Real-time inland flooding prediction system, real-time inland flooding prediction device, real-time inland flooding prediction method, real-time inland flooding prediction program, computer-readable recording medium, and device for storing the program

Abstract

It enables real-time prediction of water levels for inland flooding. [Solution] The real-time inland flooding prediction system 1000 comprises a flood volume calculation unit 31 that calculates the current flood volume in real time based on rainfall and initial water level, a predicted flood depth distribution model storage unit 22 that stores multiple predicted flood depth distribution models calculated in advance for each flood volume in the predicted area, a flood arrival time calculation unit 33 that calculates the flood arrival time when the flood is expected to reach the predicted area based on rainfall intensity, a predicted flood depth distribution model extraction unit 34 that corrects the flood volume calculated in real time by the flood volume calculation unit 31 to a corrected flood volume taking into account the flood arrival time calculated by the flood arrival time calculation unit 33 and extracts a predicted flood depth distribution model corresponding to the corrected flood volume from the multiple predicted flood depth distribution models stored in the predicted flood depth distribution model storage unit 22, and a display unit 40 that displays the predicted flood depth distribution model extracted by the predicted flood depth distribution model extraction unit 34.

Description

[Technical field]

[0001] The present disclosure relates to a real-time inland flooding prediction system, a real-time inland flooding prediction device, a real-time inland flooding prediction method, a real-time inland flooding prediction program, and a computer-readable recording medium and device storing the program. [Background technology]

[0002] In recent years, super typhoons and linear rain bands, thought to be caused by global warming and abnormal weather, have occurred in various places, raising concerns about flooding and inundation. Therefore, flood hazard maps and other information are provided for the purpose of providing safe evacuation routes, etc. In particular, various research projects are being conducted with the aim of improving the accuracy of real-time predictions of water levels when inland flooding occurs.

[0003] For example, a real-time dynamic flood simulation system has been proposed that predicts inundation based on observed rainfall (Patent Document 1). However, this real-time dynamic flood simulation system only makes predictions based on past rainfall, and there is a problem that prediction errors become large when the amount of rainfall changes from the past, for example, when the amount of rainfall becomes heavier or weaker. On the other hand, it is also possible to change the inundation prediction taking into account the predicted amount of rainfall, but in this case, the amount of calculations required depending on the complex topography of each region becomes enormous, making it difficult to make real-time inundation predictions, and if real-time inundation predictions were to be made, a considerably high level of calculation capacity would be required, resulting in a problem of soaring calculation costs. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2004-197554 A Summary of the Invention [Problem to be solved by the invention]

[0005] One of the objectives of one embodiment of the present disclosure is to provide a real-time inland flooding prediction system, a real-time inland flooding prediction device, a real-time inland flooding prediction method, a real-time inland flooding prediction program, and a computer-readable recording medium and device that store the same, which realizes real-time inland flooding water level prediction with high accuracy. Another objective of another embodiment is to provide a real-time inland flooding prediction system, a real-time inland flooding prediction device, a real-time inland flooding prediction method, a real-time inland flooding prediction program, and a computer-readable recording medium and device that store the same, which further improves accuracy by reflecting observed water level values ​​in flood depth predictions. Note that the description of these objectives and problems of the present disclosure does not preclude the existence of other objectives and problems. Furthermore, it is not necessary for one embodiment of the present disclosure to solve all of these problems. Furthermore, it is possible to extract other problems from the description of the specification, drawings, and claims of the present disclosure. Means for solving the problems and effects of the invention

[0006] The real-time inland flooding prediction system according to the first embodiment of the present disclosure is a real-time inland flooding prediction system that predicts in real time inland flooding caused by rainfall in a prediction area under conditions in which a prediction area is divided into small drainage basins, uniform rainfall is assumed in each small drainage basin, and a pre-calculated flood depth distribution is assumed to occur for each small drainage basin at each location included in the prediction area. The real-time inland flooding prediction system includes a data acquisition unit that acquires observed rainfall and predicted rainfall, a flooding volume calculation unit that calculates the current flooding volume in real time based on the observed rainfall acquired by the data acquisition unit, the predicted rainfall, and an initial water level, and a pre-calculated flooding volume for each flooding volume in the prediction area. The system comprises a predicted inundation depth distribution model storage unit for storing multiple predicted inundation depth distribution models, a flood arrival time calculation unit for calculating the flood arrival time when the flood is expected to reach the prediction area based on rainfall intensity, a predicted inundation depth distribution model extraction unit for correcting the inundation volume calculated in real time by the inundation volume calculation unit to a corrected inundation volume taking into account the flood arrival time calculated by the flood arrival time calculation unit and extracting a predicted inundation depth distribution model corresponding to the corrected inundation volume from the multiple predicted inundation depth distribution models stored in the predicted inundation depth distribution model storage unit, and a display unit for displaying the predicted inundation depth distribution model extracted by the predicted inundation depth distribution model extraction unit.With the above configuration, a real-time inland flooding prediction with improved prediction accuracy can be realized by adding not only the inundation volume but also the flood arrival time and converting it into an inundation depth distribution.

[0007] In addition, the real-time inland flooding prediction system of form 2, in addition to the above form, further includes a rainfall intensity-specific flooding volume time change storage unit that stores the results of pre-calculation by flood analysis of the relationship between flood volume according to rainfall intensity and elapsed time, and a flood depth distribution storage unit for storing the relationship between flood arrival time and flood volume.

[0008] Furthermore, in the real-time inland water flooding prediction system according to embodiment 3, in any of the above embodiments, the unit for storing water volume change over time by rainfall intensity includes, for each predetermined time, a flood analysis result in which a constant rainfall intensity of a plurality of different intensities continues as a graph during the water level increase period, and a flood analysis result in which a constant rainfall intensity of a plurality of different intensities continues as a graph during the water level decrease period. With the above configuration, a flood analysis in the case of a constant rainfall intensity continuing can be performed in advance.

[0009] Furthermore, in the real-time inland flooding prediction system of form 4, in any of the above forms, the flood arrival time calculation unit calculates the flood depth distribution for time step (w+1) by averaging multiple adjacent points among multiple plots that each show the relationship between the amount of water expected for multiple different probable rainfalls and the flood arrival time at that time, as recorded in the flood depth distribution storage unit.

[0010] Furthermore, in the real-time inland flooding prediction system according to a fifth aspect, in any one of the above-mentioned aspects, the plurality of different probable rainfalls includes at least one of a probable rainfall for a specific year and an assumed maximum scale.

[0011] Furthermore, in the real-time inland water flooding prediction system according to a sixth aspect, in any of the above aspects, the inundation depth distribution storage unit includes models of a front-concentration type rainfall waveform in which rainfall is concentrated at the beginning of a rainfall period, a middle-concentration type rainfall waveform in which rainfall is concentrated in the middle of a rainfall period, and a rear-concentration type rainfall waveform in which rainfall is concentrated at the end of a rainfall period, in order to assume the amount of inundation at different probable rainfall amounts. With the above configuration, it is possible to perform in advance a flood analysis with the assumed maximum rainfall amount as an external force.

[0012] Furthermore, the real-time inland flooding prediction system according to embodiment 7 is any of the above embodiments, further comprising a plurality of water level sensors installed at predetermined installation positions in the prediction area, the flood volume calculation unit being connected to each of the plurality of water level sensors via a network, and the flood volume calculation unit being configured to correct the flood volume calculated in real time with the measured water levels at the installation positions observed by the water level sensors. With the above configuration, more accurate real-time inland flooding predictions can be realized by correcting the flood volume at each installation position with the actually measured water levels.

[0013] Furthermore, the real-time inland flooding prediction system according to aspect 8 is any of the above aspects, wherein the flood volume calculation unit performs real-time calculation of the flood volume using the measured water level at the installation position observed by the water level sensor as the initial water level. With the above configuration, it is possible to perform a more accurate inland flooding prediction by performing a real-time inland flooding prediction using the measured water level at a time when inundation has already occurred as the initial water level, rather than being limited to a real-time inland flooding prediction using the initial water level at the start of rainfall.

[0014] Furthermore, the real-time inland flooding prediction system according to aspect 9 is any of the above aspects, wherein the flood volume calculation unit is configured to complement the water levels of the surrounding area with adjacent measured water levels among the measured water levels at the multiple installation positions observed by the multiple water level sensors. With the above configuration, more accurate real-time inland flooding prediction can be realized by complementing area information from the point measured water levels.

[0015] A real-time inland flooding prediction system according to still another embodiment is a real-time inland flooding prediction system that predicts in real time inland flooding in a prediction area caused by rainfall under conditions in which a prediction area is divided into small drainage basins, uniform rain falls in each small drainage basin, and a pre-calculated inundation depth distribution occurs for each small drainage basin at each location included in the prediction area, and includes a flood volume calculation unit for calculating the current flood volume in real time based on an initial water level, a predicted flood depth distribution model storage unit for storing a plurality of flood depth distribution models pre-calculated for each flood volume in the prediction area, and a predicted flood depth distribution model extraction unit for extracting a corresponding predicted flood depth distribution model from the plurality of predicted flood depth distribution models stored in the predicted flood depth distribution model storage unit based on the flood volume calculated in real time by the flood volume calculation unit. With the above configuration, real-time inland flooding prediction is possible without using a dynamic model during real-time operation.

[0016] A real-time inland flooding prediction device according to a tenth aspect is a real-time inland flooding prediction device which predicts in real time inland flooding in a prediction area caused by rainfall under conditions in which a prediction area is divided into small drainage basins, uniform rainfall is assumed in each small drainage basin, and a pre-calculated flood depth distribution is assumed to occur for each small drainage basin at each location included in the prediction area, and which includes a data acquisition unit which acquires observed rainfall and predicted rainfall, a flooding amount calculation unit which calculates the current flooding amount in real time based on the observed rainfall, the predicted rainfall and the initial water level acquired by the data acquisition unit, and a plurality of predicted flood depth distribution models which are pre-calculated for each flooding amount in the prediction area. The system comprises a predicted inundation depth distribution model storage unit for calculating the flood arrival time when a flood is expected to reach a prediction area based on rainfall intensity, a flood arrival time calculation unit for calculating the flood arrival time when a flood is expected to reach a prediction area based on rainfall intensity, and a predicted inundation depth distribution model extraction unit for extracting a corresponding predicted inundation depth distribution model from a plurality of predicted inundation depth distribution models stored in the predicted inundation depth distribution model storage unit based on the inundation volume calculated in real time by the inundation volume calculation unit, wherein the predicted inundation depth distribution model extraction unit is configured to correct the inundation volume calculated in real time by the inundation volume calculation unit to a corrected inundation volume taking into account the flood arrival time calculated by the flood arrival time calculation unit, and to extract a predicted inundation depth distribution model corresponding to the corrected inundation volume. With the above configuration, a real-time inland flooding prediction with improved prediction accuracy can be realized by adding not only the inundation volume but also the flood arrival time and converting it into an inundation depth distribution.

[0017] A real-time inland flooding prediction method according to an eleventh aspect is a real-time inland flooding prediction method for predicting inland flooding caused by rainfall in a prediction area in real time under the condition that a prediction area is divided into small drainage basins, uniform rain falls in each small drainage basin, and a pre-calculated flood depth distribution occurs for each small drainage basin at each position included in the prediction area, the method comprising the steps of: calculating a current flood volume in real time based on an initial water level; and calculating a flood volume of the prediction area in real time based on the flood volume calculated in real time. and extracting a corresponding predicted inundation depth distribution model from a plurality of predicted inundation depth distribution models calculated in advance for each flood volume in the area, wherein the step of extracting the corresponding predicted inundation depth distribution model may include a step of calculating the flood arrival time at which the flood is expected to reach the prediction area based on rainfall intensity in a flood arrival time calculation unit, correcting the flood volume calculated in real time to a corrected flood volume taking into account the calculated flood arrival time, and extracting a predicted inundation depth distribution model corresponding to the corrected flood volume. In this way, by adding not only the flood volume but also the flood arrival time and converting it into an inundation depth distribution, it is possible to realize a real-time inland flooding prediction with improved prediction accuracy.

[0018] Furthermore, the real-time inland flooding prediction method according to aspect 12, in any of the above aspects, includes, in the step of extracting the corresponding predicted inundation depth distribution model, a step of preparing an inundation depth distribution storage unit that stores the relationship between inundation volume and flood arrival time as inundation depth distribution data, a step of preparing a rainfall intensity-specific inundation volume time change storage unit that stores rainfall intensity-specific inundation volume time change data calculated by flood analysis as the relationship between inundation volume according to rainfall intensity and elapsed time, a step of comparing the inundation volume (V(t)) of the current time step (w), which is the inundation volume at the current time, with the rainfall intensity-specific inundation volume time change data stored in the rainfall intensity-specific inundation volume time change storage unit, and a step of calculating a time (Δw) from the inundation volume (V(t)) of the current time step (w) a step of calculating a time change (ΔV(R)) of flooding volume due to any rainfall (R) occurring up to a time step (w+1) after the elapse of time from the flooding volume time change data by rainfall intensity; a step of estimating the flooding volume (V(t+1)) of the time step (w+1) based on the calculated time change (ΔV(R)) of the flooding volume of the time step (w+1); a step of comparing the flooding volume (V(t+1)) of the time step (w+1) with the flooding depth distribution data stored in the flooding depth distribution storage unit; a step of storing a rainfall intensity curve calculated from a rainfall waveform in a rainfall intensity curve storage unit; a step of calculating the Kadoya-Fukushima formula in a quasi-linear storage type function model calculation unit; and a step of storing the rainfall intensity curve stored in the rainfall intensity curve storage unit. quasi A step of calculating the flood arrival time in a flood arrival time calculation unit, the point where the linear storage type function model intersects with the flood arrival time calculation unit, and a step of calculating the flood volume (V A (T A)), a step of calculating an averaged flood depth distribution for the time step (w+1) by averaging multiple nearby points recorded in advance in the flood depth distribution storage unit, and a step of repeating the steps from the step of comparing with the flood volume time change data by rainfall intensity to the step of calculating the averaged flood depth distribution for the time step (w+2) after a certain time (Δw) has elapsed since the time step (w+1) to calculate the averaged flood depth distribution for each time. In this way, by calling up the flood depth distribution in the real-time operation phase according to the flood volume and flood arrival time calculated from an arbitrary rainfall waveform, it is possible to reduce the calculation load in the real-time operation phase and realize real-time inland flooding prediction.

