VOCs underway monitoring, analyzing and tracing method based on clustering analysis

Abstract

The invention discloses a VOCs underway monitoring, analyzing and tracing method based on clustering analysis, and relates to the technical field of environment monitoring, and the method comprises the following specific steps: obtaining underway monitoring basic data; analyzing the component spectrum by utilizing clustering analysis; visually marking a navigation track and a component spectrum; and component spectrograms of different types of pollution are subjected to traceability analysis. According to the method, the in-depth analysis of the underway track is realized through the clustering analysis of the underway track. The method can quickly point out dominant pollution spectrograms of different geographic positions, so that the main source of pollution is quickly locked, targeted emission reduction measures can be taken in time, the environment quality is protected and improved, the method has important application value in the field of environment monitoring, the monitoring accuracy and efficiency are improved, and the monitoring cost is reduced. And powerful technical support is provided for environment management and decision making, and the method is of great significance in promoting progress of an environment monitoring technology and environment protection work.

Description

Technical Field

[0001] The present invention relates to the technical field of environmental monitoring, and particularly to a VOCs mobile monitoring analysis and source tracing method based on cluster analysis. Background Art

[0002] Traditional air pollution monitoring methods mainly involve regularly sampling at several discrete points at fixed positions, analyzing the samples to determine the pollution level and its variation over time. This method generally determines the polluted area of air pollution. To determine the pollution source, it requires manual step-by-step investigation of pollution sources, which consumes a large amount of manpower.

[0003] Traditional fixed-position monitoring has significant drawbacks:

[0004] Insufficient spatio-temporal resolution: Discrete sampling results in missing the capture of the pollution diffusion process and cannot reflect the dynamic pollution transmission;

[0005] Low source tracing efficiency: It relies on manual investigation of pollution sources, which takes several days to several weeks (typical case: the VOCs source tracing in a certain industrial park took 23 days, and the labor cost exceeded 50,000 yuan);

[0006] Lack of characteristic analysis: Traditional mobile monitoring technology can only monitor the concentration value of atmospheric VOCs in real time. A single concentration index cannot distinguish the type of pollution source and is prone to misjudgment. Therefore, we propose a VOCs mobile monitoring analysis and source tracing method based on cluster analysis. Summary of the Invention

[0007] (1) Technical problems to be solved

[0008] In view of the deficiencies of the prior art, the present invention provides a VOCs mobile monitoring analysis and source tracing method based on cluster analysis to solve the technical problem that inaccurate pollution detection and analysis lead to inability to trace the source.

[0009] (2) Technical solutions

[0010] To achieve the above object, the present invention is realized through the following technical solutions:

[0011] A VOCs mobile monitoring analysis and source tracing method based on cluster analysis,

[0012] The method includes the following steps:

[0013] 1. Obtaining basic mobile monitoring data

[0014] Obtain basic mobile monitoring data, including the monitoring time t of the monitoring point, the monitoring location (x, y), and the component spectrum s = [s 1 , s 2 ,..., s N, where N is the number of monitored species, such as N = 57, s i represents the concentration of the i-th species. For a single cruise, a series of monitoring point sequences T = {t1, t2,..., t M}, X = {x1, x2,..., x M}, Y = {y1, y2,..., y M}, S = {s1, s2,... s M} will be formed, and M is the total number of monitoring points obtained in a single cruise.

[0015] Preferably, second, using cluster analysis to analyze the component spectra

[0016] Using multivariate analysis techniques to deeply analyze the cruise trajectory and mine the information of the pollution characteristic component spectra. Perform cluster analysis on the component spectra information of all monitoring points recorded in a single cruise mission.

[0017] For each monitoring point, the above data acquisition step has obtained a component spectrum where represents the concentration of the j-th monitored species at the monitoring point i, with the unit of ppm.

