KR102641838B1 - Damage prevention system of solar photovoltaic power station using artificial intelligence from lightning stroke - Google Patents

Abstract

The present invention predicts the possibility of lightning strikes and movement paths, allowing solar power plants located in places where lightning strikes are likely to occur to quickly prepare for lightning strikes, thereby preventing casualties and breakdowns in solar power plants due to lightning strikes. This is about a lightning damage prevention system for solar power plants using artificial intelligence.
The present invention detects lightning within a certain radius and generates a lightning detection signal, and provides lightning learning including past weather information corresponding to the occurrence of lightning from a plurality of solar power plants (100) located at a certain distance apart and the Korea Meteorological Administration server (200). A control server (300) that collects data and monitors whether the current weather information for each region corresponds to the lightning occurrence pattern through a lightning learning model that learns the lightning occurrence pattern, determines whether or not lightning is likely to occur in each region, and generates lightning prediction information (300) It includes, and the control server 300 collects movement learning data including lightning occurrence point extracted through the lightning detection signal and lightning observation information in which the location of lightning is stored according to the past weather information to create a lightning movement pattern. Through the learned lightning movement learning model, path prediction information about the movement path of lightning is generated according to the lightning occurrence point and current meteorological information, and the lightning prediction information is generated through the location information of the solar power plant 100 stored in advance. Accordingly, the solar power plant 100 within a certain radius is extracted and the lightning prediction information is transmitted, and when the solar power plant 100 receives the lightning prediction information, it sends an evacuation signal or warning according to the lightning prediction information. It is characterized by performing the function of turning off the power to the inverter and connection panel while outputting a signal.

Description

Lightning damage prevention system for solar power plants using artificial intelligence {DAMAGE PREVENTION SYSTEM OF SOLAR PHOTOVOLTAIC POWER STATION USING ARTIFICIAL INTELLIGENCE FROM LIGHTNING STROKE}

The present invention relates to a lightning damage prevention system for solar power plants using artificial intelligence. More specifically, it predicts the possibility of lightning strikes and movement paths so that solar power plants located in places where lightning strikes are likely to occur can quickly respond to lightning strikes. This is about a lightning damage prevention system for solar power plants using artificial intelligence that can prevent casualties and breakdowns of solar power plants due to lightning by preparing for lightning strikes.

Lightning is most commonly caused by thunderclouds, and is a phenomenon in which the electric charge of the thundercloud is released into the atmosphere through the destruction of air insulation. In this process, an electromagnetic field is generated and propagated into the atmosphere. At this time, the thundercloud is charged with electricity, and the upper part is generally made up of positively charged ice crystals, and the lower part is made up of negatively charged water droplets.

These thunderclouds are created through a cumulus stage with strong updrafts, a maturation stage with updrafts and gusts, thunder, lightning, and showers, and a dissipation stage where downdrafts dominate and stability is restored, often in hot and humid air. This is likely to occur when rising in an unstable environment.

Lightning causes significant damage in a very small area, including power outages, fires, destruction of facilities and casualties, and the damage is gradually increasing, with approximately 1 million cases occurring each year.

In particular, damage from lightning occurs frequently when operating solar power plants. The reason is that in the case of large-scale solar power plants, multiple solar power plants are installed flatly on vast areas of land and tall structures that block sunlight are intentionally avoided or eliminated, making them easy targets for lightning strikes.

When a lightning strike occurs in a solar power plant, a direct strike may cause failure of modules, trackers, inverters, etc., and in the case of modules connected in parallel, even if only one module receives a direct strike, all connected modules are damaged. Because it cannot function, the damage increases.

In addition, when overvoltage occurs due to lightning, equipment breaks down or the lightning strikes the ground, causing overvoltage damage in a 1 to 2 km radius from the point of occurrence.

This overvoltage reduces the efficiency of solar power plants and causes breakdowns. Insulation breakdown occurs in areas where overvoltage is applied, and overcurrent flows in, damaging the device.

