Artificial Intelligence-Based Building Automation Control System and Method Interworking Disaster Forecast Data and On-Site Detection Data

Abstract

The present invention relates to an artificial intelligence-based building automatic control system and method linked with disaster forecasting and on-site detection data. The system according to the present invention comprises: an external data interface unit that collects wide-area disaster forecasting data by disaster type from at least one of a national or regional meteorological agency, a regional weather station, and a disaster information providing server; a DDC data linkage unit that collects on-site disaster detection data and disaster prevention facility status data through a direct digital controller installed in a building and transmits control outputs to the disaster prevention facility; a data alignment unit that generates a comparison dataset by disaster type by aligning wide-area disaster forecasting data and on-site disaster detection data on a common time axis; a disaster cross-judgment unit that determines whether forecast risk indicators and on-site risk indicators satisfy judgment conditions by disaster type based on the comparison dataset; and a control execution unit that determines the control mode of the disaster prevention facility according to the judgment result of the disaster cross-judgment unit. The disaster cross-judgment unit determines a local disaster state, a forecast dominance discrepancy state, or a local occurrence discrepancy state based on the combination of judgment results of forecast risk indicators and on-site risk indicators. The control execution unit selects one of a standby mode, an alarm mode, a pre-operation mode, and a main operation mode of the disaster prevention facility according to the determined state. Accordingly, disaster prevention facilities can be controlled by reflecting wide-area disaster forecast data and actual detection data from the building site together.

Description

Technology Field

[0001] The present invention relates to an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system and method.

[0002] More specifically, the present invention relates to an intelligent building automatic control system and method that aligns wide-area disaster forecast data collected from a meteorological agency or a disaster information provision server with on-site disaster detection data collected through a direct digital controller (DDC) installed in a building on a common time axis, determines a local disaster state, a forecast dominance discrepancy state, or a local occurrence discrepancy state of the building based on the aligned data, and determines a control mode of a disaster prevention facility according to the determination result.

[0003] In addition, the present invention relates to a technology for controlling at least one disaster prevention facility among a drainage pump, a water barrier, an air conditioner, an air conditioner damper, a supply fan, an exhaust fan, a heating cable, a snow removal facility, a gas valve, a district heating valve, a fire shutter, and an emergency shut-off relay, by comparing wide-area disaster forecast data with detection data at a building site for at least one type of disaster among precipitation, flooding, heatwave, heavy snow, freezing, air quality deterioration, and vibration. Background Technology

[0004] Recently, buildings have been becoming larger, more complex, and more intelligent. Accordingly, HVAC systems, water supply and drainage systems, power systems, fire protection systems, ventilation systems, and various sensors are being installed in combination within buildings. Furthermore, building management systems, building automation control systems, and direct digital controllers are widely used to manage these systems and sensors in an integrated manner.

[0005] Meanwhile, due to climate change, catastrophic environmental changes such as localized heavy rainfall, heatwaves, heavy snowfall, freezing, yellow dust, fine dust, and earthquakes are occurring frequently. These catastrophic environmental changes can directly affect the drainage capacity of buildings, indoor thermal environments, HVAC operating conditions, freezing risks, structural loads, air quality, and the operating status of disaster prevention facilities.

[0006] Accordingly, automatic control technology for buildings needs to evolve beyond simply managing indoor temperature or energy consumption to a direction that responds to disaster situations by utilizing both external disaster forecast information and actual on-site detection data.

[0007] Conventionally, a technology has been proposed to manage the safety status of construction sites by combining BIM technology with IoT sensor data. This technology can predict deterioration or abnormal conditions of structures under construction by utilizing actual rainfall, structural response data, and ground or slope-related data. However, this technology primarily focuses on the safety management of slopes, foundation excavation areas, or supporting structures during construction. Consequently, there are limitations in determining the control modes of disaster prevention equipment, such as drainage pumps, water cutoff membranes, HVAC dampers, heating cables, valves, and emergency shut-off equipment, in conjunction with direct digital controllers of buildings in operation.

[0008] Furthermore, conventional building energy management technologies have been proposed that utilize BIM data to predict building energy demand and derive equipment operation scenarios using meteorological agency data, target energy data, energy unit price data, and equipment sensor data. Such technologies are useful for improving the energy efficiency of buildings and enhancing the economic feasibility of equipment operation. However, these technologies are primarily based on energy consumption, demand, carbon dioxide emissions, or the economic feasibility of equipment operation. Consequently, there are limitations in cross-referencing wide-area disaster forecast data with on-site disaster detection data regarding disaster types such as heavy rain, heatwaves, heavy snowfall, freezing, and air quality deterioration, and in determining the standby, alarm, backup, or main operation modes of disaster prevention equipment based on the results.

[0009] Furthermore, conventional BMS-based predictive control technology can predict time-series values ​​of building variables and determine control actions for building equipment based on the predicted load or energy consumption. This technology is suitable for predicting cooling loads, heating loads, power loads, or energy consumption using weather data and historical load data. However, this technology primarily focuses on demand response, load forecasting, or energy optimization. Consequently, there are limitations in aligning disaster forecast data from meteorological agencies with on-site detection data onto a common time axis and determining the local disaster status of a building based on the agreement or disagreement between forecast risk indicators and on-site risk indicators.

[0010] Furthermore, conventional technology has been proposed in which sensors installed in a building's water supply lines detect water flow and, if the flow persists for a certain period, determine that there is a possibility of leakage or flooding, transmit an alarm to a local or central monitoring system. This technology is useful for detecting water pipe leaks or water flow anomalies in specific buildings. However, this technology is not capable of determining the localized flood risk at the actual building location by comparing external heavy rain forecasts, on-site rainfall, drainage tank water levels, drainage pump operating status, water cutoff membrane status, and flood warnings. Additionally, since it may be limited to providing response information only after a simple leakage alarm, it has limitations in actively determining the control mode of disaster prevention facilities for each type of disaster.

[0011] As such, conventional technologies provide useful functions in the fields of structural safety management at construction sites, building energy management, BMS-based load prediction, or leak alarms, respectively. However, conventional technologies have limitations in integrated building automation control, which aligns wide-area disaster forecast data with detection data from the building site on the same time axis, cross-evaluates forecast risk indicators with on-site risk indicators to classify disaster states into local disaster states, forecast dominance discrepancies, or local occurrence discrepancies, and determines the control mode of disaster prevention equipment based on the classification result.

[0012] In particular, even when the same weather forecast is provided, the degree of disaster risk can vary depending on the actual location of the building, drainage conditions, equipment status, indoor environment, road surface conditions, air quality, and sensor detection values. Therefore, controlling disaster prevention facilities based solely on wide-area forecast data may result in unnecessary operation, whereas controlling them based only on on-site detection data may make it difficult to proactively account for external disaster situations.

[0013] Therefore, intelligent building automation technology is required that utilizes wide-area disaster forecast data and on-site disaster detection data together, aligns these data on a common time axis, subdivides disaster states based on the relationship between forecast risk indicators and on-site risk indicators, and determines the control mode of disaster prevention facilities based on the results. Prior art literature

[0014] (Prior Art 001) Chinese Published Patent Application CN 122114653 A (Published May 29, 2026) (Prior Art 002) Korean Published Patent Application KR 10-2023-0013771 A (Published January 27, 2023) (Prior Art 003) U.S. Registered Patent Application US 11,803,174 B2 (Registered October 31, 2023) (Prior Art 004) U.S. Registered Patent Application US 7,057,507 B1 (Registered June 6, 2006) The problem to be solved

[0015] One objective of the present invention is to provide an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system and method capable of determining the actual disaster state of a building by utilizing wide-area disaster forecasting data and on-site disaster detection data together.

[0016] Another objective of the present invention is to provide an AI-based disaster forecasting and on-site detection data-linked building automatic control system and method capable of aligning wide-area disaster forecasting data collected from a meteorological agency or a disaster information provision server and on-site disaster detection data collected through a direct digital controller installed in a building onto a common time axis and generating a comparative dataset by disaster type.

[0017] Another objective of the present invention is to provide an AI-based disaster forecasting and field detection data-linked building automatic control system and method capable of classifying the state of a building into a local disaster state, a forecast dominance discrepancy state, or a local occurrence discrepancy state by comparing the forecast risk indicator of wide-area disaster forecasting data with the field risk indicator of field disaster detection data.

[0018] Another objective of the present invention is to provide an AI-based disaster forecasting and on-site detection data-linked building automatic control system and method that can complement the limitations of methods that collectively control disaster prevention facilities based solely on wide-area disaster forecasting data or control disaster prevention facilities retrospectively based solely on on-site detection data.

[0019] Another objective of the present invention is to provide an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system and method capable of selecting a standby mode, alarm mode, backup operation mode, or main operation mode of a disaster prevention facility according to any one of a local disaster state, a forecast dominance discrepancy state, or a local occurrence discrepancy state.

[0020] Another objective of the present invention is to provide an artificial intelligence-based disaster forecasting and on-site detection data linked building automatic control system and method that can compare forecast data and on-site detection data for a plurality of disaster types, such as heavy rain or flooding, heatwave, heavy snow or freezing, and deterioration of air quality, and determine at least one control mode among drainage pumps, water barriers, air conditioners, air conditioning dampers, supply fans, exhaust fans, heating wires, snow removal equipment, and air purification equipment according to the comparison result.

[0021] Another objective of the present invention is to provide an AI-based disaster forecasting and on-site detection data-linked building automatic control system and method that stores disaster status determination results, disaster prevention facility status data, and control mode information in a history management database, and displays wide-area disaster forecasting data and on-site disaster detection data as a graph on the same time axis. means of solving the problem

[0022] An AI-based disaster forecasting and on-site detection data-linked building automatic control system according to the present invention for solving the above problem comprises: an external data interface unit that collects wide-area disaster forecasting data by disaster type from at least one of a national or regional meteorological agency, a regional meteorological station, and a disaster information providing server; a DDC data linkage unit that collects on-site disaster detection data and disaster prevention facility status data of the building through a direct digital controller installed in the building and transmits control output to the disaster prevention facility; a data alignment unit that generates a comparison dataset by disaster type by aligning the wide-area disaster forecasting data and the on-site disaster detection data on a common time axis; a disaster cross-judgment unit that determines whether the forecast risk indicator of the wide-area disaster forecasting data and the on-site risk indicator of the on-site disaster detection data satisfy the judgment conditions by disaster type based on the comparison dataset; and a control execution unit that determines the control mode of the disaster prevention facility according to the judgment result of the disaster cross-judgment unit. The disaster cross-judgment unit determines a local disaster state when both the forecast risk indicator and the on-site risk indicator satisfy the judgment conditions by disaster type, and when the forecast risk indicator satisfies the judgment conditions by disaster type but the on-site risk indicator is the disaster If the judgment conditions by type are not satisfied, it is determined to be a forecast dominance mismatch state; if the forecast risk indicator does not satisfy the judgment conditions by disaster type but the field risk indicator satisfies the judgment conditions by disaster type, it is determined to be a local occurrence mismatch state; and the control execution unit selects one of the standby mode, alarm mode, pre-operation mode, and main operation mode of the disaster prevention facility according to any one of the local disaster state, the forecast dominance mismatch state, and the local occurrence mismatch state. Effects of the invention

[0023] According to the present invention, the disaster state of a building can be determined by utilizing wide-area disaster forecast data and disaster detection data from the building site together. Accordingly, control decisions that correspond to the actual disaster situation of the building are possible, compared to a method that relies solely on external forecast information or on-site detection information.

[0024] Furthermore, according to the present invention, wide-area disaster forecast data collected from meteorological agencies or disaster information provision servers and on-site disaster detection data collected through direct digital controllers installed in buildings can be aligned on a common time axis to generate a comparison dataset by disaster type. Accordingly, forecast data and on-site data having different sources and collection cycles can be compared based on the same judgment criteria.

[0025] In addition, according to the present invention, by comparing the forecast risk indicator of wide-area disaster forecast data with the field risk indicator of field disaster detection data, the state of a building can be classified into a local disaster state, a forecast dominance discrepancy state, or a local occurrence discrepancy state. Accordingly, it is possible to distinguish between cases where both the forecast and the site are dangerous, cases where only the forecast is dangerous, and cases where only the site is dangerous, and to perform control decisions appropriate to each situation.

[0026] Furthermore, according to the present invention, problems such as collectively operating disaster prevention facilities based solely on wide-area disaster forecast data or operating disaster prevention facilities retrospectively based solely on on-site detection data can be reduced. Accordingly, in the event of a forecast dominance discrepancy, unnecessary main operation can be suppressed while performing alarms or preliminary operation, and in the event of a local occurrence discrepancy, on-site disaster situations not indicated in the forecast can be responded to.

[0027] In addition, according to the present invention, a standby mode, alarm mode, pre-operation mode, or main operation mode of a disaster prevention facility can be selected depending on any one of a local disaster state, a forecast dominance discrepancy state, or a local occurrence discrepancy state. Accordingly, the disaster prevention facility can be controlled in stages according to the degree and type of the disaster state.

[0028] In addition, according to the present invention, for a plurality of disaster types such as heavy rain or flooding, heatwaves, heavy snow or freezing, and deterioration of air quality, forecast data and on-site detection data for each disaster type can be compared, and a control mode of the corresponding disaster prevention equipment can be determined based on the comparison result. Accordingly, at least one of a drainage pump, a water barrier, an air conditioner, an air conditioning damper, a supply fan, an exhaust fan, a heating cable, a snow removal facility, and an air purification facility can be controlled to suit the disaster type.

[0029] In addition, according to the present invention, disaster status determination results, disaster prevention facility status data, and control mode information are stored in a history management database, and wide-area disaster forecast data and on-site disaster detection data can be displayed as a graph on the same time axis. Accordingly, the manager can intuitively check the difference between forecast data and on-site data, disaster status determination results, and the control history of the disaster prevention facility. Brief explanation of the drawing

[0030] FIG. 1 is an overall configuration diagram of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to one embodiment of the present invention. FIG. 2 is a detailed block diagram of a direct digital controller according to one embodiment of the present invention. FIG. 3 is a detailed block diagram of a data matching unit and a disaster cross-determination unit according to an embodiment of the present invention. FIG. 4 is a disaster state classification flowchart according to one embodiment of the present invention. FIG. 5 is a flowchart of a control mode determination unit according to an embodiment of the present invention. FIG. 6 is a flowchart for responding to heavy rain or flood disasters according to one embodiment of the present invention. FIG. 7 is a detailed block diagram of a control execution unit according to one embodiment of the present invention. FIG. 8 is an example screen diagram of a monitoring interface according to one embodiment of the present invention. FIG. 9 is an overall flowchart of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control method according to one embodiment of the present invention. FIG. 10 is a flowchart of field safety margin-based operating threshold correction according to one embodiment of the present invention. FIG. 11 is a flowchart for identifying vulnerable buildings based on repeated alarms according to one embodiment of the present invention. Specific details for implementing the invention

[0031] It should be noted that in the following description, only the parts necessary for understanding the embodiments of the present invention are explained, and the description of other parts will be omitted to the extent that it does not deviate from the gist of the present invention.

[0032] The terms and words used in the specification and claims described below should not be interpreted as being limited to their ordinary or dictionary meanings, but should be interpreted in a meaning and concept consistent with the technical spirit of the invention, based on the principle that the inventor may appropriately define the concept of the terms to best describe his invention. Accordingly, the embodiments described in this specification and the configurations illustrated in the drawings are merely preferred embodiments of the invention and do not represent all aspects of the technical spirit of the invention; therefore, it should be understood that various equivalents and modifications capable of replacing them may exist at the time of filing this application.

[0033] Hereinafter, embodiments of the present invention will be described in more detail with reference to the attached drawings.

[0034] FIG. 1 is an overall configuration diagram of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to an embodiment of the present invention, FIG. 3 is a detailed block diagram of a data matching unit and a disaster cross-determination unit according to an embodiment of the present invention, FIG. 4 is a disaster state classification flowchart according to an embodiment of the present invention, and FIG. 5 is a control mode determination flowchart of a control execution unit according to an embodiment of the present invention.

[0035] According to FIGS. 1, 3 to 5, an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to an embodiment of the present invention includes an external data interface unit (110), a DDC data linkage unit (190), a data matching unit (120), a disaster cross-determination unit (150), and a control execution unit (170). The external data interface unit (110) collects wide-area disaster forecasting data by disaster type from at least one of a national or regional meteorological agency, a regional meteorological station, and a disaster information providing server. The DDC data linkage unit (190) collects on-site disaster detection data and disaster prevention facility status data of a building through a direct digital controller (210) installed in the building, and transmits control outputs to the disaster prevention facility (300). The data matching unit (120) aligns the wide-area disaster forecasting data and the on-site disaster detection data on a common time axis to generate a comparison dataset by disaster type. The disaster cross-judgment unit (150) determines whether the forecast risk indicator of the wide-area disaster forecast data and the field risk indicator of the field disaster detection data satisfy the judgment conditions for each disaster type based on the comparison dataset. The control execution unit (170) determines the control mode of the disaster prevention equipment (300) according to the judgment result of the disaster cross-judgment unit (150).

[0036] The external data interface unit (110) can be connected to communicate with a weather agency server (10) or a disaster information provision server (20). The external data interface unit (110) can collect precipitation forecasts, temperature forecasts, snowfall forecasts, freezing forecasts, air quality forecasts, or weather special report data from the weather agency server (10). Additionally, the external data interface unit (110) can collect disaster notification data related to heavy rain, flooding, heatwaves, heavy snowfall, freezing, air quality deterioration, or earthquakes from the disaster information provision server (20). The external data interface unit (110) converts the collected data into wide-area disaster forecast data by disaster type or transmits it in a data format that can be processed by the data matching unit (120).

[0037] The DDC data linkage unit (190) is connected to communicate with a direct digital controller (210) installed in a building. The direct digital controller (210) can be connected to field sensors (400) and disaster prevention equipment (300) placed within the building. The DDC data linkage unit (190) can collect field disaster detection data related to water level, leakage, rainfall, indoor temperature, outdoor temperature, humidity, air quality, vibration, load, snow accumulation, or freezing through the direct digital controller (210). Additionally, the DDC data linkage unit (190) can collect disaster prevention equipment status data related to drainage pumps, water barriers, air conditioners, air conditioner dampers, supply fans, exhaust fans, heating wires, snow removal equipment, gas valves, district heating valves, fire shutters, or emergency cutoff relays through the direct digital controller (210).

[0038] The data alignment unit (120) organizes the wide-area disaster forecast data collected through the external data interface unit (110) and the field disaster detection data collected through the DDC data linkage unit (190) into a form that can be compared with each other. To this end, the data alignment unit (120) can align data having different collection cycles, time standards, data formats, or location standards to a common time axis. For example, if the wide-area disaster forecast data is provided as hourly forecast values ​​and the field disaster detection data is provided as minute-level sensor values, the data alignment unit (120) can convert both data to correspond to the same judgment interval.

[0039] The data matching unit (120) matches the aligned wide-area disaster forecast data and the field disaster detection data according to the disaster type. For example, for the heavy rain or flood disaster type, the precipitation forecast value, the field rainfall amount, the drainage tank water level, the leakage signal, or the flood signal may be matched with each other. For the heatwave disaster type, the temperature forecast value, the outdoor temperature, the indoor temperature, or the facility temperature may be matched with each other. For the heavy snow or freezing disaster type, the snow depth forecast value, the field snow depth amount, the freezing detection value, or the load detection value may be matched with each other. For the air quality disaster type, the fine dust forecast value, the yellow dust forecast value, the field outdoor air quality, or the indoor air quality may be matched with each other. Based on these correspondence relationships, the data matching unit (120) generates a comparison dataset by disaster type.

[0040] The disaster cross-judgment unit (150) calculates or determines forecast risk indicators and field risk indicators based on a comparison dataset for each disaster type. The forecast risk indicator may refer to the degree of risk derived from wide-area disaster forecast data. The field risk indicator may refer to the degree of risk derived from field disaster detection data. The judgment conditions for each disaster type may include reference values, threshold values, alarm conditions, or complex judgment conditions corresponding to each disaster type.

[0041] The disaster cross-determination unit (150) determines the state of the building as a local disaster state when both the forecast risk indicator and the on-site risk indicator satisfy the judgment conditions for each type of disaster. In this case, since both the wide-area disaster forecast and the detection values ​​at the building site indicate a dangerous state, the disaster cross-determination unit (150) can determine that the actual disaster risk is confirmed at the location of the building.

[0042] The disaster cross-judgment unit (150) determines the condition of the building as a forecast-dominant discrepancy state when the forecast risk indicator satisfies the judgment conditions for each disaster type, but the on-site risk indicator does not satisfy the judgment conditions for each disaster type. In this case, there may be a state where a disaster risk exists in the external forecast, but the risk detection value has not yet been confirmed at the building site. Therefore, the disaster cross-judgment unit (150) can distinguish the discrepancy between the forecast and the site and transmit it to the control execution unit (170).

[0043] The disaster cross-judgment unit (150) determines the state of the building as a local occurrence discrepancy state when the forecast risk indicator does not satisfy the judgment conditions for each type of disaster, but the on-site risk indicator satisfies the judgment conditions for each type of disaster. In this case, the disaster risk may be low or not indicated in the external forecast, but the actual risk detection value may be confirmed at the building site. Therefore, the disaster cross-judgment unit (150) can distinguish the possibility of a local disaster occurring that does not appear in the forecast as a separate state.

[0044] The control execution unit (170) selects a control mode of the disaster prevention facility (300) according to the judgment result of the disaster cross-judgment unit (150). The control execution unit (170) may select one of the standby mode, alarm mode, pre-operation mode, and main operation mode of the disaster prevention facility (300) according to any one of the local disaster state, forecast dominance discrepancy state, and local occurrence discrepancy state. The standby mode may be a mode that maintains the disaster prevention facility (300) in a general standby state. The alarm mode may be a mode that provides an alarm to an administrator or control system. The pre-operation mode may be a mode that switches the disaster prevention facility (300) to a ready state before full operation. The main operation mode may be a mode that actually operates the disaster prevention facility (300) as a disaster response operation.

[0045] When a local disaster state is determined, the control execution unit (170) can switch the disaster prevention facility (300) corresponding to the disaster type to the main driving mode. Since both the forecast risk indicator and the field risk indicator indicate a dangerous state, the control execution unit (170) can generate a control output for the disaster prevention facility (300) and transmit it to the direct digital controller (210) through the DDC data linkage unit (190). The direct digital controller (210) can drive the corresponding disaster prevention facility (300) according to the received control output.

[0046] If the control execution unit (170) determines that the forecast dominance is inconsistent, it can switch the disaster prevention facility (300) to an alarm mode or a preliminary operation mode. In this case, since the field detection value does not yet satisfy the judgment conditions for each disaster type, the control execution unit (170) can control the disaster prevention facility (300) to perform an alarm provision or preliminary preparation operation without immediately switching it to the main operation mode. Accordingly, the control execution unit (170) can determine the control mode by reflecting the field conditions while performing a preemptive response based on the external forecast.

[0047] When the control execution unit (170) determines that a local occurrence mismatch state has been detected, it provides a local occurrence alarm to the manager and can switch the disaster prevention facility (300) to an alarm mode, a backup operation mode, or a main operation mode. In this case, even if the risk is not indicated in the external forecast, the field detection value indicates a dangerous state, so the control execution unit (170) can determine the need to control the disaster prevention facility (300) based on the field risk indicator.

