Method for optimizing arrangement of water level gauges, method for predicting water level in manhole, and system for predicting water level in manhole

Abstract

[Problem] To more efficiently predict the water level in a manhole. [Solution] Provided is a method for optimizing the arrangement of water level gauges, the method including: a data acquisition step for acquiring water level data items acquired by respective water level gauges provided in a plurality of manholes in a predetermined area; a classification step for classifying the plurality of manholes into one or more clusters by clustering the water level data items; and a determination step for determining one or more representative manholes for each cluster from the plurality of manholes included in the one or more clusters.

Description

Method for optimizing the placement of water level gauges, method for predicting water levels in manholes, and system for predicting water levels in manholes

[0001] The present disclosure relates to a method for optimizing the placement of water level gauges, a method for predicting water levels in manholes, and a system for predicting water levels in manholes.

[0002] In recent years, due to increased rainfall caused by torrential rain and thunderstorms, there have been a series of incidents in which water has overflowed from manholes, causing flooding damage. Therefore, in areas where flooding is expected, it is necessary to measure fluctuations (especially increases) in water levels in manholes in real time and predict flooding. For example, Patent Literature 1 discloses a technology for predicting the water level and inflow volume at pumping stations using information such as the water level in sewer mains, the water level at pumping stations, and rainfall.

[0003] Japanese Patent Application Laid-Open No. 2002-285634

[0004] However, the technology described in Patent Document 1 is only capable of predicting the amount of inflow to a pumping station, and is premised on using information on the water level in sewer pipes. Therefore, it is not possible to predict overflow onto the ground. Furthermore, to predict overflow onto the ground, it is necessary to monitor the water level in manholes, which are spaces connecting sewer pipes to the ground. However, since there are many manholes, monitoring the water level in all of the manholes and predicting water level fluctuations would require installing many water level gauges, which is difficult to achieve.

[0005] Therefore, the present disclosure has been made in consideration of the above problems, and an object of the present disclosure is to provide a new and improved method for optimizing the placement of water level gauges, a method for predicting the water level in a manhole, and a system for predicting the water level in a manhole, which are capable of more efficiently predicting the water level in a manhole.

[0006] In order to solve the above problem, according to one aspect of the present disclosure, there is provided a method for optimizing the placement of water level gauges, the method including: a data acquisition step of acquiring each piece of water level data acquired by each water level gauge provided in a plurality of manholes set up in a specified area; a classification step of classifying the plurality of manholes into one or more clusters by clustering each piece of water level data; and a determination step of determining one or more representative manholes for each cluster from the plurality of manholes included in the one or more clusters.

[0007] Furthermore, in order to solve the above problem, according to another aspect of the present disclosure, there is provided a method for predicting the water level of a manhole, the method including: a past data acquisition step of acquiring past time-series water level data of a manhole located in a specified area and predicted rainfall data for the specified area; and an output step of inputting the acquired past time-series water level data of the manhole and predicted rainfall data into a specified engine, thereby outputting predicted information on the water level of the manhole at a future point in time.

[0008] Furthermore, in order to solve the above problem, according to another aspect of the present disclosure, there is provided a manhole water level prediction system, which includes a past data acquisition unit that acquires past time-series water level data of a manhole included in a target area and predicted rainfall data for the specified area, and an output unit that outputs predicted information for the water level of the manhole at a future point in time by inputting the acquired past time-series water level data of the manhole and the predicted rainfall data into a specified engine.

[0009] As described above, according to the present disclosure, it is possible to more efficiently predict the water level in a manhole.

