitoring devices (first monitoring devices 1a), and determines the state of the equipment group based on the secondary reference state information and the secondary current state information. Here, the functions of the monitoring unit may be performed by either the server 30 or the information terminal 40, or may be shared between the server 30 and the information terminal 40 as appropriate. This makes it possible to accurately detect signs of abnormalities occurring in each piece of equipment 3 based on whether the equipment 3 is maintained in a harmonious state, and further to accurately detect signs of abnormalities occurring in each equipment group based on whether an equipment group including multiple pieces of equipment 3 is in harmony.

[0154] Furthermore, the equipment operation state monitoring method of the first embodiment may measure state information of each piece of equipment 3, derive primary reference state information representing the state information during a predetermined period in the past based on the state information, derive primary current state information representing the state information during a monitoring period after the predetermined period in the past based on the state information, determine the state of each piece of equipment 3 based on the primary reference state information and the primary current state information, derive secondary reference state information representing the primary reference state information for an equipment group formed by grouping multiple pieces of equipment 3 based on the multiple pieces of primary reference state information, derive secondary current state information representing the primary current state information for the equipment group based on the multiple pieces of primary current state information, and determine the state of the equipment group based on the secondary reference state information and the secondary current state information. This makes it possible to accurately detect signs of abnormality in each piece of equipment 3 based on whether the equipment 3 is maintained in a harmonious state, and further to accurately detect signs of abnormality in each equipment group based on whether the equipment group including multiple pieces of equipment 3 is in harmony.

[0155] (Second embodiment) 14 is an explanatory diagram showing an example of harmony determination according to a second embodiment of the present invention. In the second embodiment, the presence or absence of signs of an abnormality is detected for a facility 3 or a facility group including multiple facilities 3, based on status information acquired over a predetermined period of time. In the monitoring process in the second embodiment, it is not necessary to distinguish between reference status information representing status information over a predetermined period of time in the past and current status information representing status information after the predetermined period of time in the past.

[0156] As described above, the status information is stored in the server memory unit 30b, and each monitoring device 1 included in each functional layer (functional first layer L1 to functional Nth layer LN) can extract status information acquired during a specific period, for example, using a timestamp, or new status information derived based on the status information (i.e., a two-dimensional vector obtained by projecting onto a coordinate system defined by an eigenvector derived by eigenvalue analysis).

[0157] Here, the equipment 3 or equipment group to be monitored may take multiple states during a manufacturing process or the like, and may transition between states as the process progresses. In the equipment operation EBSM state base map shown in Figure 14, the process involves seven transitions, from the first state ST1, which is the start process state, to the end process state ST8. State transitions can be represented by mapping two-dimensional vectors calculated using eigenvectors. It should be noted that known clustering tools can be used to spatially separate the states of the equipment 3 or equipment group.

[0158] The monitoring device 1 of the second embodiment counts the number of transitions from each extracted state to other states between the start process state and the end process state. If the equipment 3 or equipment group is in a harmonious state, the number of transitions is seven, as described above. However, as the harmony is lost, the number of state transitions increases. Specifically, there should be only one transition from the first state ST1 to the second state ST2, but as the harmony is lost, multiple transitions between the first state ST1 and the second state ST2 may occur. That is, the transitions are as follows: first state ST1 → second state ST2 → first state ST1 → second state ST2.

[0159] In the second embodiment, where X1 is the number of times a state transition has occurred and X2 is the number of samples used to obtain state information, the transition rate RX is defined as X1 / X2, and if the value of RX exceeds a predetermined threshold, it is determined that there is a sign of an abnormality. As in the first embodiment, in the second embodiment, base information may be acquired in advance, and transitions between states may be detected based on reference state information obtained from the base information and current state information obtained during the monitoring period. In this case, harmony may be determined taking into account, for example, the time a specific state is maintained.

[0160] While specific embodiments of the equipment operational status monitoring device and equipment operational status monitoring system Sy1 according to the present invention have been described above, these are merely examples, and the present invention is not limited to these embodiments. For example, the above-described series of processes may be executed by hardware or software. Furthermore, as long as the equipment operational status monitoring system Sy1 (EBSM) is provided with the functionality to execute the above-described series of processes as a whole, there are no limitations on the type of functional blocks used to realize these functions.

