KR102668287B1 - Monitorable electric vehicle charging station fire prevention distribution panel and the operating method - Google Patents

Abstract

The present invention relates to a fire prevention distribution panel for electric vehicle charging stations that can be monitored. More specifically, the monitoring involves monitoring a plurality of charging stations in real time with a thermal imaging camera and immediately activating an alarm device when an abnormality occurs to minimize safety risks. This is about a possible electric vehicle charging station fire prevention distribution board.

Description

Monitorable electric vehicle charging station fire prevention distribution panel and the operating method}

The present invention relates to a fire prevention distribution panel for electric vehicle charging stations that can be monitored. More specifically, the monitoring involves monitoring multiple charging stations in real time with a thermal imaging camera and immediately activating an alarm device when an abnormality occurs to minimize safety risks. This is about a possible electric vehicle charging station fire prevention distribution board.

In general, the risk of fire occurring at electric vehicle charging stations is directly related to the ever-increasing penetration rate of electric vehicles.

Battery overheating due to high-speed charging, battery overheating, defects in charging station electrical facilities, arcs occurring at charger contact points, and wear of charging ports and plugs are considered to be some of the main causes of this fire risk.

Various efforts are being made to prevent fires at electric vehicle charging stations. Examples include the development of non-conductive water-based fire extinguishers, distribution of chargers with safety devices, attachment of overcharge prevention devices, and improvement of systems for fire response.

In particular, overheating and deterioration occurring in distribution boards can affect electric vehicles, charging stations, and even large public buildings.

Existing monitoring methods do not effectively solve these problems.

The present invention detects problems in real time and responds immediately by integrating a thermal imaging camera with a tilting function, a data collection device, an AI function unit, a tilting controller, and an alarm device through an advanced algorithm of the distribution board.

Therefore, the present invention can present a new paradigm that increases the safety of electric vehicle charging stations.

The present invention was created to improve the above-described problems. The present invention monitors a plurality of electric vehicle chargers in real time through a thermal imaging camera in the distribution board, and immediately activates an alarm device when an abnormality occurs, thereby minimizing safety risks. The purpose is to provide a fire prevention distribution panel for electric vehicle charging stations that can be monitored.

In order to achieve the above object, an embodiment of the present invention includes a distribution panel 10 that is responsible for supplying power and providing stable energy to other components; A thermal imaging camera 20 coupled to the distribution panel 10 to obtain thermal image information, including a thermal imaging camera, photographs the external charging part of the charging station 80 where a plurality of thermal imaging cameras are installed, and is installed within the distribution panel 10. A digital earth leakage circuit breaker that controls power supply, photographs the exterior of the charging station 80 where a plurality of thermal imaging cameras are installed, and automatically distinguishes between earth leakage and abnormal current in controlling the power supply within the distribution panel 10. (15) maintains the power supply from the distribution board (10) during abnormal currents that do not damage equipment or power supply, and stops power only in case of an electric leak, and the distribution board (10) operates together with the digital earth leakage circuit breaker (15). Therefore, when a temperature exceeding a preset value occurs in the electric vehicle charging unit or an abnormal current that interferes with power supply is detected, an alarm is transmitted remotely, and the distribution board 10 is equipped with the thermal imaging camera 20 and the digital earth leakage circuit breaker 15. ), the processor 50 analyzes the collected thermal image information and abnormal current information and provides a function to remotely transmit the alarm.

Additionally, a distribution panel that is responsible for power supply and provides stable energy to other components; A tiltable camera capable of acquiring areas and thermal information for analysis, and shooting from multiple angles; In order to analyze the data collected from the data collection device and make a decision, the AI includes an AI function that improves performance over time using machine learning algorithms; A data collection device designed to monitor image information of the distribution board in real time; and a control command to be provided to the AI function unit to block the distribution board before a defect causes damage or causes a safety hazard by detecting a fire early through the AI function unit. In the direction in which the AI function unit moves the camera through learning, the tracking data converted by the trajectory input function by the tracking input module is sent to the tilting controller, and the tilting controller adjusts the camera movement direction based on the tracking data. When the movement is completed, the camera is operated from the adjusted position, and the tilting controller installed to enable rotation within the distribution panel has the function of measuring the temperature of a specific area. It monitors risk factors with multiple cameras and provides a voice alarm. , the AI function unit utilizes a trajectory input function that receives trajectory information from a camera and monitors only the temperature corresponding to the trajectory according to the input trajectory.

