KR20230165427A - Artificial intelligence fire spread prediction system - Google Patents

Abstract

The present invention relates to an artificial intelligence fire spread prediction system, which includes a complex sensor unit fixedly installed in the building to detect fire; An image detection unit that is fixedly installed in the building and detects fire using thermal images; A weather information collection unit that collects weather information; A database unit that stores building information, weather information, detection information from complex sensors, and thermal image information in a predetermined area; It includes a server that analyzes information collected from the complex sensor unit, image detection unit, and weather information collection unit to detect fire, analyze fire, predict the spread of fire, and provide the information to an external emergency situation system. In an embodiment of the present invention, accident-type chemicals can be read through smoke and flame images and complex sensors, and fire spread can be predicted using weather information such as building geographic information and wind direction (wind direction/speed) using artificial intelligence. You can.

Description

Artificial intelligence fire spread prediction system {Artificial intelligence fire spread prediction system}

The present invention relates to an artificial intelligence fire spread prediction system. More specifically, it reads accident-type chemicals through smoke and flame images and complex sensors, and detects fire spread using building geographic information and weather information such as wind. It is about an artificial intelligence fire spread prediction system that predicts.

Generally, when a fire occurs in a large-scale factory or large-scale industrial facility consisting of several buildings, members working in the large-scale industrial facility can recognize the fact of the fire through a fire alarm, but the exact location of the fire occurred and the It is difficult to accurately determine the situation, and it is also difficult to spread the fact and location of the fire and the situation. In other words, even if someone discovered the location of the fire and rang the emergency bell, it was impossible to accurately propagate the location of the fire in large-scale industrial plants due to the noise and the work being done in sections over long distances.

In particular, fires in factories that manufacture chemicals generate a lot of toxic gas, make it difficult to extinguish the fire, and cause significant damage to nearby residents, causing significant damage to life and property.

In addition, in the case of large-scale businesses engaged in the manufacture of base paper, corrugated cardboard, and boxes using waste paper, large and small fires frequently occur due to the collection, distribution, processing, and manufacturing of paper and dust, and the failure to respond initially to such fires is due to The spread of fire causes damage to life and property.

Meanwhile, in addition to factories, fire surveillance is also required in apartments, buildings, schools, ports, airports, water and sewage treatment plants, substations, heavy industrial plants, subways, chemical complexes, and steel mills.

The technical task to be achieved by the present invention is to improve the inconveniences of the prior art, by reading accident-type chemicals through smoke and flame images and complex sensors, and by using building geographic information and weather information such as wind with artificial intelligence. The goal is to provide an artificial intelligence fire spread prediction system that predicts the spread of fire.

In addition, the technical task to be achieved by the present invention is to improve the inconveniences of the prior art, by using time-series thermal image data and smoke data to select an AI learning model and to analyze abnormalities and problems of complex sensors. , which provides an artificial intelligence fire spread prediction system.

In addition, the technical problem that the present invention seeks to achieve is to improve the inconveniences of the prior art, using weather and chemical information to select an AI learning model, predict fire spread, and use artificial intelligence for fire extinguishing and evacuation. The goal is to provide an intelligent fire spread prediction system.

The artificial intelligence fire spread prediction system according to the characteristics of the present invention to solve these problems is,

An artificial intelligence fire spread prediction system that predicts the spread of fire in the event of a fire in a certain building and provides it to the outside,

A complex sensor unit fixedly installed in the building to detect fire;

An image detection unit that is fixedly installed in the building and detects fire using thermal images;

A weather information collection unit that collects weather information;

A database unit that stores building information, weather information, detection information from complex sensors, and thermal image information in a predetermined area;

It includes a server that analyzes information collected from the complex sensor unit, image detection unit, and weather information collection unit to detect fire, analyze fire, predict the spread of fire, and provide the information to an external emergency situation system.

The server may consist of a single server or multiple servers.

The complex sensor unit,

A carbon dioxide sensor that detects carbon dioxide, a dust sensor that detects dust, an internal flame sensor that detects flames inside the building, a temperature sensor that detects temperature, a hydrogen sensor that detects hydrogen, an external flame sensor that detects flames outside the building, At least one of the following: an electrical sensor that detects electricity including arcs, an ethylene sensor that detects ethylene, a fluorocarbon sensor that detects fluorocarbons, a sound sensor that detects sound, a moving body detection sensor that detects moving bodies, and a smoke sensor that detects smoke. Includes.

