Abnormality detection system, abnormality detection device, abnormality detection method, and program

Abstract

To enable abnormality detection suited to actual usage conditions at each measurement point in a water supply system, thereby achieving earlier and more accurate abnormality detection than conventional abnormality detection using upper and lower limit values.SOLUTION: An abnormality detection system comprises an abnormality detection device and a measurement sensor. The measurement sensor transmits measurement data obtained by measuring values of predetermined determination elements at measurement points in a water supply system to the abnormality detection device. The abnormality detection device comprises: a storage unit that stores the measurement data; a determination model generation unit that generates a determination model indicating a normal region observed in a predetermined cycle using measurement data corresponding to the predetermined cycle in the measurement data in the storage unit; and a determination unit that determines whether a value indicated by target data of determination is normal or abnormal by using the determination model.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to an abnormality detection system, an abnormality detection device, an abnormality detection method, and a program.

Background Art

[0002] There is a monitoring system that monitors the state of a water supply system. In this monitoring system, the occurrence of abnormalities in the water supply system (for example, failure of a pressure reducing valve or water leakage) is monitored. Specifically, the monitoring system acquires sensor data from, for example, a water pressure sensor or a flow rate sensor installed near a branch point between a main water distribution pipe and a sub water distribution pipe, and detects an abnormality in the water supply system when sensor data exceeding a threshold value that defines the upper and lower limits of a normal range is acquired.

[0003] However, the use of the water supply system usually varies depending on the time zone and day of the week, which are different for each consumer (for example, each household or store, users of the water supply system). Therefore, there is a problem that it is difficult to detect an abnormality that conforms to the actual situation within a normal range defined by a uniform threshold value.

[0004] Note that Patent Document 1 discloses an information processing device that determines whether each device in a water treatment system is normal or abnormal using measurement values measured over a past continuous period. Specifically, in the technology of Patent Document 1, a determination model is constructed for each predetermined period, and based on the determination model and the measurement values, it is determined whether the state of each device in the water treatment system is normal or abnormal.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] However, the technology of Patent Document 1 detects abnormalities in each device based on measurement values indicating the state of each device, and does not consider detecting abnormalities occurring in the water supply system. Therefore, even if the technology of Patent Document 1 is adopted, it is difficult to solve the above problem of detecting abnormalities in accordance with the actual usage situation in the water supply system.

[0007] The present invention has been made in view of the above points, and an object thereof is to detect abnormalities in accordance with the actual usage situation at each measurement point in the water supply system.

Means for Solving the Problem

[0008] This application includes a plurality of means for solving at least a part of the above problems. For example, they are as follows.

[0009] In order to solve the above problems, an abnormality detection system according to an aspect of the present invention is an abnormality detection system including an abnormality detection device and a measurement sensor. The measurement sensor transmits measurement data obtained by measuring the value of a predetermined determination element at a measurement point in the water supply system to the abnormality detection device. The abnormality detection device includes a storage unit that stores the measurement data, a determination model generation unit that generates a determination model indicating a normal region found in a period using the measurement data corresponding to a predetermined period among the measurement data in the storage unit, and a determination unit that determines whether the value indicated by the data to be determined is normal or abnormal using the determination model.

[0010] Further, in the above abnormality detection system, the abnormality detection device may generate the determination model based on unsupervised machine learning using the measurement data stored in the storage unit.

[0011] Further, in the above abnormality detection system, the determination element includes a first determination element and a second determination element of different types, and the first determination element and the second determination element may be any of water pressure, flow rate, and water quality.

[0012] In addition, in the above-described abnormality detection system, the determination element includes one type of determination element, and any one of water pressure, flow rate, and water quality may be sufficient.

[0013] In addition, in the above-described abnormality detection system, the period may include at least any one of by time zone, by day of the week, by week, by month, by season, or a combination of these periods, and an arbitrary period designated by the user.

[0014] In addition, in the above-described abnormality detection system, it further includes a display information generation unit that generates display information including the positional relationship between the normal region and the plotted position of the target data and the result of the determination, and the display information generation unit may display the display information on a predetermined display device.

[0015] In addition, in the above-described abnormality detection system, it further includes a display information generation unit that generates display information displaying a graph indicating the degree of normality or abnormality of the value of the target data according to the distance from the normal region to the plotted position of the target data, and the display information generation unit may display the display information on a predetermined display device.

