A method for processing pollutant emission monitoring data for early warning of polluting enterprises

Abstract

The present invention discloses a method for processing pollutant monitoring data for early warning of polluting enterprises, comprising the following steps: A. sampling pollutants at the pollutant discharge end, analyzing and obtaining the pollutant types and the emission data of each type of pollutant at the pollutant discharge end; B. sampling the peripheral environment of the pollutant discharge end, obtaining the concentration change data of each type of pollutant; C. establishing a spatial distribution map of the pollutant concentration in the peripheral environment; D. establishing a diffusion trend model of each type of pollutant in the spatial distribution map; E. using the diffusion trend model to predict the change of the spatial distribution of pollutants. The present invention can improve the deficiencies of the prior art, simplify the computational complexity of the pollutant discharge monitoring prediction process, and improve its robustness.

Description

Technical Field

[0001] The invention relates to the technical field of environmental monitoring, and in particular to a method for processing pollutant emission monitoring data for early warning of emission of polluting enterprises. Background Art

[0002] With the improvement of my country's environmental protection requirements, it is now required that not only the emission indicators of pollutants from polluting enterprises meet national standards, but also the total emission limit is determined based on the actual environmental data of each region, and then the emission quota is issued to each polluting enterprise according to the limit, requiring the pollutant emission enterprises to discharge pollutants strictly in accordance with the emission quota. This requires the existing monitoring system to quickly and accurately predict the impact of future pollution discharge processes on the environment. There are many existing prediction methods, such as the use of neural network models, genetic algorithms, etc. Although the existing algorithms can obtain relatively accurate prediction results, their computational complexity is large and their robustness is insufficient. When the initial conditions fluctuate slightly, the prediction results may change dramatically. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a pollution emission monitoring data processing method for emission early warning of polluting enterprises, which can solve the shortcomings of the existing technology, simplify the computational complexity of the pollution emission monitoring prediction process, and improve its robustness.

[0004] In order to solve the above technical problems, the technical solutions adopted by the present invention are as follows.

[0005] A method for processing pollutant emission monitoring data for early warning of polluting enterprises comprises the following steps:

[0006] A. Sampling pollutants at the sewage discharge end, analyzing and obtaining the types of pollutants at the sewage discharge end and the emission data of each type of pollutant;

[0007] B. Take samples of the surrounding environment of the sewage discharge end to obtain the concentration change data of each type of pollutant;

[0008] C. Establish a spatial distribution map of pollutant concentrations in the surrounding environment;

[0009] D. Establish a spreading trend model for each type of pollutant in the spatial distribution map;

[0010] E. Use the spread trend model to predict changes in the spatial distribution of pollutants.

[0011] Preferably, the types of pollutants include air pollutants and water pollutants. Air pollutants include dust, sulfur dioxide, nitrogen dioxide, carbon monoxide and volatile organic compounds. Water pollutants include suspended particulate matter, oxygen-consuming organic matter, inorganic nitrogen and heavy metal elements.

[0012] Preferably, in step C, establishing a spatial distribution map of the concentration of pollutants in the surrounding environment includes the following steps:

[0013] The concentration point data of each type of pollutant are expanded into concentration surface data one by one, and then the concentration surface data are visualized to obtain the spatial image of each type of pollutant. The spatial images belonging to air pollutants are merged, and the spatial images belonging to water pollutants are merged.

[0014] Preferably, expanding the concentration point data into concentration surface data comprises the following steps:

[0015] Assume that the direction vector between the i-th concentration point data and the sewage discharge end is v i According to the difference between the i-th concentration point data and the corresponding pollutant concentration at the sewage discharge end, the corresponding pollutant diffusion rate in the horizontal direction and the vertical direction are fitted, and the two sets of diffusion rates are combined into the pollutant diffusion rate matrix d at the i-th concentration point data position i , the external diffusion state matrix of the i-th concentration point data is S i , S i =K i ·v i ·d i , K i is the state correction matrix, and the influence range of the i-th concentration point data is ||S i || F >L, L is the lower limit threshold of the influence coefficient;

