A method for preparing a biomass-based activated carbon catalytic material

Abstract

The present application relates to the technical field of carbon-based flue gas purification catalysts, and specifically discloses a preparation method of a biomass-based activated carbon catalyst carrier material, which collects product carbonization images at two different times through a camera, processes the product carbonization images using image processing technology based on deep learning, extracts product carbonization features in the two images respectively, intelligently estimates a product carbonization degree coefficient based on the differences in product carbonization features between the two images, and then further calculates the time needed for complete carbonization of the product by combining the time interval between the two time points and the product carbonization degree coefficient. In this way, by real-time monitoring and accurate control of the carbonization process, the problems of uneven carbonization or overcarbonization in traditional processes can be effectively avoided, thereby improving the performance and stability of the catalyst.

Description

Technical Field

[0001] This application relates to the field of carbon-based flue gas purification catalyst technology, and more specifically, to a method for preparing a biomass-based activated carbon catalyst support material. Background Technology

[0002] Nitrogen oxides (NOx) are a major component of air pollution. Their emissions not only lead to environmental problems such as acid rain and photochemical smog, but also have serious impacts on human health. With increasing environmental awareness and increasingly stringent emission standards, the performance optimization and innovative preparation technologies of denitrification catalysts, as key materials for reducing NOx emissions, have become a research hotspot. Among various denitrification catalysts, biomass-based activated carbon, due to its wide availability, renewability, and environmental friendliness, has become an important direction for developing highly efficient denitrification catalysts.

[0003] Chinese patent CN110433788A discloses a denitrification catalyst using biomass-based activated carbon as a carrier and its preparation method. The method uses biomass-based activated carbon as the substrate, and involves crushing, washing, and drying the biomass raw material, followed by high-temperature carbonization and chemical activation treatment. This results in a catalyst with excellent denitrification performance, achieving the goals of resource conservation and environmental protection. The carbonization process is a crucial step in the preparation of biomass-based activated carbon, directly affecting the performance of the final catalyst. However, traditional carbonization processes typically rely on empirically set time parameters, lacking real-time monitoring and precise control of the carbonization degree. This can easily lead to uneven or over-carbonization, thus affecting the catalyst's performance.

[0004] Therefore, an optimized method for preparing biomass-based activated carbon catalytic materials is desired. Summary of the Invention

[0005] To address the aforementioned technical problems, this application is proposed. Embodiments of this application provide an optimized method for preparing biomass-based activated carbon catalysts. This method involves acquiring product carbonization images at two different times using a camera, and processing these images using deep learning-based image processing technology to extract the product carbonization features from each image. Based on the differences in these features between the two images, the carbonization degree coefficient is intelligently estimated. Furthermore, by combining the time interval between the two time points and the carbonization degree coefficient, the remaining time required for complete carbonization is calculated. In this way, by real-time monitoring and precise control of the carbonization process, the problems of uneven or excessive carbonization in traditional processes can be effectively avoided, thereby improving the performance and stability of the catalyst.

[0006] Accordingly, according to one aspect of this application, a method for preparing a biomass-based activated carbon catalytic material is provided, comprising:

[0007] S1: Take biomass raw materials, crush and wash them in sequence, and then dry them at 105℃ for 2 hours;

[0008] S2: Take out the dried product, use N2 as a protective gas, and heat it to 500-800℃ at a heating rate of 10℃ / min for carbonization treatment. After 1-2 hours, the biomass-based activated carbon precursor is obtained.

[0009] S3: The biomass-based activated carbon precursor is immersed in a saturated KOH solution for 2 hours;

[0010] S4: Take out the impregnated product and heat it to 400-800℃ for 0.5-2h to obtain biomass-based activated carbon;

[0011] S5: Wash the activated biomass-based activated carbon with 1-2M hydrochloric acid solution, rinse with deionized water until neutral, and then dry to constant weight to obtain biomass-based activated carbon powder particles.

[0012] S6: The biomass-based activated carbon powder particles and binder are mixed evenly at a weight ratio of 100:(0.5-1.5) and then kneaded to form a biomass-based activated carbon catalytic material carrier.

[0013] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, step S2 includes: acquiring product carbonization images at a first time point and a second time point respectively using a camera to obtain a first product carbonization image and a second product carbonization image; determining the product carbonization time based on the first time point and the second time point; extracting product carbonization features from the first product carbonization image and the second product carbonization image respectively to obtain a first product carbonization coding feature map and a second product carbonization coding feature map; performing pixel-level spatial constraint enhancement processing on the first product carbonization coding feature map and the second product carbonization coding feature map respectively to obtain a first product carbonization spatial enhancement coding feature map and a second product carbonization spatial enhancement coding feature map; generating a product carbonization degree coefficient based on the feature differences between the first product carbonization spatial enhancement coding feature map and the second product carbonization spatial enhancement coding feature map; and predicting the time required for complete product carbonization based on the product carbonization degree coefficient and the product carbonization time.

[0014] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, the carbonization features of the first product carbonization image and the second product carbonization image are extracted to obtain the first product carbonization coding feature map and the second product carbonization coding feature map, respectively. This includes: passing the first product carbonization image and the second product carbonization image through a product carbonization feature extractor based on a deep separable convolutional neural network model to obtain the first product carbonization coding feature map and the second product carbonization coding feature map.