[0019] Furthermore, in the real-time inland water flooding prediction method according to aspect 13, in any of the above aspects, in the step of preparing the inundation depth distribution storage unit, the step of creating a rainfall intensity formula for each of a plurality of different probable rainfalls, and the step of performing flood analysis for each of the created rainfall intensity formulas for each probable rainfall using a front-concentration type rainfall waveform in which rainfall is concentrated at the beginning of a rainfall period, a middle-concentration type rainfall waveform in which rainfall is concentrated in the middle of a rainfall period, and a rear-concentration type rainfall waveform in which rainfall is concentrated at the end of a rainfall period as external forces, and outputting the flood analysis results at predetermined time intervals. This makes it possible to perform in advance a flood analysis using the expected maximum rainfall as an external force.

[0020] Furthermore, in the real-time inland flooding prediction method of aspect 14, in any of the above aspects, the multiple different probable rainfalls include at least one of a probable rainfall for a specific year and an expected maximum scale.

[0021] Furthermore, the real-time inland water flooding prediction method according to aspect 15, in any of the above aspects, includes, in the step of preparing the inundation depth distribution storage unit, a step of performing inundation analyses using a plurality of different intensities of continuous constant rainfall intensity as a graph for the water level increase period as an external force of rainfall and outputting the flood analysis results every predetermined time, and a step of performing inundation analyses using a plurality of different intensities of continuous constant rainfall intensity as a graph for the water level decrease period as an external force of rainfall and outputting the flood analysis results every predetermined time. This makes it possible to perform in advance a flood analysis for the case where a constant rainfall intensity continues.

[0022] A real-time inland flooding prediction method according to a sixteenth aspect, in any of the above aspects, further includes the steps of acquiring water levels at each installation location observed by a plurality of water level sensors installed at predetermined installation locations in the prediction area, and correcting the amount of flooding calculated in real time with the water levels at the installation locations observed by the water level sensors. This allows for more accurate real-time prediction of inland flooding by correcting the amount of flooding at each installation location with the actually measured water level.

[0023] Furthermore, in the real-time inland flooding prediction method according to aspect 17, in any of the above aspects, the step of correcting the flooded volume includes a step of performing real-time calculation of the flooded volume using the measured water level at the installation position observed by the water level sensor as the initial water level. This enables more accurate inland flooding prediction by performing real-time inland flooding prediction using the measured water level at a time when flooding has already occurred as the initial water level, rather than being limited to real-time inland flooding prediction using the initial water level at the start of rainfall.

[0024] Furthermore, in the real-time prediction method for inland flooding according to aspect 18, in any of the above aspects, the step of correcting the flood volume includes a step of complementing the water levels of the surrounding area with adjacent measured water levels among the measured water levels at the multiple installation positions observed by the multiple water level sensors. This makes it possible to realize more accurate real-time prediction of inland flooding by complementing the surrounding water levels from the actually measured measured water levels.

[0025] Furthermore, the real-time inland flooding prediction program of form 20 is a real-time inland flooding prediction program for predicting in real time inland flooding in a predicted area caused by rainfall under conditions in which the predicted area is divided into small drainage basins and uniform rain falls in each small drainage basin, and a pre-calculated flood depth distribution occurs for each small drainage basin at each location included in the predicted area, and is configured to enable a computer to realize the following functions: calculating the current flood volume in real time based on the initial water level; extracting a corresponding predicted flood depth distribution model from a plurality of predicted flood depth distribution models pre-calculated for each flood volume in the predicted area based on the flood volume calculated in real time; and calculating the flood arrival time when the flood is expected to reach the predicted area based on the rainfall intensity, correcting the flood volume calculated in real time to a corrected flood volume taking into account the calculated flood arrival time, and extracting a predicted flood depth distribution model corresponding to the corrected flood volume. This makes it possible to realize real-time inland flooding predictions with improved prediction accuracy by adding not only the flood volume but also the flood arrival time and converting it into a water depth distribution.

[0026] Furthermore, the computer-readable recording medium or the device storing the program according to the 20th aspect stores the program according to the above aspect. The recording medium includes a CD-ROM, a CD-R, a CD-RW, a flexible disk, a magnetic tape, an MO, a DVD-ROM, a DVD-RAM, a DVD-R, a DVD+R, a DVD-RW, a DVD+RW, a Blu-ray (registered trademark), a HD DVD (AOD), or other magnetic disks, an optical disk, an optical magnetic disk, a semiconductor memory, and other media capable of storing a program. The program includes a program stored in the recording medium and distributed, as well as a program distributed by downloading via a network line such as the Internet. The device storing the program includes a general-purpose or dedicated device in which the program is implemented in an executable state in the form of software or firmware. Furthermore, each process or function included in the program may be executed by a program software executable by a computer, or each process may be realized by hardware such as a predetermined gate array (FPGA, ASIC), or a form in which the program software and a partial hardware module that realizes some elements of the hardware are mixed. [Brief description of the drawings]

[0027] [Figure 1] 1 is a block diagram showing a real-time inland flooding prediction system according to embodiment 1. [Diagram 2] This is a block diagram showing a network-connected real-time inland flooding prediction system. [Diagram 3] FIG. 1 is a schematic diagram showing a real-time inland flood inundation prediction method according to embodiment 1. [Figure 4] 1 is a flowchart showing a real-time inland water flooding prediction method according to embodiment 1. [Diagram 5] 1 is a graph showing a flood analysis when a constant rainfall intensity continues. [Figure 6] 1 is a graph showing the change in water volume over time by rainfall intensity. [Figure 7]7A to 7D are graphs showing exception processing for taking delay times into consideration. [Figure 8] This is a graph of the water level increase period used in the basic process of calculating the water volume at the next time step. [Figure 9] This is a graph of the water level decline period used in the basic process of calculating the water volume at the next time step. [Figure 10] 1 is a graph showing the change in water volume over time. [Figure 11] 13 is a graph showing the comparison results of maximum flood depth. [Figure 12] 13 is a graph showing a comparison result of maximum flow flux. [Figure 13] This is a graph showing the comparison of flood depth every hour on the hour. [Figure 14] FIG. 13 is an image showing the distribution of flood depth at 11pm on the 20th using the method of Comparative Example 1. [Figure 15] FIG. 13 is an image showing the distribution of flood depth at 11:00 p.m. on the 20th using the method according to Example 1. [Figure 16] FIG. 11 is a schematic diagram showing a real-time inland flooding prediction method according to the second embodiment. [Figure 17] 13 is a flowchart showing a procedure of a real-time operation phase according to the second embodiment. [Figure 18] FIG. 11 is a conceptual diagram showing how to calculate flood depth at each time step in the second embodiment. [Figure 19] 1 is a graph showing the change in water volume over time. [Figure 20] This is a graph comparing maximum flood depths. [Figure 21] 1 is a graph comparing maximum flow fluxes. [Figure 22] This is a graph comparing the flood depth every hour on the hour. [Figure 23] This is a graph showing the accuracy rate of color combinations for each hour on the hour. [Figure 24] FIG. 13 is an image showing the maximum flood depth distribution using the method of Comparative Example 2. [Diagram 25] FIG. 13 is an image showing the maximum flood depth distribution using the method of Example 2. [Figure 26] FIG. 13 is an image diagram showing the difference in maximum flood depth distribution between Comparative Example 2 and Example 2. [Figure 27] This is a graph showing the accuracy rate of above-floor flooding for each hour on the hour. [Figure 28] This is an image showing the distribution of flood depth using the method of comparison example 2 at 5:00 p.m. on the 19th. [Figure 29] This is an image showing the distribution of flood depth using the method of Example 2 at 17:00 on the 19th. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0028] Hereinafter, the embodiments of the present disclosure will be described based on the drawings. However, the embodiments shown below are examples for embodying the technical ideas of the present disclosure, and the present disclosure is not limited to the following. In addition, this specification never specifies the members shown in the claims to the members of the embodiments. In particular, the dimensions, materials, shapes, and relative positions of the components described in the embodiments are not intended to limit the scope of the present disclosure to those alone, and are merely explanatory examples, unless otherwise specified. Note that the size and positional relationship of the members shown in each drawing may be exaggerated to clarify the explanation. Furthermore, in the following explanation, the same name and symbol indicate the same or similar members, and detailed explanation will be omitted as appropriate. Furthermore, each element constituting the present disclosure may be configured as a form in which multiple elements are composed of the same member and one member serves multiple elements, or conversely, the function of one member can be shared among multiple members.

[0029] The real-time inland flooding prediction device used in the embodiments of the present disclosure is connected to a computer, printer, external storage device, and other peripheral devices for operation, control, display, and other processing, etc., via a serial connection such as IEEE1394, RS-232x, RS-422, RS-423, RS-485, USB, or a parallel connection, or via a network such as 10BASE-T, 100BASE-TX, or 1000BASE-T, electrically, magnetically, or optically. The connection is not limited to a physical connection using a wire, but may be a wireless connection using radio waves, infrared rays, optical communications, etc., such as wireless LAN such as IEEE802.1x, Bluetooth (registered trademark), or other NFC. Furthermore, a memory card, a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc. may be used as a recording medium for exchanging data and saving settings. In this specification, the real-time inland flooding prediction device is used to include not only the real-time inland flooding prediction device itself, but also an inland flooding prediction system that combines peripheral devices such as a computer and an external storage device with it. [Embodiment 1]

[0030] The real-time inland flooding prediction system predicts inland flooding in a prediction area that is the target of the inundation prediction. FIG. 1 shows a real-time inland flooding prediction system 1000 according to the first embodiment of the present disclosure. The real-time inland flooding prediction system 1000 shown in this figure predicts a flood depth distribution map of a prediction area based on rainfall data in a certain prediction area, and displays it on the display unit 40. Here, the prediction area refers to the entire area in which the real-time inland flooding prediction system 1000 predicts the presence or absence of flooding and the flood risk. The range of this prediction area is not particularly specified. The range of the prediction area can be arbitrarily determined by the system user within the target area, and can be, for example, a specific city, town, or village unit. The flood depth distribution map is overlaid on a map to display the locations where flooding has occurred and the water depth. In addition, the real-time inland flooding prediction system divides the prediction area into small drainage basins, assumes that uniform rain will fall within each small drainage basin, and predicts in real time inland flooding in the prediction area caused by rainfall under conditions assuming that a pre-calculated flood depth distribution will occur for each small drainage basin at each location included in the prediction area.

[0031] The real-time inland water flooding prediction system 1000 includes a real-time inland water flooding prediction device 100. The real-time inland water flooding prediction device 100 can be configured with dedicated hardware, or by installing an inland water flooding prediction program on a general-purpose or dedicated computer. This real-time inland water flooding prediction system 1000 includes a data acquisition unit 10, a data storage unit 20, a calculation unit 30, a display unit 40, and an operation unit 50. (Data Acquisition Unit 10)

[0032] The data acquisition unit 10 is a component for acquiring data from the outside. Here, it functions as a component for acquiring rainfall data. The rainfall data is typically rainfall per unit time. The rainfall is observed rainfall (past rainfall) observed in the past. The rainfall date may also include predicted rainfall (future rainfall) that predicts future rainfall.

[0033] These rainfall data are acquired via a network. This makes it easy to update the data to the latest information. For this reason, the data acquisition unit 10 has a communication function for connecting to a general-purpose network such as the Internet, or a specific network via a dedicated line, etc.

[0034] The rainfall data includes the analyzed rainfall and the short-term precipitation forecast for the predicted area. The analyzed rainfall is the amount of rainfall that has actually fallen within a specified time period from the current time. The short-term precipitation forecast is the amount of rainfall that is predicted to fall within a specified time period from the current time. In this embodiment, the specified time period is every hour. However, it may be every 30 minutes or any other arbitrary time unit.

[0035] For example, the rainfall data distributed by the Meteorological Business Support Center includes analyzed rainfall and short-term precipitation forecasts, and is updated every 30 minutes to predict rainfall up to six hours ahead for an area divided into 1 km meshes of topographical data. Therefore, the data acquisition unit 10 sequentially acquires such data and sends it to the calculation unit 30. The rainfall data acquired by the data acquisition unit 10 can also be stored in the data storage unit 20. For example, the data storage unit 20 may be provided with a rainfall data storage unit for storing rainfall data.