[0048] The control execution unit (170) transmits a control output corresponding to the selected control mode to the DDC data linkage unit (190). The DDC data linkage unit (190) can convert the control output received from the control execution unit (170) into a control signal or control command that can be processed by the direct digital controller (210). The direct digital controller (210) can change the operating state of the fire protection equipment (300) according to the converted control signal or control command.

[0049] In this way, the artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to the present embodiment aligns wide-area disaster forecasting data and on-site disaster detection data on a common time axis, subdivides the disaster state according to the relationship between the forecasting risk indicator and the on-site risk indicator, and determines the control mode of the disaster prevention equipment (300) according to the subdivided disaster state.

[0050] FIG. 1 is an overall configuration diagram of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to one embodiment of the present invention, and FIG. 3 is a detailed block diagram of a data matching unit and a disaster cross-determination unit according to one embodiment of the present invention.

[0051] According to FIGS. 1 and 3, the external data interface unit (110) according to the present embodiment converts wide-area disaster forecast data collected from at least one of a national or regional meteorological agency, a regional weather station, and a disaster information providing server into a data format including at least one of a disaster type code, forecast time, forecast target location, forecast value, and warning grade, and provides it to the data matching unit (120).

[0052] The external data interface unit (110) is connected to communicate with a weather agency server (10) or a disaster information provision server (20). The weather agency server (10) may include a national weather agency server, a regional weather agency server, or a regional weather station server. The disaster information provision server (20) may include a public disaster information server, a local government disaster information server, or a disaster notification provision server. The external data interface unit (110) collects disaster forecast information corresponding to the location of a building or a managed area from these external servers.

[0053] Wide-area disaster forecast data may include data regarding at least one type of disaster among heavy rain, flooding, heatwave, heavy snowfall, freezing, deterioration of air quality, strong winds, earthquakes, or disaster alerts. For example, wide-area disaster forecast data may include precipitation forecast values, precipitation probability, heavy rain advisories, temperature forecast values, heatwave advisories, snowfall forecast values, freezing forecast values, fine dust forecast values, yellow dust forecast values, wind speed forecast values, or earthquake-related alert information.

[0054] The external data interface unit (110) classifies the collected wide-area disaster forecast data by disaster type. For example, the external data interface unit (110) may correspond data regarding precipitation or heavy rain warnings to heavy rain or flood disaster types, correspond data regarding temperature forecast values ​​or heat wave warnings to heat wave disaster types, and correspond data regarding snowfall forecast values ​​or freezing forecast values ​​to heavy snowfall or freezing disaster types. Additionally, the external data interface unit (110) may correspond data regarding fine dust forecast values ​​or yellow dust forecast values ​​to air quality disaster types.

[0055] The external data interface unit (110) can assign a disaster type code to the wide-area disaster forecast data. The disaster type code can be used as identification information for the data matching unit (120) to match the wide-area disaster forecast data with the field disaster detection data by disaster type. For example, heavy rain or flood disaster types, heatwave disaster types, heavy snow or freezing disaster types, and air quality disaster types can be distinguished by different disaster type codes.

[0056] The external data interface unit (110) can assign forecast time information to wide-area disaster forecast data or extract forecast time information received from an external server. The forecast time information may include the forecast announcement time, the forecast application start time, the forecast application end time, or the forecast target time interval. The data alignment unit (120) can use this forecast time information to align the wide-area disaster forecast data with the field disaster detection data on a common time axis.

[0057] The external data interface unit (110) can assign forecast target location information to wide-area disaster forecast data or extract forecast target location information received from an external server. The forecast target location information may include administrative district information, grid coordinate information, latitude and longitude information, observation station identification information, or regional code information. The data matching unit (120) can select or match wide-area disaster forecast data corresponding to the location of a building using the forecast target location information.

[0058] The external data interface unit (110) extracts forecast values ​​from wide-area disaster forecast data. Depending on the type of disaster, the forecast values ​​may include at least one of precipitation amount, probability of precipitation, temperature, wind chill temperature, snowfall amount, possibility of freezing, fine dust concentration, ultrafine dust concentration, yellow dust concentration, wind speed, intensity, or disaster risk index. The external data interface unit (110) can convert the forecast values ​​into a unit or format that the data matching unit (120) can process.

[0059] The external data interface unit (110) can extract the warning level of the wide-area disaster forecast data. The warning level may include stage information such as advisory, warning, special report, interest, caution, alert, or severe. The warning level can be used by the disaster cross-judgment unit (150) when calculating forecast risk indicators or determining whether judgment conditions by disaster type are satisfied.

[0060] The external data interface unit (110) converts wide-area disaster forecast data, which has different data formats for each external server, into an internal processing format. For example, the external data interface unit (110) can convert tabular data, API response data, text notification data, or message format data into structured data including a disaster type code, forecast time, forecast target location, forecast value, and alarm level. The converted data is provided to the data matching unit (120).

[0061] The external data receiving unit (121) of the data matching unit (120) receives the converted wide-area disaster forecast data from the external data interface unit (110). The external data receiving unit (121) can provide the received wide-area disaster forecast data to at least one of the time alignment unit (125), the location matching unit (127), the sensor type matching unit (129), the disaster type matching unit (131), or the comparison dataset generation unit (133).

[0062] In this way, the external data interface unit (110) according to the present embodiment does not merely receive wide-area disaster forecast data provided from an external server, but can convert it into a data format including a disaster type code, forecast time, forecast target location, forecast value, and alarm grade, and provide it to the data matching unit (120). Accordingly, the data matching unit (120) can process the wide-area disaster forecast data into a form that can be compared with on-site disaster detection data by disaster type.

[0063] FIG. 1 is an overall configuration diagram of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to one embodiment of the present invention, FIG. 2 is a detailed block diagram of a direct digital controller according to one embodiment of the present invention, and FIG. 9 is an overall flowchart of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control method according to one embodiment of the present invention.

[0064] According to FIGS. 1, 2 and 9, the DDC data linkage unit (190) according to the present embodiment collects on-site disaster detection data and disaster prevention equipment status data through a direct digital controller (210) installed in a building, and generates or converts a control output transmitted to the direct digital controller (210). On-site disaster detection data includes at least one of water level data, leakage data, rainfall data, indoor temperature data, outdoor temperature data, humidity data, air quality data, vibration data, load data, snow accumulation data, and freezing data. Disaster prevention equipment status data includes at least one of drainage pump status data, water barrier status data, air conditioner status data, air conditioner damper status data, heating cable status data, snow removal equipment status data, and valve status data. The DDC data linkage unit (190) converts sensor data and equipment status data collected from the direct digital controller (210) into a packet format that the data matching unit (120) can process, and converts the control mode determined by the control execution unit (170) into output control data that the direct digital controller (210) can process.

[0065] The direct digital controller (210) is connected to the field sensor (400) and the fire protection equipment (300) installed in the building. The direct digital controller (210) may be installed in at least one of the building's machine room, electrical room, basement, rooftop, outside air intake, water supply and drainage equipment area, air conditioning equipment area, or fire protection equipment area. The direct digital controller (210) receives a detection signal from the field sensor (400), receives the operating status or abnormal status of the fire protection equipment (300), and transmits a control output to the fire protection equipment (300).

[0066] The direct digital controller (210) may include a sensor input unit (220), an equipment status input unit (250), and an output control unit (270). The sensor input unit (220) receives field disaster detection data from a field sensor (400). The equipment status input unit (250) receives at least one of the operating state, standby state, fault state, alarm state, or manual operation state of the disaster prevention equipment (300). The output control unit (270) controls the operation of the disaster prevention equipment (300) according to the control output transmitted through the DDC data linkage unit (190).

[0067] The sensor input unit (220) may include a water level sensor input unit (221), a leak sensor input unit (223), a rainfall sensor input unit (225), an indoor temperature sensor input unit (227), an outdoor temperature sensor input unit (229), a humidity sensor input unit (231), an air quality sensor input unit (233), a vibration sensor input unit (235), a load sensor input unit (237), a snow accumulation sensor input unit (239), and a freezing sensor input unit (241). The water level sensor input unit (221) receives water level data from a drainage tank, a sump, a basement, or a flood risk area. The leak sensor input unit (223) receives leak detection data from pipes, floors, machine rooms, or electrical rooms. The rainfall sensor input unit (225) receives field rainfall data from a rainfall sensor installed around the building or on the rooftop.

[0068] The indoor temperature sensor input unit (227) receives indoor temperature data from inside the building. The outdoor temperature sensor input unit (229) receives outdoor temperature data from outside the building. The humidity sensor input unit (231) receives humidity data from inside or outside the building. The air quality sensor input unit (233) receives air quality data regarding at least one of fine dust, ultrafine dust, carbon dioxide, volatile organic compounds, or harmful gases.

[0069] The vibration sensor input unit (235) receives vibration data related to earthquakes, structural vibrations, or equipment vibrations. The load sensor input unit (237) receives load data acting on a rooftop, awning, structural member, or snow removal target area. The snow accumulation sensor input unit (239) receives snow accumulation data on the rooftop of a building, an external passageway, or a snow removal target area. The freezing sensor input unit (241) receives freezing data on a road surface, piping, an outside air intake, or an external equipment area.

[0070] The equipment status input section (250) may include a drainage pump status input section (251), a water barrier status input section (253), an air conditioner status input section (255), an air conditioning damper status input section (257), a heating wire status input section (259), a snow removal equipment status input section (261), and a valve status input section (263). The drainage pump status input section (251) receives the operating status, stopped status, fault status, or overload status of the drainage pump. The water barrier status input section (253) receives the open status, closed status, driving status, or abnormal status of the water barrier. The air conditioner status input section (255) receives the operating status, stopped status, cooling status, heating status, or abnormal status of the air conditioner.

[0071] The air conditioning damper status input unit (257) receives the opening status of the outside air damper, exhaust damper, or circulation damper. The heating wire status input unit (259) receives the energized status, cutoff status, temperature status, or fault status of the heating wire. The snow removal equipment status input unit (261) receives the operating status, standby status, or abnormal status of the snow removal equipment. The valve status input unit (263) receives the open / closed status of the gas valve, district heating valve, or other fluid control valve.

[0072] The output control unit (270) may include a drainage pump output control unit (271), a water barrier output control unit (273), an air conditioner output control unit (275), an air conditioning damper output control unit (277), a supply fan output control unit (279), an exhaust fan output control unit (281), a heating wire output control unit (283), a snow removal equipment output control unit (285), a gas valve output control unit (287), a district heating valve output control unit (289), a fire shutter output control unit (291), and an emergency cutoff relay output control unit (293). The output control unit (270) generates or transmits an output signal for the disaster prevention equipment (300) according to a control mode selected by the control execution unit (170).

[0073] The drainage pump output control unit (271) can control the drainage pump to standby, pre-operation, or main operation state. The water barrier output control unit (273) can control the water barrier to open, closed, or partially closed state. The air conditioner output control unit (275) can control the operation mode, airflow, or heating and cooling operation of the air conditioner. The air conditioning damper output control unit (277) can control the opening of the air conditioning damper to regulate the amount of outside air inflow, exhaust volume, or circulating air volume.

[0074] The supply fan output control unit (279) and the exhaust fan output control unit (281) can control the operating state of the fans to regulate the indoor ventilation or exhaust volume. The heating wire output control unit (283) can control the energization state of the heating wire to prevent or eliminate freezing in areas at risk of freezing. The snow removal equipment output control unit (285) can control the operating state of the snow removal equipment according to the snow accumulation condition or snow accumulation forecast. The gas valve output control unit (287) and the district heating valve output control unit (289) can shut off or open the valves according to a disaster condition or an equipment malfunction condition. The fire shutter output control unit (291) and the emergency shut-off relay output control unit (293) can control the fire shutter or the emergency shut-off relay according to disaster response or emergency shut-off conditions.

[0075] The DDC data linkage unit (190) converts field disaster detection data and disaster prevention facility status data collected from the direct digital controller (210) into a data format that can be processed by the data matching unit (120) or the control execution unit (170). For example, the DDC data linkage unit (190) can generate a packet including sensor identification information, facility identification information, measurement time, measurement value, status value, alarm value, and building identification information. The generated packet can be provided to the data matching unit (120) and compared with wide-area disaster forecast data.

[0076] The DDC data linkage unit (190) may include a communication gateway for communication with the direct digital controller (210). The communication gateway can perform signal format conversion and packet routing between an upper communication protocol for communication with the upper management system, the data matching unit (120), or the control execution unit (170), and a lower serial communication protocol for communication with the lower distributed direct digital controller (210). For example, the communication gateway may use the MODBUS / TCP protocol as the upper communication protocol and the RS-485-based MODBUS-RTU protocol as the lower serial communication protocol.

[0077] The communication gateway can convert control command or data request packets received from the upper communication network into serial communication frames that the lower direct digital controller (210) can process. Additionally, the communication gateway can convert sensor measurement values, facility status values, or alarm values ​​received from the direct digital controller (210) into packet formats that the upper communication network can process. For example, the communication gateway can communicate with the direct digital controller (210) according to serial communication conditions having a communication speed of 9600 bps, 8 data bits, no parity, and 1 stop bit. However, the above communication conditions are merely examples and may be changed according to the facility configuration of the building, communication distance, communication network quality, or the specifications of the direct digital controller (210).

[0078] When multiple direct digital controllers (210) are distributed, the DDC data linkage unit (190) can route packets based on the communication address, facility identification information, or sensor identification information of each direct digital controller (210). For example, the communication gateway can collect drainage tank water level data and drainage pump status data from a direct digital controller (210) installed in the basement, collect snow accumulation data and heating wire status data from a direct digital controller (210) installed on the rooftop, and collect air conditioner status data and air conditioner damper status data from a direct digital controller (210) installed in the air conditioning room. In this case, the DDC data linkage unit (190) can packetize the collected data along with building identification information, zone identification information, sensor identification information, and measurement time and provide it to the data matching unit (120).

[0079] The sensor input section (220) may include an analog input channel or a digital input channel corresponding to a physical interface for each sensor. For example, the indoor temperature sensor input section (227) or the outdoor temperature sensor input section (229) may receive a change in the resistance value of the platinum resistance thermometer RTD Pt100. The water level sensor input section (221) may receive a DC 4-20mA current loop signal from the drainage tank level sensor. The vibration sensor input section (235) or the load sensor input section (237) may receive a DC 0-5V voltage signal. The water leakage sensor input section (223), flood-related input, or equipment malfunction input may receive a voltage-free contact signal isolated by a photocoupler.

[0080] The direct digital controller (210) can convert a physical signal input through the sensor input unit (220) into a digital measurement value, status value, or alarm value. For example, the direct digital controller (210) can convert a change in the resistance value of the RTD Pt100 into a temperature value, convert a DC 4-20mA current loop signal into a water level value, and convert a DC 0-5V voltage signal into a vibration value or load value. In addition, the direct digital controller (210) can convert the open or closed state of a voltage-free contact signal into a leak alarm, a flood alarm, or an equipment malfunction alarm.

[0081] The DDC data linkage unit (190) collects digital measurement values, status values, or alarm values ​​converted from the direct digital controller (210) and can format them into field disaster detection data or disaster prevention facility status data that can be processed by the data matching unit (120). For example, the DDC data linkage unit (190) can generate a packet including the unit, scale, sensor address, facility address, data collection time, and communication status of the measurement value. The data matching unit (120) can match the field disaster detection data with wide-area disaster forecast data on a common time axis using the measurement time and sensor identification information included in this packet.

[0082] Additionally, when the DDC data linkage unit (190) converts the control mode determined by the control execution unit (170) into output control data that can be processed by the direct digital controller (210), it can generate an output packet including the communication address, output channel, control command code, and control mode information of the equipment to be controlled. The direct digital controller (210) can operate the output control unit (270) according to the output channel and control command code included in the output packet. Accordingly, at least one of the drainage pump output control unit (271), water barrier output control unit (273), air conditioner output control unit (275), air conditioner damper output control unit (277), heating wire output control unit (283), or snow removal equipment output control unit (285) can be controlled according to the selected control mode.

[0083] The DDC data linkage unit (190) converts the control mode determined by the control execution unit (170) into output control data that can be processed by the direct digital controller (210). For example, if the control execution unit (170) selects the main driving mode of the drainage pump, the DDC data linkage unit (190) can generate output control data to be transmitted to the drainage pump output control unit (271). If the control execution unit (170) selects the alarm mode or the pre-driving mode of the air conditioning damper, the DDC data linkage unit (190) can generate output control data to be transmitted to the air conditioning damper output control unit (277).

[0084] In this way, the direct digital controller (210) and the DDC data linkage unit (190) according to the present embodiment can collect field disaster detection data from the field sensor (400), collect status data of the disaster prevention facility (300), and transmit the control mode of the control execution unit (170) to the disaster prevention facility (300). Accordingly, field data that can be compared with wide-area disaster forecast data is secured through the direct digital controller (210), and the control output of the disaster prevention facility (300) based on the disaster status determination result can be transmitted to the field facility.

[0085] FIG. 1 is an overall configuration diagram of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to an embodiment of the present invention, FIG. 3 is a detailed block diagram of a data matching unit and a disaster cross-determination unit according to an embodiment of the present invention, and FIG. 9 is an overall flowchart of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control method according to an embodiment of the present invention.

[0086] According to FIGS. 1, 3 and 9, the data matching unit (120) according to the present embodiment aligns wide-area disaster forecast data received from the external data interface unit (110) and field disaster detection data received from the DDC data linkage unit (190) on a common time axis, matches the building location with the forecast target location, and creates a comparison dataset by disaster type by matching the disaster type of the wide-area disaster forecast data with the sensor type of the field disaster detection data.

[0087] The data matching unit (120) may include an external data receiving unit (121), a field data receiving unit (123), a time alignment unit (125), a location matching unit (127), a sensor type matching unit (129), a disaster type matching unit (131), and a comparison dataset generation unit (133). The external data receiving unit (121) receives wide-area disaster forecast data from the external data interface unit (110). The field data receiving unit (123) receives field disaster detection data from the DDC data linkage unit (190). The external data receiving unit (121) and the field data receiving unit (123) transmit data collected from different sources in an internal processing format of the data matching unit (120).

[0088] The external data receiving unit (121) can receive at least one of the disaster type code, forecast time, forecast target location, forecast value, and warning level included in the wide-area disaster forecast data. If the received wide-area disaster forecast data is data formatted by the external data interface unit (110), the external data receiving unit (121) can provide it as is to the time alignment unit (125), the location matching unit (127), or the disaster type matching unit (131). Additionally, if necessary, the external data receiving unit (121) can remove or correct missing values, duplicate values, or abnormal values ​​from the received wide-area disaster forecast data.

[0089] The field data receiving unit (123) receives field disaster detection data from the DDC data linkage unit (190). The field disaster detection data may include at least one of water level data, leakage data, rainfall data, indoor temperature data, outdoor temperature data, humidity data, air quality data, vibration data, load data, snow accumulation data, and freezing data. The field data receiving unit (123) can extract sensor identification information, measurement time, measurement value, status value, alarm value, and building identification information included in each field disaster detection data.

[0090] The time alignment unit (125) aligns the wide-area disaster forecast data and the field disaster detection data on a common time axis. The wide-area disaster forecast data may include the forecast announcement time, the forecast application start time, the forecast application end time, or the forecast target time interval. The field disaster detection data may include the sensor measurement time or the DDC collection time. The time alignment unit (125) uses this time information to align the wide-area disaster forecast data and the field disaster detection data so that they are compared in the same judgment interval.

[0091] The time alignment unit (125) can perform resampling when the data collection cycles are different. For example, if wide-area disaster forecast data is provided in 1-hour intervals and field disaster detection data is collected in 1-minute intervals, the time alignment unit (125) can convert the field disaster detection data into an average value, maximum value, minimum value, cumulative value, or representative value corresponding to the forecast time interval. Conversely, if the wide-area disaster forecast data includes multiple time intervals, the time alignment unit (125) can classify the field disaster detection data corresponding to each forecast time interval.

[0092] The location matching unit (127) matches the forecast target location of the wide-area disaster forecast data with the location of the building. The forecast target location may include administrative district information, grid coordinate information, latitude and longitude information, observation station identification information, or area code information. The location of the building may include building identification information, address information, latitude and longitude information, management district information, or point code. The location matching unit (127) selects wide-area disaster forecast data corresponding to the building using this location information.

[0093] The location matching unit (127) may select the forecast grid, observation station, administrative district, or regional code closest to the building when the forecast target location does not completely match the location of the building. Additionally, when multiple forecast target locations correspond to a single building, the location matching unit (127) may select or weight-combine wide-area disaster forecast data based on at least one of the distance from the building, altitude, topographic conditions, or administrative district priority.

[0094] The sensor type matching unit (129) matches the sensor type of the field disaster detection data to the data item to be used for determining the disaster type. For example, water level data, leakage data, and rainfall data may correspond to heavy rain or flood disaster types. Indoor temperature data, outdoor temperature data, and humidity data may correspond to heatwave disaster types. Snow accumulation data, freezing data, and load data may correspond to heavy snow or freezing disaster types. Air quality data may correspond to air quality disaster types. Vibration data may correspond to disaster types related to earthquakes or structural abnormalities.

[0095] The sensor type matching unit (129) can duplicately match the corresponding sensor type to multiple comparison datasets according to the judgment conditions for each disaster type when a single sensor type can be used for multiple disaster types. For example, ambient temperature data can be used to determine not only heatwave disaster types but also freezing disaster types. Load data can be used to determine not only heavy snow disaster types but also structural abnormalities. The sensor type matching unit (129) can process these multiple correspondence relationships according to a pre-set matching table or rule.

[0096] The disaster type matching unit (131) matches the disaster type of the wide-area disaster forecast data with the sensor type of the on-site disaster detection data. The disaster type matching unit (131) may use the disaster type code assigned by the external data interface unit (110) and the sensor type information classified by the sensor type matching unit (129). The disaster type matching unit (131) matches the wide-area disaster forecast data with the on-site disaster detection data for at least one of heavy rain or flooding, heatwave, heavy snow or freezing, deterioration of air quality, strong wind, earthquake, or other disaster types.

[0097] The comparison dataset generation unit (133) generates a comparison dataset by disaster type using the processing results of the time alignment unit (125), the location matching unit (127), the sensor type matching unit (129), and the disaster type matching unit (131). The comparison dataset by disaster type may be data that combines wide-area disaster forecast data and on-site disaster detection data corresponding to one disaster type into a single comparison unit. The comparison dataset generation unit (133) may include at least one of building identification information, disaster type code, judgment time interval, forecast value, on-site detection value, alarm grade, sensor status value, or correction value in the comparison dataset.

[0098] For example, a comparison dataset for a heavy rain or flood disaster type may include at least one of precipitation forecast values, heavy rain warning information, on-site rainfall amount, drainage tank water level, leakage signals, and flood signals. A comparison dataset for a heatwave disaster type may include at least one of temperature forecast values, heatwave warning information, outdoor temperature, indoor temperature, humidity, and air conditioner status. A comparison dataset for a heavy snow or freezing disaster type may include at least one of snowfall forecast values, freezing forecast values, on-site snowfall amount, freezing detection values, load data, and heating cable status. A comparison dataset for an air quality disaster type may include at least one of fine dust forecast values, yellow dust forecast values, outdoor air quality data, indoor air quality data, and air conditioning damper status.

