Information management method and system for preparation of biochar-based fertilizer

Abstract

The present application relates to the field of information management technology, and discloses an information management method and system for the preparation of biochar-based fertilizer. The method includes: numerically processing the parameters of the preparation process of biochar-based fertilizer to obtain a parameter vector set; performing local optimization and global optimization processing on the parameter vector set to obtain a first preparation parameter combination; matrixing and standardizing the first preparation parameter combination to obtain a preparation process characteristic matrix; performing characteristic vector centrality processing on the preparation process characteristic matrix to obtain a target parameter execution mode; obtaining the carbonized biochar and the carbonization process parameter set according to the target parameter execution mode, and performing spectral analysis on the carbonized biochar to obtain a biochar quality data set; performing parameter optimization on the first preparation parameter combination according to the biochar quality data set and the carbonization process parameter set to obtain a second preparation parameter combination. The present application realizes the intelligent management of the biochar-based fertilizer preparation process.

Description

Technical Field

[0001] The present application relates to the field of information management technology, and in particular to an information management method and system for preparing biochar-based fertilizer. Background Art

[0002] As a new type of fertilizer, biochar-based fertilizer has significant advantages in improving soil structure, increasing crop yields and reducing environmental pollution. However, the preparation process of biochar-based fertilizer involves many parameters, such as raw material type, pyrolysis temperature, additives, etc. There are complex interactions between these parameters, which brings challenges to the optimization of the preparation process. Traditional preparation methods often rely on experience and trial and error, which makes it difficult to achieve precise control and efficient production.

[0003] With the rapid development of information technology, it has become possible to apply information management methods to the preparation process of biochar-based fertilizers. However, there are still some problems with the current information management methods, such as inaccurate numerical processing of parameters, inefficient optimization algorithms, and incomplete feature extraction. These problems limit the further optimization of the preparation process of biochar-based fertilizers and the improvement of product quality. Summary of the invention

[0004] The present application provides an information management method and system for the preparation of biochar-based fertilizer, which are used to realize intelligent management of the biochar-based fertilizer preparation process.

[0005] In a first aspect, the present application provides an information management method for preparing biochar-based fertilizer, the information management method for preparing biochar-based fertilizer comprising:

[0006] The parameters of the biochar-based fertilizer preparation process are numerically processed to obtain a parameter vector set;

[0007] Performing local optimization and global optimization processing on the parameter vector set to obtain a first preparation parameter combination;

[0008] Matrixing and standardizing the first preparation parameter combination to obtain a preparation process characteristic matrix;

[0009] Performing eigenvector centrality processing on the characteristic matrix of the preparation process to obtain a target parameter execution mode;

[0010] Acquire the carbonized biochar and a carbonization process parameter set according to the target parameter execution mode, and perform spectral analysis on the carbonized biochar to obtain a biochar quality data set;

[0011] The first preparation parameter combination is optimized according to the biochar quality data set and the carbonization process parameter set to obtain a second preparation parameter combination.

[0012] In a second aspect, the present application provides an information management system for the preparation of biochar-based fertilizers, the information management system for the preparation of biochar-based fertilizers comprising:

[0013] A numerical module is used to perform numerical processing on the parameters of the preparation process of biochar-based fertilizer to obtain a parameter vector set;

[0014] A processing module, used for performing local optimization and global optimization processing on the parameter vector set to obtain a first preparation parameter combination;

[0015] A matrixing module, used for performing matrixing and standardization processing on the first preparation parameter combination to obtain a preparation process characteristic matrix;

[0016] An execution module, used for performing characteristic vector centrality processing on the characteristic matrix of the preparation process to obtain a target parameter execution mode;

[0017] An analysis module, used for acquiring the carbonized biochar and a carbonization process parameter set according to the target parameter execution mode, and performing spectral analysis on the carbonized biochar to obtain a biochar quality data set;

[0018] A parameter optimization module is used to optimize the first preparation parameter combination according to the biochar quality data set and the carbonization process parameter set to obtain a second preparation parameter combination.

[0019] The third aspect of the present application provides an information management device for the preparation of biochar-based fertilizers, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the information management device for the preparation of biochar-based fertilizers executes the above-mentioned information management method for the preparation of biochar-based fertilizers.

[0020] A fourth aspect of the present application provides a computer-readable storage medium, wherein instructions are stored in the computer-readable storage medium, and when the computer-readable storage medium is run on a computer, the computer executes the above-mentioned information management method for preparing biochar-based fertilizer.

