Prediction device

Abstract

A prediction device (10) according to the present disclosure comprises a prediction unit (143) that inputs, to a prediction model for predicting the temporal or monetary cost required to restore facilities damaged due to a disaster, prediction target data comprising information relating to the damage of facilities in a target area due to a disaster, information relating to the service and maintenance of the facilities in the target area, and information relating to the disaster resistance of the target area, and predicts the temporal or monetary cost required to restore facilities damaged in the target area due to a disaster as the disaster risk of facilities in the target area.

Description

Prediction device

[0001] The present disclosure relates to a prediction device.

[0002] In recent years, due to the impact of climate change such as global warming, natural disasters occur every year across the country, causing great damage to social infrastructure facilities (hereinafter simply referred to as "facilities"). In particular, heavy rain and landslides have a statistically increasing occurrence frequency, and disaster prevention and mitigation efforts including pre-disaster prevention are required.

[0003] On the other hand, the number of engineers involved in the construction and maintenance management of facilities has been decreasing year by year. Also, due to population decline and low birth rate and aging population, it is assumed that the decrease in engineers will further progress in the future. Therefore, more efficient and effective maintenance management of facilities and disaster response are required. In addition, as the population outflow from urban areas and the concentration of productivity progress, regional disparities are expanding, so it is assumed that the maintenance management of facilities and disaster response in local areas will become more difficult.

[0004] Non-Patent Document 1 describes a technique for evaluating the physical damage risk of facilities due to disasters by using disaster damage data of facilities caused by past disasters. Also, Non-Patent Document 2 describes a technique for evaluating the risk due to disasters from the perspective of the recovery speed of facilities based on the damage scale of facilities caused by past disasters (earthquakes) and the recovery scale from the damage of the facilities due to the disasters.

[0005] Construction of an AI for predicting infrastructure damage against various disasters, Development of a recovery prediction model for supply system lifelines during earthquakes, NTT News Release (2024.4), 56th Annual Academic Lecture Meeting of the Japan Society of Civil Engineers (2001.10)

[0006] The techniques described in Non-Patent Documents 1 and 2 focus only on physical vulnerabilities such as damage and damage scale of facilities. That is, in the techniques described in Non-Patent Documents 1 and 2, comprehensive evaluations based on the supply and maintenance of facilities and regional disparities such as productivity are not sufficiently considered. As described above, as the population outflow from urban areas and the concentration of productivity progress, regional disparities are expanding, so a more multi-faceted prediction of the disaster risk of facilities considering the supply and maintenance of facilities and the disaster resistance of the region is required.

[0007] In light of the problems described above, the purpose of this disclosure is to provide a predictive device that can predict the disaster risk of equipment from a more multifaceted perspective, taking into account the operation and maintenance of the equipment, as well as the disaster resilience of the region.

[0008] One embodiment of the prediction device is a prediction device that predicts the disaster risk of facilities in a target area due to a disaster, and includes a prediction unit that inputs prediction target data consisting of information on the damage to the facilities in the target area due to the disaster, information on the operation and maintenance of the facilities in the target area, and information on the disaster resilience of the target area into a prediction model that predicts the time or monetary cost required to restore the facilities damaged by the disaster in the target area as the disaster risk of the facilities in the target area.

[0009] According to this disclosure, it is possible to predict the disaster risk of equipment from a more multifaceted perspective, taking into account the operation and maintenance of the equipment, as well as the disaster resilience of the region.

[0010] This figure shows an example of the configuration of a prediction device according to one embodiment of this disclosure. This figure shows an example of the configuration of the disaster DB shown in Figure 1. This figure shows an example of the configuration of the equipment DB shown in Figure 1. This figure shows an example of the configuration of the regional DB shown in Figure 1. This is a flowchart showing an example of the operation when the prediction device shown in Figure 1 constructs a prediction model. This figure shows an example of the training data generated by the data generation unit in Figure 1. This is a flowchart showing an example of the operation when the prediction device shown in Figure 1 predicts disaster risk.

[0011] Embodiments of this disclosure will be described below with reference to the drawings.

[0012] Figure 1 is a diagram showing an example of the configuration of a prediction device 10 according to one embodiment of the present disclosure. The prediction device 10 according to the present disclosure constructs a prediction model that predicts the disaster risk (the time or monetary cost required to restore damaged facilities) of social infrastructure facilities (facilities) such as utility poles, underground conduits, and transmission towers due to disasters, and uses the constructed prediction model to predict the disaster risk of facilities in a target area. As shown in Figure 1, the prediction device 10 according to this embodiment comprises an input unit 11, an output unit 12, a storage unit 13, and a control unit 14.

