Source-filth relationship analysis method and system for high-salt sand and dust natural filth source area

Abstract

The invention provides a source-filth relation analysis method and system for a high-salt type sand and dust natural filth source area, and relates to the technical field of insulation detection, and the method comprises the steps: setting sampling points in the high-salt type sand and dust natural filth source area, collecting a source sample, and carrying out the pretreatment; based on an isotope mass balance model and in combination with the preprocessed source sample, determining a # imgabs 1 # value of # imgabs 0 # in the soil sample, and extracting oxygen isotope data; constructing a deep neural model based on a self-attention mechanism; and training a deep neural model through the data of the source sample and the oxygen isotope data, outputting to obtain a relation model between the source-filth characteristics of the high-salt sand and dust and the environmental change, and performing visual display. According to the method, a sand and dust sample, a rainfall sample and a soil sample are analyzed, a relation model between the source-filth characteristics of high-salt sand and dust and environmental changes is constructed, the sources of nitrate and components thereof can be accurately identified, and data support is provided for environmental governance.

Description

Technical Field

[0001] The present application relates to the field of insulation detection technology, and in particular to a source-pollution relationship analysis method and system for a high-salinity dust natural pollution source area. Background Art

[0002] In the Tarim Basin, most transmission lines pass through desert areas with frequent sandstorms. These extreme environments pose a severe challenge to the safety of the power system. The impact of sandstorms and natural pollution in saline areas on the external insulation of high-voltage equipment is a key factor in ensuring the reliable operation of the power grid. Sandstorms not only cause a large amount of salt, sulfur, calcium and other substances to accumulate on the surface of the equipment, but also form pollution flash when the humidity increases, causing equipment failure and power outages. Under extreme climatic conditions, power facilities face safety hazards and maintenance needs. Therefore, it is particularly important to conduct research on the source-pollution relationship in high-salinity sand and dust natural pollution source areas.

[0003] At present, the research on pollution sources in dust and salinized areas is relatively weak, and effective monitoring and analysis methods have not yet been established. There has been no effective measurement method for the analysis of the spatial distribution characteristics of this type of natural pollution. Summary of the invention

[0004] The purpose of the present invention is to provide a source-pollution relationship analysis method and system for high-salt dust natural pollution source areas in order to fill the gap in the technical research on pollution sources in dust and saline areas.

[0005] The above-mentioned purpose of the present application is achieved through the following technical solutions: S1: Set up sampling points in the high-salinity dust natural pollution source area, collect source samples and pre-treat them; source samples include: dust samples, precipitation samples and soil samples; S2: Based on the isotope mass balance model, the isotope mass in soil samples was determined by combining the pre-treated source samples. of Value, extract oxygen isotope data; S3: Construct a deep neural model based on the self-attention mechanism; train the deep neural model through the source sample data and oxygen isotope data, output the relationship model between the source-pollution characteristics of high-salinity dust and environmental changes and visualize it.

[0006] Optionally, step S1 includes: The high-salinity dust natural pollution source areas include: power transmission lines and substations; Within the preset range of transmission lines and substations, dust end members and external insulation pollution sampling points are set up to obtain source samples of high-salt natural pollution and human-influenced pollution.

[0007] Optionally, the preprocessing includes the following steps: analyzing nitrate data and organic component data in the source sample.

[0008] Optionally, the preprocessing includes the following steps: analyzing nitrate data and organic component data in the source sample.

[0009] Optionally, step S3 includes: The soil sample of The value can be expressed by the following formula:

[0010] in, In soil samples of value; Atmospheric deposition of The value of dust sample of value; Indicates the biological nitrification of Value, that is, the value of organic component data of value; and Represents the atmospheric deposition and nitrification effects on soil The sum of the contribution ratio is 1.

[0011] Optionally, step S5 includes: The deep learning model includes: a convolutional neural network and a self-attention mechanism module; The convolutional neural network includes: an input layer, a hidden layer and an output layer.

