Tail gas treatment system and method for the production process of refinery catalysts

Abstract

This application relates to the field of tail gas treatment. Specifically, it discloses a tail gas treatment system and method for the production process of refining catalysts. First, it obtains the infrared spectrogram of the tail gas treatment in the production process of refining catalysts collected by an infrared spectrometer and the waveform diagram of the particulate echo signals at multiple predetermined time points collected by a particulate lidar. Then, using deep learning technology, it extracts features and conducts correlation analysis on the two. Finally, a classifier is used to obtain a classification result to determine whether the removal degree of the particulate matter meets the standard, thereby realizing an automated monitoring and evaluation process, and then timely discovering problems and taking measures for improvement.

Description

Technical Field

[0001] This application relates to the field of tail gas treatment, and more specifically, to a tail gas treatment system and method for the production process of refining catalysts. Background Art

[0002] Refining catalysts are crucial materials in the refining industry. By providing active sites on their surfaces, the catalysts can accelerate chemical reactions in the refining process, such as cracking, reforming, and hydrogenation, thereby improving the product quality and yield. The calcination stage in the production process of refining catalysts is a key step, where the catalyst is heated at high temperature to promote the combination of metal oxides and carriers to form active particles.

[0003] The production process of refining catalysts involves a series of complex steps, including raw material preparation, shaping, drying, calcination, and activation. In the initial stage of construction of existing refining catalyst devices, no supporting tail gas purification unit was considered, and the tail gas was directly discharged into the atmosphere. However, the calcination process generates tail gas, which may contain harmful gases and particulate matter, such as nitrogen oxides and volatile organic compounds.

[0004] Therefore, a tail gas treatment system and method for the production process of refining catalysts are desired. Summary of the Invention

[0005] To solve the above technical problems, this application is proposed. Embodiments of this application provide a tail gas treatment system and method for the production process of refining catalysts, which first obtain the infrared spectrogram of the tail gas treatment in the production process of refining catalysts collected by an infrared spectrometer and the waveform diagram of the particulate echo signals at multiple predetermined time points collected by a particulate lidar, then use deep learning technology to perform feature extraction and correlation analysis on the two, and finally obtain a classification result through a classifier to determine whether the removal degree of particulate matter meets the standard, so as to realize an automated monitoring and evaluation process, and then timely discover problems and take measures for improvement.

[0006] According to one aspect of this application, a tail gas treatment system for the production process of refining catalysts is provided, which includes:

[0007] A catalyst production process data acquisition module, configured to obtain the infrared spectrogram of the tail gas treatment in the production process of refining catalysts collected by an infrared spectrometer and the waveform diagram of the particulate echo signals at multiple predetermined time points collected by a particulate lidar;

[0008] A catalyst production process data extraction module, configured to extract a global correlation feature vector of the infrared spectrum of the tail gas treatment and a particulate echo feature vector from the infrared spectrogram of the tail gas treatment in the production process of refining catalysts collected by the infrared spectrometer and the waveform diagram of the particulate echo signals at multiple predetermined time points collected by the particulate lidar;

[0009] An exhaust gas particle removal degree judgment module, which is used to judge whether the removal degree of the particles meets the standard based on the global correlation feature vector of the exhaust gas treatment infrared spectrum and the particle echo feature vector.

[0010] According to another aspect of the present application, a method for treating exhaust gas in the production process of a refining catalyst is provided, which includes:

[0011] Obtain the infrared spectrum diagram of the exhaust gas treatment in the production process of the refining catalyst collected by an infrared spectrometer and the waveform diagram of the particle echo signals at multiple predetermined time points collected by a particle lidar.

[0012] Extract the global correlation feature vector of the exhaust gas treatment infrared spectrum and the particle echo feature vector from the infrared spectrum diagram of the exhaust gas treatment in the production process of the refining catalyst collected by the infrared spectrometer and the waveform diagram of the particle echo signals at multiple predetermined time points collected by the particle lidar.

[0013] Based on the global correlation feature vector of the exhaust gas treatment infrared spectrum and the particle echo feature vector, judge whether the removal degree of the particles meets the standard.

[0014] Compared with the prior art, a system and method for treating exhaust gas in the production process of a refining catalyst provided by the present application first obtain the infrared spectrum diagram of the exhaust gas treatment in the production process of the refining catalyst collected by an infrared spectrometer and the waveform diagram of the particle echo signals at multiple predetermined time points collected by a particle lidar, then use deep learning technology to perform feature extraction and correlation analysis on the two, and finally obtain a classification result through a classifier to judge whether the removal degree of the particles meets the standard, so as to realize an automated monitoring and evaluation process, and then timely discover problems and take measures for improvement. Description of the Drawings

[0015] By describing the embodiments of the present application in more detail in conjunction with the drawings, the above and other objects, features, and advantages of the present application will become more obvious. The drawings are used to provide a further understanding of the embodiments of the present application, and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation to the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0016] Figure 1 It is a block diagram of a system for treating exhaust gas in the production process of a refining catalyst according to an embodiment of the present application.

[0017] Figure 2 It is a block diagram of a catalyst production process data extraction module in a system for treating exhaust gas in the production process of a refining catalyst according to an embodiment of the present application.

[0018] Figure 3 It is a block diagram of an exhaust gas particle removal degree judgment module in an exhaust gas treatment system for a refinery catalyst production process according to an embodiment of the present application.