Accordingly, in order to prevent lightning damage to solar power plants, in Publication Patent No. 10-2017-0123912 (publication date: November 9, 2017), solar panels and direct current transmitted from the solar panels are converted into alternating current. An inverter that converts, a current cutoff unit that selectively connects the solar panel and the inverter, an electric field sensor that measures the intensity of the atmospheric electric field, and are connected to the current cutoff unit and the electric field sensor, respectively, and measure the electric field sensor. A lightning damage prevention system for a solar power plant is disclosed, which includes a control unit that controls the current blocking unit according to the intensity of the atmospheric electric field.

However, in the related art, as in the above-mentioned published patents, there is a problem in that it is difficult to prepare in advance for preventing damage due to lightning because the current blocking unit is controlled when lightning is detected. In other words, lightning does not occur in one place and dissipate immediately, but is moved and then dissipated, so if a lightning strike occurs at one solar power plant, the probability of lightning striking other solar power plants located nearby increases, resulting in another lightning strike. Damage may occur, but in the past, there is a problem that it is difficult to prepare for it in advance.

Accordingly, there is a need for a technical solution that predicts the path of lightning and allows solar power plants located in the path of lightning to prepare in advance for lightning.

The present invention was proposed to solve the above-mentioned problems, and predicts the possibility of lightning strikes and movement paths to secure preparation time for lightning strikes in order to effectively protect solar power plants and nearby workers located in places where lightning strikes are likely to occur. The purpose is to provide a lightning damage prevention system for solar power plants using artificial intelligence.

delete

Meanwhile, the purpose is to provide a lightning damage prevention system for solar power plants using artificial intelligence that can minimize damage to the operation of solar power plants by predicting the disappearance of lightning.

The problems to be solved by the present invention are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below.

The lightning damage prevention system for solar power plants using artificial intelligence according to the present invention to achieve the above purpose includes a plurality of solar power plants that are located at a certain distance apart and detect lightning within a certain radius to generate a lightning detection signal ( 100), and collect lightning learning data including past weather information corresponding to lightning occurrences from the Korea Meteorological Administration server 200 and monitor whether the current weather information for each region corresponds to the lightning occurrence pattern through a lightning learning model that learned the lightning occurrence pattern. and a control server 300 that determines whether lightning may occur in each area and generates lightning forecast information. The control server 300 communicates with the Korea Meteorological Administration server 200 and the solar power plant 100. A server communication unit 301 for this purpose, a solar power plant DB unit 302 that stores solar power plant information including location information of the solar power plant in advance, and a regional past weather information received and stored from the Korea Meteorological Administration server 200. A weather information collection unit 303 to receive current weather information for each region in real time, collects lightning learning data including past weather information corresponding to lightning strikes among the past weather information, and detects lightning strikes based on the lightning learning data. A lightning prediction learning unit 304 for generating a lightning learning model by learning the corresponding lightning occurrence pattern, and a lightning strike learning model to monitor whether the current weather information corresponds to the lightning occurrence pattern and determine whether or not lightning is likely to occur. a lightning monitoring unit 305 for extracting lightning prediction points and generating lightning prediction information; and, when it is determined that lightning has occurred, the sun located within a certain radius based on the lightning prediction point through the solar power plant information. A lightning prediction notification unit 306 for extracting a photovoltaic power plant and transmitting the lightning prediction information, and receiving a lightning detection signal from the solar power plant 100 that detects lightning after transmitting the lightning prediction information, determines the lightning occurrence point. A lightning occurrence extraction unit 307 for extracting and generating lightning occurrence information, and a lightning observation collection unit for receiving and storing lightning observation information that observes the location of lightning according to past weather information from the Korea Meteorological Administration server 200. (308) and a lightning movement learning unit (308) to collect movement learning data including the lightning observation information and learn lightning movement patterns for lightning movement conditions based on the movement learning data to generate a lightning movement learning model ( 309) and a movement path prediction unit 310 for predicting the movement path of lightning according to the current weather information based on the lightning occurrence information through the lightning movement learning model and generating path prediction information, and the solar power A lightning movement notification unit 311 for extracting a solar power plant according to the route prediction information through power plant information and transmitting the route prediction information, and using the lightning movement learning model to determine whether the lightning is scheduled to dissipate and determine whether the lightning is about to be extinguished. A lightning extinction prediction unit ( 312), and the solar power plant 100 includes a lightning detection unit 101 for detecting lightning that occurs within a preset radius, generating a lightning detection signal, and then transmitting it to the control server 300. , a power plant communication unit 102 for communicating with the control server 300, and when receiving the lightning prediction information or the route prediction information from the control server 300, according to the lightning prediction information or the route prediction information. It may be configured to include a hazard warning unit 103 for outputting an evacuation signal or warning signal, and a control unit 104 for cutting off power to the inverter and the connection panel when the evacuation signal or warning signal is output.