[0016] In addition, in the above-described abnormality detection system, the storage unit stores setting information in which the time zone and day of the week when the determination is executed are associated and registered with information specifying the period of the determination model applied to the determination, and the determination unit may select the determination model at the time of executing the determination based on the setting information.

[0017] In addition, in the above-described abnormality detection system, it may further include a display information generation unit that generates display information in which the normal region of the determination model set for each date and time when the determination is made is superimposed and displayed with a graph indicating the value of the measurement data.

[0018] Further, an abnormality detection device according to another aspect of the present invention includes a storage unit that stores measurement data obtained by measuring the value of a predetermined determination element at a measurement point of a water supply system, and among the measurement data in the storage unit, a determination model generation unit that generates a determination model indicating a normal region found in a period using the measurement data corresponding to a predetermined period, and a determination unit that determines whether the value indicated by the data to be determined is normal or abnormal using the determination model.

[0019] Further, an abnormality detection method according to another aspect of the present invention is an abnormality detection method performed by an abnormality detection device. The abnormality detection device includes a storage step of storing measurement data obtained by measuring the value of a predetermined determination element at a measurement point of a water supply system, a determination model generation step of generating a determination model indicating a normal region found in a period using the measurement data corresponding to a predetermined period among the measurement data stored in the storage step, and a determination step of determining whether the value indicated by the data to be determined is normal or abnormal using the determination model.

[0020] Further, a program according to another aspect of the present invention is a program that causes a computer to function as an abnormality detection device. The computer functions as a storage unit that stores measurement data obtained by measuring the value of a predetermined determination element at a measurement point of a water supply system, a determination model generation unit that generates a determination model indicating a normal region found in a period using the measurement data corresponding to a predetermined period among the measurement data in the storage unit, and a determination unit that determines whether the value indicated by the data to be determined is normal or abnormal using the determination model.

Advantages of the Invention

[0021] According to the present invention, it becomes possible to detect an abnormality in accordance with the actual usage situation at each measurement point of the water supply system.

[0022] Note that other problems, configurations, effects, etc. will be clarified by the description of the following embodiments.

Brief Description of the Drawings

[0023]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Embodiments for Carrying Out the Invention

[0024] Hereinafter, embodiments which are examples of the present invention will be described with reference to the drawings.

[0025] <First Embodiment> <Schematic Configuration of Abnormality Detection System 1000> FIG. 1 is a diagram showing an example of the schematic configuration of an abnormality detection system 1000 according to this embodiment. As shown in the figure, the abnormality detection system 1000 includes an abnormality detection device 100 and a measurement sensor 200, which are connected to be communicable with each other via a predetermined network N. Note that the network N is, for example, the Internet, an intranet, a WAN (Wide Area Network), a mobile phone network, or the like.

[0026] In such an abnormality detection system 1000, a measurement sensor 200 installed at a predetermined point (measurement point) in the water supply pipe periodically measures the water pressure, flow rate, etc. at each point, and transmits the measurement data to the abnormality detection device 100. Further, the abnormality detection device 100 accumulates the measurement data acquired via the network N in its own database.

[0027] Further, the abnormality detection device 100 extracts measurement data corresponding to a predetermined period (for example, a predetermined time zone or day of the week) from the accumulated past measurement data, and performs machine learning using the measurement data to generate a determination model indicating a certain pattern (normal region) found in the period.

[0028] Also, the abnormality detection device 100 uses the latest measurement data acquired from the measurement sensor 200 as data to be determined (hereinafter sometimes referred to as "target data"), and performs processing for detecting an abnormality at each measurement point based on a comparison with the normal region (normal range) indicated by the determination model.

[0029] FIG. 2 is a diagram showing an example of the normal region indicated by the determination model. The illustrated example is a determination model corresponding to a predetermined period (for example, the 8 o'clock hour on every Monday), with the vertical axis indicating the value of water pressure and the horizontal axis indicating the value of flow rate. When the target data 301 deviates from the normal region 300, the abnormality detection device 100 detects an abnormality at the measurement point where the target data 301 was measured. Note that the conventional normal region 310 is defined by a uniform threshold indicating upper and lower limits, and thus has a larger area than the normal region 300 of the determination model. Therefore, even when measurement data that should have an abnormality detected is received, it may be determined as a normal state. In contrast, since the normal region 300 of the determination model 124 has a tight shape (area) according to the actual usage situation, when measurement data different from the normal state is detected, the abnormality detection device 100 can detect the abnormality more accurately (earlier) than when using the conventional normal region 310.