[0019] Figure 4 It is a flowchart of an exhaust gas treatment method for a refinery catalyst production process according to an embodiment of the present application.

[0020] Figure 5 It is a block diagram of an electronic device according to an embodiment of the present application. Detailed implementation manners

[0021] Next, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the exemplary embodiments described herein.

[0022] Exemplary system

[0023] Figure 1 It is a block diagram of an exhaust gas treatment system for a refinery catalyst production process according to an embodiment of the present application. As Figure 1 shown, the exhaust gas treatment system 100 for a refinery catalyst production process according to an embodiment of the present application includes: a catalyst production process data acquisition module 110, configured to acquire an infrared spectrogram of exhaust gas treatment in a refinery catalyst production process collected by an infrared spectrometer and waveforms of particulate echo signals at multiple predetermined time points collected by a particulate lidar; a catalyst production process data extraction module 120, configured to extract a global correlation feature vector of exhaust gas treatment infrared spectrum and a particulate echo feature vector from the infrared spectrogram of exhaust gas treatment in a refinery catalyst production process collected by the infrared spectrometer and the waveforms of particulate echo signals at multiple predetermined time points collected by the particulate lidar; an exhaust gas particle removal degree judgment module 130, configured to judge whether the removal degree of particulate matter meets the standard based on the global correlation feature vector of exhaust gas treatment infrared spectrum and the particulate echo feature vector.

[0024] In the tail gas treatment system 100 for the production process of the above-mentioned oil refining catalyst, the catalyst production process data acquisition module 110 is used to acquire the infrared spectrogram of the tail gas treatment in the production process of the oil refining catalyst collected by an infrared spectrometer and the waveform diagram of the particulate echo signals at multiple predetermined time points collected by a particulate lidar. It should be understood that oil refining catalysts play a crucial role in the oil refining industry. These catalysts can promote chemical reactions in the oil refining process, such as cracking, reforming, and hydrogenation, etc., by providing active sites on their surfaces, thereby improving the quality and yield of products. In the production process of oil refining catalysts, the calcination stage is a key step. In this stage, the catalyst is heated at a high temperature to promote the combination of metal oxides and carriers to form active particles, thus ensuring the activity and stability of the catalyst. However, the production process of oil refining catalysts involves multiple complex steps, including raw material preparation, shaping, drying, calcination, and activation, etc. The current problem is that some oil refining catalyst devices do not consider the corresponding tail gas purification unit during the initial construction stage, resulting in the direct emission of the tail gas generated during the calcination process into the atmosphere. These tail gases may contain harmful gases and particulate matters, such as nitrogen oxides and volatile organic compounds, causing potential environmental pollution and health risks. To solve this problem, it is crucial to effectively treat the tail gas generated during the production process of oil refining catalysts. In the technical solution of this application, by acquiring the infrared spectrogram of the tail gas treatment in the production process of the oil refining catalyst collected by an infrared spectrometer and the waveform diagram of the particulate echo signals at multiple predetermined time points collected by a particulate lidar, and combining deep learning technology, it is judged whether the removal degree of particulate matter meets the standard. This method realizes an automated monitoring and evaluation process, can timely discover problems and take necessary measures for improvement, thereby ensuring that the production meets the standard requirements and improving production efficiency and product quality.

[0025] Specifically, the infrared spectrogram provides characteristic spectral information of various chemical components in the tail gas. By analyzing this information, the content and types of harmful substances in the tail gas can be understood, thereby evaluating the effect of tail gas treatment. The waveform diagram of the particulate echo signal reflects the distribution and concentration of particulate matter in the air over time. By monitoring these signals, the removal situation of particulate matter can be understood in real time. Among them, the infrared spectrogram is collected by an infrared spectrometer, and the structure and composition of substances are analyzed by using the characteristics of substances absorbing, emitting, or scattering infrared light. During the production process of refining catalysts, the chemical components in the tail gas will produce specific absorption peaks on the spectrogram. By analyzing the positions and intensities of these absorption peaks, the presence and concentration of various components in the tail gas can be determined. Such analysis helps to monitor the effect of tail gas treatment during the catalyst production process, promptly discover problems, and make adjustments and improvements. On the other hand, the waveform diagram of the particulate echo signal collected by the particulate lidar provides information about the concentration and distribution of particulate matter. The characteristics of the particulate echo signal can reflect key parameters such as the morphology, size, and concentration of particulate matter. By monitoring the changes in these signals, the removal situation of particulate matter can be understood in real time. In this way, the chemical components and particulate matter emissions during the production process of refining catalysts can be analyzed and monitored to evaluate the catalyst performance, monitor the pollutant emission levels, and optimize the production process to ensure the safety and efficiency of the environment and production.

[0026] In the above-mentioned tail gas treatment system 100 for the production process of refining catalysts, the catalyst production process data extraction module 120 is used to extract the global correlation feature vector of the tail gas treatment infrared spectrum and the particulate echo feature vector from the infrared spectrogram of the tail gas treatment in the production process of refining catalysts collected by the infrared spectrometer and the waveform diagrams of the particulate echo signals at multiple predetermined time points collected by the particulate lidar. It should be understood that the global correlation feature vector of the tail gas treatment infrared spectrum is extracted by analyzing the infrared spectrogram. It contains the overall characteristic information of the chemical components during the tail gas treatment process, can reflect the mutual correlation of various components in the tail gas, and the overall effect of tail gas treatment. This global correlation feature vector is crucial for judging the comprehensive effect of tail gas treatment and can help determine the removal degree and treatment efficiency of harmful substances in the tail gas. On the other hand, the particulate echo feature vector is extracted from the waveform diagram of the particulate echo signal. These feature vectors can reveal important information such as the distribution, concentration, and particle size of particulate matter in the tail gas. By analyzing the particulate echo feature vector, the removal situation and treatment effect of particulate matter can be understood, and it helps to evaluate whether the removal degree of particulate matter in the tail gas meets the relevant standards. Combining these two feature vectors can comprehensively consider the removal of chemical components and particulate matter during the tail gas treatment process, thereby more accurately judging whether the effect of tail gas treatment meets the expected requirements.