[0161] Furthermore, the locations where the functional blocks in the monitoring device 1 are installed are not particularly limited. For example, the equipment 3 and part of the first monitoring device 1a shown in FIG. 3 may be installed overseas, while the other equipment 3 and monitoring device 1 are installed domestically. The location where the server 30 shown in FIGS. 4 and 5 is installed is also not limited, and may be either domestic or overseas. Regarding the functional blocks, for example, the harmony determination process may be performed in Japan, and the determination results may be reported overseas. Furthermore, the processing of each functional block may be performed not only by a dedicated monitoring device 1, but also by a user terminal (information terminal 40), the server 30, etc. Furthermore, one functional block may be configured as a single piece of hardware, a single piece of software, or a combination of these. [Industrial Applicability]

[0162] The equipment operation status monitoring device (monitoring device 1), equipment operation status monitoring system Sy1, and equipment operation status monitoring method of the present invention can accurately detect signs of abnormality in individual equipment 3 based on whether the equipment 3 is maintained in a harmonious state, and can also accurately detect signs of abnormality in an equipment group including multiple pieces of equipment 3 based on whether the equipment group is maintained in a harmonious state.Therefore, they can be widely used to monitor the status of a production line 90 including multiple pieces of manufacturing equipment and equipment 3, a chemical plant including multiple pieces of equipment 3, a factory including multiple production lines 90 and chemical plants, etc., a factory group including these multiple factories, and a group of these multiple factory groups. [Explanation of symbols]

[0163] 1. Monitoring device (equipment operation status monitoring device) 1a 1st monitoring device 1b Second monitoring device 1c Third monitoring device 3 Equipment 3a 1st equipment group 3b 2nd equipment group 5 sensors 5S sensor group 11a First control section 12a Second control section 30 servers 40 Information terminal 50 Network 60 Robots 70 Conveyor Group 80 AGVs 90 production lines L1 Functional layer 1 L2 Functional second layer L3 Functional third layer Sy1 Equipment operation status monitoring system

Claims

1. At least one sensor provided in each piece of equipment to measure status information of the equipment; a functional first layer including a plurality of first equipment operation state monitoring devices that determine the state of the equipment based on the output of the sensor; provided as an upper layer of the functional first layer, a functional second layer including at least one second equipment operational status monitoring device that determines the status of a first equipment group formed by grouping a plurality of the equipment based on outputs of the plurality of first equipment operational status monitoring devices included in the functional first layer; A facility operation status monitoring system comprising:

2. When N is an integer of 2 or more, Provided as an upper layer of the functional N-1 layer, 2. The equipment operation status monitoring system according to claim 1, further comprising a functional Nth layer including at least one Nth equipment operation status monitoring device that determines the status of a group of equipment to be monitored based on outputs of a plurality of N-1th equipment operation status monitoring devices included in the functional N-1th layer.

3. 3. The equipment operation status monitoring system according to claim 2, wherein the outputs of the first equipment operation status monitoring device through the N-1th equipment operation status monitoring device are of a common dimension.

4. the first equipment operation state monitoring device performs eigenvalue analysis on the input state information, and outputs new state information by reducing the dimension of the state information based on a first eigenvector and a second eigenvector extracted by the eigenvalue analysis; The equipment operation status monitoring system according to claim 3, characterized in that each of the second equipment operation status monitoring device to the N-1 equipment operation status monitoring device performs eigenvalue analysis on the status information output by the operation status monitoring device included in the layer immediately below, and outputs new status information by reducing the dimension of the status information based on the first eigenvector and the second eigenvector extracted by the eigenvalue analysis.

5. The first equipment operation state monitoring device is a first communication unit that receives the state information measured by the sensor; a first control unit that derives reference state information that represents the state information during a predetermined period in the past and stores the reference state information in a predetermined storage unit, and generates current state information that represents the state information after the predetermined period in the past; Equipped with The first control unit 3. The facility operation status monitoring system according to claim 2, wherein the status of the facility is determined based on the generated current status information and the stored reference status information.

6. The second equipment operation state monitoring device is 6. The equipment operation state monitoring system according to claim 5, wherein the reference state information and the current state information derived by a plurality of the first equipment operation state monitoring devices are input.

7. When N is an integer of 2 or more, The Nth equipment operation status monitoring device included in the functional Nth layer is a second communication unit that receives the reference state information and the current state information output from the N-1 equipment operation state monitoring device included in the functional N-1 layer as a layer below the functional N layer; a second control unit that derives new reference state information representative of the reference state information and stores it in a predetermined storage unit, and also generates new current state information representative of the current state information; Equipped with The second control unit is 7. The equipment operation status monitoring system according to claim 6, wherein the status of the equipment group as the monitoring target is determined based on the generated new current status information and the stored new reference status information.