The thermal imaging camera 20 of the distribution board 10 measures external heat of charging stations connected in parallel.

The distribution board 10 includes a tiltable camera 20 that acquires image information of the neighboring charging station 80 and is capable of taking pictures from various angles; It includes an AI function unit 40 that analyzes image data collected from the tiltable camera 20 to determine whether there is a fire, and detects fire early through the AI function unit 40 to prevent defects from being damaged. The AI function unit 40 gives a control command to block the distribution board before it causes damage or causes a safety hazard.

In the present invention, a plurality of thermal imaging cameras of a monitorable electric vehicle charging fire prevention distribution board can measure detailed heat distribution by tilting within a certain range to photograph the exterior of the distribution board to which a plurality of cameras are connected and attached.

When an abnormality outside the normal value range is discovered in the information of the data collection device, an exponential moving average or exponentially weighted moving average is used as an alarm technique, and an alarm is generated only when the recent change in abnormal value is outside the normal value range.

The present invention includes the step of requesting analysis of passion information data from the data collection device to the AI function unit; Fire detection confirmation step by analysis of thermal information data of AI function department; Distribution panel blocking execution step by AI function unit control command; The AI function transmits tracking data to the tilting controller; A step of adjusting the movement direction of the tiltable camera of the tilting controller by the AI function unit; Monitoring including the alarm generation step of the alarm device by the AI function unit is possible.

The AI function unit further includes monitoring only the temperature of the corresponding trajectory according to the input trajectory using the trajectory input function.

The AI function unit operates using a principal component analysis method.

The AI function probabilistically evaluates the possibility of fire by combining statistical methods and machine learning.

Through a configuration in which the distribution board, the thermal imaging camera, the digital earth leakage circuit breaker, the remote power control function, and the charger are organically connected, the present invention achieves the following effects.

First, it accurately detects abnormal currents such as surges and induced lightning generated from lightning to prevent malfunctions in charging stations and vehicles.

Second, the remote power control function allows you to conveniently manage the power of electric chargers installed every quarter and provide immediate response in emergency situations.

Third, the cost of replacing the charger and purchasing additional parts is cheaper than other products, reducing the economic burden.

Fourth, in case of fever, the vehicle owner or charging station is immediately notified through an alarm, and efficient proactive measures are possible by monitoring the detected thermal image information in real time.

Fifth, prevent safety accidents and ensure uninterrupted power supply.

The present invention has structural and technical differentiation that simultaneously improves the safety, efficiency, and economic feasibility of charging stations through the various effects described above.

Figure 1 is a diagram showing an example of a fire at a charging station according to a conventional invention.
Figure 2 is a block diagram schematically showing the relationship between the distribution board of an electric vehicle charging station fire prevention distribution board capable of monitoring according to an embodiment of the present invention, the charging station, and a thermal imaging camera.
Figure 3 is a block diagram schematically showing the relationship between the AI function unit and the tilting controller according to an embodiment of the present invention.
Figure 4 is a block diagram schematically showing the relationship between components implemented through the AI function unit according to an embodiment of the present invention.
Figure 5 is a diagram showing the information transfer process during tilting movement according to an embodiment of the present invention.
Figure 6 is a diagram showing a process of adjusting the direction of movement using tracking data according to an embodiment of the present invention.
Figure 7 is a diagram showing the process of blocking when an alarm is generated according to an embodiment of the present invention.
Figure 8 is a diagram sequentially showing the stages of fire detection and subsequent alarm generation according to an embodiment of the present invention.

The present invention as described above will be described in detail through the attached drawings and examples.

It should be noted that the technical terms used in the present invention are only used to describe specific embodiments and are not intended to limit the present invention. In addition, the technical terms used in the present invention, unless specifically defined in a different sense in the present invention, should be interpreted as meanings generally understood by those skilled in the art in the technical field to which the present invention pertains, and are not overly comprehensive. It should not be interpreted in a literal or excessively reduced sense. Additionally, if the technical term used in the present invention is an incorrect technical term that does not accurately express the idea of the present invention, it should be replaced with a technical term that can be correctly understood by a person skilled in the art. In addition, general terms used in the present invention should be interpreted according to the definition in the dictionary or according to the context, and should not be interpreted in an excessively reduced sense.

Additionally, as used in the present invention, singular expressions include plural expressions unless the context clearly dictates otherwise. In the present invention, terms such as “consists of” or “comprises” should not be construed as necessarily including all of the various components or steps described in the invention, and some of the components or steps are included. It may not be possible, or it should be interpreted as including additional components or steps.

Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the attached drawings, but identical or similar components will be assigned the same reference numbers regardless of the reference numerals, and duplicate descriptions thereof will be omitted.

Additionally, when describing the present invention, if it is determined that a detailed description of related known technologies may obscure the gist of the present invention, the detailed description will be omitted. In addition, it should be noted that the attached drawings are only intended to facilitate easy understanding of the spirit of the present invention, and should not be construed as limiting the spirit of the present invention by the attached drawings.

Figure 2 is a block diagram schematically showing the relationship between the distribution board of the electric vehicle charging station fire prevention distribution board capable of monitoring according to an embodiment of the present invention, the charging station, and the thermal imaging camera, and Figure 3 is an AI function according to an embodiment of the present invention. It is a block diagram schematically showing the relationship between the unit and the tilting controller, and Figure 4 is a block diagram schematically showing the relationship between components implemented through the AI function unit according to an embodiment of the present invention.

As shown in Figure 2, a distribution panel 10 that is responsible for power supply and provides stable energy to other components; And a thermal imaging camera 20 coupled to the distribution board 10 to acquire thermal image information.

The thermal imaging camera captures the exterior of the charging station 80, where a plurality of thermal imaging cameras are installed, controls the power supply within the distribution board 10, and the digital earth leakage circuit breaker 15, which automatically blocks earth leakage, detects abnormal current and supplies power. stop,

The distribution board 10 operates together with the digital earth leakage circuit breaker 15 to remotely send an alarm when it detects a temperature or abnormal current above a preset value,

The distribution board 10 includes a processor that performs real-time monitoring by integrating information from the thermal imaging camera 20 and the digital earth leakage circuit breaker 15,

The processor 50 analyzes collected thermal image information and abnormal current information and provides a function to remotely send an alarm or reset the power.

The present invention is constructed by organically connecting a distribution board, a thermal imaging camera, a digital earth leakage circuit breaker, a remote power control function, and a charger. These components work in the following way:

The distribution board plays a central role in all electrical circuits and performs the function of immediately blocking an abnormal current. In particular, when a surge or induced lightning occurs due to lightning, it detects it and prevents malfunctions in the electric charger and vehicle.

Thermal imaging cameras continuously monitor the heating status of the charger and vehicle. When heat is detected, this information is immediately notified to the car owner or charging station, and has the function of monitoring the detected thermal image information in real time.

A digital earth leakage circuit breaker prevents malfunctions caused by abnormal currents other than earth leakage. This function specifically prevents the charger from stopping due to malfunction of the earth leakage circuit breaker due to abnormal current input during the charging process.

The remote power control feature allows users to remotely control the power of quarterly installed electric chargers. This allows immediate response even in emergency situations.

As shown in Figures 3 and 4, the present invention monitors a plurality of distribution boards 10 in real time through sensors, etc., and immediately activates an alarm device when an abnormal phenomenon occurs to minimize the risk of fire.

In cases where blind spots occur due to pillars or inner walls, such as in an apartment underground parking lot, a (thermal imaging) camera can be attached to the outside of the distribution panel 10 or the side of the pillar and electrically connected to the controller to prevent fire.

Looking at sensors, representative examples include tiltable cameras and (thermal imaging) cameras. In addition, the present invention integrates various sensors to detect fire or abnormal situations and provide sensors to effectively manage them.

For example, a temperature sensor and thermometer attached to the inside or outside of the distribution board 10 continuously monitor changes in external temperature.

In particular, using infrared sensors, high temperatures can be quickly detected even in the early stages of a fire.

When the infrared sensors exceed a fixed threshold, they send a signal to the AI function unit 40, and the AI function unit 40 analyzes this and adjusts the direction of the tiltable camera or activates an alarm device.

The combination of external and internal sensors increases the accuracy of fire detection.

In other words, the external sensor detects the temperature of the environment, and the internal sensor detects the temperature of the device itself.

For example, an outlier can be determined by comparing two types of data from an external sensor and an internal sensor. This outlier judgment algorithm predicts abnormal data through exponential smoothing.

As another example, optical sensors and cameras attached to the inside or outside of the distribution board 10 collect visual information.

The visual information is analyzed together with the image information in the AI function unit 40 to take action appropriate for the specific situation. That is, when the optical sensor detects a flame, the AI function unit 40 determines it to be a fire and activates the alarm device.

