A gas pollution source detection and tracing system

Abstract

The application discloses a kind of gas pollution source detection traceability system, it is related to gas pollution source instantaneous discharge traceability detection technical field;The present application determines target position by extracting the abnormal position in traceability signal, determines the detection route with target position as starting point, matches unmanned aerial vehicle for each route to collect pollutant concentration and collect environmental data, then the compensation amount is corrected pollutant concentration by matching collected environmental data, concentration distribution map is constructed and diffusion main shaft and diffusion time are determined, and the distance of emission source is calculated to lock the suspected position, finally, the real pollution source is obtained by verifying the suspected source through CFD model;In the technical scheme, the pollution area can be covered without dead angle by multi-unmanned aerial vehicle and other angle partition detection, and the concentration data is ensured to be complete and continuous;CFD model verification greatly improves the positioning accuracy of pollution source, shortens the response time of traceability in the instantaneous discharge scene, reduces the positioning error, provides key technical support for rapid disposal of leakage and reduction of pollution diffusion.

Description

Technical Field

[0001] This invention belongs to the field of gas pollution source instantaneous emission tracing and detection technology, specifically a gas pollution source detection and tracing system. Background Technology

[0002] In industrial production, oil and gas storage and transportation, and hazardous waste treatment, the instantaneous emission of gaseous pollutants (referred to as "instantaneous emissions") is a key focus and challenge in air pollution control. These emissions are characterized by their suddenness, duration, and high intensity. Examples include benzene leaks caused by accidental valve ruptures in storage tanks, the release of toxic gases due to illegal dumping by hazardous chemical transport vehicles, and instantaneous venting from overpressure reactors. The released gaseous pollutants can quickly create localized high-concentration pollution zones, posing a serious threat to the surrounding environment, human health, and public safety.

[0003] Existing gaseous pollution source tracing technologies are mainly built around the "continuous emission" scenario, relying on two core approaches: First, fixed monitoring point networks. Fixed monitoring points have limitations in coverage and deployment costs. The distance between adjacent monitoring points is often several kilometers, making it difficult to capture the rapid diffusion process of instantaneous emissions. When instantaneous emissions occur, pollutants may spread to sensitive areas before the monitoring points have completed data collection and transmission, or data may be missing due to blind spots in the monitoring points, making it impossible to support source tracing analysis. Second, the advance deployment of mobile monitoring equipment. While this approach can compensate for the lack of flexibility of fixed monitoring points, it requires advance planning of the monitoring range and time. The suddenness of instantaneous emissions is often unpredictable, making it difficult to deploy monitoring equipment to the pollution-affected area in a timely manner.

[0004] This invention provides a gas pollution source detection and tracing system to solve the above-mentioned technical problems. Summary of the Invention

[0005] This invention aims to address at least one of the technical problems existing in the prior art. To this end, this invention proposes a gaseous pollution source detection and tracing system. This technical solution solves the problems of limited coverage of traditional fixed monitoring points and the difficulty of emergency response with mobile equipment. Multiple drones and other methods can perform angular zoning detection to achieve comprehensive coverage of polluted areas without blind spots, ensuring complete and continuous concentration data. Real-time environmental data correction ensures the accuracy of pollutant concentrations and avoids tracing deviations caused by sensor interference from temperature and humidity. CFD model verification significantly improves the accuracy of pollution source location, shortening the tracing response time and reducing location errors in instantaneous discharge scenarios, providing key technical support for rapid leak response and reducing pollution spread.

[0006] To achieve the above objectives, a first aspect of the present invention provides a gas pollution source detection and tracing system, including a source detection and tracing module and several unmanned aerial vehicles connected thereto;

[0007] Source tracing control module: used to receive source tracing signals, plan detection routes in several directions based on the abnormal locations in the source tracing signals; control several drones to detect pollutant concentrations along the detection routes; and,

[0008] It is used to correct pollutant concentrations based on collected environmental data, analyze pollutant concentrations to estimate the suspected location of gaseous pollution sources, and simulate and determine the location of pollution sources based on the suspected locations of gaseous pollution sources.

[0009] In one possible implementation, detection routes in several directions are planned based on the anomaly locations in the source signal, including:

[0010] Extract the abnormal locations from the source signal and use them as target locations;

[0011] Based on the target location, at least three drones are identified; using the target location as the starting point, several detection areas are determined at equal angles on the horizontal plane, and detection routes are planned for each of the several detection areas.

[0012] In one possible implementation, several drones are controlled to detect pollutant concentrations along a detection route, including:

[0013] Associate at least one drone with each detection route;

[0014] The drone is controlled to detect the concentration of gaseous pollutants along an associated detection route; the drone is equipped with sensors including gas analysis sensors and a miniature weather station.

[0015] In one possible implementation, pollutant concentrations are corrected based on collected environmental data, including:

[0016] Identify the time of pollutant concentration collection and extract the corresponding environmental data at that time; the environmental data is collected using a miniature weather station mounted on a drone.

[0017] The pollutant concentration is corrected by matching the compensation amount of the sensors carried by the UAV with the collected environmental data.

[0018] In one possible implementation, analyzing pollutant concentrations to estimate the suspected location of a gaseous pollution source includes:

[0019] Concentration distribution maps are constructed based on pollutant concentrations, and diffusion axes and diffusion times are determined using these maps. The distance to emission sources is then estimated based on diffusion time and diffusion coefficients.