[0018] Preferably, the above method of using cluster analysis to analyze the component spectra refers to assigning a type label I = {I1, I2,.., I M} to all monitoring points, I i ∈ [1, K], where K is the number of categories. The methods of using cluster analysis to analyze the component spectra include k-means, DBSCAN, etc. Taking k-means as an example, the specific process of using cluster analysis to analyze the component spectra is explained as follows:

[0019] 1) Initialization, randomly generate a set of type labels

[0020] 2) For each category k, select all subsets s k = {s i}, I i = k, and calculate the mean of all s i vectors, which is the center point of the category

[0021] 3) Calculate the distance between the component spectrum s i of all monitoring points and the center points of each category, and update the k value with the shortest distance to the type label, that is, I i = k, where k is the category serial number marked by the center point i nearest to s ;

[0022] 4) Repeat steps 2) and 3) until the class labels of each component spectrum no longer change and the algorithm converges. The class labels at this time are the clustering results. H is the final number of iterations. The at this time is the average component spectrum of the final type k.

[0023] Preferably, III. Visualize the driving track and component spectrum

[0024] Based on the collected driving basic information and data, visualize the driving track and component spectrum diagram, and associate a specific color C k =(R k , G k , B k ) with each type. On the map, mark points according to the positions x i and y i of the monitoring points. The color of the points is the corresponding associated color C k . The driving track is colored blue, yellow, and green according to the associated categories.

[0025] Preferably, refer to the steps of the flow chart:

[0026] The first step is to actually drive with a driving vehicle. This is the monitoring result of the VOCs component monitoring equipment on the driving vehicle, and each point represents the component spectrum result of one driving;

[0027] The second and third steps are to classify through the described clustering analysis method. For example, for the spectrum diagram on the right side of the figure, the driving results of this time are clustered and analyzed into three types of characteristic component spectra: blue, green, and yellow. That is to say, there are 3 main characteristic component spectra in the driving results of this time;

[0028] The fourth step is the visualization of the clustering results. The results of blue, green, and yellow have been presented. That is, point A is blue, indicating that it is greatly affected by source type A; point B is green, indicating that it is greatly affected by source type B. The monitoring results of the first step are thus transformed into the clustering analysis results based on the pollution source spectrum.

[0029] Preferably, IV. Traceability analysis of component spectrum diagrams of different types of pollution

[0030] Establish a traceability characteristic spectrum library, which includes pollution source type A and the corresponding characteristic component spectrum s A .

[0031] For each type k, take its average component spectrum Calculate respectively The similarity with each component spectrum. When the similarity is greater than a certain threshold, it is determined that this type is the source of the corresponding pollution type. If there are multiple characteristic component spectra with similarities greater than the threshold, the pollution source with the highest similarity is selected.

[0032] Preferably, the similarity calculation method is:

[0033]

[0034] where |·| represents the modulus of the vector.

[0035] Preferably, by associating the pollution type with the characteristic component spectrum, it is possible to trace the pollution source separately according to the differences in the pollution types at different points of the entire vehicle-mounted monitoring trajectory (such as the spraying industry, vehicle exhaust, chemical production, etc.), improving the accuracy of pollution source discrimination.

[0036] (III) Beneficial effects

[0037] The technical achievements of the present invention have significantly improved the intelligent level of environmental monitoring. Through these steps, the present invention not only improves the efficiency and accuracy of environmental monitoring, but also provides strong technical support for environmental management and pollution control. By using the clustering analysis technology, this technology can accurately identify the characteristic pollution spectra presented in the vehicle-mounted monitoring trajectory and associate them with specific pollution emission source industries. This not only enhances the ability to identify pollution sources, but also improves the accuracy and efficiency of monitoring.

[0038] In addition, through the clustering analysis of the vehicle-mounted monitoring trajectory, the present invention realizes in-depth analysis of the vehicle-mounted monitoring trajectory. It can quickly point out the dominant pollution spectra in different geographical locations, thereby quickly locking the main sources of pollution. The realization of this function greatly promotes the rapid response and effective governance of environmental supervision departments to pollution sources, helps to take targeted emission reduction measures in a timely manner, and protects and improves environmental quality.

[0039] Generally speaking, the present invention has important application value in the field of environmental monitoring. It not only improves the accuracy and efficiency of monitoring, but also provides strong technical support for environmental management and decision-making, and is of great significance for promoting the progress of environmental monitoring technology and environmental protection work. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The above description is only an overview of the technical solution of the present invention. In order to understand the technical means of the present invention more clearly and implement it according to the content of the specification, the following takes the preferred embodiments of the present invention and combines with the drawings to describe in detail as follows.

