Gaseous pollutant diffusion rule modeling method and system
By using convolutional neural network modeling, the problem of inaccurate predictions in actual monitoring by traditional models is solved, and accurate prediction of gaseous pollutant concentrations is achieved, thus improving the performance of the model.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-09-01
- Publication Date
- 2026-04-14
AI Technical Summary
Traditional gaseous pollutant prediction models are difficult to accurately reflect the complexity of the atmosphere and the uncertainty of meteorological conditions in actual monitoring, resulting in inaccurate concentration predictions.
Convolutional Neural Networks (CNNs) are used to model the diffusion patterns of gaseous pollutants. By preprocessing experimental sample data of gaseous fluids, convolutional layers, pooling layers, and fully connected layers are constructed. The network is trained using the backpropagation algorithm to minimize the prediction error, and the model is optimized through evaluation metrics.
It enables effective prediction of gaseous pollutant concentrations during actual monitoring, improving the accuracy and precision of the model.
Smart Images
Figure CN121859685A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of deep learning technology, and in particular to a method and system for modeling the diffusion law of gaseous pollutants. Background Technology
[0002] In environmental monitoring, the prediction of air pollutant concentrations is crucial and plays a significant role in the prevention and control of air pollution.
[0003] Traditional gaseous pollutant prediction models are established and analyzed based on gas diffusion patterns obtained from wind tunnel experiments or data simulations. However, in actual monitoring work, due to the complexity of the atmosphere and the uncertainty of meteorological conditions, it is difficult to obtain accurate gaseous pollutant diffusion models. Therefore, how to establish a suitable model using experimental data is of great significance for predicting the concentration of gaseous pollutants in actual monitoring processes. Summary of the Invention
[0004] Based on this, the embodiments of this application provide a method and system for modeling the diffusion law of gaseous pollutants, which can effectively predict the concentration of gaseous pollutants during actual monitoring.
[0005] Firstly, a method for modeling the diffusion patterns of gaseous pollutants is provided, the method comprising:
[0006] Acquire experimental sample data of gaseous fluids and preprocess the sample data to obtain a training dataset; wherein, the experimental sample data of gaseous fluids includes the concentration of gaseous pollutants and meteorological conditions;
[0007] Construct a convolutional neural network; specifically, construct and configure convolutional layers, pooling layers, and fully connected layers.
[0008] The constructed convolutional neural network is trained using a preprocessed sample dataset; the weights and biases in the network are continuously adjusted through the backpropagation algorithm to minimize the error between the predicted results and the true values.
[0009] After training is completed, the trained convolutional neural network is used to extract features from new samples and evaluate them. When the evaluation is successful, the current convolutional neural network is output as a model of the diffusion law of gaseous pollutants.
[0010] Optionally, feature extraction and evaluation of new samples can be performed using a trained convolutional neural network, including:
[0011] The performance of the model is measured by comparing the predicted results with the true values in the test set; wherein the evaluation metrics include accuracy, recall, precision, and F1 score.
[0012] Optionally, the method further includes:
[0013] If the evaluation fails, the model parameters are adjusted, and additional training sets are added for continuous cumulative updates and iterations.
[0014] Optionally, constructing the convolutional neural network includes:
[0015] A 3x3 convolutional kernel is used in the convolutional layer to generate 64 feature maps, and the ReLU activation function is applied.
[0016] Optionally, the constructed convolutional neural network is trained using the preprocessed sample dataset, including:
[0017] Stochastic gradient descent is used for optimization training.
[0018] Secondly, a system for modeling the diffusion patterns of gaseous pollutants is provided, the system comprising:
[0019] The acquisition module is used to acquire gaseous fluid experimental sample data and preprocess the sample data to obtain a training dataset; wherein, the gaseous fluid experimental sample data includes the concentration of gaseous pollutants and meteorological conditions;
[0020] The building module is used to construct convolutional neural networks; it includes the configuration of convolutional layers, pooling layers, and fully connected layers.