[0020] The suspected location of the gaseous pollution source was calculated based on the diffusion axis and the distance to the emission source.

[0021] In one possible implementation, the location of the pollution source is determined based on a simulation of the suspected location of the gaseous pollution source, including:

[0022] The target area is determined with the suspected location as the center and a preset distance as the radius; several suspected pollution sources are selected in the target area; the preset distance is set according to the terrain complexity or wind field stability.

[0023] Several suspected emission sources were verified using a CFD model to determine the optimal emission source; the optimal emission source was then used as the gaseous pollution source, and its corresponding location was designated as the pollution source location.

[0024] In one possible implementation, before estimating the suspected location of a gaseous pollution source, pollutant concentrations are corrected based on historical meteorological data, including:

[0025] Historical meteorological data is extracted from ground meteorological stations, and the historical meteorological data is used to determine whether pollutant concentration correction is needed; if so, the correction range for pollutant concentration is determined based on the historical meteorological data.

[0026] Historical meteorological data and intelligent correction models are combined to obtain the corrected pollutant concentrations; the intelligent correction model is built based on an artificial intelligence model.

[0027] In one possible implementation, the intelligent correction model is built upon an artificial intelligence model, including:

[0028] Construct artificial intelligence models; these models include BP neural network models or RBF neural network models.

[0029] Extract the training dataset and use it to train the artificial intelligence model to obtain the intelligent correction model. The training dataset includes meteorological environmental data, diffusion characteristic data, and concentration deviation. The meteorological environmental data is used to represent extreme meteorological scenarios that affect pollutant concentrations, and the diffusion characteristic data includes the diffusion time and diffusion distance after being affected by extreme meteorological scenarios.

[0030] In one possible implementation, the correction range for pollutant concentrations is determined based on historical meteorological data, including:

[0031] Analysis determines the spatial regions affected by historical meteorological data;

[0032] By combining historical meteorological data, the spatial region is determined within the current mapping region and used as the correction region; the data range covered by the correction region is used as the correction range.

[0033] In one possible implementation, after determining the location of the gaseous pollution source, the corresponding pollution area is determined based on that location, including:

[0034] The simulation covers the area of ​​gaseous pollutants emitted instantaneously from a gaseous pollution source, and extracts the area of ​​influence from the coverage area based on a set concentration threshold; the gaseous pollutants in the area of ​​influence will cause adverse effects.

[0035] Drones are used to verify and adjust the boundaries of the affected area, and the pollution area of ​​the gas pollution source is determined based on the adjustment results.

[0036] Compared with the prior art, the beneficial effects of the present invention are:

[0037] 1. This invention determines the target location by extracting anomalies from the source tracing signal, plans a detection route starting from the target location, and assigns drones to each route to collect pollutant concentrations and environmental data. The pollutant concentration is then corrected by matching compensation amounts to the collected environmental data, constructing a concentration distribution map, determining the diffusion axis and diffusion time, and estimating the distance to the emission source to pinpoint suspected locations. Finally, the suspected source is verified through a CFD model to obtain the actual pollution source. This technical solution solves the problems of limited coverage of traditional fixed monitoring points and the difficulty of emergency response with mobile equipment. Multiple drones and other angled zoning detection can achieve comprehensive coverage of the polluted area without blind spots, ensuring complete and continuous concentration data. Real-time environmental data correction ensures the accuracy of pollutant concentrations and avoids source tracing deviations caused by sensor interference from temperature and humidity. CFD model verification significantly improves the accuracy of pollution source location, shortening the source tracing response time and reducing location errors in instantaneous discharge scenarios, providing key technical support for rapid leak handling and reducing pollution spread.

[0038] 2. This invention employs a concentration correction mechanism based on historical meteorological data and an intelligent correction model. It extracts historical data from ground-based meteorological stations to determine the need for correction, analyzes the spatial regions affected by extreme weather events and maps the current correction range, and then uses an artificial intelligence model to perform secondary optimization of pollutant concentrations. This technical solution addresses the shortcoming of traditional detection methods that neglect historical extreme weather interference. Historical meteorological data allows for tracing the impact of extreme weather on pollutant diffusion before detection, avoiding data distortion caused by heavy rain or high temperatures. The intelligent correction model, trained on multi-dimensional data, can accurately quantify concentration deviations under extreme scenarios. Precise division of the correction range avoids data redundancy caused by blind correction, ensuring the accuracy of subsequent distance calculations between the diffusion axis and emission sources, and effectively guaranteeing the reliability of instantaneous emission source tracing under complex meteorological conditions. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a schematic diagram of the system principle of the gas pollution source detection and tracing system in an embodiment of the present invention;

[0041] Figure 2 This is a schematic diagram illustrating the principle of planning inspection routes based on target locations in an embodiment of the present invention. Figure 1 ;

[0042] Figure 3 This is a schematic diagram illustrating the principle of planning inspection routes based on target locations in an embodiment of the present invention. Figure 2 ;

[0043] Figure 4 This is a schematic diagram illustrating the principle of determining a gaseous pollution source based on a suspected pollution source in an embodiment of the present invention;

[0044] Figure 5 This is a schematic diagram illustrating the principle of how extreme weather affects pollutant concentrations in an embodiment of the present invention;

[0045] Figure 6 This is a schematic diagram illustrating the principle of determining the pollution area corresponding to the gas pollution source in an embodiment of the present invention. Detailed Implementation

[0046] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Example 1:

[0048] Please see Figure 1 The first aspect of the present invention provides a gas pollution source detection and tracing system, including a source detection and tracing module and several drones connected thereto;

[0049] Source tracing control module: used to receive source tracing signals, plan detection routes in several directions based on the abnormal locations in the source tracing signals; control several drones to detect pollutant concentrations along the detection routes; and,

[0050] It is used to correct pollutant concentrations based on collected environmental data, analyze pollutant concentrations to estimate the suspected location of gaseous pollution sources, and simulate and determine the location of pollution sources based on the suspected locations of gaseous pollution sources.

