Integrated natural disaster monitoring apparatus and method based on ultra-large language model

Abstract

The present invention provides an integrated natural disaster monitoring apparatus and method based on an ultra-large language model. The integrated natural disaster monitoring apparatus based on an ultra-large language model according to the present invention comprises: a basic data collection unit for collecting basic data corresponding to each of two or more natural disasters; a natural disaster monitoring information generation unit for generating first natural disaster monitoring information for each natural disaster on the basis of a natural disaster data pool in which the collected basic data are integrated or generating second natural disaster monitoring information by integrating the generated first natural disaster monitoring information; and a natural disaster monitoring result provision unit for generating and outputting a prediction diagnosis result affecting an area of interest and decision-making information required for the area of interest by inputting natural disaster analysis request information including the first natural disaster monitoring information and the second natural disaster monitoring information to a diagnosis prediction module based on an ultra-large language model.

Description

Integrated natural disaster monitoring device and method based on a super-large language model

[0001] The present invention relates to a natural disaster integrated monitoring device and method based on a large-scale language model, and more particularly, to a natural disaster integrated monitoring device and method based on a large-scale language model for performing safety measures when a natural disaster occurs.

[0002] In the national and public utility industries (e.g., nuclear power plants), the requirements of these industries are evolving in line with the advancement of AI technology, and the level of technology is advancing to reflect this.

[0003] Furthermore, the national and public facility industries (e.g., nuclear power plants) are industries that can only be sustained by ensuring reliable safety. They are affected by climate and the environment, and response measures to natural disasters that cause these effects are continuously being researched and developed and reflected in industrial sites.

[0004] Recently, by utilizing the weather information system and forest fire surveillance system provided by the Korea Meteorological Administration and the Korea Forest Service, the climate and environmental impacts on areas where national and public facility industries (e.g., nuclear power plants) are located are being monitored, and the earthquake surveillance system equipped with a seismometer is also operated as an independent system separate from the aforementioned weather information system and forest fire surveillance system to monitor the climate and environmental impacts on areas where national and public facility industries (e.g., nuclear power plants) are located.

[0005] Based on the operation of each system, by monitoring the climate and environment, in the event of a natural disaster (e.g., weather warning, earthquake, tsunami, wildfire, typhoon, etc.), safety managers of the national and public facility industries (e.g., nuclear power plants) check the early recommendations against the pre-determined facility action level table based on the alerts notified from each of the aforementioned systems, and make decisions on whether to operate the facility in question based on the confirmed action recommendations.

[0006] This conventional natural disaster monitoring system method is operated in a distributed and independent structure based on information provided by public institutions, and is not provided in a way that provides information specialized to the area where the national and public facility industry (e.g., nuclear power plant) is located, but rather in a way that notifies on a nationwide basis. Therefore, the operators of the national and public facility industry (e.g., nuclear power plant) individually control the information notified on a nationwide basis, and the information controlled individually is collected based on their own criteria that are not systematic or three-dimensional, and then they decide whether to operate the facility in question.

[0007] In addition, even when trying to build a system that reflects the impact of climate and environment based on the area where national and public facility industries (e.g. nuclear power plants) are located, there are some parts that are not easy to actually build because when national security facilities are 'A' level facilities, external exposure of related information such as facilities is blocked, and there are limitations in building through linkage with external systems.

[0008] That is, there are various online-based LM models currently in use, such as GPT, Bard, LLaMa, and Vicuna, and in order to utilize the models, you can obtain answers by utilizing APIs in an environment where the Internet is available, but in the case of data that is not common, such as nuclear power plants, there is a problem that the model outputs incorrect answers, and although learning must be done based on the procedures, instructions, and documents of the actual facility, in the case of national infrastructure 'A' grade, there is a problem that it is difficult to export the relevant data, so as mentioned above, it is difficult to advance the system to respond to climate and environmental impacts on the national and public facility industry (e.g. nuclear power plants).

[0009] As a result, operators of national and public facility industries (e.g., nuclear power plants) frequently check climate and environmental warnings and, even when it is not necessary to stop the operation of the facility, take excessive measures to stop the operation of the facility, resulting in unnecessary damage (e.g., in the case of nuclear power plants, failure to meet planned energy production due to facility shutdown, or increased carbon emissions due to restarting of large-scale facilities that have been shut down). In other cases, there are concerns that safety accidents may occur due to insufficient information on specific regions where national and public facility industries (e.g., nuclear power plants) are located, or due to the failure to properly control information on some natural disasters as numerous natural disasters occur, resulting in appropriate measures not being taken within a given time.

[0010] This is a time when a solution that reflects this is needed.

[0011] Accordingly, the present invention was created to solve the above problems, and the purpose of the present invention is to provide a natural disaster integrated monitoring device and method based on a super-large language model that enables specialized safety measures to be taken for each area of ​​interest (e.g., the area around a nuclear power plant) when a natural disaster occurs.