8. The Nth equipment operating state monitoring device performs the following with respect to the new reference state information and the new current state information, respectively: The degree of dispersion V is calculated based on the T2 statistic. The harmony degree H is calculated based on the Q statistic. The equipment operation status monitoring system according to claim 7, characterized in that it is determined whether a harmonious state of the equipment group is maintained as the state of the equipment group based on the degree of dispersion V and the degree of harmony H.

9. The equipment operation status monitoring device included in the layer above the functional first layer is 8. The facility operation status monitoring system according to claim 7, further comprising: receiving the status information measured by the sensor.

10. At least one sensor provided in each piece of equipment to measure status information of the equipment; a first equipment operational state monitoring device that derives primary reference state information representing the state information during a predetermined period in the past, derives primary current state information representing the state information during a monitoring period after the predetermined period in the past, and determines the state of each of the equipment based on the primary reference state information and the primary current state information; a monitoring unit that derives secondary reference state information representing the primary reference state information for an equipment group formed by grouping a plurality of the equipment based on the primary reference state information output by the plurality of first equipment operational state monitoring devices, derives secondary current state information representing the primary current state information for the equipment group based on the primary current state information output by the plurality of first equipment operational state monitoring devices, and determines the state of the equipment group based on the secondary reference state information and the secondary current state information; A facility operation status monitoring system comprising:

11. At least one sensor provided in each piece of equipment to measure status information of the equipment; a monitoring unit that receives the status information for a predetermined period in the past and derives primary reference status information that represents the status information and secondary reference status information that represents the primary reference status information; a first equipment operational state monitoring device that acquires the primary reference state information from the monitoring unit, derives primary current state information that represents the state information during a monitoring period that is later than the predetermined past period, and determines the state of the equipment based on the primary reference state information and the primary current state information; a second equipment operational state monitoring device that acquires the secondary reference state information from the monitoring unit, derives secondary current state information that represents the primary current state information output by the first equipment operational state monitoring device during the monitoring period, and determines the state of an equipment group formed by grouping a plurality of the equipment based on the secondary reference state information and the secondary current state information; A facility operation status monitoring system comprising:

12. The first equipment operation state monitoring device is For the primary reference state information and the primary current state information, The degree of dispersion V is calculated based on the T2 statistic. The harmony degree H is calculated based on the Q statistic.

12. The equipment operation state monitoring system according to claim 10, wherein the equipment status is determined based on the degree of dispersion V and the degree of harmony H, to determine whether a harmonized state of the equipment is maintained.

13. The second equipment operation state monitoring device is For the secondary reference state information and the secondary current state information, The degree of dispersion V is calculated based on the T2 statistic. The harmony degree H is calculated based on the Q statistic. The equipment operation status monitoring system according to claim 11, characterized in that it is determined whether a harmonious state of the equipment group is maintained as the state of the equipment group based on the degree of dispersion V and the degree of harmony H.

14. Measure the status information of each piece of equipment, deriving primary reference state information representing the state information over a predetermined period of time in the past based on the state information; deriving primary current state information representing the state information in a monitoring period after the predetermined past period based on the state information; an equipment operation status monitoring method comprising: determining the status of each of the pieces of equipment based on the primary reference state information and the primary current state information; deriving secondary reference state information representing the primary reference state information for an equipment group formed by grouping a plurality of the pieces of equipment based on a plurality of pieces of primary reference state information; deriving secondary current state information representing the primary current state information for the equipment group based on a plurality of pieces of primary current state information; and determining the status of the equipment group based on the secondary reference state information and the secondary current state information.

15. Measure the status information of each piece of equipment, receiving the state information for a predetermined period in the past, and deriving primary reference state information representative of the state information and secondary reference state information representative of the primary reference state information; acquiring the primary reference state information, deriving primary current state information representative of the state information during a monitoring period after the predetermined past period, and determining the state of the equipment based on the primary reference state information and the primary current state information; acquiring the secondary reference state information, deriving secondary current state information representative of the primary current state information during the monitoring period, and determining the state of an equipment group formed by grouping a plurality of the equipment based on the secondary reference state information and the secondary current state information; A facility operation status monitoring method comprising:

Images