[0021] In the technical solution provided by this application, the precise numerical representation of the parameters of the preparation process is achieved by digitally encoding the types of raw materials, discretizing the pyrolysis temperature, and mathematically modeling the cooling rate. The method of combining local optimization and global optimization is adopted to improve the efficiency and accuracy of parameter optimization, and the optimal preparation parameter combination can be found more quickly. Through matrixization and standardization processing, combined with technologies such as singular value decomposition and non-negative matrix decomposition, the comprehensive extraction of preparation process characteristics is achieved. Through feature vector centrality processing, the relationship between features is considered, and the generated target parameter execution mode has strong adaptability. Combined with Fourier transform infrared spectroscopy and X-ray fluorescence spectroscopy analysis, the quality of biochar is evaluated in multiple dimensions, which improves the accuracy and reliability of quality data. Through neural network models and sensitivity analysis, dynamic optimization of preparation parameters is achieved, and the preparation plan can be continuously adjusted and improved according to actual production conditions, providing intelligent decision support for the preparation process of biochar-based fertilizers, which helps to improve production efficiency and product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0023] Figure 1 This is a schematic diagram of an embodiment of the information management method for preparing biochar-based fertilizer in the embodiment of the present application;

[0024] Figure 2 This is a schematic diagram of an embodiment of an information management system for preparing biochar-based fertilizer in an embodiment of the present application. DETAILED DESCRIPTION

[0025] The embodiments of the present application provide an information management method and system for the preparation of biochar-based fertilizer. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0026] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 , one embodiment of the information management method for preparing biochar-based fertilizer in the embodiment of the present application includes:

[0027] Step S101, numerically processing the parameters of the preparation process of biochar-based fertilizer to obtain a parameter vector set;

[0028] It is understandable that the execution subject of the present application may be an information management system for preparing biochar-based fertilizer, or a terminal or a server, which is not specifically limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0029] Specifically, the raw material types of biochar-based fertilizer are digitally encoded to generate a raw material type code set, which represents different raw material types. The raw material ratio is numerically processed according to the raw material type code set to form a raw material ratio vector to represent the ratio relationship of various raw materials. The pyrolysis temperature of the biochar-based fertilizer is discretized to generate a pyrolysis temperature discrete value set, and the discrete value can accurately represent the different temperature sections in the pyrolysis process. At the same time, the pyrolysis time of the biochar-based fertilizer is segmented and mapped to obtain a pyrolysis time vector, which divides the pyrolysis time into different time periods, and each time period corresponds to a different numerical representation. When processing the cooling rate of the biochar-based fertilizer, the cooling rate function is obtained by mathematical modeling, which can accurately describe the rate change during the cooling process. The types of additives of the biochar-based fertilizer are parameterized to generate an additive feature matrix, which includes the feature parameters of various additives. The additive dosage is normalized to generate an additive dosage normalized vector. At the same time, the production equipment parameters of the biochar-based fertilizer are feature-coded to generate an equipment parameter vector, and these codes can specifically represent the various parameters of the production equipment. When considering the environmental factors of biochar-based fertilizer, a quantitative evaluation is performed to obtain an environmental impact index vector, which reflects the impact of environmental factors on the preparation process. The raw material ratio vector, pyrolysis temperature discrete value set, pyrolysis time vector, cooling rate function, additive feature matrix, additive dosage normalization vector, equipment parameter vector and environmental impact index vector are multi-dimensionally spliced ​​to form a complete parameter vector set.

[0030] Step S102, performing local optimization and global optimization processing on the parameter vector set to obtain a first preparation parameter combination;

[0031] Specifically, principal component analysis is performed on the parameter vector set to reduce the dimension of high-dimensional data and obtain a simplified parameter space. Cluster analysis is performed on the parameter space after dimension reduction to determine the initial local optimization regions, which represent different feature subsets in the parameter space. Gradient descent is performed on the initial local optimization regions, and the local optimal solution set of each region is found by continuously adjusting the parameters. The local optimal solution set provides the optimal parameter combination in each region. Based on the local optimal solution set, an initial population of global optimization is constructed, which represents multiple potential optimal solutions in the entire parameter space. Fitness evaluation is performed on the initial population of global optimization to obtain a fitness matrix, which reflects the pros and cons of each parameter combination. According to the fitness matrix, selection, crossover and mutation operations are performed on the initial population of global optimization to obtain an evolved population. Local search is performed on the evolved population to refine parameter optimization and obtain a hybrid optimization solution set. A parameter response surface is constructed based on the hybrid optimization solution set to show the influence of different parameter combinations on the preparation process. The parameter response surface is optimized to find the Pareto optimal solution set, which represents the best trade-off between different optimization objectives. Decision analysis is performed based on the Pareto optimal solution set, and different objectives and constraints are comprehensively considered to obtain the first preparation parameter combination.