[0041] Figure 1 It is a flowchart of a VOCs vehicle-mounted monitoring analysis and tracing method based on clustering analysis of the present invention;

[0042] Figure 2 An actual navigation diagram of a navigation vehicle for a VOCs navigation monitoring, analysis and tracing method based on cluster analysis according to the present invention;

[0043] Figure 3 The component spectrum of vehicle exhaust associated with the VOCs cruise monitoring and analysis tracing method based on cluster analysis of the present invention Figure 1 ;

[0044] Figure 4 The component spectrum of vehicle exhaust associated with the VOCs cruise monitoring and analysis tracing method based on cluster analysis of the present invention Figure 2 ;

[0045] Figure 5 The component spectrum of vehicle exhaust associated with the VOCs cruise monitoring and analysis tracing method based on cluster analysis of the present invention Figure 3 . DETAILED DESCRIPTION

[0046] The embodiment of the present application provides a VOCs cruise monitoring, analysis and tracing method based on cluster analysis to solve the problem of inaccurate pollution detection and analysis in the prior art, which leads to the inability to trace the source. By applying cluster analysis technology, the technology can accurately identify the characteristic pollution spectrum presented in the cruise trajectory and associate it with specific pollution emission source industries.

[0047] Example 1

[0048] like Figure 1 As shown, the technical solution in the embodiment of the present application is to solve the problem of inaccurate pollution detection and analysis, which leads to the inability to trace the source. The overall idea is as follows:

[0049] In view of the problems existing in the prior art, the present invention provides a VOCs cruise monitoring, analysis and tracing method based on cluster analysis, which includes the following steps:

[0050] 1. Obtaining basic data for cruise monitoring

[0051] Obtain basic navigation data, including monitoring point monitoring time t, monitoring position (x, y) and component spectrum s = [s 1 ,s 2 ,...,s N ], where N is the number of monitored species, such as N = 57, s i represents the concentration of the ith species. For one cruise, a series of monitoring points will be formed, T = {t1, t2, ..., t M},X={x1,x2,...,x M},Y={y1,y2,...,y M},S={s1,s2,...sM}, M is the number of all monitoring points obtained in one navigation.

[0052] 2. Analyzing the component spectrum using cluster analysis

[0053] Multivariate analysis technology is used to deeply analyze the navigation trajectory, mine the spectrum information of pollution characteristic components, and perform cluster analysis on the component spectrum information of all monitoring points recorded in a navigation mission.

[0054] For each monitoring point, the above data acquisition step has obtained a component spectrum in, It represents the concentration of the jth monitored species at monitoring point i, in ppm.

[0055] The above cluster analysis method for parsing component spectra refers to assigning a type label I = {I1, I2, .., I M},I i ∈[1,K], where K is the number of categories. Cluster analysis methods for parsing component spectra include k-means, DBSCAN, etc. Taking k-means as an example, the specific process of clustering is explained. The specific steps are as follows:

[0056] 1) Initialization, randomly generate a set of type tags

[0057] 2) For each category k, select all subsets s belonging to this category k ={s i},I i = k, calculate all s i The mean of the vector, which is the center point of the category

[0058] 3) Calculate the component spectrum s of all monitoring points i The center point of each category The k value with the shortest distance is updated to the type tag, that is, I i = k, where k is the distance s i The nearest center point The serial number of the category marked;

[0059] 4) Repeat steps 2) and 3) until the category label of each component spectrum no longer changes and the algorithm converges. is the clustering result, H is the final number of iterations, and That is the final average component spectrum of type k.

[0060] 3. Visual marking of navigation trajectory and component spectrum

[0061] The steps are as follows:

[0062] 1) Visualization rules

[0063] Assign a unique color C k =(R k , G k , B k ) to each category k, and mark the monitoring points (xi, yi) with C on the map to form a color-segmented driving track; k

[0064] 2) Component spectrum display

[0065] For each category k, plot its average component spectrum μk to highlight characteristic pollutants (such as benzene series, alkanes, etc.);

[0066] 3) Example explanation

[0067] Blue points (type A sources): High in toluene / xylene, possibly associated with the spraying industry;

[0068] Figure 4 Green points (type B sources): High in CO / NO x , possibly associated with vehicle exhaust as shown. According to the collected driving basic information and data visualization, mark the driving track and component spectrum diagram, and associate each type with a specific color C k =(R k , G k , B k ). On the map, according to the positions x i and y i of the monitoring points, plot the points, and the point color is the corresponding associated color C k . As shown in the following table, the driving track is colored blue, yellow, and green according to the associated category.