[0012] The purpose of the invention is not limited to the purposes mentioned above, and other purposes not mentioned will be clearly understood by those skilled in the art from the description below.

[0013] In order to achieve the above object, a natural disaster integrated monitoring device based on a super-large language model according to a first aspect of the present invention includes a basic data collection unit that collects basic data corresponding to two or more natural disasters, a natural disaster monitoring information generation unit that generates first natural disaster monitoring information for each natural disaster based on a natural disaster data pool that integrates the collected basic data, or generates second natural disaster monitoring information that integrates the generated first natural disaster monitoring information, and a natural disaster monitoring result provision unit that generates and outputs a prediction diagnosis result affecting an area of ​​interest and decision-making information required for the area of ​​interest by inputting natural disaster analysis request information including the first natural disaster monitoring information and the second natural disaster monitoring information into a diagnosis and prediction module based on a super-large language model.

[0014] The above basic data collection unit may include a data collection path that receives and collects the basic data from an API (Application Programming Interface) corresponding to each of the two or more natural disasters.

[0015] The above natural disaster monitoring information generation unit can convert the first natural disaster monitoring information and the second natural disaster monitoring information based on the area of ​​interest.

[0016] The above second natural disaster monitoring information may include additional monitoring parameter information additionally derived based on the interaction between the first natural disaster monitoring information for each natural disaster based on the area of ​​interest.

[0017] The above natural disaster monitoring result provision unit sets a reference radius according to the characteristics of the equipment located in the area of ​​interest, and can adjust the monitoring level of the area of ​​interest by expanding or reducing the reference radius of the area of ​​interest based on the predicted diagnosis result.

[0018] The above decision-making information may include one or more pieces of action information that should be taken for the area of ​​interest based on the predicted diagnosis results, or may include one or more pieces of preemptive action information recommended for the area of ​​interest based on the predicted simulation results within a predetermined future period.

[0019] The above natural disaster analysis request information and facility information related to the above area of ​​interest may be input into the above diagnosis prediction module after undergoing encryption processing or equivalent security processing based on standards that meet the external security requirements of the above area of ​​interest.

[0020] The above natural disaster monitoring result provision unit generates the decision-making information for the above area of ​​interest, and if there is one or more sub-areas of interest within the influence area of ​​the above prediction diagnosis result among the surrounding areas of the above area of ​​interest, it can additionally generate and output natural disaster response information corresponding to the sub-areas of interest.

[0021] And, according to the second aspect of the present invention for achieving the above object, the integrated natural disaster monitoring method based on the ultra-large language model includes a data collection step of collecting basic data corresponding to two or more natural disasters in a natural disaster integrated monitoring device, a monitoring information generation step of generating first natural disaster monitoring information for each natural disaster based on a natural disaster data pool that integrates the collected basic data, or generating second natural disaster monitoring information that integrates the generated first natural disaster monitoring information, and a monitoring result provision step of generating and outputting a prediction diagnosis result affecting an area of ​​interest and decision-making information required for the area of ​​interest by inputting natural disaster analysis request information including the first natural disaster monitoring information and the second natural disaster monitoring information into a diagnosis prediction module based on the ultra-large language model.

[0022] The above data collection step may include a step of collecting the basic data by receiving it from an API (Application Programming Interface) corresponding to each of the two or more natural disasters.

[0023] The above surveillance information generation step may include a step of converting the first natural disaster surveillance information and the second natural disaster surveillance information based on the area of ​​interest.

[0024] The above second natural disaster monitoring information may include additional monitoring parameter information additionally derived based on the interaction between the first natural disaster monitoring information for each natural disaster based on the area of ​​interest.

[0025] The above monitoring result provision step may include a step of setting a reference radius according to the characteristics of the equipment located in the area of ​​interest, and adjusting the monitoring level of the area of ​​interest by expanding or reducing the reference radius of the area of ​​interest based on the predicted diagnosis result.

[0026] The above decision-making information may include one or more pieces of action information that should be taken for the area of ​​interest based on the predicted diagnosis results, or may include one or more pieces of preemptive action information recommended for the area of ​​interest based on the predicted simulation results within a predetermined future period.

[0027] The above-mentioned monitoring result provision step may include a step of inputting the natural disaster analysis request information and the facility information related to the area of ​​interest into the diagnosis prediction module after undergoing encryption processing or equivalent security processing based on standards that meet the external security requirements of the area of ​​interest.

[0028] The above monitoring result provision step may include a step of generating the decision-making information for the area of ​​interest, and, if there is one or more sub-areas of interest within the sphere of influence of the predicted diagnosis result among the surrounding areas of the area of ​​interest, additionally generating and outputting natural disaster response information corresponding to the sub-areas of interest.

[0029] Accordingly, the present invention has the advantage of being able to perform specialized safety measures for each area of ​​interest (e.g., the area surrounding a nuclear power plant) when a natural disaster occurs, and even when two or more natural disasters occur, by analyzing the impact of the combined natural disasters, it is possible to perform safety measures optimized for the area of ​​interest.