[0010] FIG. 1 is an explanatory diagram for explaining an overview of an information processing system according to the present embodiment. FIG. 2 is a block diagram showing an example of the functional configuration of a server 10 according to a first embodiment of the present disclosure. FIG. 3 is a diagram showing an example of clustering processing of manholes 4 by a classification unit 102 according to the present embodiment. FIG. 4 is a diagram showing an example of a result of clustering processing by a classification unit 102 according to the present embodiment. FIG. 5 is a diagram showing an example of a result of clustering processing by a classification unit 102 according to the present embodiment. FIG. 6 is a flowchart showing an example of the flow of optimization processing for the placement of water level gauges by a server 10 according to the present embodiment. FIG. 7 is a block diagram showing an example of the functional configuration of a server 10 according to a second embodiment of the present disclosure. FIG. 8 is a diagram showing an overview of processing for calculating a predicted value of water level by a learning engine 112 according to the present embodiment. FIG. 9 is a flowchart showing an example of the flow of processing for predicting water level in a manhole by a server 10 according to the present embodiment. FIG. 10 is a graph for explaining an example of the present embodiment. FIG. 11 is an explanatory diagram for explaining an example of the hardware configuration of a server 10 according to the present embodiment.

[0011] Preferred embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings. In this specification and drawings, components having substantially the same functional configurations are designated by the same reference numerals, and redundant description will be omitted.

[0012] In addition, in this specification and drawings, the manholes will be distinguished as necessary, such as referring to a representative manhole described below as a representative manhole 4A and other manholes as manholes 4B. However, when there is no need to particularly distinguish between the manholes, the manholes will be simply referred to as manholes 4.

[0013] One embodiment of the present disclosure relates to an information processing system capable of efficiently predicting manhole water levels. Such a system can be used, for example, to predict manhole water levels and predict flooding damage caused by manhole overflows to the ground. Conventional sewerage stormwater management systems combine meteorological conditions in a monitored area, a short-term rainfall forecast model, and a stormwater runoff analysis model for the monitored area. However, such systems cannot guarantee the accuracy of the runoff analysis model due to changes in topography, urban development, construction, and other factors. Furthermore, while such systems are only suitable for static analysis such as hazard maps, they have difficulty predicting flooding in real time. To achieve real-time monitoring, a technology for sequentially predicting manhole water levels is required.

[0014] However, conventional systems for measuring water levels in sewer pipes in real time, such as smart manholes, are extremely expensive. Therefore, installing such systems in all of the numerous manholes is impractical. Therefore, a water level gauge placement optimization method according to an embodiment of the present disclosure uses a clustering technique to optimize which manholes in a given area should have water level gauges. A cluster refers to a cluster in which the temporal changes in water levels at the manholes included in that cluster are similar. Therefore, the water levels of other manholes within the same cluster can be estimated using simple calculations. This makes it possible to cover anticipated flooding areas with a minimum number of water level gauges (which are expensive due to their wireless transmission capabilities). Furthermore, by inputting time-series water level data obtained from water level gauges whose placement has been optimized using this method and short-term rainfall forecast information into a pre-trained water level prediction engine, predicted manhole water levels can be output. This enables highly accurate flooding predictions for anticipated flooding areas using a limited number of water level gauges.

[0015] Fig. 1 is an explanatory diagram for explaining an overview of an information processing system according to this embodiment. The information processing system 1 according to this embodiment includes a server 10, as shown in Fig. 1. The information processing system 1 according to this embodiment can be used via, for example, another server 20 or an information terminal 30.

[0016] The server 10, the other server 20, and the information terminal 30 are connected to one another via a network.

[0017] The server 10 according to this embodiment is a server for predicting the water level of each manhole 4 in a water level prediction target area 2 that includes multiple manholes 4. Specifically, the server 10 acquires past water level data and the like for some of the multiple manholes 4 (representative manhole 4A in the example shown in FIG. 1 ) provided in sewer pipes 3 in the water level prediction target area 2, inputs the acquired data into a learning engine, and outputs prediction information regarding the water level of the manholes 4 output from the learning engine. The server 10 outputs the prediction information to, for example, another server 20 or an information terminal 30. Note that the type of sewer pipe 3 is not particularly limited, and the sewer pipe 3 may be a storm water pipe, a sewage pipe, or a combined sewer pipe.

[0018] The server 10 may be hardware such as a computer having the hardware configuration described below. The hardware that realizes the server 10 and the manner in which the hardware is provided (for example, whether it is a stationary server or a processor in an edge device) are not particularly limited, and the server 10 may be realized by one or more computers or servers. The same applies to hardware that realizes the other servers 20. Furthermore, the hardware that realizes the information terminal 30 may be a computer or server, or may be realized by any terminal such as a smartphone, mobile phone, tablet, laptop, or game console.

