Apparatus, method, and program

Abstract

To provide a device for specifying which abnormality type of abnormality has occurred among two or more predetermined abnormality types for at least one facility.SOLUTION: An acquisition unit configured to acquire measurement values corresponding to outputs of a plurality of sensors configured to monitor states of a plurality of facilities; The device is provided with a specification unit which specifies which of a plurality of predetermined abnormality types of abnormality has occurred in any of the facilities, and an output unit which outputs information indicating the facility and the abnormality type specified by the specification unit, wherein the specification unit specifies which of two or more predetermined abnormality types of abnormality has occurred in at least one facility.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to an apparatus, a method, and a program. [Background technology]

[0002] Patent document 1 and non-patent documents 1 and 2 state that "multiple status values ​​corresponding to the measurement values ​​of multiple sensors 110 are selected according to the range of change or the rate of change, and the selected status values ​​are sorted and displayed according to the range of change or the rate of change" (paragraph 0033 of patent document 1). [Prior art document] [Patent documents] [Patent Document 1] JP 2022-39101 A [Non-patent literature] [Non-Patent Document 1] Kiyoyoshi Odawara, "GRANDSIGHT and Sushi Sensor Integrated Solution," Yokogawa Technical Report Vol. 61 No. 1, 2018, pp. 21-24 [Non-Patent Document 2] Akio Kitajima, Takayuki Sugisaki, "Sushi Sensor: An IIoT Solution for Realizing Sensemaking," Yokogawa Technical Report Vol. 62 No. 2, 2019, pp. 61-68 Summary of the Invention

[0003] In a first aspect of the present invention, there is provided an apparatus comprising: an acquisition unit that acquires measurement values ​​corresponding to the outputs of a plurality of sensors that monitor the status of a plurality of pieces of equipment; an identification unit that identifies, based on the measurement values ​​acquired by the acquisition unit, in which piece of equipment among the plurality of pieces of equipment an abnormality of which type from a plurality of predetermined abnormality types has occurred; and an output unit that outputs information indicating the equipment and abnormality type identified by the identification unit.

[0004] In the above-described device, the identifying unit may identify, for at least one piece of equipment, which type of abnormality has occurred among two or more predetermined abnormality types.

[0005] In any of the above devices, the identification unit may include a calculation unit that calculates a status value for each piece of equipment based on the measurement values ​​acquired by the acquisition unit, a first identification execution unit that identifies equipment in which an abnormality has occurred based on the status value of each piece of equipment, and a second identification execution unit that identifies which type of abnormality has occurred in the equipment identified by the first identification execution unit.

[0006] In the above-described device, the first identification execution unit may identify, among the plurality of pieces of equipment, a piece of equipment in which a change in a state value has occurred that is equal to or exceeds a predetermined change width or rate of change within a predetermined time period, as equipment in which an abnormality has occurred.

[0007] Any of the above devices may further include a memory unit that stores a plurality of data sets including a data set that associates, for at least one piece of equipment, measurement values ​​when an abnormality has occurred in the past with the abnormality type of the abnormality, and the identification unit may identify the abnormality type of the abnormality that has occurred in the at least one piece of equipment by cluster analysis using each data set stored in the memory unit.

[0008] In any of the above devices, the identification unit may identify the anomaly type of an anomaly that has occurred in at least one piece of equipment using a learning model that is generated by a learning process using learning data including a data set that associates measurement values ​​when an anomaly has occurred in the past with the anomaly type of the anomaly, and that outputs an anomaly type according to the measurement values ​​supplied.

[0009] In the above device, the learning model may include a classification model provided for each anomaly type, which classifies whether or not an anomaly of the corresponding anomaly type has occurred based on the supplied measurement value.

[0010] Any of the above devices may further include a memory unit that stores a corresponding level of urgency of the abnormality in advance for each of the plurality of abnormality types, the identification unit further identifying the level of urgency associated with the identified abnormality type, and the output unit further outputting the level of urgency identified by the identification unit.

[0011] In a second aspect of the present invention, there is provided a method comprising: an acquisition step of acquiring measurement values ​​corresponding to the outputs of a plurality of sensors that monitor the status of a plurality of pieces of equipment; an identification step of identifying, based on the measurement values ​​acquired in the acquisition step, in which piece of equipment among the plurality of pieces of equipment an abnormality of which type from a plurality of predetermined abnormality types has occurred; and an output step of outputting information indicating the equipment and abnormality type identified in the identification step.

[0012] In a third aspect of the present invention, there is provided a program that, when executed by a computer, causes the computer to function as an acquisition unit that acquires measurement values ​​corresponding to the outputs of multiple sensors that monitor the status of multiple pieces of equipment, an identification unit that identifies, based on the measurement values ​​acquired by the acquisition unit, which piece of equipment among the multiple pieces of equipment has experienced an abnormality of which type from multiple predetermined abnormality types, and an output unit that outputs information indicating the equipment and abnormality type identified by the identification unit.

[0013] The above summary of the invention does not list all of the necessary features of the present invention, and subcombinations of these features may also constitute inventions. [Brief explanation of the drawings]

[0014] [Figure 1] The configuration of a monitoring system 10 according to an embodiment of the present invention is shown together with facilities 100a to 100c. [Figure 2] 1 shows a configuration of a monitoring device 140 according to an embodiment of the present invention. [Figure 3] 10 shows an operation flow of a monitoring device 140 according to an embodiment of the present invention. [Figure 4]10 shows an example of history data recorded in the storage unit 142 according to the embodiment of the present invention. [Figure 5] 10 shows another example of history data recorded in the storage unit 142 according to the embodiment of the present invention. [Figure 6] 14 shows a data set 1420 stored in the storage unit 142 according to this embodiment. [Figure 7] An example of the operation of the second specific execution unit 1432 will be described. [Figure 8] 1 shows a correspondence table 1421 stored in the storage unit 142 according to the present embodiment. [Figure 9] 5 shows a first example of a display screen 500 output by a monitoring device 140 according to an embodiment of the present invention. [Figure 10] 5 shows a second example of a display screen 500 output by the monitoring device 140 according to the present embodiment. [Figure 11] 10 shows an example of a detection condition designation screen output by the monitoring device 140 according to this embodiment. [Figure 12] 1 shows an example of a trend graph output by the monitoring device 140 according to this embodiment. [Figure 13] 10 shows an example of a detection condition change designation screen output by the monitoring device 140 according to this embodiment. [Figure 14] 10 shows an example of a learning specification screen output by the monitoring device 140 according to this embodiment. [Figure 15] 10 shows an example of a detection condition designation screen according to a modified example of the present embodiment. [Figure 16] 14 shows a second specific execution unit 1432A according to a modified example. [Figure 17] 12 illustrates an example computer 1200 in which aspects of the present invention may be embodied, in whole or in part. DETAILED DESCRIPTION OF THE INVENTION

[0015] The present invention will be described below through embodiments of the invention, but the following embodiments do not limit the scope of the invention according to the claims. Furthermore, not all of the combinations of features described in the embodiments are necessarily essential to the solution of the invention.

[0016] (Monitoring System 10) FIG. 1 shows the configuration of a monitoring system 10 according to this embodiment, along with multiple facilities 100a-c. Each of the facilities 100a-c (hereinafter also referred to as "facility 100") is installed in a plant or the like. Such plants may be, for example, industrial plants such as chemical or metal plants, plants that manage and control wellheads and surrounding areas of gas or oil fields, plants that manage and control hydroelectric, thermal, or nuclear power generation, plants that manage and control environmental power generation such as solar or wind power, or plants that manage and control water supply, sewage, dams, etc. Furthermore, each facility 100 may be installed in a building, various factories, transportation facilities, etc. Such a facility 100 may include one or more process devices, one or more power generation devices, and one or more other devices.

[0017] Each facility 100 may have one or more field devices. The field devices may be, for example, sensor devices such as pressure gauges, flow meters, and temperature sensors, valve devices such as flow control valves and on-off valves, actuator devices such as fans and motors, imaging devices such as cameras or videos that capture the status of a plant or an object such as a facility, audio devices such as microphones or speakers that collect abnormal sounds from a plant or facility or emit alarm sounds, position detection devices that output position information of devices in the facility 100, or other devices.

[0018] Each facility 100 is provided with one or more sensors 110a-i (hereinafter also referred to as "sensors 110") that function as sensor devices, imaging devices, acoustic devices, or the like to monitor the status of the facility 100. Each sensor 110 may be a sensor device built into the facility 100, or may be retrofitted to the facility 100 or installed near the facility 100. Each sensor 110 may measure a physical quantity related to the facility 100 (e.g., acceleration, speed, temperature, pressure, flow rate, vibration, etc.). Two or more sensors 110 may be built into a sensor device for measuring two or more types of physical quantities related to the facility 100. In the example shown in the figure, sensors 110a-c are provided to monitor the status of facility 100a, sensors 110d-f are provided to monitor the status of facility 100b, and sensors 110g-i are provided to monitor the status of facility 100c.