[0030] The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description of the claims.

[0031] Figure 1 is a configuration diagram showing an integrated natural disaster monitoring device according to one embodiment of the present invention.

[0032] Figure 2 is an exemplary diagram showing another example of applying the integrated natural disaster monitoring device of Figure 1.

[0033] Figure 3 is an example diagram showing the first screen in which the natural disaster integrated monitoring device of Figure 1 is operated.

[0034] Figure 4 is an example diagram showing a second screen that operates the natural disaster integrated monitoring device of Figure 1.

[0035] Figure 5 is an example diagram showing the third screen that operates the natural disaster integrated monitoring device of Figure 1.

[0036] Figure 6 is an example diagram showing the fourth screen that operates the natural disaster integrated monitoring device of Figure 1.

[0037] Figure 7 is an example diagram showing the fifth screen that operates the natural disaster integrated monitoring device of Figure 1.

[0038] Figure 8 is an example diagram showing the sixth screen that operates the natural disaster integrated monitoring device of Figure 1.

[0039] Figure 9 is a flowchart showing a natural disaster integrated monitoring method according to one embodiment of the present invention.

[0040] Fig. 10 is a flowchart showing an example of an alarm generation algorithm applied to the integrated natural disaster monitoring method of Fig. 9.

[0041] Figure 11 is a flowchart showing another example of an alarm generation algorithm applied to the integrated natural disaster monitoring method of Figure 9.

[0042] Figure 12 is a flowchart showing another example of an alarm generation algorithm applied to the integrated natural disaster monitoring method of Figure 9.

[0043] Figure 13 is a flowchart showing another example of an alarm generation algorithm applied to the integrated natural disaster monitoring method of Figure 9.

[0044] Figure 14 is a flowchart showing another example of an alarm generation algorithm applied to the integrated natural disaster monitoring method of Figure 9.

[0045] And, Fig. 15 is a flowchart showing an example of a similar natural disaster search algorithm applied to the natural disaster integrated monitoring method of Fig. 9.

[0046] The advantages and features of the present invention, and the methods for achieving them, will become clearer with reference to the embodiments described in detail below together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below and may be implemented in various different forms. These embodiments are provided only to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined only by the scope of the claims. Like reference numerals designate like elements throughout the specification.

[0047] Furthermore, the embodiments described herein will be described with reference to cross-sectional and / or schematic drawings, which are ideal illustrations of the present invention. Therefore, the form of the illustrations may be modified due to manufacturing techniques and / or tolerances. Furthermore, in each drawing illustrated in the present invention, each component may be depicted somewhat enlarged or reduced for convenience of explanation.

[0048] The integrated natural disaster monitoring device based on the ultra-large language model of the present invention can perform specialized safety measures for each area of ​​interest when a natural disaster occurs, and is configured to enable optimized safety measures for each area of ​​interest by analyzing the impact of the integrated natural disasters even when two or more natural disasters occur.

[0049] Figure 1 is a configuration diagram showing an integrated natural disaster monitoring device according to one embodiment of the present invention.

[0050] As illustrated in FIG. 1, a natural disaster integrated monitoring device (100) based on a large-scale language model includes a basic data collection unit (110) that collects basic data corresponding to two or more natural disasters, a natural disaster monitoring information generation unit (120) that generates first natural disaster monitoring information for each natural disaster based on a natural disaster data pool that integrates the collected basic data, or generates second natural disaster monitoring information that integrates the generated first natural disaster monitoring information, and a natural disaster monitoring result provision unit (130) that inputs natural disaster analysis request information including the first natural disaster monitoring information and the second natural disaster monitoring information into a diagnosis and prediction module based on a large-scale language model, thereby generating and outputting a prediction and diagnosis result affecting an area of ​​interest and decision-making information required for the area of ​​interest.

[0051] Here, the area of ​​interest refers to the area where the impact of natural disasters can be monitored and measures taken based on the monitoring results. Examples include the area surrounding national and public utility facilities. Below, we will provide a more detailed explanation, assuming the area surrounding a nuclear power plant.

[0052] Furthermore, super-large language models are models with tens to hundreds of billions of parameters, distinguishing them from giant language models with hundreds of millions to billions of parameters. For example, GPT-3, with 175 billion parameters, and GPT-4, with even more parameters, are considered super-large language models, while GPT-2, with 1.5 billion parameters, is considered a large language model.

[0053] The present invention relates to industrial facilities mainly applied to cutting-edge facilities such as nuclear power plants, and in order to generate precise decision-making information, it is preferable to base it on an ultra-large language model capable of understanding the context of technical documents, complex queries and their responses, and even zero-shot learning, rather than a large language model.

[0054] The natural disaster monitoring result provision unit (130) can provide decision-making information for one or more areas of interest, and can receive an information provision request from a user device used by an operator in a nuclear power plant located in the area of ​​interest and provide decision-making information as a response, or can automatically provide decision-making information so that an operator in a nuclear power plant located in the area of ​​interest can confirm it when a meaningful natural disaster prediction diagnosis result occurs at a predetermined cycle or for the area of ​​interest.