The database part,

Building information DB that stores building information in a given area;

Weather information DB that stores weather information;

Sensor detection information DB that stores detection information of complex sensors;

Thermal image information DB storing thermal image information;

Analysis result information DB storing fire analysis result information;

Includes a prediction information DB that stores fire spread prediction information.

The server is,

An information receiving unit that receives information collected from the complex sensor unit, image detection unit, and weather information collection unit;

a fire analysis unit that analyzes the information received from the information receiver to detect a fire and analyze the fire;

A spread prediction unit that predicts the spread of fire;

It includes a situation providing unit that provides the information analyzed by the fire analysis unit and the prediction information predicted by the spread prediction unit to an external emergency situation system.

The fire analysis department,

If ethylene or fluorocarbon is detected, it is judged to be a fire caused by chemicals.

The fire analysis department,

If the ethylene or fluorocarbons are not detected, it is judged to be a fire without chemical leakage.

The fire analysis department,

When a fire is detected in the image detection unit and the fire is not detected in the composite sensor unit,

It is determined that the composite sensor unit has failed.

The diffusion prediction unit,

Predict the direction and size of fire spread in response to the wind direction and amount of rain.

The diffusion prediction unit,

After a fire is detected, the fire pattern, weather information, and fire cause information that continues to spread or be extinguished are accumulated and stored.

It is used to predict the spread of fire by referring to the spread pattern of accumulated fire when a fire occurs.

The situation provider of the server reports the risk of suffocation, amount of smoke generated, amount of toxic substances generated, presence of people, movement speed and route of harmful substances, cause of accident, and risk level depending on the nature of chemicals generated during a fire. Provided to emergency situation system.

The situation provision department said,

If there are people at the fire scene, the primary evacuation route is provided to nearby terminals,

If there are people at the fire scene even after the evacuation time has elapsed, a rescue request message is transmitted to the external emergency system in a second way.

The situation provision department said,

If there are people inside the building at the fire scene, a rescue request message is transmitted to the external emergency system.

The situation provision unit of the server warns of the risk of suffocation when carbon dioxide and smoke increase.

The status providing unit of the server detects ethylene or fluorocarbon and warns of toxicity.

In an embodiment of the present invention, artificial intelligence fire spread is used to read accident-type chemicals through smoke and flame images and complex sensors, and to predict fire spread using building geographic information and weather information such as wind using artificial intelligence. A prediction system can be provided.

In addition, in an embodiment of the present invention, it is possible to provide an artificial intelligence fire spread prediction system that uses time-series thermal image data and smoke data to select an AI learning model and analyzes abnormalities and problems in complex sensors. there is.

In addition, in an embodiment of the present invention, an artificial intelligence fire spread prediction system can be provided that uses weather and chemical information to select an AI learning model, predicts fire spread, and uses it for fire extinguishment and evacuation.

Figure 1 is a configuration diagram of an artificial intelligence fire spread prediction system according to an embodiment of the present invention.
FIG. 2 is a configuration diagram of the composite sensor unit of FIG. 1.
FIG. 3 is a configuration diagram of the image detection unit of FIG. 1.
4 to 7 are examples of service screens provided by the artificial intelligence fire spread prediction system according to an embodiment of the present invention.

Below, with reference to the attached drawings, the present invention will be described in detail so that those skilled in the art can easily implement it. However, the present invention may be implemented in many different forms and is not limited to the examples described herein. In order to clearly explain the present invention in the drawings, parts that are not related to the description are omitted, and similar parts are given similar reference numerals throughout the specification.

Throughout the specification, when a part is said to “include” a certain element, this means that it may further include other elements rather than excluding other elements, unless specifically stated to the contrary.

Figure 1 is a configuration diagram of an artificial intelligence fire spread prediction system according to an embodiment of the present invention.

FIG. 2 is a configuration diagram of the composite sensor unit of FIG. 1.

FIG. 3 is a configuration diagram of the image detection unit of FIG. 1.

4 to 7 are examples of service screens provided by the artificial intelligence fire spread prediction system according to an embodiment of the present invention.