[0012] Optionally, the cross entropy loss function is used as a cost function to train the deep learning network model.

[0013] A high-salinity dust natural pollution source area source-pollution relationship analysis system, the system comprises: a sample collection and preprocessing module, a chemical composition analysis module, a data analysis module and a display module; The sample collection and preprocessing module is used to collect source samples and perform preprocessing; the source samples include: dust samples, precipitation samples and soil samples; The chemical composition analysis module is used to determine the content of the isotope mass balance model in soil samples in combination with pre-treated source samples. of Value, extract oxygen isotope data; The data analysis module is used to build a deep neural model based on the self-attention mechanism; the deep neural model is trained through the source sample data and oxygen isotope data, and the output is a relationship model between the source-pollution characteristics of high-salinity dust and environmental changes; The display module is used to visualize the relationship model between the source-pollution characteristics of high-salinity dust and environmental changes.

[0014] The beneficial effects of the technical solution provided by this application are: 1. By setting up sampling points in high-salt dust natural pollution source areas and collecting air, precipitation and soil samples, this study can accurately identify the sources of nitrate and its components and provide data support for environmental governance.

[0015] 2. Monitoring of nitrate and nitrogen oxides in the air can reveal the relationship between these pollutants and sandstorms, and increase sensitivity to environmental changes.

[0016] 3. By analyzing the salt content, organic composition and isotopes of the samples, a source-pollution relationship model is constructed, making the quantitative assessment of pollution sources more comprehensive and accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present application will be further described below with reference to the accompanying drawings and embodiments, in which: Figure 1 is a step diagram in an embodiment of the present application; Figure 2 It is a visualization result diagram in the embodiment of the present application; Figure 3 It is a schematic diagram of the system structure in an embodiment of the present application. DETAILED DESCRIPTION

[0018] In order to have a clearer understanding of the technical features, purposes and effects of the present application, the specific implementation methods of the present application are now described in detail with reference to the accompanying drawings.

[0019] An embodiment of the present application provides a method for analyzing the source-pollution relationship of a high-salinity dust natural pollution source area.

[0020] Please refer to Figure 1 , Figure 1 This is a step diagram of a method for analyzing the source-pollution relationship of a high-salinity dust natural pollution source area in an embodiment of the present application, comprising: S1: Set up sampling points in the high-salinity dust natural pollution source area, collect source samples and pre-treat them; source samples include: dust samples, precipitation samples and soil samples; The present application provides an embodiment as follows, in which the concentration and composition of nitrate are particularly concerned in precipitation samples and soil samples for subsequent analysis. Increase monitoring of nitrate and nitrogen oxides in the air to analyze their relationship with dust.

[0021] S2: Based on the isotope mass balance model, the isotope mass in soil samples was determined by combining the pre-treated source samples. of Value, extract oxygen isotope data; This application provides an embodiment as follows: Oxygen isotopes (such as and ) analysis and establish an analytical method for high-salt pollution components.

[0022] S3: Construct a deep neural model based on the self-attention mechanism; train the deep neural model through the source sample data and oxygen isotope data, output the relationship model between the source-pollution characteristics of high-salinity dust and environmental changes and visualize it.

[0023] The present application provides an embodiment as follows: isotope data integration: in the process of model construction, consider The oxygen isotope data of the pollution source are incorporated into the model to quantify its contribution and impact on the pollution source. The self-attention mechanism is combined with the deep learning model to improve the model performance and generalization ability, and verify the importance of isotope data in the source-pollution relationship model. Data analysis and result display, analyze the model output, generate the spatial distribution characteristics of the pollution source, and perform visual display, such as Figure 2 shown.

[0024] Step S1 includes: The high-salinity dust natural pollution source areas include: power transmission lines and substations; Within the preset range of transmission lines and substations, dust end members and external insulation pollution sampling points are set up to obtain source samples of high-salt natural pollution and human-influenced pollution.