[0027] Figure 2The block diagram of the catalyst production process data extraction module in the tail gas treatment system for the production process of the oil refining catalyst according to the embodiments of the present application. As Figure 2 shown, in a specific embodiment of the present application, the catalyst production process data extraction module 120 includes: a spectral counting feature extraction unit 121, configured to perform spectral counting feature extraction on the infrared spectrogram of the tail gas treatment in the oil refining catalyst production process collected by the infrared spectrometer to obtain a tail gas treatment infrared spectrum statistical feature vector; a spectral image feature extraction unit 122, configured to perform spectral image feature extraction on the infrared spectrogram of the tail gas treatment in the oil refining catalyst production process collected by the infrared spectrometer to obtain a tail gas treatment infrared spectrum feature vector; a spectral feature correlation unit 123, configured to correlate the tail gas treatment infrared spectrum statistical feature vector and the tail gas treatment infrared spectrum feature vector to obtain a tail gas treatment infrared spectrum global correlation feature vector; and an echo signal feature encoding unit 124, configured to perform feature encoding on the waveform diagrams of the particulate echo signals at a plurality of predetermined time points collected by the particulate lidar to obtain the particulate echo feature vector.

[0028] It should be understood that spectral counting feature extraction refers to counting and statistical analysis of the absorption peaks in the spectrogram to obtain a feature vector containing information such as the number, position, and intensity of the absorption peaks. These statistical feature vectors can reflect the content and distribution of different chemical components in the tail gas. Through spectral counting feature extraction, a tail gas treatment infrared spectrum statistical feature vector can be obtained, which contains the characteristic statistical information of various chemical components in the spectrogram. These statistical feature vectors can help identify and distinguish different chemical components, understand their relative content and change trends in the tail gas. By analyzing these statistical feature vectors, potential problems or abnormalities in the tail gas can be detected in a timely manner, providing guidance for adjusting and optimizing the tail gas treatment system.

[0029] Furthermore, spectral image feature extraction refers to image processing and analysis of the spectrogram to extract key information representing spectral features. Through spectral image feature extraction, complex spectral data can be converted into feature vectors with practical significance, which can be used to describe the characteristics of different chemical components in the tail gas treatment process, helping to identify and distinguish different compounds or functional groups. By analyzing these feature vectors, the content, change trends, and mutual relationships of various chemical components in the tail gas can be understood, thereby evaluating the progress of chemical reactions and the removal effect of harmful substances in the tail gas treatment process.

[0030] Furthermore, by correlating statistical features with image features, their advantages can be combined to improve the understanding and analysis capabilities of the tail gas treatment process. Among them, statistical features provide statistical information on the content and distribution of chemical components, while image features provide an intuitive description of spectral morphology and characteristics. By combining these two types of features, a more comprehensive understanding of the characteristics of different chemical components in the tail gas can be obtained, including information on their content, distribution, morphology, etc., thereby more accurately evaluating the tail gas treatment effect and chemical changes.

[0031] In particular, by feature encoding the particulate echo signal, real-time monitoring and identification of particulate matter can be achieved, helping to understand the distribution of particulate matter in the environment, the concentration change trend, and its impact on the environment and health. These feature vectors can provide a more comprehensive and accurate understanding of the characteristics and properties of particulate matter, providing a scientific basis and technical support for environmental monitoring, particulate matter control, and pollution prevention.

[0032] In a specific embodiment of the present application, the spectral count feature extraction unit 121 includes: extracting the tail gas treatment infrared spectrometer values at multiple predetermined wavelengths from the infrared spectrogram of the tail gas treatment in the production process of the refining catalyst collected by the infrared spectrometer; associating the tail gas treatment infrared spectrometer values at the multiple predetermined wavelengths through a tail gas treatment infrared spectrum count sequence correlation feature encoder including a one-dimensional convolutional layer and a fully connected layer to obtain the tail gas treatment infrared spectrum statistical feature vector.

[0033] It should be understood that different compounds have specific absorption characteristics in the infrared spectrum, and the intensity and position of their absorption peaks can provide information on the type and content of the compounds. By selecting multiple predetermined wavelengths, the characteristic absorption peaks of specific compounds in the infrared spectrum can be targeted for capture, thereby realizing the quantitative and qualitative analysis of these compounds. By analyzing the count values at multiple predetermined wavelengths, a comprehensive understanding of the changes of various compounds in the tail gas can be obtained, providing strong support for optimizing the production process, improving the environmental protection level, and ensuring production safety.