[0030] In this way, the abnormality detection system 1000 performs normal / abnormality determination based on the normal state (normal region 300) of the time zone and day of the week when the determination process is performed using the determination model. As a result, compared with the abnormality detection based on the comparison with the normal region 310 given by a uniform threshold value, the abnormality detection system 1000 can detect a tight abnormality according to the actual usage situation and can detect an abnormality at an earlier stage. Note that based on the determination by the abnormality detection device, abnormalities such as a failure of a pressure reducing valve or water leakage near the measurement point are detected, for example.

[0031] <Measurement sensor 200> The measurement sensor 200 is a sensor that is installed at a predetermined point in the water supply pipeline and measures water pressure and flow rate. Note that the water pressure sensor for measuring water pressure and the flow rate sensor for measuring flow rate are each separate sensors, but in this embodiment, for the sake of convenience, they will be described as one sensor having both functions.

[0032] The measurement sensor 200 is installed, for example, near the branch point between the main water distribution pipe laid downstream from the water distribution tank and the small water distribution pipes branching from the main water distribution pipe toward each water distribution block including a plurality of consumers (for example, inside the manhole buried near the branch point), and measures the water pressure and flow rate at the installation point. Note that if there is a manhole with a pressure reducing valve installed near the branch point, the measurement sensor 200 is installed so as to be able to measure the water pressure and flow rate near the pressure reducing valve.

[0033] Further, the measurement sensor 200 transmits measurement data of water pressure and flow rate to the abnormality detection device 100 via the network N. The measurement data includes, for example, measurement values indicating water pressure and flow rate, the measurement time, and position information indicating the measurement point.

[0034] Note that the timing of measurement and data transmission is arbitrary. For example, measurement and data transmission may be performed every minute or every 10 minutes, or may be performed every 30 minutes or every hour. Also, it is preferable that the measurement timings of water pressure and flow rate are synchronized.

[0035] In addition, the measurement target by the measurement sensor 200 is not limited to water pressure and flow rate. For example, any target that can be a determination factor for abnormality detection, such as water quality, is acceptable. Also, there does not necessarily have to be a plurality of measurement sensors 200. It is sufficient if at least one type of determination factor (for example, any one of water pressure, flow rate, water quality, etc.) can be measured. In this embodiment, a case where normal / abnormal determination processing is performed using the measurement data of two types of determination factors (water pressure and flow rate) measured by the measurement sensor 200 will be described as an example.

[0036] <Abnormality detection device 100> The abnormality detection device 100 is a device capable of detecting abnormalities at each measurement point. Specifically, the abnormality detection device 100 generates a determination model that sets a certain pattern found in the cycle (period) as the normal region by machine learning using past measurement data corresponding to a predetermined cycle (period), such as each time zone and day of the week. Also, the abnormality detection device 100 executes determination processing using the determination model, determines that it is normal if the target data is within the normal region of the determination model, and determines that it is abnormal if it is outside the normal region, thereby detecting abnormalities at each measurement point.

[0037] Such an abnormality detection device 100 includes, as shown in FIG. 1, a processing unit 110, a storage unit 120, and a communication unit 130. Further, the processing unit 110 has, as individual functional units that execute processing related to abnormality detection, an input reception unit 111, a data acquisition unit 112, a determination model generation unit 113, a determination unit 114, and a display information generation unit 115.

[0038] The input reception unit 111 is a functional unit that receives instructions and information input from the user via a predetermined input device (for example, the input device included in the abnormality detection device 100 or the input device of a personal computer or smartphone on which the user performs an input operation). Specifically, the input reception unit 111 receives from the user a specification of a period (for example, the past 6 months, 12 months, or 18 months, etc.) for which measurement data is to be used to generate a determination model. In addition, the input reception unit 111 receives an input from the user regarding the period of the generated determination model. The input reception unit 111 receives, for example, an input of setting information from the user, such as which determination model to apply (adopt) from among determination models corresponding to mutually different periods, according to the time zone, day of the week, etc. for which the determination process is to be performed.

[0039] The data acquisition unit 112 is a functional unit that acquires the measurement data transmitted from the measurement sensor 200. In addition, the data acquisition unit 112 stores the acquired measurement data in a database (measurement value DB) within the storage unit 120.

[0040] The determination model generation unit 113 is a functional unit that generates a determination model based on machine learning using past measurement data stored in a database (measurement value DB) within the storage unit 120. Specifically, the determination model generation unit 113 extracts, from the measurement data for a predetermined past period (for example, the past 12 months, the past 18 months, etc.) specified by the user, the measurement data corresponding to a predetermined period from the database, and generates a determination model for the corresponding period using an unsupervised machine learning model such as One Class SVM (Support Vector Machine).

