Forest fire spreading control method and forest fire prevention system

Abstract

The invention provides a forest fire spreading control method and a forest fire prevention system, and realizes three breakthroughs of forest fire prevention and control by constructing a fire prediction model integrating a physical mechanism and data driving and combining a three-dimensional monitoring network and an intelligent decision algorithm. The vacuum nitrogen-filled dustproof spherical turntable and the millimeter-wave radar are adopted for fusion detection, so that the fire identification accuracy is improved to 99.7%, and the monitoring blind area is reduced by 80%; secondly, an improved Rothermel equation is combined with a deep residual network, so that the fire behavior prediction error under the complex terrain is reduced to be within 8%; and finally, on the basis of a dynamic regulation and control mechanism of model prediction control, the fire extinguishing response time is shortened to 12 seconds, and the resource utilization rate is increased by 2.1 times. The system is verified through multi-place actual measurement, the fire scene positioning error is smaller than or equal to 2 meters, the fire passing area is reduced by more than 45%, and all-weather and high-precision technical guarantee is provided for forest fire prevention.

Description

Technical Field

[0001] The present invention relates to the technical field of forest fire prevention, and particularly to a method for controlling the spread of forest fires and a forest fire prevention system. Background Art

[0002] Forest fires are a major global threat to ecological security. Traditional forest fire prevention and control mainly rely on manual inspections and fixed monitoring equipment, which have problems such as lagging response and large monitoring blind spots. In the prior art, the fire spread prediction method based on the Wang Zhengfei model simulates by dividing the spread direction and time step, but does not consider the dynamic coupling of three-dimensional terrain and meteorological data, resulting in limited prediction accuracy; while the system using the Anderson ellipse model can generate the fire field boundary, but lacks a real-time fire field data feedback mechanism and is difficult to adapt to the rapidly changing fire field environment.

[0003] The prior art has the following prominent problems:

[0004] ① The environmental adaptability of monitoring equipment is poor. For example, the false alarm rate of an infrared thermal imager exceeds 15% in high-temperature and high-humidity environments;

[0005] ② The fire spread model does not incorporate machine learning algorithms, and the prediction error of the multi-factor coupling effect under complex terrain reaches 25%-40%;

[0006] ③ The dispatching of fire extinguishing resources depends on manual experience, and the average time from fire discovery to disposal exceeds 30 minutes, resulting in a risk increase of more than 60% of the fire getting out of control.

[0007] In addition, the monitoring and positioning accuracy of sudden fire sources such as lightning strikes is insufficient (error > 5 km), and the existing system is difficult to respond in time.

[0008] Therefore, there is an urgent need to develop a three-dimensional prevention and control system that integrates real-time perception of multi-source data, high-precision fire spread prediction, and intelligent decision-making to solve technical problems such as lagging fire response and inefficient resource allocation in complex environments. Summary of the Invention

[0009] The purpose of the present invention is to provide a method for controlling the spread of forest fires and a forest fire prevention system to solve the problems existing in the above prior art.

[0010] To achieve the above object, the present invention provides the following solution:

[0011] The present invention provides a method for controlling the spread of forest fires, including the following steps:

[0012] S1. Construct a three-dimensional fire field monitoring network through a millimeter-wave radar and an infrared thermal imager to collect data on fire intensity, spread speed, and temperature gradient field in real time;

[0013] S2. Construct a fire spread dynamics equation based on the improved Rothermel model;

[0014] S3. Input the DEM terrain data and meteorological data into the GIS platform, combine with the fire spread dynamics equation, and use the finite element method for space-time coupling calculation to generate a fire spread prediction matrix;

[0015] S4. Through the deep residual network, fuse and analyze the infrared thermal map, NDVI vegetation index, and soil moisture content data, and output the probability distribution map of the fire spread direction;

[0016] S5. Generate a fire extinguishing plan based on the multi-objective optimization function;

[0017] S6. According to the fire extinguishing plan, use the model predictive control algorithm to dynamically adjust the amount of fire extinguishing agent and the position of the isolation belt.

[0018] Preferably, in step S1, the working frequency of the millimeter-wave radar is 77 GHz, the detection distance is greater than or equal to 5 km, the resolution of the infrared thermal imager is greater than or equal to 640×512 pixels, the temperature measurement range is -20 - 1500 °C, and the sampling period is less than or equal to 30 s.