delete

delete

delete

delete

delete

delete

As described above, according to the present invention, by predicting the possibility of lightning strikes and the movement path, it is possible to secure lightning preparation time to prepare for lightning strikes at a solar power plant located in a place where lightning strikes are likely to occur before suffering damage from lightning strikes. This has the effect of protecting the solar power plant by preventing breakdown of the solar power plant.

In addition, by using lightning detection information from solar power plants that detect lightning, the accuracy of predicting the path of lightning can be improved and the intensity of lightning can be predicted to protect other solar power plants and nearby workers from lightning. It has the effect of improving the reliability of notifications about lightning strikes.

Meanwhile, by predicting the extinction of lightning, unnecessary blocking of solar power plants can be prevented, minimizing damage to solar power generation, and promoting safe and efficient operation of solar power plants.

1 is a schematic lightning damage prevention system for a solar power plant using artificial intelligence according to an embodiment of the present invention;
Figure 2 is a schematic block diagram of a solar power plant in a lightning damage prevention system for solar power plants using artificial intelligence according to an embodiment of the present invention;
Figure 3 is a schematic block diagram of a control server in a lightning damage prevention system for solar power plants using artificial intelligence according to an embodiment of the present invention;
Figure 4 shows a lightning generation extraction unit in a lightning damage prevention system for a solar power plant using artificial intelligence according to an embodiment of the present invention;
Figure 5 is a flowchart of a method for preventing lightning damage to a solar power plant using a lightning damage prevention system for a solar power plant using artificial intelligence according to an embodiment of the present invention.

Hereinafter, a lightning damage prevention system for a solar power plant using artificial intelligence according to the present invention will be described in detail with reference to the attached drawings.

Figure 1 is a schematic lightning damage prevention system for a solar power plant using artificial intelligence according to an embodiment of the present invention, and Figure 2 is a lightning damage prevention system for a solar power plant using artificial intelligence according to an embodiment of the present invention. is a schematic block diagram of a solar power plant, FIG. 3 is a schematic block diagram of a control server in a lightning damage prevention system for a solar power plant using artificial intelligence according to an embodiment of the present invention, and FIG. 4 is a schematic block diagram of the control server. It is a lightning generation extraction unit in a lightning damage prevention system for solar power plants using artificial intelligence according to an embodiment of the invention, and Figure 5 shows a lightning damage prevention system for solar power plants using artificial intelligence according to an embodiment of the present invention. This is a flowchart of the lightning damage prevention method for solar power plants.

In adding quotation marks to the components of the drawings, the same components are given the same symbols as much as possible even if they are shown in different drawings, a known function that is judged to unnecessarily obscure the gist of the present invention. and detailed description of the configuration is omitted. Additionally, directional terms such as 'top', 'bottom', 'front', 'back', 'front', 'front', 'back', etc. are used in connection with the orientation of the disclosed drawing(s). Since components of embodiments of the present invention can be positioned in various orientations, the term directional is used for illustrative purposes and is not limiting.

As shown in FIG. 1, the lightning damage prevention system for solar power plants using artificial intelligence according to a preferred embodiment of the present invention is located at a certain distance apart and detects lightning within a certain radius to generate a lightning detection signal. Current weather information for each region is generated through a lightning learning model that learns lightning occurrence patterns by collecting lightning learning data including past weather information corresponding to lightning strikes from a number of solar power plants (100) and the Korea Meteorological Administration server (200). It includes a control server 300 that monitors whether it corresponds to an occurrence pattern, determines whether lightning is likely to occur in each area, and generates lightning prediction information.