[0019] The other server 20 and the information terminal 30 are examples of information processing devices used to acquire and utilize, for example, prediction information regarding the water level in a manhole. The other server 20 and the information terminal 30 may be owned by, for example, a national or local government agency with authority over sewerage, or may be used by general users. Such prediction information may be used, for example, to take measures to prevent flooding in the water level prediction target area 2 or to notify residents and others of the possibility of flooding.

[0020] As will be described later, the server 10 may have, as another function or in addition, a function for determining which manholes 4 included in the water level prediction target area 2 should have water level gauges permanently installed. When many manholes 4 are installed in the water level prediction target area 2, it may be difficult to install water level gauges in all of them due to cost or equipment constraints. On the other hand, if water level gauges are not installed in appropriate manholes 4, it may be difficult to predict the water level in the manholes 4. Therefore, the server 10 according to one embodiment can optimize which of the limited number of manholes 4 installed in the water level prediction target area 2 should have water level gauges installed. This enables more accurate prediction of the water level in the manholes 4 within the constraints related to the installation of water level gauges.

[0021] In the following, a first embodiment will be described as an example of a method for optimizing the placement of water level gauges by the server 10. In addition, a second embodiment will be described as an example of a method for predicting the water level in a manhole by the server 10. These embodiments may be realized by the same server 10 or by different servers.

[0022] 2 is a block diagram showing an example of the functional configuration of the server 10 according to the first embodiment of the present disclosure. As shown in FIG. 2, the server 10 according to the present embodiment includes a data acquisition unit 101, a classification unit 102, a determination unit 103, and a water level / rainfall DB (database) 104.

[0023] The data acquisition unit 101 has a function of acquiring each piece of water level data acquired by each water level gauge installed in a plurality of manholes installed in a predetermined area. For example, the data acquisition unit 101 can acquire water level data accumulated in the past from the water level / rainfall DB 104 or an external server. The water level data referred to here is time-series water level data. Time-series water level data is water level data acquired continuously or intermittently over several months, for example.

[0024] The water level data can be appropriately acquired from a water level meter installed in a manhole 4 located within the water level prediction target area 2. The type of water level meter is not particularly limited in terms of the installation manner, as long as it can continuously or intermittently acquire the water level in the manhole.

[0025] The data acquisition unit 101 may also acquire precipitation data relating to the manhole. The precipitation data relating to the manhole may be, for example, data on past precipitation provided by an organization such as the Japan Meteorological Agency in the area to which the water level prediction target area 2 in which the manhole 4 is located belongs, or may be precipitation data based on a water gauge installed in the water level prediction target area 2.

[0026] The classification unit 102 has the function of classifying multiple manholes into one or more clusters by performing clustering on each piece of water level data. The classification unit 102 classifies multiple manholes 4 into one or more clusters using a method such as hierarchical clustering. Here, the classification unit 102 determines one or more representative manholes for each cluster based on the similarity determined by hierarchical clustering. Similarity, for example, refers to the similarity of the temporal changes in water level. In other words, if there are manholes whose time-series changes in water level are similar to a certain extent, the clustering process determines that the similarity is high. The threshold value can be determined appropriately in the clustering process. The clustering method is not particularly limited, but a method that can evaluate the distance within or outside a cluster is preferable.

[0027] 3 is a diagram showing an example of clustering processing of manholes 4 by the classification unit 102 according to this embodiment. The classification unit 102 performs clustering processing using a technique such as structured clustering for each of the water level data of manholes 4 acquired by the data acquisition unit 101. As a result, for example, the water level prediction target area 2 can be classified into a plurality of clusters 5 (four clusters 5 in the example shown in FIG. 3 ).