[0099] The comparison dataset generation unit (133) provides the generated comparison dataset by disaster type to the disaster cross-judgment unit (150). The disaster cross-judgment unit (150) can calculate or determine forecast risk indicators and field risk indicators based on at least one of the forecast value, alarm grade, field detection value, and sensor status value included in the comparison dataset. Accordingly, the data matching unit (120) performs the role of preprocessing data having different sources and formats into a form that the disaster cross-judgment unit (150) can judge based on the same criteria.

[0100] The data matching unit (120) may further perform missing data processing, outlier removal, unit conversion, or data reliability evaluation as needed. For example, if data from some field sensors has not been collected, the data matching unit (120) may supplement the comparison dataset using the recent normal value, adjacent sensor value, or facility status data of the corresponding sensor. Additionally, if the units of external forecast data and field detection data are different, the data matching unit (120) may generate a comparison dataset after converting them to the same unit.

[0101] In this way, the data matching unit (120) according to the present embodiment can align wide-area disaster forecast data and on-site disaster detection data on a common time axis, match building locations with forecast target locations, and correspond disaster types with sensor types to generate a comparison dataset by disaster type. Accordingly, the disaster cross-determination unit (150) can compare wide-area disaster forecast data and on-site disaster detection data in the same judgment interval and the same disaster type criteria.

[0102] FIG. 3 is a detailed block diagram of a data matching unit and a disaster cross-determination unit according to an embodiment of the present invention, FIG. 4 is a flowchart of a disaster state classification according to an embodiment of the present invention, and FIG. 9 is an overall flowchart of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control method according to an embodiment of the present invention.

[0103] According to FIGS. 3, 4 and 9, the disaster cross-determination unit (150) according to the present embodiment determines whether the forecast risk indicator of the wide-area disaster forecast data and the field risk indicator of the field disaster detection data satisfy the disaster type judgment conditions based on the disaster type comparison dataset generated by the data matching unit (120), and classifies the state of the building into one of a local disaster state, a forecast dominance discrepancy state, a local occurrence discrepancy state, or a normal state according to the combination of the judgment results of the forecast risk indicator and the field risk indicator.

[0104] The disaster cross-judgment unit (150) may include a forecast risk indicator calculation unit (151), a field risk indicator calculation unit (153), and a judgment state classification unit (155). The forecast risk indicator calculation unit (151) calculates a forecast risk indicator based on at least one of a forecast value, an alarm grade, a disaster type code, or a forecast application time interval included in wide-area disaster forecast data. The field risk indicator calculation unit (153) calculates a field risk indicator based on at least one of a sensor measurement value, a status value, an alarm value, a measurement time, or a sensor type included in field disaster detection data. The judgment state classification unit (155) compares the forecast risk indicator and the field risk indicator with judgment conditions for each disaster type and classifies the disaster state of the building according to the comparison result.

[0105] Forecast risk indicators may be values ​​or conditions representing the degree of disaster risk derived from wide-area disaster forecast data. For example, in the case of heavy rain or flood disaster types, forecast risk indicators may be calculated based on precipitation forecast values, precipitation probability, heavy rain advisory grades, or flood risk alert grades. In the case of heatwave disaster types, forecast risk indicators may be calculated based on temperature forecast values, wind chill forecast values, or heatwave advisory grades. In the case of heavy snowfall or freezing disaster types, forecast risk indicators may be calculated based on snowfall forecast values, freezing forecast values, or cold wave advisory grades. In the case of air quality disaster types, forecast risk indicators may be calculated based on fine dust forecast values, ultrafine dust forecast values, yellow dust forecast values, or air quality warning grades.

[0106] On-site risk indicators may be values ​​or conditions representing the degree of disaster risk derived from on-site disaster detection data collected at a building site. For example, in the case of heavy rain or flood disaster types, on-site risk indicators may be calculated based on on-site rainfall, drainage tank water levels, leakage signals, or flood signals. In the case of heatwave disaster types, on-site risk indicators may be calculated based on outdoor temperature, indoor temperature, humidity, or equipment temperature. In the case of heavy snowfall or freezing disaster types, on-site risk indicators may be calculated based on snow accumulation, freezing detection values, load detection values, or road surface temperature. In the case of air quality disaster types, on-site risk indicators may be calculated based on outdoor air quality, indoor air quality, particulate matter concentration, or hazardous gas concentration.

[0107] Judgment conditions by disaster type may include forecast criteria, field criteria, or composite judgment criteria corresponding to each disaster type. Judgment conditions by disaster type may include forecast judgment conditions applied to forecast risk indicators and field judgment conditions applied to field risk indicators. Forecast judgment conditions may include whether the disaster forecast value is above a threshold value, whether the alarm level is above a threshold level, or whether a disaster notification is included. Field judgment conditions may include whether the sensor measurement value is above a threshold value, whether the status value is in a dangerous state, or whether an alarm value has occurred.

[0108] The judgment state classification unit (155) first determines whether both the forecast risk indicator and the field risk indicator satisfy the judgment conditions for each type of disaster. If both the forecast risk indicator and the field risk indicator satisfy the judgment conditions for each type of disaster, the judgment state classification unit (155) determines the state of the building as a local disaster state. A local disaster state may mean a state in which a risk is confirmed in the wide-area disaster forecast data and a risk is also confirmed in the detection data at the building site.

[0109] For example, in the case of a heavy rain or flood disaster type, if the precipitation forecast value or heavy rain warning satisfies the forecast judgment condition and at the same time the drainage tank water level, leakage signal, or flood signal satisfies the field judgment condition, the judgment state classification unit (155) can determine the state of the building in question as a local disaster state. In this case, since both the forecast disaster risk and the field disaster risk are confirmed, the control execution unit (170) can determine whether to drive the disaster prevention equipment (300) corresponding to the disaster type.

[0110] The judgment state classification unit (155) determines whether the forecast risk indicator satisfies the judgment conditions for each type of disaster, but the field risk indicator does not satisfy the judgment conditions for each type of disaster. If the forecast risk indicator satisfies the judgment conditions for each type of disaster and the field risk indicator does not satisfy the judgment conditions for each type of disaster, the judgment state classification unit (155) determines the state of the building as a forecast-dominant discrepancy state. A forecast-dominant discrepancy state may mean a state in which there is a risk in the external forecast, but the risk detection value has not yet been confirmed at the building site.

[0111] For example, in the case of a heavy snow or freezing disaster type, if the snowfall forecast value or freezing forecast value satisfies the forecast judgment condition but the on-site snowfall amount, freezing detection value, or load detection value does not satisfy the on-site judgment condition, the judgment state classification unit (155) may determine the state of the building as a forecast-dominant mismatch state. In this case, the control execution unit (170) may select an alarm mode or a preliminary drive mode rather than immediately selecting the main drive mode.

[0112] The judgment state classification unit (155) determines whether the forecast risk indicator does not satisfy the judgment conditions for each type of disaster, but the field risk indicator satisfies the judgment conditions for each type of disaster. If the forecast risk indicator does not satisfy the judgment conditions for each type of disaster and the field risk indicator satisfies the judgment conditions for each type of disaster, the judgment state classification unit (155) determines the state of the building as a local occurrence discrepancy state. A local occurrence discrepancy state may mean a state in which the risk is low or not indicated in the external forecast, but an actual risk is detected at the building site.

[0113] For example, in an air quality disaster type, if the fine dust forecast value or yellow dust forecast value does not satisfy the forecast judgment condition, but the air quality sensor value in the building's outdoor air intake or indoor space satisfies the field judgment condition, the judgment state classification unit (155) can determine the state of the building as a local occurrence discrepancy state. Additionally, in a heavy rain or flood disaster type, even if there is no heavy rain forecast, if the drainage tank water level or leakage signal satisfies the field judgment condition, the judgment state classification unit (155) can determine the state as a local occurrence discrepancy state.

[0114] The judgment state classification unit (155) can determine the state of the building to be normal if both the forecast risk indicator and the field risk indicator do not satisfy the judgment conditions for each type of disaster. A normal state may mean a state in which no disaster risk is confirmed in both external forecast data and field detection data. In this case, the control execution unit (170) can maintain the disaster prevention equipment (300) in standby mode.

[0115] The disaster cross-judgment unit (150) can independently classify each judgment state for multiple types of disasters. For example, for a single building, the heavy rain or flood disaster type may be judged as a normal state, and the air quality disaster type may be judged as a local occurrence inconsistency state. In this case, the disaster cross-judgment unit (150) provides the judgment results for each disaster type to the control execution unit (170), and the control execution unit (170) can determine the control mode of different disaster prevention facilities (300) for each disaster type.

[0116] The disaster cross-judgment unit (150) may include at least one of building identification information, disaster type code, judgment time interval, forecast risk indicator, site risk indicator, judgment status, and judgment basis data in the judgment result. The judgment basis data may be information indicating whether the forecast risk indicator and the site risk indicator each satisfy the judgment conditions for each disaster type. These judgment results may be provided to the control execution unit (170), the history management database (500), or the monitoring interface unit (600).

[0117] In this way, the disaster cross-determination unit (150) according to the present embodiment does not simply determine wide-area disaster forecast data and on-site disaster detection data individually, but can classify the disaster status of a building according to the combined relationship between the forecast risk indicator and the on-site risk indicator. Accordingly, the system according to the present embodiment can distinguish between cases where both the forecast and the site are dangerous, cases where only the forecast is dangerous, cases where only the site is dangerous, and cases where both are normal, and use this to determine the subsequent control mode.

[0118] FIG. 1 is an overall configuration diagram of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to an embodiment of the present invention, FIG. 5 is a flowchart of a control mode determination of a control execution unit according to an embodiment of the present invention, and FIG. 9 is an overall flowchart of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control method according to an embodiment of the present invention.

[0119] According to FIGS. 1, 5 and 9, the control execution unit (170) according to the present embodiment determines the control mode of the disaster prevention facility (300) based on the judgment result of the disaster cross-determination unit (150). The control execution unit (170) selects one of the standby mode, alarm mode, pre-operation mode, and main operation mode of the disaster prevention facility (300) based on any one of the local disaster state, forecast dominance mismatch state, and local occurrence mismatch state determined by the disaster cross-determination unit (150). The control execution unit (170) transmits a control output corresponding to the selected control mode to the DDC data linkage unit (190), and the DDC data linkage unit (190) converts the control output into output control data that can be processed by the direct digital controller (210) and transmits it to the disaster prevention facility (300).

[0120] The control execution unit (170) receives a judgment result from the disaster cross-judgment unit (150). The judgment result may include at least one of building identification information, a disaster type code, a judgment time interval, a forecast risk indicator, a site risk indicator, a judgment status, and judgment basis data. The judgment status may be any one of a local disaster status, a forecast dominance discrepancy status, a local occurrence discrepancy status, or a normal status. The control execution unit (170) may select a disaster prevention facility (300) to be controlled by using the judgment status and the disaster type code together.

[0121] The control execution unit (170) determines whether the judgment state is a local disaster state. A local disaster state is a state in which both the forecast risk indicator and the field risk indicator satisfy the judgment conditions for each type of disaster. If the judgment state is a local disaster state, the control execution unit (170) can select the main driving mode of the disaster prevention equipment (300) corresponding to the disaster type. The main driving mode may be a control mode in which the disaster prevention equipment (300) actually performs disaster response operations.

[0122] For example, if a local disaster condition is determined in the case of a heavy rain or flood disaster type, the control execution unit (170) may select the main operation mode of a drainage pump or a water barrier. If a local disaster condition is determined in the case of a heatwave disaster type, the control execution unit (170) may select the main operation mode of an air conditioner, a supply fan, or an exhaust fan. If a local disaster condition is determined in the case of a heavy snow or freezing disaster type, the control execution unit (170) may select the main operation mode of a heating cable or a snow removal facility. If a local disaster condition is determined in the case of an air quality disaster type, the control execution unit (170) may select the main operation mode of an air conditioning damper, a supply fan, an exhaust fan, or an air purification facility.

[0123] The control execution unit (170) determines whether the judgment state is a forecast dominance mismatch state. A forecast dominance mismatch state is a state in which the forecast risk indicator satisfies the judgment conditions for each disaster type, but the field risk indicator does not satisfy the judgment conditions for each disaster type. If the judgment state is a forecast dominance mismatch state, the control execution unit (170) may select an alarm mode or a backup operation mode of the disaster prevention facility (300). The alarm mode may be a mode that provides a disaster forecast-based alarm to an administrator, a control system, or a monitoring interface unit (600). The backup operation mode may be a mode that switches the disaster prevention facility (300) to a ready state before main operation.

[0124] For example, if there is a heavy rain forecast or flood risk alert but the on-site water level, leakage signal, or flood signal does not yet satisfy the danger condition, the control execution unit (170) may switch the drainage pump to a pre-operation state or set the ready-operation state of the water barrier. If there is a heavy snow or freezing forecast but the on-site snow accumulation or freezing detection value does not satisfy the danger condition, the control execution unit (170) may select a pre-operation mode for the heating cable or snow removal equipment. If there is a forecast of air quality deterioration but the on-site air quality data does not satisfy the danger condition, the control execution unit (170) may provide an alert to the manager or select a pre-control mode for the air conditioning damper.

[0125] The control execution unit (170) determines whether the judgment state is a local occurrence mismatch state. A local occurrence mismatch state is a state in which the forecast risk indicator does not satisfy the judgment conditions for each disaster type, but the field risk indicator satisfies the judgment conditions for each disaster type. If the judgment state is a local occurrence mismatch state, the control execution unit (170) transmits a local occurrence alarm to the administrator terminal or the monitoring interface unit (600) and can select one of the alarm mode, the backup operation mode, or the main operation mode of the disaster prevention equipment (300).

[0126] For example, even if there is no forecast of heavy rain, if a rise in water level, a leakage signal, or a flooding signal is detected in the basement or drainage tank of a building, the control execution unit (170) can transmit a local occurrence alarm and select a preliminary driving mode or a main driving mode of the drainage pump or water barrier. Even if there is no forecast of air quality deterioration, if an air quality risk value is detected in the outdoor air intake or indoor space, the control execution unit (170) can change the control mode of the air conditioning damper, supply fan, or exhaust fan.

[0127] The control execution unit (170) can select a standby mode of the disaster prevention facility (300) when the judgment state is a normal state. A normal state is a state in which both the forecast risk indicator and the field risk indicator do not satisfy the judgment conditions for each type of disaster. A standby mode may be a mode that maintains the disaster prevention facility (300) in a normal operating state or a disaster response standby state. The control execution unit (170) may limit the control output of the disaster prevention facility (300) or maintain the existing operating state so that unnecessary main operation or preliminary operation is not performed in a normal state.

[0128] The control execution unit (170) selects a disaster prevention facility (300) to be controlled according to a disaster type code. For a heavy rain or flood disaster type, at least one of a drainage pump, a water barrier, an emergency shut-off relay, or a fire shutter may be the control target. For a heatwave disaster type, at least one of an air conditioner, a supply fan, an exhaust fan, or an air conditioning damper may be the control target. For a heavy snow or freezing disaster type, at least one of a heating cable, a snow removal facility, or a district heating valve may be the control target. For an air quality disaster type, at least one of an air conditioning damper, a supply fan, an exhaust fan, or an air purification facility may be the control target.

[0129] The control execution unit (170) generates a control output according to the selected control mode. The control output may include identification information of the controlled equipment, control mode information, driving intensity, driving time, output value, priority, alarm information, or control command code. For example, the control output of the driving mode may include a command to switch the fire protection equipment (300) to an actual driving state. The control output of the pre-driving mode may include a command to instruct the fire protection equipment (300) to pre-inspect, build atmospheric pressure, partially open, partially operate, or preparatory operation. The control output of the alarm mode may include a command for administrator alarms, control screen displays, or warning message generation.

[0130] The control execution unit (170) transmits the generated control output to the DDC data linkage unit (190). The DDC data linkage unit (190) converts the control output received from the control execution unit (170) into output control data that can be processed by the direct digital controller (210). The output control data can be transmitted to the output control unit (270) of the direct digital controller (210). The output control unit (270) can operate at least one of the following according to the output control data: a drainage pump output control unit, a water barrier output control unit, an air conditioner output control unit, an air conditioner damper output control unit, a supply fan output control unit, an exhaust fan output control unit, a heating wire output control unit, a snow removal equipment output control unit, a valve output control unit, a fire shutter output control unit, or an emergency cutoff relay output control unit.

[0131] The control execution unit (170) may further consider disaster prevention equipment status data during the control mode determination process. For example, if the disaster prevention equipment (300) is in a broken state or in a manual operation state, the control execution unit (170) may restrict the main operation command for the disaster prevention equipment (300) and prioritize the generation of an administrator alarm. Additionally, if multiple disaster prevention equipment (300) correspond to the same disaster type, the control execution unit (170) may select a control target based on the equipment's current state, operational availability, or priority.

[0132] The control execution unit (170) may store the control mode determination result in the history management database (500). The control mode determination result may include at least one of building identification information, disaster type code, judgment status, selected control mode, control target equipment, control output, control time, and equipment status data. Additionally, the control execution unit (170) may provide the control mode determination result to the monitoring interface unit (600) so that an administrator can check the current disaster status and the control status of the disaster prevention equipment (300).

[0133] In this way, the control execution unit (170) according to the present embodiment can select one of the standby mode, alarm mode, pre-operation mode, and main operation mode of the disaster prevention facility (300) according to the judgment state of the disaster cross-judgment unit (150), and transmit a control output corresponding to the selected control mode to the direct digital controller (210) through the DDC data linkage unit (190). Accordingly, the disaster prevention facility (300) can be controlled stepwise in response to the type and degree of the disaster state.

[0134] FIG. 1 is an overall configuration diagram of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to one embodiment of the present invention, FIG. 6 is a flowchart of a response to heavy rain or flood disaster according to one embodiment of the present invention, and FIG. 7 is a detailed block diagram of a control execution unit according to one embodiment of the present invention.

[0135] According to FIGS. 1, 6 and 7, the control execution unit (170) according to the present embodiment determines the control mode of the drainage pump or water barrier based on at least one of a precipitation forecast value, on-site rainfall amount, drainage tank water level, leakage signal, flood signal, drainage pump operation status and water barrier status for a heavy rain or flood disaster type. The control execution unit (170) may include a heavy rain or flood disaster judgment unit (171), and the heavy rain or flood disaster judgment unit (171) may control the drainage pump or water barrier to any one of a standby mode, an alarm mode, a preliminary operation mode, or a main operation mode using the judgment result of the disaster cross-judgment unit (150) and disaster prevention facility status data.

[0136] The heavy rain or flood disaster determination unit (171) may use a comparison dataset of heavy rain or flood disaster types generated by the data matching unit (120). The comparison dataset of heavy rain or flood disaster types may include at least one of precipitation forecast values, precipitation probability, heavy rain warnings, flood risk alerts, or drainage risk alerts included in the wide-area disaster forecast data. Additionally, the comparison dataset of heavy rain or flood disaster types may include at least one of field rainfall amount, drainage tank water level, sump water level, leakage signal, flood signal, or basement water level detection value included in the field disaster detection data.

[0137] The heavy rain or flood disaster judgment unit (171) first checks at least one of the following: a precipitation forecast value, on-site rainfall amount, drainage tank water level, leakage signal, flood signal, drainage pump operation status, and water barrier status. The precipitation forecast value may be included in wide-area disaster forecast data collected from a weather agency server (10) or a disaster information provision server (20). The on-site rainfall amount, drainage tank water level, leakage signal, and flood signal may be included in on-site disaster detection data collected through a direct digital controller (210). The drainage pump operation status and water barrier status may be included in disaster prevention equipment status data collected through a direct digital controller (210).

[0138] The data matching unit (120) can align precipitation forecast values ​​and field rainfall amounts on a common time axis. For example, if precipitation forecast values ​​are provided in hourly units and field rainfall amounts are collected in minute units, the data matching unit (120) can convert the field rainfall amounts into cumulative values ​​or representative values ​​corresponding to the forecast time interval. Additionally, the data matching unit (120) can generate a heavy rain or flood comparison dataset by aligning the drainage tank water level, leakage signals, and flood signals to the same judgment time interval.

[0139] The disaster cross-judgment unit (150) determines forecast risk indicators and field risk indicators based on a heavy rain or flood comparison dataset. The forecast risk indicator may be determined based on precipitation forecast values, precipitation probability, heavy rain warning grades, or flood risk notifications. The field risk indicator may be determined based on field rainfall amount, drainage tank water level, leakage signals, or flood signals. The judgment conditions for each disaster type may include at least one of a precipitation amount criterion, a drainage tank water level criterion, whether leakage is detected, whether flood is detected, or an alarm grade criterion.

[0140] The disaster cross-determination unit (150) determines the condition of the building as a local disaster state when both the forecast risk indicator and the field risk indicator satisfy the judgment conditions for a heavy rain or flood disaster type. In this case, the heavy rain or flood disaster judgment unit (171) can select the main operation mode of the drainage pump or the water barrier. For example, if the precipitation forecast value is above the reference value and the drainage tank water level or flood signal satisfies the reference conditions, the heavy rain or flood disaster judgment unit (171) can switch the drainage pump to the main operation state or switch the water barrier to the closed state.

[0141] When the heavy rain or flood disaster judgment unit (171) selects the main driving mode of the drainage pump, it transmits the control output to the DDC data linkage unit (190). The DDC data linkage unit (190) converts the control output into output control data that can be processed by the direct digital controller (210). The drainage pump output control unit (271) of the direct digital controller (210) can drive the drainage pump according to the output control data. Additionally, when the heavy rain or flood disaster judgment unit (171) selects the main driving mode of the water barrier, the water barrier output control unit (273) can close the water barrier or move it to a flood blocking position according to the output control data.

[0142] The disaster cross-judgment unit (150) can determine the condition of a building as a forecast-dominant discrepancy state if the forecast risk indicator satisfies the judgment conditions for a heavy rain or flood disaster type, but the on-site risk indicator does not satisfy the judgment conditions. In this case, the heavy rain or flood disaster judgment unit (171) can control the drainage pump or water barrier in an alarm mode or a backup mode. For example, if there is a heavy rain forecast but the water level of the drainage tank and the leakage signal do not yet satisfy the danger conditions, the heavy rain or flood disaster judgment unit (171) can provide a heavy rain forecast-based alarm to the manager and switch the drainage pump to a backup mode.

[0143] The pre-operation mode may be a mode in which the drainage pump or water barrier is switched to a ready state prior to the main operation. For example, the pre-operation mode of the drainage pump may include checking the operating status of the drainage pump, standby operation, partial operation, or automatic operation standby state. The pre-operation mode of the water barrier may include checking the operating status of the water barrier, partial closure, unlocking, or operation standby state. The heavy rain or flood disaster judgment unit (171) may switch to the main operation mode when the field risk indicator satisfies the subsequent judgment conditions in a forecast dominance discrepancy state.