[0021] The training module is used to train the constructed convolutional neural network using a preprocessed sample dataset; the weights and biases in the network are continuously adjusted through the backpropagation algorithm to minimize the error between the predicted results and the true values.
[0022] The output module is used to extract features from new samples and evaluate them after training is completed. When the evaluation is successful, the current convolutional neural network is output as a model of the diffusion law of gaseous pollutants.
[0023] Optionally, the output module extracts and evaluates features from new samples using a trained convolutional neural network, including:
[0024] The performance of the model is measured by comparing the predicted results with the true values in the test set; wherein the evaluation metrics include accuracy, recall, precision, and F1 score.
[0025] Optionally, the output module further includes:
[0026] If the evaluation fails, the model parameters are adjusted, and additional training sets are added for continuous cumulative updates and iterations.
[0027] Optionally, the building module constructs a convolutional neural network, including:
[0028] A 3x3 convolutional kernel is used in the convolutional layer to generate 64 feature maps, and the ReLU activation function is applied.
[0029] Optionally, the training module trains the constructed convolutional neural network using the preprocessed sample dataset, including:
[0030] Stochastic gradient descent is used for optimization training.
[0031] The technical solution provided in this application first acquires experimental sample data of gaseous fluid and preprocesses the sample data to obtain a training dataset; then, a convolutional neural network is constructed; the constructed convolutional neural network is trained using the preprocessed sample dataset; after training, the trained convolutional neural network is used to extract features from new samples and evaluate them. When the evaluation is successful, the current convolutional neural network is output as a model of the diffusion law of gaseous pollutants. It can be seen that the beneficial effect of this invention is that it can effectively predict the concentration of gaseous pollutants during actual monitoring. Attached Figure Description
[0032] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.
[0033] Figure 1 A flowchart illustrating a method for modeling the diffusion patterns of gaseous pollutants, as provided in this application embodiment;
[0034] Figure 2 A block diagram of a gaseous pollutant diffusion law modeling system provided in this application embodiment. Detailed Implementation
[0035] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0036] In the description of this invention, the terms “comprising,” “having,” and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units that are expressly listed, but may also include other steps or units that are not expressly listed but are inherent to these processes, methods, products, or apparatuses, or steps or units added based on further optimizations of the inventive concept.
[0037] For details, please refer to Figure 1 The diagram illustrates a flowchart of a method for modeling the diffusion law of gaseous pollutants provided in an embodiment of this application. This method may include the following steps:
[0038] Step 101: Obtain experimental sample data of gaseous fluid and preprocess the sample data to obtain a training dataset.
[0039] The experimental sample data for gaseous fluids includes the concentration of gaseous pollutants and meteorological conditions.
[0040] Use appropriate sensors or monitoring equipment to acquire relevant data from gaseous fluid experimental samples. This data may include meteorological conditions such as gaseous pollutant concentration, temperature, humidity, wind speed, and wind direction. Then, clean the collected raw data to remove outliers, noise, and missing values. Statistical methods and filtering techniques can be used to process and repair the data. Normalize the cleaned data to make the numerical ranges of different features similar. Common normalization methods include maximum-minimum normalization and standardization.
[0041] The processed sample data is used as the training set, and the training set is converted into a format suitable for model training. For convolutional neural networks, this is typically done by converting the data into tensor form.
[0042] Step 102: Construct a convolutional neural network.
[0043] The system includes convolutional layers, pooling layers, and fully connected layers.
[0044] In this embodiment, the convolutional neural network (CNN) structure may include multiple convolutional layers, pooling layers, and fully connected layers. The appropriate number of network layers can be selected based on the complexity of the specific problem and the size of the dataset. For example, 2-4 convolutional layers can be used.
[0045] A convolution kernel is a filter in a convolution operation used to extract local features from an image or data. The choice of kernel size is typically based on the size of the dataset and the size of the features to be captured. For example, for two-dimensional image data, common kernel sizes are 3x3 or 5x5.