[0034] Furthermore, the one-dimensional convolutional layer can effectively capture local features and patterns in the infrared spectrum count sequence, and learn the correlation and regularity between count values at different wavelengths through convolution operations. The fully connected layer is used to integrate and map the features extracted by the convolutional layer, and further extract high-level abstract feature representations. Through such a feature encoder, the infrared spectrum count sequence can be transformed into a feature vector with higher-level semantic information, so as to better describe and understand the compound characteristics in the tail gas treatment process. The obtained tail gas treatment infrared spectrum statistical feature vector can contain richer and more meaningful information, such as feature statistical values at different wavelengths, relative contents of compounds, positions of characteristic peaks, etc. Specifically, the tail gas treatment infrared spectrum count values at the plurality of predetermined wavelengths are arranged as an infrared spectrum count input vector; each layer of the tail gas treatment infrared spectrum count sequence correlation feature encoder including the one-dimensional convolutional layer and the fully connected layer is used to respectively perform the following operations on the input data during the forward pass of the layer: performing convolution processing on the input data based on a one-dimensional convolution kernel to obtain a convolution feature map; performing pooling processing on the convolution feature map based on a feature matrix to obtain a pooling feature map; performing non-linear activation on the pooling feature map to obtain an activation feature map; wherein, the input of the first layer of the one-dimensional convolutional neural network model is the infrared spectrum count input vector, and the output of the last layer of the one-dimensional convolutional neural network model is the tail gas treatment infrared spectrum statistical feature vector.

[0035] In a specific embodiment of the present application, the spectral image feature extraction unit 122 includes: passing the infrared spectrum image of the tail gas treatment in the production process of the refinery catalyst collected by the infrared spectrometer through the tail gas treatment infrared spectrum convolutional neural network as a noise reducer to obtain a noise-reduced tail gas treatment infrared spectrum image; passing the noise-reduced tail gas treatment infrared spectrum image through the tail gas treatment infrared spectrum convolutional neural network as a feature encoder to obtain the tail gas treatment infrared spectrum feature vector.

[0036] It should be understood that the exhaust gas treatment infrared spectral convolutional neural network, as a noise reducer, can gradually extract and learn the abstract features in the infrared spectrogram through the hierarchical processing of the convolutional neural network, so as to effectively represent the exhaust gas treatment infrared spectrogram. The structure of the neural network can automatically learn the complex patterns and rules in the data, and at the same time has strong generalization ability, which can handle the differences between different exhaust gas treatment infrared spectrograms, so as to extract the key features of the exhaust gas treatment infrared spectrogram. Specifically, the infrared spectrogram of the exhaust gas treatment in the production process of the refining catalyst collected by the infrared spectrometer is input into the encoder of the exhaust gas treatment infrared spectral convolutional neural network as a noise reducer, wherein the encoder uses a convolutional layer to perform explicit spatial encoding on the infrared spectrogram of the exhaust gas treatment in the production process of the refining catalyst collected by the infrared spectrometer to obtain image features; and, the image features are input into the decoder of the exhaust gas treatment infrared spectral convolutional neural network as a noise reducer, wherein the decoder uses a transposed convolutional layer to perform transposed convolution processing on the speech features to obtain the denoised exhaust gas treatment infrared spectrogram.

[0037] Furthermore, by processing the denoised exhaust gas treatment infrared spectrogram through the exhaust gas treatment infrared spectral convolutional neural network as a feature encoder, a deep learning model can be used to extract and learn the high-level abstract features in the infrared spectral data, so as to realize the feature representation and encoding of the compounds in the exhaust gas treatment process. Through the hierarchical stacking and parameter learning of the convolutional neural network, the spatial features and frequency domain features in the infrared spectrogram can be gradually extracted, so as to realize the high-level feature representation of the compounds in the exhaust gas treatment process. The feature vectors obtained after being processed by the exhaust gas treatment infrared spectral feature encoder contain rich information in the infrared spectrogram, such as the characteristic peaks, absorption intensities, frequency distributions, etc. of the compounds. These feature vectors can better describe and characterize the compound features in the exhaust gas treatment process, and provide a more reliable feature representation for subsequent classification, recognition and quantitative analysis.

[0038] In a specific embodiment of the present application, the echo signal feature encoding unit 124 includes: passing the waveform diagrams of the particulate matter echo signals at a plurality of predetermined time points collected by the particulate matter lidar through the particulate matter echo feature encoder using the dilated convolutional neural network to obtain the particulate matter echo feature maps; and pooling the particulate matter echo feature maps to obtain the particulate matter echo feature vectors.

[0039] It should be understood that the dilated convolutional neural network has excellent feature extraction capabilities when processing time series data, and can effectively capture the temporal features and spatial relationships in the particulate matter echo signal. By performing dilated convolution operations on the waveform diagram, the network can expand the receptive field without increasing the number of parameters, thereby better capturing the long-range dependencies and global information in the signal. Through the particulate matter echo feature encoder, the particulate matter echo signal can be transformed into a feature map with higher-level semantic information, which contains the important features and structural information of the particulate matter signal. Specifically, each layer of the first convolutional neural network model using the particulate matter echo feature encoder of the dilated convolutional neural network performs dilated convolution processing based on the first convolutional kernel, pooling processing along the channel dimension, and non-linear activation processing on the input data respectively during the forward pass of the layer, so as to output a first feature map by the last layer of the first convolutional neural network model; each layer of the second convolutional neural network model using the particulate matter echo feature encoder of the dilated convolutional neural network performs dilated convolution processing based on the second convolutional kernel, pooling processing along the channel dimension, and non-linear activation processing on the input data respectively during the forward pass of the layer, so as to output a second feature map by the last layer of the second convolutional neural network model; the first feature map and the second feature map are fused to obtain the particulate matter echo feature map.