[0144] The disaster cross-judgment unit (150) can determine the condition of a building as a local occurrence discrepancy state if the forecast risk indicator does not satisfy the judgment conditions for the heavy rain or flood disaster type, but the on-site risk indicator satisfies the judgment conditions. In this case, the heavy rain or flood disaster judgment unit (171) can transmit a local flood warning to the manager terminal or monitoring interface unit (600) and select a preliminary operation mode or a main operation mode for the drainage pump or water barrier. For example, even if there is no heavy rain forecast, if the water level of the drainage tank rises above a reference value or a leakage signal or flood signal occurs, the heavy rain or flood disaster judgment unit (171) can generate a local flood warning.

[0145] The heavy rain or flood disaster judgment unit (171) can select a different control mode depending on the degree of the on-site risk indicator in a local occurrence discrepancy state. For example, if the water level of the drainage tank reaches the warning standard, an alarm mode or a preliminary driving mode can be selected, and if the water level of the drainage tank reaches the danger standard or a flood signal occurs, a main driving mode can be selected. In addition, if a leakage signal occurs at a single point, an alarm mode can be selected, and if leakage signals occur at multiple points or flood signals occur together, a main driving mode can be selected.

[0146] The heavy rain or flood disaster judgment unit (171) may further consider disaster prevention facility status data. For example, if the drainage pump operation status is in a faulty state, the heavy rain or flood disaster judgment unit (171) may restrict the main drive command of the drainage pump and prioritize the transmission of an administrator alarm. Additionally, if the water barrier status is in a non-operational state, the heavy rain or flood disaster judgment unit (171) may prioritize the operation of the drainage pump or the control of the emergency shut-off relay instead of controlling the water barrier. Conversely, if both the drainage pump and the water barrier are in a normal state, the heavy rain or flood disaster judgment unit (171) may control the two facilities sequentially or in parallel depending on the disaster status.

[0147] The heavy rain or flood disaster judgment unit (171) can maintain the drainage pump or water barrier in standby mode when the judgment state is normal. The normal state may be a state in which the precipitation forecast value, the on-site rainfall amount, the drainage tank water level, the leakage signal, and the flood signal do not all satisfy the judgment conditions for the heavy rain or flood disaster type. In this case, the heavy rain or flood disaster judgment unit (171) can limit unnecessary operation of the drainage pump or water barrier and maintain the existing operation state.

[0148] The heavy rain or flood disaster judgment unit (171) may store the results of the heavy rain or flood response in the history management database (500). The history management database (500) may store at least one of building identification information, judgment time interval, precipitation forecast value, on-site rainfall amount, drainage tank water level, leakage signal, flood signal, drainage pump operation status, water barrier status, judgment status, selected control mode, and control output. The stored information may subsequently be used for history analysis of the same disaster type or for identifying vulnerable buildings.

[0149] Additionally, the heavy rain or flood disaster judgment unit (171) can provide the results of the heavy rain or flood response to the monitoring interface unit (600). The monitoring interface unit (600) can display precipitation forecast values, on-site rainfall amounts, drainage tank water levels, leakage signals, flood signals, and the control status of drainage pumps or water barriers. Accordingly, the manager can check forecast data, on-site data, and the control status of disaster prevention facilities for the type of heavy rain or flood disaster together.

[0150] In this way, the heavy rain or flood disaster determination unit (171) according to the present embodiment can determine the heavy rain or flood disaster state by considering the precipitation forecast value, the on-site rainfall amount, the drainage tank water level, and the leakage signal or flood signal together, and can determine the control mode of the drainage pump or water barrier. Accordingly, the system according to the present embodiment can control the drainage pump or water barrier by reflecting the wide-area heavy rain forecast and the flood risk at the building site together.

[0151] FIG. 1 is an overall configuration diagram of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to an embodiment of the present invention, FIG. 7 is a detailed block diagram of a control execution unit according to an embodiment of the present invention, and FIG. 9 is an overall flowchart of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control method according to an embodiment of the present invention.

[0152] According to FIGS. 1, 7 and 9, the control execution unit (170) according to the present embodiment determines the control mode of an air conditioner, air conditioner damper, supply fan, or exhaust fan based on at least one of a temperature forecast value, a heatwave warning, an outdoor temperature, an indoor temperature, humidity, equipment temperature, an air conditioner status, a supply fan status, and an exhaust fan status, for a heatwave disaster type. The control execution unit (170) may include a heatwave disaster judgment unit (173), and the heatwave disaster judgment unit (173) may control the air conditioner, air conditioner damper, supply fan, or exhaust fan to any one of a standby mode, an alarm mode, a pre-operation mode, or a main operation mode using the judgment result of the disaster cross-judgment unit (150) and disaster prevention equipment status data.

[0153] The heatwave disaster judgment unit (173) may use a comparison dataset of heatwave disaster types generated by the data matching unit (120). The comparison dataset of heatwave disaster types may include at least one of the temperature forecast value, perceived temperature forecast value, heatwave warning, tropical night forecast, or high temperature risk alert included in the wide-area disaster forecast data. Additionally, the comparison dataset of heatwave disaster types may include at least one of the outdoor temperature, indoor temperature, humidity, equipment temperature, air conditioner operating status, supply fan operating status, or exhaust fan operating status included in the field disaster detection data.

[0154] The heatwave disaster judgment unit (173) first checks at least one of the temperature forecast value, heatwave warning, outdoor temperature, indoor temperature, humidity, equipment temperature, air conditioner status, supply fan status, and exhaust fan status. The temperature forecast value and heatwave warning may be included in wide-area disaster forecast data collected from a weather agency server (10) or a disaster information provision server (20). The outdoor temperature, indoor temperature, humidity, and equipment temperature may be included in field disaster detection data collected through a direct digital controller (210). The air conditioner status, air conditioner damper status, supply fan status, and exhaust fan status may be included in disaster prevention equipment status data collected through a direct digital controller (210).

[0155] The data matching unit (120) can align the temperature forecast value or heat wave warning with the outdoor temperature, indoor temperature, humidity, or facility temperature on a common time axis. For example, if the temperature forecast value is provided in hourly units and the indoor temperature or outdoor temperature is collected in minute units, the data matching unit (120) can convert the field temperature data into an average value, a maximum value, or a representative value corresponding to the forecast time interval. Additionally, the data matching unit (120) can generate a heat wave comparison dataset by aligning the humidity and facility temperature to the same judgment time interval.

[0156] The disaster cross-judgment unit (150) determines forecast risk indicators and field risk indicators based on a heatwave comparison dataset. Forecast risk indicators may be determined based on temperature forecast values, perceived temperature forecast values, heatwave warning grades, or high temperature risk alerts. Field risk indicators may be determined based on outdoor temperature, indoor temperature, humidity, equipment temperature, or air conditioning equipment status. The judgment conditions for each disaster type may include at least one of temperature criteria, perceived temperature criteria, indoor temperature criteria, humidity criteria, equipment temperature criteria, or warning grade criteria.

[0157] The disaster cross-determination unit (150) determines the condition of the building as a local disaster state when both the forecast risk indicator and the on-site risk indicator satisfy the conditions for determining the heatwave disaster type. In this case, the heatwave disaster judgment unit (173) can select the main driving mode of the air conditioner, air conditioner damper, supply fan, or exhaust fan. For example, when the temperature forecast value or the heatwave warning satisfies the standard conditions and at the same time the indoor temperature or outdoor temperature is above the standard value, the heatwave disaster judgment unit (173) can select the cooling operation of the air conditioner, the increased operation of the supply fan, or the ventilation operation of the exhaust fan as the main driving mode.

[0158] When the heatwave disaster judgment unit (173) selects the main driving mode of the air conditioner, it transmits the control output to the DDC data linkage unit (190). The DDC data linkage unit (190) converts the control output into output control data that can be processed by the direct digital controller (210). The air conditioner output control unit (275) of the direct digital controller (210) can control the air conditioner's operating mode, cooling setting value, or airflow volume according to the output control data. Additionally, the supply fan output control unit (279) and the exhaust fan output control unit (281) can adjust the supply air volume or exhaust air volume according to the output control data.

[0159] The heatwave disaster judgment unit (173) can select an air conditioning damper as a control target. For example, if the outside air temperature is higher than the indoor temperature and the conditions for determining a heatwave disaster type are satisfied, the heatwave disaster judgment unit (173) can transmit a control output to the air conditioning damper output control unit (277) to limit the amount of outside air introduced. Conversely, during the night or a time period when the outside air temperature is relatively low, the opening of the air conditioning damper can be increased to perform outside air cooling or ventilation control.

[0160] The disaster cross-judgment unit (150) can determine the condition of a building as a forecast-dominant discrepancy state if the forecast risk indicator satisfies the judgment conditions for a heatwave disaster type, but the on-site risk indicator does not satisfy the judgment conditions. In this case, the heatwave disaster judgment unit (173) can control the air conditioner, air conditioner damper, supply fan, or exhaust fan in an alarm mode or a backup mode. For example, if a heatwave warning exists but the indoor temperature or facility temperature does not yet satisfy the danger conditions, the heatwave disaster judgment unit (173) can provide a heatwave forecast-based warning to the manager and switch the air conditioner to a backup cooling state.

[0161] A pre-operation mode may be a mode in which an air conditioner, an air conditioner damper, a supply fan, or an exhaust fan is switched to a ready state prior to main operation. For example, a pre-operation mode of an air conditioner may include preparation for cooling operation, pre-adjustment of a set temperature, preparation for chilled water supply, or partial cooling operation. A pre-operation mode of a supply fan or an exhaust fan may include checking the fan start-up status, low-speed operation, or pre-adjustment of ventilation volume. A pre-operation mode of an air conditioner damper may include partial closing of an outside air damper, partial opening of an exhaust damper, or adjustment of the circulating air volume.

[0162] The disaster cross-judgment unit (150) can determine the condition of a building as a local occurrence inconsistency state if the forecast risk indicator does not satisfy the judgment condition for a heatwave disaster type, but the on-site risk indicator satisfies the judgment condition. In this case, the heatwave disaster judgment unit (173) transmits a local high temperature alarm to the manager terminal or monitoring interface unit (600) and can select one of the alarm mode, pre-operation mode, or main operation mode of the air conditioner, air conditioner damper, supply fan, or exhaust fan. For example, even if there is no heatwave forecast, if the indoor temperature or equipment temperature inside the building rises above a reference value, the heatwave disaster judgment unit (173) can generate a local high temperature alarm and select the main operation mode of the air conditioner.

[0163] The heatwave disaster judgment unit (173) can select a different control mode depending on the degree of the on-site risk indicator in a localized inconsistency state. For example, if the indoor temperature reaches the warning standard, an alarm mode or a backup drive mode can be selected, and if the indoor temperature reaches the danger standard or the equipment temperature reaches the overheating standard, the main drive mode can be selected. In addition, if only the indoor temperature of a specific zone rises, the air conditioner or fan of that zone can be selectively controlled, and if a high temperature condition is detected in multiple zones, the air conditioning control mode of the entire building can be changed.

[0164] The heatwave disaster judgment unit (173) may further consider disaster prevention equipment status data. For example, if the air conditioner is in a malfunction state, the heatwave disaster judgment unit (173) may restrict the main operation command of the air conditioner and prioritize the transmission of an administrator alarm. Additionally, if the supply fan or exhaust fan is in a non-operational state, the heatwave disaster judgment unit (173) may select an air conditioner damper or other air conditioner equipment as an alternative control target. Conversely, if the air conditioner, supply fan, and exhaust fan are all in a normal state, the heatwave disaster judgment unit (173) may control these equipment sequentially or in parallel according to the heatwave disaster state.

[0165] When the heatwave disaster judgment unit (173) determines that the judgment state is normal, it can maintain the air conditioner, air conditioner damper, supply fan, or exhaust fan in standby mode or normal operation mode. The normal state may be a state in which the temperature forecast value, heatwave warning, outside temperature, indoor temperature, humidity, and equipment temperature do not all satisfy the judgment conditions for a heatwave disaster type. In this case, the heatwave disaster judgment unit (173) can limit unnecessary cooling main operation or excessive ventilation operation and maintain the existing operation state.

[0166] The heatwave disaster judgment unit (173) may store the results of the heatwave response in the history management database (500). The history management database (500) may store at least one of building identification information, judgment time interval, temperature forecast value, heatwave warning, outdoor temperature, indoor temperature, humidity, equipment temperature, air conditioner status, supply fan status, exhaust fan status, judgment status, selected control mode, and control output. The stored information may subsequently be used for history analysis of the same type of disaster or for identifying vulnerable buildings.

[0167] Additionally, the heatwave disaster judgment unit (173) can provide the results of the heatwave response to the monitoring interface unit (600). The monitoring interface unit (600) can display the temperature forecast value, the outdoor temperature, the indoor temperature, the humidity, the equipment temperature, and the control status of the air conditioner or fan. Accordingly, the manager can check the forecast data, field data, and the control status of the disaster prevention equipment for the type of heatwave disaster together.

[0168] In this way, the heatwave disaster judgment unit (173) according to the present embodiment can determine the heatwave disaster state by considering the temperature forecast value or heatwave warning together with the outdoor temperature, indoor temperature, humidity or facility temperature, and can determine the control mode of the air conditioner, air conditioning damper, supply fan or exhaust fan. Accordingly, the system according to the present embodiment can control the air conditioning-related disaster prevention equipment by reflecting the wide-area heatwave forecast and the high temperature condition of the building site together.

[0169] FIG. 1 is an overall configuration diagram of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to an embodiment of the present invention, FIG. 7 is a detailed block diagram of a control execution unit according to an embodiment of the present invention, and FIG. 9 is an overall flowchart of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control method according to an embodiment of the present invention.

[0170] According to FIGS. 1, 7 and 9, the control execution unit (170) according to the present embodiment determines the control mode of the heating cable or snow removal equipment based on at least one of a snowfall forecast value, a freezing forecast value, an ambient temperature, a field snowfall amount, a freezing detection value, load data, a heating cable status, and a snow removal equipment status for a heavy snowfall or freezing disaster type. The control execution unit (170) may include a heavy snowfall or freezing disaster judgment unit (175), and the heavy snowfall or freezing disaster judgment unit (175) may control the heating cable or snow removal equipment to any one of a standby mode, an alarm mode, a preliminary driving mode, or a main driving mode using the judgment result of the disaster cross-judgment unit (150) and disaster prevention equipment status data.

[0171] The heavy snow or freezing disaster determination unit (175) may use a comparison dataset of heavy snow or freezing disaster types generated by the data matching unit (120). The comparison dataset of heavy snow or freezing disaster types may include at least one of snow depth forecast values, snow depth probability, cold wave warnings, freezing forecast values, road surface freezing risk alerts, or heavy snow warnings included in wide-area disaster forecast data. Additionally, the comparison dataset of heavy snow or freezing disaster types may include at least one of field snow depth, freezing detection values, ambient temperature, road surface temperature, load data, or structural vibration data included in field disaster detection data.

[0172] The heavy snow or freezing disaster judgment unit (175) first checks at least one of the snow accumulation forecast value, freezing forecast value, ambient temperature, on-site snow accumulation amount, freezing detection value, load data, heating cable status, and snow removal equipment status. The snow accumulation forecast value, freezing forecast value, or heavy snow warning may be included in wide-area disaster forecast data collected from a weather agency server (10) or a disaster information provision server (20). The on-site snow accumulation amount, freezing detection value, ambient temperature, road surface temperature, and load data may be included in on-site disaster detection data collected through a direct digital controller (210). The heating cable status and snow removal equipment status may be included in disaster prevention equipment status data collected through a direct digital controller (210).

[0173] The data matching unit (120) can align snowfall forecast values ​​or freezing forecast values ​​with field snowfall amounts, freezing detection values, ambient temperature, or load data on a common time axis. For example, if snowfall forecast values ​​are provided in hourly or daily units and field snowfall amounts are collected in minute or hourly units, the data matching unit (120) can convert field snowfall amounts into cumulative values, maximum values, or representative values ​​corresponding to the forecast time interval. Additionally, the data matching unit (120) can generate a heavy snowfall or freezing comparison dataset by aligning ambient temperature, road surface temperature, freezing detection values, and load data with the same judgment time interval.

[0174] The disaster cross-judgment unit (150) determines forecast risk indicators and field risk indicators based on a heavy snow or freezing comparison dataset. The forecast risk indicator may be determined based on snow depth forecast values, freezing forecast values, cold wave warning grades, heavy snow warning grades, or road surface freezing risk notifications. The field risk indicator may be determined based on field snow depth, freezing detection values, ambient temperature, road surface temperature, load data, or structural vibration data. The judgment conditions for each disaster type may include at least one of snow depth criteria, freezing detection criteria, temperature criteria, load criteria, vibration criteria, or warning grade criteria.

[0175] The disaster cross-determination unit (150) determines the condition of the building as a local disaster state when both the forecast risk indicator and the field risk indicator satisfy the judgment conditions for the heavy snow or freezing disaster type. In this case, the heavy snow or freezing disaster judgment unit (175) can select the main operation mode of the heating cable or snow removal equipment. For example, if the snowfall forecast value or the heavy snow warning satisfies the standard conditions and at the same time the field snowfall amount or load data is greater than or equal to the standard value, the heavy snow or freezing disaster judgment unit (175) can select the main operation mode of the snow removal equipment. Additionally, if the freezing forecast value or the cold wave warning satisfies the standard conditions and at the same time the freezing detection value or the road surface temperature satisfies the standard conditions, the heavy snow or freezing disaster judgment unit (175) can select the main operation mode of the heating cable.

[0176] When the heavy snow or freezing disaster judgment unit (175) selects the main driving mode of the heating cable, it transmits the control output to the DDC data linkage unit (190). The DDC data linkage unit (190) converts the control output into output control data that can be processed by the direct digital controller (210). The heating cable output control unit (283) of the direct digital controller (210) can control the energization state or heating intensity of the heating cable according to the output control data. Additionally, when the heavy snow or freezing disaster judgment unit (175) selects the main driving mode of the snow removal facility, the snow removal facility output control unit (285) can drive the snow removal facility according to the output control data.

[0177] The disaster cross-judgment unit (150) can determine the condition of a building as a forecast-dominant discrepancy state if the forecast risk indicator satisfies the judgment conditions for a heavy snow or freezing disaster type, but the on-site risk indicator does not satisfy the judgment conditions. In this case, the heavy snow or freezing disaster judgment unit (175) can control the heating cable or snow removal equipment in an alarm mode or a backup mode. For example, if a heavy snow forecast or freezing forecast exists but the on-site snow accumulation, freezing detection value, or load data does not yet satisfy the danger conditions, the heavy snow or freezing disaster judgment unit (175) can provide a heavy snow or freezing forecast-based alarm to the manager and switch the heating cable or snow removal equipment to a backup mode.

[0178] The pre-operation mode may be a mode in which the heating cable or snow removal equipment is switched to a ready state prior to the main operation. For example, the pre-operation mode of the heating cable may include checking the power supply status, low-power preheating, zone-specific preheating, or an automatic power supply standby state. The pre-operation mode of the snow removal equipment may include checking the snow removal equipment's operational status, setting snow removal zones, partial operation, or operation standby state. The heavy snow or freezing disaster judgment unit (175) may switch to the main operation mode when the field risk indicator satisfies subsequent judgment conditions in a forecast dominance discrepancy state.

[0179] The disaster cross-judgment unit (150) may determine the condition of a building as a local occurrence inconsistency state if the forecast risk indicator does not satisfy the judgment conditions for the heavy snow or freezing disaster type, but the on-site risk indicator satisfies the judgment conditions. In this case, the heavy snow or freezing disaster judgment unit (175) transmits a local snow accumulation warning or a local freezing warning to the manager terminal or monitoring interface unit (600) and may select one of the warning mode, pre-operation mode, or main operation mode of the heating cable or snow removal equipment. For example, even if there is no heavy snow forecast, if the on-site snow accumulation on the rooftop or external passageway increases above a reference value, the heavy snow or freezing disaster judgment unit (175) may generate a local snow accumulation warning and select the main operation mode of the snow removal equipment. Additionally, even if there is no freezing forecast, if a freezing detection value occurs in the road surface or piping area, the heavy snow or freezing disaster judgment unit (175) may generate a local freezing warning and select the main operation mode of the heating cable.

[0180] The snow or freezing disaster judgment unit (175) can select a different control mode depending on the degree of the on-site risk indicator in a local occurrence discrepancy state. For example, if the on-site snow accumulation reaches the warning standard, an alarm mode or a preliminary drive mode can be selected, and if the on-site snow accumulation reaches the danger standard or the load data is greater than the standard value, the main drive mode can be selected. In addition, if the freezing detection value occurs in a single zone, the heating wire of that zone can be selectively driven, and if the freezing detection value occurs in multiple zones, the heating wires or snow removal equipment of multiple zones can be driven together.

[0181] The heavy snow or freezing disaster judgment unit (175) may further consider disaster prevention facility status data. For example, if the heating cable status is in a broken state or a cutoff state, the heavy snow or freezing disaster judgment unit (175) may restrict the main operation command of the heating cable and prioritize the transmission of an administrator alarm. Additionally, if the snow removal facility status is in an inoperable state, the heavy snow or freezing disaster judgment unit (175) may select at least one of the district heating valve, heating cable, or emergency cutoff relay as an alternative control target instead of controlling the snow removal facility. Conversely, if both the heating cable and the snow removal facility are in a normal state, the heavy snow or freezing disaster judgment unit (175) may control these facilities sequentially or in parallel according to the disaster state.

[0182] The heavy snow or freezing disaster judgment unit (175) can maintain the heating cable or snow removal equipment in standby mode when the judgment state is normal. The normal state may be a state in which the snowfall forecast value, freezing forecast value, on-site snowfall amount, freezing detection value, ambient temperature, and load data do not all satisfy the judgment conditions for the heavy snow or freezing disaster type. In this case, the heavy snow or freezing disaster judgment unit (175) can limit unnecessary heating cable power supply or snow removal equipment operation and maintain the existing operating state.

[0183] The heavy snow or freezing disaster judgment unit (175) may store the results of the response to heavy snow or freezing in the history management database (500). The history management database (500) may store at least one of building identification information, judgment time interval, snowfall forecast value, freezing forecast value, on-site snowfall amount, freezing detection value, ambient temperature, load data, heating wire status, snow removal equipment status, judgment status, selected control mode, and control output. The stored information may subsequently be used for history analysis of the same disaster type or for identifying vulnerable buildings.

[0184] Additionally, the heavy snow or freezing disaster judgment unit (175) can provide the results of the response to heavy snow or freezing to the monitoring interface unit (600). The monitoring interface unit (600) can display snowfall forecast values, freezing forecast values, on-site snowfall amount, freezing detection values, load data, and the control status of heating wires or snow removal equipment. Accordingly, the manager can check forecast data, on-site data, and the control status of disaster prevention equipment for the type of heavy snow or freezing disaster together.

[0185] In this way, the snowfall or freezing disaster determination unit (175) according to the present embodiment can determine the snowfall or freezing disaster state by considering the snowfall forecast value or freezing forecast value together with the on-site snowfall amount, freezing detection value, and ambient temperature or load data, and can determine the control mode of the heating cable or snow removal equipment. Accordingly, the system according to the present embodiment can control the heating cable or snow removal equipment by reflecting the wide-area snowfall or freezing forecast and the risk of snowfall or freezing at the building site together.