[0036] The source of rainfall data is not particularly specified, and for example, rainfall data distributed by the Japan Meteorological Agency or the Meteorological Operations Support Center may be used, or the data may be collected directly by installing a unique rainfall observation device, etc. Furthermore, the predetermined time is not limited to one hour, and may be shorter (e.g., 30 minutes) or longer (e.g., two hours). (Data storage unit 20)

[0037] The data storage unit 20 is a component for holding various data. This data storage unit 20 includes a flood depth distribution storage unit 21, a predicted flood depth distribution model storage unit 22, a rainfall intensity-specific water volume change over time storage unit 23, a rainfall intensity curve storage unit 26, and a quasi-linear storage function model storage unit 27. (Flood depth distribution preservation section 21)

[0038] The inundation depth distribution storage unit 21 stores a group of inundation depth distributions. The inundation depth distribution indicates the relationship between the flood arrival time and the amount of water. This inundation depth distribution can be expressed as a flood depth distribution map that indicates the flooding situation in a predicted area at a specific time. This inundation depth distribution group is generated in advance by the calculation unit 50 or an external computer, etc., and stored in the inundation depth distribution storage unit 21.

[0039] The flood depth distribution storage unit 21 can include multiple rainfall waveform models to estimate the amount of flooding with different probable rainfall amounts. Specifically, it includes models of a front-concentration type rainfall waveform in which rainfall is concentrated at the beginning of a rainfall period, a middle-concentration type rainfall waveform in which rainfall is concentrated in the middle of a rainfall period, and a rear-concentration type rainfall waveform in which rainfall is concentrated at the end of a rainfall period. This allows flood analysis to be performed in advance with the expected maximum rainfall amount as an external force.

[0040] The predicted flood depth distribution model storage unit 22 is a component for storing multiple predicted flood depth distribution models that are calculated in advance for each amount of flooding in the predicted area. (Storage section for time-dependent changes in water volume by rainfall intensity 23)

[0041] The unit 23 for storing the time variation of flood volume by rainfall intensity is a component for storing the results of a calculation performed in advance by flood analysis of the relationship between the flood volume according to the rainfall intensity and the elapsed time. Here, the unit 23 for storing the time variation of flood volume by rainfall intensity stores the data on the time variation of flood volume by rainfall intensity. The data on the time variation of flood volume by rainfall intensity is, for example, a graph with the horizontal axis representing the elapsed time and the vertical axis representing the flood volume. Furthermore, the data on the time variation of flood volume by rainfall intensity can be created for each of the water level rise period when the external force of rainfall is greater than the drainage capacity of the predicted area and the water level fall period when the external force of rainfall is less than the drainage capacity of the predicted area, as shown in FIG. 16 described later, and stored in the unit 23 for storing the time variation of flood volume by rainfall intensity. That is, the unit 23 for storing the time-dependent changes in flood volume by rainfall intensity can store, for each predetermined time, the results of flood analysis in which a constant rainfall intensity of multiple different intensities continues as a graph during the water level rise period, and the results of flood analysis in which a constant rainfall intensity of multiple different intensities continues as a graph during the water level fall period, for each predetermined time. This makes it possible to perform in advance a flood analysis in which a constant rainfall intensity continues.

[0042] The rainfall intensity curve storage unit 26 is a member for storing the rainfall intensity curve calculated from the rainfall waveform.

[0043] The quasi-linear storage function model storage unit 27 is a component for storing the quasi-linear storage function model. The quasi-linear storage function model is one of the runoff calculations that calculates the flood arrival time in a watershed where land use types such as rice fields, building sites, and forests are mixed. In the quasi-linear storage function model, the intersection of the rainfall intensity curve and the two curves of the Kadoya-Fukushima formula is calculated as the flood arrival time. (Computation unit 30)

[0044] The calculation unit 30 is a component for performing various calculations, and here it functions as a flood volume calculation unit 31, a flood arrival time calculation unit 33, a predicted inundation depth distribution model extraction unit 34, and a quasi-linear storage type function model calculation unit 35.

[0045] The flood water volume calculation unit 31 is a component for calculating the current flood water volume in real time based on the observed rainfall amount acquired by the data acquisition unit 10, the predicted rainfall amount, and the initial water level.

[0046] The flood arrival time calculation unit 33 is a component for calculating the flood arrival time that is expected to reach the forecast area based on the rainfall intensity. Here, the point where the rainfall intensity curve stored in the rainfall intensity curve storage unit 26 intersects with the quasi-linear storage function model stored in the quasi-linear storage function model storage unit 27 is calculated as the flood arrival time.

[0047] The predicted inundation depth distribution model extraction unit 34 is a component that corrects the flood volume calculated in real time by the flood volume calculation unit 31 to a corrected flood volume taking into account the flood arrival time calculated by the flood arrival time calculation unit 33, and extracts a predicted inundation depth distribution model corresponding to the corrected flood volume from among the multiple predicted inundation depth distribution models stored in the predicted inundation depth distribution model storage unit 22.

[0048] The quasi-linear reservoir type function model calculation unit 35 is a component for calculating a quasi-linear reservoir type function model, specifically, the Kadoya-Fukushima formula.

[0049] With this configuration, it is possible to realize real-time inland flooding predictions with improved prediction accuracy by adding not only the flood volume but also the flood arrival time and converting it into a water depth distribution. (Flood arrival time calculation unit 33)

[0050] The flood arrival time calculation unit 33 can calculate the flood depth distribution for time step (w+1) by averaging multiple adjacent points among multiple plots that show the relationship between the flood volume assumed for multiple different probable rainfalls and the flood arrival time at that time, which are recorded in the inundation depth distribution storage unit 21. Here, the multiple different probable rainfalls can include at least one of the probable rainfall for a specific year and the maximum expected scale. Examples of the probable rainfall for a specific year include a 2-year, 10-year, 50-year, 100-year, and 500-year probability.

[0051] The calculation unit 30 also includes a database access unit 32 for accessing the data storage unit 20. The database access unit 32 is an interface for accessing the data storage unit 20, such as the inundation depth distribution storage unit 21. The database access unit 32 may be configured to access a storage medium, such as a hard disk or semiconductor memory, included in the real-time inland water flooding prediction device 100, or may be configured to access a database connected externally to the real-time inland water flooding prediction device 100. For example, the real-time inland water flooding prediction device 100 and the real-time inland water flooding prediction system 1000 can also be constructed by using an external storage, such as a NAS or file server connected via a network line, as the data storage unit 20.

[0052] Furthermore, as shown in the real-time inland flooding inundation prediction system 1000' in Figure 2, the calculation unit 30 can also function as a flood depth distribution group creation unit 33 that creates a flood depth distribution group, and a pre-calculation graph creation unit 34 that creates a pre-calculation graph. However, since it is sufficient to create the flood depth distribution group and pre-calculation graph in advance and store them in the data recording unit 20, they do not necessarily have to be created by the real-time inland flooding inundation prediction device 100. For example, it is sufficient to create them in a calculator such as an external computer or server and register them in the data storage unit 20 or another database. In this case, there is no need to provide the flood depth distribution group creation unit 33 or pre-calculation graph creation unit 34 in the real-time inland flooding inundation prediction device 100. Figure 1 shows an example of such a configuration.

[0053] For example, a CPU can be used for such a computing unit 30. In addition to a CPU, a graphics processing unit (GPU), for example, a GPGPU, can also be used. By using such a GPGPU (General-Purpose computing on Graphics Processing Units), the load on the CPU can be reduced, the computing speed can be improved, and the prediction results can be made more accurate, while the system user can easily check the prediction results on a display unit 40, which will be described later. For example, CUDA (product name) can be used for the GPGPU. (Display section 40)

[0054] The display unit 40 is a component for displaying the predicted flood depth distribution model extracted by the predicted flood depth distribution model extraction unit 34. The display unit 40 can display future flood area predictions and past flood depth distribution. Displays such as a liquid crystal display, organic electroluminescence (EL), or CRT can be used for such a display unit 40. The display unit may also serve as an input unit as a touch panel.

[0055] The display unit 40 has a map display area for displaying a map of the predicted area. In the map display area, the map is displayed at a predetermined magnification. The map magnification can be enlarged or reduced by operating the zoom buttons provided in the corners of the map display area. Alternatively, the map may be enlarged or reduced by operating the scroll button of a mouse constituting the operation unit 50. For data management purposes, grid lines are displayed vertically and horizontally at a predetermined width, and the map is displayed in blocks divided by the lines. Hereinafter, the areas divided by the lines are referred to as blocks (reference area meshes). The map display magnification can be changed as appropriate. (Operation unit 50)

[0056] The operation unit 50 is a member for performing various operations, and is composed of, for example, a keyboard, a pointing device such as a mouse, a console, etc. In addition, by using a touch panel, the operation unit 50 and the display unit 40 can be integrated. (Server-Client)

[0057] Furthermore, by using the calculation unit as a server device and the display unit 40 as a display screen of a client device connected to this server device, a user can access the server device from outside using the client device and view the prediction results and the like obtained by the calculation unit. An example of this is shown in Fig. 2. In this case, the operation unit 50 can use operations such as a touch panel or a mouse for the client device CL connected to the server device SV via a network.

[0058] The display unit 40 displays a user interface screen of the inland flooding prediction program. As mentioned above, the user interface screen of the inland flooding prediction program is displayed on the display unit 40 connected to a standalone computer, as well as on the display screen of a client device that accesses a server device on which the inland flooding prediction program is installed, as shown in Figure 2. The rendering can be web-based, making it versatile. The client device can be a smartphone, mobile phone, personal computer, or other device equipped with a communication function to access the server device.

[0059] With this configuration, rather than calculating and creating future flood depth distribution maps each time, a flood depth distribution map can be selected and displayed from a collection of multiple pre-calculated flood depth distribution maps according to their correspondence, making it possible to display the flood depth distribution map with a light load, and real-time prediction of inland flooding can be realized without increasing the calculation performance. (Water level sensor WS)

[0060] The real-time inland flood prediction system 1000 may also include multiple water level sensors WS. Each water level sensor WS is installed at a predetermined installation position in the prediction area to detect the presence or absence of flooding. The water level sensor WS may also have the function of a water level meter that measures the water level.

[0061] Each water level sensor WS is connected to the real-time inland flooding prediction device via a network. In this case, the data acquisition unit 10 functions as an input interface that acquires information such as the presence or absence of flooding and the water level acquired by the water level sensor WS in addition to rainfall data. This allows the amount of flooding at each installation location to be corrected with the actually measured water level, thereby realizing more accurate real-time inland flooding prediction.

[0062] The flood volume calculation unit 31 of the real-time inland flooding prediction device can correct the flood volume calculated in real time with the measured water level at the installation location observed by the water level sensor WS. This allows the flood volume at each installation location to be corrected with the actually measured water level, thereby realizing more accurate real-time prediction of inland flooding.

[0063] The flood volume calculation unit 31 may be configured to perform real-time calculation of the flood volume using the measured water level at the installation position observed by the water level sensor WS as the initial water level. With this configuration, real-time prediction of inland flooding is possible using the measured water level at which inundation has already occurred as the initial water level, rather than being limited to real-time prediction of inland flooding using the initial water level at the start of rainfall, thereby enabling more accurate prediction of inland flooding.

[0064] Furthermore, the flood volume calculation unit 31 may be configured to complement the water levels of the surrounding areas with adjacent measured water levels among the measured water levels at multiple installation positions observed by multiple water level sensors WS. This allows for more accurate real-time prediction of inland flooding by complementing area information from the points of the measured water levels that are actually measured. In particular, the accuracy of the flood depth distribution can be improved by calculating and complementing area data from point data of the measured water levels at multiple installation positions. As an interpolation method for obtaining area data from such point data, known methods such as TIN interpolation (Triangulated Irregular Network), Kriging method, spline method, inverse distance weighting method, natural neighbor interpolation method, improved Shepard method, radial basis function, average method, TOPOGRID, etc. can be used as appropriate.

[0065] In this way, if a water level gauge is installed within a prediction area for which inundation prediction is performed by the real-time inland flooding prediction system 1000, the data from this water level gauge can be used to improve the accuracy of the real-time inland flooding prediction. The examples in Figures 1 and 2 show how multiple water level sensors WS are installed within the prediction area.

[0066] Here, the water level sensor WS is a device that observes the water level, also called a flood sensor. The water level sensor WS may be a device that measures the water level as a numerical value, or may be a device that detects when the water level reaches a preset specified water level. For example, the specified water level may be set to a water level that is determined to be flooded, and a flood sensor that is set to operate when the water level reaches this specified water level may be used as the water level sensor WS.

[0067] Such water level sensors WS are installed at multiple locations within the predicted flooding area. Examples include roads such as main roads in urban areas, designated evacuation sites, public facilities such as city halls, ponds, rivers, etc. The locations where water level sensors WS are installed are called water level observation locations.