[0051] The source tracing and detection module is the core device for UAV control and data analysis. It can wirelessly communicate and transmit data with several UAVs, and can also analyze the data detected by the UAVs through built-in algorithms. UAVs can interact with each other to collaboratively complete tasks. The system also includes a data storage module for storing data.

[0052] In a preferred embodiment, a detection route in several directions is planned based on the abnormal location in the source tracing signal, including: extracting the abnormal location in the source tracing signal as the target location; determining at least three drones based on the target location; using the target location as the starting point, determining several detection areas at equal angles on the horizontal plane, and planning detection routes in the several detection areas respectively.

[0053] The generation of source tracing signals primarily relies on public feedback, enterprise emergency alarms, remote sensing hotspots, and data monitoring and early warning systems. The public can report odors, and the source tracing detection module generates a source tracing signal based on the public's location after receiving the report. Enterprises can proactively alarm for potential gas pollution situations, and the source tracing detection module generates a source tracing signal based on the location provided by the enterprise when it alarms. If remote sensing satellites detect abnormal locations, such as high VOC concentrations detected by the Sentinel-5P satellite, the source tracing detection module generates a source tracing signal based on that abnormal location. If permanently installed air monitoring equipment issues an early warning, the source tracing detection module can generate a source tracing signal based on the location of the air monitoring equipment.

[0054] Permanent gas detection equipment is installed according to relevant requirements to monitor air quality within the area. For example, cities with a population of over 3 million and a built-up area of ​​over 200 square kilometers are required to have one monitoring point for every 25-30 square kilometers of built-up area, with a minimum of eight points. Although this gas detection equipment is fixed, the distance between adjacent equipment is relatively large, making it impossible to trace the source of instantly emitted pollutants. Therefore, its early warning system can be used as the basis for generating source tracing signals.

[0055] The generation of source tracing signals includes both active early warning and passive monitoring. When an active early warning is not triggered—for example, if members of the public or businesses discover gas pollution but cannot report it—and no other timely remedial measures are taken, the gas pollution cannot be detected in time. This allows the polluted area to expand as the pollutants spread, resulting in greater losses. Therefore, passive monitoring is also implemented, such as collecting and analyzing airborne pollutants through remote sensing hotspots and permanent air monitoring equipment. If gas pollution is passively detected, active source tracing is initiated from the monitoring location to determine the pollution's origin. The combination of active early warning and passive monitoring ensures that gas pollution can be detected and traced in a timely manner.

[0056] The source tracing signal includes anomaly locations, which can be identified from the signal. However, these anomaly locations may not be the actual locations of the gaseous pollution source. Instead, the gaseous pollutants may have diffused to these anomaly locations and been identified. Therefore, these anomaly locations should all be used as the starting point for tracing the gaseous pollution source. If the gaseous pollutants have diffused to these anomaly locations, the gaseous pollution source may be located in any direction centered on these anomaly locations.

[0057] To improve the efficiency of source tracing, a detection route should be planned around the abnormal location. The number of drones should be no less than three. If there are fewer than three drones, it will be impossible to determine the overall concentration distribution based on the detected pollutant concentration, and the test of the drones' endurance will be more severe.

[0058] It is worth noting that gaseous pollutants emitted in an instant generally do not spread very far, and a single drone can detect the pollutant concentration along the detection route. Considering that the drone's endurance is affected by various factors, a drone swarm can be matched to each detection route, that is, multiple drones complete the pollutant concentration detection for that inspection route. Of course, the pollutant concentration can be detected in a relay manner, or it can be started simultaneously in a coordinated manner to detect the pollutant concentration.

[0059] Figure 2 This is a schematic diagram illustrating the principle of planning inspection routes based on target locations. Figure 1 If three (or groups of) drones are selected for source tracing and detection, the plane can be divided into three detection areas at an angle of 120°. Figure 2 Position A is the target position, i.e., the abnormal position contained in the source signal. Lines X1, X2, and X3 are all on the same plane and their starting point is the target position. They are used to divide the plane into several detection areas at equal angles. The angle between any two lines is 120°, dividing the plane into three detection areas: one between lines X1 and X2, one between lines X2 and X3, and one between lines X1 and X3.

[0060] After determining the detection area, a detection route can be planned around the boundary line, such as... Figure 2 The detection route XJ1 is planned around line X1. The UAV can start from position A and perform spiral detection around line X1 along detection route XJ1. Thus, the UAV's detection range covers two detection areas on both sides of line X1. Alternatively, a detection route can be planned within the detection area, such as... Figure 2 The detection route XJ2 is planned in the detection area between line X2 and line X3. The UAV can start from position A and perform spiral detection in the detection area. The coverage of the UAV is the entire detection area.