[0028] 4 and 5 are diagrams showing an example of the results of the clustering process performed by the classification unit 102 according to this embodiment. For example, if there are 15 manholes in the water level prediction target area 2, the results of FIGS. 4 and 5 show that manholes No. 4, No. 5, and No. 13 form a cluster independent of the other manholes. Furthermore, as shown in FIG. 5, the other manholes can also be classified into multiple clusters. For example, the clusters other than the above three manholes can be classified into Cluster A, which contains only manhole No. 9. Furthermore, within Cluster A, the cluster containing only manhole No. 6 can be classified into Cluster A1 and Cluster A2. The granularity of the clusters can be set as appropriate, and for example, as described below, the granularity of the clusters can be determined based on the upper limit of the number of water level gauges that can be installed.

[0029] The determination unit 103 has a function of determining one or more representative manholes for each cluster from among the manholes included in the cluster. For example, as shown in Figure 3, the determination unit 103 determines one or more representative manholes 4A for each cluster 5 based on the clustering results obtained by the classification unit 102.

[0030] For example, as shown in Figures 4 and 5, manholes 4, 5, and 13 may be determined as representative manholes 4A where water level gauges should be installed. Because manholes 4, 5, and 13 are in clusters independent of the other manholes, it is difficult to estimate the water levels of these manholes from the results of the other manholes. Therefore, these manholes are determined as representative manholes 4A to optimize the placement of water level gauges.

[0031] Furthermore, when an upper limit on the number of water level gauges to be placed is set, the determination unit 103 may determine a representative manhole based on the clustering results. For example, if the upper limit on the number of water level gauges to be placed is five, as shown in Fig. 5, the determination unit 103 may determine, as representative manholes, manhole number 9, which is the most isolated from other clusters when viewed from the top of the structured tree, and manhole number 8, which most closely encompasses adjacent manholes when viewed from the bottom of the structured tree. This allows for the most effective placement that reflects the clustering results.

[0032] Similarly, if the upper limit of the number of water level gauges to be placed is seven, in addition to the five manholes mentioned above, manhole No. 1 included in Cluster A1 and manhole No. 6 included in Cluster A2 may be determined as representative manholes. Manhole No. 1 and manhole No. 6 are the most isolated from the other manholes, and therefore are preferably selected to further improve accuracy.

[0033] Next, the optimization process for water level gauge placement by the server 10 according to this embodiment will be described. FIG. 6 is a flowchart showing an example of the flow of the optimization process for water level gauge placement by the server 10 according to this embodiment. First, in step S101, the data acquisition unit 101 of the server 10 acquires past water level data, etc. for multiple manholes. Then, in step S103, the classification unit 102 of the server 10 performs a clustering process on this water level data, etc., to classify the multiple manholes into one or more clusters. Then, in step S105, the determination unit 103 of the server 10 determines a representative manhole for each of the clustered clusters.

[0034] The first embodiment of the present disclosure has been described above. According to this embodiment, even if water level gauges are not installed in all manholes provided in the water level prediction target area 2, by installing a water level gauge in at least one manhole included in each cluster as a representative manhole, sufficient data can be obtained for the manhole water level prediction described below. This makes it possible to effectively monitor and predict water levels even when there are constraints such as cost and installation. Furthermore, by using precipitation (also referred to as rainfall) data in the clustering process, it is possible to associate changes in water level with precipitation, making it possible to perform optimization that reflects the effects of precipitation.

[0035] Next, a second embodiment of the present disclosure will be described. Fig. 7 is a block diagram showing an example of the functional configuration of a server 10 according to the second embodiment of the present disclosure. As shown in Fig. 7, the server 10 according to this embodiment includes a data acquisition unit 111, a learning engine 112, an output unit 113, and a water level / rainfall DB (database) 114.

[0036] The data acquisition unit 111 has a function of acquiring past time-series water level data for a manhole located in a predetermined area (water level prediction target area 2) and predicted rainfall data for the predetermined area. For example, the data acquisition unit 111 may acquire past time-series water level data for a manhole and precipitation data for the manhole as data necessary for training the learning engine 112. The precipitation data for the manhole may be the same as the precipitation data described in the first embodiment. Furthermore, as described below, the data acquisition unit 111 may acquire past time-series water level data for a manhole adjacent to the predetermined manhole. The predicted rainfall data for the predetermined area may be data acquired, for example, from the Japan Meteorological Agency or may be data appropriately predicted using a rain gauge or the like installed in the predetermined area. The time-series water level data and rainfall data may be data acquired, for example, every few minutes. These data may be stored in the water level / rainfall DB 114 or may be acquired externally as appropriate.