[0015] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, the first product carbonization coding feature map and the second product carbonization coding feature map are subjected to pixel-level spatial constraint enhancement processing to obtain the first product carbonization spatial enhancement coding feature map and the second product carbonization spatial enhancement coding feature map, respectively. This includes: performing feature divergence along the channel dimension on the first product carbonization coding feature map to obtain a set of first product carbonization pixel-level feature vectors; calculating the semantic association score between any two first product carbonization pixel-level feature vectors in the set of first product carbonization pixel-level feature vectors to obtain a first product carbonization pixel-level semantic association score topology matrix; based on the pixel spatial location information of the first product carbonization coding feature map, performing spatial attention attenuation modulation on the first product carbonization pixel-level semantic association score topology matrix to obtain a first product carbonization semantic association spatial soft constraint score topology feature matrix; and based on the first product carbonization semantic association spatial soft constraint score topology feature matrix, performing pixel-level spatial enhancement processing on the first product carbonization coding feature map to obtain the first product carbonization spatial enhancement coding feature map.

[0016] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, calculating the semantic association score between any two first product carbonized pixel particle size feature vectors in the set of first product carbonized pixel particle size feature vectors to obtain the semantic association score topology matrix of the first product carbonized pixel particle size includes: performing nonlinear mapping on each first product carbonized pixel particle size feature vector in the set of first product carbonized pixel particle size feature vectors to obtain a set of mapped first product carbonized pixel particle size feature vectors; calculating the Poincaré distance between any two mapped first product carbonized pixel particle size feature vectors in the set of mapped first product carbonized pixel particle size feature vectors as the semantic association score to obtain the semantic association score topology matrix of the first product carbonized pixel particle size.

[0017] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, based on the pixel spatial location information of the first product carbonization encoding feature map, spatial attention attenuation modulation is performed on the first product carbonization pixel particle size semantic association score topology matrix to obtain the first product carbonization semantic association spatial soft constraint score topology feature matrix. This includes: performing spatial attention attenuation modulation on the first product carbonization pixel particle size semantic association score topology matrix based on the spatial distance between any two first product carbonization pixel particle size feature vectors in the set of first product carbonization pixel particle size feature vectors to obtain the first product carbonization semantic association spatial soft constraint score topology matrix; and performing dilated convolution encoding on the first product carbonization semantic association spatial soft constraint score topology matrix to obtain the first product carbonization semantic association spatial soft constraint score topology feature matrix.

[0018] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, the spatial attention attenuation modulation of the semantic association score topology matrix of the first product carbonization pixel is performed based on the spatial distance between any two first product carbonization pixel particle size feature vectors in the set of first product carbonization pixel particle size feature vectors to obtain the spatial soft constraint score topology matrix of the first product carbonization semantic association. This includes: calculating the natural exponential function value of the Euclidean distance between any two first product carbonization pixel particle size feature vectors in the set of first product carbonization pixel particle size feature vectors to obtain the spatial distance topology matrix of the first product carbonization pixel; multiplying the spatial distance topology matrix of the first product carbonization pixel by a preset attenuation modulation constant to obtain the spatial attention attenuation modulation matrix; and calculating the first product carbonization pixel particle size semantic association score topology matrix and the spatial attention attenuation modulation matrix by position points to obtain the spatial soft constraint score topology matrix of the first product carbonization semantic association.

[0019] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, based on the semantic association spatial soft constraint score topological feature matrix of the first product carbonization, pixel-level spatial enhancement processing is performed on the first product carbonization coding feature map to obtain the first product carbonization spatial enhancement coding feature map. This includes: inputting the set of first product carbonization pixel-level feature vectors and the first product carbonization semantic association spatial soft constraint score topological feature matrix into a graph convolutional coding module to obtain a set of context-related enhanced first product carbonization pixel-level feature vectors; and reshaping the feature shape of the set of context-related enhanced first product carbonization pixel-level feature vectors to obtain the first product carbonization spatial enhancement coding feature map.

[0020] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, a product carbonization degree coefficient is generated based on the feature differences between the first product carbonization space enhancement coding feature map and the second product carbonization space enhancement coding feature map. This includes: inputting the first product carbonization space enhancement coding feature map and the second product carbonization space enhancement coding feature map into a product carbonization degree measurement network to obtain a product carbonization degree change coding feature map; and inputting the product carbonization degree change coding feature map into a decoder-based product carbonization degree estimation module to obtain the product carbonization degree coefficient.

[0021] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, predicting the time required for complete carbonization of the product based on the product carbonization degree coefficient and the product carbonization time includes: dividing the product carbonization degree coefficient by the product carbonization time to obtain the time required for complete carbonization of the product.

[0022] According to another aspect of this application, a biomass-based activated carbon catalytic material is provided, which is prepared by the aforementioned method for preparing biomass-based activated carbon catalytic materials.

[0023] Compared with existing technologies, the biomass-based activated carbon catalyst and its preparation method provided in this application acquire product carbonization images at two different times using a camera. Deep learning-based image processing technology is then used to process these images, extracting the product carbonization features from each image. Based on the differences in these features between the two images, the carbonization degree coefficient is intelligently estimated. Furthermore, by combining the time interval between the two time points and the carbonization degree coefficient, the time required for complete carbonization is calculated. This real-time monitoring and precise control of the carbonization process effectively avoids the problems of uneven or excessive carbonization in traditional processes, thereby improving the catalyst's performance and stability. Attached Figure Description

[0024] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0025] Figure 1 This is a flowchart of the carbonization process in the preparation method of biomass-based activated carbon catalytic material according to an embodiment of this application.