[0069] Steps referring to the flow chart:

[0070] Figure 2 Comprehensively shown, the first step is to actually drive with a driving vehicle (points in the map area in the figure), which is the monitoring result of the VOCs component monitoring equipment on the driving vehicle, and each point represents the component spectrum result of one drive;

[0071] The second and third steps are to classify through the described clustering analysis method. For example, for the spectrum diagram on the right side of the figure, the driving results of this time are clustered and analyzed into three types of characteristic component spectra: blue, green, and yellow. That is to say, there are 3 main characteristic component spectra in the driving results of this time (including those mentioned in the text with a similarity threshold exceeding 0.75);

[0072] Figures 3 - 5 ​​As shown, the visualization of the clustering results in the fourth step has presented the results in blue, green, and yellow. That is, point A is blue, indicating that it is greatly affected by source type A; point B is green, indicating that it is greatly affected by source type B. Thus, the monitoring results of the first step have been transformed into the results of cluster analysis based on the pollution source spectra.

[0073] IV. Source Tracing Analysis of Component Spectra of Different Types of Pollution

[0074] Establish a source tracing characteristic spectrum library, which includes pollution source type A and the corresponding characteristic component spectrum s A 。

[0075] For each type k, take its average component spectrum Calculate respectively the similarity with each component spectrum. When the similarity is greater than a certain threshold, determine that this type is the source of the corresponding pollution type. If there are multiple characteristic component spectra with similarities greater than the threshold, select the pollution source with the highest similarity. The similarity calculation method is:

[0076]

[0077] where |·| represents the norm of the vector.

[0078] By associating the pollution type with the characteristic component spectrum, it is possible to trace and find the pollution sources (such as the spraying industry, vehicle exhaust, chemical production, etc.) respectively according to the differences in pollution types at different points of the entire driving track, improving the accuracy of pollution source discrimination.

[0079] The technical achievements of the present invention have significantly improved the intelligent level of environmental monitoring. Through these steps, the present invention not only improves the efficiency and accuracy of environmental monitoring, but also provides strong technical support for environmental management and pollution control. By using the cluster analysis technology, this technology can accurately identify the characteristic pollution spectra presented in the driving track and associate them with specific pollution emission source industries. This not only enhances the ability to identify pollution sources, but also improves the accuracy and efficiency of monitoring.

[0080] In addition, through the cluster analysis of the driving track, this technology realizes the in-depth analysis of the driving track. It can quickly point out the dominant pollution spectra at different geographical locations, thus quickly locking the main sources of pollution. The realization of this function has greatly promoted the rapid response and effective treatment of pollution sources by environmental supervision departments, helping to take targeted emission reduction measures in a timely manner to protect and improve environmental quality.

[0081] Generally speaking, the present invention has important application value in the field of environmental monitoring. It not only improves the accuracy and efficiency of monitoring, but also provides strong technical support for environmental management and decision-making, which is of great significance for promoting the progress of environmental monitoring technology and environmental protection work.

[0082] Finally, it should be noted that: Obviously, the above embodiments are merely examples given to clearly illustrate the present invention, rather than limitations on the implementation manners. For those of ordinary skill in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to enumerate all the implementation manners here. And the obvious changes or modifications derived therefrom are still within the protection scope of the present invention.

Claims

1. A VOCs cruise monitoring, analysis and tracing method based on cluster analysis, characterized by: The specific steps are as follows:

1. Obtaining basic data for cruise monitoring; 2. Analyze the component spectrum using cluster analysis; 3. Visually mark the navigation trajectory and component spectrum; 4. Component spectrum tracing analysis of different types of pollution.