[0037] The learning engine 112 is an engine that has a learning model that uses the acquired current water level data of the one manhole, the water levels of nearby manholes, and future predicted rainfall data as input values ​​and calculates predicted values ​​of future water level data as output values. Below, we will explain the learning model of the learning engine 112 separately for training and for predictive use.

[0038] The data acquired during training of the learning model of the learning engine 112 may include not only past time-series water level data and past precipitation data, but also, for example, past time-series water level data of manholes adjacent to a given manhole. A manhole adjacent to a given manhole may be, for example, a manhole located at least either upstream or downstream of the given manhole in the sewer pipe 3. These data may be summarized into a variable such as a vector as a composite explanatory variable.

[0039] Then, using the actual manhole's past water level data, the adjacent manhole's past water level data, the past cumulative precipitation data, and the resulting manhole's water level at a future time point as training data, a learning model can be generated by training using, for example, nonlinear multiple regression analysis such as support vector regression or machine learning such as a neural network. Note that such a learning model may be updated as appropriate even during prediction. For example, such training may utilize water level data and precipitation data acquired by the water level gauges in each manhole when optimizing the placement of the water level gauges in the above embodiment.

[0040] A learning engine (Trained Water Level Prediction Engine) based on a learning model trained on these data is used to calculate a predicted value of the water level in the manhole r minutes from the present time (r is any positive number). Figure 8 is a diagram showing an overview of the process of calculating a predicted value of the water level by the learning engine 112 according to this embodiment. As shown in Figure 8, the learning engine 112 uses past water level data l of the manhole to be predicted. Ｐ，ｔ and past water level data for nearby manholes. Ｐｒｅｆ，ｔand the predicted precipitation data after r minutes x ｔ＋ｒ Using this, the predicted value of the manhole water level after r minutes l Ｐ，ｔ，ｒ The learning engine 112 may, but need not, use past precipitation data as input values. Such predicted values ​​of the manhole water level may be calculated, for example, every few minutes. Furthermore, the past water level data of nearby manholes may be obtained directly from a water level meter, or may be estimated based on the water level meter of the representative manhole determined using the technology of the first embodiment and the correlation between the water levels of nearby manholes and the representative manhole obtained during the clustering process. Even in the latter case, the accuracy of the prediction can be improved because it reflects time-series data.

[0041] The learning engine 112 may also use past water level data for a manhole to calculate a predicted value for the water level of another manhole that is different from the first manhole. In this case, for example, the other manhole may be a manhole adjacent to the first manhole. By using the above-mentioned correlation and the predicted value for the water level of a representative manhole where a water level meter is installed, it is possible to predict the water level of a nearby manhole that does not have a water level meter.

[0042] The output unit 113 has a function of outputting predicted information of the water level of a manhole at a future time point calculated by the learning engine 112. That is, the output unit 113 can output the predicted value of the water level of the manhole described above and information based on the predicted value (e.g., the fact that the water level in the manhole has reached a dangerous level, the magnitude of the change in the water level, etc.). This output predicted information can be transmitted to the server 20, the information terminal 30, etc., for example.

[0043] Next, a description will be given of the process of predicting the water level of a manhole performed by the server 10 according to this embodiment. FIG. 9 is a flowchart showing an example of the process of predicting the water level of a manhole performed by the server 10 according to this embodiment. First, in step S111, the data acquisition unit 111 of the server 10 acquires past water level data and predicted rainfall data for multiple manholes. Next, in step S113, the learning engine 112 of the server 10 inputs the acquired data and calculates predicted water level data for the manhole in the future. Then, in step S115, the output unit 113 of the server 10 outputs predicted information for the water level of the manhole based on the calculated predicted water level data.