[0016] When the influence range of a concentration point data does not include the influence range of other concentration point data, this concentration point data and its diffusion rate matrix are used to fit the concentration data within the influence range; when the influence range of a concentration point data includes the influence range of other concentration point data, all the relevant concentration point data within the overlapping range and their corresponding diffusion rate matrices are used to fit the concentration data within the overlapping influence range; when there is an area that is not included in the influence range of any concentration point data, the center point of this area is calculated, and the direction vector between the concentration point data closest to the center point and the center point is used to correct the diffusion rate matrix of the concentration point data, and then the corrected diffusion rate matrix and the corresponding concentration point data are used to fit the concentration data of this area.

[0017] Preferably, imaging the concentration surface data comprises the following steps:

[0018] Each data point in the concentration surface data is taken as a pixel point of the image data, and the concentration data is linearly changed to the initial grayscale value G of the corresponding pixel. The grayscale correction function F of each pixel is fitted using the coordinate position of the measured concentration point data and its diffusion rate matrix. Suppose the influence weight of the current pollutant in this type of pollutant is W, and the three-dimensional data information contained in each pixel point of the spatial image is (G, W, F).

[0019] Preferably, the merging of the spatial images comprises the following steps:

[0020] The grayscale values ​​of different pixels in the corresponding positions are weighted averaged by the influence weights to obtain the merged grayscale G ’ , the grayscale correction functions of different pixels in the corresponding position are weighted averaged by the influence weights to obtain the combined grayscale correction function F ’ .

[0021] Preferably, in step D, establishing a diffusion trend model of each type of pollutant in the spatial distribution map includes the following steps:

[0022] According to the linearity, the grayscale correction function F ’ Group them and combine the grayscale correction function F for each group ’ Perform Laplace transform, then merge the frequency domain components that are linearly correlated with the frequency domain characteristics to obtain a grayscale change frequency domain correlation mapping set, then perform inverse Laplace transform on the grayscale change frequency domain correlation mapping set to obtain a grayscale change path set, and mark the grayscale change path set in the spatial distribution map.

[0023] Preferably, in step E, using the diffusion trend model to make predictions includes the following steps:

[0024] The future emission data of the sewage discharge end is converted into grayscale data, and then input into the diffusion trend model, so that the grayscale data is transmitted according to the grayscale change path set. During the transmission process, the pixels on the path and the corresponding pollutant types are combined into a grayscale correction function F. ’ Grayscale fitting is performed on the pixels within the included range, and finally the fitted grayscale data is used to determine the trend of pollutant concentration changes within the monitoring range.

[0025] The beneficial effect brought about by adopting the above technical solution is that the present invention greatly reduces the amount of calculation and improves the prediction robustness by converting the pollution data into image data of uniform format and nature for processing and prediction. DETAILED DESCRIPTION

[0026] The present invention first samples and monitors the emission data of pollutants and environmental pollution data. The sampled data include air pollutants and water pollutants, among which air pollutants include dust, sulfur dioxide, nitrogen dioxide, carbon monoxide and volatile organic compounds, and water pollutants include suspended particulate matter, oxygen-consuming organic matter, inorganic nitrogen, heavy metal elements and other pollutant types.

[0027] According to the existing technology, a simulation model (such as the ECA-CNN model) is established, the collected data is used to train the model, and then the simulation operation is performed. Because a large number of cycles of training and simulation are required, the amount of calculation is large, and because of the robustness problem, in order to ensure the authenticity of the simulation results, the training set data needs to be frequently updated, and then the cycle training is repeated, which further increases the amount of calculation. Therefore, we abandoned the simulation operation method using traditional simulation models, and used data visualization operations to achieve data normalization. At the same time, we used the characteristics of image data to improve the robustness of the prediction system, quickly fit the diffusion path of pollutants, and thus achieve pollution discharge prediction. Specifically include:

[0028] 1. Expand the concentration point data into concentration surface data.