[0046] Pooling layers are used to reduce the spatial size of feature maps while retaining key information. Common pooling methods include max pooling and average pooling. Max pooling selects the maximum value within each region as the pooling result, while average pooling takes the average value within the region. The choice of which pooling method to use depends on the requirements of the problem and the characteristics of the data.
[0047] Activation functions are used to introduce non-linear properties, enabling models to learn more complex features. Common activation functions include ReLU, Sigmoid, and Tanh. For example, the ReLU activation function can be used to perform non-linear mapping after convolutional layers.
[0048] Step 103: Train the constructed convolutional neural network using the preprocessed sample dataset.
[0049] In this process, the weights and biases in the network are continuously adjusted through the backpropagation algorithm to minimize the error between the predicted results and the true values.
[0050] Step 104: After training is completed, the new samples are used to extract features and evaluate them by using the trained convolutional neural network. When the evaluation is passed, the current convolutional neural network is output as a model of the diffusion law of gaseous pollutants.
[0051] The following is a general model evaluation process in the embodiments of this application:
[0052] The established model is used to predict the samples in the test set, and the prediction results are obtained. For example, the diffusion law value output by the model is predicted based on the input meteorological conditions and gaseous pollutant concentrations.
[0053] Evaluation metrics are calculated to measure model performance by comparing predicted results with true values in the test set. Common evaluation metrics include accuracy, recall, precision, and F1 score. These metrics help evaluate a model's predictive ability, classification ability, or regression accuracy.
[0054] The performance of the model is analyzed and interpreted based on the values of the evaluation indicators. The evaluation results of different models are compared to identify potential problems or directions for improvement.
[0055] Based on the evaluation results, the model can be adjusted and optimized. For example, methods such as adjusting model parameters, increasing the amount of training data, and introducing regularization can be tried to improve model performance.
[0056] If the model is not ideal enough, multiple iterations of model evaluation and optimization can be performed until a satisfactory performance level is achieved.
[0057] like Figure 2 This application also provides a system for modeling the diffusion patterns of gaseous pollutants. The system includes:
[0058] The acquisition module is used to acquire gaseous fluid experimental sample data and preprocess the sample data to obtain a training dataset; the gaseous fluid experimental sample data includes the concentration of gaseous pollutants and meteorological conditions.
[0059] The building module is used to construct convolutional neural networks; it includes the configuration of convolutional layers, pooling layers, and fully connected layers.
[0060] The training module is used to train the constructed convolutional neural network using a preprocessed sample dataset; the weights and biases in the network are continuously adjusted through the backpropagation algorithm to minimize the error between the predicted results and the true values.
[0061] The output module is used to extract features from new samples and evaluate them after training is completed. When the evaluation is successful, the current convolutional neural network is output as a model of the diffusion law of gaseous pollutants.
[0062] In optional embodiments of this application, the output module extracts features from and evaluates new samples using a trained convolutional neural network, including:
[0063] The performance of the model is measured by comparing the predicted results with the true values in the test set; the evaluation metrics include accuracy, recall, precision, and F1 score.
[0064] In an optional embodiment of this application, the output module further includes:
[0065] If the evaluation fails, the model parameters are adjusted, and additional training sets are added for continuous cumulative updates and iterations.
[0066] In an optional embodiment of this application, the construction module constructs a convolutional neural network, including:
[0067] A 3x3 convolutional kernel is used in the convolutional layer to generate 64 feature maps, and the ReLU activation function is applied.
[0068] In an optional embodiment of this application, the training module trains the constructed convolutional neural network using a preprocessed sample dataset, including:
[0069] Stochastic gradient descent is used for optimization training.