The solar power plants 100 are installed at a certain distance from each other and may be configured by region.

The solar power plant 100 includes a solar panel unit composed of a plurality of solar cells, an inverter cut-off unit for blocking the power of an inverter that converts the electrical energy generated by the solar panel unit into alternating current, and the solar panel unit. It may include a connection panel blocking unit connected between the panel unit and the inverter to cut off the power of the connection panel that supplies the electric energy to the inverter.

Here, the solar panel unit can operate in a power generation mode in which the front surfaces of one side and the other side of the square-shaped panel support are positioned to be inclined at a certain angle, or in a safety mode in which the rear surfaces of one side and the other side of the panel support part are positioned close to each other. there is.

As shown in FIG. 2, the solar power plant 100 includes a lightning detection unit ( 101), a power plant communication unit 102 for communicating with the control server 300, and when receiving lightning prediction information or route prediction information from the control server 300, the lightning prediction information or the route prediction information Accordingly, it includes a hazard warning unit 103 for outputting an evacuation signal or warning signal, and a control unit 104 for cutting off the power to the inverter and the connection panel when the warning signal is output.

The lightning detection unit 101 may use a Time-Of-Arrival (TOA) method or a Magnetic Direction Finding (MDF) method.

The TOA (Time-Of-Arrival) method can observe lightning using two sensors, and when a lightning strike occurs, the latitude, longitude, and distance of the lightning can be estimated using the time the lightning arrived at each sensor. there is.

In addition, the TOA method preferably uses three or more lightning sensors to obtain lightning location accuracy.

The MDF (Magnetic Direction Finding) method can detect the location of lightning using the direction in which the lightning occurred. In other words, the MDF method is a method of determining the location of a lightning strike from two antennas by measuring the direction in which electromagnetic waves generated from a lightning strike arrive.

By combining the TOA method and the MDF method, the problem of inaccuracy in the location of lightning can be prevented, making it possible to determine the exact location of lightning, and can be helpful in predicting the path of lightning in the future. It is possible to improve the prediction rate.

The lightning detection unit 101 may use the AS3935 Franklin Lightning Sensor.

The AS3935 Franklin Lightning sensor has a built-in RF receiver that detects electrical emissions from lightning strikes and converts the RF signal into an estimate of the distance to the lightning strike through an algorithm, allowing it to calculate the estimated distance to the lightning strike ranging from 40 km to 1 km.

Here, the lightning detection unit 101 preferably has a detection radius set in advance around the solar power plant 100, and generates a lightning detection signal by detecting lightning that occurs within a lightning detection range of a preset radius. do.

It is preferable that the lightning detection signal be generated by including not only the lightning detection time when the lightning was detected but also information on the location where the lightning occurred.

The lightning detection signal generated by the lightning detection unit 101 is transmitted to the control server 300.

The power plant communication unit 102 is for communicating with the control server 300, and it is desirable to use a wireless network.

When receiving the lightning prediction information or the path prediction information from the control server 300, the danger warning unit 103 outputs an evacuation signal to notify workers of evacuation or a warning signal to notify workers of caution. .

Here, the danger warning unit 103 can output the evacuation signal or warning signal according to the lightning prediction information or the path prediction information, and determines whether the location of the solar power plant 100 is included in a dangerous section. To make a decision, set a risk zone in advance.

The risk section is preferably set within a certain distance based on the lightning prediction point included in the lightning prediction information or the lightning movement path included in the path prediction information.

That is, the risk warning unit 103 outputs the evacuation signal when the location of the solar power plant 100 is in a danger section based on the lightning prediction information or the path prediction information, and the solar power plant 100 ), if the location is not a danger zone based on the lightning prediction information or the path prediction information, the warning signal can be output.

The control unit 104 determines and controls whether to cut off power to the inverter, connection panel, and communication equipment when the evacuation signal or warning signal is output.

The control unit 104 cuts off power to the inverter, connection panel, and communication equipment when an evacuation signal is output from the danger warning unit 103, and when a warning signal is output from the danger warning unit 103, the inverter, the connection panel, etc. It is desirable to monitor without turning off the power of and communication equipment.