In addition, biosensors and biological sensors attached to the inside or outside of the distribution board 10 are particularly useful in ensuring the safety of people flocking to a large charging station.

Biosensors and biological sensors monitor people's biometric data in real time and can quickly detect, for example, a sudden increase in heart rate due to surprise from a fire or other abnormal symptoms.

Motion sensors, microphones, and sound sensors attached to the inside or outside of the distribution board 10 are also particularly useful for detecting abnormal situations such as fires caused by intrusions or explosions caused by natural fires. They detect abnormalities in sound or movement and send a signal to the AI function unit (40).

In addition, CCTV, vision sensors, and image capture modules attached to the inside or outside of the distribution board 10 or to the ceiling of the building provide high-quality image data.

Image data provides detailed data, especially when the AI function unit 40 analyzes image information, thereby increasing the accuracy of analysis.

As described above, the sensors used in the present invention each perform a special function and role, and the sensors are organically connected to quickly and accurately detect and respond to complex abnormal situations.

Looking specifically, the distribution board 10 plays a central role in supplying power to peripheral devices (electric vehicle charging stations).

In the present invention, the external heat of the distribution board 10 is measured by a tiltable camera, and the heat information is transmitted to a data collection device.

For example, the distribution board 10 is a key component of power supply and is connected to various electric circuits.

In the present invention, the external heat of the distribution panel 10 is measured in real time by a tiltable camera installed on the side facing each other.

At this time, a thermal imaging camera can be used as a method of measuring heat, but by taking advantage of the fact that the camera's image sensor is sensitive to a certain range of heat in addition to light, the reaction rate of the camera sensor to heat can be determined through learning (experiment). And, using this data, an artificial intelligence algorithm can be applied that converts the color data of commonly photographed images into heat information data.

The measured heat information is transmitted to the data collection device, where initial heat information data is processed and stored.

In this process, the tracking data transmitted to the AI function unit 40 is used to detect fire, external defects, or other abnormalities early.

In addition, a real image camera and a thermal image camera can be used simultaneously, and a pair of real image cameras can be installed to measure depth, making it easy to identify external defects.

The tiltable camera 20 measures the external heat of the distribution panel 10 and is adjusted by the rotatable tilting controller 50, which is based on tracking data processed by the AI function unit 40.

The tiltable camera 20 precisely measures the external heat of the distribution panel 10 through a high-resolution thermal sensor and lens.

The tiltable camera 20 can be adjusted to various angles and directions by a rotatable tilting controller.

The tilting controller utilizes a step motor and gear mechanism to finely adjust the camera's position.

The AI function unit 40 tracks abnormal temperatures or other risk factors of the distribution board 10 in real time through a tracking input module.

This tracking data is transmitted to the tilting controller, which appropriately adjusts the camera's movement direction and angle.

This process enables detection and response to abnormalities within a short period of time, resulting in increased safety.

Additionally, the tiltable camera can monitor various distribution panels 10 simultaneously, contributing to efficient deployment of resources. This reduces maintenance costs and improves the overall efficiency of the system.

These features of the tiltable camera, combined with the advanced algorithms of the AI function unit 40, can quickly and accurately detect a variety of hazards beyond simple heat measurement. This enables highly reliable monitoring and provides a better experience for users.

The data collection device 30 processes real-time image information received from a tiltable camera. Real-time image information is transmitted to the AI function unit 40 and analyzed, and the thermal information data and tracking data are predicted using exponential smoothing.

For example, the data collection device 30 serves to process real-time image information received from a tiltable camera. The data collection device 30 has high-speed data processing capabilities and sufficient storage space. Image information collected in real time goes through an initial filtering process to remove unnecessary data. Afterwards, the filtered data is transmitted to the AI function unit 40.

The AI function unit 40 analyzes the thermal information data and determines the status of the neighboring distribution board 10. In particular, when an abnormality such as fire, deterioration, or overheating is detected, a control command (tilting command through tracking data) to immediately block the distribution board 10 is executed. This rapid response can greatly reduce risk by immediately identifying an accident or disaster.

The data collection device 30 also predicts abnormal data from the thermal information data using the exponential smoothing method.

Exponential smoothing allows for more accurate predictions by assigning higher weight to recent data (thermal information data, tracking data, etc.).

Through predictive methods, it is possible to predict future risk factors beyond simple abnormal data. When the predicted abnormal data exceeds a certain threshold, the alarm device operates and immediately generates an alarm.