[0186] FIG. 1 is an overall configuration diagram of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to an embodiment of the present invention, FIG. 7 is a detailed block diagram of a control execution unit according to an embodiment of the present invention, and FIG. 9 is an overall flowchart of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control method according to an embodiment of the present invention.

[0187] According to FIGS. 1, 7 and 9, the control execution unit (170) according to the present embodiment determines the control mode of the air conditioning damper, supply fan, or exhaust fan based on at least one of the following: fine dust forecast value, ultrafine dust forecast value, yellow dust forecast value, outdoor air quality data, indoor air quality data, air conditioning damper status, supply fan status, and exhaust fan status, for air quality disaster types. The control execution unit (170) may include an air quality disaster judgment unit (177), and the air quality disaster judgment unit (177) may control the air conditioning damper, supply fan, or exhaust fan to any one of a standby mode, an alarm mode, a pre-operation mode, or a main operation mode using the judgment result of the disaster cross-judgment unit (150) and disaster prevention equipment status data.

[0188] The air quality disaster judgment unit (177) may use a comparison dataset of air quality disaster types generated by the data matching unit (120). The comparison dataset of air quality disaster types may include at least one of fine dust forecast values, ultrafine dust forecast values, yellow dust forecast values, ozone forecast values, air quality warning grades, or air pollution notifications included in the wide-area disaster forecast data. Additionally, the comparison dataset of air quality disaster types may include at least one of outdoor air quality data, indoor air quality data, fine dust concentration, ultrafine dust concentration, carbon dioxide concentration, volatile organic compound concentration, or harmful gas concentration included in the field disaster detection data.

[0189] The air quality disaster judgment unit (177) first checks at least one of the fine dust forecast value, ultrafine dust forecast value, yellow dust forecast value, outdoor air quality data, indoor air quality data, air conditioning damper status, supply fan status, and exhaust fan status. The fine dust forecast value, ultrafine dust forecast value, yellow dust forecast value, or air quality warning grade may be included in wide-area disaster forecast data collected from a meteorological agency server (10) or a disaster information provision server (20). Outdoor air quality data and indoor air quality data may be included in field disaster detection data collected through a direct digital controller (210). The air conditioning damper status, supply fan status, and exhaust fan status may be included in disaster prevention equipment status data collected through a direct digital controller (210).

[0190] The data matching unit (120) can align fine dust forecast values, ultrafine dust forecast values, or yellow dust forecast values ​​with outdoor air quality data or indoor air quality data on a common time axis. For example, if the air quality forecast values ​​are provided in hourly or daily units and the outdoor air quality data or indoor air quality data is collected in minute units, the data matching unit (120) can convert the field air quality data into an average value, a maximum value, or a representative value corresponding to the forecast time interval. Additionally, the data matching unit (120) can generate an air quality comparison dataset by aligning the air conditioning damper status, the supply fan status, and the exhaust fan status to the same judgment time interval.

[0191] The disaster cross-judgment unit (150) determines forecast risk indicators and field risk indicators based on an air quality comparison dataset. The forecast risk indicator may be determined based on fine dust forecast values, ultrafine dust forecast values, yellow dust forecast values, ozone forecast values, or air quality warning grades. The field risk indicator may be determined based on outdoor air quality data, indoor air quality data, fine dust concentration, ultrafine dust concentration, carbon dioxide concentration, volatile organic compound concentration, or hazardous gas concentration. The judgment conditions for each disaster type may include at least one of fine dust concentration criteria, ultrafine dust concentration criteria, yellow dust concentration criteria, indoor air quality criteria, outdoor air quality criteria, or warning grade criteria.

[0192] The disaster cross-determination unit (150) determines the condition of the building as a local disaster state when both the forecast risk indicator and the on-site risk indicator satisfy the conditions for determining the air quality disaster type. In this case, the air quality disaster judgment unit (177) can select the main driving mode of the air conditioning damper, supply fan, or exhaust fan. For example, if the fine dust forecast value or yellow dust forecast value satisfies the standard condition and at the same time the outdoor air quality data or indoor air quality data is greater than or equal to the standard value, the air quality disaster judgment unit (177) can select the main driving mode of the air conditioning damper to restrict the inflow of outdoor air. Additionally, if the indoor air quality data satisfies the standard condition, the air quality disaster judgment unit (177) can select the main driving mode of the supply fan or exhaust fan.

[0193] When the air quality disaster judgment unit (177) selects the main driving mode of the air conditioning damper, it transmits the control output to the DDC data linkage unit (190). The DDC data linkage unit (190) converts the control output into output control data that can be processed by the direct digital controller (210). The air conditioning damper output control unit (277) of the direct digital controller (210) can control the opening of the outside air damper, exhaust damper, or circulation damper according to the output control data. For example, if the outside air quality deteriorates, the air conditioning damper output control unit (277) can reduce the opening of the outside air damper and increase the circulation air ratio.

[0194] The air quality disaster judgment unit (177) can select a supply fan or an exhaust fan as a control target. For example, if indoor air quality data deteriorates above a reference value, the air quality disaster judgment unit (177) can transmit a control output to the supply fan output control unit (279) or the exhaust fan output control unit (281). The supply fan output control unit (279) can control the amount of air supplied to the indoor space. The exhaust fan output control unit (281) can control the amount of exhaust to discharge indoor contaminated air. At this time, if the outdoor air quality is poor, exhaust or circulation control can be performed while limiting the amount of outdoor air introduced.

[0195] The disaster cross-judgment unit (150) can determine the condition of a building as a forecast-dominant discrepancy state if the forecast risk indicator satisfies the judgment condition for the air quality disaster type, but the on-site risk indicator does not satisfy the judgment condition. In this case, the air quality disaster judgment unit (177) can control the air conditioning damper, supply fan, or exhaust fan in an alarm mode or a backup drive mode. For example, if the fine dust forecast value or yellow dust forecast value satisfies the standard condition but the outdoor air quality data or indoor air quality data does not yet satisfy the danger condition, the air quality disaster judgment unit (177) can provide an air quality forecast-based alarm to the manager and select a backup control mode for the air conditioning damper.

[0196] The pre-operation mode may be a mode in which the air conditioning damper, supply fan, or exhaust fan is switched to a ready state prior to the main operation. For example, the pre-operation mode of the air conditioning damper may include partial closing of the outside air damper, partial opening of the circulation damper, or pre-limiting of the amount of outside air introduced. The pre-operation mode of the supply fan or exhaust fan may include checking the fan start-up status, low-speed operation, checking filter differential pressure, or pre-adjusting the ventilation amount. The air quality disaster judgment unit (177) may switch to the main operation mode when the field risk indicator satisfies the subsequent judgment conditions in a forecast-dominant discrepancy state.

[0197] The disaster cross-determination unit (150) can determine the condition of a building as a local occurrence inconsistency state if the forecast risk indicator does not satisfy the judgment condition for the air quality disaster type, but the on-site risk indicator satisfies the judgment condition. In this case, the air quality disaster judgment unit (177) can transmit a local air quality alarm to the manager terminal or monitoring interface unit (600) and select one of the alarm mode, pre-operation mode, or main operation mode of the air conditioning damper, supply fan, or exhaust fan. For example, even if there is no forecast of air quality deterioration, if the air quality sensor value of the outside air intake unit or indoor space rises above a reference value, the air quality disaster judgment unit (177) can generate a local air quality alarm and change the control mode of the air conditioning damper, supply fan, or exhaust fan.

[0198] The air quality disaster judgment unit (177) can select a different control mode depending on the degree of the on-site risk indicator in a localized inconsistency state. For example, if the indoor air quality data reaches the warning standard, an alarm mode or a preliminary drive mode can be selected, and if the indoor air quality data reaches the danger standard, a main drive mode can be selected. In addition, if only the outdoor air quality data deteriorates, the outdoor air damper can be controlled first, and if the indoor air quality data deteriorates together, the supply fan, exhaust fan, or air conditioning damper can be controlled together.

[0199] The air quality disaster judgment unit (177) may further consider disaster prevention equipment status data. For example, if the air conditioning damper is in a broken state or cannot be driven, the air quality disaster judgment unit (177) may restrict the main driving command of the air conditioning damper and send an administrator alarm first. In addition, if the supply fan or exhaust fan is in a non-operable state, the air quality disaster judgment unit (177) may select other air conditioning equipment or air purification equipment as alternative control targets. Conversely, if the air conditioning damper, supply fan, and exhaust fan are all in a normal state, the air quality disaster judgment unit (177) may control these equipment sequentially or in parallel according to the air quality disaster state.

[0200] The air quality disaster judgment unit (177) can maintain the air conditioning damper, supply fan, or exhaust fan in standby mode or normal operation mode when the judgment state is normal. The normal state may be a state in which the fine dust forecast value, ultrafine dust forecast value, yellow dust forecast value, outdoor air quality data, and indoor air quality data do not all satisfy the judgment conditions for the air quality disaster type. In this case, the air quality disaster judgment unit (177) can limit unnecessary air conditioning damper control or fan operation and maintain the existing operation state.

[0201] The air quality disaster judgment unit (177) may store the air quality response results in the history management database (500). The history management database (500) may store at least one of building identification information, judgment time interval, fine dust forecast value, ultrafine dust forecast value, yellow dust forecast value, outdoor air quality data, indoor air quality data, air conditioning damper status, supply fan status, exhaust fan status, judgment status, selected control mode, and control output. The stored information may subsequently be used for history analysis of the same disaster type or for identifying vulnerable buildings.

[0202] Additionally, the air quality disaster judgment unit (177) can provide air quality response results to the monitoring interface unit (600). The monitoring interface unit (600) can display fine dust forecast values, ultrafine dust forecast values, yellow dust forecast values, outdoor air quality data, indoor air quality data, and the control status of air conditioning dampers, supply fans, or exhaust fans. Accordingly, the manager can check forecast data, field data, and the control status of disaster prevention equipment for air quality disaster types together.

[0203] In this way, the air quality disaster judgment unit (177) according to the present embodiment can determine the air quality disaster state by considering the fine dust forecast value, the ultrafine dust forecast value, or the yellow dust forecast value together with the outdoor air quality data or the indoor air quality data, and can determine the control mode of the air conditioning damper, the supply fan, or the exhaust fan. Accordingly, the system according to the present embodiment can control the air conditioning-related disaster prevention equipment by reflecting the wide-area air quality forecast and the air quality status of the building site together.

[0204] FIG. 1 is an overall configuration diagram of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to one embodiment of the present invention, FIG. 9 is an overall flowchart of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control method according to one embodiment of the present invention, and FIG. 11 is a flowchart of a repetitive alarm-based vulnerable building identification according to one embodiment of the present invention.

[0205] According to FIGS. 1, 9 and 11, the artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to the present embodiment may include a history management database (500). The history management database (500) stores at least one of wide-area disaster forecasting data, on-site disaster detection data, disaster prevention equipment status data, the judgment result of the disaster cross-judgment unit (150), and the control mode determination result of the control execution unit (170). The history management database (500) may manage the stored history information by correlating it with at least one of building identification information, disaster type code, and judgment time interval.

[0206] The history management database (500) can store wide-area disaster forecast data collected through the external data interface unit (110). The wide-area disaster forecast data may include at least one of a disaster type code, forecast time, forecast target location, forecast value, and warning grade. For example, the history management database (500) can store a precipitation forecast value for a heavy rain or flood disaster type, a temperature forecast value for a heatwave disaster type, a snowfall forecast value or freezing forecast value for a heavy snow or freezing disaster type, and a fine dust forecast value or yellow dust forecast value for an air quality disaster type.

[0207] The history management database (500) can store field disaster detection data collected through the DDC data linkage unit (190). The field disaster detection data may include at least one of water level data, leakage data, rainfall data, indoor temperature data, outdoor temperature data, humidity data, air quality data, vibration data, load data, snow accumulation data, and freezing data. The history management database (500) can store such field disaster detection data along with sensor identification information, measurement time, measurement value, status value, alarm value, or building identification information.

[0208] The history management database (500) can store disaster prevention equipment status data. The disaster prevention equipment status data may include at least one of drainage pump status data, water barrier status data, air conditioner status data, air conditioner damper status data, supply fan status data, exhaust fan status data, heating wire status data, snow removal equipment status data, gas valve status data, district heating valve status data, fire shutter status data, or emergency shut-off relay status data. The history management database (500) can store the disaster prevention equipment status data along with equipment identification information, status collection time, operation status, failure status, alarm status, or manual operation status.

[0209] The history management database (500) can store a comparison dataset by disaster type generated by the data matching unit (120). The comparison dataset by disaster type may include wide-area disaster forecast data and on-site disaster detection data aligned on a common time axis. For example, the history management database (500) can store precipitation forecast values, on-site rainfall, drainage tank water level, and leakage signals for a single judgment time interval as a single heavy rain or flood comparison dataset. Additionally, the history management database (500) can store temperature forecast values, outdoor temperature, indoor temperature, and humidity as a single heatwave comparison dataset.

[0210] The history management database (500) can store the judgment results of the disaster cross-judgment unit (150). The judgment results may include at least one of forecast risk indicators, field risk indicators, whether judgment conditions by disaster type are satisfied, judgment status, and judgment basis data. The judgment status may be any one of a local disaster status, a forecast dominance discrepancy status, a local occurrence discrepancy status, or a normal status. The history management database (500) can store these judgment results by correlating them with building identification information, disaster type codes, and judgment time intervals.

[0211] The history management database (500) can store the control mode determination result of the control execution unit (170). The control mode determination result may include at least one of the controlled equipment, selected control mode, control output, control time, control duration, or control priority. The control mode may be any one of a standby mode, an alarm mode, a backup drive mode, and a main drive mode. For example, the history management database (500) can store the history of the drainage pump being controlled in the main drive mode or the history of the water barrier being controlled in the backup drive mode for a heavy rain or flood disaster type.

[0212] The history management database (500) can store history information by classifying it according to disaster type. For example, history information for a heavy rain or flood disaster type may include precipitation forecast values, on-site rainfall amount, drainage tank water level, leakage signal, flood signal, drainage pump status, water barrier status, and control mode determination results. History information for a heatwave disaster type may include temperature forecast values, outdoor temperature, indoor temperature, humidity, equipment temperature, air conditioner status, and fan control results. History information for a heavy snow or freezing disaster type may include snow accumulation forecast values, freezing forecast values, on-site snow accumulation amount, freezing detection value, load data, heating cable status, and snow removal equipment control results. History information for an air quality disaster type may include fine dust forecast values, yellow dust forecast values, outdoor air quality data, indoor air quality data, air conditioning damper status, and fan control results.

[0213] The history management database (500) can store history information by classifying it according to building identification information. The building identification information may include building name, building code, address information, location information, management zone information, or facility group identification information. When a single management server manages multiple buildings, the history management database (500) can store different forecast data, on-site detection data, judgment results, and control history for each building, even for the same type of disaster.

[0214] The history management database (500) can store history information by judgment time interval. The judgment time interval may be a time interval in which the data alignment unit (120) aligns wide-area disaster forecast data and on-site disaster detection data on a common time axis. The history management database (500) can store comparison datasets by disaster type, forecast risk indicators, on-site risk indicators, judgment status, and control mode determination results by judgment time interval. Accordingly, changes in disaster status and changes in disaster prevention equipment control for the same building can be managed in chronological order.

[0215] The history management database (500) can store the disaster state determination results and the control mode determination results in correspondence with each other. For example, it can store which disaster prevention equipment was controlled in the main operation mode during the time period determined to be a local disaster state. Additionally, it can store which disaster prevention equipment was controlled in the alarm mode or the backup operation mode during the time period determined to be a forecast dominance mismatch state. Additionally, it can store which local alarm occurred and which disaster prevention equipment was controlled during the time period determined to be a local occurrence mismatch state.

[0216] The history management database (500) can store disaster prevention equipment status data and control mode determination results together. For example, if the drainage pump status data indicates a failure state even though the control execution unit (170) has selected the main operation mode, the history management database (500) can store the main operation command and the drainage pump failure state together. Additionally, if the control execution unit (170) selects the preliminary operation mode and then switches to the main operation mode due to an increase in the field risk indicator, the history management database (500) can store the history of the switch from preliminary operation to main operation.

[0217] The history management database (500) can provide stored history information to the monitoring interface unit (600). The monitoring interface unit (600) can query and display wide-area disaster forecast data, field disaster detection data, judgment status, and control mode determination results stored in the history management database (500). For example, the monitoring interface unit (600) can display forecast risk indicators and field risk indicators for a specific judgment time interval on a graph of the same time axis, and display the judgment status and control mode of that time interval together.

[0218] The history management database (500) can be used to identify vulnerable buildings based on repeated warnings. For example, there may be cases where the risk warning of the same type of disaster is repeatedly included in the field disaster detection data or disaster prevention facility status data, even though the forecast risk indicator of the wide-area disaster forecast data is a low-risk weather condition that is below the initial disaster threshold for each type of disaster. In this case, the history management database (500) can provide history data to calculate the cumulative number of occurrences of risk warnings for the same building and the same type of disaster.

[0219] The control execution unit (170) can calculate the cumulative number of occurrences of risk alerts of the same disaster type by referring to the history information stored in the history management database (500). The control execution unit (170) can determine whether the calculated cumulative number of occurrences exceeds the statistical cumulative threshold frequency for each disaster type. If the cumulative number of occurrences exceeds the statistical cumulative threshold frequency for each disaster type, the control execution unit (170) can register a building corresponding to the building identification information as a vulnerable building for the same disaster type. The history management database (500) can store the result of the vulnerable building registration.

[0220] The history management database (500) can store the results of registering vulnerable buildings along with at least one of building identification information, disaster type code, cumulative occurrence count, statistical cumulative threshold frequency, time of registration, and basis data for registration. For example, if a drainage tank water level alarm or leakage signal occurs repeatedly in a specific building even though the forecast for heavy rain in the area is low, the history management database (500) can store the result of registering the building as a vulnerable building for the heavy rain or flood disaster type.

[0221] The history management database (500) may subsequently be referenced in determining the control mode of the control execution unit (170). For example, if a specific building is registered as a vulnerable building for a specific type of disaster, the control execution unit (170) may prioritize selecting an alarm mode or a backup operation mode in a state of forecast dominance discrepancy or local occurrence discrepancy for the same type of disaster. However, the method of utilizing the results of such vulnerable building registration may be omitted or changed as necessary.

[0222] In this way, the history management database (500) according to the present embodiment can store wide-area disaster forecast data, on-site disaster detection data, disaster prevention facility status data, judgment results, and control mode determination results by correlating them with building identification information, disaster type codes, and judgment time intervals. Additionally, the history management database (500) can provide the stored history information to the monitoring interface unit (600) or provide it as basic data for identifying vulnerable buildings based on repeated alarms.

[0223] FIG. 1 is an overall configuration diagram of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to one embodiment of the present invention, and FIG. 8 is an example screen diagram of a monitoring interface according to one embodiment of the present invention.

[0224] According to FIGS. 1 and 8, the artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to the present embodiment may include a monitoring interface unit (600). The monitoring interface unit (600) displays at least one of wide-area disaster forecasting data, on-site disaster detection data, the judgment result of the disaster cross-judgment unit (150), and the control mode determination result of the control execution unit (170). The monitoring interface unit (600) may display the wide-area disaster forecasting data and the on-site disaster detection data as a graph on the same time axis, and may display the judgment status by disaster type and the control mode of the disaster prevention equipment (300) together.

[0225] The monitoring interface unit (600) can display wide-area disaster forecast data collected through the external data interface unit (110). The wide-area disaster forecast data may include at least one of a disaster type code, forecast time, forecast target location, forecast value, and warning grade. For example, the monitoring interface unit (600) can display a precipitation forecast value for a heavy rain or flood disaster type, a temperature forecast value for a heatwave disaster type, a snowfall forecast value or freezing forecast value for a heavy snow or freezing disaster type, and a fine dust forecast value or yellow dust forecast value for an air quality disaster type.

[0226] The monitoring interface unit (600) can display field disaster detection data collected through the DDC data linkage unit (190). The field disaster detection data may include at least one of water level data, leakage data, rainfall data, indoor temperature data, outdoor temperature data, humidity data, air quality data, vibration data, load data, snow accumulation data, and freezing data. The monitoring interface unit (600) can display each field disaster detection data along with sensor identification information, measurement time, measurement value, status value, or alarm value.

[0227] The monitoring interface unit (600) can display wide-area disaster forecast data and on-site disaster detection data on a graph on the same time axis. For example, for heavy rain or flood disaster types, the precipitation forecast value and the on-site rainfall amount can be displayed together on a graph on the same time axis. In addition, the drainage tank water level, leakage signal, or flood signal can be displayed together as display elements corresponding to the same time interval. Accordingly, the manager can check changes in forecast data and changes in on-site detection data on the same time basis.

[0228] The monitoring interface unit (600) can display the temperature forecast value, outdoor temperature, indoor temperature, humidity, or facility temperature on a graph on the same time axis for the type of heatwave disaster. For example, the monitoring interface unit (600) can display the temperature forecast value and the indoor temperature on different curves or display lines, and display the heatwave judgment criterion value on the reference line. In addition, the air conditioner status, supply fan status, or exhaust fan status can be provided as status indicators corresponding to the time interval of the graph.

[0229] The monitoring interface unit (600) can display snowfall forecast values, freezing forecast values, on-site snowfall amount, freezing detection values, ambient temperature, or load data for heavy snowfall or freezing disaster types. For example, the monitoring interface unit (600) can display snowfall forecast values ​​and on-site snowfall amount on a graph on the same time axis, and display load data or freezing detection values ​​as status icons, alarm displays, or numerical data. In addition, the status of heating cables or snow removal equipment can be displayed together.

[0230] The monitoring interface unit (600) can display fine dust forecast values, ultrafine dust forecast values, yellow dust forecast values, outdoor air quality data, or indoor air quality data for air quality disaster types. For example, the monitoring interface unit (600) can display outdoor air quality data and indoor air quality data on a graph on the same time axis and display air quality judgment criteria values ​​as a baseline. In addition, the air conditioning damper status, supply fan status, or exhaust fan status can be displayed together.

[0231] The monitoring interface unit (600) can display the judgment result of the disaster cross-judgment unit (150). The judgment result may include at least one of forecast risk indicators, field risk indicators, whether judgment conditions by disaster type are satisfied, judgment status, and judgment basis data. The judgment status may be any one of a local disaster status, a forecast dominance discrepancy status, a local occurrence discrepancy status, or a normal status. The monitoring interface unit (600) can display these judgment statuses by classifying them according to disaster type.

[0232] For example, if the judgment status of a heavy rain or flood disaster type is a local disaster state, the monitoring interface unit (600) can display the local disaster state in the status display area for the corresponding disaster type. In addition, it can display the precipitation forecast value used as the basis for judgment, the field rainfall amount, the drainage tank water level, the leakage signal, or the flood signal together. If there is a forecast dominance mismatch state, the monitoring interface unit (600) can display information that the forecast risk indicator satisfies the judgment condition but the field risk indicator does not satisfy the judgment condition. If there is a local occurrence mismatch state, the monitoring interface unit (600) can display information that the forecast risk indicator does not satisfy the judgment condition but the field risk indicator satisfies the judgment condition.