[0061] Figure 3This is a schematic diagram illustrating the principle of planning inspection routes based on target locations. Figure 2 If four drones (or groups of drones) are selected for source tracing and detection, the plane can be divided into four detection areas at an angle of 90°. Figure 3 In the diagram, position A is the target position, i.e., the abnormal position contained in the source signal. Lines X1, X2, X3, and X4 are all on the same plane and their starting point is the target position. They are used to divide the plane into several detection areas at equal angles. The angle between any two lines is 90°, dividing the plane into four detection areas: one detection area is between lines X1 and X2, another between lines X2 and X3, another between lines X3 and X4, and the third between lines X4 and X1.

[0062] After determining the detection area, a detection route can be planned around the boundary line, such as... Figure 3 The detection route SJ1 is planned around line X1. The UAV can start from position A and perform spiral detection around line X1 using the detection route SJ1. Therefore, the UAV's detection range covers two detection areas on both sides of line X1. Alternatively, a detection route can be planned within the detection area, such as... Figure 3 The detection route XJ2 is planned in the detection area between line X2 and line X3. The UAV can start from position A and perform spiral detection in the detection area. The coverage of the UAV is the entire detection area.

[0063] It should be noted that the equal angle on the plane is to limit the direction of the drone's source tracing, not to limit the drone's flight altitude. That is, when the drone is conducting source tracing detection along the detection route, its flight altitude can be adaptively controlled so as not to affect the source tracing of gaseous pollution.

[0064] In a preferred embodiment, controlling a plurality of drones to detect pollutant concentrations along a detection route includes: associating at least one drone with each detection route; controlling the drones to detect the concentrations of gaseous pollutants along the associated detection routes; wherein the sensors carried by the drones include gas analysis sensors and miniature weather stations.

[0065] To detect gaseous pollutants in the air, drones need to be equipped with various sensors, including but not limited to gas molecule sensors and optical sensors. Gas analysis sensors include photoionization sensors (PID), electrochemical sensors, non-dispersive infrared sensors (NDIR), and metal oxide semiconductor sensors (MOS). PID sensors are used to detect volatile organic compounds (VOCs, such as benzene, toluene, and formaldehyde); electrochemical sensors are used to detect toxic gases (such as H2S, Cl2, SO2, NO2, and CO); NDIR sensors are used to detect infrared-reactive gases (such as CO2, CH4, and CO); and MOS sensors are used to detect broad-spectrum gases (such as VOCs, alcohol, and ammonia). Optical sensors include differential absorption spectrometers and infrared thermal imagers. Differential absorption spectrometers are used to detect nearby gaseous pollutants (such as SO2, NO2, and O3) and pollution source plumes, while infrared thermal imagers are used to detect leaked gas clouds (such as VOCs and SF6).

[0066] It should also be noted that, in addition to the sensors mentioned above for detecting gaseous pollutants, the drone also needs to carry sensors necessary for flight and data transmission, such as lidar and positioning modules. Of course, it also needs to carry a miniature weather station to correct for pollutant concentrations. This miniature weather station can measure wind speed, temperature, humidity, and air pressure during the drone's flight to correct the collected pollutant concentrations, thereby improving the reliability of air quality data and providing accurate data for subsequent identification of gaseous pollutants.

[0067] In a preferred embodiment, correcting pollutant concentration based on collected environmental data includes: identifying the time of pollutant concentration collection and extracting the environmental data corresponding to the collection time; wherein the environmental data is collected by a micro weather station mounted on a drone; and matching the compensation amount of the sensors mounted on the drone with the collected environmental data, and correcting the pollutant concentration using the compensation amount to obtain the corrected pollutant concentration.

[0068] Sensors on drones are easily affected by the surrounding environment. Therefore, a miniature weather station on the drone collects environmental data in real time to correct the detected pollutant concentrations and ensure accuracy. For example, electrochemical sensors are affected by ambient humidity; when detecting H2S in humidity above 80%, the detection value may be 20% higher. Photoionization sensors absorb ultraviolet light at high humidity (>90%), reducing ionization efficiency; if the humidity increases from 50% to 90%, the toluene detection value may be 15%-30% lower.

[0069] The offset of sensor measurements under various environmental conditions relative to a standard environment can be obtained through testing. A fitting curve can be established using environmental data as the independent variable and the offset as the dependent variable; this fitting curve can be built into the drone. After collecting pollutant concentration data, the drone calls the corresponding fitting curve and calculates the offset using the collected environmental data. This offset is the compensation amount. The measured pollutant concentration and the compensation amount are then superimposed to obtain the corrected pollutant concentration.

[0070] Some sensors may be affected by the interaction of multiple environmental factors. Simulations are used to obtain the offset of sensor measurements under various environmental conditions relative to a standard environment. Environmental data (including data on multiple environmental factors such as temperature, humidity, and air pressure) is used as input data, and the offset is used as output data. An artificial intelligence model (such as a backpropagation neural network model) is then trained to obtain a data compensation model. After the drone collects pollutant concentration data, it calls the corresponding data compensation model and calculates the offset using the collected environmental data. This offset is the compensation amount. The measured pollutant concentration and the compensation amount are then superimposed to obtain the corrected pollutant concentration.

[0071] In a preferred embodiment, analyzing pollutant concentrations to estimate the suspected location of a gaseous pollution source includes: constructing a concentration distribution map based on the pollutant concentrations; determining the diffusion axis and diffusion time using the concentration distribution map; estimating the distance to the emission source based on the diffusion time and diffusion coefficient; and calculating the suspected location of the gaseous pollution source based on the diffusion axis and the distance to the emission source.