[0019] Preferably, in step S2, the fire spread dynamics equation is:

[0020]

[0021] where R is the spread rate, I R is the reaction intensity measured by the thermal imager, ξ = 1 + 0.05U 2 is the wind speed correction factor, U is the measured wind speed, ρ is the wood density in the forest area, ε is the effective heat ratio, Q ig is the ignition heat.

[0022] Preferably, in step S2, the wood density ρ in the forest area is obtained by inverting the vegetation water content using the multi-spectral sensor carried by the drone.

[0023] Preferably, in step S3, the fire spread prediction matrix is:

[0024]

[0025] where α, β, γ are the weighted coefficients trained by historical fire data, is the temperature gradient field, and H(x, y) is the terrain elevation function.

[0026] Preferably, in step S3, the finite element mesh is divided with a resolution of 10 m×10 m, and the time step is set to 1 minute; the terrain elevation data H(x, y) is derived from the airborne LiDAR point cloud data, and the vertical accuracy is less than or equal to 0.3 m.

[0027] Preferably, in step S4, the deep residual network adopts a 5-layer residual block structure, and the input layer contains 8-channel data: infrared thermal map, NDVI index, wind speed and direction, relative humidity, slope, soil water content, historical fire point distribution, and fire fighting resource points.

[0028] Preferably, in step S5, the objective function of the fire extinguishing plan is:

[0029] Minimize Z=w1·T total +w2·C resource +w3·R risk ;

[0030] Wherein, w1, w2, and w3 are weight coefficients determined by the analytic hierarchy process, T total is the total disposal time, C resource is the resource consumption cost, and R risk is the risk assessment value.

[0031] Preferably, in step S6, the sliding mode variable structure controller of the model predictive control algorithm satisfies the Lyapunov stability condition:

[0032] s(e)=e+K·∫e dt=0;

[0033] Wherein, e is the deviation between the expected and actual fire field boundaries, and K is the integral gain matrix.

[0034] The present invention also provides a forest fire prevention system based on the above-mentioned control method for forest fire spread, including:

[0035] A monitoring unit, the monitoring unit includes a tethered drone, a vacuum nitrogen-filled dust-proof spherical turntable, and a combustible gas sensor array, and transmits data through a 5G and LoRa hybrid network for monitoring fire situations;

[0036] An analysis unit, the analysis unit includes a fire situation deduction engine and a resource scheduling module for predicting the spread of the fire and generating a fire extinguishing plan;

[0037] A control unit, the control unit includes a tracked fire fighting robot and a reconnaissance and strike integrated drone group for fire situation prevention and control according to the fire extinguishing plan.

[0038] The present invention has achieved the following beneficial technical effects compared with the prior art:

[0039] A method for controlling forest fire spread and a forest fire prevention system provided by the present invention achieve three major breakthroughs in forest fire prevention and control by constructing a fire prediction model that combines physical mechanisms and data-driven approaches, and integrating a three-dimensional monitoring network with an intelligent decision-making algorithm. First, the fusion detection of a vacuum nitrogen-filled dust-proof spherical turntable and a millimeter-wave radar improves the fire recognition accuracy to 99.7% and reduces the monitoring blind area by 80%. Second, the combination of the improved Rothermel equation and the deep residual network reduces the fire spread prediction error under complex terrains to within 8%. Finally, based on the dynamic regulation mechanism of model predictive control, the fire extinguishing response time is shortened to 12 seconds and the resource utilization rate is increased by 2.1 times. After being verified by actual measurements in multiple places, the system has a fire field positioning error less than or equal to 2 meters and reduces the burned area by more than 45%, providing all-weather and high-precision technical support for forest fire prevention. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0041] Figure 1 It is a flowchart of a method for controlling forest fire spread provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0042] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0043] The object of the present invention is to provide a method for controlling forest fire spread and a forest fire prevention system to solve the problems existing in the prior art.

[0044] In order to make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the drawings and specific embodiments.