[0233] The monitoring interface unit (600) can display the control mode determination result of the control execution unit (170). The control mode determination result may include at least one of the controlled equipment, the selected control mode, the control output, the control time, the control duration, or the control priority. The control mode may be any one of a standby mode, an alarm mode, a preliminary operation mode, and a main operation mode. The monitoring interface unit (600) can display the judgment status by disaster type and the selected control mode in correspondence.

[0234] For example, if the control execution unit (170) selects the main driving mode of the drainage pump for a heavy rain or flood disaster type, the monitoring interface unit (600) may display the control status of the drainage pump as the main driving mode. Additionally, if the control execution unit (170) selects the preliminary driving mode of the heating cable for a heavy snow or freezing disaster type, the monitoring interface unit (600) may display the control status of the heating cable as the preliminary driving mode. If the control execution unit (170) changes the control mode of the air conditioning damper for an air quality disaster type, the monitoring interface unit (600) may display the opening state of the air conditioning damper or a control command.

[0235] The monitoring interface unit (600) can query and display history information stored in the history management database (500). The history information may include at least one of wide-area disaster forecast data, on-site disaster detection data, disaster prevention facility status data, the judgment result of the disaster cross-judgment unit (150), and the control mode determination result of the control execution unit (170). The monitoring interface unit (600) can display history information of a specific building, a specific disaster type, or a specific judgment time interval according to the manager's query conditions.

[0236] The monitoring interface unit (600) may divide the screen area into multiple display areas. For example, the monitoring interface unit (600) may include a building selection area, a disaster type selection area, a time interval selection area, a co-time axis graph area, a judgment status display area, a disaster prevention equipment status display area, and a control mode display area. The building selection area may be an area for selecting a building to be monitored among multiple buildings. The disaster type selection area may be an area for selecting any one of the following disaster types: heavy rain or flooding, heatwave, heavy snow or freezing, or deterioration of air quality.

[0237] The same time axis graph area may be an area that displays wide-area disaster forecast data and on-site disaster detection data based on the same time. The judgment status display area may be an area that displays the judgment result of the disaster cross-judgment unit (150). The disaster prevention equipment status display area may be an area that displays the status of at least one of a drainage pump, water barrier, air conditioner, air conditioner damper, supply fan, exhaust fan, heating wire, snow removal equipment, valve, fire shutter, or emergency cutoff relay. The control mode display area may be an area that displays a control mode selected by the control execution unit (170).

[0238] The monitoring interface unit (600) can display a judgment reference value or a threshold value on a graph. For example, in the case of a heavy rain or flood disaster type, a precipitation reference value or a drainage tank water level reference value can be displayed as a reference line. In the case of a heatwave disaster type, an indoor temperature reference value or an outdoor temperature reference value can be displayed as a reference line. In the case of a heavy snow or freezing disaster type, a snow accumulation reference value, a load reference value, or a freezing reference value can be displayed as a reference line. In the case of an air quality disaster type, a fine dust concentration reference value or an indoor air quality reference value can be displayed as a reference line.

[0239] The monitoring interface unit (600) can display the time of alarm occurrence or the time of control mode switching on the same time axis graph. For example, the time when a forecast dominant discrepancy state occurs, the time when a local discrepancy state occurs, the time when the disaster prevention equipment (300) is switched to a preliminary operation mode or a main operation mode can be displayed on the graph. Accordingly, the manager can check the temporal relationship between forecast data, field data, judgment status, and control mode switching.

[0240] The monitoring interface unit (600) can provide an alarm message to the administrator. The alarm message may include at least one of a disaster type, a judgment status, the time of alarm occurrence, judgment basis data, a disaster prevention facility to be controlled, and a selected control mode. For example, if a local occurrence discrepancy status is determined, the monitoring interface unit (600) can provide an alarm message stating that the field detection value satisfies a danger condition, even though it is not indicated in the forecast. If a forecast-dominant discrepancy status is determined, the monitoring interface unit (600) can provide an alarm message stating that there is a danger in the external forecast, but the field detection value does not yet satisfy a danger condition.

[0241] The monitoring interface unit (600) can receive confirmation input or action input from an administrator. For example, the administrator can input whether to confirm a specific alarm, whether to perform manual control, whether to request an inspection, or a request to change the control mode through the monitoring interface unit (600). However, the administrator's input function may be omitted as necessary, and the monitoring interface unit (600) may perform only a display function.

[0242] The monitoring interface unit (600) can display the results of the registration of vulnerable buildings. For example, if a specific building is registered in the history management database (500) as a vulnerable building for a specific type of disaster, the monitoring interface unit (600) can display the vulnerability status of the building by disaster type, the cumulative number of occurrences, the statistical cumulative threshold frequency, and the basis for registration data. Accordingly, the manager can verify information on vulnerable buildings identified by repeated alarms.

[0243] The monitoring interface unit (600) can display the results of the operation threshold correction based on the site safety margin. For example, if the control execution unit (170) delays the entry of the main operation mode of the disaster prevention facility (300) based on the site safety margin and corrects the operation threshold to a dynamic operation threshold, the monitoring interface unit (600) can display the initial disaster limit point, the dynamic operation threshold, the site detection value, and the main operation delay status together. However, this display function may be omitted as necessary.

[0244] Additionally, the control execution unit (170) may not maintain the list of manually approved targets for high-risk control targets as a fixed list, but may dynamically update it based on at least one of the disaster type, judgment status, disaster prevention equipment status data, past control history, frequency of danger alarm occurrence, and change trends of on-site disaster detection data. For example, the control execution unit (170) may set at least one of a water barrier, a main gas supply valve, a parking lot ramp heating wire, a fire shutter, and an emergency cutoff relay as a basic high-risk control target, and if a repeated alarm occurs in a specific building or a failure history of the disaster prevention equipment (300) accumulates, the disaster prevention equipment (300) may be added to the list of manually approved targets.

[0245] The control execution unit (170) may use an artificial intelligence analysis model to update the list of manual approval targets. The artificial intelligence analysis model may use at least one of the wide-area disaster forecast data, field disaster detection data, disaster prevention facility status data, judgment status, control mode determination result, and manager approval or rejection history stored in the history management database (500) as input. The artificial intelligence analysis model may calculate the risk of malfunction, risk of over-control, risk of delayed control, or the need for manager verification for a specific disaster prevention facility (300). If the calculated risk of malfunction, risk of over-control, risk of delayed control, or the need for manager verification is greater than or equal to a reference value, the control execution unit (170) may add the corresponding disaster prevention facility (300) to the list of manual approval targets, or adjust the approval waiting time, approval requirement conditions, or preliminary alarm conditions.

[0246] The monitoring interface unit (600) may provide a settings screen for editing the list of manual approval targets. Through the monitoring interface unit (600), the administrator may set at least one of the following: whether manual approval is applied per disaster prevention facility, the approval waiting time, the type of disaster requiring approval, the status of determining approval requirement, an alternative control action upon rejection, and an exception handling action upon non-response. The control execution unit (170) saves the edited result entered through the settings screen as a manual approval control policy, and subsequently, the manual approval control policy may be referenced when determining the control mode for high-risk control targets.

[0247] Additionally, the control execution unit (170) may temporarily change the list of manual approval targets when specific situation conditions occur. Specific situation conditions may include at least one of the following: when the disaster prevention equipment status data includes a failure state, a communication abnormality state, an inability to operate state, or a danger alarm; when the on-site disaster detection data shows a rapid increasing trend; when a danger alarm of the same disaster type is repeated more than a preset number of times; when control of the corresponding disaster type is required in a building registered as a vulnerable building; or when equipment requiring on-site verification by a manager is selected as a control target. In this case, the control execution unit (170) may temporarily switch even a disaster prevention equipment (300) originally classified as an autonomous control target to a manual approval target. Additionally, the control execution unit (170) may switch even a disaster prevention equipment (300) originally classified as a manual approval target to an automatic emergency control target if emergency shutdown conditions for protecting lives, protecting equipment, or preventing secondary damage are satisfied.

[0248] In this way, the monitoring interface unit (600) according to the present embodiment can display wide-area disaster forecast data and on-site disaster detection data as a graph on the same time axis, and can display the judgment result of the disaster cross-judgment unit (150) and the control mode determination result of the control execution unit (170) together. In addition, the monitoring interface unit (600) can set or display a list of manual approval targets and manual approval control policies for high-risk control targets. Accordingly, the manager can check forecast data by disaster type, on-site data, judgment status, control status of disaster prevention equipment (300), and manual approval conditions for high-risk control targets on a single screen.

[0249] FIG. 9 is an overall flowchart of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control method according to one embodiment of the present invention.

[0250] According to FIG. 9, a computer-readable recording medium according to the present embodiment may have a program recorded thereon that enables a computer to perform an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control method. The program may include a command for an external data interface unit (110) to collect wide-area disaster forecasting data by disaster type from at least one of a national or regional meteorological agency, a regional meteorological station, and a disaster information providing server; a command for a DDC data linkage unit (190) to collect on-site disaster detection data and disaster prevention facility status data through a direct digital controller (210) installed in a building; a command for a data alignment unit (120) to align wide-area disaster forecasting data and on-site disaster detection data on a common time axis to generate a comparison dataset by disaster type; a command for a disaster cross-judgment unit (150) to determine whether forecast risk indicators and on-site risk indicators satisfy judgment conditions by disaster type based on the comparison dataset; and a command for a control execution unit (170) to select a control mode of a disaster prevention facility (300) according to the judgment result.

[0251] The program may include a command to cause the external data interface unit (110) to collect wide-area disaster forecast data. When the command is executed, the external data interface unit (110) may collect wide-area disaster forecast data by disaster type from a weather agency server (10) or a disaster information providing server (20). The wide-area disaster forecast data may include at least one of precipitation forecast values, temperature forecast values, snowfall forecast values, freezing forecast values, fine dust forecast values, yellow dust forecast values, and disaster notification or warning grades. The program may further include a command to convert the collected wide-area disaster forecast data into an internal processing format that includes at least one of a disaster type code, forecast time, forecast target location, forecast value, and warning grade.

[0252] The program may include a command for the DDC data linkage unit (190) to collect field disaster detection data and disaster prevention equipment status data through the direct digital controller (210). When the command is executed, the DDC data linkage unit (190) may collect at least one of water level data, leakage data, rainfall data, indoor temperature data, outdoor temperature data, humidity data, air quality data, vibration data, load data, snow accumulation data, or freezing data from the direct digital controller (210). Additionally, the DDC data linkage unit (190) may collect at least one of drainage pump status data, water barrier status data, air conditioner status data, air conditioner damper status data, heating cable status data, snow removal equipment status data, or valve status data.

[0253] The program may include a command that causes the data alignment unit (120) to align the wide-area disaster forecast data and the field disaster detection data to a common time axis. When the command is executed, the data alignment unit (120) may align both data to the same judgment time interval based on the forecast time or forecast application time interval of the wide-area disaster forecast data and the measurement time or collection time of the field disaster detection data. For example, if the collection cycles of the wide-area disaster forecast data and the field disaster detection data are different, the program may calculate an average value, a maximum value, a minimum value, a cumulative value, or a representative value to convert both data into a form that can be compared.

[0254] The program may include a command that causes the data matching unit (120) to match the building location with the forecast target location. When the command is executed, the data matching unit (120) can compare the building identification information, address information, latitude and longitude information, management area information, or point code with the administrative district information, grid coordinate information, observation station identification information, or area code information of the wide-area disaster forecast data. The data matching unit (120) can select the wide-area disaster forecast data corresponding to the building using this location information.

[0255] The program may include a command that causes the data matching unit (120) to match a disaster type with a sensor type. When the command is executed, the data matching unit (120) may match the disaster type code of the wide-area disaster forecast data with the sensor type of the field disaster detection data. For example, the program may match a precipitation forecast value, field rainfall amount, drainage tank water level, leakage signal, or flood signal to a heavy rain or flood disaster type. Additionally, the program may match a temperature forecast value, outdoor temperature, indoor temperature, or humidity to a heatwave disaster type; match a snow depth forecast value, freezing forecast value, field snow depth amount, freezing detection value, or load data to a heavy snow or freezing disaster type; and match a fine dust forecast value, yellow dust forecast value, outdoor air quality data, or indoor air quality data to an air quality disaster type.

[0256] The program may include a command that causes the data matching unit (120) to generate a comparison dataset by disaster type. When the command is executed, the data matching unit (120) can generate a comparison dataset by grouping wide-area disaster forecast data and on-site disaster detection data aligned on a common time axis by disaster type. The comparison dataset may include at least one of building identification information, disaster type code, judgment time interval, forecast value, alarm grade, on-site detection value, sensor status value, or correction value.

[0257] The program may include a command that causes the disaster cross-determination unit (150) to determine forecast risk indicators and field risk indicators. When the command is executed, the disaster cross-determination unit (150) may calculate or determine forecast risk indicators from wide-area disaster forecast data included in a comparison dataset, and may calculate or determine field risk indicators from field disaster detection data included in a comparison dataset. Forecast risk indicators may be calculated based on forecast values, alarm grades, or disaster notifications. Field risk indicators may be calculated based on sensor measurements, status values, or alarm values.

[0258] The program may include a command that causes the disaster cross-judgment unit (150) to determine whether the forecast risk indicator and the field risk indicator satisfy the judgment conditions for each disaster type. When the command is executed, the disaster cross-judgment unit (150) may determine whether the forecast risk indicator satisfies the forecast judgment conditions for each disaster type and whether the field risk indicator satisfies the field judgment conditions for each disaster type. The judgment conditions for each disaster type may include at least one of precipitation criteria, water level criteria, temperature criteria, humidity criteria, snow accumulation criteria, freezing criteria, load criteria, air quality criteria, or warning grade criteria.

[0259] The program may include a command that causes the disaster cross-judgment unit (150) to classify the disaster state according to a combination of judgment results. When such command is executed, the disaster cross-judgment unit (150) may determine a local disaster state if both the forecast risk indicator and the field risk indicator satisfy the judgment conditions for each disaster type. The disaster cross-judgment unit (150) may determine a forecast dominant inconsistency state if the forecast risk indicator satisfies the judgment conditions for each disaster type but the field risk indicator does not satisfy the judgment conditions for each disaster type. The disaster cross-judgment unit (150) may determine a local occurrence inconsistency state if the forecast risk indicator does not satisfy the judgment conditions for each disaster type but the field risk indicator satisfies the judgment conditions for each disaster type. The disaster cross-judgment unit (150) may determine a normal state if both the forecast risk indicator and the field risk indicator do not satisfy the judgment conditions for each disaster type.

[0260] The program may include a command that causes the control execution unit (170) to select a control mode of the disaster prevention facility (300) according to the judgment result of the disaster cross-judgment unit (150). When the command is executed, the control execution unit (170) may select one of the standby mode, alarm mode, pre-operation mode, and main operation mode of the disaster prevention facility (300) according to any one of the local disaster state, forecast dominance mismatch state, and local occurrence mismatch state. For example, the main operation mode may be selected in the local disaster state, the alarm mode or pre-operation mode may be selected in the forecast dominance mismatch state, and the alarm mode, pre-operation mode, or main operation mode may be selected in the local occurrence mismatch state.

[0261] The program may include a command that causes the control execution unit (170) to select a disaster prevention facility (300) corresponding to a disaster type. When the command is executed, the control execution unit (170) may select a drainage pump or a water barrier as a control target for a heavy rain or flood disaster type. The control execution unit (170) may select an air conditioner, an air conditioning damper, a supply fan, or an exhaust fan as a control target for a heatwave disaster type. The control execution unit (170) may select a heating cable or a snow removal facility as a control target for a heavy snow or freezing disaster type. The control execution unit (170) may select an air conditioning damper, a supply fan, or an exhaust fan as a control target for an air quality disaster type.

[0262] The program may include a command that causes the DDC data linkage unit (190) to convert a control mode selected by the control execution unit (170) into output control data that can be processed by the direct digital controller (210). When the command is executed, the DDC data linkage unit (190) can generate output control data including control target equipment identification information, control mode information, driving intensity, driving time, output value, or control command code. The direct digital controller (210) can change the operating state of the disaster prevention equipment (300) according to the output control data.

[0263] The program may further include a command that causes the history management database (500) to store the judgment result and the control mode determination result. When the command is executed, the history management database (500) may store at least one of the wide-area disaster forecast data, on-site disaster detection data, disaster prevention facility status data, the judgment result of the disaster cross-judgment unit (150), and the control mode determination result of the control execution unit (170) by correlating them with building identification information, disaster type code, and judgment time interval.

[0264] The program may further include a command that causes the monitoring interface unit (600) to display wide-area disaster forecast data and on-site disaster detection data as a graph on the same time axis. When such a command is executed, the monitoring interface unit (600) can display forecast data by disaster type, on-site data, judgment status, and the control mode of the disaster prevention equipment (300) on a single screen. Additionally, the monitoring interface unit (600) can display the alarm occurrence time, the control mode switching time, or the judgment criterion value together on the graph.

[0265] A computer-readable recording medium may be any medium capable of storing a program. For example, a computer-readable recording medium may include at least one of a hard disk, a solid-state drive, flash memory, ROM, RAM, a memory card, an optical disk, a magnetic disk, or a storage medium accessible via a network. The program may be executed on at least one of a server, a control device, a building management device, an edge computing device, or a cloud computing device.

[0266] In this way, the program recorded on the computer-readable recording medium according to the present embodiment can enable the computer to perform the collection of wide-area disaster forecast data, the collection of on-site disaster detection data, alignment of a common time axis, generation of a comparison dataset by disaster type, classification of disaster status, and selection of a control mode for disaster prevention equipment (300). Accordingly, the artificial intelligence-based disaster forecast and on-site detection data-linked building automatic control method according to the present embodiment can be implemented by the execution of the program.

[0267] FIG. 1 is an overall configuration diagram of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to one embodiment of the present invention, and FIG. 10 is a flowchart of on-site safety margin-based operating threshold correction according to one embodiment of the present invention.

[0268] According to FIGS. 1 and 10, the control execution unit (170) according to the present embodiment delays the entry of the disaster prevention facility (300) into the main operation mode when at least one of the forecast risk indicator of the wide-area disaster forecast data and the field risk indicator of the field disaster detection data reaches the initial disaster threshold for each disaster type, provided that the disaster prevention facility status data is in a normal state and the field safety margin is confirmed from the field disaster detection data, and corrects the operation threshold of the disaster prevention facility (300) to a dynamic operation threshold higher than the initial disaster threshold based on the field safety margin. The control execution unit (170) then switches the disaster prevention facility (300) to the main operation mode when the field disaster detection data reaches the dynamic operation threshold or when the disaster prevention facility status data includes a danger alarm.

[0269] The initial disaster threshold may be a pre-set reference value for each disaster type. For example, in the case of heavy rain or flood disaster types, the initial disaster threshold may be a reference value corresponding to the precipitation forecast value, on-site rainfall amount, drainage tank water level, leakage signal, or flood signal. In the case of heatwave disaster types, the initial disaster threshold may be a reference value corresponding to the temperature forecast value, indoor temperature, outdoor temperature, humidity, or equipment temperature. In the case of heavy snow or freezing disaster types, the initial disaster threshold may be a reference value corresponding to the snow depth forecast value, on-site snow depth amount, freezing detection value, load data, or outdoor temperature. In the case of air quality disaster types, the initial disaster threshold may be a reference value corresponding to the fine dust forecast value, yellow dust forecast value, outdoor air quality data, or indoor air quality data.

[0270] The control execution unit (170) first determines whether at least one of the forecast risk indicator and the field risk indicator has reached the initial disaster threshold for each disaster type. The forecast risk indicator may be a risk indicator calculated from wide-area disaster forecast data. The field risk indicator may be a risk indicator calculated from field disaster detection data. If the forecast risk indicator and the field risk indicator have not both reached the initial disaster threshold, the control execution unit (170) may maintain the existing control mode.

[0271] The control execution unit (170) determines whether the disaster prevention facility status data collected from the DDC data linkage unit (190) is in a normal state when at least one of the forecast risk indicator and the field risk indicator reaches an initial disaster threshold. The disaster prevention facility status data may include at least one of drainage pump status data, water barrier status data, air conditioner status data, air conditioner damper status data, supply fan status data, exhaust fan status data, heating cable status data, snow removal facility status data, valve status data, fire shutter status data, or emergency cutoff relay status data.

[0272] The fact that the disaster prevention equipment status data is in a normal state may mean that the disaster prevention equipment (300) to be controlled is not in a broken state, inoperable state, communication impossible state, overload state, alarm state, or manual shutdown state. For example, if the drainage pump is in a normal standby state, the water barrier is in an operable state, and the air conditioner or heating wire is in an operable state, the control execution unit (170) may determine that the disaster prevention equipment status data is in a normal state. Conversely, if the disaster prevention equipment status data includes a broken state or a danger alarm, the control execution unit (170) may switch the disaster prevention equipment (300) to the main operation mode or generate an administrator alarm without delaying entry into the main operation mode.

[0273] The control execution unit (170) determines whether a field safety margin is confirmed from the field disaster detection data when the disaster prevention facility status data is in a normal state. The field safety margin may refer to a margin of safety before the current field detection value reaches the actual danger zone. For example, in the case of a heavy rain or flood disaster type, the field safety margin may include a remaining water level margin before the drainage tank water level reaches the flood danger level, a margin for the operation of the drainage pump, or a drainage processing capacity. In the case of a heatwave disaster type, the field safety margin may include a temperature margin before the indoor temperature or facility temperature reaches the danger temperature. In the case of a heavy snow or freezing disaster type, the field safety margin may include a snow accumulation or load margin before the field snow accumulation or load data reaches the danger standard. In the case of an air quality disaster type, the field safety margin may include a concentration margin before the indoor air quality data reaches the danger standard.

[0274] The control execution unit (170) may switch the disaster prevention equipment (300) to the main operation mode if the site safety margin is not confirmed. For example, if the site detection value has reached the initial disaster threshold and there is no margin or the margin to the actual risk standard is very small, the control execution unit (170) does not delay entry into the main operation mode. In addition, even if the data capable of determining the site safety margin is missing or has low reliability, the control execution unit (170) may select the main operation mode considering safety aspects.

[0275] The control execution unit (170) delays the entry of the disaster prevention facility (300) into the main operation mode when the disaster prevention facility status data is in a normal state and the site safety margin is confirmed. Delaying the entry into the main operation mode may mean that the disaster prevention facility (300) is not immediately switched to the main operation state, but is maintained in any one of a standby mode, an alarm mode, or a backup operation mode. For example, if the heavy rain forecast has reached the initial disaster threshold but the water level of the drainage tank is sufficiently low and the drainage pump is in a normal state, the control execution unit (170) may delay the entry of the drainage pump into the main operation mode.

[0276] The control execution unit (170) corrects the operating threshold of the disaster prevention facility (300) to a dynamic operating threshold that is higher than the initial disaster threshold based on the site safety margin. The dynamic operating threshold may be a corrected threshold calculated by reflecting at least one of the site safety margin, the status of the disaster prevention facility, the type of disaster, the location of the building, the facility processing capacity, or past control history. The dynamic operating threshold may be set to a value higher than the initial disaster threshold. Here, a higher value may mean that the criterion for the disaster prevention facility (300) to switch to the main operating mode is delayed compared to the initial criterion.

[0277] For example, the control execution unit (170) can calculate a dynamic operation threshold according to the following mathematical formula 1.