[0040] Furthermore, the process of pooling the particulate matter echo feature map aims to generate a particulate matter echo feature vector by reducing the data dimension and extracting the most significant features. Pooling operations are widely used in deep learning, which can effectively reduce the amount of data, improve the computational efficiency, and retain the most important feature information. Pooling operations are usually achieved by taking the maximum or average value of a certain region in the feature map. Among them, max pooling can retain the most significant features, while average pooling can smooth the feature map and reduce the influence of noise. Through the pooling operation, the size of the feature map can be effectively reduced, the risk of overfitting can be reduced, and at the same time, the important feature information can be retained, making the feature vector more representative and robust.

[0041] In the tail gas treatment system 100 during the production process of the above-mentioned oil refining catalyst, the tail gas particle removal degree judgment module 130 is used to judge whether the removal degree of the particulate matter meets the standard based on the global correlation feature vector of the tail gas treatment infrared spectrum and the echo feature vector of the particulate matter. It should be understood that the global correlation feature vector of the tail gas treatment infrared spectrum reflects the overall characteristics and content distribution of various compounds in the tail gas, providing comprehensive information about the tail gas treatment effect. These feature vectors can include information such as the characteristic statistical values and relative contents of various compounds, reflecting the overall situation of the compounds in the tail gas. On the other hand, the echo feature vector of the particulate matter captures the echo signal characteristics of the particulate matter at different time points, reflecting information such as the concentration, size, and distribution of the particulate matter. By encoding the particulate matter echo feature map through a dilated convolutional neural network, an abstract feature representation of the particulate matter echo can be extracted. Considering the global correlation feature vector of the tail gas treatment infrared spectrum and the echo feature vector of the particulate matter comprehensively, the information of the compounds in the tail gas and the characteristics of the particulate matter can be combined, so as to more comprehensively evaluate whether the removal degree of the particulate matter meets the standard.

[0042] Figure 3 It is a block diagram of the tail gas particle removal degree judgment module in the tail gas treatment system during the production process of the oil refining catalyst according to the embodiment of the present application. In a specific embodiment of the present application, the tail gas particle removal degree judgment module 130 includes: a tail gas treatment feature fusion unit 131, which is used to fuse the global correlation feature vector of the tail gas treatment infrared spectrum and the echo feature vector of the particulate matter to obtain a tail gas particulate matter removal judgment classification feature matrix; a tail gas treatment feature optimization unit 132, which is used to perform a probability responsiveness adjustment based on a high-dimensional feature space on the tail gas particulate matter removal judgment classification feature matrix to obtain an optimized tail gas particulate matter removal judgment classification feature matrix; a removal degree standard judgment classification unit 133, which is used to pass the optimized tail gas particulate matter removal judgment classification feature matrix through a classifier to obtain a classification result, so as to judge whether the removal degree of the particulate matter meets the standard.

[0043] It should be understood that the global correlation feature vector of the exhaust gas treatment infrared spectrum contains the statistical features of the infrared spectrometer values at different wavelengths during the exhaust gas treatment process, and can reflect the contents and characteristics of different compounds. The particulate echo feature vector describes the echo signal features of the particulate matter at different time points, and can provide information about the particulate matter concentration and distribution. By fusing these two feature vectors, the infrared spectrum and particulate echo information can be comprehensively considered to achieve a comprehensive evaluation of the particulate matter removal effect. The obtained exhaust gas particulate matter removal judgment classification feature matrix after fusion combines the infrared spectrum and particulate echo information, can more comprehensively describe the characteristics and content distribution of the particulate matter in the exhaust gas, and provides more dimensional information for the evaluation of the particulate matter removal effect. In addition, by fusing the infrared spectrum and particulate echo information together, the limitations of a single information source can be overcome, and the evaluation accuracy and stability of the particulate matter removal effect can be improved.

[0044] Further, perform a probability responsiveness adjustment based on a high-dimensional feature space on the exhaust gas particulate matter removal judgment classification feature matrix to obtain an optimized exhaust gas particulate matter removal judgment classification feature matrix, including: summing the feature values at each position of the exhaust gas particulate matter removal judgment classification feature matrix to obtain the sum of the feature values of the exhaust gas particulate matter removal judgment classification feature matrix; adding one to the sum of the feature values of the exhaust gas particulate matter removal judgment classification feature matrix and then dividing by the feature distribution scale of the exhaust gas particulate matter removal judgment classification feature matrix to obtain a first probability value, where the feature distribution scale of the exhaust gas particulate matter removal judgment classification feature matrix is the width of the exhaust gas particulate matter removal judgment classification feature matrix multiplied by the height of the exhaust gas particulate matter removal judgment classification feature matrix; calculating an exponential function with the natural constant as the base using the feature value at a predetermined position of the exhaust gas particulate matter removal judgment classification feature matrix as the exponent of the natural constant to obtain a first exponential function value; calculating the absolute value of the difference between the first exponential function value minus one and the feature value at the predetermined position of the exhaust gas particulate matter removal judgment classification feature matrix to obtain a first absolute value; activating the first absolute value through a softmax function to obtain a second probability value; multiplying the first probability value and the second probability value by the feature value at the predetermined position of the exhaust gas particulate matter removal judgment classification feature matrix respectively and then calculating the weighted sum to obtain the feature value at the corresponding position of the optimized exhaust gas particulate matter removal judgment classification feature matrix.