The configuration and operating principle of this data collection device 30 greatly contributes to real-time monitoring and minimizing safety risks, which are the purposes of the present invention.

Additionally, data analysis and prediction capabilities using artificial intelligence algorithms ensure efficient resource deployment and high accuracy.

The AI function unit 40 analyzes image information received from the data collection device 30. A fire is detected early and the corresponding control command is executed on the distribution board (10). Additionally, exponential smoothing is used to predict abnormal data.

For example, the AI function unit 40 deeply analyzes image information received from the data collection device 30.

In the in-depth analysis process, machine learning algorithms and pattern recognition technology that learn by quantifying image information are used to detect abnormalities such as fire at an early stage.

The AI function unit 40 immediately executes a control command to the distribution board 10 in response to the detected fire.

The control command includes safety measures such as shutting off the distribution board 10, thereby preventing potential damage or dangerous situations. This rapid response contributes to efficient resource deployment and reduced operating costs.

In another embodiment, the tiltable camera is used to measure the external heat of the distribution panel 10. The camera measures the surface temperature of the distribution panel 10 through an infrared sensor and converts the surface temperature information into digital data (thermal information data, etc.).

The data collection device 30 receives digital data, such as surface temperature information, and stores and analyzes it in real time.

The data collection device 30 also collects image information input from the camera, in addition to the external heat of the distribution board 10. This collected information is delivered to the AI function unit 40, where a fire detection algorithm is applied.

The tilting controller (50) adjusts the moving direction of the camera based on tracking data received from the AI function unit (40). Through this, temperature-related image information of a specific area is accurately measured.

For example, due to the advanced functionality and efficiency of the tilting controller, the tilting controller of the present invention uses tracking data received from the AI function unit 40 to quickly and accurately adjust the moving direction of the camera.

In the process of adjustment, the AI function unit 40 utilizes the trajectory input function by the tracking input module (templated tracking information) through learning.

This trajectory input function continuously monitors temperature changes in specific areas and immediately activates an alarm device when an abnormality is detected. This function of the tilting controller allows effective monitoring of various parts of the distribution panel 10, increasing safety and reducing maintenance costs.

Here, the trajectory input function refers to an algorithm that is learned by the AI function unit 40 and continuously monitors temperature changes in a specific area.

The monitoring algorithm immediately and automatically activates an alarm device when the temperature of a specific area exceeds or falls below a certain standard, improving the safety and maintenance efficiency of the distribution board (10).

The tracking input module refers to a component that processes tracking data received from the AI function unit 40 and provides parameters necessary for the trajectory input function.

The tracking input module provides information to the tilting controller to quickly and accurately adjust the camera's direction of movement, and through this, it accurately measures temperature-related image information of a specific area.

Here, the trajectory input function is a function in which the AI function unit 40 continuously monitors temperature changes in a specific area based on data obtained through learning.

The monitoring function serves to immediately activate an alarm device when an abnormality is detected.

The tracking input module added to the monitoring function is a part that receives tracking data from the AI function unit 40 and processes it.

The tracking input module is essential for quickly and accurately adjusting the camera's movement direction using the trajectory input function.

Therefore, the trajectory input function and tracking input module allow the tilting controller to effectively monitor various parts of the distribution panel 10, resulting in increased safety and reduced maintenance costs.

In addition, the tilting controller goes beyond simply controlling the camera's direction of movement and provides real-time tracking data analysis and the camera's quick response capabilities.

The alarm device 60 operates according to the results processed by the AI function unit 40. When an abnormality is detected, a voice alarm is immediately issued.

The alarm device 60 is directly connected to the AI function unit 40.

The AI function unit 40 analyzes the thermal information data from the distribution board 10 in real time and detects abnormalities or fires at an early stage.

When the abnormal phenomenon is detected, the AI function unit 40 immediately transmits a signal to the alarm device. The alarm device receives this signal and generates a voice alarm, which is designed so that people can respond immediately.

The operating principle of this alarm device 60 directly contributes to the purpose and effect of the present invention.

Due to the quick analysis of the AI function unit 40 and the immediate alarm generation of the alarm device, potential risks can be quickly recognized and appropriate measures can be taken. Second, this alarm device increases reliability through high-accuracy thermal information data analysis. Since the analysis results through the advanced algorithm of the AI function unit 40 are directly transmitted to the alarm device, the possibility of false alarms is minimized.