[0013] The input unit 11 includes at least one input interface. The input interface is, for example, a physical key, a capacitive key, a pointing device, a touchscreen integrated with a display, or a microphone. The input unit 11 accepts operations to input data used for the operation of the prediction device 10. The input unit 11 receives, for example, various information stored in the storage unit 13, which will be described later. The input unit 11 also receives, for example, parameters related to the execution of the operation of the control unit 14, which will be described later, and instructions for the execution of the operation of the control unit 14.

[0014] The input unit 11 may be connected to the prediction device 10 as an external input device, instead of being provided within the prediction device 10. Any connection method can be used, such as USB, HDMI®, or Bluetooth®. "USB" is an abbreviation for Universal Serial Bus. "HDMI®" is an abbreviation for High-Definition Multimedia Interface.

[0015] The output unit 12 includes at least one output interface. The output interface is, for example, a display. The display is, for example, an LCD or an organic EL display. "LCD" is an abbreviation for Liquid Crystal Display. "EL" is an abbreviation for Electro Luminescence. The output unit 12 outputs (displays) data obtained by the operation of the prediction device 10. The output unit 12 outputs, for example, the processing status and prediction results of the disaster risk prediction of facilities in the target area by the control unit 14, which will be described later. The output unit 12 outputs, for example, information stored in the storage unit 13.

[0016] The output unit 12 may be connected to the prediction device 10 as an external output device, instead of being provided within the prediction device 10. Any connection method can be used, such as USB, HDMI®, or Bluetooth®.

[0017] The memory unit 13 includes at least one semiconductor memory, at least one magnetic memory, at least one optical memory, or any combination thereof. The semiconductor memory is, for example, RAM, ROM, or flash memory. "RAM" is an abbreviation for Random Access Memory. "ROM" is an abbreviation for Read Only Memory. The RAM is, for example, SRAM or DRAM. "SRAM" is an abbreviation for Static Random Access Memory. "DRAM" is an abbreviation for Dynamic Random Access Memory. The ROM is, for example, EEPROM. "EEPROM" is an abbreviation for Electrically Erasable Programmable Read Only Memory. The flash memory is, for example, SSD. "SSD" is an abbreviation for Solid-State Drive. The magnetic memory is, for example, HDD. "HDD" is an abbreviation for Hard Disk Drive. The storage unit 13 functions, for example, as a main memory, auxiliary memory, or cache memory. The storage unit 13 stores information used for the operation of the prediction device 10, which is input via the input unit 11, and information obtained by the operation of the prediction device 10. The storage unit 13 functions, for example, as a disaster DB 131, equipment DB 132, regional DB 133, prediction model DB 134, and disaster risk DB 135.

[0018] Disaster DB131 is a database that stores data on the damage to facilities caused by past disasters or disasters predicted to occur in the future, as well as the resources required to restore the damaged facilities.

[0019] Figure 2 shows an example of the configuration of disaster DB131.

[0020] As shown in Figure 2, the disaster DB131 stores the following information in association with each other: "Disaster ID", "Disaster Type", "Equipment Type", "Affected Area", "Number of Affected Equipment", "Damage Density", "Amount of Damage", "Amount of Restoration", "Restoration Period", and "Restoration Operation".

[0021] "Disaster ID" is a unique value that can individually identify past disasters or disasters predicted to occur in the future. "Disaster Type" indicates the type of disaster that causes damage to the equipment (e.g., earthquake, landslide, river flooding, flood, typhoon, snow damage, etc.) identified by the "Disaster ID". "Affected Area" is a set of coordinates indicating the area in which equipment has been damaged or is expected to be damaged by the disaster identified by the "Disaster ID". "Number of Damaged Equipment" is a value indicating the number of equipment that has been damaged or is expected to be damaged within the area indicated by the "Affected Area" due to the disaster identified by the "Disaster ID". "Damage Density" is a value indicating the installation density of equipment that has been damaged or is expected to be damaged within the area indicated by the "Affected Area" due to the disaster identified by the "Disaster ID". "Damage Amount" is a value indicating the scale (amount) of damage to equipment that has been damaged or is expected to be damaged within the area indicated by the "Affected Area" due to the disaster identified by the "Disaster ID". "Restoration Cost" is a value indicating the amount required to restore equipment that has been damaged or is expected to be damaged in the area indicated by "Affected Area" due to the disaster identified by "Disaster ID". "Restoration Operation" is a value indicating the operation required to restore equipment that has been damaged or is expected to be damaged in the area indicated by "Affected Area" due to the disaster identified by "Disaster ID".