[0041] Note that for a predetermined period, there are various types such as each time zone from 0:00 to 24:00 (by time zone), such as the 8 o'clock hour or the 12 o'clock hour, each day of the week from Monday to Sunday (by day of the week), each week from the 1st week to the 4th week (or the 5th week) (by week), each month from January to December (by month), and each season of spring, summer, autumn, and winter (by season). Also, the period includes, for example, a combination of multiple different types of periods such as 15:00 on Monday or Friday of the second week of May. Further, the period includes, for example, an arbitrary period (custom) specified by the user, such as 8:15 to 19:20 or Monday to Wednesday.

[0042] The determination model generation unit 113 generates a determination model 124 corresponding to the type or all types of periods selected by the user from among such periods and stores it in the storage unit 120 (determination model DB). Note that the determination model generation unit 113 generates a determination model for each measurement location. Specifically, when generating a determination model corresponding to a certain location, the determination model generation unit 113 narrows down the data extracted from the storage unit 120 (measurement value DB) based on the position information of the measurement data obtained from the measurement sensor 200 at that location, thereby generating a determination model for the corresponding measurement location.

[0043] Also, when the latest measurement data for the corresponding period is obtained from the corresponding measurement location for the generated determination model, the determination model generation unit 113 re-performs machine learning using the data, thereby performing update processing so that each determination model is always in the latest state.

[0044] The determination unit 114 is a functional unit that determines the normality / anomaly of each measurement location using the determination model. Specifically, the determination unit 114 inputs the target data into the determination model, and if the value of the target data is within the normal region of the determination model, it determines it as normal, and if it is outside the normal region, it determines it as abnormal. Note that the determination unit 114 selects the determination model to be applied to the determination process based on the setting information.

[0045] The display information generation unit 115 is a functional unit that generates display information (screen information). The display information generation unit 115 generates, for example, the display information of a dialog box for receiving input of setting information from the user. Also, the display information generation unit 115 generates, for example, the display information indicating the normal area shown by the determination model and the plot positions of the past measurement data and the target data. Further, the display information generation unit 115 outputs the generated display information to a predetermined display device (for example, the display provided in the abnormality detection device 100 or the display of a personal computer, smartphone, etc. on which the user performs an input operation).

[0046] Next, the storage unit 120 will be described. The storage unit 120 is a functional unit that stores various types of information. Specifically, the storage unit 120 has a measurement value DB 121, a determination model DB 123, and setting information 125.

[0047] The measurement value DB (database) 121 is a database that stores measurement data 122 (such as water pressure data and flow rate data). Note that abnormal data that can be removed as preprocessing of the data, such as measurement failures and extreme abnormal values at the start-up of the measurement sensor 200, will be removed from the measurement value DB 121 in advance. Also, it is assumed that the measurement data 122 when an abnormality is detected in the normal / abnormal determination process has been removed from the measurement value DB 121.

[0048] The determination model DB (database) 123 is a database that stores the determination model 124 generated by the determination model generation unit 113. Note that in the determination model DB 123, for example, the generated determination models 124 for each period are stored separately for each measurement location.

[0049] The setting information 125 is information regarding the selection of the determination model 124 applied to the determination process. In the setting information 125, for example, the time zone and day of the week when the determination process is executed (the first item) and the information specifying the period of the determination model 124 to be applied (the second item) are registered in association with each other. Note that, as described above, the information specifying the period of the determination model 124 includes those classified by time zone, day of the week, week, month, season, custom, and combinations of these periods.

[0050] For example, when the information “8:00 to 23:00 / Monday to Friday” and “classified by time zone” are registered in the setting information 125 for the first item and the second item respectively, the determination unit 114 executes a determination process in which the determination model 124 corresponding to the period classified by time zone is applied to the determination process executed from 8:00 to 23:00 on weekdays. Specifically, for example, in the determination process executed at around 8 o'clock, the determination model 124 corresponding to the period at around 8 o'clock (the determination model 124 generated based on the past measurement data 122 measured at around 8 o'clock) is selected, and for example, in the determination process executed at around 23 o'clock, the determination model 124 corresponding to the period at around 23 o'clock (the determination model 124 generated based on the past measurement data 122 measured at around 23 o'clock) is selected and applied to the determination process.