In addition, the control unit 104 cuts off the power to the inverter, connection panel, and communication equipment when a lightning detection signal is generated in the lightning detection unit 101 even if the warning signal is output.

As shown in FIG. 3, the control server 300 includes a server communication unit 301 for communicating with the Korea Meteorological Administration server 200 and the solar power plant 100, and location information of the solar power plant in advance. A solar power plant DB unit 302 that stores solar power plant information, a weather information collection unit 303 for receiving regional past weather information and current weather information from the Korea Meteorological Administration server 200, and a weather information collection unit 303 that stores the past weather information. A lightning prediction learning unit 304 for collecting lightning learning data, learning lightning occurrence patterns corresponding to lightning occurrences, and generating a lightning learning model, and the current weather information is applied to the lightning occurrence pattern through the lightning learning model. A lightning monitoring unit 305 to monitor whether a lightning strike is applicable, determine whether a lightning strike is likely to occur, extract a lightning prediction point, and generate lightning prediction information, and, if it is determined that a lightning strike has occurred, a location within a certain radius based on the lightning prediction point. A lightning prediction notification unit 306 for transmitting the lightning prediction information to the solar power plant, and receiving a lightning detection signal from the solar power plant 100 that detects the lightning after transmitting the lightning prediction information to provide lightning occurrence information. A lightning occurrence extraction unit 307 for generating a lightning strike observation collection unit 308 for receiving and storing lightning observation information observing the location of lightning according to past weather information from the Korea Meteorological Administration server 200; A lightning movement learning unit 309 for collecting movement learning data including lightning observation information and learning lightning movement patterns to create a lightning movement learning model, and a lightning movement learning model based on the lightning occurrence information. A movement path prediction unit 310 for predicting the movement path of lightning according to weather information and generating route prediction information, and a lightning movement notification for extracting solar power plants according to the route prediction information and transmitting the route prediction information. It includes a unit 311 and a lightning extinction prediction unit 312 for determining whether the lightning is scheduled to disappear using the lightning movement learning model, predicting the life cycle of the lightning, and generating extinction prediction information.

The server communication unit 301 uses a wireless network to communicate with the Korea Meteorological Administration server 200 and the solar power plant 100.

The solar power plant DB unit 302 stores solar power plant information, including location information of a plurality of solar power plants, in advance.

The solar power plant information may further include not only location information of the solar power plant, but also identification information and manager information to identify the solar power plant from other solar power plants.

The weather information collection unit 303 receives and stores past weather information for each region from the Korea Meteorological Administration server 200. The past weather information may include past temperature, past humidity, past occurrence of thunderclouds, past precipitation, past wind, past lightning, etc.

In addition, the weather information collection unit 303 receives current weather information, including the occurrence of thunderclouds for each region, temperature, and humidity, in real time through Open-API. Since the weather information collection unit 303 collects current weather information for each region through Open-API, it is possible to shorten the time to determine whether lightning is likely to occur.

Here, the weather information collection unit 303 can accumulate and store the current weather information as past weather information.

The lightning prediction learning unit 304 collects lightning learning data including past weather information corresponding to lightning occurrences among the past weather information. Based on the lightning learning data, a lightning learning model is created by learning the lightning occurrence pattern corresponding to the lightning occurrence.

The lightning occurrence pattern is preferably a pattern for weather information such as temperature, precipitation, wind, and humidity on the date of the lightning strike using the past weather information.

The lightning monitoring unit 305 determines whether a lightning strike is likely to occur by determining whether the current weather information corresponds to the lightning occurrence pattern based on the lightning learning model. That is, the lightning monitoring unit 305 determines that lightning is likely to occur if the current weather information corresponds to the lightning occurrence pattern, and if the current weather information does not correspond to the lightning occurrence pattern, it is determined that lightning is likely to occur. It is judged that the possibility of this occurring is low.

At this time, when the lightning monitoring unit 305 determines that lightning has occurred based on the current weather information, it extracts a lightning prediction point using the area of the current weather information and generates lightning prediction information.