[0022] Referring again to Figure 1, the equipment DB132 is a database that stores data related to the operation and maintenance of the equipment.

[0023] Figure 3 shows an example of the configuration of equipment DB132.

[0024] As shown in Figure 3, the equipment DB132 stores the following information in association with each other: "Equipment ID", "Installation Location", "Equipment Type", "Number of Services Provided", "Importance of Services Provided", "Maintenance Difficulty", "Recovery Difficulty", and "Average Maintenance Operation".

[0025] "Equipment ID" is a unique value that allows for the individual identification of equipment. "Installation Location" is a coordinate value or set of coordinates indicating the installation location of the equipment identified by "Equipment ID". "Equipment Type" indicates the type of equipment identified by "Equipment ID" that has been damaged or is expected to be damaged by a disaster (e.g., underground conduits, underground manholes, transmission towers, utility poles, cables, roads, railway tracks, etc.). "Number of Services Handled" is a value indicating the amount of services handled by the equipment identified by "Equipment ID" (e.g., number of lines, number of contracts, number of households, etc.). "Importance of Services Handled" is a value indicating the management priority of the services handled by the equipment identified by "Equipment ID". "Maintenance Difficulty" is a value indicating the difficulty of maintenance of the equipment identified by "Equipment ID" based on its installation environment. For example, "2" is set for maintenance at river crossings or railway crossings where maintenance is difficult, "1" is set for maintenance at road crossings, and "0" is set for maintenance at other locations. "Recovery Difficulty" is a value that indicates the difficulty of recovery work when the equipment identified by "Equipment ID" is damaged. For example, a "2" is set for high-difficulty replacement with new equipment, a "1" is set for switching services to accommodate available equipment, and a "0" is set for switching services to accommodate redundant equipment. "Average Maintenance Operation" is a value that indicates the number of people involved in the recovery of the equipment identified by "Equipment ID".

[0026] Referring again to Figure 1, Regional DB133 is a database that stores data on the disaster resilience of a region.

[0027] Figure 4 shows an example of the configuration of the regional DB133.

[0028] As shown in Figure 4, the regional DB133 stores the following information in association with each other: "Regional ID", "Coordinates", "Labor Force Population", "Total Population", "Fiscal Strength Index", "Transportation Index", "Housing Density", "Percentage of Wooden Structures", "Percentage of Forest Area", "Percentage of River Area", and "Average Elevation".

[0029] The "Regional ID" is a unique value that can identify a region within a predetermined range. The "Regional ID" may be a value that uniquely identifies a region divided into a certain rectangular size. The "Regional ID" may be a value that uniquely identifies a region divided according to, for example, a regional mesh statistical division. Alternatively, the "Regional ID" may be a value that uniquely identifies a region divided according to administrative areas such as national, prefectural, or municipal boundaries, or according to boundary divisions for maintenance and management of social infrastructure services such as road networks and power distribution networks, in accordance with the intended use of the equipment's disaster risk. The "Coordinates" are a set of coordinates that indicate the location of the region identified by the "Regional ID". The set of coordinates indicated by the "Coordinates" is a set of coordinates that can be matched with the coordinates indicated by the "Installation Location" included in the Equipment DB132. The "Labor Force Population" is the labor force population in the region identified by the "Regional ID". The "Total Population" is the total population in the region identified by the "Regional ID". The "Fiscal Strength Index" is a value that indicates the economic soundness of the local government in the region identified by the "Regional ID". The "Transportation Index" is a value that shows the volume of transportation by rail, road, ship, and aircraft, and the resulting economic effect (value added), in the region identified by the "Regional ID". The "House Density" is a value that shows the density of houses in the region identified by the "Regional ID" and its surroundings. The "Proportion of Wooden Structures" is a value that shows the proportion of wooden structures in the region identified by the "Regional ID" and its surroundings. The "Forest Area Proportion" is a value that shows the proportion of forest area in the region identified by the "Regional ID" and its surroundings. The "River Area Proportion" is a value that shows the proportion of river area in the region identified by the "Regional ID" and its surroundings. The "Average Elevation" is the average elevation in the region identified by the "Regional ID" and its surroundings.

[0030] In the disaster database 131 shown in Figure 2, a "disaster area" associated with a single "disaster ID" may include multiple areas identified by multiple "regional IDs" included in the regional database 133. In this case, the disaster database 131 stores the "disaster area," "number of damaged facilities," "disaster density," "amount of damage," "amount of recovery," "recovery period," and "recovery operation" for each area identified by each of the multiple "regional IDs." By doing this, the information stored in the disaster database 131 can be aggregated based on the "disaster area" information, on a unit basis for areas identified by "regional IDs" included in the regional database 133.

