Environment-Adaptive Electrical Signal Correlation-Based Edge-Cloud Integrated System for Predicting and Managing Electrical Accidents and Fires in Smart Farms

Abstract

The present invention relates to an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system, comprising: a sensor unit for detecting arcs, leakage currents, and thermal images of the equipment of the smart farm; an arc accident detection unit that removes environmental noise from series arc and ground fault arc signals generated in the power lines of the smart farm equipment and detects only actual arc signals; a leakage current detection unit that detects resistive leakage currents generated due to insulation breakdown, condensation, and cable degradation of the power equipment of the smart farm, while minimizing false positives occurring in high humidity and condensation environments; and an image analysis unit that corrects temperature information collected from a thermal imaging camera in response to solar interference and greenhouse light environments and analyzes the risk of overheating and ignition of the electrical equipment. It includes a control unit that aligns different physical-based signals received from the arc fault detection unit, leakage current detection unit, and image analysis unit onto a single time axis, calculates the risk level through integrated analysis, and controls a three-stage response of warning, partial shutdown, or total shutdown accordingly, and performs correlation analysis between arc, leakage, harmonics, and thermal images to significantly reduce false alarms and non-detection compared to a single technology, and enables high-reliability early prediction of complex electrical accidents.

Description

Technology Field

[0001] The present invention relates to an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system. More specifically, it relates to an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system that performs correlation analysis on multiple signals such as arc signals, leakage current, harmonics, voltage, thermal imaging, and environmental sensors in the safety management of power facilities in smart farms and agricultural facilities, and predicts, identifies, and responds to fires and electrical accidents at an early stage through an environment-adaptive correction algorithm and an edge-cloud linked analysis structure. Background Technology

[0002] In general, driven by demographic changes and advancements in information and communication technology (ICT), the winds of innovation are having a significant impact on the agricultural and manufacturing sectors. In particular, methods of integrating IT into corporate management and production to enhance productivity are spreading to various fields, such as smart farms and smart factories.

[0003] Smart farms are currently attracting attention as they integrate information and communication technology into agricultural production, processing, distribution, and consumption to automatically manage crop growth environments remotely and increase production efficiency. This serves as a means to enhance agricultural competitiveness in preparation for the decline and aging of the domestic farming population, and is part of efforts to improve price competitiveness through agricultural products grown near demand centers, such as urban farming. In particular, various attempts are being made in parallel to secure the competitiveness of farms, such as cultivating high-value crops or growing existing crops using organic farming methods through smart farms.

[0004] Furthermore, as a smart factory, the technology of attaching sensors to each piece of equipment and analyzing the status of the equipment or factory in real time using the sensing data collected by those sensors is accelerating its commercialization and efficiency due to the rapid advancement of sensors and data technology.

[0005] Therefore, significant investment is being made in actual smart farm and smart factory facilities. In addition, as efforts are being made to achieve economies of scale within a certain range to gain a competitive edge, indoor facilities such as smart farms and smart factories are trending toward becoming larger and super-large.

[0006] Meanwhile, agricultural facilities (smart greenhouses, livestock barns, vertical farms, solar-integrated greenhouses, etc.) use various loads such as LED ballasts, ventilation fans, boilers, and pumps for extended periods.

[0007] However, the reliability of existing power safety devices is significantly reduced due to the following special environmental characteristics.

[0008] 1. High humidity and condensation environments: Increased frequency of earth leakage circuit breakers malfunctions, and phase delay and distortion of sensor signals occur

[0009] 2. Solar Interference and Temperature Variations: Thermal imaging-based monitoring generates false alarms due to solar reflection—accelerating insulation degradation due to increased power line harmonics.

[0010] 3. Limitations of Single-Signal Based Safety Devices: Conventional Earth Leakage Circuit Breakers (ELCBs) and Arc Breakers (AFCIs) are single-signal based, unable to detect complex events (insulation degradation, micro-arc, ground fault arc), and suffer from serious problems of malfunction or non-detection.

[0011] 4. Limitations of real-time capabilities in cloud-based IoT systems: Inability to respond immediately to the site in the event of network failures – Potential failure in fire response due to latency issues in rural areas.

[0012] Therefore, next-generation smart farm safety management technology is required that combines environment-adaptive signal correction, multi-electrical signal correlation analysis, edge terminal-based immediate response, and cloud-based long-term AI analysis. The problem to be solved

[0013] The technical objective of the present invention is to solve the problems of the past by providing an environment-adaptive smart farm edge-cloud integrated fire and electrical accident prediction and management system based on electrical signal correlation analysis that predicts complex accidents early based on correlation analysis between electrical signals.

