Rainfall monitoring system and rainfall monitoring program

Abstract

To enable easy and proper monitoring as to whether or not 24-hour cumulative rainfall amount exceeds an inspection implementation standard.SOLUTION: A rainfall amount monitoring system comprises: a power plant information database 31 which stores the power plant information including the location of each power plant; a power plant extracting task 41 which extracts a power plant having rain as a rainfall power plant based on the wide area rain cloud radar information and the power plant information; and a heavy rain monitoring task 42 which acquires the rainfall amount data of an observation station around the rainfall power station from among rainfall amount data of a plurality of observation stations and calculates the 24-hour cumulative rainfall amount which is a rainfall amount per 24 hours.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a rainfall monitoring system and a rainfall monitoring program for monitoring the amount of rainfall in the vicinity of a power plant. [Background technology]

[0002] In thermal power plants, when a certain heavy rain occurs, a patrol inspection is conducted to check whether there are any abnormalities in the civil engineering facilities of the thermal power plant based on the civil engineering and construction related facility maintenance guidelines stipulated by each electric power company, etc. Here, a certain heavy rain is when the rainfall (24-hour cumulative rainfall) in the 24 hours (1 day) going back from a certain time exceeds a certain cumulative rainfall (inspection implementation standard), and the occurrence of the certain heavy rain is monitored as follows. That is, for example, from the rainfall data of each point and each observation station provided on the website (homepage) of the Ministry of Land, Infrastructure, Transport and Tourism Disaster Prevention Information Center, hourly rainfall data around each power plant is obtained to monitor whether the certain heavy rain occurs.

[0003] In many cases, civil engineers are not stationed at each power plant, and if they are working at the power company's headquarters, they monitor the rain cloud conditions in their jurisdiction (for example, the Chugoku region) over the Internet and obtain rainfall data around the power plant where it is thought to be raining from the website of the Ministry of Land, Infrastructure, Transport and Tourism's Disaster Prevention Information Center. They then calculate the cumulative rainfall over 24 hours based on the obtained hourly rainfall data, and understand and confirm whether it exceeds the inspection implementation standard.

[0004] Meanwhile, a rainfall estimation system is known that can estimate the ever-changing rainfall amount at any point with high accuracy (for example, see Patent Document 1). This system learns the relationship between echo intensity and rainfall amount at a fixed point based on echo intensity information from a weather radar and rainfall information from each rain gauge, and estimates the rainfall amount at any point from the echo intensity of the same point according to the learning results. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 07-146375 Summary of the Invention [Problem to be solved by the invention]

[0006] However, the conventional method of monitoring the rain cloud conditions in the jurisdiction over the Internet, obtaining rainfall data from a website for areas around the power plant where rain is thought to be falling, and calculating the 24-hour cumulative rainfall required a great deal of time and effort. Moreover, since this was done by humans, there was a risk that the accuracy and judgment of the rain cloud conditions monitoring would vary, and it was not possible to monitor properly. Furthermore, there have been many sudden heavy rains in recent years, and it is thought that there will be many more cases in the future where the 24-hour cumulative rainfall exceeds the inspection implementation standard in a short period of time. Therefore, it was desirable to be able to monitor the rainfall easily and properly.

[0007] On the other hand, the rainfall estimation system described in Patent Document 1 is capable of estimating rainfall at any point, but cannot monitor whether the 24-hour accumulated rainfall around the power plant exceeds the inspection implementation standard.

[0008] SUMMARY OF THE PRESENT EMBODIMENTS The present invention aims to provide a rainfall monitoring system and a rainfall monitoring program that can easily and appropriately monitor whether or not a 24-hour accumulated rainfall exceeds an inspection implementation standard. [Means for solving the problem]

[0009] In order to solve the above problem, the invention of claim 1 is Map with locations illustrated A power plant information storage means for storing power plant information including the following: A rain cloud radar image is overlaid on the map to identify locations within the rain clouds. This rainfall monitoring system is characterized by comprising a power plant extraction means for extracting a power plant as a rainfall power plant, and a heavy rainfall monitoring means for acquiring rainfall data from an observation station surrounding the rainfall power plant from among rainfall data from multiple observation stations, and calculating a 24-hour cumulative rainfall, which is the amount of rainfall per 24 hours.