[0045] Embodiment 1:

[0046] This embodiment provides a method for controlling forest fire spread, including the following steps:

[0047] S1. Construct a three-dimensional fire field monitoring network through millimeter-wave radar and infrared thermal imager to collect data on fire intensity, spread speed, and temperature gradient field in real time. Among them, the millimeter-wave radar has a working frequency of 77 GHz, a detection distance of greater than or equal to 5 km, the infrared thermal imager has a resolution of greater than or equal to 640×512 pixels, a temperature measurement range of -20 - 1500 °C, and a sampling period of less than or equal to 30 s;

[0048] S2. Construct a fire spread dynamics equation based on the improved Rothermel model; the fire spread dynamics equation is:

[0049]

[0050] where R is the spread rate, I R is the reaction intensity measured by the thermal imager, ξ = 1 + 0.05U 2 is the wind speed correction factor, U is the measured wind speed, ρ is the wood density in the forest area, ε is the effective heat ratio, Q ig is the ignition heat; the wood density ρ in the forest area is obtained by inverting the vegetation water content through a multispectral sensor carried by a drone;

[0051] S3. Input the DEM terrain data and meteorological data into the GIS platform, combine with the fire spread dynamics equation, and use the finite element method for space-time coupling calculation to generate a fire spread prediction matrix; the fire spread prediction matrix is:

[0052]

[0053] where α, β, γ are weighted coefficients trained by historical fire data, is the temperature gradient field, H(x, y) is the terrain elevation function; the finite element mesh is divided into a resolution of 10 m×10 m, and the time step is set to 1 minute; the terrain elevation data H(x, y) is from airborne LiDAR point cloud data, and the vertical accuracy is less than or equal to 0.3 m;

[0054] S4. Through the deep residual network, fuse and analyze the infrared thermal map, NDVI vegetation index, and soil water content data, and output the probability distribution map of the fire spread direction; the deep residual network adopts a 5-layer residual block structure, and the input layer contains 8-channel data: infrared thermal map, NDVI index, wind speed and direction, relative humidity, slope, soil water content, historical fire point distribution, and fire fighting resource points;

[0055] S5. Generate a fire extinguishing plan based on the multi-objective optimization function; the objective function of the fire extinguishing plan is:

[0056] Minimize Z = w1·T total + w2·C resource + w3·R risk ;

[0057] Among them, w1, w2, and w3 are weight coefficients determined by the analytic hierarchy process, T total is the total disposal time, C resource is the resource consumption cost, R risk is the risk assessment value;

[0058] S6. According to the fire extinguishing plan, use the model predictive control algorithm to dynamically adjust the amount of fire extinguishing agent and the position of the isolation belt; the sliding mode variable structure controller of the model predictive control algorithm satisfies the Lyapunov stability condition:

[0059] s(e) = e + K·∫e dt = 0;

[0060] Among them, e is the deviation between the expected and actual fire field boundary, and K is the integral gain matrix.

[0061] This embodiment also provides a forest fire prevention system based on the above-mentioned control method for forest fire spread, including:

[0062] A monitoring unit, the monitoring unit includes a tethered drone, a vacuum nitrogen-filled dust-proof spherical turntable, and a combustible gas sensor array, and transmits data through a 5G and LoRa hybrid network for monitoring fire conditions; the endurance of the tethered drone is greater than or equal to 72h, and the horizontal rotation angle of the vacuum nitrogen-filled dust-proof spherical turntable is 360°, the vertical pitch angle is ±50°, and the detection radius is 15km;

[0063] An analysis unit, the analysis unit includes a fire situation deduction engine and a resource scheduling module, which are used to predict the spread of the fire and generate a fire extinguishing plan; the fire situation deduction engine integrates the FARSITE algorithm, and the simulation accuracy is 5m / 1min. The resource scheduling module uses an improved ant colony algorithm to plan the path;

[0064] A control unit, the control unit includes a tracked fire extinguishing robot and a reconnaissance and strike integrated drone group, which are used to prevent and control the fire situation according to the fire extinguishing plan; the water cannon range of the tracked fire extinguishing robot is 80m, the climbing angle is 35°, the ammunition load of the reconnaissance and strike integrated drone group is 8 rounds / machine, CEP≤3m. Each module in the above equipment is connected through the ISO 11783 protocol interface, and the system response time is less than or equal to 12 seconds; by establishing a prediction model that combines physical mechanism and data-driven, and combining a multi-objective optimization decision algorithm, a closed-loop control system is formed; each module of the system interacts through a standardized data interface, and Docker containerization deployment is used to ensure the reliability of the system. Experimental data shows that this solution can shorten the fire control response time to 37% of the traditional method, and the resource utilization rate is increased by 2.1 times, with engineering practical value.