[0019] The monitoring system 10 includes one or more gateway devices 120a-b, a monitoring device 140, and a terminal 150. The one or more gateway devices 120a-b (hereinafter also referred to as "gateway devices 120") are communicatively connected to multiple sensors 110 and receive output values ​​from each sensor 110. In this embodiment, each gateway device 120 communicates with at least one sensor 110 wirelessly, for example, using a Low Power Wide Area (LPWA) protocol such as LoRa. Alternatively, each gateway device 120 may communicate with at least one sensor 110 via a wired connection, or may communicate with at least one sensor 110 using a communication protocol such as HART (registered trademark), BRAIN, Foundation Fieldbus (registered trademark), or ISA100.11a. Each gateway device 120 may be connected to two or more sensors 110 and collect output values ​​from these two or more sensors 110. In the example shown in the figure, the gateway device 120a is connected to the sensors 110a to 110f, and the gateway device 120b is connected to the sensors 110g to 110i.

[0020] Each gateway device 120 is connected to a network 130, and transmits the output of a sensor 110 assigned to each gateway device 120 among the multiple sensors 110 to a monitoring device 140 via the network 130, which is a wide area network such as the Internet or a WAN. In the example shown in the figure, each gateway device 120 is connected to the network 130 via a wireless network such as a mobile phone line. Alternatively, each gateway device 120 may be connected to the network 130 via a wired network such as Ethernet (registered trademark).

[0021] The monitoring device 140 is connected to one or more gateway devices 120 via a network 130 and monitors one or more facilities 100 using output values ​​from multiple sensors 110. The monitoring device 140 may be realized by a computer such as a personal computer (PC), tablet computer, smartphone, workstation, server computer, or general-purpose computer, or may be realized by a computer system in which multiple computers are connected. Such a computer system is also considered a computer in the broad sense. The monitoring device 140 may be a dedicated computer designed for monitoring facilities, or may be dedicated hardware realized by dedicated circuits.

[0022] In the example shown in the figure, the monitoring device 140 is a cloud computing system that provides a facility monitoring environment via a cloud computer service to one or more customers who monitor equipment such as a plant. Alternatively, the monitoring device 140 may be installed on the premises of a specific facility such as a plant, and connected to each sensor 110 or each gateway device 120 via a local area network to provide an on-premise facility monitoring environment. Alternatively, the gateway device 120 connected to at least one sensor 110 may also function as the monitoring device 140.

[0023] The terminal 150 is connected to the monitoring device 140 via the network 130 and displays an equipment monitoring screen output by the monitoring device 140. The terminal 150 also inputs instructions from a supervisor of a plant or the like that has one or more pieces of equipment 100 and transmits them to the monitoring device 140. The terminal 150 may be located within the premises of a facility such as a plant, or may be located in a location remote from the facility such as a plant. The equipment monitoring screen may be displayed on a display or the like provided in the monitoring device 140, and instructions from the supervisor may be input using this display or the like.

[0024] (Monitoring device 140) 2 shows the configuration of a monitoring device 140 according to this embodiment. The monitoring device 140 includes an acquisition unit 141, a storage unit 142, an identification unit 143, an output unit 144, an input unit 145, and a learning processing unit 146.

[0025] ((Acquisition part 141)) The acquiring unit 141 acquires measurement values ​​corresponding to the outputs of the multiple sensors 110 that monitor the states of the multiple pieces of equipment 100. The measurement values ​​corresponding to the outputs of the sensors 110 may be the output values ​​of the sensors 110 themselves, or may be values ​​obtained by performing a predetermined calculation on the output values. As an example, the measurement values ​​may be scores for each principal component (i.e., principal component scores) obtained by performing principal component analysis on the output values ​​from the multiple sensors 110. The acquiring unit 141 may acquire the output values ​​from the sensors 110 as measurement values, or may acquire measurement values ​​by performing a predetermined calculation on the output values ​​from the sensors 110. The acquiring unit 141 may acquire measurement values ​​for each piece of equipment 100.

[0026] The acquisition unit 141 may be connected to one or more gateway devices 120 via the network 130, and may acquire output values ​​of the multiple sensors 110 from the gateway device 120. The acquisition unit 141 may continuously acquire the latest output value of each sensor 110, for example, at a predetermined cycle (every second, every minute, every hour, etc.).

[0027] The acquiring unit 141 may supply the acquired measurement value to the storage unit 142 and the calculating unit 1430. As an example in the present embodiment, the acquiring unit 141 may supply the acquired measurement value to the storage unit 142 and the identifying unit 143 in association with identification information (also referred to as equipment ID) of the equipment 100 where the measurement value was measured.

[0028] ((Storage unit 142)) The storage unit 142 is connected to the acquisition unit 141. The storage unit 142 is, for example, an external storage device such as a hard disk drive, and sequentially stores the measurement values ​​acquired by the acquisition unit 141 in association with timing information such as the time or date related to the measurement values, such as the measurement time by the sensor 110, the acquisition time by the gateway device 120, or the acquisition time by the monitoring device 140. In this way, the storage unit 142 may store history data of measurement values ​​of multiple sensors 110.

[0029] The storage unit 142 may store multiple data sets 1420 for at least one facility 100 (each facility 100 in this embodiment, as an example). The multiple data sets 1420 may include a data set 1420 when an abnormality occurs. The data set 1420 when an abnormality occurs may associate measurement values ​​when an abnormality occurred in the past with the abnormality type of the abnormality, and may include, for example, measurement values ​​corresponding to the output from the sensor 110 when the abnormality occurred and identification information (also referred to as an abnormality ID) of the abnormality type. The data set 1420 when an abnormality occurs may further include the name of the abnormality type. The multiple data sets 1420 may further include a data set 1420 when normal. The data set 1420 when normal may include measurement values ​​corresponding to the output from the sensor 110 when no abnormality occurs and an abnormality ID indicating that no abnormality exists. However, the data set 1420 when normal does not have to be stored in the storage unit 142.

[0030] The type of abnormality occurring in the equipment 100 may be common to multiple pieces of equipment 100 or may differ between the pieces of equipment 100. For example, in equipment 100 having pipes, an abnormal type of abnormality called a pipe rupture may occur. In equipment 100 having actuators, an abnormal type of abnormality called a stop due to breakage may occur. In equipment 100 having a conveyor belt, an abnormal type of abnormality called a break in the conveyor belt may occur. In equipment 100 having conveyor rolls, an abnormal type of abnormality called resonance of the conveyor rolls may occur. In equipment 100 that communicates via LAN, an abnormal type of abnormality called a communication interruption (for example, a periodic communication interruption) may occur. In equipment 100 having a liquid pump, an abnormal type of abnormality called dry operation (also referred to as idling) may occur. In equipment 100 having a sealed portion with a shaft seal (also referred to as a shaft seal), an abnormal type of abnormality called a leak from the sealed portion may occur. In equipment 100 that contains liquid inside, an abnormal type of abnormality called cavitation may occur. In equipment 100 having a hydraulic pump, an abnormal type of abnormality may occur, which is a decrease in performance (a decrease in flow rate, for example) due to oil deterioration.

[0031] The storage unit 142 may store a correspondence table 1421 in which the urgency of an anomaly is associated with each of a plurality of anomaly types in advance. In the present embodiment, as an example, the storage unit 142 may store an association table 1421 in which the anomaly ID of each anomaly type that may occur in the equipment 100, the name of the anomaly type, the nature of the anomaly, and the urgency of the anomaly are associated with each other. The nature of the anomaly may indicate how the anomaly occurred, and may indicate whether the anomaly is an anomaly that cannot be predicted by prior inspection or monitoring (also referred to as a sudden failure), an anomaly that occurs intermittently and repeatedly (also referred to as an intermittent failure), or an anomaly caused by gradual deterioration of mechanical characteristics that can be predicted to some extent by prior inspection or monitoring (also referred to as a degradation failure). The urgency of the anomaly may indicate how quickly an operator should respond to the anomaly, and may be, for example, one of "high," "medium," and "low."

[0032] ((Specific Section 143)) The identification unit 143 identifies, based on the measurement values ​​acquired by the acquisition unit 141, in which of the multiple pieces of equipment 100 an abnormality of which type has occurred among a plurality of predetermined abnormality types. The identification unit 143 may identify in which piece of equipment 100 in the monitoring system 10 an abnormality has occurred, or may identify which type of abnormality has occurred among a plurality of predetermined abnormality types. The identification unit 143 may identify which type of abnormality has occurred among two or more predetermined abnormality types for at least one piece of equipment 100, and in the present embodiment, as an example, may identify which type of abnormality has occurred for each of the multiple pieces of equipment 100. The identification unit 143 includes a calculation unit 1430, a first identification execution unit 1431, a second identification execution unit 1432, and a third identification execution unit 1433.

[0033] (((calculation unit 1430))) The calculation unit 1430 calculates a state value for each piece of equipment 100 based on the measurement values ​​acquired by the acquisition unit 141. The calculation unit 1430 may calculate a state value for each piece of equipment 100 according to the measurement values ​​supplied from the acquisition unit 141. In this embodiment, the calculation unit 1430 has a model for calculating a state value for each piece of equipment 100 from the measurement values ​​acquired for that piece of equipment 100, and calculates a state value for each piece of equipment 100 from at least one measurement value using the model. This model may calculate, based on the measurement values ​​for each piece of equipment 100, a diagnosis value indicating the result of diagnosing the equipment 100, such as the health or normality of the equipment 100, or the degree of abnormality of the equipment 100, as the state value of that equipment 100.