[0026] Figure 2 This is a schematic diagram of the data flow during the carbonization process in the preparation method of biomass-based activated carbon catalytic material according to an embodiment of this application.

[0027] Figure 3 This is a flowchart of sub-step S24 of the carbonization process in the preparation method of biomass-based activated carbon catalytic material according to an embodiment of this application.

[0028] Figure 4 This is a flowchart of sub-step S243 of the carbonization process in the preparation method of biomass-based activated carbon catalytic material according to an embodiment of this application.

[0029] Figure 5 This is a flowchart of sub-step S25 of the carbonization process in the preparation method of biomass-based activated carbon catalytic material according to an embodiment of this application. Detailed Implementation

[0030] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not specifically singular and may include plural forms. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0031] While this application makes various references to certain modules of the systems according to embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The modules described are merely illustrative, and different aspects of the systems and methods may use different modules.

[0032] Flowcharts are used in this application to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0033] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0034] This application proposes a method for preparing biomass-based activated carbon catalytic material, comprising: S1: drying biomass raw materials at 105℃ for 2 hours; S2: taking out the dried product, using N2 as a protective gas, and heating it to 500-800℃ at a heating rate of 10℃ / min for carbonization treatment, obtaining a biomass-based activated carbon precursor after 1-4 hours; S3: immersing the biomass-based activated carbon precursor in a saturated KOH solution for 2 hours; S4: [The text abruptly ends here, so the translation stops as well.] The product is removed and activated by heating to 400-800℃ for 0.5-2 hours to obtain biomass-based activated carbon; S5: The activated biomass-based activated carbon is washed with 1-2M hydrochloric acid solution, then rinsed with deionized water until neutral, and then dried to constant weight to obtain biomass-based activated carbon powder particles; S6: The biomass-based activated carbon powder particles and binder are mixed evenly at a weight ratio of 100:(0.5-1.5) and kneaded into a shape to obtain a biomass-based activated carbon catalyst carrier.

[0035] In this preparation process, carbonization is a crucial step in the preparation of biomass-based activated carbon, directly affecting the performance of the final catalyst. However, traditional carbonization processes typically rely on empirically set time parameters, lacking real-time monitoring and precise control of the carbonization degree, which can easily lead to uneven or over-carbonization, thus affecting catalyst performance. To address these technical problems, this application proposes an optimized method for preparing biomass-based activated carbon catalysts. This method involves acquiring product carbonization images at two different times using a camera, and then processing these images using deep learning-based image processing technology to extract the product carbonization features from both images. Based on the differences in these features between the two images, the carbonization degree coefficient is intelligently estimated. Furthermore, by combining the time interval between the two time points and the carbonization degree coefficient, the time required for complete carbonization is calculated. In this way, by monitoring and precisely controlling the carbonization process in real time, the problems of uneven or over-carbonization in traditional processes can be effectively avoided, thereby improving the performance and stability of the catalyst.

[0036] Figure 1 This is a flowchart of the carbonization process in the preparation method of biomass-based activated carbon catalytic material according to an embodiment of this application. Figure 2 This is a schematic diagram of the data flow during the carbonization process in the preparation method of biomass-based activated carbon catalyst material according to embodiments of this application. Figure 1 and Figure 2As shown, the preparation method of biomass-based activated carbon catalyst material according to an embodiment of this application includes the following steps: S21, acquiring product carbonization images at a first time point and a second time point respectively using a camera to obtain a first product carbonization image and a second product carbonization image; S22, determining the product carbonization time based on the first time point and the second time point; S23, extracting product carbonization features from the first product carbonization image and the second product carbonization image respectively to obtain a first product carbonization coding feature map and a second product carbonization coding feature map; S24, performing pixel-level spatial constraint enhancement processing on the first product carbonization coding feature map and the second product carbonization coding feature map respectively to obtain a first product carbonization spatial enhancement coding feature map and a second product carbonization spatial enhancement coding feature map; S25, generating a product carbonization degree coefficient based on the feature differences between the first product carbonization spatial enhancement coding feature map and the second product carbonization spatial enhancement coding feature map; S26, predicting the time required for complete product carbonization based on the product carbonization degree coefficient and the product carbonization time.

[0037] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, step S21 involves acquiring carbonization images of the product at a first time point and a second time point using a camera to obtain a first product carbonization image and a second product carbonization image. It should be understood that during the carbonization process, the product exhibits a carbonization rate at different stages. For example, the carbonization rate is faster in the initial stage of carbonization, while it significantly decreases as carbonization approaches completion. Therefore, based on consideration of the variation in carbonization rate, the technical solution of this application effectively estimates the degree of carbonization of the product by comparing and analyzing the product carbonization images at two different time points.

[0038] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, step S22 determines the carbonization time of the product based on the first time point and the second time point. It should be understood that since the carbonization rate refers to the change in the degree of carbonization of the product per unit time, in order to more accurately understand the carbonization rate of the product, this application further determines the carbonization time of the product based on the time interval between the first time point and the second time point, thereby providing an important basis for predicting the time required for complete carbonization of the product. In a specific example of this application, the second time point is the current time point, and the first time point is any time point before the current time point.