[0072] When a drone collects pollutant concentrations along a detection route, it can abort the operation if the pollutant concentration is extremely low (or the value is empty). The pollutant concentrations (of the same type of pollutant) collected by each drone are preprocessed, including outlier removal, spatiotemporal alignment, and interpolation completion. The preprocessed pollutant concentration data is then combined with geographic information tools to generate a concentration distribution map, which can show the current distribution of gaseous pollutants.

[0073] After constructing the concentration distribution map, the diffusion axis and diffusion time can be determined based on the map. The diffusion axis refers to the direction in which the pollutant concentration gradient changes most significantly. First, locate the point with the maximum concentration on the concentration distribution map and the surrounding continuous high-concentration area as the core diffusion zone. Draw multiple rays from the core diffusion zone outwards and calculate the concentration attenuation rate per unit distance along each ray, such as the concentration decrease per 100 meters. The direction of the ray with the largest attenuation rate is taken as the diffusion axis, and its reverse direction is the potential direction of the gaseous pollution source.

[0074] After determining the principal axis of diffusion, concentration values ​​are extracted from the maximum concentration outwards along the principal axis to obtain the concentration decay curve. Two points are determined on the concentration decay curve according to the formula: concentration value = maximum concentration / e (where e is the natural constant). Half the distance between these two points is taken as the diffusion standard deviation σ. This standard deviation can also be calculated by fitting a Gaussian curve using linear regression. The formula σ = K·t is then used. n The diffusion time t is obtained by solving for K, where K is the diffusion coefficient, which is determined by the environmental medium and the intensity of turbulence and can be obtained by empirical formulas (such as the Pasquill-Gifford curve in atmospheric diffusion, based on the relationship between stability level σ and t), and n is the diffusion index, which is approximately 0.5 in the atmosphere.

[0075] Next, based on the principle that pollutant concentration decreases exponentially along the diffusion axis with increasing distance from the gaseous pollution source, this can be expressed by the formula... The distance between the gaseous pollution source and the core diffusion zone is calculated; where C1 is the pollutant concentration in the core diffusion zone, or the maximum concentration can be taken; C2 is the theoretical concentration of the gaseous pollution source emitted instantaneously.

[0076] It should be noted that the theoretical concentration C2 can be determined based on the type of pollutant corresponding to the pollutant concentration and experience. It can also be based on the law of conservation of mass, that is, the total amount of pollutants released by the emission source is equal to the sum of the masses of pollutants in the diffusion field (ignoring losses such as sedimentation and chemical reactions).

[0077] In a preferred embodiment, the location of a pollution source is determined by simulating the suspected location of a gaseous pollution source, including: determining a target area with the suspected location as the center and a preset distance as the radius; selecting several suspected pollution sources in the target area; wherein the preset distance is set according to the terrain complexity or wind field stability; verifying several suspected emission sources using a CFD model to determine the optimal emission source; and taking the optimal emission source as the gaseous pollution source, with the corresponding location as the pollution source location.

[0078] The preset distance is related to terrain and wind field. Under flat terrain and stable wind field conditions, the calculation error is generally within 10-20 meters. However, under complex terrain (such as tall buildings or valleys) and fluctuating wind field conditions, the calculation error may reach as high as 30-50 meters. The preset distance should be larger than the calculation error to cover the error range.

[0079] After determining the target area based on suspected locations and preset distances, key geographical information, including fixed facilities and road access, is marked within the target area. Fixed facilities mainly include potential emission sources such as storage tanks, reactors, chimneys, valves, and pipelines, while road access mainly includes factory roads and vehicle parking areas. Based on pollutant concentrations, several suspected pollution sources are identified from fixed facilities and road access, while geographical features unlikely to produce the corresponding gaseous pollution are excluded.

[0080] Several suspected pollution sources were selected as gaseous pollution sources. Historical meteorological data and topographic parameters were collected, and a CFD model was used to simulate the diffusion process of gaseous pollutants. The simulated process was compared with the actual concentration distribution map, and the suspected pollution source with the smallest deviation (such as root mean square error, mean relative error, etc.) was identified as the gaseous pollution source. After determining the gaseous pollution source, its location needed to be physically verified on-site, and the relevant data was archived.

[0081] The CFD (Computational Fluid Dynamics) model is the core tool. Essentially, it solves the governing equations of fluid motion (atmosphere in this invention) and mass transport through numerical calculations, accurately simulating the diffusion patterns of pollutants in the atmosphere. This allows for the verification of suspected pollution sources and the identification of optimal source locations. It can also estimate the data range that needs correction based on historical meteorological data. CFD models are mature tools, and their specific usage will not be elaborated upon here.

[0082] Figure 4 This is a schematic diagram illustrating the principle of identifying gaseous pollution sources based on suspected pollution sources. Figure 4 Rectangle M in the diagram represents the suspected location of the gas pollution source. The target area, centered at location M and with a preset distance as the radius, is defined as follows: Figure 4 The gray circular area is used to analyze and determine the type of gaseous pollutant corresponding to the pollutant concentration. Based on this pollutant type, locations S1, S2, and S3 are marked within the target area. Facilities at these three locations may release the corresponding type of gaseous pollutant. Initial emission parameters for locations S1, S2, and S3, including source strength, emission height, and emission duration, also need to be obtained. CFD models are used to simulate the gaseous pollutant diffusion at locations S1, S2, and S3, with the simulation duration being the emission duration plus the diffusion time. During the simulation, simulated concentrations at the locations and times corresponding to the pollutant concentrations collected by the UAV are extracted. By calculating the deviation between the simulated concentrations and the pollutant concentrations collected by the UAV, the final identified gaseous pollution source among locations S1, S2, and S3 is determined.