[0045] In particular, in the technical solution of the present application, considering that two different types of instruments (infrared spectrometer and particulate lidar) are used to collect data, the data types and dimensions collected by each instrument are different. The infrared spectrogram provides information on chemical composition, while the particulate echo signal provides information on physical characteristics. This diversity of data types leads to the high dimensionality of the feature space. The spectrometer values extracted from the infrared spectrogram and the feature maps extracted from the particulate echo signal waveform both contain rich information. During the process of converting this information into the exhaust particulate removal judgment classification feature matrix, a large number of features will be generated, further increasing the dimensionality of the feature space. Moreover, the data in the actual industrial process often contains noise and outliers. These noise and outliers may be amplified during the feature extraction and encoding process, resulting in a more dispersed distribution in the feature space. In order to enable the model to have good generalization ability, sufficient information needs to be captured in the feature space to distinguish different particulate removal effects. This usually means that the model needs to find a suitable decision boundary in the high-dimensional space, and the determination of this boundary may be affected by the divergent distribution. Based on this, in the technical solution of the present application, for the distribution divergence in the high-dimensional feature semantic space formed by the exhaust particulate removal judgment classification feature matrix, in the technical solution of the present application, a probability responsiveness adjustment based on the high-dimensional feature space is performed on the exhaust particulate removal judgment classification feature matrix.

[0046] Specifically, performing a probability responsiveness adjustment based on the high-dimensional feature space on the exhaust particulate removal judgment classification feature matrix to obtain an optimized exhaust particulate removal judgment classification feature matrix includes:

[0047] Performing a probability responsiveness adjustment based on the high-dimensional feature space on the exhaust particulate removal judgment classification feature matrix with the following optimization formula to obtain the optimized exhaust particulate removal judgment classification feature matrix, where the optimization formula is:

[0048]

[0049] where m i,j is the eigenvalue at the (i, j) position of the exhaust particulate removal judgment classification feature matrix, W and H are respectively the width and height of the exhaust particulate removal judgment classification feature matrix, α and β represent weighting coefficients, softmax represents the normalized exponential function, and m i,j ' is the eigenvalue at the (i, j) position of the optimized exhaust particulate removal judgment classification feature matrix.

[0050] Here, for the distribution divergence of the exhaust particulate removal judgment classification feature matrix in the high-dimensional feature semantic space, in the technical solution of this application, the exhaust particulate removal judgment classification feature matrix is adjusted based on the probability responsiveness of the high-dimensional feature space, and the label probability responsiveness of each feature in the exhaust particulate removal judgment classification feature matrix is adjusted to better capture and reflect the internal structure and class distribution of the data. Specifically, it uses statistical analysis methods to evaluate the response probabilities of each feature in the exhaust particulate removal judgment classification feature matrix for different labels, and then adjusts the feature weights according to these probabilities, so that the model can more sensitively reflect the correlation between the feature distribution and the label. This can effectively reduce the classification uncertainty caused by the divergence of the feature distribution, so that the feature distribution of the exhaust particulate removal judgment classification feature matrix is transferred to the range with a stable and structured boundary in the target domain, and the stability of the inductive iteration of the classification solution is improved.

[0051] Furthermore, by inputting the feature matrix into the classifier model, the trained model can be used to classify the feature matrix, so as to judge whether the removal degree of the particulate matter meets the standard. In the exhaust gas treatment scenario during the production process of refining catalysts, the exhaust particulate removal judgment classification feature matrix contains the comprehensive information of the global correlation features of the exhaust gas treatment infrared spectrum and the particulate echo features. Through the processing of the classifier, the effect of particulate matter removal during the exhaust gas treatment process can be judged based on these feature matrices, so as to realize the automatic evaluation and monitoring of the particulate matter removal effect, improving work efficiency and accuracy. This method combines the advantages of deep learning and machine learning, and provides an efficient and reliable solution for the quality assessment of particulate matter removal during the exhaust gas treatment process.

[0052] In summary, in the embodiment of this application, first, the infrared spectrum diagram of the exhaust gas treatment during the production process of refining catalysts collected by the infrared spectrometer and the waveform diagram of the particulate echo signals at multiple predetermined time points collected by the particulate lidar are obtained, then deep learning technology is used to extract features and perform correlation analysis on the two, and finally a classification result is obtained through the classifier to judge whether the removal degree of the particulate matter meets the standard, so as to realize an automatic monitoring and evaluation process, and then timely discover problems and take measures for improvement.

[0053] As described above, the tail gas treatment system 100 for the production process of the oil refining catalyst according to the embodiments of the present application can be implemented in various terminal devices. In one example, the tail gas treatment system 100 for the production process of the oil refining catalyst can be integrated into the terminal device as a software module and / or a hardware module. For example, the tail gas treatment system 100 for the production process of the oil refining catalyst can be a software module in the operating system of the terminal device, or can be an application program developed for the terminal device; of course, the tail gas treatment system 100 for the production process of the oil refining catalyst can also be one of the numerous hardware modules of the terminal device.

[0054] Alternatively, in another example, the tail gas treatment system 100 for the production process of the oil refining catalyst and the terminal device can also be separate devices, and the tail gas treatment system 100 for the production process of the oil refining catalyst can be connected to the terminal device through a wired and / or wireless network, and transmit and interact information according to a predefined data format.