[0039] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, step S23 involves extracting the carbonization features of the first and second product carbonization images to obtain a first product carbonization coding feature map and a second product carbonization coding feature map. In a specific example of this application, step S23 includes: passing the first and second product carbonization images through a product carbonization feature extractor based on a deep separable convolutional neural network model to obtain the first and second product carbonization coding feature maps. That is, in order to extract the carbonization features from the first and second product carbonization images respectively for subsequent carbonization degree estimation, this application uses a deep separable convolutional neural network model as the product carbonization feature extractor. By performing depthwise convolution and pointwise convolution operations on the first and second product carbonization images respectively, carbonization information such as color, texture, and shape of the product in the images can be efficiently extracted, thereby obtaining the first and second product carbonization coding feature maps. It is understandable that deep separable convolutional neural networks (DSNNs) effectively reduce the computational complexity of the model by decomposing the standard convolution operation into two steps: depthwise convolution and pointwise convolution, while preserving both the depth and breadth of feature extraction. Specifically, in the depthwise convolution stage, the network focuses on learning the independent spatial features of each feature channel of the image, while in the pointwise convolution stage, the network integrates cross-channel features by applying 1x1 convolution kernels to each feature channel, thereby achieving a balance between the depth and breadth of features. In this way, the deep separable convolutional neural network model can effectively capture subtle changes in the carbonization process, providing accurate feature support for the subsequent generation of carbonization degree coefficients.

[0040] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, step S24 involves performing pixel-level spatial constraint enhancement processing on the first product carbonization coding feature map and the second product carbonization coding feature map to obtain a first product carbonization spatial enhancement coding feature map and a second product carbonization spatial enhancement coding feature map. It should be understood that, considering the possible redundancy or irrelevant features in the first and second product carbonization coding feature maps, this application proposes a feature enhancement method to further enhance the distinguishability of features and improve the accuracy of carbonization degree estimation. This method performs pixel-level spatial constraint enhancement processing on the first and second product carbonization coding feature maps to highlight features closely related to the carbonization degree characterization, while suppressing the influence of irrelevant or noisy features. This avoids misjudgments caused by abrupt changes in local features and improves the model's sensitivity to key information.

[0041] Figure 3This is a flowchart of sub-step S24 of the carbonization process in the preparation method of biomass-based activated carbon catalytic material according to an embodiment of this application. Figure 3 As shown, step S24 includes the following steps: S241, performing feature divergence on the first product carbonization coding feature map along the channel dimension to obtain a set of first product carbonization pixel granularity feature vectors; S242, calculating the semantic association score between any two first product carbonization pixel granularity feature vectors in the set of first product carbonization pixel granularity feature vectors to obtain a first product carbonization pixel granularity semantic association score topology matrix; S243, based on the pixel spatial location information of the first product carbonization coding feature map, performing spatial attention attenuation modulation on the first product carbonization pixel granularity semantic association score topology matrix to obtain a first product carbonization semantic association spatial soft constraint score topology feature matrix; S244, based on the first product carbonization semantic association spatial soft constraint score topology feature matrix, performing pixel granularity spatial enhancement processing on the first product carbonization coding feature map to obtain a first product carbonization spatial enhancement coding feature map.

[0042] In a specific example of this application, step S242 includes: performing a nonlinear mapping on each of the first product carbonized pixel granularity feature vectors in the set of first product carbonized pixel granularity feature vectors to obtain a set of mapped first product carbonized pixel granularity feature vectors; calculating the Poincaré distance between any two mapped first product carbonized pixel granularity feature vectors in the set of mapped first product carbonized pixel granularity feature vectors as the semantic association score to obtain the first product carbonized pixel granularity semantic association score topology matrix. That is, by calculating the semantic association score between any two product carbonized pixel granularity feature vectors in the set, the product carbonized pixel granularity semantic association score topology matrix is ​​constructed, thereby revealing the inherent contextual semantic association of the image. Specifically, in this process, considering that hyperbolic space can more effectively capture the nonlinear relationships between data compared to Euclidean space, this application first maps the feature vectors of each product carbonized pixel granularity from Euclidean space to Poincaré space, so that it adapts to the geometric characteristics of hyperbolic space while maintaining the original semantic information. The Poincaré distance between any two mapped product carbonized pixel granularity feature vectors is calculated as the semantic association score between them, so as to utilize the geometric characteristics of Poincaré space to enhance the nonlinear association between features, improve the model's ability to distinguish between different categories of features, and thus more accurately simulate the inherent semantic association structure of the image.

[0043] Figure 4 This is a flowchart of sub-step S243 of the carbonization process in the preparation method of biomass-based activated carbon catalytic material according to an embodiment of this application. Figure 4As shown, step S243 includes the following steps: S2431, performing spatial attention attenuation modulation on the semantic association scoring topology matrix of the first product carbonized pixel based on the spatial distance between any two first product carbonized pixel granularity feature vectors in the set of first product carbonized pixel granularity feature vectors to obtain the first product carbonized semantic association spatial soft constraint scoring topology matrix; S2432, performing dilated convolutional encoding on the first product carbonized semantic association spatial soft constraint scoring topology matrix to obtain the first product carbonized semantic association spatial soft constraint scoring topology feature matrix. In a specific example of this application, step S2431 includes: calculating the natural exponential function value of the Euclidean distance between any two first product carbonized pixel granularity feature vectors in the set of first product carbonized pixel granularity feature vectors to obtain a first product carbonized pixel granularity spatial distance topology matrix; multiplying the first product carbonized pixel granularity spatial distance topology matrix by a preset attenuation modulation constant to obtain a spatial attention attenuation modulation matrix; and calculating the first product carbonized pixel granularity semantic association score topology matrix and the spatial attention attenuation modulation matrix by position points to obtain the first product carbonized semantic association spatial soft constraint score topology matrix.