[0034] The calculation unit 1430 may store a state value calculated based on a measurement value at a certain time in the storage unit 142 in association with the measurement value. Furthermore, the calculation unit 1430 may output at least one measurement value for the equipment 100 as a state value of the equipment 100. In this way, each "state value" may be the above-mentioned diagnostic value or the measurement value itself.

[0035] The calculation unit 1430 may supply the state value calculated for each facility 100 to the first identification execution unit 1431. As an example in the present embodiment, the calculation unit 1430 may associate the state value of each facility 100 with the facility ID of the facility 100 and supply the state value to the first identification execution unit 1431. The calculation unit 1430 may supply the measurement value for each facility 100 supplied from the acquisition unit 141 to the second identification execution unit 1432.

[0036] (((First specific execution unit 1431))) The first identification execution unit 1431 identifies equipment 100 in which an abnormality has occurred based on the state values ​​of each equipment 100. The first identification execution unit 1431 may identify, among the multiple equipment 100, equipment 100 in which a change in state value has occurred that is equal to or greater than a predetermined change width or rate of change within a predetermined time period, as equipment 100 in which an abnormality has occurred. The first identification execution unit 1431 may detect, among the state values ​​of each equipment 100, a state value in which a change in state value has occurred that is equal to or greater than a predetermined change width or rate of change within a predetermined time period, and may identify equipment 100 corresponding to the detected state value as equipment 100 in which an abnormality has occurred.

[0037] Here, the predetermined length of time may be the length of a determination period for determining a change in the state value. The predetermined change width or rate of change may be a determination threshold for determining whether to detect a state value. The first identification execution unit 1431 may detect a state value that has changed by more than a predetermined change width or rate of change during a predetermined time period as a state value indicating the occurrence of an abnormality. For example, the first identification execution unit 1431 may be configured to detect a state value that has dropped by 0.3 or more over a two-hour period. The first identification execution unit 1431 may determine, for each of the multiple state values, whether there is a change of more than a predetermined change width or rate between the latest state value and a state value a predetermined length of time ago (e.g., a state value two hours ago), and detect a state value determined to have such a change. In this embodiment, the first identification execution unit 1431 detects a state value based on whether there has been a change in the state value of more than a predetermined change width. Alternatively, the first identification execution unit 1431 may detect the state value based on whether or not there is a change in the state value that is equal to or greater than a predetermined rate of change. Note that the first identification execution unit 1431 may determine whether or not there is a change between the maximum and minimum values ​​of the state value within the determination period that is equal to or greater than a predetermined range or rate of change.

[0038] The first identification execution unit 1431 may identify the equipment 100 corresponding to the detected status value as the equipment 100 in which an abnormality has occurred, among the multiple equipment 100. The first identification execution unit 1431 may supply the equipment ID of the identified equipment 100 to the second identification execution unit 1432.

[0039] (((Second specific execution unit 1432))) The second identification execution unit 1432 identifies which type of anomaly has occurred in the equipment 100 identified by the first identification execution unit 1431. The second identification execution unit 1432 may identify the anomaly type of the anomaly that has occurred in at least one piece of equipment 100 (in this embodiment, each piece of equipment 100 is used as an example) by cluster analysis using each data set 1420 stored in the storage unit 142. A specific cluster analysis method will be described in detail later. The second identification execution unit 1432 may supply the anomaly ID of the identified anomaly type and the equipment ID of the equipment 100 in which the anomaly has occurred to the third identification execution unit 1433.

[0040] (((Third specific execution unit 1433))) The third identification execution unit 1433 identifies the urgency associated with the abnormality type identified by the second identification execution unit 1432. The third identification execution unit 1433 may identify the urgency associated with the identified abnormality type in the correspondence table 1421. The third identification execution unit 1433 may supply information indicating the identified urgency and abnormality type, and the facility ID of the facility 100 in which the abnormality has occurred, to the output unit 144. The information indicating the abnormality type may be the abnormality ID, or may be the name of the abnormality type or the nature of the abnormality.

[0041] ((output unit 144)) The output unit 144 outputs information indicating the equipment 100 and the abnormality type identified by the identification unit 143. The output unit 144 may further output the urgency associated with the abnormality type identified by the identification unit 143. The output unit 144 may output information indicating the identified equipment 100, the abnormality type, etc. to the terminal 150, and may display it via the terminal 150. The output unit 144 may display the identified equipment 100, the abnormality type, and the urgency in descending order of urgency.

[0042] ((Input unit 145)) The input unit 145 inputs various instructions to the monitoring device 140 from the terminal 150. For example, the input unit 145 inputs an instruction to the calculation unit 1430 to learn a model used to calculate a state value for at least one of the multiple state values.

[0043] The input unit 145 may input a designation of the length of a determination period (a predetermined time length) and a determination threshold (at least one of a predetermined change range or a change rate) for at least one of the multiple state values. The input unit 145 sets the designated time length, change range, etc. in the first determination execution unit 1431 as the time length, change range, etc. that the first determination execution unit 1431 uses to detect the state value.

[0044] Furthermore, the input unit 145 may input, for at least one of the multiple state values, a designation of a change period for changing the detection conditions (for example, at least one of a predetermined time length and a predetermined change width or change rate) of the state value to be detected by the first identification execution unit 1431, and a designation of a change value of the detection conditions. When an instruction to change the detection conditions is input via the terminal 150 by a monitor or the like, the first identification execution unit 1431 detects the state value for which the change period is designated by changing at least one of the predetermined time length and the predetermined change width or change rate to the designated change value during the change period.

[0045] Furthermore, the input unit 145 may input a designation of an exclusion period for excluding at least one of the plurality of state values ​​from detection by the first identification execution unit 1431. When a designation of an exclusion period by an observer or the like is input via the terminal 150, the first identification execution unit 1431 prevents detection of the state value for which the exclusion period is designated during the exclusion period.

[0046] ((Learning processing unit 146)) The learning processing unit 146 is connected to the storage unit 142 and the calculation unit 1430. In response to an instruction to learn the model of the calculation unit 1430 being input from the terminal 150 via the input unit 145, the learning processing unit 146 performs processing to generate a model by learning based on history data of multiple measurement values ​​stored in the storage unit 142. The learning processing unit 146 may perform learning processing on a model used to calculate any one of the state values ​​of each facility 100. Then, the learning processing unit 146 provides the learned model to the calculation unit 1430.

[0047] According to the above-described monitoring device 140, it is possible to identify which equipment 100 has experienced an abnormality of which type based on the acquired measurement values, and to output information indicating the identified equipment 100 and the abnormality type. Therefore, unlike when an operator determines which equipment 100 has experienced an abnormality and which type of abnormality, it is possible to accurately and quickly identify and notify which equipment 100 has experienced an abnormality and which type of abnormality.

[0048] Furthermore, for at least one piece of equipment 100, it is possible to identify which type of abnormality has occurred out of two or more predetermined abnormality types, so even if two or more types of abnormalities can occur in one piece of equipment 100, it is possible to accurately identify which type of abnormality has occurred and report it.

[0049] Furthermore, a status value for each facility 100 is calculated based on the acquired measurement values, and the facility 100 in which an abnormality has occurred is identified based on the status value of each facility 100, and it is then determined which type of abnormality has occurred in the identified facility 100. This makes it possible to prevent the type of abnormality from being identified for a facility 100 in which no abnormality has occurred.

[0050] Furthermore, among the multiple pieces of equipment 100, the equipment 100 in which a change in the status value that is equal to or exceeds a predetermined change width or change rate within a predetermined time period is identified as the equipment 100 in which an abnormality has occurred. Therefore, the equipment 100 in which a change in the status value with a higher degree of importance has occurred can be preferentially identified as the equipment 100 in which an abnormality has occurred. Therefore, unlike a case in which the equipment 100 in which a minor change in the status value has occurred is identified, it is possible to speed up the response to the abnormality in the equipment 100, increase the operating rate of the plant, etc., and reduce the occurrence of abnormalities.

[0051] Furthermore, the urgency of the abnormality is previously associated with each of a plurality of abnormality types and stored, and the urgency associated with the identified abnormality type is further output. Therefore, unlike when an operator determines the urgency, an appropriate urgency can be notified. This allows for quicker response to abnormalities in the facility 100, improving the availability of the plant, etc., and reducing the occurrence of abnormalities.

[0052] Furthermore, for at least one piece of equipment 100, the anomaly type of an anomaly that has occurred in the equipment 100 is identified by cluster analysis using a data set 1420 that associates measurement values ​​when an anomaly has occurred in the past with the anomaly type of the anomaly. Therefore, the anomaly type can be identified more accurately.

[0053] 3 shows the operation flow of the monitoring device 140 according to this embodiment. The monitoring device 140 performs the processes of steps S301 to S313 to support the monitoring of each facility 100. Note that in this figure, the operation when an abnormality occurs in any of the facilities 100 will be described.

[0054] In step S301, the acquisition unit 141 acquires measurement values ​​corresponding to the outputs of the multiple sensors 110. The acquisition unit 141 may acquire the latest output values ​​from the sensors 110 as the measurement values, or may acquire the results of a predetermined operation performed on the output values ​​as the measurement values. The acquisition unit 141 may store the acquired measurement values ​​as a new data set 1420 in the storage unit 142. The anomaly types included in the data set 1420 may be stored in step S311, which will be described later.