[0055] Exemplary method

[0056] Figure 4 FIG. is a flowchart of the method for treating the tail gas in the production process of the oil refining catalyst according to the embodiments of the present application. As Figure 4 shown, the method for treating the tail gas in the production process of the oil refining catalyst according to the embodiments of the present application includes: S110, obtaining an infrared spectrogram of the tail gas treatment in the production process of the oil refining catalyst collected by an infrared spectrometer and a waveform diagram of particulate echo signals at a plurality of predetermined time points collected by a particulate lidar; S120, extracting a global correlation feature vector of the tail gas treatment infrared spectrum and a particulate echo feature vector from the infrared spectrogram of the tail gas treatment in the production process of the oil refining catalyst collected by the infrared spectrometer and the waveform diagram of the particulate echo signals at the plurality of predetermined time points collected by the particulate lidar; S130, based on the global correlation feature vector of the tail gas treatment infrared spectrum and the particulate echo feature vector, determining whether the removal degree of the particulate matter reaches the standard.

[0057] Here, those skilled in the art can understand that the specific operations of each step in the above method for treating the tail gas in the production process of the oil refining catalyst have been described in detail above with reference to Figure 1 to FIG. 3 of the tail gas treatment system for the production process of the oil refining catalyst, and therefore, the repeated description thereof will be omitted.

[0058] Exemplary electronic device

[0059] Next, reference is made to Figure 5 to describe the electronic device according to the embodiments of the present application.

[0060] Figure 5A schematic structural diagram of an electronic device provided by an embodiment of the present application. The electronic device 10 includes: a processor 11, a memory 12, a bus 13, and a communication interface 14. The processor 11, the communication interface 14, and the memory 12 are connected through the bus 13. The processor 11 is configured to execute an executable module stored in the memory 12, such as a computer program.

[0061] The memory 12 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. Through at least one communication interface 14 (which can be wired or wireless), a communication connection is established between the system network element and at least one other network element. The Internet, wide area network, local area network, metropolitan area network, etc. can be used.

[0062] The bus 13 may be an ISA bus, a PCI bus, an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 only a bidirectional arrow is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0063] Among them, the memory 12 is used to store a program. After receiving an execution instruction, the processor 11 executes the program. The method executed by the device defined by the flow process disclosed in any one of the foregoing embodiments of the present invention can be applied to the processor 11 or implemented by the processor 11.

[0064] The processor 11 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor 11 or the instructions in the form of software. The above-mentioned processor 11 may be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it may also be a digital signal processor (DSP for short), an application-specific integrated circuit (ASIC for short), a field-programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present invention can be directly embodied as being executed and completed by the hardware decoding processor, or completed by the combination of the hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory 12, and the processor 11 reads the information in the memory 12 and combines its hardware to complete the steps of the above method.

[0065] Corresponding to the above method, an embodiment of the present application further provides a computer-readable storage medium. The computer-readable storage medium stores machine-executable instructions. When the computer-executable instructions are called and run by the processor, the computer-executable instructions cause the processor to run the steps of the above method.

[0066] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For another example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other may be through some communication interfaces. The indirect coupling or communication connection of the devices or units may be in an electrical, mechanical, or other form.

[0067] The unit described as a separation component may or may not be physically separated. The component shown as a unit may or may not be a physical unit, that is, it may be located in one place or may be distributed across multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0068] In addition, each functional unit in the embodiments provided in this application can be integrated into a processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit.