[0055] The basic data collection unit (110) may include a data collection path that collects basic data from an API (Application Programming Interface) corresponding to two or more natural disasters, and examples of the APIs described above may include APIs of public institutions' systems, such as APIs of a weather observation system and APIs of a forest fire monitoring system.

[0056] The first natural disaster monitoring information refers to information categorized by natural disaster, such as weather warnings, typhoon information, earthquake information, and forest fire information, and the second natural disaster monitoring information refers to information that integrates two or more pieces of natural disaster monitoring information and can be used as information to monitor the impact of natural disasters that are difficult to confirm with only the first natural disaster monitoring information described above.

[0057] The natural disaster monitoring information generation unit (120) can convert the first natural disaster monitoring information and the second natural disaster monitoring information based on the area of ​​interest, which can be used as information for specialized analysis of the impact of natural disasters on the area of ​​interest based on the first natural disaster monitoring information and the second natural disaster monitoring information.

[0058] In the case of an earthquake occurring due to natural disaster A, the first natural disaster monitoring information may include information on the reference location where natural disaster A occurred, which is '42 km northwest of Gilju, North Hamgyong Province, North Korea', information on the distance between the nuclear power plant within the area of ​​interest and the reference location of the aforementioned natural disaster A, information on the type and intensity of the earthquake, and derived information on the geological layer between the aforementioned reference location and the area of ​​interest and the seismic impact transmitted to this geological layer.

[0059] In addition, the second natural disaster surveillance information may be generated as information that integrates the first natural disaster surveillance information that is provided continuously or discontinuously in response to natural disaster A, or may be generated by supplementing the first natural disaster surveillance information provided at a specific point in time or updating it by adding information items by looking up the first natural disaster surveillance information provided at a specific point in time and information on past earthquake history that affected the area of ​​interest.

[0060] Figure 2 is an exemplary diagram showing another example of applying the integrated natural disaster monitoring device of Figure 1.

[0061] The second natural disaster surveillance information may include additional surveillance parameter information derived from the interaction between the first natural disaster surveillance information for each natural disaster based on the area of ​​interest.

[0062] For example, in the event of a forest fire and a typhoon, if the location of the forest fire is within an area set as an area of ​​interest or within a wide-area of ​​interest that may affect the area of ​​interest, and the area of ​​interest or wide-area of ​​interest is included in the area of ​​strong winds caused by the typhoon, and the impact of the forest fire and typhoon on the area of ​​interest must be monitored, soil erosion information for analyzing the possibility of landslide occurrence according to the interaction between the movement path of the forest fire and the wind speed / wind direction of the typhoon can be set as the additional monitoring parameter information described above.

[0063] As illustrated in FIG. 2, when natural disasters corresponding to the first area of ​​interest occur, such as natural disaster A (e.g., earthquake), natural disaster B (e.g., forest fire), and natural disaster C (e.g., typhoon), information on the impact and necessary measures that may be caused by the natural disasters on the first area of ​​interest is provided for each natural disaster, and as in the example described above, information on integrated analysis of natural disaster A (e.g., earthquake), natural disaster B (e.g., forest fire), and natural disaster C (e.g., typhoon) can be provided, so that natural disasters affecting the first area of ​​interest can be monitored in an integrated manner at a glance, and even when multiple natural disasters occur at once, it is possible to respond without missing any considerations or measures.

[0064] The natural disaster monitoring information generation unit (120) may be configured to use an offline-based GIS (Geographic Information System) because it is not easy to utilize public maps due to the characteristics of national facilities such as nuclear power plants located in areas of interest, and it is possible to implement user-defined GIS location information so that it can be utilized for specialized monitoring of areas of interest (e.g. nuclear power plants).

[0065] The natural disaster monitoring result provision unit (130) sets a reference radius according to the characteristics of the equipment located in the area of ​​interest, and can adjust the monitoring level of the area of ​​interest by expanding or reducing the reference radius of the area of ​​interest based on the prediction diagnosis results.

[0066] The area of ​​interest can be set to a nuclear power plant, in which case the radius of the transmission line and the surrounding radius based on the location of the nuclear power plant can be set, and the settings for the first natural disaster monitoring information and the second natural disaster monitoring information and the decision-making information can be set to match the characteristics of the nuclear power plant.

[0067] Decision information may include one or more action information items that should be taken for the area of ​​interest based on the predicted diagnosis results, or one or more preemptive action information items that are recommended for the area of ​​interest based on the predicted simulation results within a predetermined future period.

[0068] By simulating the forest fire movement path based on wind speed / direction information when a forest fire occurs, it is possible to provide information on the prediction of the threat impact and damage situation of power transmission lines in the area of ​​interest from the current point in time, and based on this, one or more pieces of preemptive action information can be generated.