[0029] Since the sampling process is point sampling, the first step in data visualization is to solve the problem of how to convert the sampling point data into surface data. Since the diffusion of pollution in the environmental space has its own characteristics, the diffusion of pollutants must be simulated when generating surface data. Let the direction vector between the i-th concentration point data and the sewage discharge end be v i According to the difference between the i-th concentration point data and the corresponding pollutant concentration at the sewage discharge end, the corresponding pollutant diffusion rate in the horizontal direction and the vertical direction are fitted, and the two sets of diffusion rates are combined into the pollutant diffusion rate matrix d at the i-th concentration point data position i , the external diffusion state matrix of the i-th concentration point data is S i , S i =K i ·v i ·d i , K i is the state correction matrix, and the influence range of the i-th concentration point data is ||S i || F >L, where L is the lower limit threshold of the influence coefficient.

[0030] When the influence range of a concentration point data does not include the influence range of other concentration point data, this concentration point data and its diffusion rate matrix are used to fit the concentration data within the influence range; when the influence range of a concentration point data includes the influence range of other concentration point data, all the relevant concentration point data within the overlapping range and their corresponding diffusion rate matrices are used to fit the concentration data within the overlapping influence range; when there is an area that is not included in the influence range of any concentration point data, the center point of this area is calculated, and the direction vector between the concentration point data closest to the center point and the center point is used to correct the diffusion rate matrix of the concentration point data, and then the corrected diffusion rate matrix and the corresponding concentration point data are used to fit the concentration data of this area.

[0031] 2. Visualize the concentration surface data.

[0032] The key to data visualization is how to make the image data simulate the distribution state of pollutants. Since the grayscale data of traditional grayscale images is one-dimensional and cannot effectively simulate the distribution state of pollutants, we have specially designed a three-dimensional data information structure containing initial grayscale values, influence weights and grayscale correction functions to simulate the distribution state of pollutants. Take each data point in the concentration surface data as a pixel point of the image data, linearly change the concentration data to the initial grayscale value G of the corresponding pixel, and use the measured coordinate position of the concentration point data and its diffusion rate matrix to fit the grayscale correction function F of each pixel point. Assume that the influence weight of the current pollutant in this type of pollutant is W, and the three-dimensional data information contained in each pixel point of the spatial image is (G, W, F).

[0033] 3. Merge multiple layers of spatial images.

[0034] The grayscale values ​​of different pixels in the corresponding positions are weighted averaged by the influence weights to obtain the merged grayscale G ’ , the grayscale correction functions of different pixels in the corresponding position are weighted averaged by the influence weights to obtain the combined grayscale correction function F ’ .

[0035] 4. Establish a diffusion trend model for each type of pollutant in the spatial distribution map.

[0036] The most important thing in the diffusion trend model is the diffusion path of pollutants in space. Since the combined grayscale correction function contains the diffusion rates of different types of pollutants in different directions and regions, they can be separated by frequency domain transformation. Therefore, the combined grayscale correction function F is linearly converted to ’ Group them and combine the grayscale correction function F for each group ’ Perform Laplace transform, then merge the frequency domain components that are linearly correlated with the frequency domain characteristics to obtain a grayscale change frequency domain correlation mapping set, then perform inverse Laplace transform on the grayscale change frequency domain correlation mapping set to obtain a grayscale change path set, and mark the grayscale change path set in the spatial distribution map.

[0037] 5. Use the diffusion trend model for prediction.

[0038] The future emission data of the sewage discharge end is converted into grayscale data, and then input into the diffusion trend model, so that the grayscale data is transmitted according to the grayscale change path set. During the transmission process, the pixels on the path and the corresponding pollutant types are combined into a grayscale correction function F. ’ Grayscale fitting is performed on the pixels within the included range, and finally the fitted grayscale data is used to determine the trend of pollutant concentration changes within the monitoring range.

[0039] The present invention solves the technical problems existing in the existing prediction methods by means of data visualization, is suitable for direct upgrading based on the existing prediction system, and reduces the hardware cost of system update.