[0010] The invention of claim 2 is the rainfall monitoring system according to claim 1, Monitoring The means is characterized by notifying a predetermined contact person when the 24-hour accumulated rainfall exceeds a predetermined accumulated rainfall.

[0011] The invention of claim 3 is the rainfall monitoring system according to claim 1 or 2, Monitoring The means is characterized in that, when the 24-hour accumulated rainfall does not exceed a predetermined accumulated rainfall, it estimates whether or not the predetermined accumulated rainfall will be exceeded in the future, and, if it is estimated that the predetermined accumulated rainfall will be exceeded, it estimates how many hours are left until the predetermined accumulated rainfall is exceeded.

[0012] The invention of claim 4 is the rainfall monitoring system according to claims 1 to 3, wherein the power plant extraction means is the rain cloud radar. image The method is characterized in that it uses a learning model for raincloud radar analysis that has been machine-learned based on past performance data so that the rainfall power plant is output when the following is input:

[0013] The invention of claim 5 is characterized in that, in the rainfall monitoring system described in claim 3, the heavy rain judgment means uses a learning model for heavy rain estimation that is machine-learned based on past actual data so that, when the rainfall data is input, an estimation result of whether or not the specified cumulative rainfall will be exceeded in the future and how many hours it will take to exceed that amount is output.

[0014] The invention of claim 6 is a computer for each power plant. Map with locations illustrated A power plant information storage means for storing power plant information including the following: A rain cloud radar image is overlaid on the map to identify locations within the rain clouds. This is a rainfall monitoring program that functions as a power plant extraction means that extracts a power plant as a rainfall power plant, and a heavy rainfall monitoring means that obtains rainfall data from observation stations surrounding the rainfall power plant from among rainfall data from multiple observation stations, and calculates the 24-hour cumulative rainfall, which is the rainfall per 24 hours.

[0015] The invention of claim 5 is the rainfall monitoring system according to claim 3, MonitoringThe means is characterized in that it uses a learning model for estimating heavy rain that has been machine-learned based on past actual data so that, when the rainfall data is input, an estimation result of whether the specified cumulative rainfall will be exceeded in the future and how many hours it will take to exceed that amount is output.

[0016] The invention of claim 8 provides the rainfall monitoring program according to claim 6 or 7, wherein the power plant extraction means is a rain cloud radar. image The method is characterized in that it uses a learning model for raincloud radar analysis that has been machine-learned based on past performance data so that the rainfall power plant is output when the following is input:

[0017] The invention of claim 9 is the rainfall monitoring program according to claim 7, Monitoring The means is characterized in that it uses a learning model for estimating heavy rain that has been machine-learned based on past actual data so that, when the rainfall data is input, an estimation result of whether the specified cumulative rainfall will be exceeded in the future and how many hours it will take to exceed that amount is output. Effect of the Invention

[0018] According to the inventions described in claims 1 and 6, a precipitation power plant is extracted from wide-area raincloud radar information, and precipitation data around the precipitation power plant is obtained from precipitation data from multiple observation stations, and the 24-hour cumulative precipitation is calculated. In this way, precipitation data around the precipitation power plant is automatically obtained and the 24-hour cumulative precipitation is calculated, so it becomes possible to easily and appropriately monitor whether the 24-hour cumulative precipitation at each power plant exceeds the inspection implementation standard. As a result, even if there are frequent sudden heavy rains and the 24-hour cumulative precipitation exceeds the inspection implementation standard in a short period of time, it becomes possible to easily and appropriately grasp the situation and appropriately and quickly deal with it.

[0019] According to the invention described in claim 2, if the 24-hour accumulated rainfall exceeds a predetermined accumulated rainfall (e.g., an inspection implementation standard), a specified contact point is notified of this, enabling the contact point to take appropriate and prompt action.

[0020] According to the inventions described in claims 3 and 7, when the 24-hour cumulative rainfall does not exceed a predetermined cumulative rainfall, it is estimated whether or not the predetermined cumulative rainfall will be exceeded in the future, making it possible to deal with future situations appropriately and quickly. Moreover, when it is estimated that the predetermined cumulative rainfall will be exceeded, it is further estimated how many hours it will take to exceed the predetermined cumulative rainfall, making it possible to take more appropriate and quicker measures according to the estimated time.