[0069] When a typhoon occurs, the history of similar typhoons in the past is searched and the typhoon's path is simulated, thereby providing information on the threat impact and damage situation prediction for nuclear power plants in the area of ​​interest from the present point in time, and based on this, one or more pieces of preemptive action information can be generated.

[0070] The integrated natural disaster monitoring device (100) of the present invention can provide a main screen through which a service manager who provides integrated natural disaster monitoring and action service through the integrated natural disaster monitoring device (100) or a service user in an area of ​​interest (e.g., an area around a nuclear power plant) can manage and use the integrated natural disaster monitoring and action service.

[0071] The main screen may include a menu list that allows you to mute alarms, set general service items, and manage basic information of key service factors (change, add, delete, and hide location information of natural disasters, location information of nuclear power plants within the area of ​​interest, and location information of substations connected to nuclear power plants within the area of ​​interest), a key location navigator function that allows you to zoom in on a selected area of ​​interest among multiple areas of interest, a real-time basic information display function that displays information set as basic display items, such as real-time weather information for each area of ​​interest, and a function that displays geographic information, including the location of nuclear power plants located within the area of ​​interest and the location information of substations and transmission lines connected to nuclear power plants.

[0072] Figure 3 is an example diagram showing the first screen in which the natural disaster integrated monitoring device of Figure 1 is operated.

[0073] As illustrated in Fig. 3, the integrated natural disaster monitoring device (100) can display the location of a forest fire on the management screen of a service manager or the use screen of a service user when a forest fire occurs as a natural disaster (1).

[0074] Here, the location of a forest fire may be displayed on the service user's screen by setting whether or not to display it based on the surveillance level, including the distance between the location of the forest fire and the area of ​​interest.

[0075] For this purpose, information (2) on the distance between the location where a forest fire occurred and the nuclear power plant in the area of ​​interest is managed, and it is also possible to display on the screen the first natural disaster surveillance information including the location of the forest fire and weather information that can be used to assess the environmental impact and determine the extent of the forest fire's duration.

[0076] The integrated natural disaster monitoring device (100) can generate and provide second natural disaster monitoring information by integrating information, including the first natural disaster monitoring information described above. To this end, based on the location where a forest fire occurred, a simulation is run to predict the forest fire situation 3 hours and 6 hours from the current point in time. As a result, the second natural disaster monitoring information, including wind information around the forest fire 3 hours in advance and wind information around the forest fire 6 hours in advance, can be displayed on the service manager screen or the service user screen.

[0077] In addition, information on the location where a forest fire occurred, the location where it is spreading, and the adjacent area of ​​interest around the forest fire, as well as information on the distance to the adjacent area of ​​interest, can be included in the first natural disaster surveillance information and displayed on the service manager screen or the service user screen. It is also possible to process the second natural disaster surveillance information described above based on the adjacent area of ​​interest and display it on the service manager screen or the service user screen.

[0078] The integrated natural disaster monitoring device (100) can visualize the location of a typhoon in real time by displaying the typhoon's actual and expected movement paths and strong wind / storm radius on a map using an offline-based GIS (Geographic Information System).

[0079] In addition, the integrated natural disaster monitoring device (100) can display the basic information of the typhoon described above, the current location and the expected location change thereafter, and can provide first natural disaster monitoring information including detailed typhoon information that can determine the degree of influence of the typhoon at the current time for each area of ​​interest.

[0080] If there are multiple typhoons that have occurred, you can select a specific typhoon and it is also possible to provide detailed information about the selected typhoon, such as its direction of travel, central pressure, and radius.

[0081] As previously explained, it provides basic and detailed information about the typhoon, and the collected first natural disaster monitoring information that predicts future changes in the typhoon can also be displayed on the service manager screen or service user screen. In addition, information related to the typhoon's influence on each area of ​​interest and the typhoon response level table can be displayed together.

[0082] Additionally, typhoon details, including the typhoon's direction of movement, center, and radius, can be displayed on the service manager screen or service user screen at each predicted mutation point.

[0083] FIG. 4 is an exemplary diagram showing a second screen in which the natural disaster integrated monitoring device of FIG. 1 is operated, and FIG. 5 is an exemplary diagram showing a third screen in which the natural disaster integrated monitoring device of FIG. 1 is operated.

[0084] As shown in Figure 4, in order to predict the movement path of the typhoon currently being monitored, typhoons that have moved along a similar path among the typhoons that have occurred in the past can be searched and visualized on the screen.

[0085] In addition, as shown in Fig. 5, the future movement of a typhoon currently being monitored can be simulated (1), and based on the simulation results, the period from the occurrence to the disappearance of the typhoon can be displayed, and a timeline function can be provided to view a specific point in time during the displayed period through means such as cursor movement by a service manager or service user (2).

[0086] Even when searching for a typhoon simulation at a specific point in the future as mentioned above, predicted information such as the direction of movement, central pressure, and radius of the typhoon can be displayed as detailed information (3).