[0069] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, an electronic device, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0070] It should be noted that similar reference numerals and letters indicate similar items in the drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. In addition, the terms "first", "second", "third", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0071] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A tail gas treatment system for a refinery catalyst production process, characterized in that: include: The catalyst production process data acquisition module is used to obtain the infrared spectrum of the tail gas treatment in the oil refining catalyst production process collected by the infrared spectrometer and the waveform of the particle echo signal at multiple predetermined time points collected by the particle laser radar instrument; A catalyst production process data extraction module is used to extract the exhaust gas treatment infrared spectrum global correlation feature vector and the particulate matter echo feature vector from the infrared spectrum of the exhaust gas treatment in the oil refining catalyst production process collected by the infrared spectrometer and the waveform of the particulate matter echo signal at multiple predetermined time points collected by the particulate matter laser radar instrument; An exhaust gas particle removal degree judgment module is used to judge whether the particle removal degree reaches the standard based on the exhaust gas treatment infrared spectrum global correlation feature vector and the particle echo feature vector; Wherein, the catalyst production process data extraction module includes: A spectrum counting feature extraction unit is used to extract spectrum counting features from the infrared spectrum of the tail gas treatment in the oil refining catalyst production process collected by the infrared spectrometer to obtain a statistical feature vector of the infrared spectrum of the tail gas treatment; A spectral image feature extraction unit is used to extract spectral image features from the infrared spectrum of the tail gas treatment in the oil refining catalyst production process collected by the infrared spectrometer to obtain a tail gas treatment infrared spectrum feature vector; A spectral feature correlation unit, used for correlating the exhaust gas treatment infrared spectrum statistical feature vector with the exhaust gas treatment infrared spectrum feature vector to obtain the exhaust gas treatment infrared spectrum global correlation feature vector; an echo signal feature encoding unit, used for performing feature encoding on the waveform diagrams of the particle echo signals at a plurality of predetermined time points collected by the particle laser radar instrument to obtain the particle echo feature vector; Wherein, the spectrum counting feature extraction unit comprises: Extracting infrared spectrum count values ​​of tail gas treatment at multiple predetermined wavelengths from the infrared spectrum of tail gas treatment in the oil refining catalyst production process collected by the infrared spectrometer; The exhaust gas treatment infrared spectrum count values ​​at the plurality of predetermined wavelengths are passed through an exhaust gas treatment infrared spectrum count sequence associated feature encoder comprising a one-dimensional convolution layer and a fully connected layer to obtain the exhaust gas treatment infrared spectrum statistical feature vector; Wherein, the echo signal characteristic encoding unit includes: The waveform diagram of the particle echo signal at a plurality of predetermined time points collected by the particle laser radar instrument is obtained by using a particle echo feature encoder of a dilated convolutional neural network to obtain a particle echo feature diagram; Pooling the particle echo feature map to obtain the particle echo feature vector; Wherein, the exhaust gas particle removal degree judgment module includes: An exhaust gas treatment feature fusion unit, used for fusing the exhaust gas treatment infrared spectrum global correlation feature vector and the particle echo feature vector to obtain an exhaust gas particle removal judgment classification feature matrix; An exhaust treatment feature optimization unit, used for performing a probability responsiveness adjustment on the exhaust particulate matter removal judgment classification feature matrix based on a high-dimensional feature space to obtain an optimized exhaust particulate matter removal judgment classification feature matrix; A removal degree standard judgment classification unit is used to pass the optimized exhaust particulate matter removal judgment classification feature matrix through a classifier to obtain a classification result to judge whether the particulate matter removal degree meets the standard; Wherein, the exhaust gas treatment characteristic optimization unit comprises: Summing the eigenvalues ​​of each position of the exhaust particulate matter removal judgment classification feature matrix to obtain the sum of the eigenvalues ​​of the exhaust particulate matter removal judgment classification feature matrix; The sum of the eigenvalues ​​of the exhaust particulate matter removal judgment classification feature matrix plus one is divided by the characteristic distribution scale of the exhaust particulate matter removal judgment classification feature matrix to obtain a first probability value, wherein the characteristic distribution scale of the exhaust particulate matter removal judgment classification feature matrix is ​​the width of the exhaust particulate matter removal judgment classification feature matrix multiplied by the height of the exhaust particulate matter removal judgment classification feature matrix; Using the characteristic value of the predetermined position of the exhaust particulate matter removal judgment classification characteristic matrix as the exponent of the natural constant, an exponential function with the natural constant as the base is calculated to obtain a first exponential function value; Subtracting one from the first exponential function value and then calculating the absolute value of the eigenvalue at the predetermined position of the exhaust particulate matter removal judgment classification feature matrix to obtain a first absolute value; Activating the first absolute value through a softmax function to obtain a second probability value; The first probability value and the second probability value are respectively multiplied by the eigenvalues ​​of predetermined positions of the exhaust particulate matter removal judgment classification feature matrix, and then the weighted sum is calculated to obtain the eigenvalues ​​of corresponding positions of the optimized exhaust particulate matter removal judgment classification feature matrix.

2. The tail gas treatment system for the production process of oil refining catalyst according to claim 1, characterized in that: The spectral image feature extraction unit comprises: The infrared spectrum of the tail gas treatment in the oil refining catalyst production process collected by the infrared spectrometer is passed through the tail gas treatment infrared spectrum convolution neural network as a noise reducer to obtain a tail gas treatment infrared spectrum after noise reduction; The exhaust gas treatment infrared spectrum image after noise reduction is passed through an exhaust gas treatment infrared spectrum convolutional neural network as a feature encoder to obtain the exhaust gas treatment infrared spectrum feature vector.

3. The tail gas treatment system for the production process of oil refining catalyst according to claim 2, characterized in that: The waveform diagram of the particle echo signal at a plurality of predetermined time points collected by the particle laser radar instrument is obtained by using a particle echo feature encoder of a dilated convolutional neural network to obtain a particle echo feature diagram, including: Each layer of the first convolutional neural network model of the particle echo feature encoder using the dilated convolutional neural network performs dilated convolution processing based on the first convolution kernel, pooling processing along the channel dimension, and nonlinear activation processing on the input data in the forward pass of the layer, so as to output a first feature map from the last layer of the first convolutional neural network model; Each layer of the second convolutional neural network model of the particle echo feature encoder using the atrous convolutional neural network performs atrous convolution processing based on a second convolution kernel, pooling processing along the channel dimension, and nonlinear activation processing on the input data in the forward pass of the layer, so as to output a second feature map from the last layer of the second convolutional neural network model; The first characteristic map and the second characteristic map are fused to obtain the particle echo characteristic map.

4. A method for treating tail gas in a refinery catalyst production process, using the tail gas treatment system in a refinery catalyst production process according to claim 1, characterized in that: include: Obtaining an infrared spectrum of tail gas treatment during the oil refining catalyst production process collected by an infrared spectrometer and a waveform of particulate matter echo signals at multiple predetermined time points collected by a particulate matter laser radar instrument; Extracting the exhaust gas treatment infrared spectrum global correlation feature vector and the particulate matter echo feature vector from the infrared spectrum of the exhaust gas treatment in the oil refining catalyst production process collected by the infrared spectrometer and the waveform of the particulate matter echo signal at multiple predetermined time points collected by the particulate matter laser radar instrument; Based on the exhaust gas treatment infrared spectrum global correlation feature vector and the particle echo feature vector, it is judged whether the particle removal degree reaches the standard.

Images