[0014] Furthermore, the technical objective of the present invention is to solve conventional problems by providing an environment-adaptive correction algorithm to eliminate noise based on high humidity, condensation, and solar interference, and to provide an environment-adaptive electrical signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system capable of applying Fast Fourier Transform (FFT), adaptive thresholding, and phase correction circuits.

[0015] Furthermore, the technical objective of the present invention is to solve the problems of the past by providing an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system that enables real-time field response based on edge terminals to perform independent operations in the event of network failure, and allows for warnings based on risk levels, as well as partial and total shutdowns.

[0016] Furthermore, the technical objective of the present invention is to solve the problems of the past by providing an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system that compares patterns of multiple greenhouses / facilities through cloud-based Macro AI analysis and OTA updates, and enables automatic updating of a long-term learning-based risk detection model.

[0017] Furthermore, the technical objective of the present invention is to solve conventional problems by providing an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system applicable to various sites, such as smart farms, buildings, and solar power facilities. means of solving the problem

[0018] An environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system according to the features of the present invention for solving these problems,

[0019] As an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system connected via a network with at least one smart farm edge terminal,

[0020] A sensor unit for detecting arc, leakage current, and thermal images of the equipment of the smart farm above;

[0021] An arc fault detection unit that removes environmental noise from series arc and ground fault arc signals generated in the power line of the smart farm facility among the output signals of the sensor unit and detects only the actual arc signal;

[0022] A leakage current detection unit that detects resistive leakage current caused by insulation breakdown, condensation, and cable degradation of the power equipment of the smart farm among the output signals of the sensor unit, while minimizing false positives occurring in high humidity and condensation environments;

[0023] An image analysis unit that corrects temperature information collected from a thermal imaging camera among the output signals of the sensor unit in response to solar interference and greenhouse light environments, and analyzes the risk of overheating and ignition of electrical equipment;

[0024] It includes a control unit that aligns different physical-based signals received from the arc fault detection unit, leakage current detection unit, and image analysis unit onto a single time axis, performs integrated analysis to calculate the risk level, and controls a three-stage response of warning, partial shutdown, or total shutdown accordingly.

[0025] The edge terminal above performs a risk-based 3-stage response according to the control of the control unit above.

[0026] It further includes a communication unit for communicating with the above edge terminal using one of LAN, WiFi, LTE, 5G, or NB-IoT methods.

[0027] The above control unit trains a long-term data-based AI model and distributes it to the edge terminal via OTA.

[0028] The above edge terminal can operate independently when the network is disconnected, and always preprocesses the signal from the sensor unit, performs local correlation analysis calculations, calculates the risk level, and performs a three-stage response based on the risk level.

[0029] The above network supports LAN, WiFi, LTE, 5G, and NB-IoT methods, and provides bidirectional data transmission and OTA functions.

[0030] The above risk is calculated by the following mathematical formula.

[0031]

[0032] Here, Earc is arc energy, and Ileak( ) is the phase-corrected leakage current, ΔT is the thermal image-based temperature rise, and Env is the environment sensor-based correction index. w1 is the arc sensitivity, w2 is the dielectric breakdown sensitivity, w3 is the heat risk contribution, and w4 is the environmental factor correction value. Effects of the invention

[0033] In an embodiment of the present invention, by performing correlation analysis between arcs, leakage, harmonics, and thermal images, a smart farm edge-cloud integrated fire and electrical accident prediction and management environment system based on environmentally adaptive electrical signal correlation analysis can be provided, which significantly reduces false alarms and non-detection compared to a single technology and enables high-reliability early prediction of complex electrical accidents.

[0034] In addition, an embodiment of the present invention can provide an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management environment system that improves responsiveness to field characteristics through environment-adaptive correction.

[0035] In addition, an embodiment of the present invention can provide an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management environment system capable of immediate on-site response even in the event of a network failure (Edge Autonomous Safety).

[0036] In addition, an embodiment of the present invention can provide an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management environment system capable of establishing a cloud-based national safety management system.