[0083] Example 2:

[0084] When gaseous pollutants are dispersed by the wind, they may settle or transform rapidly if they encounter extreme weather, resulting in a decrease in the concentration of pollutants in the air. This can lead to inaccurate concentration data collected by drones, and may even affect the tracing of gaseous pollution sources because the continuous diffusion characteristics of pollutant concentrations are disrupted.

[0085] The following are examples illustrating how extreme weather events can affect air pollutant concentrations:

[0086] Impacts of Heavy Rainfall: Heavy rainfall, including downpours, can dissolve and wash away some water-soluble gaseous pollutants (such as sulfur dioxide and hydrogen chloride). Sulfur dioxide undergoes a series of chemical reactions in rainwater to form sulfurous acid, some of which may be further oxidized to sulfuric acid. As the pollutants fall to the ground, their form and concentration change. In cities, for example, if sulfur dioxide emitted from factories encounters heavy rain during its diffusion process, its atmospheric concentration will decrease significantly due to the precipitation, and it will be converted into acidic substances that enter surface water and soil.

[0087] Impacts of High Temperatures and Drought: Under extreme weather conditions of high temperatures and drought, on the one hand, photochemical reactions of atmospheric pollutants are intensified. For example, nitrogen oxides and volatile organic compounds undergo complex photochemical reactions under strong sunlight and high temperatures, producing secondary pollutants such as ozone and peroxyacetyl nitrate (PAN). On the other hand, high temperatures accelerate the volatilization of pollutants in soil and on the ground surface. Some organic pollutants that were originally adsorbed in the soil may volatilize into the atmosphere, increasing the concentration and complexity of pollutants in the atmosphere. Near chemical industrial parks, high temperatures may cause residual organic pollutants in the soil to volatilize, interacting with pollutants emitted from the parks and leading to more severe air pollution.

[0088] Figure 5 This is a schematic diagram illustrating how extreme weather conditions affect pollutant concentrations. Figure 5 The center point M represents the gaseous pollution source. The horizontal solid line represents the change in pollutant concentration during wind diffusion when the gaseous pollutants are emitted instantaneously and are not affected by extreme weather. If the diffusion is affected by extreme weather, such as a storm, the pollutant concentration in the affected area will decrease. Figure 5 As shown by the horizontal dashed line.

[0089] exist Figure 5 In the scenario shown, if a drone, starting from location A, detects pollutant concentrations along its route, the concentration detected when passing through an area affected by a storm might be lower than at location A. This disrupts the continuous diffusion pattern of pollutant concentrations, making it difficult to construct an accurate concentration distribution map based on the drone's data, thus hindering the tracing of gaseous pollution sources. Of course, if the extreme weather occurs before the drone's detection, it will also affect the accuracy of the concentration distribution map. For example... Figure 5 If the gaseous pollutants encountered extreme weather before the drone began detection, and the affected area had already spread along the direction of the arrow when the drone began detection, then the pollutant concentration detected by the drone would still be partially inaccurate and insufficient to construct an accurate concentration distribution map.

[0090] In a preferred embodiment, before estimating the suspected location of the gas pollution source, the pollutant concentration is corrected based on historical meteorological data, including: extracting historical meteorological data from ground meteorological stations, determining whether the pollutant concentration needs to be corrected based on the historical meteorological data; if so, determining the correction range of the pollutant concentration based on the historical meteorological data; and combining the historical meteorological data with an intelligent correction model to obtain the corrected pollutant concentration; wherein the intelligent correction model is constructed based on an artificial intelligence model.

[0091] As mentioned above, gaseous pollutants may settle or transform during diffusion due to local extreme weather. When drones collect the concentration of these pollutants, the values ​​will be significantly lower than the actual concentration. These data will create a false impression and reduce the accuracy of tracing the source of gaseous pollution.

[0092] After collecting pollutant concentration data, historical meteorological data is extracted from ground-based meteorological stations (or meteorological platforms). This historical meteorological data covers the diffusion paths and diffusion times of gaseous pollutants. Based on the historical meteorological data, it is determined whether there are meteorological scenarios affecting pollutant concentrations. If so, the corresponding data needs to be corrected; otherwise, no correction is required. Using the previous example, the greater the rainfall (e.g., moderate to heavy rain), the more intense the reaction between sulfur dioxide in the air and rainwater, potentially forming sulfurous acid or even sulfuric acid. If the rainfall is light, its impact on sulfur dioxide conversion is relatively small. Therefore, if the rainfall is light, it is determined not to be an (extreme) meteorological scenario affecting gaseous pollutant concentrations; if the rainfall reaches moderate or heavy rain levels, it is determined to be an (extreme) meteorological scenario affecting gaseous pollutant concentrations.