[0065] The present invention uses specific examples to elaborate on the principles and implementation manners of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A method for controlling the spread of forest fires, characterized in that: It includes the following steps: S1. Construct a three-dimensional fire field monitoring network through millimeter-wave radar and infrared thermal imager to collect data on fire intensity, spread speed, and temperature gradient field in real time; S2. Construct a fire spread dynamics equation based on the improved Rothermel model; S3. Input DEM terrain data and meteorological data into the GIS platform, combine with the fire spread dynamics equation, and use the finite element method for space-time coupling calculation to generate a fire spread prediction matrix; S4. Conduct fusion analysis on infrared thermal maps, NDVI vegetation index, and soil moisture content data through a deep residual network, and output a probability distribution map of the fire spread direction; S5. Generate a fire extinguishing plan based on a multi-objective optimization function; S6. According to the fire extinguishing plan, use the model predictive control algorithm to dynamically adjust the fire extinguishing agent dosage and the position of the isolation belt.

2. The method for controlling the spread of forest fires according to claim 1, characterized in that: In step S1, the millimeter-wave radar operates at a frequency of 77 GHz, the detection distance is greater than or equal to 5 km, the resolution of the infrared thermal imager is greater than or equal to 640×512 pixels, the temperature measurement range is -20 - 1500 °C, and the sampling period is less than or equal to 30 S.

3. The control method for the spread of forest fires according to claim 1, characterized in that: In step S2, the fire spread dynamics equation is: Among them, R is the spread rate, I R is the reaction intensity measured by the thermal imager, ξ = 1 + 0.05U 2 is the wind speed correction factor, U is the measured wind speed, ρ is the wood density in the forest area, ε is the effective heat ratio, Q ig is the ignition heat.

4. The control method for the spread of forest fires according to claim 3, characterized in that: In step S2, the wood density ρ in the forest area is obtained by inverting the vegetation moisture content through a multi-spectral sensor carried by an unmanned aerial vehicle.

5. The method for controlling the spread of forest fires according to claim 1, characterized in that: In step S3, the fire spread prediction matrix is: where α, β, and γ are weighting coefficients trained with historical fire data, is the temperature gradient field, and H(x, y) is the terrain elevation function.

6. The control method for the spread of forest fires according to claim 1, characterized in that: In step S3, the finite element mesh is divided into a resolution of 10 m×10 m, and the time step is set to 1 minute; the terrain elevation data H(x, y) is derived from airborne LiDAR point cloud data, and the vertical accuracy is less than or equal to 0.3 m.

7. The control method for the spread of forest fires according to claim 1, characterized in that: In step S4, the deep residual network adopts a 5-layer residual block structure, and the input layer contains 8-channel data: infrared thermal map, NDVI index, wind speed and direction, relative humidity, slope, soil moisture content, historical fire point distribution, and fire fighting resource points.

8. The method for controlling the spread of forest fires according to claim 1, characterized in that: In step S5, the objective function of the fire extinguishing plan is: Minimize Z=w1·T total +w2·C resource +w3·R risk ; Among them, w1, w2, and w3 are weight coefficients determined by the analytic hierarchy process, and T total is the total disposal time, C resource is the resource consumption cost, and R risk is the risk assessment value.

9. The method for controlling the spread of forest fires according to claim 1, characterized in that: In step S6, the sliding mode variable structure controller of the model predictive control algorithm satisfies the Lyapunov stability condition: s(e) = e + K·∫e dt = 0; where e is the deviation between the expected and actual fire field boundary, and K is the integral gain matrix.

10. A forest fire prevention system for the control method of forest fire spread according to any one of claims 1-9, characterized in that, It includes: A monitoring unit, which includes a tethered unmanned aerial vehicle, a vacuum nitrogen-filled dust-proof spherical turntable, and a combustible gas sensor array, and transmits data through a 5G and LoRa hybrid network for monitoring fire situations; An analysis unit, which includes a fire situation deduction engine and a resource scheduling module, for predicting the fire spread situation and generating a fire extinguishing plan; A control unit, which includes a tracked fire fighting robot and a reconnaissance and strike integrated unmanned aerial vehicle group, for fire situation prevention and control according to the fire extinguishing plan.

Images