2. A VOCs cruise monitoring, analysis and tracing method based on cluster analysis as claimed in claim 1, characterized in that: Step 1: Acquisition of basic data for cruise monitoring Obtain basic navigation data, including monitoring point monitoring time t, monitoring position (x, y) and component spectrum s = [s 1 ,s 2 ,...,s N ], where N is the number of monitored species, such as N = 57, s i represents the concentration of the ith species. For one cruise, a series of monitoring points will be formed, T = {t1, t2, ..., t M },X={x1,x2,...,x M },Y={y1,y2,...,y M },S={s1,s2,...s M }, M is the number of all monitoring points obtained in one navigation.

3. A VOCs cruise monitoring, analysis and tracing method based on cluster analysis as claimed in claim 2, characterized in that: Step 2: Analyze component spectrum using cluster analysis Multivariate analysis technology is used to deeply analyze the navigation trajectory, mine the spectrum information of pollution characteristic components, and perform cluster analysis on the component spectrum information of all monitoring points recorded in a navigation mission. For each monitoring point, the above data acquisition step has obtained a component spectrum in, It represents the concentration of the jth monitored species at monitoring point i, in ppm.

4. A VOCs cruise monitoring, analysis and tracing method based on cluster analysis as claimed in claim 3, characterized in that: The method of using cluster analysis to analyze the component spectrum in the above step 2 refers to assigning a type label I = {I1, I2, .., I M },I i ∈[1,K], where K is the number of categories. The specific process of cluster analysis to analyze the component spectrum is explained as follows: 1) Initialization, randomly generate a set of type tags 2) For each category k, select all subsets s belonging to this category k ={s i },I i = k, calculate all s i The mean of the vector, which is the center point of the category 3) Calculate the component spectrum s of all monitoring points i The center point of each category The k value with the shortest distance is updated to the type tag, that is, I i = k, where k is the distance s i The nearest center point The serial number of the category marked; 4) Repeat steps 2) and 3) until the category label of each component spectrum no longer changes and the algorithm converges. is the clustering result, H is the final number of iterations, and That is the final average component spectrum of type k.

5. A VOCs cruise monitoring, analysis and tracing method based on cluster analysis as claimed in claim 4, Features: The steps of step 3. visually marking the navigation trajectory and component spectrum are as follows: 1) Visualization rules Assign a unique color C to each category k k =(R k ,G k ,B k ), with C k Mark the monitoring points (xi, yi) to form a color-segmented navigation trajectory; 2) Component spectrum display For each class k, plot its average component spectrum μk.

6. A VOCs cruise monitoring, analysis and tracing method based on cluster analysis as claimed in claim 5, characterized in that: Among them, 1) visualization rules, based on the collected basic navigation information and data visualization marking navigation trajectory and component spectrum, associate each type with a specific color C k =(R k ,G k ,B k ), on the map, according to the location of the monitoring point x i and i Draw points, point colors are corresponding Associated color C k , 7. A VOCs cruise monitoring, analysis and tracing method based on cluster analysis as claimed in claim 5, characterized in that: Steps of step three: The first step is to use the UAV to actually cruise. This is the monitoring result of the VOCs component monitoring equipment on the UAV. Each point represents the component spectrum result of one cruise. The second and third steps are to classify by the described cluster analysis method; The fourth step is to visualize the clustering results by distinguishing them with different colors.

8. The VOCs cruise monitoring, analysis and tracing method based on cluster analysis as claimed in claim 5, characterized in that: Step 4. Spectral traceability analysis of components of different types of pollution Establish a traceability characteristic spectrum library, which includes pollution source type A and corresponding characteristic component spectrum s A , for each type k, take its average component spectrum Calculate separately The similarity with each component spectrum. When the similarity is greater than a certain threshold, the type is judged as the source of the corresponding pollution type. If there are multiple characteristic component spectra with similarities greater than the threshold, the pollution source with the highest similarity is selected. The similarity calculation method is: where |·| represents the magnitude of the vector.

9. A VOCs cruise monitoring, analysis and tracing method based on cluster analysis as claimed in claim 8, characterized in that: By associating the pollution type with the characteristic component spectrum, the source of pollution can be traced back according to the differences in pollution types at different points in the entire flight trajectory.

Images