[0055] In step S303, the calculation unit 1430 calculates a state value of each of the multiple pieces of equipment 100 using a model associated with that equipment. In the present embodiment, as an example, the model associated with the equipment 100 defines a calculation method for inputting one or more measurement values ​​of each sensor 110 assigned to the equipment 100 at a certain time point (e.g., the current time point or the latest time point) and calculating a diagnostic value of the equipment 100 at a certain time point (e.g., the current time point or the latest time point) from these measurement values. The calculation unit 1430 may use such a model to calculate a state value from the measurement value (also referred to as the most recent measurement value) acquired in step S301. Note that, when it is specified that any measurement value be output as the state value, the calculation unit 1430 may output that measurement value as the state value as is.

[0056] In step S305, the first identification execution unit 1431 identifies which of the multiple pieces of equipment 100 has experienced an abnormality. The first identification execution unit 1431 may perform the identification based on the most recent measurement value, and in the present embodiment, as an example, the identification may be performed based on a state value calculated for each piece of equipment 100 by the calculation unit 1430 using the most recent measurement value.

[0057] The first identification execution unit 1431 may detect a state value that has changed by a predetermined amount or rate of change or more within a predetermined time period, among the multiple state values ​​calculated by the calculation unit 1430. The first identification execution unit 1431 may identify, among the multiple pieces of equipment 100, the equipment 100 corresponding to the detected state value as the equipment 100 in which an abnormality has occurred.

[0058] In step S307, the second identification execution unit 1432 identifies which type of abnormality has occurred in the equipment 100 (also referred to as the abnormal equipment 100) identified in step S303. The second identification execution unit 1432 may perform the identification based on the most recent measurement value.

[0059] For example, the second identification execution unit 1432 may identify the type of abnormality that has occurred in the abnormal equipment 100 by performing cluster analysis using each data set 1420 stored in the storage unit 142 for the abnormal equipment 100. The second identification execution unit 1432 may read the data set 1420 from the storage unit 142 and perform cluster analysis.

[0060] Here, in each data set 1420 in the storage unit 142 according to this embodiment, measurement values ​​are associated with an anomaly ID of an anomaly type. The second identification execution unit 1432 may read each data set 1420 for the faulty equipment 100 from the storage unit 142, perform cluster analysis on a set of measurement values ​​associated with each anomaly ID, and classify each measurement value into multiple clusters. The second identification execution unit 1432 may perform cluster analysis using a conventionally known method, and may perform cluster analysis using the k-means method with the number of clusters set equal to the number of anomaly IDs included in the data set 1420 for the faulty equipment 100, or may perform cluster analysis using another method without specifying the number of clusters. Measurement values ​​included in each cluster may share the same anomaly ID associated in the data set 1420. The second identification execution unit 1432 may associate the anomaly ID associated in the data set 1420 with each measurement value included in the cluster for the center point of each cluster. The second identification execution unit 1432 may identify, among the center points of each cluster, the center point closest to the most recent measurement value for the faulty equipment 100, and identify the abnormality ID associated with the center point as the abnormality ID of the occurred abnormality. Alternatively, the second identification execution unit 1432 may identify, among the measurement values ​​in each data set 1420 for the faulty equipment 100, the measurement value closest to the most recent measurement value for the faulty equipment 100, and identify the abnormality ID associated with the measurement value in the data set 1420 as the abnormality ID of the occurred abnormality.

[0061] The second identification execution unit 1432 may perform cluster analysis including the most recent measurement value for the faulty equipment 100. In this case, the second identification execution unit 1432 may read each data set 1420 for the faulty equipment 100 from the storage unit 142, and perform cluster analysis on a set of measurement values ​​associated with the respective abnormality IDs (i.e., measurement values ​​in the data set 1420) and the most recent measurement value. The second identification execution unit 1432 may identify, as the abnormality ID of the occurred abnormality, an abnormality ID associated in the data set 1420 with another measurement value included in the same cluster as the most recent measurement value.

[0062] In step S309, the third identification execution unit 1433 identifies the urgency associated with the identified abnormality type in the correspondence table 1421. The third identification execution unit 1433 may further identify the name of the abnormality type associated with the identified abnormality type in the correspondence table 1421, etc.

[0063] In step S311, the output unit 144 outputs information indicating the equipment 100, the abnormality type, and the urgency level identified in steps S305, S307, and S309. The output unit 144 may output information further including the most recent measurement value and status value. The output unit 144 may display the information on the terminal 150. The output unit 144 may store the abnormality ID and the name of the identified abnormality type in the data set 1420 stored in the storage unit 142 in step S301.

[0064] In step S313, the input unit 145 determines whether or not an instruction has been input from the terminal 150. If it is determined that no instruction has been input (step S313; No), the process may proceed to step S301 described above. As a result, the monitoring device 140 repeats the process from S301 for the measurement value at the next time point. If it is determined that an instruction has been input (step S313; Yes), the process may proceed to step S315.

[0065] In step S315, the input unit 145 etc. performs processing according to the input instruction. For example, when an instruction is input of the length of a determination period (a predetermined time length) and a determination threshold (at least one of a predetermined change range or change rate) for at least one of the state values ​​for each facility 100, the input unit 145 may set the instructed time length, change range, etc. in the first identification execution unit 1431 as the time length, change range, etc. that the first identification execution unit 1431 uses to detect the state value. When an instruction for a change period for changing the detection conditions (e.g., at least one of a predetermined time length and a predetermined change width or change rate) of the state value by the first identification execution unit 1431 and an instruction for a change value of the detection conditions are input for at least one of the state values ​​for each facility 100, the first identification execution unit 1431 may change at least one of the predetermined time length and the predetermined change width or change rate to the instructed change value for the specified state value during the instructed change period. When an instruction for an exclusion period for excluding at least one of the state values ​​for each facility 100 from detection by the first identification execution unit 1431 is input, the first identification execution unit 1431 may set the state value for which the exclusion period is specified not to be detected during the exclusion period. After the processing of step S315 is completed, the processing may proceed to the above-mentioned step S303, and the processing from step S303 onwards may be performed again using the set conditions, etc. The specific processing of step S315 will be described later with reference to FIGS.

[0066] In the above operation, the second identification execution unit 1432 identifies the anomaly type of the anomaly that occurred in the abnormal equipment 100 by cluster analysis using the data sets stored in the storage unit 142, and therefore identifies the anomaly type included in any of the data sets 1420. However, the second identification execution unit 1432 may also identify a new anomaly type that is not included in any of the data sets 1420. As an example, the second identification execution unit 1432 may perform cluster analysis on the measurement values ​​of each data set 1420 to calculate a measurement value indicating the center point of each cluster, and may identify that a new anomaly type has occurred if the distance between the most recent measurement value and the measurement value of each center point is greater than a reference distance. In this case, the output unit 144 may output information indicating the equipment 100 identified in step S305 and that a new anomaly type has occurred.

[0067] In addition, although it has been described that the length of the judgment period and the judgment threshold are changed in response to an input instruction in step S315, other processing may also be performed. For example, if a correction instruction for the anomaly type output in step S311 is input to the input unit 145 in step S313, the input unit 145 may correct the anomaly ID and the name of the anomaly type stored in the data set 1420 in step S311 in accordance with the correction instruction. As an example, the anomaly ID and name stored in the storage unit 142 may be changed to an anomaly ID and name of another data set 1420 already stored in the storage unit 142, or may be changed to a new anomaly ID and name not stored in the storage unit 142.

[0068] 4 shows an example of history data stored in the storage unit 142 according to this embodiment. The storage unit 142 stores history data including measurement values ​​corresponding to output values ​​output from the multiple sensors 110 at multiple times, such as at predetermined intervals, and multiple state values ​​calculated for the multiple pieces of equipment 100.

[0069] The historical data shown in this figure has a time field for recording time, and multiple equipment fields for recording measurement values ​​and status values ​​for each of multiple pieces of equipment 100. Each equipment field includes at least one measurement value field for recording the measurement value of the corresponding piece of equipment 100, and a status value field for recording the status value of the corresponding piece of equipment 100. Here, the time recorded in the time field may be the time associated with the measurement value or the status value. Note that in this figure, measurement values ​​of acceleration, speed, and temperature are stored for each of the equipment 100 "equipment (1)" and the equipment 100 "equipment (2)."

[0070] 5 shows another example of history data recorded in the storage unit 142 according to this embodiment. In this figure, for the facility 100 "facility (1)", the acceleration, speed, and temperature corresponding to the output values ​​from the two sensors 110 are stored.

[0071] 6 shows a data set 1420 stored in the storage unit 142 according to this embodiment. The storage unit 142 according to this embodiment may store a data set 1420 when an abnormality occurs. Each data set 1420 includes a field for recording a measurement value, a field for recording an abnormality ID, and a field for recording the name of the abnormality type.