[0278] [Mathematical Formula 1]

[0279] Th_t = Th_base + ΔTh × δ_s

[0280] Here, Th_t may represent the dynamic operating threshold at the current time of judgment. Th_base may represent a preset base threshold for each disaster type. ΔTh may represent the threshold correction range calculated based on the site safety margin. δ_s may represent the local safety factor of the building.

[0281] The local safety factor δ_s can be calculated by considering the location of the building, drainage capacity, whether the disaster prevention facility is in a normal state, the reliability of the field sensor, past alarm history, whether the building is registered as vulnerable, and the risk level by disaster type. For example, if the disaster prevention facility (300) is in a normal state, the field safety margin is sufficient, and the reliability of the field sensor is high, the control execution unit (170) can set the local safety factor δ_s relatively large. Conversely, if there are abnormal signs in the disaster prevention facility (300), the reliability of the field sensor is low, or the building is registered as a vulnerable building for the same disaster type, the control execution unit (170) can set the local safety factor δ_s relatively small.

[0282] The threshold correction margin ΔTh can be calculated based on the difference between the current field detection value and the actual risk threshold. For example, in the case of heavy rain or flood disaster types, ΔTh can be calculated based on the remaining water level margin between the current drainage tank water level and the flood risk water level. In the case of heatwave disaster types, ΔTh can be calculated based on the temperature margin between the current indoor temperature or facility temperature and the risk temperature threshold. In the case of heavy snow or freezing disaster types, ΔTh can be calculated based on the remaining margin between the current snow accumulation or load data and the risk threshold. In the case of air quality disaster types, ΔTh can be calculated based on the concentration margin between the current indoor air quality data and the risk concentration threshold.

[0283] The control execution unit (170) can determine whether to switch the disaster prevention facility (300) to the main operation mode based on the dynamic operation threshold Th_t calculated by mathematical formula 1. That is, if the field disaster detection data does not reach the dynamic operation threshold Th_t, the control execution unit (170) can maintain a delay in entering the main operation mode of the disaster prevention facility (300). Conversely, if the field disaster detection data reaches the dynamic operation threshold Th_t, the control execution unit (170) can switch the disaster prevention facility (300) to the main operation mode.

[0284] For example, in the case of a heavy rain or flood disaster type, if the initial disaster threshold is set based on a specific drainage tank water level, the control execution unit (170) can correct the criteria for entering the main drive mode to a dynamic operation threshold higher than the initial disaster threshold if there is sufficient margin between the current drainage tank water level and the flood risk water level and the drainage pump is in a normal state. In this case, the drainage pump may not switch to the main drive mode until the field water level reaches the dynamic operation threshold.

[0285] In the case of a heatwave disaster type where the initial disaster threshold is set based on indoor temperature, the control execution unit (170) can adjust the criteria for entering the main driving mode of the air conditioner to a dynamic operation threshold higher than the initial disaster threshold if the rate of increase in indoor temperature is low, the air conditioner condition is normal, and there is room in the equipment temperature. In the case of a heavy snow or freezing disaster type, the control execution unit (170) can set the dynamic operation threshold of the heating cable or snow removal equipment if it confirms that the structural load margin or the heating cable condition is normal, even if the field snow accumulation or load data approaches the initial disaster threshold. In the case of an air quality disaster type, the control execution unit (170) can also set the dynamic operation threshold of the air conditioning-related equipment if there is room in the indoor air quality data and the air conditioning damper or fan condition is normal.

[0286] The control execution unit (170) continues to collect field disaster detection data and disaster prevention facility status data even after the dynamic operation threshold is set. The control execution unit (170) determines whether the field disaster detection data reaches the dynamic operation threshold. For example, if the drainage tank water level, indoor temperature, field snowfall amount, load data, or air quality data reaches the dynamic operation threshold, the control execution unit (170) can switch the disaster prevention facility (300) to the main operation mode without further delaying entry into the main operation mode.

[0287] Additionally, the control execution unit (170) determines whether the disaster prevention equipment status data includes a danger alarm. The danger alarm may include at least one of a failure alarm, an overload alarm, an inability to operate alarm, a communication abnormality alarm, an equipment temperature abnormality alarm, or a status abnormality alarm of the disaster prevention equipment (300). If the disaster prevention equipment status data includes a danger alarm, the control execution unit (170) may switch the disaster prevention equipment (300) to the main operation mode or provide an alarm to the manager, even if the field disaster detection data has not reached the dynamic operation threshold.

[0288] The control execution unit (170) may maintain a delay in entering the main operation mode of the disaster prevention equipment (300) when the field disaster detection data does not reach the dynamic operation threshold and the disaster prevention equipment status data does not include a danger alarm. In this case, the control execution unit (170) may repeatedly check the field disaster detection data and the disaster prevention equipment status data while maintaining a standby mode, an alarm mode, or a backup operation mode. Even while maintaining the delay state, the control execution unit (170) may display the delay state, the initial disaster threshold, the dynamic operation threshold, and the current field detection value on the monitoring interface unit (600).

[0289] The control execution unit (170) may store the correction result of the dynamic operation threshold in the history management database (500). The history management database (500) may store at least one of building identification information, disaster type code, initial disaster threshold, site safety margin, dynamic operation threshold, whether the main operation is deferred, deferral start time, deferral end time, site disaster detection data, disaster prevention equipment status data, and whether the main operation mode is switched. The stored information may subsequently be used to reset the threshold correction criteria for each disaster type or to analyze the control history of the same building.

[0290] The control execution unit (170) can set dynamic operating thresholds differently for each type of disaster. For example, in the case of a heavy rain or flood disaster, drainage capacity and the rate of water level rise may be considered. In the case of a heatwave disaster, the rate of temperature rise, indoor heat load, and the operational status of the air conditioner may be considered. In the case of a heavy snow or freezing disaster, the rate of snow accumulation increase, load margin, and the condition of the heating cable may be considered. In the case of an air quality disaster, the difference between the outdoor air quality and the indoor air quality, the ventilation status, and the condition of the air conditioner damper may be considered.

[0291] When calculating the dynamic operation threshold, the control execution unit (170) may consider not only the field safety margin but also the trend of change in the forecast risk indicator. For example, even if the forecast risk indicator has reached an initial disaster threshold, if the forecast value subsequently shows a decreasing trend, the control execution unit (170) may delay entry into the main operation mode and set the dynamic operation threshold. Conversely, if the forecast risk indicator shows a rapidly increasing trend, the control execution unit (170) may not set the dynamic operation threshold or may set a small correction range.

[0292] Additionally, the control execution unit (170) may optionally use a time series prediction model to correct the field safety margin, dynamic operation threshold, or base threshold. The time series prediction model may include at least one of a long short-term memory neural network or a gated recurrent unit. The long short-term memory neural network may be a Long Short-Term Memory (LSTM) model, and the gated recurrent unit may be a Gated Recurrent Unit (GRU) model. However, the time series prediction model is not limited thereto and may include any lightweight artificial intelligence model capable of predicting disaster risk or facility risk using wide-area disaster forecast data, field disaster detection data, and disaster prevention facility status data accumulated in chronological order.

[0293] The time series prediction model can use time series data from a past period stored in a history management database (500) as input. For example, the time series prediction model can use at least one of the wide-area disaster forecast data, on-site disaster detection data, disaster prevention facility status data, judgment status, control mode determination result, and risk alarm occurrence history from the past 12 months as an input sequence. The input sequence can be defined as shown in Equation 2 below.

[0294] [Mathematical Formula 2]

[0295] X_t = [F_t, S_t, E_t, C_t, H_t]

[0296] Here, X_t may represent a multidimensional time series input vector at time t. F_t may represent wide-area disaster forecast data or forecast risk indicators at time t. S_t may represent on-site disaster detection data or on-site risk indicators at time t. E_t may represent disaster prevention facility status data at time t. C_t may represent the control mode determination result at time t. H_t may represent risk alarm history information prior to time t or at time t.

[0297] A time series prediction model can calculate a disaster-inducing density prediction value based on an input vector X_t. The disaster-inducing density prediction value may be a value indicating the degree of probability that a disaster warning, equipment malfunction, or a state requiring full operation will occur in the building for a specific type of disaster. For example, the disaster-inducing density prediction value can be calculated as shown in Equation 3 below.

[0298] [Mathematical Formula 3]

[0299] _t = f_θ(X_t)

[0300] Here, _t may represent the predicted value of the disaster-inducing density at time t. f_θ may represent a time series prediction model with a learning parameter θ. If the time series prediction model is an LSTM model, f_θ may be a function including the input gate, forget gate, output gate, and memory state of the LSTM cell. If the time series prediction model is a GRU model, f_θ may be a function including the update gate, reset gate, and hidden state.

[0301] The control execution unit (170) is the disaster-inducing density prediction value The site safety margin or dynamic operation threshold can be corrected using _t. For example, the disaster-inducing density prediction value If _t is greater than or equal to a reference value, the control execution unit (170) may reduce the correction range of the dynamic operation threshold or limit the delay in entering this operation mode. Conversely, the disaster-inducing density prediction value When _t is less than the reference value, the disaster prevention equipment status data is in a normal state, and the field safety margin is confirmed, the control execution unit (170) can maintain or increase the correction range of the dynamic operation threshold.

[0302] A time series forecasting model can be trained to reduce the error between the actual disaster-inducing density value and the predicted disaster-inducing density value. For example, the loss function of a time series forecasting model can be defined as the mean squared error as shown in Equation 4 below.

[0303] [Mathematical Formula 4]

[0304] L = 1 / M Σ_{i=1}^{M} (y_i - _i)^2

[0305] Here, L can represent the loss function. M can represent the number of data samples used for training. y_i can represent the actual disaster-inducing density value for the i-th training sample. _i may represent the predicted value of the disaster-inducing density of the time series prediction model for the i-th training sample. The actual disaster-inducing density value y_i may be calculated based on at least one of whether an actual danger alert occurs, whether the current driving mode is switched, whether an abnormality occurs in the disaster prevention equipment, or whether the field disaster detection data exceeds a danger threshold.

[0306] The control execution unit (170) can update the learning parameter θ of the time series prediction model so that the loss function L is reduced. For example, the control execution unit (170) can periodically retrain or fine-tune the time series prediction model based on past data stored in the history management database (500). At this time, the training of the time series prediction model can be performed by building, by disaster type, or by disaster prevention facility. Accordingly, the control execution unit (170) can reflect the on-site disaster detection characteristics or the operating characteristics of the disaster prevention facility that appear differently by building even under the same disaster forecast conditions.

[0307] Additionally, the control execution unit (170) can calculate a long-term performance decline coefficient based on the long-term operation history of the disaster prevention facility (300) and correct the basic threshold value for each disaster type according to the calculated long-term performance decline coefficient. The long-term performance decline coefficient can be calculated based on at least one of the following: the usage period of the disaster prevention facility (300), the frequency of repeated alarm occurrences, the failure history, the response delay to the control command, the abnormal pattern of the sensor measurement value, whether there is a normal response after switching to the main driving mode, and whether the building is registered as vulnerable.

[0308] For example, regarding the same type of heavy rain or flood disaster, if there is a delay in starting the drainage pump, repeated occurrence of drainage tank water level alarms, or accumulation of drainage pump failure alarms, the control execution unit (170) may determine that the long-term performance of the drainage pump is degraded. Additionally, regarding the same type of heavy snow or freezing disaster, if there is a delay in powering the heating cable, repeated occurrence of freezing alarms, or accumulation of response delays of the snow removal equipment, the control execution unit (170) may determine that the long-term performance of the heating cable or snow removal equipment is degraded. Additionally, regarding the same type of air quality disaster, if there is a delay in driving the air conditioning damper or repeated abnormal alarms of the supply fan or exhaust fan, the control execution unit (170) may determine that the long-term performance of the air conditioning-related equipment is degraded.

[0309] The control execution unit (170) can correct the basic threshold value for each disaster type based on the long-term performance decline coefficient. For example, the control execution unit (170) can calculate the corrected basic threshold value according to the following mathematical formula 5.

[0310] [Mathematical Formula 5]

[0311] Th_base_new = Th_base_old × (1 - γ × s_a)

[0312] Here, Th_base_new may refer to the corrected base threshold. Th_base_old may refer to the base threshold before correction. γ may refer to the correction sensitivity coefficient. s_a may refer to the long-term performance decay coefficient.

[0313] The long-term performance degradation coefficient s_a can be normalized to a value between 0 and 1. A value closer to 0 may indicate that the long-term performance degradation of the disaster prevention facility (300) is small. A value closer to 1 may indicate that the long-term performance degradation of the disaster prevention facility (300) is large. The correction sensitivity coefficient γ may be a coefficient for controlling the extent to which the long-term performance degradation coefficient s_a is reflected in the basic threshold correction. γ may be set differently depending on the type of disaster, the type of disaster prevention facility (300), the use of the building, or the manager's settings.

[0314] The control execution unit (170) can adjust the basic threshold to be lowered as the long-term performance degradation coefficient s_a increases. That is, if the long-term performance degradation of the disaster prevention facility (300) is judged to be significant, the control execution unit (170) can apply a corrected basic threshold lower than the existing basic threshold so that the alarm mode, pre-operation mode, or main operation mode is selected at an earlier time. Accordingly, even if the response performance of the disaster prevention facility (300) is degraded due to long-term operation, the control execution unit (170) can conservatively adjust the control judgment criteria for each disaster type.

[0315] For example, if the long-term performance decay coefficient s_a of the drainage pump increases in the case of a heavy rain or flood disaster type, the control execution unit (170) may lower the basic threshold for the drainage tank water level. In this case, an alarm mode or a pre-operation mode may be selected even before the drainage tank water level reaches the existing threshold. If the long-term performance decay coefficient s_a of the heating cable or snow removal equipment increases in the case of a heavy snow or freezing disaster type, the control execution unit (170) may lower the basic threshold for the snow accumulation amount, freezing detection value, or load data. If the long-term performance decay coefficient s_a of the air conditioning damper, supply fan, or exhaust fan increases in the case of an air quality disaster type, the control execution unit (170) may lower the basic threshold for indoor air quality or outdoor air quality.

[0316] The control execution unit (170) can incorporate the corrected base threshold Th_base_new into the calculation of the dynamic operation threshold. For example, the control execution unit (170) can calculate the dynamic operation threshold Th_t using the corrected base threshold Th_base_new instead of the base threshold Th_base_old before correction. In this case, the dynamic operation threshold can be a value that reflects not only the current on-site safety margin of the disaster prevention equipment (300) but also the long-term performance degradation state.

[0317] The control execution unit (170) can store the long-term performance decline coefficient s_a and the basic threshold correction result in the history management database (500). The history management database (500) can store at least one of building identification information, disaster type code, disaster prevention equipment identification information, long-term performance decline coefficient, basic threshold before correction, corrected basic threshold, correction sensitivity coefficient, correction time, and correction basis data. The correction basis data may include the number of repeated alarm occurrences, failure history, response delay time, sensor abnormality pattern, or vulnerable building registration information.

[0318] The control execution unit (170) can provide the results of the basic threshold correction to the monitoring interface unit (600). The monitoring interface unit (600) can display the basic threshold before correction, the corrected basic threshold, the long-term performance decline coefficient, and the basis for correction data. Accordingly, the manager can verify the reason why the judgment criteria for each disaster type have been conservatively adjusted for a specific building or specific disaster prevention facility (300).

[0319] However, the basic threshold correction based on the long-term performance decline coefficient may be applied optionally. The control execution unit (170) may not perform the basic threshold correction if history data is not sufficiently accumulated, if performance degradation of the disaster prevention facility (300) is not confirmed, or according to the manager's settings. Additionally, the control execution unit (170) may limit the lower limit of the corrected basic threshold so that a threshold lower than the statutory safety standard or management standard is not set.

[0320] Additionally, the control execution unit (170) may use the output value of the time series prediction model for long-term correction of the basic threshold. For example, if a risk warning of the same type of disaster occurs repeatedly in a specific building, or if the response performance of the disaster prevention equipment (300) deteriorates over time, the control execution unit (170) may calculate a long-term performance decay coefficient and correct the basic threshold. The correction of the basic threshold can be performed as shown in Equation 5 below.

[0321] [Mathematical Formula 5]

[0322] Th_base_new = Th_base_old × (1 - γ × s_a)

[0323] Here, Th_base_new may mean a corrected base threshold. Th_base_old may mean a base threshold before correction. γ may mean a correction sensitivity coefficient. s_a may mean a long-term performance decay coefficient. The long-term performance decay coefficient s_a may be calculated based on at least one of the elapsed usage period of the fire protection equipment (300), failure history, frequency of repeated alarm occurrences, response delay to control commands, abnormal pattern of sensor measurements, or whether the building is registered as vulnerable.

[0324] The control execution unit (170) can adjust the basic threshold to be lowered as the long-term performance decay coefficient s_a increases. For example, if a repetitive failure alarm or response delay occurs in a drainage pump of a specific building, the control execution unit (170) can lower the basic threshold for the heavy rain or flood disaster type so that the alarm mode, pre-operation mode, or main operation mode is selected at an earlier time. Additionally, if the response performance of the heating cable or snow removal equipment of a specific building deteriorates, the control execution unit (170) can lower the basic threshold for the heavy snow or freezing disaster type.

[0325] However, a time series prediction model, mean squared error-based learning, and a basic threshold correction based on a long-term performance decay coefficient may be applied optionally. The control execution unit (170) may calculate a dynamic operating threshold using only pre-set rule-based logic without using a time series prediction model, depending on the size of the building, the degree of accumulation of historical data, the type of disaster prevention equipment (300), the type of disaster, or the manager settings.

[0326] The control execution unit (170) can switch to the main operation mode according to the administrator's input even when a dynamic operation threshold has been set. For example, if the administrator approves the main operation switch or inputs a deferral release through the monitoring interface unit (600), the control execution unit (170) can switch the fire prevention equipment (300) to the main operation mode regardless of whether the dynamic operation threshold has been reached. However, this administrator input function may be omitted as necessary.

[0327] The control execution unit (170) can readjust the dynamic operation threshold if the site safety margin decreases after setting the dynamic operation threshold. For example, if the rate of increase in the drainage tank water level increases or the rate of increase in the indoor temperature increases, the control execution unit (170) can lower the dynamic operation threshold to a value closer to the initial disaster limit. Additionally, if the disaster prevention equipment status data changes from a normal state to a caution state, the control execution unit (170) can release the delay in entering the main operation mode.

[0328] In this embodiment, dynamic operating threshold correction based on the site safety margin is not necessarily applied to all types of disasters. The control execution unit (170) may apply dynamic operating threshold correction only to specific types of disasters according to the type of disaster, building characteristics, type of disaster prevention equipment (300), manager settings, or legal standards. For example, dynamic operating threshold correction may be omitted in disaster types or equipment abnormal conditions where immediate risk to human life is expected, and the disaster prevention equipment (300) may switch to this operating mode immediately upon reaching the initial disaster limit.

[0329] In this way, the control execution unit (170) according to the present embodiment may delay the entry of the disaster prevention facility (300) into the main operation mode and correct the operation threshold to a dynamic operation threshold when the disaster prevention facility status data is in a normal state and the field safety margin is confirmed, even if the forecast risk indicator or the field risk indicator reaches the initial disaster threshold. Additionally, the control execution unit (170) may switch the disaster prevention facility (300) to the main operation mode when the field disaster detection data reaches the dynamic operation threshold or when the disaster prevention facility status data includes a danger alarm.

[0330] FIG. 1 is an overall configuration diagram of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system according to one embodiment of the present invention, FIG. 9 is an overall flowchart of an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control method according to one embodiment of the present invention, and FIG. 11 is a flowchart of a repetitive alarm-based vulnerable building identification according to one embodiment of the present invention.

[0331] According to FIGS. 1, 9 and 11, the control execution unit (170) according to the present embodiment calculates the cumulative number of occurrences of the same disaster type of risk warning stored in the history management database (500) when the risk indicator of the wide-area disaster forecast data is below the initial disaster threshold for each disaster type under low-risk weather conditions, and the same disaster type of risk warning is repeatedly included in the field disaster detection data or disaster prevention facility status data. When the cumulative number of occurrences exceeds the statistical cumulative threshold frequency for each disaster type, the building corresponding to the building identification information is registered as a vulnerable building for the same disaster type.

[0332] The history management database (500) can store at least one of wide-area disaster forecast data, on-site disaster detection data, disaster prevention facility status data, the judgment result of the disaster cross-judgment unit (150), and the control mode determination result of the control execution unit (170). The history management database (500) can store these data in correspondence with building identification information, disaster type codes, and judgment time intervals. The building identification information may include building name, building code, address information, location information, management zone information, or facility group identification information. The disaster type code may be information for distinguishing disaster types such as heavy rain or flooding, heatwave, heavy snow or freezing, or deterioration of air quality.

[0333] The control execution unit (170) first determines whether the forecast risk indicator of the wide-area disaster forecast data is a low-risk weather condition that is below the initial disaster threshold for each disaster type. The forecast risk indicator may be calculated based on wide-area disaster forecast data collected from a weather agency server (10) or a disaster information provision server (20). The initial disaster threshold for each disaster type may be a pre-set forecast reference value for each disaster type. For example, in the case of a heavy rain or flood disaster type, a precipitation forecast value or a heavy rain warning grade may be used as the initial disaster threshold. In the case of a heatwave disaster type, a temperature forecast value or a heatwave warning grade may be used as the initial disaster threshold. In the case of a heavy snow or freezing disaster type, a snow depth forecast value, a freezing forecast value, or a cold wave warning grade may be used as the initial disaster threshold. In the case of an air quality disaster type, a fine dust forecast value, an ultrafine dust forecast value, or a yellow dust forecast value may be used as the initial disaster threshold.

[0334] Low-risk weather conditions may be conditions where the risk of the corresponding disaster type is judged to be low based solely on wide-area disaster forecast data. For example, if the precipitation forecast value is lower than the heavy rain judgment criterion, or if a heat wave warning has not been issued, or if the snowfall forecast value is lower than the heavy snowfall judgment criterion, or if the fine dust forecast value is lower than the air quality deterioration criterion, the control execution unit (170) may determine that the corresponding disaster type is a low-risk weather condition. If the control execution unit (170) is not a low-risk weather condition, it may terminate the repeated warning-based vulnerable building identification procedure.

[0335] The control execution unit (170) determines whether the same type of disaster risk alarm is included in the field disaster detection data or disaster prevention facility status data when the weather conditions are low risk. The field disaster detection data may include at least one of water level data, leakage data, rainfall data, indoor temperature data, outdoor temperature data, humidity data, air quality data, vibration data, load data, snow accumulation data, or freezing data collected through the direct digital controller (210). The disaster prevention facility status data may include at least one of drainage pump status data, water barrier status data, air conditioner status data, air conditioner damper status data, heating cable status data, snow removal facility status data, valve status data, fire shutter status data, or emergency cutoff relay status data.