[0021] According to the inventions described in claims 4 and 8, rainfall power plants are extracted from rainfall radar information using a machine-learned learning model for rainfall radar analysis, so it is possible to accurately and appropriately extract rainfall power plants. As a result, it is possible to properly (without misidentification) monitor whether the 24-hour cumulative rainfall at each power plant exceeds the inspection implementation standard without omission.

[0022] According to the inventions of claims 5 and 9, a machine-learned model for estimating heavy rainfall is used to estimate whether or not the cumulative rainfall over the next 24 hours will exceed a predetermined cumulative rainfall amount and how many hours until that amount is exceeded, making it possible to make accurate and proper estimates. As a result, it becomes possible to more properly monitor whether or not the cumulative rainfall over the next 24 hours exceeds the inspection implementation standard at each power plant. [Brief description of the drawings]

[0023] [Figure 1] 1 is a schematic block diagram showing a rainfall monitoring system according to an embodiment of the present invention; [Diagram 2] FIG. 2 is a data structure diagram of a power plant information database of the rainfall monitoring system of FIG. 1. [Diagram 3] FIG. 3 is a diagram showing the locations of each power plant on a map based on the information in the power plant information database of FIG. 2. [Figure 4] FIG. 2 is a diagram showing an example of a raincloud radar image according to an embodiment of the present invention. [Diagram 5] This is a diagram that overlays the map in Figure 3 with the rain cloud radar image in Figure 4. [Figure 6]FIG. 2 is a functional block diagram showing a schematic configuration of a learning model for raincloud radar analysis in the rainfall monitoring system of FIG. 1. [Figure 7] FIG. 2 is a diagram showing rainfall data from a plurality of observation stations according to an embodiment of the present invention. [Figure 8] FIG. 2 is a diagram showing meteorological data of a rain-fed power plant in an embodiment of the present invention. [Figure 9] 1. FIG. 4 is a diagram showing a first example of a 24-hour accumulated rainfall calculated in a heavy rainfall monitoring task of the rainfall monitoring system of FIG. [Figure 10] 1. FIG. 4 is a diagram showing a second example of a 24-hour accumulated rainfall calculated in the heavy rainfall monitoring task of the rainfall monitoring system of FIG. [Figure 11] 2 is a functional block diagram showing a schematic configuration of a learning model for estimating heavy rain in the rainfall monitoring system of FIG. 1. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0024] Hereinafter, the present invention will be described based on the illustrated embodiment.

[0025] Fig. 1 is a schematic block diagram showing a rainfall monitoring system 1 according to an embodiment of the present invention. This rainfall monitoring system 1 is a system that monitors the amount of rainfall around a power plant, and in this embodiment, a case where the rainfall monitoring system 1 is configured with one computer server will be described, but it may be configured with multiple computers and storage devices. In this embodiment, the power plant to be monitored is a thermal power plant, and when heavy rain occurs at the thermal power plant, it is necessary to subsequently patrol the power plant to check whether there are any abnormalities in the civil engineering facilities of the thermal power plant, so the system mainly monitors whether heavy rain occurs.

[0026] Here, heavy rain is defined as rainfall in which the amount of rainfall (24-hour accumulated rainfall) going back 24 hours (one day) from a certain time exceeds a predetermined accumulated rainfall (inspection implementation standard, 150 mm in this embodiment), or rainfall in which the amount of rainfall per hour exceeds a predetermined hourly rainfall (40 mm in this embodiment).In addition, in this embodiment, it is assumed that an electric power company that mainly supplies electricity to a specific region of Japan (Chugoku region, Kanto region, Shikoku region, etc.) monitors thermal power plants installed in its own jurisdiction, and the Chugoku region will be explained below as an example.

[0027] The rainfall monitoring system 1 mainly comprises an input unit 21, a display unit 22, a communication unit 23, a camera 24, a memory unit 3, a main task 4, and a central processing unit 5 that controls these components.

[0028] The input unit 21 is an interface for inputting various information and commands, and the display unit 22 is a display for displaying various data and information, etc. The communication unit 23 is an interface for communicating with the outside via the Internet network, a telephone communication network, etc., and the camera 24 is a photographing device for taking images (still images, videos).

[0029] The memory unit 3 mainly comprises a power plant information database (power plant information storage means) 31, a learning model for raincloud radar analysis 32, a track record database for raincloud radar analysis 33, a learning model for heavy rain estimation 34, and a track record database for heavy rain estimation 35. Here, the power plant information database 31 will be explained, and the learning models 32, 34 and the track record databases 33, 35 will be described later.