[0044] The second embodiment of the present disclosure has been described above. According to this embodiment, when predicting the water level of manholes located in the water level prediction target area 2, it is possible to predict the water level of manholes included in the water level prediction target area 2 simply by using a learning model that uses manhole water level data and precipitation data, without using topographical information about the water level prediction target area 2 or model information related to sewer pipes. Furthermore, even if not all manholes included in the water level prediction target area 2 are equipped with water level gauges, if there is water level data for a representative manhole identified using, for example, the technology described in the first embodiment, it is possible to predict the water levels of not only the representative manhole but also nearby manholes.

[0045] Although the above embodiment predicts the water level in a manhole using predicted rainfall data, the present technology is not limited to such an example. For example, the learning model of the learning engine 112 can be configured without using predicted rainfall data, and the learning engine 112 can calculate a predicted value for the water level in a manhole at a future time point using past manhole water level data. This makes it possible to predict the water level when rainfall does not need to be taken into consideration. Furthermore, in the above embodiment, the water level in a manhole is predicted using a water level gauge whose placement is optimized using the technology according to the first embodiment, but the present technology is not limited to such an example. For example, the technology according to this embodiment can of course be applied even when all manholes are equipped with water level gauges.

[0046] Furthermore, the predicted precipitation data used to predict water levels may be data obtained from organizations such as the Japan Meteorological Agency, but it may also be processed data like past predicted precipitation data. Generally, predicted precipitation data obtained from the Japan Meteorological Agency and other weather data providers is provided at intervals (e.g., every 30 minutes), making it difficult to predict at shorter intervals. Therefore, for example, precipitation data measured after the time corresponding to the predicted precipitation data may be used as the predicted precipitation data. For example, if predicted precipitation data for 20 minutes later (i.e., 3:20 PM) is provided at 3:00 PM, and precipitation data is actually measured at 3:20 PM, the precipitation data measured at 3:20 PM may be used as the predicted precipitation data for 3:20 PM. This enables more accurate water level predictions. Furthermore, the predicted precipitation data actually input to the learning engine 112 may be a net value, or may be data with a positive or negative margin added to the predicted precipitation. For example, if the predicted precipitation is 30 mm, 45 mm (50% more) and 15 mm (50% less) may be used as input values. This allows for water level prediction information that takes volatility into account.

[0047] FIG. 10 is a graph illustrating an example of this embodiment. The solid line in FIG. 10 indicates the actual water level, the dashed line indicates the predicted water level of the manhole obtained by inputting data into the learning engine 112 that is 50% larger than the predicted rainfall data described above, and the dashed line indicates the predicted water level of the manhole obtained by inputting data into the learning engine 112 that is 50% smaller than the predicted rainfall data described above. As shown in this graph, the predicted rainfall data, the manhole water level data, and the water level data of the manhole adjacent to the manhole are acquired at five-minute intervals, and the predicted water level and the actual measured water level are output at five-minute intervals. Furthermore, as described above, the predicted rainfall data is appropriately replaced with actual rainfall data. As shown in the graph in FIG. 10, there is little discrepancy between the predicted water level and the actual measured water level, and they are nearly identical. In this way, by using the learning engine described above to calculate the manhole's past water level data and predicted rainfall data, it is possible to predict the manhole's water level with high accuracy.

[0048] The information processing described above is realized by cooperation between software and the hardware of the server 10, which will be described below. The hardware configuration described below can also be applied to the server 20 and the information terminal 30.

[0049] 11 is an explanatory diagram illustrating an example of the hardware configuration of the server 10 according to this embodiment. As shown in FIG. 10, the server 10 includes a control unit 11, a memory 12, a storage 13, a transmission / reception unit 14, an input / output unit 15, etc., which are electrically connected to one another via a bus 16.

[0050] The control unit 11 is a computing device that controls the overall operation of the server 10, controls the transmission and reception of data between each element, and performs information processing necessary for application execution and authentication processing. For example, the control unit 11 is a CPU (Central Processing Unit) and / or a GPU (Graphics Processing Unit), and executes programs for this system stored in the storage 13 and deployed in the memory 12 to perform various information processing as described above.