[0037] In addition, an embodiment of the present invention provides an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system that can be extended to data centers and solar power generation buildings in addition to smart farms. Brief explanation of the drawing

[0038] FIG. 1 is a configuration diagram of an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electric accident prediction and management system according to an embodiment of the present invention. FIG. 2 is a schematic diagram showing the sensor unit in an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electric accident prediction and management system according to an embodiment of the present invention. FIG. 3 is a schematic flowchart of the operation of an arc accident detection unit in an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electric accident prediction and management system according to an embodiment of the present invention. FIG. 4 is a schematic flowchart of the operation of a leakage current detection unit in an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electric accident prediction and management system according to an embodiment of the present invention. FIG. 5 is a schematic flowchart of the operation of the image analysis unit in an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electric accident prediction and management system according to an embodiment of the present invention. FIG. 6 is a schematic flowchart of the operation of the control unit in an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electric accident prediction and management system according to an embodiment of the present invention. Specific details for implementing the invention

[0039] Embodiments of the present invention are described below with reference to the attached drawings so that those skilled in the art can easily implement them. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. Furthermore, in order to clearly explain the present invention in the drawings, parts unrelated to the explanation have been omitted, and similar parts throughout the specification are denoted by similar reference numerals.

[0040] Throughout the specification, when a part is described as "including" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0041] FIG. 1 is a configuration diagram of an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electric accident prediction and management system according to an embodiment of the present invention.

[0042] Referring to FIG. 1, the environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system according to an embodiment of the present invention is,

[0043] As an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electric accident prediction and management system (100) connected via a network (200) and at least one smart farm edge terminal (300),

[0044] A sensor unit (310) for detecting arc, leakage current, and thermal image of the equipment of the smart farm above;

[0045] An arc fault detection unit (111) that removes environmental noise from the series arc and ground fault arc signals generated in the power line of the smart farm facility among the output signals of the sensor unit (310) and detects only the actual arc signal;

[0046] A leakage current detection unit (112) that detects resistive leakage current caused by insulation breakdown, condensation, and cable deterioration of the power equipment of the smart farm among the output signals of the sensor unit (310), while minimizing false positives occurring in high humidity and condensation environments;

[0047] An image analysis unit (113) that corrects temperature information collected from a thermal imaging camera among the output signals of the sensor unit (310) in response to solar interference and greenhouse light environment, and analyzes the risk of overheating and ignition of electrical equipment;

[0048] It includes a control unit (114) that aligns different physical-based signals received from the arc fault detection unit (111), leakage current detection unit (112), and image analysis unit (113) onto a single time axis, performs integrated analysis to calculate the risk level, and controls a three-stage response of warning, partial shutdown, or total shutdown accordingly.

[0049] The edge terminal (300) responds to three levels of risk based on control by the control unit (114).

[0050] The above edge terminal (300) further includes a communication unit (115) for communicating with one of LAN, WiFi, LTE, 5G, and NB-IoT methods.

[0051] The above control unit (114) learns a long-term data-based AI model and distributes it to the edge terminal (300) via OTA.

[0052] The above edge terminal (300) can operate independently when the network is disconnected, and preprocesses the signal of the sensor unit (310), performs local correlation analysis calculations, calculates the risk level, and performs a three-stage response based on the risk level.

[0053] The above network supports LAN, WiFi, LTE, 5G, and NB-IoT methods, and provides bidirectional data transmission and OTA functions.

[0054] The above risk is calculated by the following mathematical formula 1.

[0055] [Mathematical Formula 1]

[0056]

[0057] Here, Earc is arc energy, Ileak(-orr) is phase-corrected leakage current, ΔT is the thermal image-based temperature rise, and Env is the environment sensor-based correction index. w1 is arc sensitivity, w2 is dielectric breakdown sensitivity, w3 is the heat risk contribution, and w4 is the environmental factor correction value.

[0058] FIG. 2 is a schematic diagram showing the sensor unit in an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electric accident prediction and management system according to an embodiment of the present invention.

[0059] Referring to FIG. 2, the sensor unit (310) includes a current sensor (CT), a voltage sensor, a leakage current sensor (ZCT), a broadband current sensor for harmonic analysis, a thermal imaging camera, and an environmental sensor (temperature, humidity, condensation, solar radiation, etc.).

[0060] FIG. 3 is a schematic flowchart of the operation of an arc accident detection unit in an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electric accident prediction and management system according to an embodiment of the present invention.

[0061] Referring to FIG. 3, the arc fault detection unit (111) is a device that detects only the actual arc signal with high reliability by separating the series arc and ground fault arc signals generated in the power line of the smart farm facility from various environmental noises (inverter switching, harmonics, solar interference signals, etc.).

[0062] The signal input from the sensor unit (310) is a high-speed sampling current signal (AC / DC), a voltage waveform, and a raw signal including harmonics / noise.