[0044] It is understandable that, considering that adjacent image regions often have stronger correlations than distant regions in practical applications, this application further applies spatial attention attenuation modulation to the aforementioned pixel-level semantic association scoring topology matrix based on the spatial distance between any two product carbonized pixel granularity feature vectors. This simulates the correlation attenuation characteristics between local image regions, making the model more focused on the feature associations of adjacent regions. Furthermore, dilated convolution technology is used to perform dilated convolutional encoding on the spatially soft-constrained scoring matrix to capture the semantic association topology features between the granular features of each product carbonized pixel, generating a product carbonization semantic association spatial soft-constrained scoring topology feature matrix.

[0045] In a specific example of this application, step S244 includes: inputting the set of pixel-level feature vectors of the first product carbonization and the soft-constraint score topological feature matrix of the first product carbonization semantic association space into a graph convolutional coding module to obtain a set of context-enhanced pixel-level feature vectors of the first product carbonization; and reshaping the feature shape of the set of context-enhanced pixel-level feature vectors of the first product carbonization to obtain a spatially enhanced coded feature map of the first product carbonization. That is, graph convolutional coding technology is used to process the set of pixel-level feature vectors of the product carbonization and the soft-constraint score topological feature matrix of the product carbonization semantic association space to utilize the characteristics of graph convolution to update the pixel-level feature representation of the image based on the semantic association information between pixels. By learning the inherent semantic association structure of the product carbonization image, the global semantic context-association feature expression of the pixel-level features of the product carbonization image is enhanced. Furthermore, by reshaping the feature shapes of the carbonized pixel granular features of each product after contextual association enhancement, the original image structure is restored to obtain the first product carbonized spatial enhancement coding feature map and the second product carbonized spatial enhancement coding feature map, thereby achieving refined enhancement of local detail features of the image. It is worth mentioning that graph convolutional coding is a deep learning model for processing graph-structured data. It extends the concept of convolutional neural networks to graph structures, enabling direct operations on graph data. The core idea of ​​graph convolutional coding is to use the structural information of the graph to update the representation of nodes, thereby capturing the complex relationships between nodes. In graph convolutional coding, the feature representation of each node includes not only its own features but also the features of its neighboring nodes. This aggregation of information is achieved through graph convolution operations, which allows the model to perform feature extraction and propagation on the graph.

[0046] Accordingly, the carbonization coding feature map of the first product is processed using the following feature enhancement formula to obtain the carbonization spatially enhanced coding feature map of the first product, wherein the feature enhancement formula is:

[0047]

[0048]

[0049] T = AtrousConv(S′)

[0050] V′=GCN(V,T)

[0051] X′=reshape[V′,(H,W,C)]

[0052] Where X is the carbonization encoding feature map of the first product, H, W, and C represent the height, width, and number of channels of the carbonization encoding feature map of the first product, respectively, reshape[·] represents feature shape reshaping, and V represents the set of carbonization pixel-level feature vectors of the first product. i and v j Let represent the i-th and j-th first product carbonized pixel particle size feature vectors in the set of first product carbonized pixel particle size feature vectors, respectively. The value of HW is the number of feature vectors in the set of first product carbonized pixel particle size feature vectors, and v i ∈R C The value of C is the length of the feature vector of the carbonized pixel size of the first product, W1 is the weight transformation matrix, W2 is the weight transformation vector, and h i and h j Let represent the i-th and j-th mapped first product carbonized pixel granularity feature vectors in the set of mapped first product carbonized pixel granularity feature vectors, respectively. arccosh(·) represents the inverse hyperbolic cosine function in hyperbolic space, and ||·|| represents the norm of the feature vector. P (h i ,h j S represents the Poincaré distance between the feature vector of the carbonized pixel of the first product after the i-th mapping and the feature vector of the carbonized pixel of the first product after the j-th mapping. ij This represents the semantic association score value at position (i,f) in the topological matrix of the semantic association score of the carbonized pixel granularity of the first product, where θ is a preset attenuation modulation constant, and d(v i ,v j S' represents the Euclidean distance between the ith carbonized pixel granularity feature vector of the first product and the jth carbonized pixel granularity feature vector of the first product, and S' is the soft-constraint score topology matrix of the semantic association space of the first product carbonization. ij ′ represents the feature value at position (i,j) in the soft constraint score topology matrix of the semantic association space of the first product carbonization, AtrousConv(·) represents dilated convolutional coding, T is the soft constraint score topology feature matrix of the semantic association space of the first product carbonization, GCN(·) represents graph convolution processing, V′ represents the set of context association enhanced pixel granularity feature vectors of the first product carbonization, and X′ represents the spatial enhancement coding feature map of the first product carbonization.

[0053] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, step S25 generates a product carbonization degree coefficient based on the feature differences between the first product carbonization space enhancement coding feature map and the second product carbonization space enhancement coding feature map. Figure 5 This is a flowchart of sub-step S25 of the carbonization process in the preparation method of biomass-based activated carbon catalytic material according to an embodiment of this application. Figure 5 As shown, step S25 includes the following steps: S251, inputting the first product carbonization spatial enhancement coding feature map and the second product carbonization spatial enhancement coding feature map into the product carbonization degree measurement network to obtain the product carbonization degree change coding feature map; S252, inputting the product carbonization degree change coding feature map into the decoder-based product carbonization degree estimation module to obtain the product carbonization degree coefficient.