[0051] Also, for example, when the information “8:00 to 15:00 / Monday to Sunday” and “classified by time zone / classified by day of the week” are registered in the setting information 125 for the first item and the second item respectively, the determination unit 114 executes a determination process in which the determination model 124 corresponding to the period classified by time zone and by day of the week is applied to the determination process executed from 8:00 to 15:00 on each day of the week. Specifically, for example, in the determination process executed at around 8 o'clock on Monday, the determination model 124 corresponding to the period at around 8 o'clock on Monday (the determination model 124 generated based on the past measurement data 122 measured at around 8 o'clock on Monday) is selected, and for example, in the determination process executed at around 15 o'clock on Wednesday, the determination model 124 corresponding to the period at around 15 o'clock on Wednesday (the determination model 124 generated based on the past measurement data 122 measured at around 15 o'clock on Wednesday) is selected and applied to the determination process.

[0052] Note that these are just examples, and for each time slot of each day of the week, information for selecting any one of the determination models 124 stored in the determination model DB 123 is registered in the setting information 125.

[0053] <Abnormality detection process> FIG. 3 is a flowchart showing an example of the abnormality detection process executed by the abnormality detection device 100. The abnormality detection process is started, for example, when an execution instruction from the user is received via the input reception unit 111. Note that, hereinafter, the process for one measurement point will be described, but it is assumed that the abnormality detection device 100 executes the abnormality detection process for a plurality of measurement points substantially simultaneously.

[0054] When the process is started, the determination unit 114 selects a determination model 124 to be applied to the determination process using the setting information 125 (step S10). Specifically, the determination unit 114 selects the determination model 124 to be applied to the determination process by specifying the period of the determination model 124 associated with the first item including the current time and day of the week from the second item of the setting information 125.

[0055] Next, the determination unit 114 extracts the data to be determined from the measurement value DB 121 (step S20). Specifically, the determination unit 114 extracts the latest measurement data 122 at the target measurement point from the measurement value DB 121.

[0056] Next, the determination unit 114 inputs the data to be determined into the determination model 124 to specify the position (plot position) of the data to be determined with respect to the normal region (step S30).

[0057] Next, the determination unit 114 determines whether or not the data to be determined has deviated from the normal region (step S40). Specifically, the determination unit 114 determines whether or not the position of the data to be determined is outside the normal region based on the positional relationship between the specified position of the data to be determined and the normal region.

[0058] And when it is determined that there is a deviation (Yes in step S40), the determination unit 114 shifts the process to step S50. On the other hand, when it is determined that there is no deviation (No in step S40), the determination unit 114 returns the process to step S10, and selects the determination model 124 by specifying the period of the determination model 124 from the setting information 125 based on the time and day of the week at that time. As a result, since the determination model 124 with a period corresponding to the time zone and day of the week of the process is reselected based on the setting information 125, the abnormality detection device 100 can repeatedly execute the determination process based on the optimal determination model 124 associated therewith according to the occasion.

[0059] In step S50, the determination unit 114 detects an abnormality at the measurement point and shifts the process to step S60.

[0060] In step S60, the display information generation unit 115 generates display information including the normal region of the determination model 124 and the plot position of the target data, and the determination result, and causes it to be displayed on a predetermined display device.

[0061] FIG. 4 is a diagram showing an example of display information indicating a determination result. As shown in the figure, the display information of the determination result includes information 321 regarding the measurement point, the period 322 of the applied determination model 124, the normal region 300 of the determination model 124, the target data (plot position of the target data) 301, and the determination result (in this case, "An abnormality has been detected", etc.) 323. When the display information generation unit 115 generates display information of a determination result in which an abnormality has been detected, the display information may be output together with a predetermined alert sound (beep sound) from the speaker of the device used by the user.

[0062] Also, when the process of step S60 is executed, the display information generation unit 115 shifts the process to step S10, and in the abnormality detection device 100, the processes of steps S10 to S60 are repeatedly executed.

[0063] The above described the abnormality determination process. According to the abnormality detection device that executes such a process, it becomes possible to detect an abnormality in accordance with the actual usage situation at each measurement point of the water supply. In particular, the abnormality detection device makes a normal / abnormal determination based on a normal state (normal region) corresponding to a predetermined cycle such as the time zone or day of the week for performing the determination process. Therefore, it becomes possible to detect a tight abnormality in accordance with the actual usage situation, and the abnormality can be detected at an earlier stage.