Therefore, the alarm device plays an important role in the present invention and provides a structure that monitors safety conditions in real time and responds immediately when an abnormality occurs.

As another embodiment, the present invention collects various data (heat information data, etc.) from a fire detection camera, temperature sensor, smoke detector, fire smell detector, and fire sound detector and uses it for fire detection.

Sensors, including fire detection cameras, temperature sensors, smoke detectors, fire smell detectors, and fire sound detectors, have unique characteristics and detection capabilities and can detect outliers in different variables.

The data collection process takes place in real time and is networked for fast and accurate detection.

For example, in the process of determining key independent variables, the collected data is analyzed.

Data such as temperature, smoke, and fire smell can be determined as key independent variables, which are classified through machine learning algorithms.

The independent variable has excellent scalability and reusability, and can be used in other models such as fire diagnosis models.

In the diagnostic model building stage, the AI function unit 40 creates a fire detection model based on key independent variables.

The fire detection model combines statistical methods and machine learning to probabilistically evaluate the possibility of fire.

In particular, an optimized model is selected through various algorithms and parameter adjustments.

In the abnormality detection stage, the AI function unit 40 detects data abnormalities using the built fire detection model.

If the temperature exceeds a certain threshold or smoke and fire odor data show outliers, a notification mechanism is activated.

The abnormality detection of the AI function unit 40 uses an exponential moving average or exponentially weighted moving average to detect the amount of change in the abnormal value.

The electric vehicle charging station fire prevention distribution board (including an AI function that uses thermal information data and tracking data), which can be monitored through the process of detecting changes in abnormal values, performs fast and accurate fire detection and analysis.

Each sensor information (data) included in the AI function unit 40 and the models and algorithms using it are organically connected to each other, and the system is capable of simultaneously monitoring and adjusting the model in real time for each data and the model that analyzed it.

Therefore, the present invention ensures high efficiency and stability in preventing and responding to fires in the distribution board (10).

Figure 5 is a diagram showing a tilting movement process according to an embodiment of the present invention, Figure 6 is a diagram showing a process of adjusting the movement direction with tracking data according to an embodiment of the present invention, etc., and Figure 7 is a diagram showing a tilting movement process according to an embodiment of the present invention. It is a diagram showing the process of executing blocking by generating an alarm according to an embodiment of the present invention, and FIG. 8 is a diagram sequentially showing the stages of fire detection and the corresponding alarm generation according to an embodiment of the present invention.

As shown in Figure 5, the present invention includes an external heat generation step (S101, S102), a heat distribution measurement step (S103, S104), a tilting movement command step (S105, S106), and a detailed heat distribution measurement step (S107, S108). ), etc.

If a fire is detected in step S102, the fire in panel 1 (11) is detected in panel 2 (12), and the fire information in panel 1 (11) is collected as much as possible through steps S104, S106, and S108.

As shown in Figure 6, the present invention includes an external heat measurement step (S201), a real-time image information transmission step (S202), an image information analysis request step (S203), a fire or external defect detection step (S204), and a tracking data transmission step. It consists of (S205), movement direction adjustment step (S206), temperature-related image information measurement step (S207), alarm generation step (S208), and abnormal data prediction step (exponential smoothing method) (S209).

As shown in Figure 7, the present invention includes an image information analysis request step (S303), an early fire risk detection step (S304), a tracking data transmission step (S306), a movement direction adjustment step (S307), and temperature-related image information measurement. It consists of a step (S308), an alarm generation step (S309), and a blocking execution step (S310).

As shown in Figure 8, the present invention includes a step (S403) in which the data collection device 30 requests the AI function unit to analyze thermal information data; Fire detection confirmation step (S404) by analyzing the fire information data of the AI function unit 40; The AI function unit 40 executes blocking the distribution board 10 by a control command (S405); The AI function unit 40 transmits tracking data to the tilting controller 50 (S406); Adjusting the movement direction of the tiltable camera 20 of the tilting controller 50 by the AI function unit (S407); Alarm generation step (S408) of the alarm device 60 by the AI function unit; It consists of etc.

Specifically, the monitorable electric vehicle charging station fire prevention distribution panel 10 according to the present invention integrates various sensors to ensure high accuracy and quick response.

Fire detection cameras, temperature sensors, smoke detectors, fire smell detectors, and fire sound detectors ensure diversity of sensors and capture various characteristics of fire.

Fire detection data collection is mainly carried out through a sensor array, and these sensors transmit information to the data collection device 30 in real time.