[0093] In a preferred embodiment, the intelligent correction model is constructed based on an artificial intelligence model, including: constructing an artificial intelligence model; wherein the artificial intelligence model includes a BP neural network model or an RBF neural network model; extracting a training dataset, and using the training dataset to train the artificial intelligence model to obtain the intelligent correction model; wherein the training dataset includes meteorological environmental data, diffusion characteristic data, and concentration deviation, wherein the meteorological environmental data is used to represent extreme meteorological scenarios affecting pollutant concentrations, and the diffusion characteristic data includes diffusion time and diffusion distance after being affected by extreme meteorological scenarios.

[0094] The intelligent correction model is constructed through an artificial intelligence model. First, construct an artificial intelligence model, such as a BP neural network model or an RBF neural network model, using the classic three-layer structure, including an input layer, a hidden layer, and an output layer.

[0095] A training dataset is collected or simulated, including different combinations of meteorological and environmental data, diffusion characteristic data, and corresponding pollutant deviation data, which includes pollutant type and concentration deviation. After standardizing the training dataset, an artificial intelligence model is trained and labeled as an intelligent correction model. When pollutant concentration correction is needed, historical meteorological data for the corresponding time is input into the intelligent correction model to obtain the output pollutant deviation data. The concentration deviation of pollutants of the same type is extracted from this data, and the pollutant concentration is corrected based on this deviation to obtain the corrected pollutant concentration.

[0096] It is important to note that gaseous pollutants are continuously diluted during diffusion; that is, the pollutant concentration decreases with increasing diffusion distance. Therefore, it is necessary to incorporate diffusion time, diffusion distance, and meteorological parameters (such as average wind speed) during the diffusion process when training the intelligent correction model. This is essential to accurately assess the amount of pollutant concentration reduction caused by extreme weather events.

[0097] For example, taking a rainfall scenario, the model simulates pollutant concentrations under different rainfall amounts, diffusion times, and diffusion distances. By comparing these concentrations with those under a non-rainfall scenario at the same diffusion time and distance, the concentration deviation can be calculated. The model inputs are rainfall scenarios (moderate rain, heavy rain, etc.), rainfall per unit time, diffusion time, and diffusion distance. The concentration deviation of the pollutant concentration calculated from this set of data is then used as the model output. By simulating several input data sets and corresponding output data sets, these model input and output data can be integrated into a training dataset. This training dataset can then be used to train the artificial intelligence model.

[0098] It should be noted that if a general intelligent correction model for multiple gaseous pollutants is trained, the concentration deviation and pollutant type can be integrated into the model's output data. In other words, the intelligent correction model can simultaneously output the impact of an extreme weather scenario on the concentration of multiple gaseous pollutants.

[0099] In a preferred embodiment, determining the correction range for pollutant concentration based on historical meteorological data includes: analyzing and determining the spatial region affected by historical meteorological data; combining historical meteorological data to determine the spatial region within the current mapping region as the correction region; and using the data range covered by the correction region as the correction range.

[0100] When performing calibration, considering that gaseous pollutants move with the wind, it is necessary to estimate the impact range of extreme weather when calibrating the concentration of gaseous pollutants based on historical meteorological data, and then combine the historical meteorological data to calibrate the pollutant concentration in the corresponding range collected by the drone.

[0101] For example, a gaseous pollutant might be at location one at time one, and after a period of diffusion, it might be at location two at time two. At location one, extreme weather conditions could cause a decrease in pollutant concentration. Therefore, meteorological data from the extreme weather at location one needs to be used to correct the pollutant concentration at location two to avoid abnormal pollutant concentrations caused by environmental factors.

[0102] When estimating the concentration of pollutants affected by extreme weather, historical meteorological data can be input into a CFD model to simulate the diffusion of gaseous pollutants and infer the current location of pollutant concentrations affected by extreme weather. The pollutant concentrations collected at these locations then need to be corrected. For example... Figure 5 In the simulation, based on historical meteorological data, if gaseous pollutants diffuse in the direction of the arrow, then the pollutant concentration data in the area indicated by the arrow needs to be corrected.

[0103] Example 3:

[0104] In a preferred embodiment, after determining the location of the gas pollution source, the pollution area corresponding to the gas pollution source is determined based on the location of the pollution source, including: simulating the coverage area of ​​gaseous pollutants emitted instantaneously by the gas pollution source, and extracting the influence area from the coverage area according to a set concentration threshold; wherein the gaseous pollutants in the influence area will cause adverse effects; using a drone to verify and adjust the boundary of the influence area, and determining the pollution area of ​​the gas pollution source based on the adjustment results.

[0105] While tracing the source of gaseous pollution, the system can also flexibly adjust or reschedule unmanned aerial vehicles to pinpoint contaminated areas, enabling timely treatment of gaseous pollution within those areas. Concentration thresholds are determined based on the type of pollutant, with different thresholds for different pollutant types. Adverse effects refer to the impact on humans, animals, public facilities, etc., caused when pollutant concentrations exceed the concentration threshold.

[0106] Figure 6 A schematic diagram illustrating the principle for determining the pollution area corresponding to a gaseous pollution source. Assume... Figure 6 The outermost solid ellipse represents the coverage area obtained from the CFD model simulation, while the dashed ellipse represents the influence area extracted from the coverage area by setting a concentration threshold. Theoretically, the coverage area is the area affected by the gaseous pollution source during the diffusion process. However, considering that the limitations of the simulation may lead to an inaccurate influence area, this will affect the treatment of the impact of gaseous pollutants.