[0064] Also, the abnormality detection device determines the normality / abnormality of each measurement point using a determination model based on a plurality of different types of determination elements (for example, water pressure and flow rate). As a result, since the correlation relationship established between the determination elements is taken into account in the determination of the abnormality detection, it is possible to determine the normality / abnormality more accurately compared to the case of using a determination model generated based on one determination element. Further, the abnormality detection device updates the determination model using the measurement data acquired from the measurement point, and selects an optimal determination model according to the timing of the process based on the setting information and executes the determination process. That is, since the abnormality detection device can continuously perform the determination process using the optimal determination model generated and updated by itself without requiring a user operation for the determination process, the user can save the labor for operation as much as possible.

[0065] Note that the abnormality detection device 100 may generate display information indicating the degree of normality / abnormality in addition to the display information indicating the determination result, and display it on a predetermined display device.

[0066] FIG. 5 is a diagram showing an example of display information indicating normality / abnormality. In the illustrated example, the vertical axis indicates normality / abnormality, and the horizontal axis indicates the measurement date and time of measurement data 122. Also, reference line 331 is a line that divides normal and abnormal. When the value of measurement data 122 is normal, that is, when it is within the normal region 300 of determination model 124, graph 332 is drawn so as to be located above reference line 331. When the value of measurement data 122 is abnormal, that is, when it is located outside the normal region 300 of determination model 124, graph 332 is drawn so as to be located below reference line 331. Also, graph 332 is drawn such that the greater the degree of normality or abnormality, the greater the deviation from reference line 331. Note that display information generation unit 115 calculates, for example, a center line of the normal region that connects the points within the normal region 300 that are farthest from the outer contour of normal region 300, and determines the magnitude of the deviation from reference line 331 according to the distance from the plotted position of measurement data 122 to the center line, and draws graph 332 of the display information. Also, the method of determining the magnitude of the deviation is not limited to this, and any distance calculated based on normal region 300 and the data to be determined may be used.

[0067] According to such display information indicating normality / abnormality, the degrees of normality and abnormality can be clearly shown. As a result, even when graph 332 is located on the abnormal side, the user can instantaneously determine the degree of abnormality based on the magnitude of the deviation from reference line 331. As a result, the user can use it for various countermeasures, such as changing the countermeasure method according to the degree (magnitude) of abnormality.

[0068] Note that the display information indicating normality / abnormality may be displayed, for example, on one screen together with the display information of the determination result shown in FIG. 4.

[0069] <Second Embodiment> In the above-described embodiment, the determination model 124 using two determination factors, i.e., water pressure and flow rate, is used to determine the normality / abnormality of the measurement point. However, the present invention is not limited thereto. For example, it is also possible to determine the normality / abnormality of the measurement point based on one determination factor (e.g., water pressure, flow rate, or other determination factors). Note that since the basic configuration of the abnormality detection system 1000 according to the present embodiment and the processes performed by each functional unit are the same as those of the first embodiment described above, detailed descriptions are omitted, and the following description will focus on the differences.

[0070] Specifically, the abnormality detection device 100 generates a determination model 124 by the same method as in the first embodiment using the measurement data 122 of water pressure, flow rate, or other determination factors (e.g., water quality, etc.). Further, the abnormality detection device 100 uses the generated determination model 124 to compare the position of the target data with respect to the normal region of the determination model 124, and in the same manner as in the first embodiment, determines the normality / abnormality of the measurement point where the target data is measured, thereby detecting an abnormality.

[0071] Also, the abnormality detection device 100 generates display information indicating the determination result and display information indicating the normality / abnormality degree in the same manner as in the first embodiment, and causes the information to be displayed on a predetermined display device.

[0072] FIG. 6 is a diagram showing an example of the normal region indicated by a determination model using one determination factor (water pressure). The illustrated example is the determination model 124 corresponding to a predetermined period (e.g., 8 o'clock on every Monday), where the vertical axis indicates the value of water pressure and the horizontal axis indicates time. As shown in the figure, the normal region 340 of the determination model 124 based on one determination factor is different in shape (range) from the normal region 300 of the determination model 124 based on two determination factors. However, even so, compared with the uniform threshold values (the illustrated threshold upper limit value and threshold lower limit value) indicating the upper and lower limits, it has a tight shape (area) suitable for the actual usage situation.

[0073] Thus, even when the abnormality detection device 100 uses the determination model 124 based on one determination factor, it can accurately determine the normality / abnormality of the measurement point.