[0063] The above arc fault detection unit (111) performs band-stop filtering.

[0064] Arc signals generally have an irregular high-frequency energy distribution of 2 kHz to 100 kHz. On the other hand, inverter switching, motors, LED ballasts, etc., generate repetitive signals in a specific frequency band. Therefore, the arc fault detection unit (111) eliminates the 60 Hz fundamental wave, low-order harmonics such as the 3rd, 5th, and 7th harmonics, the fixed switching frequency of the inverter / LED ballast (e.g., 4 kHz, 8 kHz, etc.), and fixed frequency interference generated by the solar inverter.

[0065] Specifically, the arc fault detection unit (111) performs an FFT transformation, analyzes the energy of each frequency component, matches it with a predefined "Forbidden Band Table," removes the designated frequency component band using a notch filter, and resynthesizes only the remaining components to use as an arc candidate signal. As a result, only a pure arc candidate signal remains after removing environmental and equipment interference components.

[0066] Then, the arc fault detection unit (111) executes a phase alignment algorithm. The arc signal may have a different phase from the leakage current and voltage. Additionally, phase delay of the ZCT sensor occurs due to condensation and changes in humidity. Therefore, phase alignment is performed.

[0067] The phase correction operation method of the above arc accident detection unit (111) is as follows.

[0068] First, the voltage waveform is set to the reference phase (0°), and the current and leakage current waveforms are phase-estimated based on the Hilbert transform. Then, the phase delay is corrected using the values ​​from environmental sensors (humidity and temperature), and all signals are aligned on the same time axis. Through this, phase correction is performed by accurately comparing whether the arc and leakage current occurred at the same time.

[0069] In addition, the arc fault detection unit (111) performs solar interference compensation. Abnormal harmonics may increase in power lines due to strong light reflection inside a solar inverter or greenhouse. Therefore, compensation is required.

[0070] The correction method for the above arc accident detection unit (111) is as follows.

[0071] First, pattern analysis based on time of day and solar radiation is performed, followed by normalization and adaptive threshold adjustment. Finally, a multiple regression correction formula based on weather and solar radiation sensor values ​​is applied. Through this process, "high-frequency noise caused by sunlight" and "actual arc high frequencies" are clearly distinguished.

[0072] In addition, the arc accident detection unit (111) performs the final arc determination logic.

[0073] First, determine whether the high-frequency energy change rate ΔEarc is above a threshold value, determine whether the irregularity of the pattern (Entropy value) exceeds a reference value, and check whether it occurs simultaneously with leakage current after phase alignment (τcoinc ≤ 500ms). If all three conditions are satisfied, determine that it is a series arc or a ground fault arc.

[0074] FIG. 4 is a schematic flowchart of the operation of a leakage current detection unit in an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electric accident prediction and management system according to an embodiment of the present invention.

[0075] Referring to FIG. 4, the leakage current detection unit (112) (ZCT-based Environment-Adaptive Leakage Current Detection Unit) is an environment-adaptive detection unit that detects resistive leakage current (Igr) caused by insulation breakdown, condensation, and cable deterioration of power equipment in a smart farm, while minimizing false positives that occur in high humidity and condensation environments.

[0076] The above leakage current detection unit (112) detects a minute current (in the mA range) leaking outside the wire from the Zero-Phase Current Transformer (ZCT) at the input of the sensor unit (310). It then converts it into a real-time waveform through an ADC.

[0077] The above leakage current detection unit (112) performs a phase correction operation adaptive to a humidity / condensation environment. An "inductive phase delay" occurs in the ZCT sensor under humidity / condensation conditions, and the leakage current detection unit (112) corrects this.

[0078] The correction algorithm procedure of the leakage current detection unit (112) is as follows.

[0079] First, environmental data is input from the temperature and humidity sensors of the sensor unit (310), and an "environmental correction coefficient χ" is calculated based on the possibility of condensation (Dew Point calculation), and the leakage current waveform Iraw(t) is converted into the following mathematical formula 2.

[0080] [Mathematical Formula 2]

[0081]

[0082] As a result, the dry environment is χ - 0, the high humidity environment is χ increase, and the condensation environment is automatically corrected to the maximum value of χ.

[0083] In addition, the leakage current detection unit (112) minimizes false positives under high humidity and condensation conditions.

[0084] The above leakage current detection unit (112) performs base noise modeling to identify whether the leakage current is an actual leakage current or noise caused by humidity.