[0336] A danger alert may be information generated when on-site disaster detection data or disaster prevention facility status data satisfies on-site judgment conditions for each disaster type, or when the facility status data includes an abnormal condition. For example, in the case of heavy rain or flood disaster types, danger alerts may include drainage tank water level alerts, leakage alerts, flood alerts, drainage pump failure alerts, or water cutoff curtain operation abnormality alerts. In the case of heatwave disaster types, danger alerts may include indoor high temperature alerts, facility overheating alerts, air conditioner abnormality alerts, or fan operation abnormality alerts. In the case of heavy snow or freezing disaster types, danger alerts may include on-site snow accumulation alerts, freezing alerts, overload alerts, heating cable failure alerts, or snow removal facility abnormality alerts. In the case of air quality disaster types, danger alerts may include indoor air quality deterioration alerts, outdoor air quality deterioration alerts, air conditioning damper abnormality alerts, or fan abnormality alerts.

[0337] The control execution unit (170) may terminate the vulnerable building identification procedure if the field disaster detection data or disaster prevention facility status data does not contain a risk warning of the same disaster type. Conversely, if a risk warning of the same disaster type is included even under low-risk weather conditions, the control execution unit (170) calculates the cumulative number of occurrences of risk warnings for the same building and the same disaster type by referring to the history information stored in the history management database (500).

[0338] The cumulative occurrence count may refer to the number of times a risk warning of the same type of disaster occurs in the same building during a predetermined statistical calculation period. The statistical calculation period may be on a daily, weekly, monthly, seasonal, or period set by the manager. For example, the control execution unit (170) may calculate the number of times a drainage tank water level warning of the heavy rain or flood disaster type occurred in a specific building during the last 30 days, even though low-risk weather conditions were present. Additionally, the control execution unit (170) may calculate the number of times an indoor air quality warning of the air quality disaster type occurred in a specific building during the last 1 year, even though low-risk weather conditions were present.

[0339] When calculating the cumulative occurrence count, the control execution unit (170) can filter history information based on building identification information and disaster type codes. For example, if multiple buildings are managed by a single management system, the control execution unit (170) can select only the history information corresponding to the building identification information of a specific building. Additionally, the control execution unit (170) can calculate the cumulative occurrence count by selecting only the risk alarms corresponding to the same disaster type code. Accordingly, it is possible to prevent alarms of different buildings or different disaster types from being mixed into a single cumulative occurrence count.

[0340] The control execution unit (170) determines whether the cumulative number of occurrences exceeds the statistical cumulative threshold frequency for each disaster type. The statistical cumulative threshold frequency may be a reference value for determining whether the repeated alarm of a specific disaster type in a specific building has exceeded the normal range. The statistical cumulative threshold frequency may be set differently for each disaster type. For example, the statistical cumulative threshold frequency for a heavy rain or flood disaster type may be set based on the number of occurrences of a drainage tank water level alarm, a leakage alarm, or a flood alarm during a certain period. The statistical cumulative threshold frequency for a heatwave disaster type may be set based on the number of repeated occurrences of an indoor high temperature alarm or an equipment overheating alarm. The statistical cumulative threshold frequency for a heavy snow or freezing disaster type may be set based on the number of repeated occurrences of a freezing alarm, an overload alarm, or a heating wire abnormality alarm. The statistical cumulative threshold frequency for an air quality disaster type may be set based on the number of repeated occurrences of an indoor air quality alarm or an air conditioning damper abnormality alarm.

[0341] The statistical cumulative threshold frequency may be set based on at least one of the average number of alarms in all buildings, the average number of alarms in buildings of the same use, the average number of alarms in buildings of the same area, the history of past alarm occurrences, or a standard value set by the manager. For example, if the average number of alarms in general buildings during the same period for a specific type of disaster is low, but the alarm occurs repeatedly only in a specific building, the control execution unit (170) may determine that the building is likely vulnerable to the type of disaster.

[0342] The control execution unit (170) may allow the history management database (500) to maintain the cumulative storage of risk alerts of the same disaster type if the cumulative number of occurrences does not exceed the statistical cumulative threshold frequency for each disaster type. In this case, the control execution unit (170) may not register the building as a vulnerable building and may continue to check whether additional risk alerts of the same disaster type occur thereafter. The history management database (500) may continue to accumulate and store additional risk alerts.

[0343] The control execution unit (170) registers a building corresponding to the building identification information as a vulnerable building for the same disaster type when the cumulative number of occurrences exceeds the statistical cumulative threshold frequency for each disaster type. A vulnerable building may refer to a building in which on-site warnings or facility warnings for a specific disaster type occur repeatedly, even though it is in a low-risk state according to wide-area disaster forecast data. For example, a building in which drainage tank water level warnings occur repeatedly even under conditions where precipitation forecast values ​​are low may be registered as a vulnerable building for heavy rain or flood disaster types. Additionally, a building in which indoor air quality warnings occur repeatedly even under conditions where air quality forecasts are good may be registered as a vulnerable building for air quality disaster types.

[0344] The history management database (500) stores the results of registering vulnerable buildings. The results of registering vulnerable buildings may include at least one of building identification information, disaster type code, cumulative occurrence count, statistical cumulative threshold frequency, time of registration, registration basis data, and recent risk warning history. The registration basis data may include the time of occurrence, warning type, on-site detection value, disaster prevention equipment status value, or judgment status of a risk warning that occurred under low-risk weather conditions.

[0345] For example, the history management database (500) may include a vulnerable building history table for identifying vulnerable buildings based on repeated alarms. The vulnerable building history table may include building identification information, registration date, type of recognized disaster event, cumulative number of event occurrences, and a vulnerable building flag. Building identification information may be stored in the BLDG_ID field. The registration date may be stored in the REG_DATE field. The type of recognized disaster event may be stored in the COGN_EVENT_TYPE field. The cumulative number of event occurrences may be stored in the ACCUM_EVT_CNT field. The vulnerable building flag may be stored in the VULN_FLAG field.

[0346] The BLDG_ID field is a value used to identify buildings for which risk alerts have occurred repeatedly. The BLDG_ID field may correspond to at least one of the following: building code, building name, address code, management zone code, or facility group identification information. The REG_DATE field may indicate the time information when the vulnerable building registration result was generated or updated. The COGN_EVENT_TYPE field may indicate a disaster type, such as heavy rain or flooding, heatwave, heavy snow or freezing, or air quality deterioration. The ACCUM_EVT_CNT field may indicate the cumulative number of risk alerts of the same disaster type occurring under low-risk weather conditions. The VULN_FLAG field may indicate whether the building has been registered as a vulnerable building because the cumulative number of occurrences exceeds the statistical cumulative threshold frequency for each disaster type.

[0347] The history management database (500) may further store statistical cumulative threshold frequency, time of the most recent event, reason for vulnerability registration, and control policy identification information in the vulnerability building history table. The statistical cumulative threshold frequency may be stored in the THRESH_FREQ field. The time of the most recent event may be stored in the LAST_EVT_TIME field. The reason for vulnerability registration may be stored in the VULN_REASON field. The control policy identification information may be stored in the CTRL_POLICY_ID field. The THRESH_FREQ field may be used as a reference value to be compared with ACCUM_EVT_CNT. The LAST_EVT_TIME field may indicate the time of the most recent risk alarm of the same disaster type. The VULN_REASON field may indicate the repeating alarm type, on-site detection value, or disaster prevention equipment status value that serves as the basis for the vulnerability building registration. The CTRL_POLICY_ID field may be a value for identifying a control policy or manual approval control policy that can be applied based on the vulnerability building registration result.

[0348] The control execution unit (170) can calculate the number of cumulative event occurrences for the same building and the same disaster type by referring to the vulnerable building history table of the history management database (500). For example, the control execution unit (170) can search for history records where BLDG_ID and COGN_EVENT_TYPE are the same and update the number of risk alarms that occurred under low-risk weather conditions to ACCUM_EVT_CNT. If the updated ACCUM_EVT_CNT exceeds THRESH_FREQ, the control execution unit (170) can set VULN_FLAG to a vulnerable state and store REG_DATE and VULN_REASON together.

[0349] The control execution unit (170) may refer to the VULN_FLAG stored in the vulnerable building history table for subsequent control mode determination or updating of the list of manual approval targets. For example, if a building corresponding to a specific BLDG_ID is registered as a vulnerable building for a heavy rain or flood disaster type, the control execution unit (170) may prioritize selecting an alarm mode or a backup operation mode in a state of forecast dominance discrepancy or local occurrence discrepancy for the same disaster type. Additionally, if control of a high-risk control target is required in a building where the VULN_FLAG is in a vulnerable state, the control execution unit (170) may add the corresponding disaster prevention facility to the list of manual approval targets or strengthen the approval requirements.

[0350] However, the above field names are examples for explaining the internal schema of the history management database (500). The history management database (500) may use different field names, different table structures, or different data storage formats within the scope of performing the same function.

[0351] The control execution unit (170) can provide the results of the vulnerable building registration to the monitoring interface unit (600). The monitoring interface unit (600) can indicate that a specific building has been registered as a vulnerable building for a specific type of disaster. Additionally, the monitoring interface unit (600) can display the cumulative number of occurrences, the statistical cumulative threshold frequency, the registration basis data, and the recent alarm history together. Accordingly, the manager can verify which buildings are vulnerable to which types of disasters.

[0352] The control execution unit (170) may refer to the results of the vulnerable building registration for determining the control mode thereafter. For example, if a specific building is registered as a vulnerable building for a heavy rain or flood disaster type, the control execution unit (170) may prioritize selecting an alarm mode or a backup operation mode in a forecast dominance discrepancy or local occurrence discrepancy state for the same disaster type. Additionally, if a specific building is registered as a vulnerable building for an air quality disaster type, the control execution unit (170) may provide a pre-warning to the monitoring interface unit (600) when the air quality-related field detection value approaches the caution standard.

[0353] However, the method of utilizing the results of the vulnerable building registration is not limited to this. The control execution unit (170) may use the results of the vulnerable building registration for at least one of the following: requesting an inspection of the disaster prevention equipment (300), notifying an administrator, adjusting the priority of the control mode, correcting judgment conditions by disaster type, or generating a history analysis report. Additionally, the results of the vulnerable building registration may not be used for automatic control by the control execution unit (170) as needed, but may be used only as reference information displayed to the administrator.

[0354] The control execution unit (170) can periodically update the results of the vulnerable building registration. For example, if a risk warning of the same type of disaster is no longer repeated for a certain period after a specific building is registered as a vulnerable building, the control execution unit (170) may release the vulnerable building registration status or change it to a pending status. Conversely, if the risk warning continues to be repeated, the control execution unit (170) may raise the vulnerability level or provide a request for inspection to the manager. This method of updating may vary depending on the manager's settings, criteria by disaster type, or statistical information stored in the history management database (500).

[0355] In this embodiment, the identification of vulnerable buildings based on repeated alarms is not necessarily applied to all types of disasters. The control execution unit (170) may perform the vulnerable building identification procedure only for specific types of disasters depending on the type of disaster, building characteristics, sensor installation status, type of disaster prevention equipment (300), or manager settings. For example, the identification of vulnerable buildings based on repeated alarms may be performed only for heavy rain or flood disaster types and air quality disaster types, and the procedure may be omitted for other types of disasters.

[0356] Additionally, when the control execution unit (170) calculates the cumulative number of occurrences of danger alarms, it may reflect the duration of the alarm, the alarm intensity, the alarm occurrence zone, or the facility condition as weights. For example, a higher weight may be assigned to an alarm that lasts for a long time rather than an alarm that is released after a short period. Additionally, a higher weight may be assigned to an alarm that occurs simultaneously from multiple sensors or multiple zones rather than an alarm that occurs from a single sensor. In this case, the control execution unit (170) may determine whether the statistical cumulative threshold frequency is exceeded by using not only the simple number of occurrences but also the weighted cumulative score.

[0357] In this way, the control execution unit (170) according to the present embodiment can identify a building in which a risk warning of the same type of disaster occurs repeatedly even when there are low-risk weather conditions in the wide-area disaster forecast data. In addition, the control execution unit (170) can register a building corresponding to the building identification information as a vulnerable building for the same type of disaster by comparing the cumulative number of occurrences of risk warnings stored in the history management database (500) with the statistical cumulative threshold frequency for each type of disaster.

[0358] Meanwhile, the embodiments disclosed in this specification and drawings are merely specific examples provided to aid understanding and are not intended to limit the scope of the invention. It is obvious to those skilled in the art that other variations based on the technical concept of the invention are possible in addition to the embodiments disclosed herein. Explanation of the symbols

[0359] 10: Weather agency server 20: Disaster Information Provision Server 110: External Data Interface Section 120: Data alignment section 121: External data receiver 123: Field data receiver 125: Time alignment section 127: Location Matching Section 129: Sensor type matching section 131: Disaster Type Matching Section 133: Comparison Dataset Generation Section 150: Disaster Cross-Judgment Unit 151: Forecast Risk Indicator Calculation Unit 153: On-site Risk Indicator Calculation Unit 155: Judgment state classification unit 170: Control execution unit 171: Heavy Rain or Flood Disaster Judgment Division 173: Heatwave Disaster Assessment Department 175: Heavy Snow or Freezing Disaster Assessment Unit 177: Air Quality Disaster Assessment Division 190: DDC Data Interconnection Unit 210: Direct Digital Controller 220: Sensor input section 221: Water level sensor input 223: Leakage sensor input 225: Rainfall sensor input 227: Indoor temperature sensor input 229: Outdoor temperature sensor input 231: Humidity sensor input 233: Air quality sensor input 235: Vibration sensor input 237: Load sensor input section 239: Snow sensor input 241: Freezing sensor input 250: Equipment Status Input Section 251: Drainage pump status input section 253: Water barrier status input section 255: Air Conditioner Status Input 257: Air Conditioning Damper Status Input 259: Heating wire status input section 261: Snow removal equipment status input section 263: Valve Status Input 270: Output control unit 271: Drainage pump output control unit 273: Water barrier output control unit 275: Air Conditioner Output Control Unit 277: Air Conditioning Damper Output Control Unit 279: Supply fan output control unit 281: Exhaust fan output control unit 283: Heating wire output control unit 285: Snow removal equipment output control unit 287: Gas valve output control unit 289: District heating valve output control unit 291: Fire shutter output control unit 293: Emergency cutoff relay output control unit 300: Disaster prevention facilities 400: Field sensor 500: History Management Database 600: Monitoring Interface Section

Claims

Claim 1 In an AI-based building automatic control system linked with disaster forecasting and on-site detection data, the system comprises: an external data interface unit that collects wide-area disaster forecasting data by disaster type from at least one of a national or regional meteorological agency, a regional weather station, and a disaster information provision server; a DDC data linkage unit that collects on-site disaster detection data and disaster prevention facility status data of the building through a direct digital controller installed in the building and transmits control outputs to the disaster prevention facility; a data alignment unit that aligns the wide-area disaster forecasting data and the on-site disaster detection data on a common time axis to generate a comparison dataset by disaster type; a disaster cross-judgment unit that determines whether the forecast risk indicator of the wide-area disaster forecasting data and the on-site risk indicator of the on-site disaster detection data satisfy judgment conditions by disaster type based on the comparison dataset; a history management database that stores the wide-area disaster forecasting data, the on-site disaster detection data, the disaster prevention facility status data, and the judgment result of the disaster cross-judgment unit together with building identification information and a disaster type code; and a monitoring system that displays at least one of the history information stored in the history management database, the judgment result of the disaster cross-judgment unit, the result of registering vulnerable buildings, and the result of determining the control mode of the disaster prevention facility. Interface section;The control execution unit determines the control mode of the disaster prevention facility based on the judgment result of the disaster cross-judgment unit and the history information stored in the history management database. The control execution unit calculates the cumulative number of occurrences of a risk warning based on the history of the risk warning of the same disaster type stored in the history management database when, under low-risk weather conditions where the forecast risk indicator of the wide-area disaster forecast data is below the initial disaster threshold for each disaster type, the field disaster detection data or the disaster prevention facility status data contains a risk warning of the same disaster type under the low-risk weather conditions where the field disaster detection data or the disaster prevention facility status data is included. When the cumulative number of occurrences exceeds the statistical cumulative threshold frequency for each disaster type, the building corresponding to the building identification information is registered as a vulnerable building for the same disaster type. For the building registered as a vulnerable building, when the forecast risk indicator or the field disaster detection data reaches a preventive active threshold value set to a level lower than the initial disaster threshold, a control output corresponding to a preset low-risk or medium-risk equipment group among the disaster prevention facilities is transmitted to the DDC data linkage unit to automatically induce and control preventive activation. The result of the registration of the vulnerable building and the result of the control of the preventive activation are provided to the monitoring interface unit. An AI-based disaster forecasting and on-site detection data-linked building automatic control system characterized by the statistical cumulative threshold frequency being set based on statistical analysis or AI analysis of risk warning history by disaster type stored in the history management database. Claim 2 An AI-based disaster forecasting and field detection data-linked building automatic control system according to claim 1, wherein the disaster cross-determination unit determines a local disaster state when both the forecast risk indicator and the field risk indicator satisfy the judgment conditions for each disaster type, determines a forecast dominance discrepancy state when the forecast risk indicator satisfies the judgment conditions for each disaster type but the field risk indicator does not satisfy the judgment conditions for each disaster type, and determines a local occurrence discrepancy state when the forecast risk indicator does not satisfy the judgment conditions for each disaster type but the field risk indicator satisfies the judgment conditions for each disaster type. Claim 3 In paragraph 2, the control execution unit calculates a disaster-inducing density prediction value using a time-series prediction model that takes as input sequence at least one of wide-area disaster forecast data, field disaster detection data, disaster prevention facility status data, judgment status, control mode determination result, and risk alarm occurrence history for a past period stored in the history management database, corrects at least one of a field safety margin, a dynamic operation threshold, or a basic threshold by disaster type based on the disaster-inducing density prediction value, and the time-series prediction model is trained to reduce the mean squared error between the actual disaster-inducing density value and the disaster-inducing density prediction value, characterized by an artificial intelligence-based disaster forecast and field detection data-linked building automatic control system. Claim 4 In paragraph 2, the control execution unit defers entry into the main operation mode of the disaster prevention facility when at least one of the forecast risk indicator of the wide-area disaster forecast data and the field risk indicator of the field disaster detection data reaches a basic threshold value by disaster type, the disaster prevention facility status data is in a normal state, and a field safety margin is confirmed from the field disaster detection data, and calculates a dynamic operation threshold value according to the following mathematical formula: Th_t = Th_base + ΔTh × δ_s, where Th_t is the dynamic operation threshold value at the current judgment point, Th_base is the basic threshold value by disaster type, ΔTh is the threshold value correction range according to the field safety margin, and δ_s is the local safety factor of the building, and switches the disaster prevention facility to the main operation mode when the field disaster detection data reaches the dynamic operation threshold value or when a danger alarm is included in the disaster prevention facility status data. This characterizes an artificial intelligence-based disaster forecast and field detection data-linked building automatic control system. Claim 5 In paragraph 3, the control execution unit calculates a long-term performance decline coefficient based on at least one of the usage period of the disaster prevention facility, the frequency of repeated alarm occurrences, the failure history, the response delay to a control command, and whether the building is registered as vulnerable, and corrects a basic threshold value for each disaster type according to the following mathematical formula based on the long-term performance decline coefficient, Th_base_new = Th_base_old × (1 - γ × s_a), where Th_base_new is the corrected basic threshold value, Th_base_old is the basic threshold value before correction, γ is the correction sensitivity coefficient, and s_a is the long-term performance decline coefficient, characterized by an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system. Claim 6 In paragraph 2, the control execution unit updates a list of manual approval targets for high-risk control targets based on at least one of a disaster type, judgment status, disaster prevention facility status data, past control history, frequency of risk alarm occurrence, and change trends of on-site disaster detection data, and the monitoring interface unit provides a setting screen for editing at least one of the list of manual approval targets, approval waiting time, approval requirements, and exception handling actions in case of non-response, characterized by an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system. Claim 7 In claim 6, the control execution unit outputs administrator approval request information through the monitoring interface unit when the main driving mode for the high-risk control target is selected, transmits a control output corresponding to the main driving mode to the DDC data linkage unit when an approval input is received within a preset approval waiting time, and cancels the control output corresponding to the main driving mode and generates a preliminary alarm output when an approval input is not received within the preset approval waiting time or a rejection input is received. This characterizes an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control system. Claim 8 In claim 1, the DDC data linkage unit includes a communication gateway that performs serial communication or fieldbus communication with the direct digital controller and performs Ethernet-based communication with the control execution unit, and the communication gateway converts sensor input data and equipment status data received from the direct digital controller into packet data that the control execution unit can process, and converts control commands received from the control execution unit into output control data corresponding to the direct digital controller, characterized by an artificial intelligence-based disaster forecasting and field detection data linkage type building automatic control system. Claim 9 An AI-based disaster forecasting and on-site detection data-linked building automatic control system according to claim 1, wherein the data matching unit resamples the wide-area disaster forecasting data and the on-site disaster detection data at equal time intervals, matches the forecast location of the wide-area disaster forecasting data with the location information of the building, and matches the disaster type of the wide-area disaster forecasting data with the sensor type of the on-site disaster detection data to generate the comparison dataset. Claim 10 In an AI-based building automatic control method linked with disaster forecasting and on-site detection data, the method comprises: a step in which an external data interface unit collects wide-area disaster forecasting data by disaster type from at least one of a national or regional meteorological agency, a regional weather station, and a disaster information providing server; a step in which a DDC data linkage unit collects on-site disaster detection data and disaster prevention facility status data of the building through a direct digital controller installed in the building and transmits a control output for the disaster prevention facility; a step in which a data alignment unit aligns the wide-area disaster forecasting data and the on-site disaster detection data on a common time axis to generate a comparison dataset by disaster type; a step in which a disaster cross-judgment unit determines whether the forecast risk indicator of the wide-area disaster forecasting data and the on-site risk indicator of the on-site disaster detection data satisfy the judgment conditions by disaster type based on the comparison dataset; a step in which a history management database stores the wide-area disaster forecasting data, the on-site disaster detection data, the disaster prevention facility status data, and the judgment result of the disaster cross-judgment unit together with building identification information and a disaster type code; and a control execution unit determines the control mode of the disaster prevention facility based on the judgment result of the disaster cross-judgment unit and the history information stored in the history management database. Step; the control execution unit, when a risk warning of the same disaster type is included in the field disaster detection data or the disaster prevention facility status data under low-risk weather conditions where the forecast risk indicator of the wide-area disaster forecast data is below the initial disaster threshold for each disaster type, calculates the cumulative number of occurrences of a risk warning based on the risk warning history of the same disaster type stored in the history management database; the control execution unit, when the cumulative number of occurrences exceeds the statistical cumulative threshold frequency for each disaster type, registers the building corresponding to the building identification information as a vulnerable building for the same disaster type;A method for automatic building control based on artificial intelligence, comprising: a step in which, for a building registered as a vulnerable building, the control execution unit transmits a control output corresponding to a preset low-risk or medium-risk equipment group among the disaster prevention facilities to the DDC data linkage unit to automatically induce and control a preventive activation when the forecast risk indicator or the field disaster detection data reaches a preventive active threshold value set to a level lower than the initial disaster threshold; and a step in which a monitoring interface unit displays at least one of the registration result of the vulnerable building and the control result of the preventive activation, wherein the statistical cumulative threshold frequency is set based on statistical analysis of risk warning history by disaster type stored in the history management database or artificial intelligence analysis. Claim 11 A computer-readable recording medium having a program recorded thereon for performing an artificial intelligence-based disaster forecasting and on-site detection data-linked building automatic control method pursuant to Paragraph 10.

Images