[0030] The power plant information database 31 is a database that stores power plant information including the location of each thermal power plant, and as shown in Fig. 2, for each power plant name 311, which is the identification information of the thermal power plant, a location 312, a rain gauge 313, a seismic intensity meter 314, an anemometer 315, and others 316 are stored. The location 312 stores the latitude and longitude of the location of the thermal power plant, and further, as shown in Fig. 3, a map illustrating the locations of each thermal power plant G1 to G8 is stored.

[0031] The rain gauge 313 stores which meter is used to measure the amount of rainfall at this thermal power plant. Specifically, in this embodiment, among the observation stations whose rainfall data is disclosed and provided on the website (homepage) of the Ministry of Land, Infrastructure, Transport and Tourism Disaster Prevention Information Center, identification information (name) of an observation station close to this thermal power plant is stored. In other words, the amount of rainfall at the thermal power plant is the amount of rainfall measured at an observation station close to this thermal power plant. The website discloses the accumulated amount of rainfall since it started to rain, but this is different from the 24-hour accumulated amount of rainfall.

[0032] Similarly, the seismic intensity meter 314 stores information about which instrument is used to measure the seismic intensity of an earthquake that occurred at this thermal power plant, for example, information indicating that it is a seismic intensity meter installed at the power plant and identification information (name) of the local meteorological observatory. The anemometer 315 stores information about which instrument is used to measure the wind speed and wind direction at this thermal power plant, for example, information indicating that it is a anemometer installed at the power plant and identification information (name) of the local meteorological observatory.

[0033] Air temperature and temperature are also measured at each of the thermal power plants G1 to G8, and in this embodiment, air temperature data and wind speed data including wind direction are successively acquired and stored and accumulated in the storage unit 3. As will be described later, rainfall data is acquired as necessary in the heavy rain monitoring task 42, but the rainfall data at each of the thermal power plants G1 to G8 may also be acquired successively and stored and accumulated in the storage unit 3.

[0034] Main task 4 is a task that calculates the 24-hour cumulative rainfall at thermal power plants that are currently or have previously experienced rainfall, and includes a power plant extraction task (power plant extraction means) 41, a heavy rain monitoring task (heavy rain monitoring means) 42, a first learning task 43, and a second learning task 44.

[0035] The power plant extraction task 41 is a task program that extracts thermal power plants where rain is falling as rainfall power plants based on wide-area raincloud radar information and power plant information. That is, based on the raincloud radar image and the location and position of each thermal power plant G1-G8, the thermal power plants G1-G8 where rain is falling or are estimated to have fallen are extracted as rainfall power plants. Specifically, in this embodiment, a raincloud radar image of the Chugoku region as shown in FIG. 4 provided on the Internet is acquired, and the thermal power plants G1-G8 located within the raincloud C are extracted (determined). For example, as shown in FIG. 5, the raincloud radar image is superimposed on a map showing the positions of each thermal power plant G1-G8, and the second thermal power plant G2 and the third thermal power plant G3 located within the raincloud C are extracted as rainfall power plants.

[0036] Here, the rainfall monitoring system 1 may directly access a website that discloses and provides rainfall radar images on the Internet to acquire and monitor the rainfall radar images, or the rainfall radar images acquired from the website using another computer or the like may be imported into the rainfall monitoring system 1. Alternatively, the rainfall radar images may be photographed by the camera 24 and loaded into the rainfall monitoring system 1, and any other form may be used. Also, in order to properly calculate and monitor the 24-hour accumulated rainfall, which will be described later, extraction of such rainfall power plants, i.e., activation of the power plant extraction task 41, is performed periodically.

[0037] In the power plant extraction task 41, a learning model 32 for precipitation radar analysis that is machine-learned based on past performance data is used to extract precipitation power plants when precipitation radar information (precipitation radar image) is input. This learning model 32 for precipitation radar analysis is created by a first learning task 43.

[0038] That is, the first learning task 43 uses past performance data recorded and accumulated in the raincloud radar analysis performance database 33 to create a raincloud radar analysis learning model 32 by a known machine learning algorithm such as a neural network. This raincloud radar analysis performance database 33 is a database in which, as input information, past raincloud radar images and performance data including thermal power plants where rain actually fell when the raincloud radar images were in a certain state are recorded and accumulated. The past performance data includes data created based on actual raincloud radar images and thermal power plants where rain actually fell, as well as data created for advance training, etc.