[0085] Here, the ZCT noise pattern in the humidity range of 70~95% is stored in a database, and the presence of noise is determined by comparing it with the real-time waveform.

[0086] Then, the leakage current detection unit (112) performs a voltage phase comparison.

[0087] Here, if the leakage current has a resistive relationship with the voltage (same phase), it is determined to be "actual leakage," and if the phase difference is large, it is determined to be "humidity / condensation-based inductive noise."

[0088] And the above leakage current detection unit (112) extracts the resistive component as shown in the following mathematical formula 3.

[0089] [Mathematical Formula 3]

[0090]

[0091] This allows it to detect only actual dangerous leakage current.

[0092] FIG. 5 is a schematic flowchart of the operation of the image analysis unit in an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electric accident prediction and management system according to an embodiment of the present invention.

[0093] Referring to FIG. 5, the image analysis unit (113) (Thermal Image Analysis Unit) is a device that corrects temperature information collected from the thermal imaging camera of the sensor unit (310) in response to solar interference and greenhouse light environment, and evaluates the risk of overheating and ignition of electrical equipment.

[0094] The above image analysis unit (113) receives the thermal camera image (resistivity / temperature data) of the sensor unit (310), the solar radiation sensor, and the output of the greenhouse internal light intensity sensor as input data.

[0095] In addition, the image analysis unit (113) performs solar interference correction (Adaptive Thresholding). When sunlight is reflected inside the glass greenhouse, a false high-temperature area is created in the camera. To correct this, first, the correction formula of the solar irradiance mathematical formula 4 is applied to all pixel T(x,y) of the image.

[0096] [Mathematical Formula 4]

[0097]

[0098] Then, the adaptive threshold is automatically adjusted.

[0099] In this case, the threshold automatically increases as the light intensity increases. Additionally, only the actual heat generation area is selected.

[0100] Next, reflectance according to greenhouse material (PE film, glass, etc.) Corrects.

[0101] As a result, only the actual heat source is accurately detected.

[0102] In addition, the above image analysis unit (113) analyzes the overheating pattern.

[0103] First, calculate the temperature change as in mathematical formula 5.

[0104] [Mathematical Formula 5]

[0105]

[0106] Then, the calculated temperature change is compared with the normal heating pattern DB of the specific equipment.

[0107] Next, concentrated fever is detected using hot spot clustering (K-means).

[0108] And it classifies types such as high temperature, dispersion, and linear patterns.

[0109] FIG. 6 is a schematic flowchart of the operation of the control unit in an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electric accident prediction and management system according to an embodiment of the present invention.

[0110] Referring to FIG. 6, the control unit (114) performs signal fusion and risk determination.

[0111] That is, multi-sensor signal fusion logic is performed, and the control unit (114) aligns different physical-based signals, such as arc, leakage current, and thermal image, onto a single time axis and performs integrated analysis in the next procedure.

[0112] First, the control unit (114) performs phase alignment.

[0113] For example, the control unit (114) aligns different physical-based signals received from the arc fault detection unit (111), the leakage current detection unit (112), and the image analysis unit (113) onto a single time axis.

[0114] Next, the control unit (114) extracts the following feature quantity.

[0115] For example, arc (high-frequency energy Earc), leakage current (resistive component) ), increase in thermal imaging temperature ( ), environmental value (RH), G (solar radiation), Extracts (solar altitude), etc.

[0116] The next control unit (114) determines simultaneity ( ).

[0117] At this point, the key is to determine whether different types of abnormal signals occurred within 0.5 seconds, while identifying complex accidents (such as ground fault arcs).

[0118] In addition, the control unit (114) calculates the risk using a risk calculator.

[0119] The risk level is calculated using a formula like Mathematical Formula 1.

[0120] [Mathematical Formula 1]

[0121]

[0122] Here, Earc is arc energy, Ileak(-orr) is phase-corrected leakage current, ΔT is the thermal image-based temperature rise, and Env is the environment sensor-based correction index. w1 is arc sensitivity, w2 is dielectric breakdown sensitivity, w3 is the heat risk contribution, and w4 is the environmental factor correction value.

[0123] Here, a three-step response is performed when the risk level exceeds the threshold.

[0124] The three-stage response (Warning, Partial Shutdown, Full Shutdown) is as follows.

[0125] Step 1 is the Warning stage.

[0126] Here, monitoring is enhanced after text / app notifications, dashboard (120) warning lights are turned on and logs are recorded.

[0127] Step 2 is the Partial Shutdown stage.