It is preferable that the lightning prediction information includes not only the lightning prediction point but also the expected lightning occurrence time zone. Here, the lightning strike expected time zone is a time zone in which lightning is likely to occur at the lightning strike prediction point using the wind of the current weather information.

The lightning prediction notification unit 306 is located within a certain radius based on the lightning prediction point among the solar power plant information stored in the solar power plant DB unit 302 when the lightning monitoring unit 305 determines that lightning has occurred. extract solar power plants.

The lightning prediction notification unit 306 transmits the lightning prediction information to the extracted solar power plant. The reason for transmitting the lightning prediction information to solar power plants located within a certain radius based on the lightning prediction point is to minimize damage to the solar power plant when an error occurs in the lightning prediction point.

As described above, the lightning monitoring unit 305 monitors the current weather information through the lightning learning model and transmits lightning prediction information to the relevant solar power plant through the lightning prediction notification unit 306, so that the solar power plant can be exposed to sunlight before lightning occurs. The probability of protecting the photovoltaic power plant can be increased, and the risk of having to deal with lightning strikes in locations adjacent to the photovoltaic power plant can be reduced, minimizing human casualties and financial damage due to breakdowns.

The lightning occurrence extraction unit 307 receives a lightning detection signal from the solar power plant 100 that detects lightning after transmitting the lightning prediction information and extracts the lightning occurrence point.

For example, as shown in FIG. 4, if the solar power plants that detected lightning are solar power plants A, solar power plants B, and solar power plants C, the location information where the lightning occurred in each lightning detection signal is provided. Use this method to extract the lightning occurrence point.

When the lightning occurrence point is extracted from the lightning occurrence extraction unit 307 as described above, the exact lightning occurrence point can be extracted, and the intensity of the lightning can also be extracted.

Lightning occurrence information is generated including the extracted lightning occurrence point.

The lightning observation collection unit 308 receives and stores lightning observation information containing the location of lightning according to past weather information from the Korea Meteorological Administration server 200.

At this time, it is preferable that the past weather information further includes the moving direction of past thunderclouds, past wind speed, and past air pressure troughs, which are conditions under which lightning moves. That is, the lightning observation collection unit 308 receives past weather information including the moving direction of thunderclouds, wind speed, and pressure trough, which are conditions for lightning movement, from the Korea Meteorological Administration server 200.

The lightning mobile learning unit 309 collects mobile learning data including lightning observation information stored in the lightning observation collection unit 308, and uses the collected observation learning data to determine the location of lightning according to the current weather information. A lightning movement learning model is created by learning changing lightning movement patterns.

The movement path prediction unit 310 predicts the path along which lightning moves according to the current weather information based on the lightning occurrence point of the lightning occurrence information through the lightning movement learning model. In other words, the movement path prediction unit 310 extracts the direction and speed of lightning according to the current weather information through the lightning movement learning model, and then uses the lightning occurrence point as the starting point to progress the extracted lightning. By applying the direction and speed, the path prediction information is generated by predicting the point where lightning moves.

By generating path prediction information as described above, the movement path of lightning can be predicted according to the current weather information, thereby improving the accuracy of identifying the location where lightning strikes next after lightning has occurred.

The lightning movement notification unit 311 extracts a solar power plant from the solar power plant information according to the path prediction information. That is, from the solar power plant information, solar power plants within a certain radius are extracted centered on the point where lightning stored in the path prediction information moves.

The lightning movement notification unit 311 transmits the route prediction information to the extracted solar power plant.

The lightning extinction prediction unit 312 predicts the life cycle of lightning using the current movement condition observation information based on the lightning movement learning model. Current weather information can also be used to predict the life cycle of the lightning strike.

The life cycle of the lightning can be predicted by determining whether the lightning is scheduled to dissipate through the barometric pressure valley of the current movement condition observation information.

The lightning extinction prediction unit 312 generates extinction prediction information when the lightning is scheduled to disappear and then transmits the extinction prediction information to the solar power plant 100 extracted by the path prediction information. By transmitting the extinction prediction information to the solar power plant 100 extracted based on the route prediction information, solar power generation of the relevant solar power plant 100 can be maintained, thereby preventing a decrease in power generation.