[0031] Furthermore, all items included in the equipment DB132 shown in Figure 3 are associated with one of the "regional IDs" included in the regional DB133 based on the "installation location" information. This makes it possible to aggregate data at the regional level identified by the "regional ID".

[0032] Referring again to Figure 1, the prediction model DB 134 stores the prediction model constructed by the control unit 14, which will be described later. The disaster risk DB 135 stores the prediction results of the disaster risk of the equipment in the target area by the control unit 14, which will be described later.

[0033] The control unit 14 includes at least one processor, at least one programmable circuit, at least one dedicated circuit, or any combination thereof. The processor is a general-purpose processor such as a CPU (Central Processing Unit) or GPU (Graphics Processing Unit), or a dedicated processor specialized for a specific process. The programmable circuit is, for example, an FPGA (Field-Programmable Gate Array). The dedicated circuit is, for example, an ASIC (Application Specific Integrated Circuit).

[0034] The control unit 14 controls each part of the prediction device 10 and executes processes related to the operation of the prediction device 10. As shown in Figure 1, the control unit 14 functions as a data generation unit 141, a model construction unit 142, and a prediction unit 143.

[0035] The data generation unit 141 generates training data used to train the prediction model and prediction target data used to predict the disaster risk of facilities in the target area.

[0036] As explained with reference to Figure 2, the disaster DB 131 stores information about equipment damage, including multiple items (in the example shown in Figure 2, "type of disaster," "type of equipment," "area of ​​damage," "number of damaged equipment," and "damage density"). Also, as explained with reference to Figure 2, the disaster DB 131 includes indicators that show the time or monetary cost required to restore the damaged equipment (in the example shown in Figure 2, "amount of damage," "amount of restoration," "restoration period," and "restoration operation").

[0037] Furthermore, as explained with reference to Figure 3, the equipment DB132 stores information regarding the operation and maintenance of the equipment, including multiple items (in the example shown in Figure 3, "equipment ID," "installation location," "equipment type," "number of services accommodated," "importance of services accommodated," "difficulty of recovery," and "average maintenance operation").

[0038] Furthermore, as explained with reference to Figure 4, the regional DB133 stores information about the disaster resilience of a region, including multiple items (in the example shown in Figure 4, these include "coordinates," "labor force population," "total population," "fiscal volume index," "transportation index," "house density," "proportion of wooden structures," "proportion of forest area," "proportion of river area," and "average elevation").

[0039] The data generation unit 141 generates training data that associates information on damage to facilities caused by past disasters in disaster-stricken areas, information on the operation and maintenance of facilities in the disaster-stricken areas, information on the disaster resilience of the disaster-stricken areas, and an index indicating the time or monetary cost required to restore the facilities damaged by the past disasters.

[0040] Specifically, when generating training data, the input unit 11 accepts the following information: the affected area, the type of past disaster in the affected area, one or more items included in the information regarding damage to facilities in the affected area caused by those past disasters, one or more items included in the information regarding the operation and maintenance of facilities in the affected area, one or more items included in the information regarding the disaster resilience of the affected area, and an indicator showing the cost.

[0041] The data generation unit 141 obtains information on specified items from the disaster DB 131, equipment DB 132, and regional DB 133 regarding specified past disasters in the designated disaster area. The data generation unit 141 then generates training data that associates information on one or more specified items included in the information on damage to equipment in the disaster area, information on one or more specified items included in the information on the operation and maintenance of equipment in the disaster area, information on one or more specified items included in the information on the disaster resilience of the disaster area, and an index indicating the time or monetary cost required to restore equipment damaged by past disasters.

[0042] Furthermore, the data generation unit 141 generates predictive data that associates information regarding damage to facilities in the target area due to disasters, information regarding the operation and maintenance of facilities in the target area, and information regarding the disaster resilience of the target area.

[0043] Specifically, when generating prediction data, the input unit 11 accepts the following specifications: the area to be predicted for disaster risk, the type of disaster to be predicted, one or more items included in the information regarding damage to facilities in the area caused by the disaster, one or more items included in the information regarding the operation and maintenance of facilities in the area, and one or more items included in the information regarding the disaster resilience of the area.

[0044] The data generation unit 141 acquires information on specified items regarding a specified disaster in a specified target area from the disaster DB 131, the facility DB 132, and the area DB 133. Then, for the specified disaster in the specified mode area, the data generation unit 141 associates information on one or more specified items included in the information on the damage to facilities in the target area, information on one or more specified items included in the information on the supply and maintenance of facilities in the target area, and information on one or more specified items included in the information on the disaster resistance of the target area, and generates prediction target data.