[0039] As shown in Fig. 6, this first learning task 43 uses machine learning and deep learning using a neural network to create a neural network with, for example, a raincloud radar image as an input layer, a rainfall power plant as an output layer, and an image analysis process from the raincloud radar image to the rainfall power plant as an intermediate layer based on the performance data recorded in the raincloud radar analysis performance database 33. Then, the first learning task 43 uses the performance data of the raincloud radar analysis learning model 32 as learning data to learn various parameters in the intermediate layer. That is, the first learning task 43 learns various parameters in the intermediate layer so that the rainfall power plants are extracted from the raincloud radar image without omission and with high accuracy.

[0040] The heavy rain monitoring task 42 is a task program that acquires rainfall data from observation stations around the rainfall power plant from among the rainfall data from multiple observation stations installed in various places, and calculates the 24-hour cumulative rainfall. Here, in this embodiment, the rainfall data from multiple observation stations is rainfall data stored and provided on the website of the Ministry of Land, Infrastructure, Transport and Tourism Disaster Prevention Information Center, which provides rainfall data from each observation station nationwide, but it may be data from any resource. The rainfall monitoring system 1 may directly access the website to search for and acquire the desired rainfall data, or may access the website from another computer or the like to search for and acquire the desired rainfall data from a data group that stores only the required rainfall data (for example, rainfall data from observation stations close to each of the thermal power plants G1 to G8). In this embodiment, the rainfall data is hourly rainfall (hourly rainfall) as shown in FIG. 7, but any data may be used as long as it can calculate the 24-hour cumulative rainfall described later.

[0041] Among the rainfall data of such a plurality of observation stations, the rainfall data of the observation stations around the rainfall power plant extracted in the power plant extraction task 41, that is, the observation station stored in the rain gauge 313, is acquired. For example, as described above, the second thermal power plant G2 and the third thermal power plant G3 are extracted as rainfall power plants, and the observation station close to the second thermal power plant G2 is the observation station B, and the observation station close to the third thermal power plant G3 is the observation station C. In this case, among the rainfall data of the observation stations in various places shown in FIG. 7, the rainfall data of the observation station B is acquired for the second thermal power plant G2, and the rainfall data of the observation station C is acquired for the third thermal power plant G3. In this embodiment, as shown in FIG. 8, data on temperature, wind direction, and wind speed are also acquired and stored in association with the acquired rainfall data (matching the date and time). Note that the weather data including the rainfall, temperature, wind direction, and wind speed as shown in FIG. 8 may be stored and accumulated in advance not only for the rainfall power plant but also for all the thermal power plants G1 to G8.

[0042] Next, the 24-hour cumulative rainfall is calculated based on the obtained rainfall data. Here, the 24-hour cumulative rainfall is the rainfall per 24 hours, and is the cumulative rainfall in the 24 hours going back from any time. For example, as shown in FIG. 9, the 24-hour cumulative rainfall is calculated by adding up the hourly rainfall from 0:00 to 23:00 on February 1, going back 24 hours from 23:00 on February 1. Similarly, the 24-hour cumulative rainfall is calculated by adding up the hourly rainfall from 2:00 on February 1 to 1:00 on February 2. Furthermore, the 24-hour cumulative rainfall going back 24 hours from the current time is calculated. In this way, the 24-hour cumulative rainfall is the cumulative rainfall in the 24 hours going back from any time in the rainfall period, and may span across days.

[0043] Here, the time interval and period for which the 24-hour cumulative rainfall is calculated are set so that it is possible to properly monitor without omission whether the 24-hour cumulative rainfall exceeds a predetermined cumulative rainfall. For example, the 24-hour cumulative rainfall is calculated every two hours, and the 24-hour cumulative rainfall is calculated from the start of rainfall (when the precipitation exceeds zero) to the end of rainfall (until the state of zero precipitation continues for several hours).

[0044] Next, if the calculated 24-hour accumulated rainfall exceeds a predetermined accumulated rainfall (150 mm), the information is notified to a predetermined contact. That is, as shown in FIG. 10, if any of the calculated 24-hour accumulated rainfall exceeds the inspection implementation standard, the information is notified to a terminal of a person in charge or a department registered in advance along with the value of the 24-hour accumulated rainfall. In this embodiment, it is also monitored whether each hourly rainfall exceeds a predetermined hourly rainfall (40 mm), and if it does exceed the standard, the information is notified to a terminal of a person in charge or a department registered in advance along with the value of the hourly rainfall, even if the 24-hour accumulated rainfall does not exceed the inspection implementation standard.