[0074] Note that, similar to the first embodiment, the abnormality detection device 100 may generate display information indicating the degree of abnormality and display it on a predetermined display device. In addition to these display information, the abnormality detection device 100 can also generate display information in the form shown in FIG. 7.

[0075] FIG. 7 is a diagram showing an example of display information according to another form. The illustrated example shows a normal region 350 of a determination model 124 based on one determination element (in this case, water pressure) and a graph 351 showing the values of the measurement data 122. In the same figure, the vertical axis represents the water pressure, and the horizontal axis represents the measurement date and time of the measurement data 122. Also, in this form, the normal region 350 of the determination model 124 is drawn in a vertically long rectangular shape based on the values of the water pressure at each date and time (the width in the vertical axis direction of the normal region at that date and time). The graph 351 shows the values of the measurement data 122 of the water pressure at each date and time and is drawn superimposed on the corresponding position on the rectangular normal region 350.

[0076] Note that the display information generation unit 115 draws each vertically long rectangular normal region 350 having a size in the vertical axis direction determined based on the values of the water pressure at each date and time (the width in the vertical axis direction of the normal region at that date and time) included in the normal region 340 shown in FIG. 6. Also, the display information generation unit 115 specifies the position on the normal region 350 corresponding to the value of the measurement data 122 at each date and time and draws a graph 351 showing the value of the measurement data 122 at that position. Note that the illustrated upper threshold and lower threshold are lines indicating the upper and lower limits of the conventional normal region 310.

[0077] In the illustrated example, since the graph 351 showing the value of the measurement data 122 deviates from the normal region 350 between 10 / 5 and 10 / 6, an abnormality at the measurement point is detected at this time, and display information (for example, FIG. 4) indicating the determination result of the abnormality detection is displayed on a predetermined display device.

[0078] In such a display form, since the normal region 350 of the determination model 124 set for each date and time can be superimposed on the graph 351 showing the values of the measurement data 122, the daily transition of the measurement data 122 with respect to the normal region 350 can be clearly shown. As a result, for example, when the graph 351 changes day by day near the upper limit or the lower limit of the normal region 350, it is possible to detect the precursor before the occurrence of an abnormality and take measures such as going to the corresponding measurement point for inspection.

[0079] <Hardware Configuration> FIG. 8 is a diagram showing an example of the hardware configuration of the abnormality detection device 100. As shown in the figure, the abnormality detection device 100 includes an input device 410, a display device 420, a processor 430, a memory 440, a storage 450, and a communication device 460, and each component is connected by a bus 470. Also, the power supply (not shown) may be a primary battery or a rechargeable secondary battery, and can be replaced when it is consumed or deteriorated.

[0080] The input device 410 is an input device such as a keyboard, a mouse, or a touch panel. The display device 420 is a display device such as a liquid crystal display or an organic display.

[0081] The processor 430 is an arithmetic device such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit), and executes processing according to a program recorded in the memory 440 or the storage 450. In the abnormality detection device 100, processing is performed by the processor 430 operating according to a program read onto the memory 440 or the storage 450. The processing unit 110 realizes each function when the processor 430 executes the program.

[0082] The memory 440 is a storage device such as a RAM (Random Access Memory) or a flash memory, and functions as a storage area where programs and data are temporarily read out. The storage 450 is a rewritable storage device. The storage unit 120 has its functions realized by the memory 440 or the storage 450. Note that the functions of the storage unit 120 may also be realized by an external storage device connected via the communication device 460.

[0083] The communication device 460 is an interface for communicatively connecting the abnormality detection device 100 to an external device. Note that the communication device 460 may originally output the content to be output to the display device 420 to the external device based on communication with the external device.

[0084] Note that the processing of each component of the abnormality detection device 100 may be executed by one piece of hardware or by a plurality of pieces of hardware. Also, the processing of each component of the abnormality detection device 100 may be realized by one program or by a plurality of programs.

[0085] Also, the examples of the above-described embodiments have been described in detail for easy understanding of the present invention, and the present invention is not limited to those having all the configurations described herein. Also, it is possible to replace a part of the configuration of an example of one embodiment with the configuration of an example of another. Also, it is possible to add the configuration of an example of another embodiment to the configuration of an example of one embodiment. Also, it is possible to add, delete, or replace a part of the configuration of an example of each embodiment with another configuration. Also, the above-described respective configurations, functions, processing units, processing means, etc. may be realized in hardware by designing a part or all of them, for example, by an integrated circuit. Also, the control lines and information lines in the drawings show those considered necessary for explanation, and do not necessarily show all of them. It may be considered that almost all the configurations are interconnected.