[0087] The natural disaster integrated monitoring device (100) displays the contents of a special weather report at the bottom right of the screen when a special weather report is issued, and can provide a function to visualize the special report information on a map depending on the type of special report.

[0088] At this time, if the location of the issued weather warning is included in the area of ​​interest, the weather warning information can be provided as visual information matching the area of ​​interest.

[0089] The integrated natural disaster monitoring device (100) provides a function for retrieving information on earthquakes occurring from points affecting an area of ​​interest, and can display the location where the actual earthquake occurred on a map, along with information on the distance from the location of the earthquake to the area of ​​interest. Here, if there are multiple areas of interest, distance information between each of the multiple areas of interest and the location of the earthquake can be displayed on the screen.

[0090] In addition, by displaying on the screen earthquake information including the magnitude and scale of the earthquake that occurred, analysis information on the earthquake's impact on the area of ​​interest, and decision-making information to support decision-making in the area of ​​interest, it is possible to guide service managers or service users on measures to take in response to natural disasters.

[0091] Figure 6 is an example diagram showing the fourth screen that operates the natural disaster integrated monitoring device of Figure 1.

[0092] Even if it is not a major natural disaster, by additionally monitoring natural ecosystem and environmental trends occurring around the area of ​​interest, it is possible to develop a mechanism that can comprehensively collect natural disaster analysis request information input to a diagnosis and prediction module based on a super-large language model and utilize it to derive previously unforeseen natural disaster analysis results.

[0093] The jellyfish information in Fig. 6 includes the frequency of jellyfish appearance on the coast near the area of ​​interest, and may further include information that can be utilized to determine the correlation between the predicted jellyfish appearance probability and natural disasters.

[0094] Figure 7 is an example diagram showing the fifth screen that operates the natural disaster integrated monitoring device of Figure 1.

[0095] As illustrated in FIG. 7, when a specific natural disaster such as an earthquake occurs, the integrated natural disaster monitoring device (100) inputs related natural disaster analysis request information into a diagnosis and prediction module based on a super-large language model, thereby reflecting analysis of past similar cases and analysis of the degree of correlation with the current natural disaster, and can generate and provide prediction and diagnosis results affecting an area of ​​interest and decision-making information required for the area of ​​interest.

[0096] The natural disaster integrated monitoring device (100) can input natural disaster analysis request information and facility information related to the area of ​​interest into the diagnosis prediction module after undergoing encryption processing or equivalent security processing based on standards that meet the external security requirements of the area of ​​interest.

[0097] In addition, the natural disaster integrated monitoring device (100) can generate decision-making information for an area of ​​interest, and if there is one or more sub-areas of interest within the influence area of ​​the predicted diagnosis result among the surrounding areas of the area of ​​interest, it is also possible to additionally generate and output natural disaster response information corresponding to the sub-areas of interest.

[0098] Figure 8 is an example diagram showing the sixth screen that operates the natural disaster integrated monitoring device of Figure 1.

[0099] As illustrated in FIG. 8, the integrated natural disaster monitoring device (100) can change the settings for management or use by natural disaster, adjust the alarm occurrence conditions and cycle, and check the history of natural disasters that have occurred in the past or the history of natural disasters that have affected each area of ​​interest and the response history for each area of ​​interest.

[0100] Figure 9 is a flowchart showing a natural disaster integrated monitoring method according to one embodiment of the present invention.

[0101] As illustrated in Fig. 9, the integrated natural disaster monitoring method based on a super-large language model proceeds by collecting basic data corresponding to two or more natural disasters in a natural disaster integrated monitoring device (100) (S100).

[0102] Based on the natural disaster data pool that integrates the basic data collected in step S100, first natural disaster surveillance information is generated for each natural disaster, or second natural disaster surveillance information is generated by integrating the first natural disaster surveillance information that has been generated (S102).

[0103] And, by inputting natural disaster analysis request information including the first natural disaster surveillance information and the second natural disaster surveillance information into a diagnosis prediction module based on a super-large language model, the prediction diagnosis results affecting the area of ​​interest and the decision-making information required for the area of ​​interest are generated and output (S104).

[0104] The integrated natural disaster monitoring method based on the ultra-large language model of the present invention is not limited to the processes mentioned above, and should be interpreted as being expanded upon by the invention contained in FIGS. 1 to 8 and their detailed descriptions. Accordingly, the detailed steps of the integrated natural disaster monitoring method based on the ultra-large language model that are not reflected in the flowchart of FIG. 9 can be interpreted as being expanded upon by the invention contained in FIGS. 1 to 8 and their detailed descriptions.

[0105] Fig. 10 is a flowchart showing an example of an alarm generation algorithm applied to the integrated natural disaster monitoring method of Fig. 9.

[0106] As illustrated in Fig. 10, when new data is input through the basic data collection unit (110) (S200), the distance between each process plant and the location of the forest fire can be calculated (S202).