[0051] The memory 12 includes a main memory configured with a volatile storage device such as a DRAM (Dynamic Random Access Memory) and an auxiliary memory configured with a non-volatile storage device such as a flash memory or an HDD (Hard Disc Drive). The memory 12 is used as a work area for the control unit 11, and also stores a BIOS (Basic Input / Output System) that is executed when the server 10 is started, various setting information, etc.

[0052] The storage 13 stores various programs such as application programs, etc. A database that stores data used for each process may be constructed in the storage 13.

[0053] The transceiver 14 connects the server 10 to a network and / or a blockchain network. The transceiver 14 may include a short-range communication interface such as Bluetooth (registered trademark) and BLE (Bluetooth Low Energy).

[0054] The input / output unit 15 includes information input devices such as a keyboard, a microphone, and a mouse, and output devices such as a display and a speaker.

[0055] The bus 16 is commonly connected to the above elements and transmits, for example, address signals, data signals and various control signals.

[0056] The information processing performed by an external device such as the information processing device 10 may be performed on-premise or in a cloud format. As an external device in a cloud format, the above-described functions and processes may be provided in the form of, for example, SaaS (Software as a Service) or cloud computing.

[0057] In the above embodiment, the information processing device 10 performs various storage and control operations, but multiple external devices may be used instead of the information processing device 10. That is, various pieces of information and programs may be distributed and stored in multiple external devices using blockchain technology or the like.

[0058] The above embodiment is not limited to the information processing system 1, and may be an information processing method or an information processing program. The information processing method includes each step of the information processing system 1. The information processing program causes at least one computer to execute each step of the information processing system 1.

[0059] Finally, while various embodiments of the present disclosure have been described, they are presented as examples and are not intended to limit the scope of the invention. The novel embodiments may be embodied in various other forms, and various omissions, substitutions, and modifications may be made without departing from the spirit of the invention. Such embodiments and modifications are intended to be included within the scope and spirit of the invention, as well as within the scope of the inventions and their equivalents as defined in the claims.

[0060] The following aspects are also included in the scope of the present disclosure. (Item 1) A method for optimizing the placement of water level gauges, comprising: a data acquisition step of acquiring water level data obtained by each of the water level gauges provided in a plurality of manholes provided in a predetermined area; a classification step of classifying the plurality of manholes into one or more clusters by clustering the water level data; and a determination step of determining one or more representative manholes for each cluster from the plurality of manholes included in the one or more clusters. (Item 2) A method for optimizing the placement of water level gauges according to item 1, wherein, in the determination step, the one or more representative manholes are determined for each cluster based on the results of the clustering. (Item 3) A method for optimizing the placement of water level gauges according to item 2, wherein the clustering used in the classification step is hierarchical clustering, and in the determination step, the one or more representative manholes are determined for each cluster based on the proximity obtained by the hierarchical clustering. (Item 4) A method for optimizing the placement of water level gauges according to item 2 or 3, wherein in the determining step, if an upper limit on the number of water level gauges to be placed is set, the representative manhole is determined based on the results of the clustering. (Item 5) A method for optimizing the placement of water level gauges according to any one of items 1 to 4, wherein in the data obtaining step, precipitation data related to the manhole is further obtained, and in the classifying step, clustering is performed using the precipitation data and each of the water level data. (Item 6) A method for optimizing the placement of water level gauges according to any one of items 1 to 5, wherein the water level data is time-series water level data.(Item 7) A method for predicting water levels in a manhole, comprising: a past data acquisition step of acquiring past time-series water level data for a manhole located in a predetermined area and predicted rainfall data for the predetermined area, and an output step of outputting predicted information for the water level of the one manhole at a future point in time by inputting the acquired past time-series water level data for the one manhole and the predicted rainfall data into a predetermined engine. (Item 8) A method for predicting water levels in a manhole according to Item 7, wherein in the past data acquisition step, past time-series water level data for another manhole close to the one manhole is further acquired, and in the output step, the past time-series water level data of the other manhole is input into the predetermined engine. (Item 9) The method for predicting the water level in a manhole according to item 7 or 8, wherein the past time-series water level data includes estimated water level data estimated based on water level data acquired by a water level gauge installed in a manhole different from the one manhole. (Item 10) The method for predicting the water level in a manhole according to any one of items 7 to 9, wherein the predetermined engine is an engine based on a learning model obtained by learning using at least the past time-series water level data of the one manhole and past rainfall data related to the target area. (Item 11) The method for predicting the water level in a manhole according to item 10, wherein the predetermined engine is obtained by further learning using past time-series water level data of other manholes adjacent to the one manhole. (Item 12) A method for predicting a water level in a manhole according to any one of items 7 to 11, wherein in the output step, predicted information on the water levels of other manholes included in the target area at a future time point is output together with or instead of the one manhole.(Item 13) A manhole water level prediction system comprising: a past data acquisition unit that acquires past time-series water level data of a manhole included in a target area and predicted rainfall data for the specified area; and an output unit that outputs predicted information for the water level of the manhole at a future point in time by inputting the acquired past time-series water level data of the manhole and the predicted rainfall data into a specified engine.