Referring to Figure 1, the artificial intelligence fire spread prediction system according to an embodiment of the present invention is,

An artificial intelligence fire spread prediction system (100) that predicts the spread of fire in the event of a fire in a predetermined building and provides it to the external emergency situation system (200),

A complex sensor unit 140 that is fixedly installed in the building to detect fire;

An image detection unit 130 that is fixedly installed in the building and detects fire using thermal images;

A weather information collection unit 150 that collects weather information;

A database unit 120 that stores building information, weather information, detection information from complex sensors, and thermal image information in a predetermined area;

Detects fire by analyzing information collected from the complex sensor unit 140, image detection unit 130, and weather information collection unit 150, analyzes the fire, and predicts the spread of the fire to provide an external emergency situation system. It includes a server 110 provided to 200.

The server 110 may consist of a single server or multiple servers.

Referring to Figure 2, the complex sensor unit 140,

A carbon dioxide sensor (141) that detects carbon dioxide, a dust sensor (142) that detects dust, an internal flame sensor (143) that detects flames inside the building, a temperature sensor (144) that detects temperature, and a hydrogen sensor that detects hydrogen. (145), an external flame sensor (146) that detects flames outside the building, an electrical sensor (147) that detects electricity including arcs, an ethylene sensor (148) that detects ethylene, a fluorocarbon sensor (148) that detects fluorocarbons ( 149), it includes at least one of a sound sensor 150 for detecting sound, a moving body detection sensor 151 for detecting a moving body, and a smoke sensor 152 for detecting smoke.

Referring to FIG. 3, the image detection unit 130 includes a general camera 131 for capturing general images and a thermal imaging camera 132 for capturing thermal images.

The database unit 120,

Building information DB 121 that stores building information in a predetermined area;

Weather information DB 122 that stores weather information;

Sensor detection information DB 123, which stores detection information of complex sensors;

Thermal image information DB 124 for storing thermal image information;

Analysis result information DB 125 for storing fire analysis result information;

It includes a prediction information DB 126 that stores fire spread prediction information.

The server 110,

An information receiver 111 that receives information collected from the complex sensor unit 140, the image detection unit 130, and the weather information collection unit 150;

a fire analysis unit 112 that detects a fire by analyzing the information received from the information reception unit 111 and analyzes the fire;

A spread prediction unit 113 that predicts the spread of fire;

It includes a situation providing unit 114 that provides the information analyzed by the fire analysis unit and the prediction information predicted by the spread prediction unit 113 to the external emergency situation system 200.

The fire analysis unit 112,

If ethylene or fluorocarbon is detected, it is judged to be a fire caused by chemicals.

The fire analysis unit 112,

If the ethylene or fluorocarbons are not detected, it is judged to be a fire without chemical leakage.

The fire analysis unit 112,

When a fire is detected in the image detection unit 130 and the fire is not detected in the composite sensor unit 140,

It is determined that the complex sensor unit 140 is broken.

The diffusion prediction unit 113,

Predict the direction and size of fire spread in response to the wind direction and amount of rain.

The diffusion prediction unit 113,

After a fire is detected, the fire pattern, weather information, and fire cause information that continues to spread or be extinguished are accumulated and stored.

It is used to predict the spread of fire by referring to the spread pattern of accumulated fire when a fire occurs.

The situation providing unit 114 of the server 110 determines the risk of suffocation, the amount of smoke generated, the amount of toxic substances generated, the presence of people, the movement speed and path of harmful substances, the cause of the accident, and the risk level, depending on the nature of the chemicals generated during the fire. is provided to the terminal of a person nearby or to the external emergency situation system (200).

The situation providing unit 114,

If there are people at the fire scene, the primary evacuation route is provided to nearby terminals,

If there are people at the fire scene even after the elapse of the time corresponding to the evacuation time, a rescue request message is transmitted to the external emergency situation system 200 in a second way.

The situation providing unit 114,

If there are people inside the building at the fire scene, a rescue request message is transmitted to the external emergency situation system 200.

The situation providing unit 114 of the server 110 warns of the risk of suffocation when carbon dioxide and smoke increase.

The situation providing unit 114 of the server 110 detects ethylene or fluorocarbon and warns of toxicity.