[0068] The water level sensor WS outputs a water level signal indicating the water level to the flood prediction device. The water level sensor WS is connected to the flood prediction device by wire or wirelessly. As shown in FIG. 1, each water level sensor WS may be connected to the flood prediction device directly or via a repeater. In the example of FIG. 2, multiple water level sensors WS are wirelessly connected to the flood prediction device via a repeater RT. For example, the multiple water level sensors WS and the repeater RT are connected by a standardized communication method such as WiFi, Bluetooth such as BLE, or ZigBee (all of which are product or service names), and the repeater RT and the flood prediction device are connected by public lines such as telephone lines or optical fibers. [Real-time inland flooding prediction method]

[0069] Here, we will explain a real-time inland flooding prediction method that predicts inland flooding caused by rainfall in a prediction area in real time, under conditions where a prediction area is divided into small drainage basins, uniform rain falls in each small drainage basin, and a pre-calculated inundation depth distribution occurs for each small drainage basin at each location included in the prediction area. First, the flooding volume calculation unit 31 calculates the current flooding volume in real time based on the initial water level. Next, based on the flooding volume calculated in real time by the flooding volume calculation unit 31, the corresponding predicted flooding depth distribution model is extracted by the predicted flooding depth distribution model extraction unit 34 from multiple predicted flooding depth distribution models calculated in advance for each flooding volume in the prediction area. The process of extracting the corresponding predicted inundation depth distribution model includes a process in which the flood arrival time calculation unit 33 calculates the flood arrival time that the flood is expected to reach the prediction area based on the rainfall intensity, corrects the flood volume calculated in real time to a corrected flood volume taking into account the calculated flood arrival time, and extracts a predicted inundation depth distribution model that corresponds to the corrected flood volume in a predicted inundation depth distribution model extraction unit 34. In this way, by adding not only the flood volume but also the flood arrival time and converting it into an inundation depth distribution, it is possible to realize real-time inland flooding prediction with improved prediction accuracy.

[0070] In addition, in the process of extracting the corresponding predicted flood depth distribution model by the predicted flood depth distribution model extraction unit 34, the following steps are performed: preparing a flood depth distribution storage unit 21 that stores the relationship between flood volume and flood arrival time as flood depth distribution data; preparing a rainfall intensity-specific flood volume time change storage unit 23 that stores the relationship between the flood volume according to the rainfall intensity and the elapsed time calculated by flood analysis; and storing the flood volume at the current time (known a step of comparing the flood volume (V(t)) of the current time step (w), which is the amount of flooding (V(t)) at the current time step (w), with the flood volume time change data by rainfall intensity stored in the rainfall intensity-specific flood volume time change storage unit 23; and a step of calculating the time change (ΔV(R)) of the flood volume due to an arbitrary rainfall amount (R) occurring from the flood volume (V(t)) at the current time step (w) to the next time step (w+1) after a certain time (Δw) has elapsed, from the flood volume (V(t)) at the current time step (w), from the rainfall intensity-specific flood volume time change data. A process of estimating the flood volume (V(t+1)) at the time step (w+1) based on the time change (ΔV(R)) of the flood volume at the calculated time step (w+1), a process of comparing the flood volume (V(t+1)) at the time step (w+1) with the flood depth distribution data stored in the flood depth distribution storage unit 21, a process of storing the rainfall intensity curve calculated from the rainfall waveform in the rainfall intensity curve storage unit 26, and a process of applying the Kadoya-Fukushima formula to a quasi-linear storage type function A step of calculating the flood arrival time in a model calculation unit 35, a step of calculating the flood arrival time in a flood arrival time calculation unit 33 from the point where the rainfall intensity curve stored in the rainfall intensity curve storage unit 26 and the quasi-linear storage type function model stored in the quasi-linear storage type function model storage unit 27 intersect, and a step of calculating the flood arrival time calculated in the flood arrival time calculation unit 33 against the flood depth distribution data stored in the flood depth distribution storage unit 21 to calculate the flood volume (V A (T A)), a step of calculating an averaged flood depth distribution for the time step (w+1) by averaging multiple nearby points recorded in advance in the flood depth distribution storage unit 21, and a step of repeating the steps from the step of comparing with the flood volume time change data by rainfall intensity to the step of calculating the averaged flood depth distribution for the time step (w+2) after a certain time (Δw) has elapsed since the time step (w+1) to calculate the averaged flood depth distribution for each time. In this way, by calling up the flood depth distribution in the real-time operation phase according to the flood volume and flood arrival time calculated from an arbitrary rainfall waveform, it is possible to reduce the calculation load in the real-time operation phase and realize real-time inland flooding prediction. [Comparative Example]

[0071] In order to realize a real-time inland flooding prediction system that can determine in real time whether or not flooding occurs above floor level, dynamic calculations using two-dimensional unsteady flow calculations have been used in the past. However, this method requires a large calculation load, making it difficult to predict flooding over a wide area.

[0072] Therefore, in this embodiment, in the preparation phase, the flood depth distribution calculated in advance using a mechanical model is stored, and in the real-time operation phase, a statistical model is adopted that calls up these flood depth distributions according to any rainfall waveform. This makes it possible to reduce the calculation load and calculation time in the real-time operation phase, and to evaluate flooding above floor level. Note that all flood analysis models used in this embodiment do not target external flooding accompanied by levee breaches, but target flooding caused by overflow, overflow, and poor drainage (referred to as "inland flooding" in this specification). [Real-time inland flooding prediction system]

[0073] An overview of the real-time inland flooding prediction method according to embodiment 1 is shown in Fig. 3. The real-time inland flooding prediction system 1000 comprises a flood depth distribution storage unit 21, a rainfall intensity-specific flood volume time change storage unit 23, and a calculation unit 30. [Real-time inland flooding prediction method]

[0074] Hereinafter, each step of the real-time inland flooding prediction method according to the first embodiment will be described with reference to the flowchart in Fig. 4. The real-time inland flooding prediction method can be broadly divided into an advance preparation phase and a real-time operation phase. Here, the advance preparation phase will be described. (Preparation Phase)

[0075] In the advance preparation phase, the inundation depth distribution storage unit 21 and the rainfall intensity-specific flooding volume time change storage unit 23 are prepared. This corresponds to the step of preparing inundation depth distribution data and rainfall intensity-specific flooding volume time change data in step S401 of the flowchart in FIG. 4. The inundation depth distribution storage unit 21 is a component for storing the results of a flood analysis in which the assumed maximum rainfall volume is an external force. The rainfall intensity-specific flooding volume time change storage unit 23 is a component for storing the results of a flood analysis in which a constant rainfall intensity continues. The results of a flood analysis in which the assumed maximum rainfall volume is an external force may also be stored in the rainfall intensity-specific flooding volume time change storage unit 23. For the flood analysis in which the assumed maximum rainfall volume is an external force and the flood analysis in which a constant rainfall intensity continues, a flood analysis model using a conventional two-dimensional unsteady flow calculation can be used.

[0076] Storage media such as fixed disks and semiconductor memories can be used for the inundation depth distribution storage unit 21 and the rainfall intensity-specific water volume change over time storage unit 23. It goes without saying that separate media can be prepared for the inundation depth distribution storage unit 21 and the rainfall intensity-specific water volume change over time storage unit 23, or the data can be stored in separate areas on the same medium. (Flood analysis using expected maximum rainfall as an external force)

[0077] This section explains flood analysis using the expected maximum rainfall as an external force. In accordance with existing guidelines, a rainfall waveform of the maximum expected scale is created, and flood analysis is performed using this as an external force. The input conditions and output values ​​(time-dependent changes in flood depth and flood volume) of this flood analysis are shown in Equations 1 and 2. In this flood analysis, the initial condition is a situation where there is no flood depth, and calculations are performed up to the time when the maximum flood volume occurs.

[0078] [Number 1] TIFF0007629248000002.tif10147[number 2] TIFF0007629248000003.tif11147In the above formula, i,j: Grid number in the x,y direction for flood analysis [-]; dx,dy: Grid size in x and y directions for flood analysis [m]; T A,int : The time interval of the rainfall waveform of the expected maximum rainfall amount, and each time [h] discretized for each output time interval as a result of the flood analysis using the rainfall waveform of the expected maximum rainfall amount as an external force; hs0(i,j): Flood depth distribution when there is no water depth (represents that all water depths at grid number (i,j) are 0.0m); Rmax(T): Expected maximum rainfall at time T [mm / h]; hs A,int (i,j,T int ): Flood depth [m] of the flood analysis result with the expected maximum rainfall as the external force at grid number (i,j) of the flood analysis result output time Tint; f{hs(i,j),R(T)}: A function that returns the flood depth [m] at grid number (i,j) at time T as the flood analysis result when the initial flood depth at grid number (i,j) is hs(i,j) and R(T) is given as the rainfall external force at time T; V A,int (T int ): The flood volume (m2) at the output time Tint as a result of flood analysis using the expected maximum rainfall as an external force 3 ). (Flood depth distribution preservation section 21)

[0079] Based on the above-mentioned flood analysis with the expected maximum rainfall as an external force, T A,int The flood depth distribution output for each time is used, and the flood depth is linearly interpolated in the time direction to obtain the flood depth at the output time T A,int and the next time T A,int+1 Any time T between A Flood depth distribution hs A (i,j,T A ) and flood volume V A (T A) is then calculated. The relationship between the amount of water inundation and the flood depth distribution is stored in the flood depth distribution storage unit 21 so that it can be used in the real-time operation phase. By performing time-consuming mechanical calculations in advance and storing the results in the flood depth distribution storage unit 21 in this way, the necessary flood depth distribution data can be retrieved from the stored results in the real-time operation phase, enabling processing with a low load in a short time, thereby realizing real-time prediction of inland flooding. (Flood analysis when a certain rainfall intensity continues)

[0080] In addition, to grasp the distribution of inundation depth after a given amount of rainfall, a flood analysis is performed when a certain rainfall intensity continues. This is shown in Figure 5. Furthermore, the input conditions and output values ​​(time-dependent changes in inundation depth and water volume) of this flood analysis are shown in Equations 3, 4, and 5.

[0081] In this flood analysis, multiple flood analyses are performed for two types of flooding: high water level (Equation 3, Equation 5) and low water level (Equation 4, Equation 5). In the high water level flood analysis, the initial condition is a state where there is no inundation depth, and the external rainfall force of multiple constant rainfall intensities is applied. The result of the high water level flood analysis is assumed to be a state where there is no inundation volume, and the inundation volume increases over time until the inundation volume becomes steady. On the other hand, in the low water level flood analysis, the initial condition is the inundation depth distribution when the maximum inundation volume occurs at the expected maximum, and the external rainfall force of multiple constant rainfall intensities is applied. The low water level flood analysis is assumed to be a state where the drainage volume of the target basin is greater than the rainfall applied as the external rainfall force. Therefore, the result of the low water level flood analysis is assumed to be a state where the inundation volume decreases over time from the maximum inundation volume at the expected maximum, until the inundation volume becomes steady.

[0082] [Number 3] TIFF0007629248000004.tif10147[number 4] TIFF0007629248000005.tif19147[number 5] TIFF0007629248000006.tif19147In the above formula, T B: Time elapsed in flood analysis when a constant rainfall intensity continues [h]; R const (T): Rainfall intensity [mm / h] of the rainfall waveform at time T when a constant rainfall intensity continues (because a constant rainfall intensity continues, it is a constant value regardless of time T); T p,A :Time when maximum flood volume occurred in flood analysis using maximum expected rainfall as external force [h]; hs B,up (i,j,T,R): Flood depth [m] at grid number (i,j) at time T when the initial condition is no initial water level and a constant rainfall intensity R [mm / h] continues for T hours; V B,up (T,R): Flood volume (m 3 ); hs B,down (i,j,T,R): Flood depth [m] at grid number (i,j) at time T when constant rainfall intensity R [mm / h] continues for T hours, with the initial condition being the flood depth distribution when the maximum flooding volume occurs during the expected maximum rainfall; V B,down (T,R): Flood volume (m 3 ). (Storage section for time-dependent changes in water volume by rainfall intensity 23)

[0083] From the above-mentioned flood analysis when the constant rainfall intensity continues, (V B,up or down (T,R)-T B In the flood analysis for the above-mentioned constant rainfall intensity, the vertical axis of this graph shows the relationship between V B,up or down The same amount of flooding V A (T A ), and even V A (T A ) to hs A (i,j,T A ), these correspondences are organized and stored. This is shown in Figure 6. Here, (V A (T A )orV B,up -T B ) is called the "water level rise graph", and (V A (T A )orVB,down -T B ) is called the "water level fall period graph." These two graphs are stored in the rainfall intensity-specific water volume change over time storage unit 23 so that they can be used in the real-time operation phase. The use of these two graphs will be described in detail in the real-time operation phase below. (Real-time operation phase)

[0084] With the flood depth distribution storage unit 21 and the flood volume time change storage unit 23 by rainfall intensity prepared as described above, i.e. with the flood depth distribution data and the flood volume time change data by rainfall intensity prepared in step S401 of the flowchart in Figure 4, the advance preparation phase ends. From step S402 onwards, the actual operation of the real-time inland water flooding prediction system begins, i.e. the real-time operation phase begins. In the real-time operation phase, the flood depth distribution for an arbitrary rainfall waveform is calculated without using a mechanical model, from the information stored in the "flood depth distribution storage unit 21" and the "flood volume time change storage unit 23" during the advance preparation phase. This is described in detail below. (Arbitrary rainfall waveform input)

[0085] First, in step S402 in FIG. 4, the amount of water in the current time step (w) (V(t)) is compared with the data on the time change of water in the current time step (w) and the data on the time change of water in the current time step by rainfall intensity. Here, the rainfall waveform is given as an input condition for the real-time operation phase. This rainfall waveform is assumed to be from the time when the rain starts to the time when the rain stops, and in this disclosure, the Thiessen division for the basin is not performed, and the average rainfall in the basin is used. Basically, the observed and forecasted rainfall delivered by the Japan Meteorological Agency is used, and the observed rainfall is used from the time when the rain starts to the current time, and the forecasted rainfall is used from the current time to the time when the rain stops. Except for the exceptional processing described later, the graph of the water level increase period (graphs in FIG. 7A and FIG. 8) is used during the rainfall increase period of this rainfall waveform, and the graph of the water level decrease period (graphs in FIG. 7D and FIG. 9) is used during the rainfall decrease period, and the basic processing (FIGS. 8 and 9) for calculating the amount of water in the next time step is performed. (Taking into account delay times)