Explanation of Signs

[0086] 1000... Abnormality detection system, 100... Abnormality detection device, 110... Processing unit, 111... Input reception unit, 112... Data acquisition unit, 113... Determination model generation unit, 114... Determination unit, 115... Display information generation unit, 120... Memory unit, 121... Measurement value DB, 122... Measurement data, 123... Determination model DB, 124... Determination model, 125... Setting information, 130... Communication unit, 200... Measurement sensor, 410... Input device, 420... Display device, 430... Processor, 440... Memory, 450... Storage, 460... Communication device, 470... Bus, N... Network

Claims

1. An abnormality detection system having an abnormality detection device and a measurement sensor, wherein the measurement sensor transmits measurement data obtained by measuring the value of a predetermined determination element at a measurement point of a water supply pipe to the abnormality detection device, and the abnormality detection device includes: a storage unit that stores the measurement data; a determination model generation unit that generates a determination model indicating a normal region observed in a period using the measurement data corresponding to a predetermined period among the measurement data in the storage unit; and a determination unit that determines whether the value indicated by the target data for determination is normal or abnormal using the determination model. An abnormality detection system characterized by the above.

2. The abnormality detection system according to claim 1, wherein the abnormality detection device generates the determination model based on unsupervised machine learning using the measurement data stored in the storage unit. An abnormality detection system characterized by the above.

3. The abnormality detection system according to claim 1, wherein the determination element includes a first determination element and a second determination element of different types, and the first determination element and the second determination element are any one of water pressure, flow rate, and water quality. An abnormality detection system characterized by the above.

4. The abnormality detection system according to claim 1, wherein the determination element includes one type of determination element and is any one of water pressure, flow rate, and water quality. An abnormality detection system characterized by the above.

5. The abnormality detection system according to claim 1, wherein the period includes at least any one of time-of-day, day-of-week, week, month, season, or a combination of these periods, and an arbitrary period specified by the user. An abnormality detection system characterized by the above.

6. The abnormality detection system according to any one of claims 1 to 4, further comprising a display information generation unit that generates display information including the positional relationship between the normal region and the plotted position of the target data and the result of the determination, wherein the display information generation unit displays the display information on a predetermined display device. An abnormality detection system characterized by the above.

7. The abnormality detection system according to any one of claims 1 to 4, further comprising a display information generation unit that generates display information displaying a graph indicating the degree of normality or abnormality of the value of the target data according to the distance from the normal region to the plotted position of the target data, wherein the display information generation unit displays the display information on a predetermined display device. ​ An abnormality detection system characterized by the following.

8. The abnormality detection system according to claim 1, wherein the storage unit stores setting information in which the time zone and day of the week when the determination is executed are associated with information specifying the period of the determination model applied to the determination and registered. The determination unit selects the determination model at the time of executing the determination based on the setting information. An abnormality detection system characterized by the following.

9. The abnormality detection system according to claim 4, further comprising a display information generation unit that generates display information in which the normal region of the determination model set for each date and time when the determination is made is superimposed on a graph showing the value of the measurement data. An abnormality detection system characterized by the following.

10. A storage unit that stores measurement data obtained by measuring the value of a predetermined determination element at a measurement point of a water supply system, a determination model generation unit that generates a determination model showing a normal region found in a period using the measurement data corresponding to a predetermined period among the measurement data in the storage unit, and a determination unit that determines whether the value indicated by the target data for determination is normal or abnormal using the determination model. An abnormality detection device characterized by the following.

11. An abnormality detection method performed by an abnormality detection device, wherein the abnormality detection device performs a storage step of storing measurement data obtained by measuring the value of a predetermined determination element at a measurement point of a water supply system, a determination model generation step of generating a determination model showing a normal region found in a period using the measurement data corresponding to a predetermined period among the measurement data stored in the storage step, and a determination step of determining whether the value indicated by the target data for determination is normal or abnormal using the determination model. An abnormality detection method characterized by the following.

12. A program for causing a computer to function as an abnormality detection device, wherein the computer is caused to function as a storage unit that stores measurement data obtained by measuring the value of a predetermined determination element at a measurement point of a water supply system, a determination model generation unit that generates a determination model showing a normal region found in a period using the measurement data corresponding to a predetermined period among the measurement data in the storage unit, and a determination unit that determines whether the value indicated by the target data for determination is normal or abnormal using the determination model. A program characterized by the following.

Images