[0045] The model construction unit 142 constructs a prediction model by learning the learning data generated by the data generation unit 141. Specifically, the model construction unit 142 constructs a prediction model by learning the learning data with an index indicating the time or monetary cost required for the restoration of damaged facilities due to past disasters included in the learning data as the objective variable. That is, the prediction model is constructed by learning the learning data in which information on the damage of facilities in a disaster area where a disaster has occurred in the past, information on the supply and maintenance of facilities in the disaster area, information on the disaster resistance of the disaster area, and information on the time or monetary cost required for the restoration of damaged facilities due to past disasters are associated, so as to predict the time or monetary cost required for the restoration of damaged facilities due to disasters.

[0046] The model construction unit 142 stores the constructed prediction model in the prediction model DB 134.

[0047] Note that in response to an input to the input unit 11, the data generation unit 141 may generate a plurality of learning data in which at least any one of the type of disaster, one or more items included in the information on the damage to facilities, one or more items included in the information on the supply and maintenance of facilities, one or more items included in the information on the disaster resistance, and the index indicating the cost is different. In this case, the model construction unit 142 constructs a plurality of prediction models by learning each of the generated plurality of learning data, and stores them in the prediction model DB 134.

[0048] The prediction unit 143 inputs the prediction target data generated by the data generation unit 141 into the prediction model constructed by the model construction unit 142, and predicts the disaster risk of the facilities in the target area due to disasters. Specifically, the prediction unit 143 inputs, into the prediction model, the prediction target data including information on damage to the facilities in the target area due to disasters, information on the supply and maintenance of the facilities in the target area, and information on the disaster resistance of the target area, and predicts, as the disaster risk of the facilities in the target area, the time or monetary cost required for the restoration of the damaged facilities due to disasters in the target area.

[0049] As described above, in the prediction model DB 134, a plurality of prediction models constructed by learning a plurality of learning data, each of which has at least one different item among the type of disaster, one or more items included in the information on damage to the facilities, one or more items included in the information on the supply and maintenance of the facilities, one or more items included in the information on the disaster resistance, and the index indicating the cost, are stored. The prediction unit 143 selects a prediction model corresponding to the prediction target data from the plurality of prediction models, inputs the prediction target data into the selected model, and predicts the disaster risk. That is, the prediction unit 143 uses, among the plurality of prediction models, a prediction model constructed by learning the learning data including the same type of disaster as the prediction target data, one or more items included in the information on damage to the facilities, one or more items included in the information on the supply and maintenance of the facilities, and one or more items included in the information on the disaster resistance, and having the same prediction target, to predict the disaster risk.

[0050] The functions of the prediction device 10 are realized by causing a processor as the control unit 14 to execute the program according to the present embodiment. That is, the functions of the prediction device 10 are realized by software. The program causes a computer to execute the operations of the prediction device 10, thereby causing the computer to function as the prediction device 10. That is, the computer functions as the prediction device 10 by executing the operations of the prediction device 10 according to the program.

[0051] The program can be stored on a non-temporary computer-readable medium. Examples of non-temporary computer-readable mediums include flash memory, magnetic recording devices, optical discs, magneto-optical recording media, or ROM. The program can be distributed, for example, by selling, transferring, or leasing portable media such as SD (Secure Digital) cards, DVDs (Digital Versatile Discs), or CD-ROMs (Compact Disc Read Only Memory) on which the program is stored. The program may also be distributed by storing it in server storage and transferring it from the server to other computers. The program may also be provided as a program product.

[0052] A computer, for example, stores a program stored on a portable medium or a program transferred from a server in its main memory. Then, the computer reads the program stored in the main memory with its processor and executes the processing according to the read program. The computer may also read the program directly from the portable medium and execute the processing according to the program. The computer may also execute the processing according to the received program sequentially each time a program is transferred to it from a server. Processing may also be performed by a so-called ASP (Application Service Provider) type service, which does not transfer programs from the server to the computer, but realizes its function only through execution instructions and result retrieval. A program includes information used for processing by an electronic computer that is equivalent to a program. For example, data that is not a direct instruction to the computer but has the nature of defining the computer's processing falls under "equivalent to a program".

[0053] Some or all of the functions of the prediction device 10 may be implemented by a programmable circuit or a dedicated circuit as the control unit 14. In other words, some or all of the functions of the prediction device 10 may be implemented by hardware.