[0128] Here, only hazardous equipment can be selectively blocked, but one LED line, one ventilation fan, etc. can be blocked.

[0129] Step 3 is the Full Shutdown step.

[0130] Here, it executes immediately upon the simultaneous occurrence of multiple signals, and as a final protection step for fire prevention, it is possible to interlock with firefighting equipment.

[0131] These embodiments of the present invention can be applied in various ways.

[0132] For example, it can be applied to the early detection of ground fault arcs in smart greenhouses.

[0133] In this case, if leakage current increases in the LED ballast line and an arc signal is detected, When the conditions are met, partial blocking can be performed at the edge terminal (300).

[0134] In addition, early detection of insulation degradation in livestock barn electric heaters is possible.

[0135] For example, if an anomaly in FFT-based harmonic components is detected, it is determined to be insulation degradation, and an OTA policy update is possible after pattern comparison in the cloud.

[0136] In addition, it can detect a rapid rise in temperature in the solar greenhouse.

[0137] For example, a thermal imaging camera can detect a sudden rise in temperature, immediately shut down the system on-site, and transmit the history to the cloud.

[0138] In an embodiment of the present invention, correlation analysis between arcs, leakage, harmonics, and thermal images is performed, significantly reducing false alarms and non-detection compared to a single technology, and enabling highly reliable early prediction of complex electrical accidents.

[0139] In addition, in an embodiment of the present invention, field characteristic responsiveness is improved through environment-adaptive correction.

[0140] In addition, in the embodiments of the present invention, immediate on-site response is possible even in the event of a network failure.

[0141] In addition, in the embodiments of the present invention, it is possible to establish a cloud-based national safety management system.

[0142] In addition, in the embodiments of the present invention, it is possible to extend it to data centers and solar power generation buildings in addition to smart farms.

[0143] Although embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements by those skilled in the art using the basic concept of the present invention as defined in the following claims also fall within the scope of the present invention.

Claims

Claim 1 An environment-adaptive electric signal correlation-based smart farm edge-cloud integrated fire and electrical accident prediction and management system connected via a network to at least one smart farm edge terminal, comprising: a sensor unit for detecting arcs, leakage currents, and thermal images of the smart farm equipment; an arc accident detection unit that removes environmental noise from series arc and ground fault arc signals generated from the power lines of the smart farm equipment among the output signals of the sensor unit and detects only actual arc signals; a leakage current detection unit that detects resistive leakage currents generated due to insulation breakdown, condensation, and cable degradation of the smart farm power equipment among the output signals of the sensor unit, while minimizing false positives occurring in high humidity and condensation environments; an image analysis unit that corrects temperature information collected from a thermal imaging camera among the output signals of the sensor unit in response to solar interference and greenhouse light environments, and analyzes the risk of overheating and ignition of electrical equipment; and aligns different physical-based signals received from the arc accident detection unit, the leakage current detection unit, and the image analysis unit onto a single time axis, performs integrated analysis to calculate the risk level, and accordingly An environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system comprising a control unit that controls a three-stage response of warning, partial shutdown, or total shutdown, wherein the edge terminal performs a three-stage response based on risk level according to the control of the control unit. Claim 2 An environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system according to claim 1, further comprising a communication unit for communicating with the edge terminal via one of LAN, WiFi, LTE, 5G, or NB-IoT methods. Claim 3 In claim 1, the control unit learns an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system that learns a long-term data-based AI model and distributes it to the edge terminal via OTA. Claim 4 An environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system according to any one of claims 1 to 3, wherein the edge terminal is capable of independent operation when the network is disconnected, and preprocesses the signal of the sensor unit at all times, performs local correlation analysis calculations, calculates the risk level, and provides a three-stage response based on the risk level. Claim 5 In paragraph 4, the above network supports LAN, WiFi, LTE, 5G, and NB-IoT methods and provides an environment-adaptive electrical signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system that provides bidirectional data transmission and OTA functions. Claim 6 In claim 5, the above risk is calculated by the following mathematical formula, an environment-adaptive electric signal correlation analysis-based smart farm edge-cloud integrated fire and electrical accident prediction and management system. Here, Earc is arc energy, and Ileak( ) is the phase-corrected leakage current, ΔT is the thermal image-based temperature rise, and Env is the environment sensor-based correction index. w1 is the arc sensitivity, w2 is the dielectric breakdown sensitivity, w3 is the heat risk contribution, and w4 is the environmental factor correction value.

Images