[0072] 7 shows an example of the operation of the second identification execution unit 1432. The second identification execution unit 1432 may identify which type of abnormality has occurred for each piece of equipment 100. In the example shown in this figure, the second identification execution unit 1432 identifies an abnormality of the abnormality type "cavitation" in equipment (1) based on the measurement values ​​at each time, identifies an abnormality of the abnormality type "dry operation" in equipment (2), identifies an abnormality of the abnormality type "pipe rupture" in equipment (3), and shows a case where no abnormality has been identified in equipment (4).

[0073] 8 shows a correspondence table 1421 stored in the storage unit 142 according to this embodiment. The correspondence table 1421 may store, for each of a plurality of abnormality types, an abnormality ID, the name of the abnormality type, the type of abnormality, and the urgency of the abnormality in association with each other.

[0074] 9 shows a first example of a display screen 500 output by the monitoring device 140 according to this embodiment. The output unit 144 performs display processing for displaying the identification result by the identification unit 143.

[0075] Display screen 500 includes list 510 and setting button 515. List 510 displays entries corresponding to at least one status value in order sorted by urgency. For each entry, list 510 displays the urgency of the abnormality, the equipment name, the name of the abnormality type, and the current status value.

[0076] In this embodiment, the list 510 displays each entry in descending order of urgency. The status value may be an index indicating the health of the corresponding equipment 100. In the example shown in the figure, the "cooling pump" with a status value of -0.54 and the "mechanical pump" with a status value of -0.44 are experiencing an abnormality with a "high" urgency and an abnormality type of "pipe burst." This allows the monitoring device 140 to prioritize displaying equipment 100 experiencing an abnormality with a high urgency on the terminal 150, thereby enabling faster inspection and other measures.

[0077] The output unit 144 may change the display format depending on the level of urgency. For example, the output unit 144 may emphasize the higher the level of urgency by changing the color of the text or background to a more noticeable color such as red, by making the text larger, by making the text bold, by adding auxiliary lines such as underlines, or by changing the display format in other ways.

[0078] The setting button 515 is a button for issuing an instruction to display a detection condition specification screen for specifying at least one of the length of the determination period (two hours in this figure) and the determination threshold value used for detecting the state value by the first identification execution unit 1431. The detection condition specification screen will be described later with reference to FIG.

[0079] Furthermore, the output unit 144 may perform processing to display a graph display button 520 in association with each of at least one state value sorted on the display screen 500. The graph display button 520 is a button for instructing the display of a trend graph related to each state value. The display of the trend graph will be described later with reference to FIG. 12.

[0080] The output unit 144 may perform processing to display a change button 530 in association with each of at least one status value sorted on the display screen 500. The change button 530 is a button for issuing an instruction to display a detection condition change specification screen for changing the detection condition of the first identification execution unit 1431 for the corresponding status value during a specified change period. The detection condition change specification screen will be described later with reference to FIG. 13.

[0081] The output unit 144 may perform processing to display an exclusion button 540 in association with each of at least one status value sorted on the display screen 500. The exclusion button 540 is a button for issuing an instruction to display a screen for excluding the corresponding status value from detection by the first identification execution unit 1431 for a specified exclusion period. This screen will be described later with reference to FIG. 13.

[0082] The output unit 144 may perform processing to display a learning button 550 in association with each of the sorted state values. The learning button 550 is a button for instructing learning of a model that calculates each state value. In this embodiment, the learning button 550 is used to instruct displaying a learning specification screen for specifying the learning of a model that calculates each state value. The learning specification screen will be described later with reference to FIG. 14.

[0083] The output unit 144 may display the display screen 500 as one of the display components included in the monitoring screen, such as a dashboard, a window, or a sub-window, etc. In this way, the output unit 144 can display, on the monitoring screen, a plurality of display screens 500 each having different parameters, such as a time length or a change width, used to detect the state value.

[0084] FIG. 10 illustrates a second example of a display screen 500 output by the monitoring device 140 according to this embodiment. As illustrated in this figure, the output unit 144 performs processing to display that no detected status values ​​exist, i.e., that there are no abnormal facilities 100, in response to the first identification execution unit 1431 detecting none of the multiple status values. In this example, the output unit 144 displays on the terminal 150 that none of the facilities 100 meet the detection conditions in response to the detection of none of the status values ​​of any of the facilities 100. This allows the output unit 144 to clearly indicate that there are no noteworthy facilities 100, thereby reducing the management burden on a plant or the like operating normally. The display indicating that there are no detected status values ​​may be clearly displayed by, for example, displaying "0 hits" to indicate that there are zero detected cases, or may implicitly indicate that there are no detected status values ​​by displaying a list with no entries and only a title line, or may indicate that there are no detected status values ​​by other methods.

[0085] 11 shows an example of a detection condition specification screen output by the monitoring device 140 according to this embodiment. The output unit 144 performs processing to display the detection condition specification screen in this figure in response to pressing of the setting button 515 (see FIG. 9) on the display screen 500. This detection condition specification screen is a screen for inputting specifications of the length of the determination period (a predetermined time length) and the determination threshold (a predetermined change width or rate of change) used by the first identification execution unit 1431 to detect the state value.

[0086] The detection condition specification screen uses a "Determination Period" field to accept the length of the determination period in hours, for example. Alternatively, the detection condition specification screen may accept the length of the determination period in other units such as seconds, minutes, days, weeks, or months.

[0087] The detection condition specification screen also uses a "threshold filter" field to accept a judgment threshold to be used during the judgment period. In the "threshold filter" field, "Only equipment with a drop of (blank) or more" accepts input of a judgment threshold for the drop of the status value in the blank field. For example, as shown in this figure, if the "threshold filter" field is specified as "Only equipment with a drop of 0.2 or more," the first identification execution unit 1431 will detect status values ​​that have dropped by 0.2 or more during the judgment period (e.g., 2 hours) specified in the "judgment period" field. Note that "Do not filter (display all)" in the "threshold filter" field specifies that all status values ​​are to be displayed. When this item is selected, the first identification execution unit 1431 will detect all status values.

[0088] 3 , the detection condition specification screen may have a save button for saving the specification and setting it in the first identification execution unit 1431, and a cancel button for canceling the specification. When the detection condition specification is input via the detection condition specification screen in step S313 of FIG. 3 , the detection condition input unit 145 sets the specified detection condition in the first identification execution unit 1431 in step S315. This allows the first identification execution unit 1431 to detect the state value using the detection condition input via the detection condition specification screen in response to re-execution of step S305. Note that the output unit 144 may display, as the detection condition specification screen, a screen for specifying only one of the length of the judgment period or the judgment threshold, and the input unit 145 may set the specified length of the judgment period or the judgment threshold in the first identification execution unit 1431.

[0089] 12 shows an example of a trend graph output by the monitoring device 140 according to this embodiment. In response to pressing of a graph display button 520 (see FIG. 9 ) on the display screen 500 that corresponds to the status value of a certain facility 100, the output unit 144 performs processing to display a trend graph related to the status value of that facility 100. Specifically, the output unit 144 outputs, to the display screen, axes, axis titles, axis scales, etc. for the horizontal and vertical axes of the trend graph. In addition, the output unit 144 reads out, from the storage unit 142, the status values ​​of the display target at each time during the display period of the trend graph and plots them on the trend graph.

[0090] The trend graph shows the change over time of the corresponding state value. In this figure, the horizontal axis represents time, and the vertical axis represents the state value. In addition to the graph of the change over time of the state value itself, the output unit 144 may display additional information, such as auxiliary lines or annotations, on the terminal 150, which allows visual confirmation of the range or rate of change of the state value. The trend graph of this example emphasizes points on the graph corresponding to the state value at the start and end of the judgment period determined to satisfy the detection condition by circling them. The trend graph of this example also includes dashed lines along the vertical axis indicating the start and end timing of the judgment period determined to satisfy the detection condition. The trend graph of this example also includes dashed lines along the horizontal axis indicating the level of the state value at the start of the judgment period determined to satisfy the detection condition. The trend graph of this example also uses speech bubbles or the like to numerically indicate the range of decline in the state value.

[0091] Furthermore, the output unit 144 may perform processing to display a graph of the state values ​​of the target facility 100, as well as a graph of the measurement values ​​of one or more sensors 110 assigned to the facility 100. Specifically, the output unit 144 reads out the measurement values ​​of the display target at each time within the display period of the trend graph from the storage unit 142 and plots them on the trend graph. In this example, the trend graph shows not only the transition of the state values ​​of the facility 100, but also the transition of the measurement values ​​of the temperature and acceleration related to the facility 100 during the same period.

[0092] The output unit 144 described above can display a trend graph of the state value of the specified facility 100 among the state values ​​of the facility 100 detected by the first identification execution unit 1431. This allows the output unit 144 to more quickly display changes in important state values ​​detected by the first identification execution unit 1431, thereby enabling anomalies in a plant or the like to be resolved quickly.

[0093] 13 shows an example of a detection condition change specification screen output by the monitoring device 140 according to this embodiment. In response to pressing of a change button 530 (see FIG. 9) associated with a certain status value on the display screen 500, the output unit 144 performs processing to display the detection condition change specification screen for specifying a change to the detection condition for that status value.