[0054] Next, the operation of the prediction device 10 according to this embodiment will be described. Figure 5 is a flowchart showing an example of the operation of the prediction device 10 according to this embodiment when constructing a prediction model. Before predicting disaster risk, it is necessary to construct a prediction model at least once using the flow shown in Figure 5 for the type of disaster to be predicted and the indicator of the time or monetary cost required to restore damaged facilities due to the disaster as a disaster risk.

[0055] The input unit 11 accepts the specification of parameters related to the construction of the prediction model (step S11).

[0056] Specifically, the input unit 11 accepts the designation of the type of disaster to be studied and the designation of past disasters. The designation of past disasters is done by specifying the "Disaster ID" of the disaster DB 131. Multiple "Disaster IDs" associated with disasters of the same type as the designated disaster type may be specified.

[0057] Furthermore, the input unit 11 accepts the specification of target information to be used for training the prediction model. Specifically, the input unit 11 accepts the specification of one or more items included in the information on damage to equipment, one or more items included in the information on the use and maintenance of equipment, and one or more items included in the information on disaster resilience, to be used for training the prediction model. By accepting the specification of items of each piece of information to be used for training the prediction model, if there are items that are partially missing, those items can be excluded from the training data.

[0058] Furthermore, the input unit 11 accepts the specification of a disaster risk indicator (an indicator showing the time or monetary cost required to restore damaged equipment) to be treated as the target variable when training the prediction model. Specifying a disaster risk indicator involves selecting one indicator from among the indicators showing the time or monetary cost required to restore damaged equipment stored in the disaster DB 131 (in the example shown in Figure 2, these are "damage amount," "restoration amount," "restoration period," and "restoration operation").

[0059] The data generation unit 141 acquires information corresponding to the specified parameters (type of disaster, past disasters, target information, and disaster risk indicators) from the disaster DB 131, equipment DB 132, and regional DB 133, and generates training data based on the acquired information (step S12).

[0060] Figure 6 shows an example of training data generated by the data generation unit 141. In Figure 6, an example is shown where earthquake is specified as the type of disaster. In Figure 6, the target information is shown as follows: "Labor force population," "Fiscal strength index," "Transportation index," "House density," "Wooden structure density," "Forest area ratio," "River area ratio," and "Average elevation" stored in the regional DB 133; "Equipment type," "Number of services provided," "Importance of services provided," "Maintenance difficulty," "Restoration difficulty," and "Average maintenance operation" stored in the equipment DB 132; and "Number of damaged facilities" and "Damage density" stored in the disaster DB 131. In Figure 6, an example is shown where "Restoration cost" stored in the disaster DB 131 is specified as an indicator of cost.

[0061] As shown in Figure 6, the data generation unit 141 generates training data that associates information on one or more specified items included in information on damage to facilities in the affected area, information on one or more specified items included in information on the operation and maintenance of facilities in the affected area, information on one or more specified items included in information on the disaster resilience of the affected area, and an index showing the costs of past disasters. The data generation unit 141 aggregates each piece of information for each region identified by the "regional ID" included in the regional DB 133 to generate the training data.

[0062] Referring again to Figure 5, the model building unit 142 builds a predictive model that predicts disasters by listing an indicator showing a specified cost, based on the training data generated by the data generation unit 141 (step S13). Specifically, the model building unit 142 builds a predictive model by training the training data with the indicator showing cost included in the training data as the target variable. The model building unit 142 builds a predictive model that predicts disasters by listing an indicator showing a specified cost, based on the training data, based on any method such as regression analysis, decision trees, neural networks, and methods applying these. The model building unit 142 may change the training method depending on the disaster risk indicator used as the target variable and the quantity of training data.

[0063] The model building unit 142 stores the constructed prediction model in the prediction model DB 134 (step S14), and then terminates the process.

[0064] Figure 7 is a flowchart showing an example of the operation of the prediction device 10 according to this embodiment when predicting disaster risk, and is a diagram for explaining the prediction method executed by the prediction device 10 according to this embodiment.

[0065] The input unit 11 accepts the specification of parameters related to disaster risk prediction (step S21).

[0066] Specifically, the input unit 11 accepts the designation of the type of disaster to be predicted and the target area for which the disaster risk is predicted. The designation of the target area is done by specifying the "Region ID" of the region DB 133. Multiple "Region IDs" may be specified depending on the area for which the disaster risk is predicted.

[0067] Furthermore, the input unit 11 accepts the designation of target information to be used for disaster risk prediction. Specifically, the input unit 11 accepts the designation of one or more items included in the information on damage to equipment, one or more items included in the information on the operation and maintenance of equipment, and one or more items included in the information on disaster resilience, to be used for disaster risk prediction. By accepting the designation of items for each piece of information used for disaster risk prediction, if some items are missing, those items can be excluded from the prediction target data.