[0061] 10 Server 20 Server 30 Information terminal 101 Data acquisition unit 102 Classification unit 103 Decision unit 111 Data acquisition unit 112 Learning engine 113 Output unit

Claims

1. A method for optimizing the placement of water level gauges, comprising: a data acquisition step of acquiring water level data obtained by each of the water level gauges provided in a plurality of manholes provided in a specified area; a classification step of classifying the plurality of manholes into one or more clusters by clustering the respective water level data; and a determination step of determining one or more representative manholes for each cluster from the plurality of manholes included in the one or more clusters.

2. A method for optimizing the placement of water level gauges as described in claim 1, wherein in the determination step, the one or more representative manholes are determined for each cluster based on the results of the clustering.

3. A method for optimizing the placement of water level gauges as described in claim 2, wherein the clustering used in the classification step is hierarchical clustering, and in the determination step, the one or more representative manholes are determined for each cluster based on the closeness obtained by the hierarchical clustering.

4. A method for optimizing the placement of water level gauges as described in claim 2 or 3, wherein in the determination step, when an upper limit on the number of water level gauges to be placed is set, the representative manhole is determined based on the results of the clustering.

5. A method for optimizing the placement of water level gauges as claimed in any one of claims 1 to 4, wherein in the data acquisition step, precipitation data relating to the manholes is further acquired, and in the classification step, clustering is performed using the precipitation data and each of the water level data.

6. A method for optimizing the placement of water level gauges according to any one of claims 1 to 5, wherein the water level data is time-series water level data.

7. A method for predicting water levels in a manhole, comprising: a past data acquisition step of acquiring past time-series water level data of a manhole located in a specified area and predicted rainfall data for the specified area; and an output step of outputting predicted information for the water level of the manhole at a future time point by inputting the acquired past time-series water level data and predicted rainfall data of the manhole into a specified engine.

8. A method for predicting water levels in a manhole as described in claim 7, wherein in the past data acquisition step, past time series water level data of another manhole adjacent to the one manhole is further acquired, and in the output step, the past time series water level data of the other manhole is input to the specified engine.

9. A method for predicting the water level in a manhole as described in claim 7 or 8, wherein the past time-series water level data includes estimated water level data estimated based on water level data obtained by a water level gauge provided in a manhole different from the one manhole.

10. A method for predicting water levels in a manhole as described in any one of claims 7 to 9, wherein the predetermined engine is an engine based on a learning model obtained by learning using at least past time-series water level data for the one manhole and past rainfall data for the target area.

11. A method for predicting the water level in a manhole as described in claim 10, wherein the specified engine is further obtained by learning using past time-series water level data of other manholes adjacent to the one manhole.

12. A method for predicting water levels in a manhole as claimed in any one of claims 7 to 11, wherein in the output step, predicted information on water levels at future times for other manholes included in the target area is output together with or instead of the one manhole.

13. A manhole water level prediction system comprising: a past data acquisition unit that acquires past time-series water level data of a manhole included in a target area and predicted rainfall data for the specified area; and an output unit that outputs predicted information of the water level of the manhole at a future time point by inputting the acquired past time-series water level data of the manhole and the predicted rainfall data into a specified engine.

Images