[0025] The preprocessing comprises the following steps: analyzing nitrate data and organic component data in the source sample.

[0026] Step S2 includes: using ion exchange method to combine nitrate data and organic matter composition data to determine the content of nitrate in soil samples. of Oxygen isotope data.

[0027] This application provides an embodiment as follows, for There are three main types of pretreatment methods for (nitrate) trioxygen isotope analysis: ion exchange, denitrification bacteria, and chemical conversion. The treatment method used in this invention is ion exchange, which separates When the sample solution flows through the anion exchange column containing positively charged groups, The ions interact with the positive charges on the resin in the column and are adsorbed, while other ions are not captured. Competition for binding sites leads to is gradually replaced by ions in the eluent, eventually After elution from the column After high temperature cracking, and The oxygen isotope analysis of the gases was performed by a continuous flow-isotopic ratio mass spectrometer (CF-IRMS) system. This method is suitable for samples in high-concentration nitrate environments and avoids the exchange of oxygen atoms at high temperatures, thus improving the test accuracy.

[0028] The effect of precipitation on soil of The influence of precipitation on soil nitrate concentration: The increase of precipitation can dilute the nitrate concentration in soil, and rain can also bring nitrogen compounds from the atmosphere into soil. Therefore, the amount and frequency of precipitation will affect the nitrate concentration in soil. The concentration and In seasons or regions with high precipitation, atmospheric deposition has a great impact on soil The contribution of of Microbial activity in soil of Impact of soil microbial activity, especially nitrification and denitrification, can affect The production and conversion of nitrifying bacteria into , while denitrifying bacteria Convert to or These processes will change of values, because different microbial processes involve different enzymes and reaction mechanisms, resulting in By measuring the microbial community structure and functional gene abundance related to nitrogen cycle in soil, we can obtain the microbial response to nitrogen cycle. contribution.

[0029] Step S3 includes: The soil sample of The value can be expressed by the following formula:

[0030] in, In soil samples of value; Atmospheric deposition of The value of dust sample of value; Indicates the biological nitrification of Value, that is, the value of organic component data of value; and Represents the atmospheric deposition and nitrification effects on soil The sum of the contribution ratio is 1.

[0031] The present application provides an embodiment as follows: when the precipitation increases, the soil moisture increases accordingly, and the microbial nitrification rate accelerates, thereby producing a large amount of , reducing soil On the contrary, when precipitation decreases, microbial nitrification almost stops and the soil The proportion of The value increases.

[0032] Step S5 includes: The deep learning model includes: a convolutional neural network and a self-attention mechanism module; The convolutional neural network includes: an input layer, a hidden layer and an output layer.

[0033] The present application provides an embodiment as follows: a convolutional neural network is selected, and the input features include Based on the oxygen isotope data and other environmental variables, multiple hidden layers are designed to extract deep and complex feature relationships, and the relu function is used as the activation function to ensure computational efficiency, sparsity and convergence while avoiding gradient disappearance.

[0034] The self-attention mechanism is introduced to perform linear transformation on the input features, generate query, key and value vectors, calculate the similarity between the query and all keys, generate attention weights, and finally perform weighted summation on the value vectors to form an output with attention. This enhances the model's ability to capture long-term dependencies. The model structure consists of an input layer, a self-attention mechanism layer, a deep neural network layer, and an output layer. Figure 2 Specifically include: (1) Input layer The input layer contains several features related to high-salinity dust pollution, including but not limited to: Environmental factors: wind speed, wind direction, temperature, humidity, air pressure, etc.

[0035] Characteristics of sandstorms: frequency, intensity, particle size distribution, etc. of sandstorms.

[0036] Geographical and regional characteristics: distribution of saline-alkali land, degree of desertification, soil salt concentration, location of power facilities, etc.

[0037] Time characteristics: sampling time period (before, during and after dust emission).