[0107] Afterwards, it is determined whether the calculation result of step S202 exceeds the preset alarm setting distance (S204), and if the determination result of S204 exceeds the distance, the distance between the transmission line and transmission tower of the nuclear power plant and the location of the forest fire is additionally calculated (S206).

[0108] If the result calculated in step S206 is less than the preset alarm setting distance (S208), a forest fire alarm can be output (S210).

[0109] However, if the calculation result of step S204 is less than the preset alarm setting distance, step S208 may be additionally performed to determine whether to output a forest fire alarm.

[0110] Figure 11 is a flowchart showing another example of an alarm generation algorithm applied to the integrated natural disaster monitoring method of Figure 9.

[0111] As illustrated in Fig. 11, when new data is input through the basic data collection unit (110) (S300), the distance between each process plant and the earthquake location can be calculated (S302).

[0112] Thereafter, it is determined whether the calculation result of step S302 is less than a preset alarm setting distance (S304), and if the determination result of S304 is less than the preset alarm setting distance, a forest fire alarm can be output (S306).

[0113] However, if the calculation result of step S304 is greater than the preset alarm setting distance, the process may be terminated without going through an additional algorithm to determine whether to output a forest fire alarm.

[0114] Figure 12 is a flowchart showing another example of an alarm generation algorithm applied to the integrated natural disaster monitoring method of Figure 9.

[0115] As illustrated in Figure 12, when new data is input through the basic data collection unit (110) (S400), the distance between each process plant and the typhoon location can be calculated (S402).

[0116] Thereafter, it is determined whether the calculation result of step S402 is less than a preset warning setting distance (S404), and if the determination result of S404 is less than the preset warning setting distance, a typhoon entry warning can be output (S406).

[0117] However, if the calculation result of step S404 is greater than the preset warning setting distance, the process can be terminated without going through an additional algorithm to determine whether to output a typhoon approach warning.

[0118] Figure 13 is a flowchart showing another example of an alarm generation algorithm applied to the integrated natural disaster monitoring method of Figure 9.

[0119] As illustrated in Fig. 13, new data is input through the basic data collection unit (110) (S500), and determination of the expected number of typhoons related to the input new data is initiated (S502 and S504).

[0120] Thereafter, in step S504, if there is a predicted typhoon related to the location of the nuclear power plant within the area of ​​interest, the distance between the process plant location and the predicted typhoon location can be calculated (S506).

[0121] At step S506, if there are multiple predicted typhoons, a process for specifying one typhoon may be added.

[0122] Afterwards, it can be determined whether the distance obtained by subtracting the typhoon radius from the distance between the process plant location and the expected typhoon location is less than the preset alarm setting distance (S508).

[0123] If the determination result at step S508 is that the distance between the process plant location and the expected typhoon location minus the typhoon radius is less than the preset warning setting distance, an approaching warning for the typhoon can be output (S510).

[0124] Figure 14 is a flowchart showing another example of an alarm generation algorithm applied to the integrated natural disaster monitoring method of Figure 9.

[0125] As illustrated in Fig. 14, new data is input through the basic data collection unit (110) (S600), and determination of the number of weather warnings related to the input new data is initiated (S602 and S604).

[0126] Thereafter, in step S604, if there is no weather warning related to the process plant location within the area of ​​interest, it is possible to additionally determine whether there is a weather warning for the weather warning issued area including the process plant location (S606).

[0127] In step S606, if there is a weather warning for a weather warning issued area including the process plant location, the weather warning can be added to the list for subsequent algorithm operation (S608).

[0128] In step S604, if there is a weather warning related to the location of the process plant within the area of ​​interest, a process of additionally determining whether there are multiple weather warnings can be performed (S610).

[0129] In step S610, if there are multiple weather warnings, a process for specifying one weather warning may be added.

[0130] Afterwards, an alert for the corresponding weather report can be output (S612).

[0131] Although the embodiments of the present invention have been described with reference to the above and the attached drawings, those skilled in the art will understand that the present invention can be implemented in other specific forms without altering the technical concept or essential characteristics thereof. Therefore, the embodiments described above should be understood to be illustrative in all respects and not restrictive.

[0132] In addition, the present invention provides a natural disaster integrated monitoring device and method based on a large-scale language model that enables specialized safety measures to be taken for each area of ​​interest (e.g., the area around a nuclear power plant) when a natural disaster occurs, and thus, the invention has sufficient potential for commercialization or operation and is clearly and realistically feasible, and thus has industrial applicability.

Claims

1. 2 Basic data collection department that collects basic data corresponding to natural disasters of grade 1 or higher; A natural disaster surveillance information generation unit that generates first natural disaster surveillance information for each natural disaster based on a natural disaster data pool that integrates the collected basic data, or generates second natural disaster surveillance information that integrates the generated first natural disaster surveillance information; and A natural disaster integrated monitoring device based on a super-large language model, which inputs natural disaster analysis request information including the first natural disaster monitoring information and the second natural disaster monitoring information into a diagnosis and prediction module based on a super-large language model, thereby generating and outputting prediction and diagnosis results affecting an area of ​​interest and decision-making information required for the area of ​​interest.