[0068] The input unit 11 also accepts the designation of an indicator of predicted disaster risk (an indicator showing the time or monetary cost required to restore damaged equipment). Designating a disaster risk indicator involves selecting one indicator from among those stored in the disaster DB 131 that show the time or monetary cost required to restore damaged equipment (in the example shown in Figure 2, these are "damage amount," "restoration amount," "restoration period," and "restoration operation").

[0069] The data generation unit 141 obtains information corresponding to the specified parameters (type of disaster, target area, target information) from the disaster DB 131, equipment DB 132, and regional DB 133, and generates prediction target data based on the obtained information (step S22).

[0070] The prediction unit 143 inputs the generated data to be predicted into the prediction model to predict the disaster risk (step S23). As described above, multiple prediction models are constructed by training multiple learning models with different target information, etc., and are stored in the prediction model DB 134. The prediction unit 143 selects a prediction model from among the multiple stored prediction models that corresponds to the specified disaster type, target area, target information, and disaster risk indicator, inputs the data to be predicted into the selected prediction model, and predicts the disaster risk.

[0071] The prediction unit 143 stores the disaster risk prediction results in the disaster risk DB 135 (step S24), and then terminates the process.

[0072] As described above, the prediction device 10 according to this embodiment includes a prediction unit 143. The prediction unit 143 inputs prediction target data, consisting of information on damage to facilities in the target area due to a disaster, information on the operation and maintenance of facilities in the target area, and information on the disaster resilience of the target area, into a prediction model that predicts the time or monetary cost required to restore facilities damaged by a disaster in the target area, and predicts the time or monetary cost required to restore facilities damaged by a disaster in the target area as the disaster risk of the facilities in the target area.

[0073] By predicting disaster risk using information on the operation and maintenance of facilities in the target area, as well as information on the disaster resilience of the target area, it is possible to predict the disaster risk of facilities from a more multifaceted perspective, taking into account the operation and maintenance of facilities and the disaster resilience of the area.

[0074] The following additional information is disclosed regarding the embodiments described above.

[0075] [Addendum 1] A prediction device for predicting disaster risk to facilities in a target area due to a disaster, comprising a control unit, wherein the control unit is configured to input prediction target data consisting of information on damage to facilities in the target area due to the disaster, information on the operation and maintenance of facilities in the target area, and information on the disaster resilience of the target area into a prediction model that predicts the time or monetary cost required to restore facilities damaged by the disaster in the target area as the disaster risk to the facilities in the target area.

[0076] [Appendix Item 2] A prediction device as described in Appendix Item 1, wherein the control unit generates learning data that associates information on damage to the equipment due to past disasters in disaster-stricken areas where disasters have occurred in the past, information on the operation and maintenance of the equipment in the disaster-stricken areas, information on the disaster resilience of the disaster-stricken areas, and an index indicating the time or monetary cost required to restore the equipment damaged by past disasters, and constructs the prediction model by learning the learning data with the index indicating the time or monetary cost required to restore the equipment damaged by past disasters as the target variable.

[0077] [Appendix 3] The prediction device described in Appendix 2, wherein the information relating to damage to the equipment, the information relating to the use and maintenance of the equipment, and the information relating to disaster resilience each include multiple items, and further comprises an input unit that accepts the designation of the affected area, the type of past disaster in the affected area, one or more items included in the information relating to damage to the equipment in the affected area due to the past disaster, one or more items included in the information relating to the use and maintenance of the equipment in the affected area, one or more items included in the information relating to the disaster resilience of the affected area, and an index indicating the cost, and the control unit generates the learning data in which, for the designated past disaster in the designated affected area, the information relating to the designated one or more items included in the information relating to damage to the equipment in the affected area, the information relating to the designated one or more items included in the information relating to the use and maintenance of the equipment in the affected area, the information relating to the designated one or more items included in the information relating to the disaster resilience of the affected area, and the index indicating the designated cost.

[0078] [Appendix 4] A prediction device according to appendix 2 or 3, wherein the control unit further generates prediction target data in which information relating to damage to the facilities in the target area due to the disaster, information relating to the operation and maintenance of the facilities in the target area, and information relating to the disaster resilience of the target area.