[0045] On the other hand, if none of the 24-hour cumulative rainfall exceeds the inspection implementation standard, it estimates whether the 24-hour cumulative rainfall will exceed the inspection implementation standard, and if it is estimated that it will, it estimates how many hours until it will exceed it. In other words, although there has been no 24-hour cumulative rainfall that exceeds the inspection implementation standard up to the current time, it estimates whether the rainfall will continue to exceed the inspection implementation standard in the future and when the inspection implementation standard will be exceeded, based on the rainfall data up to the current time, i.e., the amount of rainfall and past data. Then, if it is estimated that the 24-hour cumulative rainfall will exceed the inspection implementation standard, information including how many hours until it will exceed it is notified to a terminal of a pre-registered person in charge or in charge of a department.

[0046] For this estimation, a heavy rain estimation learning model 34 that has been machine-learned based on past performance data is used so that when rainfall data is input, an estimation result of whether or not a certain cumulative rainfall will be exceeded in the future and how many hours it will take to exceed that amount is output. Furthermore, in this embodiment, estimation is performed based not only on rainfall data but also on meteorological data including temperature, wind direction, and wind speed. This heavy rain estimation learning model 34 is created by a second learning task 44.

[0047] That is, the second learning task 44 creates a learning model 34 for heavy rain estimation by a known machine learning algorithm such as a neural network using past performance data recorded and accumulated in a heavy rain estimation performance database 35. This heavy rain estimation performance database 35 is a database in which performance data including previously acquired weather data and the history and fluctuation trends of each 24-hour accumulated rainfall corresponding to this weather data (including cases where the rainfall exceeds the inspection implementation standard) are recorded and accumulated as input information. The past performance data includes data created based on the actual weather data and the history of the actual 24-hour accumulated rainfall, as well as data created in advance training, etc.

[0048] As shown in Fig. 11, this second learning task 44 uses machine learning and deep learning using a neural network to create a neural network based on the performance data recorded in the heavy rain estimation performance database 35, with, for example, weather data as an input layer, an estimation result of how many hours will or will not exceed the inspection implementation standard in the future as an output layer, and an analysis process from the weather data to the estimation result as an intermediate layer. Then, the second learning task 44 uses the performance data of the heavy rain estimation learning model 34 as learning data to learn various parameters in the intermediate layer. That is, the second learning task 44 learns various parameters in the intermediate layer so that an accurate estimation result can be obtained based on the weather data.

[0049] Next, the operation of the rainfall monitoring system 1 configured as above will be described.

[0050] First, the power plant extraction task 41 is periodically started, and thermal power plants G1 to G8 where rain is falling are extracted as rainfall power plants based on the acquired and input raincloud radar image. Next, the heavy rain monitoring task 42 is started, and rainfall data from observation stations around the rainfall power plants is acquired, and the 24-hour cumulative rainfall is calculated. As a result, if the 24-hour cumulative rainfall exceeds the inspection implementation standard, or if the hourly rainfall exceeds a predetermined hourly rainfall, the fact is notified to a terminal of a person in charge or a department registered in advance. Here, depending on the degree of rainfall and the timing of starting the power plant extraction task 41, it may be confirmed and monitored that the 24-hour cumulative rainfall has exceeded the inspection implementation standard after the rain has stopped, or it may be confirmed while it is still raining.

[0051] On the other hand, if the 24-hour cumulative rainfall does not exceed the inspection implementation standard, it is estimated whether the 24-hour cumulative rainfall will exceed the inspection implementation standard and how many hours it will take to exceed it. As a result, if it is estimated that the 24-hour cumulative rainfall will exceed the inspection implementation standard, information including how many hours it will take to exceed it is notified to a pre-registered person in charge or a terminal of a department in charge. Next, the power plant extraction task 41 is started again, and the same process is repeated. Here, even if it is confirmed again that the 24-hour cumulative rainfall has exceeded the inspection implementation standard, the person in charge is not notified again. On the other hand, the number of hours it will take to exceed the standard is estimated again, and if the value (how many hours) is different from the previous time, the person in charge is notified of the estimated result, but if it is the same as the previous time, the person in charge is not notified.