[0094] The detection condition change specification screen is a screen for inputting the specification of a change period for changing the detection conditions for a target state value and the specification of at least one change value of the length of the judgment period or the judgment threshold. The detection condition change specification screen accepts the specification of the change period using a "period" field. In the example shown in the figure, when "No deadline" is selected, the change period input unit 145 inputs the specification of an unlimited change period from the current time. Furthermore, when "Until (blank)" is selected, the change period input unit 145 inputs the specification of a change period from the current time to the date and time input in (blank). Alternatively, the detection condition change specification screen may allow the user to input the start and end dates and times of the change period, or may allow the user to specify multiple change periods.

[0095] The detection condition change specification screen also accepts the specification of a judgment threshold using a "threshold" field. In the example shown in this figure, upon receiving input in the (blank) field for "Do not display in rankings if the value is less than the drop amount (blank)," the change period input unit 145 inputs the value entered in the (blank) field as the judgment threshold for the drop amount of the status value. The change period input unit 145 receives input in the "period" field and the "threshold" field, and sets the judgment threshold for the corresponding status value to the specified value during the specified change period. This judgment threshold becomes the threshold used by the first identification execution unit 1431 to detect (or not detect) the status value. As a result, for a status value for which a change period is specified, the first identification execution unit 1431 can change the judgment threshold to the specified change value during the change period and detect the status value.

[0096] The detection condition change specification screen may accept specification of a change period and a length of a determination period during the change period. In this case, the first identification execution unit 1431 can detect a state value for which a change period is specified by changing the length of the determination period to the specified change value during the change period.

[0097] Furthermore, in response to pressing of an exclusion button 540 (see FIG. 9 ) associated with a certain status value on the display screen 500, the output unit 144 performs processing to display an exclusion period specification screen for specifying a period during which the status value is to be excluded from detection. The exclusion period specification screen may have a "Period" field similar to the "Period" field on the detection condition change specification screen, but may have a screen configuration without a "Threshold" field. The exclusion period specification screen accepts specification of an exclusion period using the "Period" field. The exclusion period input unit 145 receives input in the "Period" field and sets the first identification execution unit 1431 to exclude the corresponding status value from detection during the specified exclusion period. This allows the first identification execution unit 1431 to prevent detection of the status value for which the exclusion period is specified during the exclusion period.

[0098] By making it possible for the monitoring device 140 to change at least one of the length of the judgment period or the judgment threshold for a target state value during the change period, if the cause of a drop or the like of the state value has already been identified, the monitoring device 140 can prevent the state value from being detected even if the state value drops or the like within a range expected for that cause. This allows the monitoring device 140 to prioritize other state values ​​detected by the first identification execution unit 1431 and display them on the terminal 150, thereby prioritizing the investigation of the cause of important changes in the other state values. Furthermore, by preventing any changes in the target state value from being detected during the exclusion period, the monitoring device 140 can prioritize the display of other state values ​​on the terminal 150 in cases where, for example, an abnormality in the equipment 100 or the sensor 110 has already been identified.

[0099] 14 shows an example of a learning specification screen output by the monitoring device 140 according to this embodiment. In response to pressing of a learning button 550 (see FIG. 9) associated with a certain state value on the display screen 500, the output unit 144 performs processing to display a learning specification screen for specifying the learning of a model that calculates that state value. Then, in response to pressing of the learning button 550, the learning processing unit 146 receives a specification regarding the learning of the model and then performs processing to generate the model by learning.

[0100] The learning specification screen according to this embodiment allows input of a specification of a period of learning data to be used in learning a model that calculates a target state value. In the example shown in the figure, the learning specification screen accepts specification of a start date and an end date of the learning data period. Alternatively, the learning specification screen may allow input of multiple learning data periods. Furthermore, the learning specification screen may allow input of hyperparameters (parameters that are set prior to learning, such as the number of layers in a neural network, the number of neurons in each layer, etc.) that do not change during model learning.

[0101] The learning specification screen may have a "Learning" button for instructing the start of model learning using the learning-related specifications. When the "Learning" button is pressed, the input unit 145 inputs the learning data period specified on the learning specification screen, provides it to the learning processing unit 146, and instructs the learning processing unit 146 to start learning.

[0102] When an instruction to start learning is given, the learning processing unit 146 acquires, from the storage unit 142, measurement values ​​of at least one sensor 110 used to calculate a target state value at each timing during a period specified by the input unit 145, and sets the acquired measurement values ​​as learning data. Here, if the target state value is, for example, the state value of the equipment 100a, the learning processing unit 146 includes, in the learning data, the measurement values ​​of the sensors 110a to 110c that monitor the state of the target equipment 100a. In this embodiment, the learning processing unit 146 sets, for each timing during the period of the learning data, a set of measurement values ​​of the sensors 110a to 110c as one sample of the learning data.

[0103] In this embodiment, the learning processing unit 146 generates a model through learning using unsupervised learning. As an example, the learning processing unit 146 may perform a process of generating a model through learning using statistical learning using at least one measurement value included in the training data. As an example of statistical learning, the learning processing unit 146 may use the training data to calculate a probability distribution of samples in a multidimensional space (a space whose dimension is the number of measurement values ​​included in each sample) and provide the calculated probability distribution to the model. Here, since abnormalities rarely occur in a plant or the like and the plant operates almost normally during the training data period, it can be estimated that the more a new set of measurement values ​​deviates from the probability distribution, the more likely the abnormality is. By using this probability distribution, when a new set of measurement values ​​is input, the model can calculate the degree to which the new set of measurement values ​​deviates from the probability distribution. For example, such a model outputs a value as a state value that decreases as the new set of measurement values ​​deviates from the probability distribution. The above model may output a state value that normalizes the degree of deviation from the probability distribution by the standard deviation of the probability distribution.

[0104] Alternatively, the learning processing unit 146 may learn a model through supervised learning. In this case, a label such as normal or abnormal is added to each sample in the learning data. For example, the input unit 145 may input, via a learning specification screen, a specification of a normal period during which the target equipment 100 was normal and an abnormal period during which the target equipment 100 was abnormal, and add a label corresponding to this specification to each sample in the learning data.

[0105] The model may be, for example, a neural network or an SVM. The input unit 145 updates the model parameters so as to reduce the error between the label and the output of the model when each sample in the training data is input to the model. For example, when a neural network is used, the input unit 145 adjusts the weights between neurons of the neural network and the biases of each neuron by a method such as backpropagation using the error between the label and the output value output by the neural network in response to the input of each sample.

[0106] By generating a model by learning by inputting a designation of the period of learning data in response to pressing the learning button 550 for a certain state value on the display screen 500, the monitoring device 140 can generate a model with higher accuracy using measurement values ​​for a period appropriate for calculating each state value. Furthermore, if the state value output by the trained model is not appropriate, the monitoring device 140 can update the model by re-learning by changing the range of measurement values ​​to be used as learning data from among the measurement values ​​stored in the storage unit 142 in response to pressing the learning button 550.

[0107] 15 shows an example of a detection condition designation screen according to a modification of this embodiment. The output unit 144 may perform processing to display the detection condition designation screen in this figure instead of the detection condition designation screen in FIG.

[0108] As in Fig. 11, this detection condition specification screen uses a "determination period" field to accept the length of the determination period. In addition, the detection condition specification screen uses a "target data" field to accept specification of whether to select the diagnosis value of each facility 100 as the status value to be displayed, or the measurement value of each sensor 110 as the status value. When it is specified to select the measurement value of each sensor 110 as the status value, the monitoring device 140 performs the various processes described in relation to Figs. 1 to 14 using the measurement value as the status value instead of the diagnosis value.

[0109] The detection condition specification screen also accepts a selection, using a "direction of change" field, as to whether a state value that has risen to or above the judgment threshold is to be the detection target, or a state value that has fallen to or above the judgment threshold is to be the detection target. When the detection condition input unit 145 inputs an instruction to set a state value that has risen to or above the judgment threshold as the detection target, the first identification execution unit 1431 detects a state value that has changed by an increase or increase rate that is equal to or greater than the judgment threshold during the judgment period. When the detection condition input unit 145 inputs an instruction to set a state value that has fallen to or above the judgment threshold as the detection target, the first identification execution unit 1431 detects a state value that has changed by an increase or decrease rate that is equal to or greater than the judgment threshold during the judgment period.

[0110] 11, the detection condition specification screen also accepts a judgment threshold using a "threshold filter" field. The detection condition specification screen also accepts, using a "moving average processing" field, specification of whether the state value (instantaneous value) at each timing is to be directly subjected to detection by the first identification execution unit 1431 and display by the output unit 144, or whether the result of taking a moving average of the state value (instantaneous value) at each timing is to be used as the state value to be processed.

[0111] When "Do not use" is selected in the "Moving average processing" field, the input unit 145 instructs the calculation unit 1430 to output the instantaneous value of the state value calculated for each timing as the state value. In response to this, the calculation unit 1430 outputs the instantaneous value of the state value calculated for each timing as the state value to be processed. When "Window size: (blank) data" is selected in the "Moving average processing" field and the number of samples of the instantaneous values ​​for which a moving average is to be taken is specified in the (blank) field, the input unit 145 instructs the calculation unit 1430 to take the moving average of the instantaneous values ​​of the state value calculated for each timing for the number of samples specified in the (blank) field, and output the result as the state value to be processed. For example, when the number of samples is specified in the (blank) field as 10 as shown in this figure, the input unit 145 instructs the calculation unit 1430 to output the average of the instantaneous values ​​of the most recent 10 samples at times t, t-1, ..., t-9 as the state value to be processed at time t. In response to receiving such an instruction, the calculation unit 1430 calculates the moving average value of the instantaneous value of the state value at each timing and outputs it as the state value to be processed.