The method of preventing lightning damage to a solar power generator using the lightning damage prevention system of a solar power generator using artificial intelligence configured as described above is, as shown in FIG. 5, the meteorological service server 200 in the control server 300. Receive and collect past and current weather information (S10).

The control server 300 creates a lightning learning model by learning the lightning occurrence pattern using lightning learning data from past weather information corresponding to lightning occurrences among the past weather information, and based on the lightning learning model, the current weather information Determine whether lightning is likely to occur. If it is determined that lightning has occurred, a lightning prediction point is extracted according to the area of the current weather information and lightning prediction information including the lightning prediction point is generated (S20).

If it is not determined that a lightning strike has occurred, new current weather information is received and the possibility of a lightning strike is determined.

The control server 300 extracts the solar power plant 100 located within a certain radius based on the lightning prediction point from the location information of the solar power plant and transmits the lightning prediction information (S30).

The solar power plant 100 outputs a warning signal or an evacuation signal according to the lightning prediction information (S40).

That is, the solar power plant 100 outputs the evacuation signal when the location of the solar power plant is included in a risk section set in advance according to the lightning prediction point of the lightning prediction information of the solar power plant, and the risk section If the location of the solar power plant is not included, the warning signal is output.

When the evacuation signal is output, the solar power plant 100 cuts off the power to the inverter and the connection panel through the control unit 104, and when the warning signal is output, it monitors the power of the inverter and the connection panel without turning off the power (S50). .

When a lightning strike occurring within a preset radius is detected, the solar power plant 100 generates a lightning detection signal including the location of the detected lightning strike and transmits it to the control server 300 (S60).

The control server 300 receives a lightning detection signal from the solar power plant 100 that detects lightning, extracts the lightning occurrence point, and generates lightning occurrence information (S70).

The control server 300 receives and collects lightning observation information storing the location of lightning according to the past weather information from the Korea Meteorological Administration server 200 (S80).

The control server 300 collects the lightning observation information as moving learning data and then learns the lightning movement pattern of the location of lightning according to past weather conditions through the moving learning data to create a lightning movement learning model, Path prediction information that predicts the lightning movement path according to the current meteorological information is generated through the lightning movement learning model (S90).

The control server 300 extracts a solar power plant from the solar power plant information according to the route prediction information and transmits the route prediction information (S100).

If the location of the solar power plant 100 is included in a danger section previously set based on the lightning movement path of the path prediction information in the solar power plant 100 that has received the path prediction information, an evacuation signal is output, and the danger section is If the location of the solar power plant 100 is not included, a warning signal is output. Thereafter, the control unit 104 of the solar power plant 100 cuts off the power to the inverter and the connection panel when the evacuation signal is output, and does not turn off the power to the inverter and the connection panel when the warning signal is output (S110).

The lightning damage prevention system for solar power generators using artificial intelligence, controlled in the same way as above, predicts the lightning strike prediction point and the lightning movement path, so that solar power located at the lightning prediction point or the lightning movement path before lightning occurs. By allowing the power plant 100 to prepare for lightning strikes in advance, it is possible to secure time to prepare for lightning strikes and has the effect of significantly reducing breakdowns in solar power plants and casualties.

In addition, through lightning detection signals from multiple solar power plants that detect lightning, it is possible to accurately identify the location where the actual lightning occurred, greatly improving accuracy when predicting the movement path of lightning, and alerting solar power plants and nearby workers in advance. The probability of being protected from lightning can be greatly improved, and the lightning detection signal can help determine the intensity of lightning.

Meanwhile, by predicting the disappearance of lightning, unnecessary blocking in solar power plants can be prevented, thereby minimizing damage to solar power generation, which has the effect of operating solar power plants safely and efficiently.

The embodiments of the present invention described above and shown in the drawings should not be construed as limiting the technical idea of the present invention. The scope of protection of the present invention is limited only by the matters stated in the claims, and those skilled in the art can improve and change the technical idea of the present invention into various forms. Therefore, if such improvements and changes are obvious to those skilled in the art, they will fall within the scope of protection of the present invention.