[0054] Specifically, in step S251, the first product carbonization spatial enhancement coding feature map and the second product carbonization spatial enhancement coding feature map are input into a product carbonization degree measurement network to obtain a product carbonization degree change coding feature map. It should be understood that, in order to achieve accurate estimation of the product carbonization degree, this application further uses a product carbonization degree measurement network to measure the difference between the first product carbonization spatial enhancement coding feature map and the second product carbonization spatial enhancement coding feature map. Specifically, the product carbonization degree measurement network quantifies the difference in carbonization characteristics between the first product carbonization spatial enhancement coding feature map and the second product carbonization spatial enhancement coding feature map by performing positional difference calculations, thereby obtaining the product carbonization degree change coding feature map.

[0055] Specifically, in step S252, the encoded feature map of the product carbonization degree change is input into the product carbonization degree estimation module based on the decoder to obtain the product carbonization degree coefficient. That is, a decoder model is used to construct the product carbonization degree estimation module. By using the decoder to extract and learn features layer by layer from the encoded feature map of the product carbonization degree change, it is mapped to the numerical representation space of the product carbonization degree, generating a quantitative index that reflects the relative change of the product carbonization degree from the first time point to the second time point, namely the product carbonization degree coefficient, thereby realizing the real-time quantitative assessment of the product carbonization degree.

[0056] Preferably, in one example of this application, inputting the product carbonization degree change coding feature map into a decoder-based product carbonization degree estimation module to obtain a product carbonization degree coefficient includes: calculating the sum of the absolute values ​​of each feature value of the product carbonization degree change coding feature map to obtain a first product carbonization degree change coding spatial structure value, and calculating the square root of the sum of the squares of each feature value of the product carbonization degree change coding feature map to obtain a second product carbonization degree change coding spatial structure value; multiplying each feature value of the product carbonization degree change coding feature map by the first product carbonization degree change coding spatial structure value and the second product carbonization degree change coding spatial structure value respectively to obtain a first product carbonization degree change coding structure reference value and a second product carbonization degree change coding structure reference value corresponding to each feature value; multiplying each feature value of the product carbonization degree change coding feature map by the scale of the product carbonization degree change coding feature map and the square root of the scale respectively to obtain... The first product carbonization degree change coding scale transformation value and the second product carbonization degree change coding scale transformation value are corresponding to each feature value; the first product carbonization degree change coding structure reference value is divided by the difference between the first product carbonization degree change coding spatial structure value and the first product carbonization degree change coding scale transformation value to obtain the first product carbonization degree change coding transformation adjustment value; the second product carbonization degree change coding structure reference value is divided by the difference between the second product carbonization degree change coding spatial structure value and the second product carbonization degree change coding scale transformation value to obtain the second product carbonization degree change coding transformation adjustment value; the weighted sum of the first product carbonization degree change coding transformation adjustment value and the second product carbonization degree change coding transformation adjustment value is calculated to obtain each feature value of the optimized product carbonization degree change coding feature map; the optimized product carbonization degree change coding feature map is input into the decoder-based product carbonization degree estimation module to obtain the product carbonization degree coefficient.

[0057] Accordingly, the optimization process of the product carbonization degree change encoding feature map is expressed by the following optimization formula:

[0058]

[0059] f 1i =(α×f i ) / (α-n×f i )

[0060]

[0061] f i ∈F∈R W′×H′×C′

[0062] f 1i ∈F1∈RW′×H′×C′

[0063] f 2i ∈F2∈R W′×H′×C′

[0064] n=W′×H′×C′

[0065] Wherein, F represents the encoded feature map of the change in the degree of carbonization of the product, f i Let represent the i-th feature value of the product carbonization degree change coding feature map, W′, H′, and C′ represent the width, height, and number of channels of the product carbonization degree change coding feature map, respectively, n is the number of feature values ​​of the product carbonization degree change coding feature map, R represents the set of real numbers, α represents the spatial structure value of the first product carbonization degree change coding feature map, β represents the spatial structure value of the second product carbonization degree change coding feature map, and f 1i F1 represents the first product carbonization degree change coding transformation adjustment value corresponding to the i-th feature value in the product carbonization degree change coding feature map, and F1 is the first product carbonization degree change coding transformation adjustment map composed of all the first product carbonization degree change coding transformation adjustment values. 2i F2 represents the second product carbonization degree change coding transformation adjustment value corresponding to the i-th feature value in the product carbonization degree change coding feature map, F2 is the second product carbonization degree change coding transformation adjustment map composed of each of the second product carbonization degree change coding transformation adjustment values, and ω is the weight hyperparameter. ⊙ represents addition by position, ⊙ represents multiplication by position, and F′ represents the optimized product carbonization degree change encoding feature map.

[0066] In other words, this application considers that the first product carbonization encoding feature map and the second product carbonization encoding feature map represent the semantics of the first and second product carbonization images at the first and second time points, respectively. After performing feature enhancement based on spatial soft constraints, both are input into a product carbonization degree measurement network to obtain the product carbonization degree change encoding feature map. Although feature enhancement based on spatial soft constraints can improve the spatial distribution saliency of the first and second product carbonization encoding feature maps, it also weakens the spatial distribution alignment between the first and second product carbonization encoding feature maps, resulting in a complex spatial structure for the product carbonization degree change encoding feature map. Therefore, it is desirable to improve its regression convergence and generalization performance based on the complex spatial structure.