[0112] As a result, even if the measurement value or diagnostic value used as the status value does not change smoothly, the monitoring device 140 can perform moving average processing to detect, sort, and display the status value based on the trend of change in the status value.

[0113] (Variation) 16 shows a second specific execution unit 1432A according to a modified example. The second specific execution unit 1432A may be used in place of the second specific execution unit 1432 in the monitoring device 140 in the above-described embodiment.

[0114] The second identification execution unit 1432A identifies the anomaly type of an anomaly that has occurred in at least one facility 100 (also referred to as the target facility 100) using the learning model 1435. The second identification execution unit 1432A may have the learning model 1435. Note that in this modification, as an example, each facility 100 in the monitoring system 10 will be described as the target facility 100.

[0115] The learning model 1435 is generated by a learning process using learning data including the data set 1420 for the target equipment 100, and outputs an abnormality type according to the supplied measurement value. In this modification, as an example, the learning model 1435 has a plurality of classification models 1436 and a determination unit 1437.

[0116] Each classification model 1436 is provided for each anomaly type, and classifies whether or not an anomaly of the corresponding anomaly type has occurred based on the supplied measurement values. Each classification model 1436 may be generated by supervised learning using a data set in which an anomaly of the corresponding anomaly type has occurred and a data set in which an anomaly of the corresponding anomaly type has not occurred. In response to the supplied measurement values, each classification model 1436 may output a classification result indicating whether or not an anomaly of the corresponding anomaly type has occurred. As an example, each classification model may be an SVM.

[0117] Here, each classification model 1436 according to this modification is provided for each anomaly type and also for each piece of equipment 100, and is generated using a data set when an anomaly of the corresponding anomaly type occurs in the corresponding piece of equipment 100 and a data set when an anomaly of the corresponding anomaly type does not occur in the corresponding piece of equipment 100. In response to the equipment ID of the corresponding piece of equipment 100 being supplied from the first identification execution unit 1431, each classification model 1436 may acquire the measurement value of the corresponding piece of equipment 100 from the calculation unit 1430 and output a classification result indicating whether or not an anomaly of the corresponding anomaly type has occurred.

[0118] The determination unit 1437 determines which type of anomaly has occurred based on the classification results from each classification model 1436. The determination unit 1437 may detect the classification model 1436 that has output a classification result indicating that an anomaly has occurred, and determine the anomaly type corresponding to the classification model 1436 as the anomaly type of the anomaly that has occurred. The determination unit 1437 may output an anomaly ID indicating the anomaly type of the anomaly that has occurred, together with the equipment ID supplied from the first identification execution unit 1431, i.e., the equipment ID of the equipment 100 in which the anomaly has occurred.

[0119] According to the above modification, the anomaly type of an anomaly that has occurred in at least one facility 100 is identified using a learning model 1435 generated by a learning process using learning data including a data set 1420 for the facility 100. Therefore, the anomaly type can be identified with high accuracy.

[0120] The learning model 1435 also includes a classification model 1436 provided for each anomaly type, which classifies whether or not an anomaly of the corresponding anomaly type has occurred based on the supplied measurement value. Therefore, the occurrence of an anomaly of each anomaly type can be reliably identified.

[0121] In the above-described modified example, the learning model 1435 is generated by a learning process using learning data including the data set 1420. Therefore, the second identification execution unit 1432A identifies any of the anomaly types included in the data set 1420 of the learning data as the anomaly type of the anomaly that occurred in the abnormal equipment 100. However, the second identification execution unit 1432 may identify a new anomaly type that is not included in any of the data sets 1420. As an example, if the first identification execution unit 1431 identifies the abnormal equipment 100 but the anomaly type cannot be identified using the learning model 1435, the second identification execution unit 1432 may identify that an anomaly of a new anomaly type has occurred. Alternatively, the second identification execution unit 1432 may identify that an anomaly of a new anomaly type has occurred when the most recent measurement value is an outlier that is different from the measurement value of each data set 1420 by more than a reference value. In this case, the second identification execution unit 1432 may have a model that is generated by unsupervised learning using the measurement values ​​of each data set 1420 in the memory unit 142 as learning data and determines whether the most recent measurement value is an outlier.

[0122] (Other variations) In the above embodiment and modified examples, the first identification execution unit 1431 has been described as detecting a state value that has changed by a predetermined amount or rate of change or more within a predetermined time period as a state value indicating the occurrence of an abnormality, but other state values ​​may be detected. For example, the first identification execution unit 1431 may detect a state value that is outside a predetermined reference range as a state value indicating the occurrence of an abnormality.

[0123] Furthermore, although the second identification execution unit 1432 has been described as identifying the abnormality type for each facility 100 in the monitoring system 10 by either performing cluster analysis or using the learning model 1435, it may also be possible to identify the abnormality type for some of the facilities 100 by performing cluster analysis, and to identify the abnormality type for other facilities 100 by using the learning model 1435.

[0124] In addition, although the identification unit 143 has been described as identifying the equipment 100 in which an abnormality has occurred using the first identification execution unit 1431 and identifying the abnormality type using the second identification execution unit 1432, the identification of the equipment 100 in which an abnormality has occurred and the identification of the abnormality type may be performed together. For example, the identification unit 143 may identify that an abnormality has occurred in at least one equipment 100 and the abnormality type of the abnormality together by cluster analysis using each dataset 1420 stored for at least one equipment 100. Alternatively, the identification unit 143 may identify that an abnormality has occurred in the at least one equipment 100 and the abnormality type of the abnormality using a learning model 1435 generated by a learning process using learning data including a dataset for at least one equipment.

[0125] Furthermore, although the monitoring device 140 has been described as including the storage unit 142, the learning processing unit 146, and the input unit 145, it may not include any of these. If the monitoring device 140 does not include the storage unit 142, the data set 1420 and the correspondence table 1421 may be stored in a storage device externally connected to the monitoring device 140. If the monitoring device 140 does not include the learning processing unit 146, the model of the calculation unit 1430 may be generated by learning processing in an external learning device.

[0126] Various embodiments of the present invention may be described with reference to flowcharts and block diagrams, where the blocks may represent (1) stages of a process in which operations are performed or (2) sections of an apparatus responsible for performing the operations. Particular stages and sections may be implemented by dedicated circuitry, programmable circuitry provided with computer-readable instructions stored on a computer-readable medium, and / or a processor provided with computer-readable instructions stored on a computer-readable medium. Dedicated circuitry may include digital and / or analog hardware circuitry, and may include integrated circuits (ICs) and / or discrete circuits. Programmable circuitry may include reconfigurable hardware circuitry, including logical AND, OR, XOR, NAND, NOR, and other logical operations, flip-flops, registers, memory elements such as field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), and the like.

[0127] A computer-readable medium may include any tangible device capable of storing instructions that are executed by an appropriate device, such that the computer-readable medium having instructions stored thereon comprises an article of manufacture containing instructions that can be executed to create means for performing the operations specified in the flowcharts or block diagrams. Examples of computer-readable media may include electronic, magnetic, optical, electromagnetic, and semiconductor storage media. More specific examples of computer-readable media may include floppy disks, diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), electrically erasable programmable read-only memory (EEPROM), static random access memory (SRAM), compact disc read-only memory (CD-ROM), digital versatile disc (DVD), Blu-ray disc, memory stick, integrated circuit card, and the like.

[0128] The computer readable instructions may include either assembler instructions, Instruction Set Architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state-setting data, or source or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk®, JAVA®, C++, etc., and conventional procedural programming languages ​​such as the “C” programming language or similar programming languages.

[0129] The computer-readable instructions may be provided to a processor or programmable circuit of a programmable data processing device, such as a computer, locally or over a wide area network (WAN) such as a local area network (LAN) or the Internet, and the computer-readable instructions may be executed to create means for performing the operations specified in the flowcharts or block diagrams. Here, the computer may be a personal computer (PC), a tablet computer, a smartphone, a workstation, a server computer, a general-purpose computer, a special-purpose computer, or the like, or may be a computer system in which multiple computers are connected. Such a computer system in which multiple computers are connected is also called a distributed computing system, and is a broad definition of computer. In a distributed computing system, the multiple computers collectively execute a program by each executing a portion of the program and passing data between the computers as needed during program execution.

[0130] Examples of processors include a computer processor, a central processing unit (CPU), a processing unit, a microprocessor, a digital signal processor, a controller, a microcontroller, etc. A computer may have one or more processors. In a multiprocessor system with multiple processors, each processor executes a portion of a program and passes data between processors as needed during program execution, allowing the multiple processors to collectively execute a program. For example, in multitasking, each of the multiple processors may execute a portion of each task in small chunks by switching tasks at time slice intervals. In this case, which portion of a program each processor executes changes dynamically. Which portion of a program each of the multiple processors executes may also be statically determined by multiprocessor-aware programming.