2. In paragraph 1, The above basic data collection unit is a natural disaster integrated monitoring device based on a large-scale language model including a data collection path that receives and collects the basic data from an API (Application Programming Interface) corresponding to each of the above two or more natural disasters.

3. In the above paragraph 1, The above natural disaster monitoring information generation unit is a natural disaster integrated monitoring device based on a super-large language model that converts the first natural disaster monitoring information and the second natural disaster monitoring information based on the above area of ​​interest.

4. In paragraph 3, The above second natural disaster monitoring information is a natural disaster integrated monitoring device based on a large-scale language model that includes additional monitoring parameter information additionally derived based on the interaction between the above first natural disaster monitoring information for each natural disaster based on the above area of ​​interest.

5. In paragraph 1, The above natural disaster monitoring result provision unit is a natural disaster integrated monitoring device based on a super-large language model that sets a reference radius according to the characteristics of the equipment located in the above area of ​​interest, and adjusts the monitoring level of the above area of ​​interest by expanding or reducing the reference radius of the above area of ​​interest based on the above prediction diagnosis result.

6. In paragraph 1, The above decision-making information is a natural disaster integrated monitoring device based on a large-scale language model that includes one or more pieces of action information that should be taken for the area of ​​interest based on the above prediction diagnosis results, or one or more pieces of preemptive action information recommended for the area of ​​interest based on the simulation results predicted within a predetermined future period.

7. In paragraph 1, A natural disaster integrated monitoring device based on a super-large language model that inputs the above natural disaster analysis request information and facility information related to the above area of ​​interest into the above diagnosis and prediction module after undergoing encryption processing or equivalent security processing based on standards that meet the external security requirements of the above area of ​​interest.

8. In paragraph 1, The above natural disaster monitoring result provision unit generates the decision-making information for the above area of ​​interest, and if there is one or more sub-areas of interest within the influence area of ​​the above prediction diagnosis result among the surrounding areas of the above area of ​​interest, an integrated natural disaster monitoring device based on a super-large language model that additionally generates and outputs natural disaster response information corresponding to the sub-areas of interest.

9. In the integrated natural disaster monitoring device, a data collection step for collecting basic data corresponding to two or more natural disasters; A surveillance information generation step for generating first natural disaster surveillance information for each natural disaster based on a natural disaster data pool that integrates the collected basic data, or generating second natural disaster surveillance information that integrates the generated first natural disaster surveillance information; A natural disaster integrated monitoring method based on a large-scale language model, comprising a monitoring result provision step for generating and outputting prediction diagnosis results affecting an area of ​​interest and decision-making information required for the area of ​​interest by inputting natural disaster analysis request information including the first natural disaster monitoring information and the second natural disaster monitoring information into a diagnosis prediction module based on a large-scale language model.

10. In paragraph 9, The above data collection step is a method for integrated natural disaster monitoring based on a large-scale language model, which includes a step of collecting the basic data provided from an API (Application Programming Interface) corresponding to each of the two or more natural disasters.

11. In paragraph 9, A natural disaster integrated monitoring method based on a super-large language model, wherein the above monitoring information generation step includes a step of converting the first natural disaster monitoring information and the second natural disaster monitoring information based on the area of ​​interest.

12. In paragraph 11, A natural disaster integrated monitoring method based on a large-scale language model, wherein the second natural disaster monitoring information includes additional monitoring parameter information derived from the interaction between the first natural disaster monitoring information for each natural disaster based on the area of ​​interest.

13. In paragraph 9, The above monitoring result provision step is a natural disaster integrated monitoring method based on a super-large language model, which includes a step of setting a reference radius according to the characteristics of the equipment located in the area of ​​interest, and adjusting the monitoring level of the area of ​​interest by expanding or reducing the reference radius of the area of ​​interest based on the prediction diagnosis result.

14. In paragraph 9, A natural disaster integrated monitoring method based on a large-scale language model, wherein the above decision-making information includes one or more pieces of action information to be taken for the area of ​​interest based on the above prediction diagnosis results, or one or more pieces of preemptive action information recommended for the area of ​​interest based on the simulation results predicted within a predetermined future period.

15. In paragraph 9, The above-mentioned monitoring result provision step is a natural disaster integrated monitoring method based on a super-large language model, which includes a step of inputting the natural disaster analysis request information and the facility information related to the area of ​​interest into the diagnosis prediction module after undergoing encryption processing or equivalent security processing based on standards that meet the external security requirements of the area of ​​interest.

16. In paragraph 9, A natural disaster integrated monitoring method based on a super-large language model, wherein the above monitoring result provision step generates the decision-making information for the area of ​​interest, and, if there is one or more sub-areas of interest within the influence area of ​​the predicted diagnosis result among the surrounding areas of the area of ​​interest, additionally generates and outputs natural disaster response information corresponding to the sub-areas of interest.

Images