[0038] (2) Self-attention mechanism layer Feature weight learning: Each input feature (such as wind speed, dust concentration, salt content, etc.) has a different impact on the output. The self-attention mechanism learns the contribution of each feature to pollution changes by assigning different weights.

[0039] Modeling temporal and spatial dependencies: Self-attention can help the model capture the dependencies between different time steps or spatial positions, such as changes in pollution characteristics before, during, and after dust emission; or the pollution diffusion relationship between different regions.

[0040] Multi-dimensional feature association: The self-attention mechanism can effectively associate the interactions between multi-dimensional features, such as the relationship between the salt content of saline-alkali land and the intensity and wind speed of sandstorms.

[0041] (3) Deep Neural Network Layer Deep networks can automatically learn complex patterns of input data and process nonlinear relationships in data. Each layer of neurons introduces nonlinear transformations through the activation function ReLU to enhance the model's expressiveness.

[0042] (4) Output layer Intensity of source-pollution relationship: By predicting the relationship between pollution sources and pollution intensity, the model can quantitatively evaluate the impact of different environmental factors on high-salinity dust pollution.

[0043] Temporal variation trend of pollution intensity: Based on the input time characteristics, the model can predict the temporal variation trend of pollution intensity and analyze the contribution of pollution sources at different stages to pollution.

[0044] Risk assessment of power facilities: Based on the predicted pollution intensity in different regions and climate conditions, the model can assess the pollution risk of power facilities.

[0045] The cross entropy loss function is used as a cost function to train the deep learning network model.

[0046] Please refer to Figure 3 , Figure 3It is a structural schematic diagram of a high-salinity dust natural pollution source area source-pollution relationship analysis system in an embodiment of the present application, the system includes: a sample collection and preprocessing module, a chemical composition analysis module, a data analysis module and a display module; The sample collection and preprocessing module is used to collect source samples and perform preprocessing; the source samples include: dust samples, precipitation samples and soil samples; The chemical composition analysis module is used to determine the content of the isotope mass balance model in soil samples in combination with pre-treated source samples. of Value, extract oxygen isotope data; The data analysis module is used to build a deep neural model based on the self-attention mechanism; the deep neural model is trained through the source sample data and oxygen isotope data, and the output is a relationship model between the source-pollution characteristics of high-salinity dust and environmental changes; The display module is used to visualize the relationship model between the source-pollution characteristics of high-salinity dust and environmental changes.

[0047] As an example, the whole process is as follows: According to the distribution of anthropogenic land use, typical salinized areas in the corridor were selected, and dust end-members and external insulation pollution sampling points were set up near transmission lines, substations, etc. to obtain source samples of high-salinity natural pollution and human-influenced pollution.

[0048] The collected samples were pre-treated, and the salt, heavy metal and organic components in the samples were analyzed, and isotope analysis was performed. of The values ​​are combined with the changes in precipitation and microbial activity to infer the source of high-salt pollution and its dynamic changes, thus establishing a method for analyzing high-salt pollution components.

[0049] Integrate the collected data and use deep neural networks to The oxygen isotope data of high-salinity dust were incorporated into the model, and a relationship model between the source-pollution characteristics of high-salinity dust and environmental changes was established, and the results were displayed visually.

[0050] Remove missing values ​​or perform interpolation to ensure data integrity. Standardize or normalize data to extract relevant oxygen isotope-related characteristics.

[0051] Select convolutional neural network, the input features include Based on the oxygen isotope data and other environmental variables, multiple hidden layers are designed to extract deep and complex feature relationships, and the relu function is used as the activation function to ensure computational efficiency, sparsity and convergence while avoiding gradient disappearance.

[0052] The self-attention mechanism is introduced to perform linear transformation on the input features, generate query, key and value vectors, calculate the similarity between the query and all keys, generate attention weights, and finally perform weighted summation on the value vectors to form an output with attention. This enhances the model's ability to capture long-term dependencies.