[0131] 17 illustrates an example of a computer 1200 in which aspects of the present invention may be embodied, in whole or in part. Programs installed on the computer 1200 may cause the computer 1200 to function as or perform operations associated with an apparatus or one or more sections of the apparatus according to embodiments of the present invention, and / or to perform a process or steps of a process according to embodiments of the present invention. Such programs may be executed by the CPU 1212 to cause the computer 1200 to perform specific operations associated with some or all of the blocks in the flowcharts and block diagrams described herein.

[0132] A computer 1200 according to this embodiment includes a CPU 1212, a RAM 1214, a graphics controller 1216, and a display device 1218, which are interconnected by a host controller 1210. The computer 1200 also includes input / output units such as a communication interface 1222, a storage device 1224 such as a hard disk drive, a DVD-ROM drive 1226, and an IC card drive, which are connected to the host controller 1210 via an input / output controller 1220. The computer also includes legacy input / output units such as a ROM 1230 and a keyboard 1242, which are connected to the input / output controller 1220 via an input / output chip 1240.

[0133] The CPU 1212 operates according to programs stored in the ROM 1230 and the RAM 1214, thereby controlling each unit. The graphics controller 1216 acquires image data generated by the CPU 1212 into a frame buffer or the like provided in the RAM 1214 or into the graphics controller 1216 itself, and causes the image data to be displayed on the display device 1218.

[0134] The communication interface 1222 communicates with other electronic devices via a network. The storage device 1224 stores programs and data used by the CPU 1212 in the computer 1200. The DVD-ROM drive 1226 reads programs or data from a DVD-ROM 1227 and provides the programs or data to the storage device 1224 via the RAM 1214. The IC card drive reads programs and data from an IC card and / or writes programs and data to an IC card.

[0135] The ROM 1230 stores therein a boot program or the like that is executed by the computer 1200 upon activation, and / or programs that depend on the hardware of the computer 1200. The input / output chip 1240 may also connect various input / output units to the input / output controller 1220 via a parallel port, a serial port, a keyboard port, a mouse port, etc.

[0136] The programs are provided by a computer-readable medium such as a DVD-ROM 1227 or an IC card. The programs are read from the computer-readable medium, installed in the storage device 1224, RAM 1214, or ROM 1230, which are also examples of computer-readable media, and executed by the CPU 1212. Information processing described in these programs is read by the computer 1200, and causes cooperation between the programs and the various types of hardware resources described above. An apparatus or a method may be configured by implementing information manipulation or processing in accordance with the use of the computer 1200.

[0137] For example, when communication is performed between the computer 1200 and an external device, the CPU 1212 may execute a communication program loaded into the RAM 1214 and instruct the communication interface 1222 to perform communication processing based on the processing described in the communication program. Under the control of the CPU 1212, the communication interface 1222 reads transmission data stored in a transmission buffer processing area provided in the RAM 1214, the storage device 1224, the DVD-ROM 1227, or a recording medium such as an IC card, and transmits the read transmission data to the network, or writes reception data received from the network to a reception buffer processing area or the like provided on the recording medium.

[0138] Furthermore, the CPU 1212 may cause all or a necessary portion of a file or database stored in an external recording medium such as the storage device 1224, the DVD-ROM drive 1226 (DVD-ROM 1227), an IC card, etc. to be read into the RAM 1214, and perform various types of processing on the data in the RAM 1214. The CPU 1212 then writes back the processed data to the external recording medium.

[0139] Various types of information, such as various types of programs, data, tables, and databases, may be stored on the recording medium and may undergo information processing. CPU 1212 may perform various types of processing on data read from RAM 1214, including various types of operations, information processing, conditional judgment, conditional branching, unconditional branching, information search / replacement, etc., as described throughout this disclosure and specified by the instruction sequences of the programs, and write the results back to RAM 1214. CPU 1212 may also search for information in a file, database, etc. on the recording medium. For example, if multiple entries each having an attribute value of a first attribute associated with an attribute value of a second attribute are stored on the recording medium, CPU 1212 may search for an entry that matches a condition specified by the attribute value of the first attribute from among the multiple entries, read the attribute value of the second attribute stored in the entry, and thereby obtain the attribute value of the second attribute associated with the first attribute that satisfies a predetermined condition.

[0140] The above-described programs or software modules may be stored in a computer-readable medium on or near the computer 1200. A recording medium such as a hard disk or RAM provided in a server system connected to a dedicated communication network or the Internet can also be used as a computer-readable medium, thereby providing the programs to the computer 1200 via the network.

[0141] Although the present invention has been described above using embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. It will be apparent to those skilled in the art that various modifications and improvements can be made to the above embodiments. It is clear from the claims that such modifications and improvements can also be included within the technical scope of the present invention.

[0142] It should be noted that the execution order of each process, such as operations, procedures, steps, and stages, in the devices, systems, programs, and methods shown in the claims, specifications, and drawings is not specifically stated as "before," "prior to," etc., and that the processes can be performed in any order unless the output of a previous process is used in a subsequent process. Even if the operational flow in the claims, specifications, and drawings is described using "first," "next," etc. for convenience, this does not mean that the processes must be performed in this order. [Explanation of symbols]

[0143] 10. Surveillance System 100 equipment 110 Sensors 120 Gateway Device 130 Network 140 Monitoring equipment 141 Acquisition Department 142 Storage section 143 Specific part 144 Output section 145 Input section 146 Learning processing unit 150 devices 500 display screen 510 List 515 Settings button 520 Graph display button 530 Change button 540 Exclude button 550 Learn button 1200 Computer 1210 host controller 1212 CPU 1214 RAM 1216 Graphics Controller 1218 Display Devices 1220 Input / Output Controller 1222 communication interface 1224 Storage device 1226 DVD-ROM drive 1227 DVD-ROM 1230 ROM 1240 Input / Output Chip 1242 keyboard 1420 datasets 1421 Corresponding Table 1430 Calculation Unit 1431 First Specific Execution Department 1432 Second Specific Execution Department 1433 Third Specific Execution Department 1435 Learning Model 1436 Classification Model 1437 Judgment section

Claims

1. an acquisition unit that acquires measurement values ​​according to outputs of a plurality of sensors that monitor the states of a plurality of pieces of equipment; an identification unit that identifies, based on the measurement value acquired by the acquisition unit, in which equipment among the plurality of equipments an abnormality of which type among a plurality of predetermined abnormality types has occurred; an output unit that outputs information indicating the equipment and the abnormality type identified by the identification unit; An apparatus comprising:

2. The apparatus according to claim 1 , wherein the identifying unit identifies which of two or more predetermined abnormality types has occurred in at least one piece of equipment.

3. The identification unit a calculation unit that calculates a state value for each piece of equipment based on the measurement values ​​acquired by the acquisition unit; a first identification execution unit that identifies a piece of equipment in which an abnormality has occurred based on a state value of each piece of equipment; a second identification execution unit that identifies an abnormality type that has occurred in the equipment identified by the first identification execution unit; 10. The apparatus of claim 1, comprising:

4. 4. The device according to claim 3, wherein the first identification execution unit identifies, among the plurality of pieces of equipment, a piece of equipment in which a change in a state value has occurred that is equal to or greater than a predetermined change width or change rate within a predetermined time period, as equipment in which an abnormality has occurred.

5. The system further includes a storage unit configured to store a plurality of data sets including a data set in which measurement values ​​when an abnormality has occurred in the past and the abnormality type of the abnormality are associated with each other for at least one piece of equipment; The device according to claim 1 , wherein the identification unit identifies an anomaly type of an anomaly that has occurred in the at least one piece of equipment by cluster analysis using each data set stored in the storage unit.

6. 2. The device according to claim 1, wherein the identification unit identifies the anomaly type of an anomaly that has occurred in at least one piece of equipment using a learning model that is generated by a learning process using learning data including a data set that associates measurement values ​​when an anomaly has occurred in the past with the anomaly type of the anomaly, and that outputs an anomaly type corresponding to the measurement value supplied.

7. The device according to claim 6 , wherein the learning model includes a classification model provided for each anomaly type, the classification model classifying whether or not an anomaly of the corresponding anomaly type has occurred based on the supplied measurement value.

8. a storage unit that stores an urgency of the abnormality in advance in association with each of the plurality of abnormality types, The identification unit further identifies an urgency level associated with the identified abnormality type, The device according to claim 1 , wherein the output unit further outputs the urgency level determined by the determination unit.

9. an acquisition stage of acquiring measurement values ​​according to outputs of a plurality of sensors that monitor the states of a plurality of pieces of equipment; an identifying step of identifying, based on the measurement values ​​acquired in the acquiring step, in which of the plurality of pieces of equipment an abnormality of which type among a plurality of predetermined abnormality types has occurred; an output step of outputting information indicating the equipment and the abnormality type identified by the identification step; A method for providing the above.

10. When executed by a computer, the computer an acquisition unit that acquires measurement values ​​according to outputs of a plurality of sensors that monitor the states of a plurality of pieces of equipment; an identification unit that identifies, based on the measurement value acquired by the acquisition unit, in which equipment among the plurality of equipments an abnormality of which type among a plurality of predetermined abnormality types has occurred; an output unit that outputs information indicating the equipment and the abnormality type identified by the identification unit; A program that functions as a

Images