The operation of the artificial intelligence fire spread prediction system according to an embodiment of the present invention with this configuration is described as follows.

First, the complex sensor unit 140 detects various fire-related information inside or outside a certain building and transmits it to the server 110. Specifically, the complex sensor unit 140 includes a carbon dioxide sensor 141 that detects carbon dioxide, a dust sensor 142 that detects dust, an internal flame sensor 143 that detects a flame inside the building, and a temperature sensor that detects temperature. Sensor 144, hydrogen sensor 145 to detect hydrogen, external flame sensor 146 to detect flames outside the building, electric sensor 147 to detect electricity including arc, ethylene sensor 148 to detect ethylene ), the fluorocarbon sensor 149 to detect fluorocarbon, the sound sensor 150 to detect sound, the moving body detection sensor 151 to detect a moving body, and the smoke sensor 152 to detect smoke detect the status and output do.

Additionally, the weather information collection unit 150 collects weather information and transmits it to the server 110.

Then, the image detection unit 130 captures general images and thermal images and transmits them to the server 110.

Referring to Figure 4, it shows a comparison between actual CFD (simpleFOAM) and DeepCFD predictions, velocity components and pressure fields, and flow absolute errors around square-based geometry 1. Figure 4 is a comparison between actual and predicted for CFD (Computational Fluid Dynamics).

Then, the information reception unit 111 of the server 110 receives the information collected from the complex sensor unit 140, the image detection unit 130, and the weather information collection unit 160 and detects the sensor of the database unit 120. They are stored in the information DB 123, the thermal image information DB 124, and the weather information DB 122, respectively.

And the fire analysis unit 112 analyzes the information received from the information receiver 111 in real time or the sensor detection information DB 123, thermal image information DB 124, and weather information DB of the database unit 120. Collected information stored in (122) and building information DB (121) is retrieved and analyzed to determine whether a fire has occurred.

Specifically, the fire analysis unit 112 determines that the fire is caused by chemicals when ethylene or fluorocarbon is detected. In particular, in the case of a chemical production or storage plant, the complex sensor unit 140 may be additionally installed with a sensor that can detect the chemical or substances generated in the event of a chemical fire in response to building information.

In addition, the fire analysis unit 112 may determine that the fire is caused by the chemical substance when a substance generated during combustion of the chemical substance is detected by the corresponding sensor.

In addition, for example, in the case of a fire in a wood factory, if ethylene or fluorocarbon is not detected, the fire analysis unit 112 determines it to be a general fire without chemical leakage.

In particular, the fire analysis unit 112 of the server 110 determines the risk of suffocation, the amount of smoke generated, the amount of toxic substances generated, the presence of people, the movement speed and path of harmful substances, and the cause of the accident, depending on the nature of the chemical substances generated during the fire. And the risk is analyzed and stored in the analysis result information DB 125 of the database unit 120. Here, the risk is set at 100% when the amount of smoke and toxic substances generated exceeds the standard value, and when the amount of smoke and toxic substances generated is 10% less than the standard value, the risk is set at 90%. This expression method can be modified in various ways.

In addition, the fire analysis unit 112 accumulates and stores the time-series change pattern of composite sensor output before the fire occurs in the database, and later compares the accumulated fire occurrence pattern with the current pattern, and if they match more than 80%, the fire is detected. The occurrence of can also be predicted.

Additionally, if necessary, the fire analysis unit 112 may use LSTM & Autoencoder models specialized for time-series data processing, or may predict fire situations using GCN, a graph-based artificial intelligence model.

Additionally, the fire analysis unit 112 determines the type of chemical substance based on time-series changes caused by the image.

For example, the shape of the flame is different for each type of chemical, and the shape of the flame is also different for wood, which is not a chemical substance.

Therefore, the fire analysis unit 112 compares the form or time-series change pattern of the flame stored in advance in the database unit 120 with the form or time-series change of the flame received as a thermal image to determine whether or not the fire is caused by chemicals. The type of chemical substance can be identified.

If necessary, the fire analysis unit 112 can predict the occurrence of a fire using an algorithm such as YOLACT or MTCNN.

In addition, the fire analysis unit 112 accumulates and stores the time-series image change pattern before the fire occurs in the database, and later compares the accumulated and stored fire occurrence pattern with the current pattern to detect the occurrence of a fire if they match more than 80%. can also be predicted.