[0040] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for processing pollutant emission monitoring data for early warning of polluting enterprises, characterized in that The following steps are involved: A. Sampling pollutants at the sewage discharge end, analyzing and obtaining the types of pollutants at the sewage discharge end and the emission data of each type of pollutant; B. Take samples of the surrounding environment of the sewage discharge end to obtain the concentration change data of each type of pollutant; C. Establish a spatial distribution map of pollutant concentrations in the surrounding environment; The following steps are included: The concentration point data of each type of pollutant are expanded into concentration surface data one by one, and then the concentration surface data are visualized to obtain the spatial image of each type of pollutant, and the spatial images belonging to air pollutants are merged, and the spatial images belonging to water pollutants are merged; The following steps are involved in imaging the concentration surface data: Each data point in the concentration surface data is taken as a pixel point of the image data, and the concentration data is linearly changed to the initial grayscale value G of the corresponding pixel. The grayscale correction function F of each pixel point is fitted using the coordinate position of the measured concentration point data and its diffusion rate matrix. The influence weight of the current pollutant in this type of pollutant is assumed to be W, and the three-dimensional data information contained in each pixel point of the spatial image is (G, W, F); The merging of spatial images includes the following steps: The grayscale values ​​of different pixels in the corresponding positions are weighted averaged by the influence weights to obtain the merged grayscale G ’ , the grayscale correction functions of different pixels in the corresponding position are weighted averaged by the influence weights to obtain the combined grayscale correction function F ’ ; D. Establish a spreading trend model for each type of pollutant in the spatial distribution map; The following steps are included: According to the linearity, the grayscale correction function F ’ Group them and combine the grayscale correction function F for each group ’ Perform Laplace transform, then merge the frequency domain components that are linearly correlated with the frequency domain characteristics to obtain a grayscale change frequency domain correlation mapping set, then perform inverse Laplace transform on the grayscale change frequency domain correlation mapping set to obtain a grayscale change path set, and mark the grayscale change path set in the spatial distribution diagram; E. Use the diffusion trend model to predict the change of the spatial distribution of pollutants; including the following steps: The future emission data of the sewage discharge end is converted into grayscale data, and then input into the diffusion trend model, so that the grayscale data is transmitted according to the grayscale change path set. During the transmission process, the pixels on the path and the corresponding pollutant types are combined to form a grayscale correction function F. ’ Grayscale fitting is performed on the pixels within the included range, and finally the fitted grayscale data is used to determine the trend of pollutant concentration changes within the monitoring range.

2. The method for processing pollutant emission monitoring data for early warning of polluting enterprises according to claim 1 is characterized by: The types of pollutants include air pollutants and water pollutants. Air pollutants include dust, sulfur dioxide, nitrogen dioxide, carbon monoxide and volatile organic compounds. Water pollutants include suspended particulate matter, oxygen-consuming organic matter, inorganic nitrogen and heavy metal elements.

3. The method for processing pollutant emission monitoring data for early warning of polluting enterprises according to claim 2 is characterized by: Expanding concentration point data to concentration surface data includes the following steps: Assume that the direction vector between the i-th concentration point data and the sewage discharge end is v i According to the difference between the i-th concentration point data and the corresponding pollutant concentration at the sewage discharge end, the corresponding pollutant diffusion rate in the horizontal direction and the vertical direction are fitted, and the two sets of diffusion rates are combined into the pollutant diffusion rate matrix d at the i-th concentration point data position i , the external diffusion state matrix of the i-th concentration point data is S i , S i =K i ·v i ·d i , K i is the state correction matrix, and the influence range of the i-th concentration point data is , L is the lower limit threshold of the influence coefficient; When the influence range of a concentration point data does not include the influence range of other concentration point data, this concentration point data and its diffusion rate matrix are used to fit the concentration data within the influence range; when the influence range of a concentration point data includes the influence range of other concentration point data, all the relevant concentration point data within the overlapping range and their corresponding diffusion rate matrices are used to fit the concentration data within the overlapping influence range; when there is an area that is not included in the influence range of any concentration point data, the center point of this area is calculated, and the direction vector between the concentration point data closest to the center point and the center point is used to correct the diffusion rate matrix of the concentration point data, and then the corrected diffusion rate matrix and the corresponding concentration point data are used to fit the concentration data of this area.