[0052] As described above, according to the rainfall monitoring system 1, a rainfall power plant is extracted from the wide-area raincloud radar information and image, rainfall data around the rainfall power plant is obtained from rainfall data from multiple observation stations, and the 24-hour cumulative rainfall is calculated. In this way, rainfall data around the rainfall power plant is automatically obtained and the 24-hour cumulative rainfall is calculated, so it is possible to easily and properly monitor whether the 24-hour cumulative rainfall exceeds the inspection implementation standard at each of the thermal power plants G1 to G8. As a result, even if there are frequent sudden heavy rains and the 24-hour cumulative rainfall exceeds the inspection implementation standard in a short period of time, it is possible to easily and properly grasp the situation and deal with it appropriately and quickly.

[0053] In addition, if the 24-hour cumulative rainfall exceeds the inspection implementation standard, a notification will be sent to a designated contact point, enabling the contact point to take appropriate and prompt action.

[0054] On the other hand, if the 24-hour cumulative rainfall does not exceed the inspection implementation standard, it is possible to estimate whether or not the inspection implementation standard will be exceeded in the future, making it possible to deal with future situations appropriately and quickly. Moreover, if it is estimated that the inspection implementation standard will be exceeded, it is also possible to estimate how many hours until the standard is exceeded, making it possible to take more appropriate and prompt action according to the estimated time.

[0055] In addition, since rainfall power plants are extracted from rainfall radar information and images using the machine-learned rainfall radar analysis learning model 32, it becomes possible to extract rainfall power plants accurately and appropriately. As a result, it becomes possible to properly (without misidentification) monitor whether the 24-hour cumulative rainfall exceeds the inspection implementation standard at each of the thermal power plants G1 to G8 without omission.

[0056] Similarly, the machine-learned heavy rainfall estimation learning model 34 is used to estimate whether or not the 24-hour cumulative rainfall will exceed the inspection implementation standard and how many hours until that will occur, making it possible to make accurate and proper estimations. As a result, it becomes possible to more properly monitor whether or not the 24-hour cumulative rainfall will exceed the inspection implementation standard at each of the thermal power plants G1 to G8.

[0057] Although the embodiment of the present invention has been described in detail above, the specific configuration is not limited to this embodiment, and even if there are design changes within the scope of the present invention, they are included in the present invention. For example, in the above embodiment, the resource of rainfall data from multiple observation stations is the website of the Ministry of Land, Infrastructure, Transport and Tourism Disaster Prevention Information Center, but other websites, databases, etc. may be used as the resource. Similarly, the information and images from any resource may be used for wide-area raincloud radar information and images.

[0058] Also, the power plant extraction task 41 is periodically started, followed by the heavy rain monitoring task 42, but the power plant extraction task 41 and the heavy rain monitoring task 42 may be started consecutively or independently at other times. For example, the power plant extraction task 41 is constantly running to constantly monitor raincloud radar information and images, and when a rainfall power plant is extracted, the heavy rain monitoring task 42 is periodically started for this rainfall power plant to monitor the 24-hour accumulated rainfall. Subsequently, when another rainfall power plant is extracted by the power plant extraction task 41, the heavy rain monitoring task 42 may also be periodically started for this rainfall power plant.

[0059] On the other hand, the rainfall monitoring system 1 as described above may be configured by installing the following rainfall monitoring program in a general-purpose computer. That is, the rainfall monitoring program causes the computer to function as a power plant information storage means (power plant information database 31) that stores power plant information including the location of each power plant, a power plant extraction means (power plant extraction task 41) that extracts power plants where rainfall is occurring as rainfall power plants based on wide-area raincloud radar information and power plant information, and a heavy rainfall monitoring means (heavy rainfall monitoring task 42) that acquires rainfall data from observation stations surrounding the rainfall power plants from among rainfall data from multiple observation stations and calculates a 24-hour accumulated rainfall, which is the amount of rainfall per 24 hours. Monitoring The means, when the 24-hour accumulated rainfall does not exceed the predetermined accumulated rainfall, estimates whether or not the predetermined accumulated rainfall will be exceeded in the future, and when it is estimated that the predetermined accumulated rainfall will be exceeded, estimates how many hours are left until the predetermined accumulated rainfall is exceeded. The power plant extraction means uses a raincloud radar analysis learning model 32 that has been machine-learned based on past performance data so that a rainfall power plant is output when raincloud radar information is input, and Monitoring The means uses a heavy rain estimation learning model 34 that has been machine-learned based on past performance data so that when rainfall data is input, an estimation result of whether or not a specified cumulative rainfall will be exceeded in the future and how many hours it will take to exceed that amount is output. [Explanation of symbols]