100: Solar power plant 101: Lightning detection unit
102: Power Plant Communication Department 103: Hazard Warning Department
104: Control unit 200: Meteorological Administration server
300: Control server 301: Server communication department
302: Solar power plant DB department 303: Weather information collection department
304: Lightning prediction learning unit 305: Lightning monitoring unit
306: Lightning prediction and notification unit 307: Lightning occurrence extraction unit
308: Lightning observation collection unit 309: Lightning mobile learning unit
310: Movement path prediction unit 311: Lightning movement notification unit
312: Lightning extinction prediction unit

Claims (7)

Lightning learning data including past weather information corresponding to lightning strikes is collected from a plurality of solar power plants (100) that are located at a certain distance apart and generate a lightning detection signal by detecting lightning within a certain radius, and the Korea Meteorological Administration server (200). It includes a control server 300 that monitors whether the current weather information for each region corresponds to the lightning occurrence pattern through a lightning learning model that learned the lightning occurrence pattern, determines whether or not lightning is likely to occur in each region, and generates lightning forecast information. ,
The control server 300 includes a server communication unit 301 for communicating with the Meteorological Administration server 200 and the solar power plant 100, and a solar power plant in which solar power plant information including location information of the solar power plant is stored in advance. A power plant DB unit 302, a weather information collection unit 303 for receiving and storing past weather information for each region from the Korea Meteorological Administration server 200 and receiving current weather information for each region in real time, and a lightning strike occurrence among the past weather information. A lightning prediction learning unit 304 for collecting lightning learning data including past weather information corresponding to the lightning strike learning data and learning the lightning occurrence pattern corresponding to the lightning strike based on the lightning learning data to generate a lightning learning model, and the lightning strike A lightning monitoring unit 305 for monitoring whether the current weather information corresponds to the lightning occurrence pattern through a learning model, determining whether lightning is likely to occur, extracting a lightning prediction point, and generating lightning prediction information, and a lightning monitoring unit 305 for generating lightning prediction information, and When it is determined, a lightning prediction notification unit 306 for extracting solar power plants located within a certain radius based on the lightning prediction point through the solar power plant information and transmitting the lightning prediction information, and the lightning prediction information After transmitting, a lightning detection signal is received from the solar power plant 100 that detects lightning, a lightning occurrence extraction unit 307 extracts the lightning occurrence point and generates lightning occurrence information, and the Meteorological Administration server 200 A lightning observation collection unit 308 to receive and store lightning observation information that observes the location of lightning according to past weather information, and a lightning observation collection unit 308 to collect moving learning data including the lightning observation information, and to detect lightning based on the moving learning data. A lightning movement learning unit 309 for generating a lightning movement learning model by learning the lightning movement pattern for moving conditions, and a lightning movement learning model to determine the level of lightning according to the current weather information based on the lightning occurrence information. A movement path prediction unit 310 for predicting the movement path and generating route prediction information, and a lightning movement for extracting solar power plants according to the route prediction information through the solar power plant information and transmitting the route prediction information. Using the notification unit 311 and the lightning movement learning model, the life cycle of the lightning is predicted by determining whether the lightning is scheduled to dissipate, and when the lightning is scheduled to dissipate, extinction prediction information is generated and then the path prediction information is provided. It includes a lightning extinction prediction unit 312 for transmitting the extinction prediction information to the solar power plant 100 extracted by,
The solar power plant 100 includes a lightning detection unit 101 for detecting lightning that occurs within a preset radius, generating a lightning detection signal and transmitting it to the control server 300, and the control server 300. ), and when receiving the lightning prediction information or the path prediction information from the power plant communication unit 102 for communication with the control server 300, an evacuation signal or warning signal is sent according to the lightning prediction information or the path prediction information. A lightning damage prevention system for a solar power generator using artificial intelligence, including a hazard warning unit 103 for output, and a control unit 104 for cutting off power to the inverter and the connection panel when the evacuation signal or warning signal is output. delete delete delete delete delete delete

Images