[0070] The gaseous pollutant diffusion law modeling system provided in this application embodiment is used to implement the above-described gaseous pollutant diffusion law modeling method. Specific limitations of the gaseous pollutant diffusion law modeling system can be found in the limitations of the gaseous pollutant diffusion law modeling method described above, and will not be repeated here. Each part of the above-described gaseous pollutant diffusion law modeling system can be implemented entirely or partially through software, hardware, or a combination thereof. Each module can be embedded in or independent of the processor in the device in hardware form, or stored in the memory of the device in software form, so that the processor can call and execute the operations corresponding to each module.
[0071] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0072] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.
Claims
1. A method for modeling the diffusion law of gaseous pollutants, characterized in that, The method includes: Acquire experimental sample data of gaseous fluids and preprocess the sample data to obtain a training dataset; wherein, the experimental sample data of gaseous fluids includes the concentration of gaseous pollutants and meteorological conditions; Construct a convolutional neural network; specifically, construct and configure convolutional layers, pooling layers, and fully connected layers. The constructed convolutional neural network is trained using a preprocessed sample dataset; the weights and biases in the network are continuously adjusted through the backpropagation algorithm to minimize the error between the predicted results and the true values. After training is completed, the trained convolutional neural network is used to extract features from new samples and evaluate them. When the evaluation is successful, the current convolutional neural network is output as a model of the diffusion law of gaseous pollutants.
2. The method for modeling the diffusion law of gaseous pollutants according to claim 1, characterized in that, Feature extraction and evaluation of new samples are performed using a trained convolutional neural network, including: The performance of the model is measured by comparing the predicted results with the true values in the test set; wherein the evaluation metrics include accuracy, recall, precision, and F1 score.
3. The method for modeling the diffusion law of gaseous pollutants according to claim 1, characterized in that, The method further includes: If the evaluation fails, the model parameters are adjusted, and additional training sets are added for continuous cumulative updates and iterations.
4. The method for modeling the diffusion law of gaseous pollutants according to claim 1, characterized in that, The construction of the convolutional neural network includes: A 3x3 convolutional kernel is used in the convolutional layer to generate 64 feature maps, and the ReLU activation function is applied.
5. The method for modeling the diffusion law of gaseous pollutants according to claim 1, characterized in that, The constructed convolutional neural network is trained using the preprocessed sample dataset, including: Stochastic gradient descent is used for optimization training.
6. A system for modeling the diffusion patterns of gaseous pollutants, characterized in that, The system includes: The acquisition module is used to acquire gaseous fluid experimental sample data and preprocess the sample data to obtain a training dataset; wherein, the gaseous fluid experimental sample data includes the concentration of gaseous pollutants and meteorological conditions; The building module is used to construct convolutional neural networks; it includes the configuration of convolutional layers, pooling layers, and fully connected layers. The training module is used to train the constructed convolutional neural network using a preprocessed sample dataset; the weights and biases in the network are continuously adjusted through the backpropagation algorithm to minimize the error between the predicted results and the true values. The output module is used to extract features from new samples and evaluate them after training is completed. When the evaluation is successful, the current convolutional neural network is output as a model of the diffusion law of gaseous pollutants.
7. The gaseous pollutant diffusion law modeling system according to claim 6, characterized in that, The output module extracts features from new samples and evaluates them using a trained convolutional neural network, including: The performance of the model is measured by comparing the predicted results with the true values in the test set; wherein the evaluation metrics include accuracy, recall, precision, and F1 score.
8. The gaseous pollutant diffusion law modeling system according to claim 6, characterized in that, The output module also includes: If the evaluation fails, the model parameters are adjusted, and additional training sets are added for continuous cumulative updates and iterations.
9. The gaseous pollutant diffusion law modeling system according to claim 6, characterized in that, The building blocks for constructing convolutional neural networks include: A 3x3 convolutional kernel is used in the convolutional layer to generate 64 feature maps, and the ReLU activation function is applied.
10. The gaseous pollutant diffusion law modeling system according to claim 6, characterized in that, The training module trains the constructed convolutional neural network using the preprocessed sample dataset, including: Stochastic gradient descent is used for optimization training.