Temperature, smoke, and fire odor data stored in the data collection device 30 are generally selected as main independent variables, and these variables serve to increase the accuracy of fire detection.

After determining key independent variables, a fire diagnosis model is constructed.

The fire diagnosis model uses machine learning algorithms to determine whether a fire is detected based on key independent variables.

The selected algorithm is optimized through cross-validation on the dataset, which improves the accuracy of the fire diagnosis model.

The fire diagnosis model of the AI function unit 40 analyzes sensor data stored in the data collection device 30 in real time.

In other words, in situations where temperature, smoke, and fire smell data are set as key independent variables, the anomaly detection algorithm quickly analyzes the patterns of these variables.

When an abnormality is detected, the notification system of the AI function unit 40 immediately notifies the manager or person in charge and, if necessary, activates an automatic fire extinguishing system.

The notification system through the notification device 60 transmits fire information through several communication channels.

Examples include text messages, emails, and voice notifications, which enable rapid response.

The notification system is organically connected to the diagnostic model and immediately generates a notification when an abnormality is detected.

The automatic fire extinguishing system inside the distribution board 10 is connected to the AI function unit 40 and operates immediately when abnormality detection is confirmed.

At this time, the fire extinguishing agent is sprayed appropriately according to the nature and location of the fire, and through this, the initial suppression of the fire is achieved.

In addition, the monitorable electric vehicle charging station fire prevention distribution board (10) maintains optimal performance through continuous updates and supplementation.

For example, sensor performance improvement, algorithm update, database expansion, etc. are progressed in the direction of increasing the stability and efficiency of the AI function unit 40.

The AI function unit 40 is an integrated solution for early detection and response to fire, and provides high safety through the diversity of sensors, accuracy of algorithms, and speed of the notification system.

The AI function unit 40 for fire detection and tracking integrates multi-step algorithms and learning mechanisms. In the initial stage, the trajectory data collected by the tracking input module is transmitted to the tilting controller. The tilting controller adjusts the moving direction of the camera based on trajectory data. Accordingly, the camera moves and photographs the target area.

It measures image information related to the temperature of a specific area learned through a rotatable tilting controller. Multiple cameras continuously monitor hazards. The AI function unit 40 monitors only the temperature according to the monitored trajectory.

The data collection device 30 predicts abnormal data that is outside the normal value range of image information using exponential smoothing. By using an exponential moving average or exponentially weighted moving average, an alarm is generated in the alarm device only when the change in recent abnormal values is outside the normal value range.

The algorithms used by the AI function unit 40 according to the present invention include exponential smoothing, exponential moving average, and exponentially weighted moving average, which are prediction mechanisms.

In its simplest form, exponential smoothing calculates the exponentially weighted average of historical data. The exponential moving average develops this method by placing greater weight on the most recent data. Exponentially weighted moving averages take this further and consider weights for multiple variables simultaneously.

The exponential smoothing method is expressed by the following formula:

The exponential moving average is as follows.

Exponentially weighted moving averages are more complex and involve multiple weights that can adjust the weights of multiple variables.

These predictive mechanisms play an important role in fire detection. When temperature or image data outside the normal range is detected, an alarm is immediately triggered in the alarm device. Therefore, the AI function unit 40 is capable of early detection of fire and immediate response.

Principal Component Analysis (PCA), a function included in the AI function unit 40, is a method of reducing high-dimensional data to low-dimensional data. PCA is a method of finding principal components that maximize the variance of data, and can be implemented using Equation 3 below.

(Where,

That is, , represents the principal component that maximizes the variance of the data, and S is the variance matrix of the principal component, representing the variance of the principal component.

is the inverse matrix of the main component, and is used when restoring the main component to the original data.

Once the main independent variables are determined through principal component analysis, a fire detection model can be built using Equation 4 below.

(Here, y is whether there is a fire, w is the weight of the fire detection model, and x is the vector of the main independent variables)

After the AI function unit 40 builds a fire detection model, it can detect anomalies. Here, “abnormality” means a state that deviates from the normal state. Using a fire detection model, input data can be analyzed to determine cases outside the normal range as abnormal.

The following Equation 5 can be used for anomaly detection.

here, is the predicted value of the fire detection model, If is outside the normal range, it can be judged as abnormal.

In addition to the formula presented above, the AI function unit 40 can use an algorithm using the following formula.

If the distribution of data (thermal information data, tracking data, etc.) is different with the normalization algorithm, the data can be standardized using normalization.