[0053] The cross entropy loss function is used as the cost function to train the deep learning network model, and the learning rate is dynamically adjusted to improve the training efficiency.

[0054] Iteratively update the model parameters in the deep learning model until the accuracy of the deep learning model in distinguishing dirt is greater than a preset accuracy threshold, and the deep learning model training is judged to be completed.

[0055] Compare the model's prediction results with the actual data to analyze the model's prediction accuracy and stability. Use Matplotlib to display the results.

[0056] The above are only exemplary embodiments of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure.

[0057] This application is intended to cover any variation, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art not described in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A method for analyzing the source-pollution relationship of a high-salinity dust natural pollution source area, characterized in that: The method comprises the following steps: S1: Set up sampling points in the high-salinity dust natural pollution source area, collect source samples and pre-treat them; source samples include: dust samples, precipitation samples and soil samples; S2: Based on the isotope mass balance model, the isotope mass in soil samples was determined by combining the pre-treated source samples. of Value, extract oxygen isotope data; S3: Construct a deep neural model based on the self-attention mechanism; train the deep neural model through the source sample data and oxygen isotope data, output the relationship model between the source-pollution characteristics of high-salinity dust and environmental changes and visualize it.

2. A method for analyzing the source-pollution relationship of a high-salinity dust natural pollution source area as claimed in claim 1, characterized in that: Step S1 includes: The high-salinity dust natural pollution source areas include: power transmission lines and substations; Within the preset range of transmission lines and substations, dust end members and external insulation pollution sampling points are set up to obtain source samples of high-salt natural pollution and human-influenced pollution.

3. The method for analyzing the source-pollution relationship of a high-salinity dust natural pollution source area according to claim 1, characterized in that: The preprocessing comprises the following steps: analyzing nitrate data and organic component data in the source sample.

4. A method for analyzing the source-pollution relationship of a high-salinity dust natural pollution source area as claimed in claim 3, characterized in that: Step S2 includes: using ion exchange method to combine nitrate data and organic matter composition data to determine the content of nitrate in soil samples. of Oxygen isotope data.

5. A method for analyzing the source-pollution relationship of a high-salinity dust natural pollution source area as claimed in claim 4, characterized in that: Step S3 includes: The soil sample of The value can be expressed by the following formula: in, In soil samples of value; Atmospheric deposition of The value of dust sample of value; Indicates the biological nitrification of Value, that is, the value of organic component data of value; and Represents the atmospheric deposition and nitrification effects on soil The sum of the contribution ratio is 1.

6. A method for analyzing the source-pollution relationship of a high-salinity dust natural pollution source area as claimed in claim 1, characterized in that: Step S5 includes: The deep learning model includes: a convolutional neural network and a self-attention mechanism module; The convolutional neural network includes: an input layer, a hidden layer and an output layer.

7. The method for analyzing the source-pollution relationship of a high-salinity dust natural pollution source area according to claim 1, characterized in that: The cross entropy loss function is used as a cost function to train the deep learning network model.

8. A high-salinity dust natural pollution source area source-pollution relationship analysis system, used to implement a high-salinity dust natural pollution source area source-pollution relationship analysis method as described in any one of claims 1 to 7, characterized in that: The system includes: a sample collection and preprocessing module, a chemical composition analysis module, a data analysis module and a display module; The sample collection and preprocessing module is used to collect source samples and perform preprocessing; the source samples include: dust samples, precipitation samples and soil samples; The chemical composition analysis module is used to determine the content of the isotope mass balance model in soil samples in combination with pre-treated source samples. of Value, extract oxygen isotope data; The data analysis module is used to build a deep neural model based on the self-attention mechanism; the deep neural model is trained through the source sample data and oxygen isotope data, and the output is a relationship model between the source-pollution characteristics of high-salinity dust and environmental changes; The display module is used to visualize the relationship model between the source-pollution characteristics of high-salinity dust and environmental changes.

Images