[0067] Therefore, in the preferred example, regarding the spatial structure information of the feature set of the product carbonization degree change coding feature map in high-dimensional space, scale-based bounding box transformations of each feature value of the product carbonization degree change coding feature map are performed using the norm-based spatial structured representation of the product carbonization degree change coding feature map as a reference window. Furthermore, bounding box attention weights based on spatial structure are adjusted for each feature value of the product carbonization degree change coding feature map. This ensures the invariance of spatial transformations (translation, scaling, and rotation) of the product carbonization degree change coding feature map under feature space interaction, thereby improving the convergence and generalization effect of the feature set of the product carbonization degree change coding feature map under complex spatial structure representations, and enhancing the accuracy of the product carbonization degree coefficients obtained by inputting it into the decoder-based product carbonization degree estimation module.

[0068] In the above-mentioned method for preparing biomass-based activated carbon catalytic materials, step S26 predicts the time required for complete carbonization of the product based on the product carbonization degree coefficient and the product carbonization time. In a specific example of this application, step S26 includes: dividing the product carbonization degree coefficient by the product carbonization time to obtain the time required for complete carbonization of the product. It should be understood that the product carbonization degree coefficient is a quantitative indicator that measures the change in the degree of carbonization of the product within the product carbonization time range. By calculating the ratio between the product carbonization degree coefficient and the product carbonization time, the carbonization rate and carbonization stage of the product can be effectively revealed, thus serving as a measure of the time still required for complete carbonization of the product. This can provide effective guidance and reference for carbonization control in industrial production processes.

[0069] In summary, the preparation method of biomass-based activated carbon catalyst material has been elucidated. This method involves acquiring product carbonization images at two different times using a camera, and then processing these images using deep learning-based image processing technology to extract the product carbonization features from each image. Based on the differences in these features between the two images, the carbonization degree coefficient is intelligently estimated. Furthermore, by combining the time interval between the two time points and the carbonization degree coefficient, the time required for complete carbonization is calculated. This real-time monitoring and precise control of the carbonization process effectively avoids the problems of uneven or over-carbonization in traditional processes, thereby improving the catalyst's performance and stability.

[0070] Furthermore, a biomass-based activated carbon catalytic material is also provided, specifically, the biomass-based activated carbon catalytic material is prepared by the above-mentioned preparation method of biomass-based activated carbon catalytic material.

[0071] The basic principles of the present invention have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in the present invention are merely examples and not limitations, and should not be considered as essential features of each embodiment of the present invention. Furthermore, the specific details of the above embodiments are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the present invention to the necessity of employing the specific details described above.

[0072] In the above embodiments, the descriptions of each embodiment have their own emphasis. Parts not detailed or described in a particular embodiment can be referred to in the relevant descriptions of other embodiments. In the several embodiments provided by this invention, it should be understood that the disclosed system can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for example, the unit division is only a logical functional division, and other division methods may exist in actual implementation. The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0073] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the present invention. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0074] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units stated in a system claim may also be implemented by a single unit through software or hardware.

[0075] Finally, it should be noted that the above description has been given for illustrative and descriptive purposes. Furthermore, the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention.

Claims

1. A method for preparing a biomass-based activated carbon catalytic material, comprising: S1: Dry the biomass raw material at 105℃ for 2 hours; S2: Take out the dried product, use N2 as a protective gas, and heat it to 500-800℃ at a heating rate of 10℃ / min for carbonization treatment. After 1-4 hours, the biomass-based activated carbon precursor is obtained; S3: Immerse the biomass-based activated carbon precursor in a saturated KOH solution for 2 hours. S4: Remove the impregnated product and activate it at 400-800℃ for 0.5-2h to obtain biomass-based activated carbon; S5: Wash the activated biomass-based activated carbon with 1-2M hydrochloric acid solution, rinse with deionized water until neutral, and then dry to constant weight to obtain biomass-based activated carbon powder particles; S6: Mix the biomass-based activated carbon powder particles with a binder at a weight ratio of 100:(0.5-1.5) and knead to form a biomass-based activated carbon catalytic material carrier, characterized in that step S2 further includes: The carbonization images of the product were acquired by a camera at the first time point and the second time point respectively to obtain the carbonization images of the first product and the carbonization images of the second product. Based on the first time point and the second time point, the carbonization time of the product is determined; The carbonization features of the first product carbonization image and the second product carbonization image are extracted respectively to obtain the first product carbonization coding feature map and the second product carbonization coding feature map. The first product carbonization coding feature map and the second product carbonization coding feature map are subjected to pixel-level spatial constraint enhancement processing to obtain the first product carbonization spatial enhancement coding feature map and the second product carbonization spatial enhancement coding feature map. Based on the feature differences between the first product carbonization space-enhanced coding feature map and the second product carbonization space-enhanced coding feature map, a product carbonization degree coefficient is generated; Based on the carbonization degree coefficient of the product and the carbonization time of the product, the time required for complete carbonization of the product is predicted.

2. The method for preparing biomass-based activated carbon catalytic material according to claim 1, characterized in that, Extract the carbonization features of the first and second product carbonization images respectively to obtain the first and second product carbonization coding feature maps, including: The carbonized images of the first and second products are processed by a carbonized product feature extractor based on a deep separable convolutional neural network model to obtain the carbonized product encoded feature map of the first and second products, respectively.