Meanwhile, when a fire is detected in the image detection unit 130 and a fire is not detected in the composite sensor unit 140, the fire analysis unit 112 determines that the composite sensor unit 140 is malfunctioning. do.

Thereafter, the fire analysis unit 112 stores the fire analysis result information in the analysis result information DB 125.

Then, the spread prediction unit 113 predicts the spread of the fire and refers to weather information. Specifically, the spread prediction unit 113 predicts the direction and size of fire spread in response to the wind direction and amount of rain.

In addition, the spread prediction unit 113 predicts the spread of fire and refers to information such as the type and amount of chemicals that caused the fire and the material of the building included in the building information.

Specifically, the diffusion prediction unit 113 predicts that in the case of highly flammable chemicals and highly flammable building materials, the diffusion speed is fast and the size of the diffusion increases, and that less flammable chemicals and building materials are flame retardant treated. In the case of the material, the diffusion speed is expected to be slow and the diffusion size to be reduced.

In addition, the diffusion prediction unit 113 predicts that when the wind blows faster than the reference value, the diffusion speed is fast and the size of the diffusion increases, and when the amount of rain is 5 mm or more per hour, the diffusion speed is slow and the size of the diffusion is reduced. It is predicted that this will happen, but for example, it can be predicted by subtracting 30% from the average diffusion rate.

In addition, the spread prediction unit 113 accumulates and stores the fire pattern and weather information and fire cause information that continuously spreads or is extinguished after the fire is detected,

It is used to predict the spread of fire by referring to the spread pattern of accumulated fire when a fire occurs.

Referring to Figure 5, a screen is provided where the manager can check or input physical property information and/or weather information. The shape of this screen can be modified in various ways.

Referring to Figure 6 or Figure 7, the direction and size of fire spread predicted by the spread prediction unit 113 are expressed, and this expression method can be modified in various ways.

Additionally, the spread prediction unit 113 stores fire spread prediction information in the prediction information DB 126.

Next, the situation providing unit 114 provides the information analyzed by the fire analysis unit 112 and the prediction information predicted by the spread prediction unit 113 to the emergency situation system 200 of a government office such as an external fire station or police station. do. Collected information may also be provided as needed. If necessary, summary evacuation information may be provided to the evacuee terminal 300. Additionally, the situation providing unit 114 may provide evacuation route information to the evacuee terminal 300 through a route with less smoke or toxic substances.

In particular, the situation providing unit 114 of the server 110 determines the risk of suffocation, the amount of smoke generated, the amount of toxic substances generated, the presence of people, the movement speed and path of harmful substances, and the cause of the accident, depending on the nature of the chemicals generated during the fire. and the risk level are provided to the terminals of nearby people or the external emergency situation system (200).

In addition, the situation providing unit 114,

If there are people at the fire scene, the primary evacuation route is provided to nearby terminals,

If there are people at the fire scene even after the elapse of the time corresponding to the evacuation time, a rescue request message is transmitted to the external emergency situation system 200 in a second way.

In addition, the situation providing unit 114,

If there are people inside the building at the fire scene, a rescue request message is transmitted to the external emergency situation system 200.

The situation providing unit 114 of the server 110 warns of the risk of suffocation when carbon dioxide and smoke increase.

In addition, the context providing unit 114 of the server 110 detects ethylene or fluorocarbon and warns of toxicity.

In the above embodiment of the present invention, accident-type chemicals can be read through smoke and flame images and complex sensors, and fire spread can be predicted using building geographic information and weather information such as wind using artificial intelligence.

Additionally, in an embodiment of the present invention, time-series thermal image data and smoke data can be used to select an AI learning model and analyze abnormalities and problems in the complex sensor.

Additionally, in an embodiment of the present invention, weather and chemical information can be used to select an AI learning model, predict fire spread, and use it for fire extinguishment and evacuation.

The embodiments of the present invention described above are not only implemented through devices and methods, but may also be implemented through a program that realizes the function corresponding to the configuration of the embodiment of the present invention or a recording medium on which the program is recorded. The implementation can be easily implemented by any expert in the technical field to which the present invention belongs based on the description of the embodiments described above.

Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements made by those skilled in the art using the basic concept of the present invention defined in the following claims are also possible. It falls within the scope of rights.

Claims (14)

An artificial intelligence fire spread prediction system that predicts the spread of fire in the event of a fire in a certain building and provides it to the outside,
A complex sensor unit fixedly installed in the building to detect fire;
An image detection unit that is fixedly installed in the building and detects fire using thermal images;
A weather information collection unit that collects weather information;
A database unit that stores building information, weather information, detection information from complex sensors, and thermal image information in a predetermined area;
Artificial intelligence that includes a server that analyzes information collected from the complex sensor unit, image detection unit, and weather information collection unit to detect fire, analyze fire, predict the spread of fire, and provide it to an external emergency situation system. Fire spread prediction system.
According to paragraph 1,
The complex sensor unit,
A carbon dioxide sensor that detects carbon dioxide, a dust sensor that detects dust, an internal flame sensor that detects flames inside the building, a temperature sensor that detects temperature, a hydrogen sensor that detects hydrogen, an external flame sensor that detects flames outside the building, At least one of the following: an electrical sensor that detects electricity including arcs, an ethylene sensor that detects ethylene, a fluorocarbon sensor that detects fluorocarbons, a sound sensor that detects sound, a moving body detection sensor that detects moving bodies, and a smoke sensor that detects smoke. An artificial intelligence fire spread prediction system including.
According to paragraph 2,
The database part,
Building information DB that stores building information in a given area;
Weather information DB that stores weather information;
Sensor detection information DB that stores detection information of complex sensors;
Thermal image information DB storing thermal image information;
Analysis result information DB storing fire analysis result information;
An artificial intelligence fire spread prediction system that includes a prediction information DB that stores fire spread prediction information.
According to paragraph 3,
The server is,
An information receiving unit that receives information collected from the complex sensor unit, image detection unit, and weather information collection unit;
a fire analysis unit that analyzes the information received from the information receiver to detect a fire and analyze the fire;
A spread prediction unit that predicts the spread of fire;
An artificial intelligence fire spread prediction system including a situation providing unit that provides the information analyzed by the fire analysis unit and the prediction information predicted by the spread prediction unit to an external emergency situation system.
According to paragraph 4,
The fire analysis department,
An artificial intelligence fire spread prediction system that determines a fire caused by chemicals when ethylene or fluorocarbon is detected.
According to clause 5,
The fire analysis department,
If the ethylene or fluorocarbon is not detected, an artificial intelligence fire spread prediction system that determines the fire to be without chemical leakage.
According to clause 6,
The fire analysis department,
When a fire is detected in the image detection unit and the fire is not detected in the composite sensor unit,
An artificial intelligence fire spread prediction system that determines the failure of the complex sensor unit.
In clause 7,
The diffusion prediction unit,
An artificial intelligence fire spread prediction system that predicts the direction and size of fire spread in response to the wind direction and amount of rain.
According to clause 8,
The diffusion prediction unit,
After a fire is detected, the fire pattern, weather information, and fire cause information that continues to spread or be extinguished are accumulated and stored.
An artificial intelligence fire spread prediction system used to predict the spread of fire by referring to the spread pattern of accumulated fire when a fire occurs.
According to clause 9,
The situation provider of the server reports the risk of suffocation, amount of smoke, amount of toxic substances, presence of people, movement speed and route of harmful substances, cause of accident, and risk level depending on the nature of chemicals generated during a fire. An artificial intelligence fire spread prediction system provided to emergency situation systems.
According to clause 10,
The situation provision department said,
If there are people at the fire scene, the primary evacuation route is provided to nearby terminals,
An artificial intelligence fire spread prediction system that sends a rescue request message to a secondary external emergency system if there are people at the fire scene even after the evacuation time has elapsed.
According to clause 11,
The situation provision department said,
An artificial intelligence fire spread prediction system that transmits a rescue request message to an external emergency system when there are people inside the building at the fire site.
According to clause 12,
The situation provision unit of the server is an artificial intelligence fire spread prediction system that warns of the risk of suffocation when carbon dioxide and smoke increase.
According to clause 13,
The situation provision unit of the server is an artificial intelligence fire spread prediction system that detects ethylene or fluorocarbon and warns of toxicity.

Images