[0060] 1. Rainfall monitoring system 3 Storage section 31 Power plant information database (power plant information storage means) 32 Learning model for rain cloud radar analysis 33. Performance database for rain cloud radar analysis 34 Learning model for estimating heavy rainfall 35. Heavy rainfall estimation database 4. Main Task 41 Power plant extraction task (power plant extraction means) 42 Heavy rain monitoring task (heavy rain monitoring method) 43 First Learning Task 44 Second Learning Task G1~G8 Thermal power plants

Claims

1. A power plant information storage means for storing power plant information including a map showing the location of each power plant; A power plant extraction means for extracting a power plant located within a rain cloud as a rainfall power plant by overlaying a rain cloud radar image on the map; A heavy rain monitoring means for acquiring rainfall data from a plurality of observation stations in the vicinity of the rain-induced power plant and calculating a 24-hour accumulated rainfall, which is the amount of rainfall per 24 hours; A rainfall monitoring system comprising:

2. The heavy rain monitoring means notifies a predetermined contact person when the 24-hour accumulated rainfall exceeds a predetermined accumulated rainfall.

2. The rainfall monitoring system according to claim 1.

3. When the 24-hour accumulated rainfall does not exceed a predetermined accumulated rainfall, the heavy rain monitoring means estimates whether the 24-hour accumulated rainfall will exceed the predetermined accumulated rainfall in the future, and when it is estimated that the 24-hour accumulated rainfall will exceed the predetermined accumulated rainfall, estimates how many hours are left until the predetermined accumulated rainfall is exceeded.

3. The rainfall monitoring system according to claim 1, wherein the rainfall monitoring system is a rainfall monitoring system.

4. The power plant extraction means uses a learning model for rain cloud radar analysis that has been machine-learned based on past performance data so that the rainfall power plant is output when the rain cloud radar image is input.

4. The rainfall monitoring system according to claim 1, wherein the rainfall monitoring system is a rainfall monitoring system.

5. The heavy rain monitoring means uses a learning model for heavy rain estimation that is machine-learned based on past performance data so that, when the rainfall data is input, an estimation result of whether or not the predetermined cumulative rainfall will be exceeded in the future and how many hours it will take to exceed the predetermined cumulative rainfall is output.

4. The rainfall monitoring system according to claim 3.

6. Computer, A power plant information storage means for storing power plant information including a map showing the location of each power plant; A power plant extraction means for extracting a power plant located within a rain cloud as a rainfall power plant by overlaying a rain cloud radar image on the map; a heavy rain monitoring means for acquiring rainfall data from a plurality of observation stations in the vicinity of the rain-induced power plant and calculating a 24-hour accumulated rainfall, which is the amount of rainfall per 24 hours; A rainfall monitoring program characterized in that the program functions as a rainfall monitoring program.

7. When the 24-hour accumulated rainfall does not exceed a predetermined accumulated rainfall, the heavy rain monitoring means estimates whether the 24-hour accumulated rainfall will exceed the predetermined accumulated rainfall in the future, and when it is estimated that the 24-hour accumulated rainfall will exceed the predetermined accumulated rainfall, estimates how many hours are left until the predetermined accumulated rainfall is exceeded.

7. The rainfall monitoring program according to claim 6.

8. The power plant extraction means uses a learning model for rain cloud radar analysis that has been machine-learned based on past performance data so that the rainfall power plant is output when the rain cloud radar image is input.

8. The rainfall monitoring program according to claim 6, wherein the rainfall monitoring program is a program for monitoring a rainfall amount.

9. The heavy rain monitoring means uses a learning model for heavy rain estimation that is machine-learned based on past performance data so that, when the rainfall data is input, an estimation result of whether or not the predetermined cumulative rainfall will be exceeded in the future and how many hours it will take to exceed the predetermined cumulative rainfall is output.

8. The rainfall monitoring program according to claim 7.

Images