[0079] [Appendix 5] A prediction device as described in Appendix 4, wherein the information relating to damage to the equipment, the information relating to the use and maintenance of the equipment, and the information relating to disaster resilience each include multiple items, the input unit receives the designation of the target area, the type of disaster for which the disaster risk is predicted, one or more items included in the information relating to damage to the equipment in the target area due to the disaster, one or more items included in the information relating to the use and maintenance of the equipment in the target area, and one or more items included in the information relating to the disaster resilience of the target area, and the control unit generates prediction target data for the designated disaster in the designated target area, in which the information relating to the designated one or more items included in the information relating to damage to the equipment in the target area, the information relating to the designated one or more items included in the information relating to the use and maintenance of the equipment in the target area, and the information relating to the designated one or more items included in the information relating to the disaster resilience of the target area.

[0080] [Appendix 6] A prediction device according to any one of Appendix 2 to 5, wherein the information relating to damage to the equipment, the information relating to the use and maintenance of the equipment, and the information relating to disaster resilience each include multiple items, the control unit constructs multiple prediction models by learning each of multiple training data sets in which at least one of the following is different: the type of disaster, one or more items included in the information relating to damage to the equipment, one or more items included in the information relating to the use and maintenance of the equipment, one or more items included in the information relating to disaster resilience, and the cost indicator, the prediction device selects a prediction model from among the multiple prediction models according to the data to be predicted, inputs the data to be predicted into the selected model and predicts the disaster risk. [Addendum 7] A prediction method performed by a prediction device for predicting disaster risk to facilities in a target area due to a disaster, wherein prediction target data consisting of information on damage to facilities in the target area due to the disaster, information on the operation and maintenance of facilities in the target area, and information on the disaster resilience of the target area is input into a prediction model for predicting the time or monetary cost required to restore facilities damaged by the disaster, and the time or monetary cost required to restore facilities damaged by the disaster in the target area is predicted as the disaster risk of facilities in the target area.

[0081] [Appendix 8] A non-temporary storage medium storing a program executable by a computer, the non-temporary storage medium storing a program that causes the computer to function as a control device as described in any one of the appendix items 1 to 6.

[0082] Although the embodiments described above are representative examples, it will be apparent to those skilled in the art that many modifications and substitutions are possible within the spirit and scope of this disclosure. Therefore, the present invention should not be construed as being limited by the embodiments described above, and various modifications or changes are possible without departing from the claims. For example, it is possible to combine multiple component blocks shown in the configuration diagram of the embodiments into one, or to divide one component block.

[0083] 10 Prediction device 11 Input unit 12 Output unit 13 Storage unit 14 Control unit 131 Disaster DB 132 Equipment DB 133 Regional DB 134 Prediction model DB 135 Disaster risk DB 141 Data generation unit 142 Model construction unit 143 Prediction unit

Claims

1. A prediction device for predicting disaster risk to facilities in a target area due to a disaster, comprising a prediction unit that inputs prediction target data consisting of information on damage to facilities in the target area due to the disaster, information on the operation and maintenance of facilities in the target area, and information on the disaster resilience of the target area into a prediction model that predicts the time or monetary cost required to restore facilities damaged by the disaster in the target area as the disaster risk of the facilities in the target area.

2. A prediction device according to claim 1, comprising: a data generation unit that generates learning data associated with information on damage to the equipment caused by past disasters in disaster-stricken areas where disasters have occurred in the past, information on the operation and maintenance of the equipment in the disaster-stricken areas, information on the disaster resilience of the disaster-stricken areas, and an index indicating the time or monetary cost required to restore the equipment damaged by past disasters; and a model construction unit that constructs the prediction model by learning the learning data with the index indicating the time or monetary cost required to restore the equipment damaged by past disasters as the target variable.

3. The prediction device according to claim 2, wherein the information relating to damage to the equipment, the information relating to the use and maintenance of the equipment, and the information relating to disaster resilience each include a plurality of items, and further comprises an input unit that accepts the specification of the affected area, the type of past disaster in the affected area, one or more items included in the information relating to damage to the equipment in the affected area due to the past disaster, one or more items included in the information relating to the use and maintenance of the equipment in the affected area, one or more items included in the information relating to the disaster resilience of the affected area, and an index indicating the cost, and the data generation unit generates the learning data in which, for the specified past disaster in the specified affected area, the information relating to the specified one or more items included in the information relating to damage to the equipment in the affected area, the information relating to the specified one or more items included in the information relating to the use and maintenance of the equipment in the affected area, the information relating to the specified one or more items included in the information relating to the disaster resilience of the affected area, and the index indicating the specified cost.

4. A prediction device according to claim 2, wherein the data generation unit further generates prediction target data, which is associated with information relating to damage to the facilities in the target area due to the disaster, information relating to the operation and maintenance of the facilities in the target area, and information relating to the disaster resilience of the target area.

Images