3. The method for preparing biomass-based activated carbon catalytic material according to claim 2, characterized in that, The first product carbonization coding feature map and the second product carbonization coding feature map are subjected to pixel-level spatial constraint enhancement processing to obtain the first product carbonization spatial enhancement coding feature map and the second product carbonization spatial enhancement coding feature map, including: The first product carbonization encoding feature map is subjected to feature divergence along the channel dimension to obtain a set of first product carbonization pixel granularity feature vectors; Calculate the semantic association score between any two first product carbonized pixel granularity feature vectors in the set of first product carbonized pixel granularity feature vectors to obtain the first product carbonized pixel granularity semantic association score topology matrix. Based on the pixel spatial location information of the first product carbonization coding feature map, spatial attention attenuation modulation is performed on the first product carbonization pixel granular semantic association scoring topology matrix to obtain the first product carbonization semantic association spatial soft constraint scoring topology feature matrix. Based on the soft constraint score topological feature matrix of the semantic association space of the first product carbonization, pixel-level spatial enhancement processing is performed on the carbonization coding feature map of the first product to obtain the spatially enhanced coding feature map of the first product carbonization.

4. The method for preparing biomass-based activated carbon catalytic material according to claim 3, characterized in that, Calculate the semantic association score between any two first product carbonized pixel granularity feature vectors in the set of first product carbonized pixel granularity feature vectors to obtain the first product carbonized pixel granularity semantic association score topology matrix, including: A nonlinear mapping is performed on each of the first product carbonized pixel particle size feature vectors in the set of first product carbonized pixel particle size feature vectors to obtain a set of first product carbonized pixel particle size feature vectors after mapping. The Poincaré distance between any two mapped first product carbonized pixel granularity feature vectors in the set of mapped first product carbonized pixel granularity feature vectors is calculated as the semantic association score to obtain the topological matrix of the semantic association score of the first product carbonized pixel granularity.

5. The method for preparing the biomass-based activated carbon catalytic material according to claim 4, characterized in that, Based on the pixel spatial location information of the first product carbonization encoded feature map, spatial attention attenuation modulation is applied to the first product carbonization pixel-level semantic association score topology matrix to obtain the first product carbonization semantic association spatial soft constraint score topology feature matrix, including: Based on the spatial distance between any two first product carbonized pixel granularity feature vectors in the set of first product carbonized pixel granularity feature vectors, the spatial attention attenuation modulation is performed on the first product carbonized pixel granularity semantic association score topology matrix to obtain the first product carbonized semantic association spatial soft constraint score topology matrix. The first product carbonization semantic association space soft constraint score topology matrix is ​​subjected to dilated convolutional encoding to obtain the first product carbonization semantic association space soft constraint score topology feature matrix.

6. The method for preparing biomass-based activated carbon catalytic material according to claim 5, characterized in that, Based on the spatial distance between any two first product carbonized pixel granularity feature vectors in the set of first product carbonized pixel granularity feature vectors, spatial attention attenuation modulation is applied to the first product carbonized pixel granularity semantic association score topology matrix to obtain the first product carbonized semantic association spatial soft constraint score topology matrix, including: Calculate the natural exponential function value of the Euclidean distance between any two first product carbonized pixel particle size feature vectors in the set of first product carbonized pixel particle size feature vectors to obtain the first product carbonized pixel particle size spatial distance topology matrix. Multiply the spatial distance topology matrix of the carbonized pixel granularity of the first product by a preset attenuation modulation constant to obtain the spatial attention attenuation modulation matrix, and calculate the semantic association score topology matrix of the carbonized pixel granularity of the first product and the spatial attention attenuation modulation matrix by the position point to obtain the spatial soft constraint score topology matrix of the semantic association of the first product carbonization.

7. The method for preparing biomass-based activated carbon catalytic material according to claim 6, characterized in that, Based on the soft-constraint score topological feature matrix of the semantic association space of the first product carbonization, pixel-level spatial enhancement processing is performed on the carbonization coding feature map of the first product to obtain the spatially enhanced coding feature map of the first product carbonization, including: The set of carbonized pixel granularity feature vectors of the first product and the topological feature matrix of the first product carbonized semantic association space soft constraint score are input into the graph convolutional encoding module to obtain the set of context association enhanced first product carbonized pixel granularity feature vectors. The feature shape of the set of carbonized pixel granular feature vectors of the first product with context association enhancement is reshaped to obtain the spatially enhanced coded feature map of the first product carbonization.

8. The method for preparing biomass-based activated carbon catalytic material according to claim 7, characterized in that, Based on the feature differences between the first product carbonization space-enhanced coding feature map and the second product carbonization space-enhanced coding feature map, a product carbonization degree coefficient is generated, including: The first product carbonization space enhancement coding feature map and the second product carbonization space enhancement coding feature map are input into the product carbonization degree measurement network to obtain the product carbonization degree change coding feature map. The encoded feature map of the product carbonization degree change is input into the product carbonization degree estimation module based on the decoder to obtain the product carbonization degree coefficient.

9. The method for preparing biomass-based activated carbon catalytic material according to claim 8, characterized in that, Based on the carbonization degree coefficient and the carbonization time of the product, the time required for complete carbonization of the product is predicted, including: Divide the carbonization degree coefficient of the product by the carbonization time of the product to obtain the time required for the product to be fully carbonized.

10. A biomass-based activated carbon catalytic material, characterized in that, The biomass-based activated carbon catalytic material is prepared by the preparation method of the biomass-based activated carbon catalytic material as described in any one of claims 1-9.

Images