As described above, normalizing data can achieve the following effects.

Since the data or distribution of data from each sensor collected by the AI function unit 40 becomes constant, it becomes easy to compare the sizes of different data, and the AI function unit 40 can improve the stability of model learning. and can improve the generalization performance of the model.

Linear regression is a method of modeling a linear relationship between two variables, but real-world data often does not follow a linear relationship. In these cases, polynomial regression can be used to model nonlinear relationships.

As an example, ensemble learning applied to the algorithm of the AI function unit 40 is a method of improving performance by combining several models.

The ensemble learning includes a method of learning multiple models using different datasets (bagging), a method of learning the next model to compensate for the errors of the previous model (boosting), and a method of combining decision tree models (random Forest).

It is expected that the accuracy and efficiency of the fire detection method can be improved by adding random forest formulas, etc.

Specifically, introducing principal component analysis (PCA) into the AI function unit 40 of the fire prevention distribution panel 10 of the electric vehicle charging station capable of monitoring according to the present invention is an effective method for reducing the data dimension and extracting key variables. am.

Principal component analysis reduces computational complexity by reducing the dimensionality of the original high-dimensional data, leaving only the main information.

In the data collection stage of the data collection device 30, information received from a plurality of (thermal imaging) cameras, CCTVs, or sensors is composed of multidimensional feature vectors. Each dimension represents a different variable, such as temperature, smoke concentration, or humidity of the space.

The AI function unit 40 finds the direction in which the variance is greatest through a process of rotating the multidimensional feature vector around the origin.

The principal components found in this way contain the largest amount of information in the original multidimensional data space.

The AI function unit 40 maps the original data to a new low-dimensional space using the principal components obtained after the data rotation process.

The AI function unit 40 predicts the possibility of fire occurrence using only the main components in the new low-dimensional space. A probabilistic classification model can be used here.

For example, the AI function unit 40 uses a classifier such as a support vector machine (SVM) to use data in a low-dimensional space (thermal information data, tracking data, fire detection camera angle data, and temperature sensors, smoke detectors, Data from fire smell detectors, fire sound detectors, etc.) are classified into fire occurrence and non-occurrence. Classifier learning uses a supervised learning method and requires labeled training data.

The AI function unit 40 can quickly analyze data using an efficient algorithm after dimension reduction through principal component analysis. Through this, the fire prevention distribution panel (10) of the electric vehicle charging station that can be monitored ensures a quick response time and minimizes inaccurate alarms.

Since the fire prediction model is created based on principal components, the AI function unit 40 of the present invention periodically updates these principal components. These updates are made by continuously collecting and analyzing new data.

This method is applied to the AI function unit 40, enabling early detection and rapid response to fire.

In fact, using data whose dimensions have been reduced through principal component analysis, quick and accurate fire prediction is possible by removing unnecessary information and analyzing only important characteristics.

The AI function unit 40 using principal component analysis realizes quick and accurate fire prediction through dimensionality reduction and extraction of important variables. This greatly contributes to safety by enabling early detection and rapid response to fire.

10: Distribution board
20: Tiltable camera
30: data collection device
40: AI function part
50: Tilting controller
60: Alarm device

Claims (8)

A step (S401) in which the data collection device 30 requests the AI function to analyze the thermal information data of the charging station 80;
Fire detection confirmation step (S402) by analyzing the fire information data of the AI function unit 40;
The AI function unit 40 executes blocking the distribution board 10 by a control command (S403);
The AI function unit 40 transmits tracking data to the tilting controller 50 (S404);
Adjusting the movement direction of the tiltable camera 20 of the tilting controller 50 by the AI function unit (S405);
A method of operating a fire prevention distribution panel at an electric vehicle charging station capable of monitoring, including an alarm generation step (S406) of the alarm device 60 by the AI function unit.
In claim 1,
The AI function unit monitors only the temperature of the charging station 80 corresponding to the trajectory according to the input trajectory using the trajectory input function. A method of operating a fire prevention distribution panel at an electric vehicle charging station capable of monitoring, further comprising:
In claim 1,
A method of operating a fire prevention distribution panel at an electric vehicle charging station capable of monitoring in which the AI function unit 40 operates using a principal component analysis method.
In claim 1,
The AI function unit 40 is a method of operating a fire prevention distribution panel at an electric vehicle charging station capable of monitoring that probabilistically evaluates the possibility of fire by combining statistical methods and machine learning. delete delete delete delete