[0086] In the input of an arbitrary rainfall waveform in the real-time operation phase, if the use of the graphs for the water level increase period and the water level decrease period is switched at the time when the rainfall waveform changes from the rainfall increase period to the rainfall decrease period, the water level will change from increasing to decreasing without considering the delay time (the time difference between the peak of rainfall and the peak of inundation). To solve this problem, even if the rainfall is decreasing, if the water level at the current time step is greater than the steady water level, the water level will be considered to rise, and the graph for the water level increase period will be used to calculate the increase in the water level at the next time step by the method described in the calculation of the water level at the next time (graph in Figure 7B). Similarly, even if the rainfall is increasing, if the water level at an arbitrary time step w is smaller than the steady water level, the water level will be considered to decrease, and the graph for the water level decrease period will be used to calculate the decrease in the water level at the next time step (graph in Figure 7C). It is believed that this makes it possible to consider the change in the water level over time taking into account the delay time. (Initial time progression settings)

[0087] Then, in the input of any rainfall waveform in the real-time operation phase described above, the start of rainfall in the rainfall waveform is set as time step w=0, and the consideration of the real-time operation phase begins. The initial condition is a situation where there is no flood depth (flood volume). In this way, in step S402 of Figure 4, the flood volume (V(t)) at the current time step (w) is compared with the flood volume time change data by rainfall intensity. (Calculation of the amount of water at the next time)

[0088] Next, in step S403, the time change in flood volume (ΔV(R)) up to the time step (w+1) is calculated from the flood volume time change data by rainfall intensity. Here, if rainfall of R [mm / h] is observed or predicted to fall within the step interval Δw from time step w to w+1, the flood volume V(w+1) for the time step is calculated by selectively using the water level rise and fall graphs in consideration of the input of any rainfall waveform in the real-time operation phase described above and the delay time. To do this, first, the water level rise and fall graphs (V B,up or down -T BSpecifically, as shown in Figures 8 and 9, the relationship curve of the recent rainfall intensity R [mm / h] calculated from the flood analysis when a constant rainfall intensity continues in the above-mentioned preparation phase and the time-dependent change data of the amount of water inundated by rainfall intensity is created. const [mm / h] and the next rainfall intensity R const+1 [mm / h](V B,up or down -T B ) relationship is obtained, and the weighted average of the rainfall intensity from these two curves is used to calculate (V B,up or down -T B ) relationship curve.

[0089] Next, in step S404, the water volume (V(t+1)) of the next time step (w+1) is estimated based on the time change (ΔV(R)) of the water volume. Here, the water volume V(w) of the current time step w and (V B,up or down -T B ) Time T at which the curves intersect B Then, at time T B Time T, ΔT hours after B +Δw, time T B +Δw and (V B,up or down -T B ) The water volume at which the curves intersect is determined as the water volume V(w+1) at the time step w+1. (Conversion of flood volume to flood depth distribution)

[0090] Furthermore, in step S405, the flood volume (V(t+1)) at the time step (w+1) is compared with the flood depth distribution data. Then, in step S406, a rainfall intensity curve is calculated from the rainfall waveform in the rainfall intensity curve calculation unit, and in step S407, the Kadoya-Fukushima formula is calculated. In step S408, the point where the rainfall intensity curve intersects with the quasi-linear storage function model is calculated as the flood arrival time. In step S409, the flood arrival time is compared with the flood depth distribution data to calculate the flood volume (V A (T A ) is calculated, and in step S410, multiple neighboring points are averaged to calculate the averaged flood depth distribution for the time step (w+1).

[0091] Specifically, V(w+1) is the same as the flood volume V(w+1). A (T A ) is calculated. Then, the flood depth distribution at the time step w+1 is calculated as hs A (i,j,T A ) as the result. (Update specifications for next time)

[0092] Furthermore, in step S411, it is determined whether or not the calculation of the flood depth distribution averaged over all time points has been completed, and if not, the process returns to step S402 and repeats.

[0093] Specifically, the parameters are updated from the next time step w+1 to the current time step w, and from V(w+1) to V(w), and the calculation of the water volume for the next time in the real-time operation phase described above, the conversion from the water volume to a flood depth distribution, and consideration of updating the parameters for the next time are repeated until the end of the rainfall in the real-time operation phase's rainfall waveform. [Example 1]

[0094] Next, the results of calculating the amount of water and the flood analysis using the real-time inland flooding prediction method according to the first embodiment are shown in Figs. 10 to 15. Here, the target area is the Iio River basin (72.0 km2) located in Tokushima City, Ishii Town, Myosa County, and Yoshinogawa City in Tokushima Prefecture. 2 The Iio River originates in Kamojima-cho, Yoshinogawa City, passes through Ishii-cho, enters Tokushima City, and joins the downstream Ayukui River. Most of the flat land, which accounts for about 70% of the basin, is alluvial plains, mainly consisting of lowlands that are floodplains of the Yoshino River, and the southern part of the basin is made up of mountains and hills. (Preparation Phase Conditions)

[0095] The conditions of the advance preparation phase for applying the method of Example 1 are described below. Both the flood analysis with the expected maximum rainfall as an external force, which will be described later, and the flood analysis in the case of a constant rainfall intensity continuing were performed with a 25m x 25m mesh and a calculation time interval of 0.75 [sec]. The calculations in the advance preparation phase required the same calculation time per calculation as the method of Comparative Example 1 in the comparison of calculation times, which will be described later. If there is a change in land use since the time of the analysis, it is necessary to re-perform the flood analysis in the advance phase accordingly. (Flood analysis using expected maximum rainfall as an external force)

[0096] In the flood analysis with the expected maximum rainfall as the external force, a centralized rainfall waveform was created using the Talbot formula from the 1-hour rainfall and 24-hour rainfall in accordance with the existing guidelines (Ministry of Land, Infrastructure, Transport and Tourism "Water Management and Land Conservation Bureau: Method for setting expected maximum external force for creating inundation forecasts (floods, inland water)" July 2015). A flood analysis was performed with this as the external force, and the results were output in 1-hour increments. (Flood analysis when a certain rainfall intensity continues)

[0097] In the flood analysis when a constant rainfall intensity continues, the rainfall external forces for the continuous constant rainfall intensity of 10, 20, 30, 40, 50, 70, 90, 110, 150, 185, and 220 [mm / h] were created when creating the graph for the water level increase period. This takes into account the range from 10 [mm / h] at which flooding begins to occur to the assumed maximum hourly rainfall of 220 [mm / h]. When creating the graph for the water level decrease period, the rainfall external forces for the continuous constant rainfall intensity of 0, 3, 6, 9, 12, 15, and 20 [mm / h] were created. This is because when flooding occurs under the initial conditions, if rainfall of 20 [mm / h] or more continues, the drainage capacity of the target watershed is exceeded and the water level rises. Flood analysis was performed using these external forces, and the results were output in one-hour increments. (Application results and verification)

[0098] The heavy rainfall (429.5 mm / day) on September 20, 2011, which was recorded at the Tokushima Local Meteorological Observatory as having a high daily rainfall in recent years, was used as an external force, and the results of flood analysis were compared between the method of Comparative Example 1 and the method of Example 1. In this heavy rainfall, although the flooded areas in the Iio River basin could not be identified, flooding above floor level in 34 houses was confirmed according to flood statistics (Water Management and Disaster Management Bureau, Ministry of Land, Infrastructure, Transport and Tourism, "Flood Statistics 2011, Basic Tables of Flood Statistics, Basic Tables of Flood Statistics on General Assets, etc." April 2015). (Changes in water volume over time)

[0099] When comparing the changes in the amount of water stored over time in the method of Comparative Example 1 and the method of Example 1, it was confirmed that the increase or decrease over time and the amount of water stored were roughly consistent, as shown in Figure 10. (Maximum flood depth and maximum flow flux)

[0100] Next, the method of Example 1 was verified based on the maximum flood depth to evaluate flooding above the floor and the maximum value of flow flux, which has the same importance as flood depth in a planar two-dimensional unsteady flow analysis. When comparing the maximum flood depth of the method of Comparative Example 1 and Example 1, it was confirmed that most of the flood depth values ​​were plotted near the y = x line, as shown in Figure 11.

[0101] In considering the maximum flood depth and maximum flow flux, the time of occurrence of maximum flood depth and flooding above floor level, and the flood depth for each hour on the hour, which will be described later, the meshes that were generated using the method of Comparative Example 1 had a flood depth of 0.5 [m] or more, which is considered to be an index of flooding above floor level (National Land Development Technology Research Center, "Guide to Inland Water Treatment Plan Formulation," p. 114, February 1995). The correlation coefficient (r) was 0.997, and the two methods had a very strong correlation. The average difference in maximum flood depth between the two methods was 0.031 [m]. This is 4% of the average maximum flood depth of 0.757 [m]. The average difference in flood depth between the two methods was 6% of 0.5 [m], which is considered to be an index of flooding above floor level, and it is considered that the maximum flood depth distribution can be evaluated. In the area where a flood depth of 0.5 [m] occurred using the method of Comparative Example 1 (1.30 km 2 ) 82% of the locations where flooding depth of 0.5 [m] occurred even with the method of Example 1 (1.07 km 2Therefore, it is considered that the method of Example 1 can evaluate whether or not flooding above the floor has occurred.

[0102] When comparing the maximum flow fluxes (FIG. 12) of the method of Comparative Example 1 and the method of Example 1, it can be seen that most of the flow flux values ​​are plotted near the y=x line, and the flow flux of the method of Comparative Example 1 is large. This is thought to be due to the fact that the flow velocity of the method of Comparative Example 1 is higher than that of Example 1 due to the problem that the inundation depth of the method of Example 1 is located upstream of the location where the inundation depth is distributed in the method of Comparative Example 1, which will be described later as the application range of the first embodiment. On the other hand, in the area of ​​FIG. 12 where the maximum flow flux is small, the method of Example 1 has a higher maximum flow flux than the method of Comparative Example 1. The flood analysis model of the first embodiment is a model in which the amount of rainfall is added to the water depth, and the flooded water flows down to a place with a low elevation. The flow flux of this area was confirmed in a place where the elevation is relatively higher than the surrounding area and the water depth is smaller than the surrounding area. In such a place, the flow velocity of the method of Comparative Example 1 is higher than that of Example 1, so the water flows downstream before the water depth becomes large. On the other hand, the flow velocity in the method of Example 1 is smaller than that in the method of Comparative Example 1, so the water does not flow downstream until the water depth becomes large. Since the flow flux is calculated by multiplying the water depth and the flow velocity, it is thought that although the flow velocity in the method of Example 1 is smaller than that in Comparative Example 1, the water depth is greater than the flow velocity, so the maximum flow flux calculated in the method of Example 1 is greater than that in Comparative Example 1. The correlation coefficient (r) was 0.914, and the two methods had a very strong correlation. The average difference in flow flux between the two methods was 0.038 [m 2 / s], which is the average maximum flow flux of 0.118 [m 2 / s], which indicates a tendency for the reproducibility to be less than that of the maximum flood depth. (Maximum flood depth and time when flooding above floor level occurred)

[0103] The average time when maximum flood depth occurred obtained for each mesh was 22:59 on the 20th using the method of Comparative Example 1, and 23:33 using the method of Example 1. It was found that Comparative Example 1 was 34 minutes earlier than Example 1. Furthermore, the average time when flooding above the floor occurred obtained for each mesh was 14:00 on the 20th using the method of Comparative Example 1, and 14:35 using the method of Example 1. It was found that Comparative Example 1 was 35 minutes earlier than Example 1. (Flood depth every hour on the hour)

[0104] In the above-mentioned maximum flood depth and time of flooding above floor, the time progresses slower in Example 1 than in Comparative Example 1, so the flood depth at an arbitrary time was examined. The 72 hours from 0:00 on September 19 to 24:00 on the 21st were output in one-hour increments, and the flood depths of both methods were plotted at the locations where the method of Comparative Example 1 caused a flood depth of 0.5 [m] or more at these times (Fig. 13). Most of the flood depth values ​​were plotted below the y = x line, and it was confirmed that the flood depth of Example 1 tends to be smaller. This is thought to be because the time progresses slower in Example 1 than in Comparative Example 1 in terms of the maximum flood depth and time of flooding above floor, so the method of Example 1 collects floodwater later than the method of Comparative Example 1, and the flood depth is calculated to be smaller. Therefore, the correlation coefficient (r) was 0.761, which is smaller than the correlation when the maximum flood depth was examined. In addition, the average difference in the flood depth of both methods was 26.3 [cm]. This is 53% of 0.5m, which is the index for flooding above floor level, and it is difficult to say that the two methods are consistent. Therefore, it is considered impossible to calculate the flood depth at any given time. (Flood depth distribution at 11pm on the 20th)

[0105] Comparing the flood depth distribution at 11pm on the 20th using the method of Comparative Example 1 (Fig. 14) with that using the method of Example 1 (Fig. 15), it was confirmed that they were roughly consistent. Comparing these figures, it can be seen that, albeit slightly, the flood depth is greater in Comparative Example 1 than in Example 1 at the locations indicated by the arrows in the enlarged views of both figures. Comparing the average values ​​of the time when maximum flood depth occurred for both methods with 11pm on the 20th, the difference is one minute in Comparative Example 1, while the difference in Example 1 is 33 minutes, which is a larger time difference than Comparative Example 1. This is thought to be the reason why the flood depth is greater in Comparative Example 1 than in Example 1. (Scope of application of Example 1)