[0107] The drone can be controlled to verify the boundary of the affected area. For example, if the pollutant concentration collected by the drone at the boundary is equal to the concentration threshold, then the boundary at that point is accurate; if the collected pollutant concentration is less than the concentration threshold, then the drone should detect inward from that location (detecting towards the center of the affected area) and determine the location where the pollutant concentration equals the concentration threshold, and use that location as the new boundary; if the collected pollutant concentration is greater than the concentration threshold, then the drone should detect outward from that location and determine the location where the pollutant concentration equals the concentration threshold, and use that location as the new boundary.

[0108] The boundaries of the impact area are adjusted based on the results of drone verification, such as... Figure 6 In the process, if the drone detects that the pollutant concentration on the boundary of the affected area is less than the concentration threshold, then the drone will detect inward to determine a new boundary. After adjustment, the final polluted area is obtained, which is the innermost solid-line ellipse.

[0109] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A gaseous pollution source detection and tracing system, characterized in that, This includes a source tracing and detection module, and several drones connected to it; Source tracing control module: used to receive source tracing signals, plan detection routes in several directions based on the abnormal locations in the source tracing signals; control several drones to detect pollutant concentrations along the detection routes; as well as, It is used to correct pollutant concentrations based on collected environmental data, analyze the pollutant concentrations to estimate the suspected location of gaseous pollution sources, and simulate and determine the location of the pollution source based on the suspected location of the gaseous pollution source. Before estimating the suspected location of the gaseous pollution source, the concentration of the pollutants is corrected based on historical meteorological data, including: Historical meteorological data is extracted from ground meteorological stations, and the historical meteorological data is used to determine whether pollutant concentration correction is needed; if so, the correction range for pollutant concentration is determined based on the historical meteorological data. The historical meteorological data and the intelligent correction model are combined for correction to obtain the corrected pollutant concentration; wherein, the intelligent correction model is constructed based on an artificial intelligence model; The intelligent correction model is built upon an artificial intelligence model and includes: Construct artificial intelligence models; these models include BP neural network models or RBF neural network models. Extract the training dataset and use the training dataset to train the artificial intelligence model to obtain the intelligent correction model; wherein, the training dataset includes meteorological environmental data, diffusion characteristic data and concentration deviation, the meteorological environmental data is used to represent extreme meteorological scenarios that affect pollutant concentration, and the diffusion characteristic data includes diffusion time and diffusion distance after being affected by extreme meteorological scenarios; The correction range for pollutant concentrations is determined based on historical meteorological data, including: The analysis determined the spatial area affected by the historical meteorological data; By combining historical meteorological data, the spatial region is determined to be within the current mapping region, which is then used as the correction region; the data range covered by the correction region is then used as the correction range.

2. The gas pollution source detection and tracing system according to claim 1, characterized in that, Based on the anomaly locations in the source tracing signal, several detection routes are planned, including: Extract the abnormal locations from the source tracing signal and use them as target locations; Based on the target location, at least three drones are identified; taking the target location as the starting point, several detection areas are determined at equal angles on the horizontal plane, and detection routes are planned in each of the several detection areas.

3. The gas pollution source detection and tracing system according to claim 1, characterized in that, Controlling several drones along a detection route to detect pollutant concentrations includes: Associate at least one drone with each of the aforementioned detection routes; The drone is controlled to detect the concentration of gaseous pollutants along the associated detection route; wherein the sensors carried by the drone include gas analysis sensors and a miniature weather station.

4. The gas pollution source detection and tracing system according to claim 1, characterized in that, Pollutant concentrations were corrected based on collected environmental data, including: The system identifies the time of collection of the pollutant concentration and extracts the environmental data corresponding to the collection time; wherein, the environmental data is collected through a miniature weather station mounted on a drone. The pollutant concentration is corrected by matching the compensation amount of the sensors carried by the UAV with the collected environmental data and using this compensation amount to obtain the corrected pollutant concentration.

5. The gas pollution source detection and tracing system according to claim 1, characterized in that, Analyzing the concentration of the pollutants to estimate the suspected location of the gaseous pollution source includes: A concentration distribution map is constructed based on the pollutant concentration, and the diffusion axis and diffusion time are determined using the concentration distribution map; the distance to the emission source is calculated based on the diffusion time and diffusion coefficient. The suspected location of the gaseous pollution source is calculated based on the diffusion axis and the distance to the emission source.

6. The gas pollution source detection and tracing system according to claim 1, characterized in that, Based on the suspected location of the gas pollution source, the location of the pollution source is determined through simulation, including: A target area is determined with the suspected location as the center and a preset distance as the radius; several suspected pollution sources are selected in the target area; wherein, the preset distance is set according to the terrain complexity or wind field stability; Several suspected emission sources were verified using a CFD model to determine the optimal emission source; the optimal emission source was then designated as the gaseous pollution source, and its corresponding location was designated as the pollution source location.

7. A gaseous pollution source detection and tracing system according to claim 1, characterized in that, After determining the location of the gaseous pollution source, the corresponding pollution area is determined based on that location, including: The coverage area of ​​gaseous pollutants emitted instantaneously from the gas pollution source is simulated, and the influence area is extracted from the coverage area based on a set concentration threshold; wherein, the gaseous pollutants in the influence area will cause adverse effects; The boundaries of the affected area are verified and adjusted using drones, and the pollution area of ​​the gas pollution source is determined based on the adjustment results.

Images