[0106] Although the method of Example 1 can calculate the change in flood volume over time, it cannot grasp the flood depth at any given time. This is thought to be because when converting from flood volume to flood depth distribution ("(7) Conversion" in the flood depth distribution storage unit 21 in Fig. 3), it is converted into a flood depth distribution during the water level rise period calculated with the expected maximum rainfall as an external force, so the flood depth is located upstream of the location where the flood depth is distributed in Comparative Example 1. However, in Example 1, although the progression of time is slower than in Comparative Example 1, if calculations are made from the occurrence of flooding to its elimination, it is thought that the flood water will eventually reach the downstream end of the basin, and it is possible to obtain the same maximum flood depth and location of above-floor flooding as in Comparative Example 1. (Comparison of calculation time)

[0107] The specifications of the personal computer used for the calculation were Intel (registered trademark) Core (trademark) i7-8700K CPU @ 3.70 GHz, and 64.00 G of RAM. In the flood analysis of Comparative Example 1 (target time 72 [h]), the calculation time was 25 hours and 6 minutes, and in the real-time operation phase of Example 1, the calculation time was 1 minute and 12 seconds. [Example 2]

[0108] As described above, the real-time inland flood prediction system according to the first embodiment utilizes a statistical model, thereby realizing inland flood prediction that would be difficult to achieve in real time using calculation processing based on a mechanical model during operation. Meanwhile, in the first embodiment, the only index for evaluating the inundation depth distribution was the flood volume, which is a property of the continuous equation. Therefore, in the second embodiment, a real-time inland flood prediction method is described below, which uses the flood arrival time, which is a property of the equation of motion, in addition to the flood volume, to evaluate the inundation depth distribution. Specifically, the system is separated into a preparation phase and a real-time operation phase, and in the real-time operation phase, the inundation depth distribution is called up taking into account two indexes, the flood volume and the flood arrival time. This is described in detail below. (Flood arrival time)

[0109] An outline of the real-time inland flooding prediction method according to the second embodiment is shown in Figure 16. Steps (1) to (6) are the same as those of the first embodiment shown in Figure 3. In step (1), the inundation depth distribution storage unit 21 and the rainfall intensity-specific flooding amount time change storage unit 23 are prepared in advance. In step (2), the flooding amount (V(t)) of the current time step (w), which is a known amount, is transferred to the rainfall intensity-specific flooding amount time change storage unit 23. In step (3), the amount of change in the flooding amount over time (ΔV(R)) when an arbitrary amount of rainfall (R) occurs is calculated. In step (4), the flooding amount (V(t+1)) of the next time step (w+1) is estimated. In step (5), the flooding amount (V(t+1)) of the next time step (w+1) is passed to the inundation depth distribution storage unit 21. In step (6), the flooding amount (V A (T A The difference from the first embodiment is that the flood arrival time calculation unit 33 adds the steps (7) and (8).

[0110] In this disclosure, the flood arrival time is calculated using tc: flood arrival time [hr], which is listed on p. 94 of "Revised New Edition Ministry of Construction River Erosion Control Technical Standards (Draft) Commentary Survey Edition" edited by the Japan River Association and supervised by the River Bureau of the Ministry of Construction. According to this document, tc is determined from the empirical formula for flood arrival time by Kadoya et al. and the actual effective rainfall intensity curve.

[0111] tc=C·A 0.22·re -0.35 however, re: Maximum average effective rainfall intensity during rainfall duration [mm / h]; A: Basin area [km 2 ]; tc: flood arrival time; C: constant determined by land use type; (In this disclosure, the above formula is called the "Kadoya-Fukushima formula.") (Flood arrival time calculation unit 33)

[0112] The flood arrival time calculation unit 33 calculates the flood arrival time (T P_A ) is calculated. Flood arrival time is calculated in accordance with the quasi-linear storage function model in Chapter 3, Section 2, p. 19 of the Ministry of Land, Infrastructure, Transport and Tourism's Water Management and Disaster Management Bureau's "River Erosion Control Technical Standards: Survey Edition" (June 2012). Finally, the flood depth is calculated by interpolating nearby known values ​​calculated in advance from the two indicators of flood volume (6) and flood arrival time (8), and the distribution of these flood depths is displayed. (Preparation Phase)

[0113] In the advance preparation phase, the flood analysis with the expected maximum rainfall as an external force stored in the inundation depth distribution storage unit 21 is different from that in the first embodiment. In addition to the flood analysis with the expected maximum rainfall ("Method of setting the expected maximum external force for creating inundation assumptions (floods, inland water) etc., Water Management and Land Conservation Bureau, Ministry of Land, Infrastructure, Transport and Tourism: p. 10 (July 2015)) as an external force, the inundation depth distribution storage unit 21 also performed flood analysis with 2-year, 10-year, 50-year, 100-year, and 500-year probability rainfall as an external force, and performed flood analysis with front-concentrated, central-concentrated, and rear-concentrated rainfall waveforms for these six cases of probability years. The results of these 18 cases in total (= 6 rainfall probabilities x 3 rainfall waveforms) (hereinafter referred to as "results of 18 cases of expected maximum rainfall, etc.") of the inundation volume and flood arrival time are stored in the inundation depth distribution storage unit 21. The unit 23 for storing the flood analysis results when a certain rainfall intensity continues and the time variation of the amount of flooding by rainfall intensity, stores the time variation of the amount of flooding, as in the first embodiment. (Real-time operation phase)

[0114] In the real-time operation phase, the flood depth distribution for an arbitrary rainfall waveform is calculated without using a dynamic model, based on the information stored in the flood depth distribution storage unit 21 and the rainfall intensity-specific water volume time change storage unit 23 during the advance preparation phase. This procedure is shown in the flowchart of Figure 17. Steps S1701 to S1704 are the same as in the first embodiment, and detailed explanations will be omitted. Steps S1705 to S1707 will be explained below. (Calculation of flood arrival time)

[0115] In step S1705 of the flowchart in Figure 17, the flood arrival time is calculated. The flood arrival time is the value where the Kadoya-Fukushima formula intersects with the rainfall intensity curve, in accordance with the quasi-linear storage function model (Ministry of Land, Infrastructure, Transport and Tourism, Water Management and Land Conservation Bureau, "River Erosion Control Technical Standards Survey Edition," Chapter 3, Section 2, p. 19 (June 2012)). The obtained result is passed from the flood arrival calculation unit to the inundation depth distribution storage unit 21. In addition, in the quasi-linear storage function, the rainfall intensity is converted to effective rainfall and used in both the rainfall intensity formula and the quasi-linear storage function model. (The flood depth distribution is calculated from the amount of water and the time it takes for the flood to reach the area.)

[0116] Next, in step S1706 of the flowchart in Figure 17, the flood depth distribution is calculated from the flood volume and flood arrival time. The concept of calculating flood depth at the next time step is shown in Figure 18. In this figure, the plots at the next time step are cross-hatched, the results of 18 cases of assumed maximum rainfall, etc. are plotted in light gray, the results of 18 cases of assumed maximum rainfall, etc. closest to the cross-hatched in each quadrant are plotted in dark gray, and the flood depth distribution without flood volume is plotted in diagonal lines. As shown in this figure, the results of 18 cases of assumed maximum rainfall, etc. at any time and any mesh are plotted in light gray on a graph with the flood volume on the vertical axis and the flood arrival time on the horizontal axis. Here, because the units of the vertical and horizontal axes are different, the distance between the vertical and horizontal axes is adjusted by normalizing (converting to deviation values). Here, to find the flood depth at the next time step, the vertical and horizontal axes are set with the plot to be found as the origin, and the results of 18 cases of the expected maximum rainfall, etc., that are closest to this plot are selected in quadrants I to IV. The results of the nearest neighbors can be found because normalization makes it possible to calculate the distance on a two-dimensional plane. Then, the ratio of the distance (R2:R3) to the point where the line connecting the nearest points in quadrant II and quadrant III intersects with the horizontal axis is found. Similarly, the ratio of the distance (R1:R4) to the point where the line connecting the nearest points in quadrant I and quadrant IV intersects with the horizontal axis is found. The ratio of the distance (R23:R14) to the origin is found between the point where the line connecting the nearest points in quadrants II and III intersects with the horizontal axis and the point where the line connecting the nearest points in quadrants I and IV intersects with the horizontal axis. Using the following equation 6, the nearest points in each of the four quadrants are interpolated to find the flood depth at the next time step. These procedures are repeated for each mesh to calculate the flood depth distribution. If there are no plots of the calculation results for 18 cases of assumed maximum rainfall, etc. in quadrants III and IV, plots are placed in the area where the flood arrival time is equivalent to the crosshatch and there is no flooding (flood depth distribution for flood depth of 0.0 [m]) (diagonal lines in Figure 18), and the dark gray plots of the calculation results for 18 cases of assumed maximum rainfall, etc. that exist in quadrants III and IV closest to the crosshatch plot at the next time step are substituted. Example 2 targets inland flooding, and level flooding on flat land is assumed.Therefore, because it is more important to evaluate the water volume in detail than the water surface gradient, we first evaluated the flood volume, which is an index of water volume, and then evaluated the flood arrival time, which is an index of water surface gradient.

[0117] [Number 6] TIFF0007629248000007.tif34147

[0118] where i,j: latitudinal and longitudinal grid numbers for inundation analysis [-]; T A, : Time interval of rainfall waveform of expected maximum rainfall and elapsed time [h] corresponding to the time step of the flood analysis using 18 cases of rainfall waveforms such as expected maximum rainfall as external force; R2:R3:Ratio of the distance to the point where the line connecting the nearest points in quadrant II and quadrant III intersects the horizontal axis [-]: R1:R4: The ratio of the distance to the point where the line connecting the nearest points in quadrant I and quadrant IV intersects with the horizontal axis [-]; R23:R14:The ratio of the distance from the origin to the point where the line connecting the neighboring points in quadrants II and III intersects with the horizontal axis to the point where the line connecting the neighboring points in quadrants I and IV intersects with the horizontal axis [-]; hs(i,j,T A ): Flood depth of grid number (i,j) at the next time step [m]; hs1~4(i,j,T A ): Flood depth [m] of 18 cases of the expected maximum rainfall etc. nearest to the plot of the time step in quadrants I to IV of grid number (i,j) at the time step; It is. (Update specifications for next time)

[0119] Finally, in step S1707 of the flowchart in Fig. 17, the specifications for the next time are updated. The specifications are updated from the next time step w+1 to the current time step w, and from V(w+1) to V(w), and the considerations in steps S1704 to S1707 are repeated until the end of the rainfall in the rainfall waveform in step S1701. (Application and Verification of Example 2) (Target area)

[0120] In Example 2, the Iio River basin (72.0 km2) located in Tokushima City, Ishii Town, Myosai County, and Yoshinogawa City in Tokushima Prefecture 2 The Iio River originates in Kamojima-cho, Yoshinogawa City, passes through Ishii-cho, enters Tokushima City, and joins the downstream Ayukui River. Most of the flat land, which accounts for about 70% of the basin, is alluvial plains, mainly consisting of lowlands that are floodplains of the Yoshino River, and the southern part of the basin is made up of mountains and hills. (Preparation Phase Conditions)

[0121] In the two flood analyses described below, the calculation mesh was 25[m]×25[m] and the calculation time interval was 0.75[sec]. The calculation time required for the preparation phase was the same as in Example 1. (Flood analysis of 18 cases including expected maximum rainfall)

[0122] The expected maximum rainfall was calculated using the Talbot formula from the 1-hour rainfall and 24-hour rainfall in accordance with the guidelines of the Ministry of Land, Infrastructure, Transport and Tourism (Ministry of Land, Infrastructure, Transport and Tourism Water Management and Land Conservation Bureau "Method of setting expected maximum external force for creating inundation assumptions (floods, inland water)" p.10 (July 2015)). Other probable rainfall (2-year, 10-year, 50-year, 100-year, 500-year) was calculated using the rainfall intensity formula described in the Tokushima Prefecture guidelines (Tokushima Prefecture Land Development Department "Rivers and Coasts of Tokushima Prefecture" p.143 (September 2022)). Flood analysis was performed for these six probable rainfall cases using front-concentration rainfall waveforms, central-concentration rainfall waveforms, and rear-concentration rainfall waveforms as external forces, and the results were output in hourly increments. (Flood analysis when a certain rainfall intensity continues)

[0123] In creating the graph for the water level increase period, rainfall external forces were created when constant rainfall intensities of 10, 20, 30, 40, 50, 70, 90, 110, 150, 185, and 220 [mm / h] continued, as in Example 1. This takes into consideration the range from 10 [mm / h] at which flooding begins to occur to the assumed maximum hourly rainfall of 220 [mm / h] (Ministry of Land, Infrastructure, Transport and Tourism Water Management and Land Conservation Bureau, "Method of Setting Assumed Maximum External Forces for Preparing Inundation Expectations (Floods, Inland Water)," p. 10 (July 2015)). On the other hand, in creating the graph for the water level decrease period, rainfall external forces were created when constant rainfall intensities of 0, 3, 6, 9, 12, 15, and 20 [mm / h] continued, as in Example 1. This is because, in a situation where flooding occurs under the initial conditions, if rainfall of 20 [mm / h] or more continues, the drainage capacity is exceeded in the target basin, causing the water level to rise. The difference from the first embodiment is that the results of the flood analysis using these external forces are output in 5-minute intervals. This improves applicability when sudden time changes occur in the rainfall waveform during the real-time operation phase. (Application results and verification)

[0124] Here, the results of flood analysis using a two-dimensional unsteady flow calculation on a plane with the heavy rainfall of September 20, 2011 (429.5 mm / day), one of the highest daily rainfalls in recent years, as an external force (Miyoshi Manabu, Tamura Takao, Muto Hironori, Aki Hirosuke, "An inland water analysis model that creates wide-area and detailed inundation depth distributions using open data maintained on a national scale", Journal of Hydraulic Engineering, Vol. 74, No. 4, pp. I 1321-I 1326 (2018)) are used as Comparative Example 2, and the results output by Example 2 (time change in water volume, maximum inundation depth, maximum flow flux, etc.) are compared. In Example 1, we focused on the flood depth above the floor (0.50 m or more) and verified Example 2. In Example 2, when evaluating flooding above the floor, we decided to add to the verification scope the display of a color scheme for flooding ranks that conforms to the Guidelines for Creating Flood Hazard Maps (Ministry of Land, Infrastructure, Transport and Tourism, Water and Disaster Management Bureau, "Guidelines for Creating Flood Hazard Maps" p.36 (revised December 2021)).

[0125] Although the flooded areas in the Iio River basin could not be identified during the heavy rainfall on September 20, 2011 (429.5 mm / day), flooding above floor level in 34 houses was confirmed according to flood statistics (Water Management and Land Conservation Bureau, Ministry of Land, Infrastructure, Transport and Tourism, "Flood Statistics 2011, Basic Table of Flood Statistics, Basic Table of Flood Statistics for General Assets, etc." April 2015). In Example 2 and Comparative Example 2, the map information digital elevation model 5 [m] was used for the flood analysis. (Changes in water volume over time)

[0126] When comparing the changes in the amount of water stored over time in Comparative Example 2 and Example 2, it was found that the increase / decrease over time and the amount of water stored were roughly consistent, as shown in Figure 19. (Maximum flood depth and maximum flow flux)

[0127] To evaluate flooding above floor level, we verified Example 2 based on the maximum flood depth and the maximum value of flow flux, which has the same importance as flood depth in two-dimensional unsteady flow analysis. The flow flux was calculated by multiplying the flood depth by the flow velocity, which was calculated in the same way as the flood depth.

[0128] Comparing the maximum flood depths of Comparative Example 2 and Example 2, it was found that most of the flood depth values ​​were plotted near the y=x line as shown in Figure 20. In addition, when considering the maximum flood depth and maximum flow flux, the time when the maximum flood depth occurred and the time when flooding above the floor occurred, and the flood depth every hour on the hour, which will be described later, the meshes generated by the method of Comparative Example 2 were targeted for a flood depth of 0.5 [m] or more, which is considered to be an index of flooding above the floor (National Land Development Technology Research Center, "Guide to the formulation of inland water treatment plans," p. 114 (February 1995)). The correlation coefficient (r) was 0.994, and the two methods had a very strong correlation. The average difference in maximum flood depth between the two methods was 0.030 [m], and Example 2 had an error of 3.81% compared to Comparative Example 2. In addition, the average difference in flood depth between the two methods was 6% of 0.5 [m], which is considered to be an index of flooding above the floor, and it is considered that the maximum flood depth distribution can be evaluated. The area where flooding depth of 0.5 m or more occurred using the method of Comparative Example 2 (2.66 km 2 ) in Example 2, 91% of the locations where flooding depths of 0.5 m or more occurred (2.41 km 2Therefore, it is considered that Example 2 can evaluate whether or not flooding above the floor has occurred.

[0129] In addition, when comparing the maximum flow fluxes of Comparative Example 2 and Example 2, it was found that most of the flow flux values ​​were plotted near the y=x line as shown in FIG. 21. The correlation coefficient (r) was 0.982, and the two methods had a very strong correlation. The average difference between the maximum flow fluxes of the two methods was 0.014 [m 2 / s], and Example 2 had an error of 27.77% compared to the method of Comparative Example 2. Although Example 1 confirmed a tendency for the flow flux of Comparative Example 2 to be large, Example 2 took the flood arrival time into account when calculating the inundation depth distribution, and therefore the tendency for the flow flux of the method of Comparative Example 2 to be large disappeared. From these, it is considered that the maximum flow flux obtained the same reproducibility as the maximum inundation depth. This improvement in reproducibility is considered to be due to the elimination of the difference in time progression by taking the flood arrival time into account. In other words, it is difficult to consider that the difference was eliminated due to spatial factors caused by using the map information digital elevation model of 5 [m] instead of 10 [m] when creating the ground elevation model. (Time of maximum flood depth and time of flooding above floor level)

[0130] The average value of the time when the maximum flood depth occurred obtained for each mesh was 20:35 on the 20th in the method of Comparative Example 2, and 20:27 in the method of Example 2. Comparative Example 2 was 8 minutes earlier than Example 2, and it can be seen that this is a time difference that does not cause problems in advising residents to evacuate in advance on the day of heavy rain. In addition, the average value of the time when flooding occurred above floor level obtained for each mesh was 12:38 on the 20th in Comparative Example 2, and 12:23 in Example 2. Comparative Example 2 was 15 minutes earlier than Example 2, and it can be seen that this is a time difference that does not cause problems in advising residents to evacuate in advance on the day of heavy rain. Note that these times were calculated by outputting the results calculated in minutes and taking the spatial average of those. (Flood depth every hour on the hour)

[0131] The above-mentioned maximum flood depth occurrence time and the time when flooding above the floor occurred suggest that the time progression of Comparative Example 2 and Example 2 may be equivalent, so here we verified the flood depth at an arbitrary time. Figure 22 shows a graph in which the flood depths of both methods at the locations where a flood depth of 0.5 [m] or more occurred in Comparative Example 2 at these times were plotted, outputting the 72 hours from 00:00 on September 19th to 24:00 on the 21st at one-hour intervals on the hour. As shown in this figure, although there is a tendency for the plots to be plotted on the y = x line, it was found that there is more variation on the y = x line than in the plot when comparing the maximum flood depths in Figure 21. However, the tendency for the flood depth to be smaller in Example 2 than in Comparative Example 2, as in Example 1, has been eliminated. The correlation coefficient (r) was 0.934, and the two methods had a very strong correlation. The average difference in flood depth for each hour on the hour between both methods was 0.121 [m], which was an error of 13.72% compared to the method of Comparative Example 2. Furthermore, the average difference in flood depth between the two methods was 24% of 0.5 m, which is considered an indicator of flooding above floor level, and is thought to capture the trend of changes in flood depth over time. (Accuracy rate of color schemes in flood hazard maps)

[0132] In the "Guide to Creating Flood Hazard Maps" by the Water Management and Land Conservation Bureau of the Ministry of Land, Infrastructure, Transport and Tourism, the flood depths are set at 0.3, 0.5, 1.0, 3.0, 5.0, 10.0, and 20.0 [m] as thresholds, and the color scheme for displaying the flood depth changes at these thresholds. The ratio of the color schemes that are the same in both Comparative Example 2 and Example 2 is defined as the color scheme accuracy rate in Example 2. In both Comparative Example 2 and Example 2, cases where there is no flood depth are excluded from the color scheme accuracy rate. The color scheme accuracy rate when the flood depth distribution is displayed every hour on the hour is shown in the graph in Figure 23. As shown in this figure, the color scheme accuracy rate is small at the beginning and end of rain when flooding is minor, but the color scheme accuracy rate increases as the amount of water inundation increases with the amount of rain. This is thought to be due to the fact that, although the width of the color scheme threshold is set finely when flooding is minor, the width of the color scheme threshold increases as the flooding depth increases. On the other hand, there were a total of 1,994,310 meshes in which flood depth occurred every hour on the hour during the calculation time (72 hours). Of these, the color schemes of 1,567,580 meshes were correct, and the color scheme accuracy rate throughout the calculation time (72 hours) was 79%. Therefore, when the color scheme according to the above-mentioned "Guide to Preparing Flood Hazard Maps" is used, it is considered that the flood depth distribution can be roughly evaluated along with the elapsed time (especially when the flood depth is large). In addition, when the color scheme accuracy rate of Comparative Example 2 and Example 2 was calculated for the maximum flood depth distribution, the color scheme of 69,474 meshes out of the 74,357 meshes in which flood depth occurred was correct. Therefore, the color scheme accuracy rate was 93%, and it is considered that the maximum flood depth distribution can be evaluated when the color scheme according to the "Guide to Preparing Flood Hazard Maps" is used.

[0133] The maximum flood depth distribution by the method of Comparative Example 2 when displayed with the color scheme according to the "Guide to Creating Flood Hazard Maps" is shown in FIG. 24, and the maximum flood depth distribution by Example 2 is shown in FIG. 25. From these figures, it can be seen that both methods are roughly consistent. Note that in FIG. 24 and FIG. 25, in order to make the coloring easier to see, the threshold values ​​of the color scheme according to the "Guide to Creating Flood Hazard Maps" are the same, but the colors are different from those in the "Guide to Creating Flood Hazard Maps". Looking at the enlarged views of FIG. 24 and FIG. 25, it can be seen that the flood depth tends to be large in fields and rivers that are one level lower than houses, and the color scheme of the flood depth is also similar. Therefore, it is considered appropriate to follow the color scheme of the "Guide to Creating Flood Hazard Maps" as a display method for evaluating flooding above floor level in the future. Meanwhile, the difference in maximum flood depth between the two methods is shown in FIG. 26. From this figure, the difference is small in the case of storage-type flooding that occurs in fields, but tends to be large in the case of runoff-type flooding that occurs in valleys and rivers. This is thought to be because, in the case of storage flooding, if the amount of water stored is the same, the water depth will match between the two methods, but in the case of flow-down flooding, it is difficult to make the relationship between water depth and flow velocity match between the two methods. (Accuracy rate of above-floor flooding locations)

[0134] In the method of Comparative Example 2, in the locations where flooding occurs above the floor level (flood depth of 0.5 m or more (National Land Development Technology Research Center, "Guide to Inland Water Treatment Plan Formulation," p. 114 (February 1995)), the accuracy rate of flooding above the floor level is also defined in Example 2 as the accuracy rate of flooding above the floor level. Note that in the method of Comparative Example 2, cases where flooding does not occur are not included in the accuracy rate of flooding above the floor level. Figure 27 shows the accuracy rate of flooding above the floor level for each hour on the hour during the calculation period (72 hours). From this figure, it can be seen that it was 97% or more immediately after the occurrence of flooding. There were a total of 2,004,099 meshes where flooding above the floor level occurred at each hour on the hour. Of these, 1,981,022 meshes had flooded locations above the floor level, and the accuracy rate of flooding above the floor level throughout the calculation period (72 hours) was 99%. Therefore, it is considered possible to evaluate flooding above the floor level over time. (Color combination accuracy rate at the time when flooding above floor level begins)

[0135] The distribution of flood depth at the time when flooding above floor level started to occur (17:00 on the 19th) using the method of Comparative Example 2 is shown in Figure 28, and the distribution of flood depth using the method of Example 2 is shown in Figure 29. Comparing the two methods, it can be seen that the locations where flooding depth occurs are different in Example 2 from Comparative Example 2. This is thought to be because, as shown in Figure 18, the distribution of flood depth in Example 2 was calculated by interpolating the results of inundation analysis for 18 cases of expected maximum rainfall, etc., so when the flooding is small and the areas flooded before water collection are scattered, the locations where flooding depth occurs are different between the two methods, resulting in a lower accuracy rate for flooding above floor level. (Comparison of calculation time)

[0136] The specifications of the personal computer used in the calculations of Example 2 and Comparative Example 2 were Intel (registered trademark) Core (trademark) i7-8700K CPU @ 3.70 GHz, and 64.00 G of RAM. In the flood analysis of Comparative Example 2 (target time 72 h), the calculation time was 25 hours and 14 minutes, and in the real-time operation phase of Example 2, the calculation time was 2 minutes and 49 seconds. Therefore, it is considered that Example 2 can reduce the calculation time to 0.19%.

[0137] As described above, in Example 2, by realizing a real-time inland water flooding prediction method that evaluates the inundation depth distribution from the flood arrival time in addition to the flood volume, it becomes possible to evaluate the inundation depth distribution from the flood volume and flood arrival time. In Example 1, an attempt was made to evaluate the inundation depth distribution using only the flood volume as an indicator, but it was not possible to calculate the inundation depth at any time or the time when inundation above floor level will occur. In Example 2, when the inundation depth distribution was evaluated using the flood arrival time as an indicator in addition to the flood volume, it was thought that the trend of the time change in the inundation depth at any time was captured. Therefore, it was thought that the inundation depth distribution could be evaluated from the flood volume and flood arrival time.

[0138] Moreover, in Example 2, it is possible to evaluate the time change in flood volume, the maximum flood depth distribution, the time when flooding above floor level occurred, and the location where flooding above floor level occurred, and it is possible to evaluate flooding above floor level. Furthermore, when the flood depth distribution at an arbitrary time was displayed using the color scheme according to the "Guide to Creating Flood Hazard Maps," the accuracy rate of the color scheme was 79%. Thus, it is considered appropriate to follow the color scheme in the said guide for the calculation of flooding above floor level and the display of flood depth distribution in Example 2.

[0139] Furthermore, according to the second embodiment, the calculation time was reduced from 25 hours 14 minutes to 2 minutes 49 seconds (0.19%). In particular, the accuracy of the method for predicting inland flooding in real time is considered to be due to the lead time required for evacuation in local communities. It is considered that the accuracy of information distribution and calculation results at time intervals that allow the lead time to be appropriately evaluated is necessary. On the other hand, in the second embodiment, the time change of the flooding amount, the distribution of maximum flood depth, the time when flooding above floor level occurs, and the location where flooding above floor level occurs can be evaluated. However, in the second embodiment, in a wide-area watershed where there is rainfall in the upstream area but not in the downstream area, it may not be possible to evaluate the phenomenon in which flooding in the upstream area flows down to the downstream area and the downstream area is flooded. Therefore, in a wide-area watershed, it is preferable to perform a fine mesh flood analysis in advance, such as the 18 cases of the expected maximum rainfall in the second embodiment. Then, flood analysis is performed for a wide watershed using a coarse mesh that does not interfere with real-time calculations, and the water volume and flood arrival time for each coarse mesh are calculated in real-time on an operational face. By calling up the inundation depth distribution for finer mesh from these two indicators, it becomes possible to predict inland flooding even in a wide watershed where there is rainfall in the upstream basin but no rainfall downstream. [Industrial Applicability]

[0140] The real-time inland flooding prediction system, real-time inland flooding prediction device, real-time inland flooding prediction method, real-time inland flooding prediction program, and computer-readable recording medium and device storing the program can be used to perform inland flooding prediction. This makes it possible to accurately predict the occurrence of inland flooding in urban areas and the like when a super typhoon or linear precipitation band occurs, and can be useful for creating evacuation routes, etc. [Explanation of symbols]

[0141] 1000, 1000'...Real-time inland flooding prediction system 100...Real-time inland flood prediction device 10…Data acquisition section 20...Data storage unit 21…Flood depth distribution preservation section 22…Predicted inundation depth distribution model storage section 23…Storage section for time-dependent changes in water volume by rainfall intensity 26...Rainfall intensity curve storage section 27...Quasi-linear storage type function model storage section 30...Arithmetic section 31...Flooding amount calculation section 32...Database access section 33…Flood arrival time calculation section 34…Predicted flood depth distribution model extraction section 35...Quasi-linear storage type function model calculation section 40…Display section 50...Operation unit CL: Client device WS...Water level sensor RT…Repeater

Claims

1. A real-time inland flooding prediction system that predicts in real time inland flooding in a predicted area caused by rainfall under conditions in which a predicted area is divided into small drainage basins, uniform rainfall is assumed in each small drainage basin, and a pre-calculated flood depth distribution is assumed to occur for each small drainage basin at each position included in the predicted area, A data acquisition unit that acquires observed rainfall and predicted rainfall; a water volume calculation unit for calculating a current water volume in real time based on the observed rainfall acquired by the data acquisition unit, the predicted rainfall, and an initial water level; a predicted flood depth distribution model storage unit for storing a plurality of predicted flood depth distribution models calculated in advance for each flooding amount in a predicted area; A flood arrival time calculation unit for calculating a flood arrival time that is expected to reach a forecast area based on rainfall intensity; a predicted inundation depth distribution model extraction unit for correcting the flood volume calculated in real time by the flood volume calculation unit to a corrected flood volume taking into account the flood arrival time calculated by the flood arrival time calculation unit, and extracting a predicted inundation depth distribution model corresponding to the corrected flood volume from among multiple predicted inundation depth distribution models stored in the predicted inundation depth distribution model storage unit; a display unit for displaying the predicted inundation depth distribution model extracted by the predicted inundation depth distribution model extraction unit; and A real-time inland flooding prediction system.

2. The real-time inland water flooding prediction system according to claim 1, further comprising: a rainfall intensity-based flood volume time change storage unit that stores the results of a relationship between the flood volume according to rainfall intensity and the elapsed time calculated in advance by flood analysis; a flood depth distribution storage unit for storing the relationship between flood arrival time and flood volume; A real-time inland flooding prediction system.

3. The real-time inland flooding prediction system according to claim 2, The rainfall intensity-based flooding amount time change storage unit The graph shows the results of a flood analysis with continuous constant rainfall of different intensities as external rainfall forces during the water level rise period. The graph shows the results of a flood analysis for the period of low water levels, where rainfall of different intensities continues at a constant intensity. A real-time inland flooding prediction system comprising:

4. The real-time inland flooding prediction system according to claim 2, This is a real-time inland flooding prediction system in which the flood arrival time calculation unit calculates the flood depth distribution for time step (w+1) by averaging multiple nearby points among multiple plots that each show the relationship between the amount of water expected for multiple different probable rainfalls and the flood arrival time at that time, as recorded in the flood depth distribution storage unit.

5. The real-time inland flooding prediction system according to claim 4, A real-time inland flooding prediction system, wherein the plurality of different probable rainfall amounts includes at least one of the probable rainfall amount for a specific year and an expected maximum scale.

6. The real-time inland flooding prediction system according to claim 4, The flood depth distribution storage unit estimates the amount of flooding for different probable rainfalls, A frontal rainfall waveform in which rain is concentrated at the beginning of the rainfall period. A centralized rainfall waveform in which rain is concentrated in the middle of the rainfall period. A rear-concentration type rainfall waveform in which rain is concentrated at the end of the rainfall period. A real-time inland flooding prediction system comprising the model.

7. The real-time inland water flooding prediction system according to any one of claims 1 to 6, further comprising: A plurality of water level sensors are installed at predetermined installation positions in the prediction area; Equipped with The water volume calculation unit is connected to each of the plurality of water level sensors via a network, A real-time inland flooding prediction system in which the water level calculation unit is configured to correct the water level calculated in real time using the measured water level at the installation location observed by the water level sensor.

8. The real-time inland flooding prediction system according to claim 7, A real-time inland flooding prediction system in which the water level calculation unit is configured to perform real-time calculation of the water level using the measured water level at the installation position observed by the water level sensor as the initial water level.

9. The real-time inland flooding prediction system according to claim 7, A real-time inland flooding prediction system in which the water volume calculation unit is configured to complement the water levels of surrounding areas with adjacent measured water levels among the measured water levels at multiple installation locations observed by the multiple water level sensors.

10. A real-time inland flooding prediction device that predicts in real time inland flooding in a predicted area caused by rainfall under conditions in which a predicted area is divided into small drainage basins, uniform rainfall is assumed in each small drainage basin, and a pre-calculated flood depth distribution is assumed to occur for each small drainage basin at each position included in the predicted area, A data acquisition unit that acquires observed rainfall and predicted rainfall; a water volume calculation unit for calculating a current water volume in real time based on the observed rainfall acquired by the data acquisition unit, the predicted rainfall, and an initial water level; a predicted flood depth distribution model storage unit for storing a plurality of predicted flood depth distribution models calculated in advance for each flooding amount in a predicted area; a flood arrival time calculation unit for calculating a flood arrival time that is expected to reach a forecast area based on a rainfall intensity curve; a predicted flood depth distribution model extraction unit for extracting a corresponding predicted flood depth distribution model from a plurality of predicted flood depth distribution models stored in the predicted flood depth distribution model storage unit based on the flood volume calculated in real time by the flood volume calculation unit; and Equipped with This real-time inland flooding prediction device is configured so that the predicted flood depth distribution model extraction unit corrects the flood volume calculated in real time by the flood volume calculation unit to a corrected flood volume by taking into account the flood arrival time calculated by the flood arrival time calculation unit, and extracts a predicted flood depth distribution model corresponding to the corrected flood volume.

11. A real-time inland flooding prediction method for predicting inland flooding in a predicted area caused by rainfall in real time under the conditions that a predicted area is divided into small drainage basins, uniform rainfall occurs in each small drainage basin, and a pre-calculated flood depth distribution occurs for each small drainage basin at each position included in the predicted area, calculating a current water volume in real time based on the initial water level; A step of extracting a corresponding predicted flood depth distribution model from a plurality of predicted flood depth distribution models calculated in advance for each flood level in the prediction area based on the flood level calculated in real time; Including, In the step of extracting the corresponding predicted flood depth distribution model, A real-time inland flood prediction method including a step of calculating the flood arrival time when the flood is expected to reach a predicted area based on rainfall intensity in a flood arrival time calculation unit, correcting the flood volume calculated in real time to a corrected flood volume taking into account the calculated flood arrival time, and extracting a predicted flood depth distribution model corresponding to the corrected flood volume.

12. A real-time inland flooding prediction method according to claim 11, In the step of extracting the corresponding predicted flood depth distribution model, preparing a flood depth distribution storage unit that stores the relationship between flood volume and flood arrival time as flood depth distribution data; preparing a storage unit for storing data on the change in flood volume over time by rainfall intensity, the data being calculated by flood analysis to show the relationship between the flood volume according to the rainfall intensity and the elapsed time; A step of comparing the water volume (V(t)) at the current time step (w), which is the water volume at the current time, with the water volume change data by rainfall intensity stored in the water volume change by rainfall intensity storage unit; A step of calculating a time change (ΔV(R)) in the amount of water caused by an arbitrary rainfall (R) occurring from the amount of water (V(t)) at the current time step (w) to the next time step (w+1) after a certain time (Δw) has elapsed, from the water amount at the current time step (w) based on the data on time changes in the amount of water by rainfall intensity; A step of estimating the water volume (V(t+1)) of the time step (w+1) based on the calculated time change (ΔV(R)) of the water volume of the time step (w+1); A step of comparing the flood volume (V(t+1)) of the time step (w+1) with the flood depth distribution data stored in the flood depth distribution storage unit; storing the rainfall intensity curve calculated from the rainfall waveform in a rainfall intensity curve storage unit; A step of calculating the Kadoya-Fukushima formula in a quasi-linear storage type function model calculation unit; A step of calculating a flood arrival time by a flood arrival time calculation unit, the point where the rainfall intensity curve stored in the rainfall intensity curve storage unit intersects with a quasi-linear storage type function model as a flood arrival time; The flood arrival time calculated by the flood arrival time calculation unit is compared with the flood depth distribution data stored in the flood depth distribution storage unit to calculate the flood volume (V A (T A ) and Calculating an averaged flood depth distribution for the time step (w+1) by averaging a plurality of nearby points that have been recorded in advance in the flood depth distribution storage unit; a step of repeating the steps from comparing the data on the amount of flooding by rainfall intensity with the data on the amount of flooding over time for a time step (w+2) that is a certain time (Δw) after the time step (w+1) to the step of calculating an averaged flood depth distribution for the time step (w+2), thereby calculating an averaged flood depth distribution for each time; A real-time inland flooding prediction method, including:

13. A real-time inland water flooding prediction method according to claim 12, In the step of preparing the flood depth distribution storage unit, A step of creating a rainfall intensity equation for each of a plurality of different probable rainfalls; For the rainfall intensity formula for each probability rainfall created above, A frontal rainfall waveform where rain is concentrated at the beginning of the rainfall period. A centralized rainfall waveform where rain is concentrated in the middle of the rainfall period. A rear-focused rainfall waveform in which rain is concentrated at the end of the rainfall period. and outputting the flood analysis results at predetermined time intervals. A real-time inland flooding prediction method, including:

14. A real-time inland water flooding prediction method according to claim 13, comprising: A real-time inland flooding prediction method, in which the plurality of different probable rainfall amounts include at least one of the probable rainfall amount for a specific year and an expected maximum scale.

15. A real-time inland water flooding prediction method according to claim 13, comprising: In the step of preparing the flood depth distribution storage unit, A step of performing flood analysis using a plurality of different intensities of continuous constant rainfall as rainfall external force as a graph of the water level rise period, and outputting the flood analysis results at predetermined time intervals; A step of performing flood analysis using a plurality of different intensities of continuous constant rainfall as rainfall external force as a graph of the water level decline period, and outputting the flood analysis results at predetermined time intervals; A real-time inland flooding prediction method, including:

16. The real-time inland flooding prediction method according to any one of claims 11 to 15, further comprising: A step of acquiring water levels at each installation location observed by a plurality of water level sensors installed at respective predetermined installation locations in the prediction area; A step of correcting the amount of water calculated in real time with the water level at the installation position observed by the water level sensor; A real-time inland flooding prediction method, including:

17. A real-time inland water flooding prediction method according to claim 16, comprising: The step of correcting the amount of water included in the A real-time inland flooding prediction method including a step of calculating the amount of water inundation in real time using the measured water level at the installation position observed by the water level sensor as the initial water level.

18. A real-time inland water flooding prediction method according to claim 16, comprising: The step of correcting the amount of water included in the A real-time inland flooding prediction method including a step of supplementing the water levels of surrounding areas with adjacent measured water levels from among the measured water levels at multiple installation locations observed by the multiple water level sensors.

19. A real-time inland flood prediction program for predicting inland flooding in a predicted area caused by rainfall in real time under the following conditions: a predicted area is divided into small drainage basins, and it is assumed that uniform rain will fall in each small drainage basin, and that a pre-calculated flood depth distribution will occur for each small drainage basin at each position included in the predicted area; A function to calculate the current water volume in real time based on the initial water level, A function of extracting a corresponding predicted flood depth distribution model from a plurality of predicted flood depth distribution models calculated in advance for each flood level in the predicted area based on the flood level calculated in real time; and A function of calculating a flood arrival time that is expected to occur in a predicted area based on rainfall intensity, correcting the flood volume calculated in real time to a corrected flood volume taking into account the calculated flood arrival time, and extracting a predicted flood depth distribution model corresponding to the corrected flood volume; A real-time inland flooding prediction program to realize this on a computer.

20. A computer-readable recording medium or storage device having the program of claim 19 recorded thereon.

Images