[0032] Step S103, matrixing and standardizing the first preparation parameter combination to obtain a preparation process characteristic matrix;

[0033] Specifically, the first preparation parameter combination is sorted to obtain an ordered parameter set, and these parameters are arranged according to their importance or relevance. The original parameter matrix is ​​constructed according to the ordered parameter set, and the matrix contains all the initial preparation parameters, and they are arranged into a matrix form according to certain rules. The original parameter matrix is ​​processed by multiple interpolation method to fill the missing data that may exist in the original matrix to obtain a complete parameter matrix. The descriptive statistics of each preparation parameter are calculated according to the complete parameter matrix to obtain a parameter statistical feature set, including the mean, variance, standard deviation and other statistics of the parameters. The parameter statistical feature set is transformed to obtain normalized parameter data. Through normalization, data of different dimensions are converted to the same dimension to make the data more uniform. The normalized parameter data is standardized to obtain a standardized parameter matrix. The standardized parameter matrix is ​​subjected to singular value decomposition to reduce the dimension of the high-dimensional data to obtain a reduced dimension parameter matrix, which retains the main characteristics and change trends of the original data. The reduced dimension parameter matrix is ​​calculated by information gain ratio to obtain a parameter weight vector, which represents the importance weight of each parameter to the preparation process. The parameter weight vector and the dimension reduction parameter matrix are matrix multiplied to obtain the weighted parameter matrix. The weighted parameter matrix is ​​decomposed into the product of two non-negative matrices by non-negative matrix decomposition to obtain the preparation process characteristic matrix, which comprehensively reflects the key parameters and characteristics of the biochar-based fertilizer preparation process.

[0034] Step S104, performing eigenvector centrality processing on the preparation process characteristic matrix to obtain a target parameter execution mode;

[0035] Specifically, a feature vector set is generated according to the feature matrix of the preparation process, and the feature vector set contains specific vector representations of various parameters in the preparation process. The cosine similarity of each vector is calculated according to the feature vector set, and a similarity matrix is ​​obtained by calculation to reflect the similarity between each vector. The centrality score is calculated for the similarity matrix to obtain an initial centrality score, which reflects the importance of each vector in the feature space. According to the initial centrality score, a feature relationship network is constructed, in which each feature vector is used as a node, and the connection weight between nodes is determined by the similarity, forming a network structure that comprehensively describes the feature relationship. Feature clustering is performed on the feature relationship network, and feature clustering results are obtained through clustering analysis. These results divide the feature vectors into several classes, each of which contains vectors with similar features. Based on the feature clustering results, the inter-class and intra-class centralities are calculated to obtain multi-scale centrality indicators, which can evaluate the importance of feature vectors at different scales. The multi-scale centrality indicators are comprehensively evaluated to calculate the comprehensive centrality score. According to the comprehensive centrality score, the feature vector set is sorted to obtain key feature sequences, which represent the most critical feature vectors in the preparation process. The feature weight analysis is performed on the key feature sequence. The weight of each key feature is determined by analysis to obtain the feature importance weight, which reflects the importance of each key feature in the entire preparation process. According to the feature importance weight, the target parameter execution mode is generated, which is the best guidance scheme for the parameter configuration and operation process of the entire preparation process.

[0036] Step S105, obtaining the carbonized biochar and the carbonization process parameter set according to the target parameter execution mode, and performing spectral analysis on the carbonized biochar to obtain a biochar quality data set;

[0037] Specifically, a carbonization process parameter set is generated according to the target parameter execution mode, including key indicators such as temperature, time, and pressure. The raw material is pyrolyzed and carbonized according to the carbonization process parameter set, and the carbonized biochar is obtained by controlling various parameters in the carbonization process. In this process, the temperature, pressure, and gas composition of the carbonization process are monitored in real time, and the parameter changes at each time point are recorded to generate a complete carbonization process parameter set. The carbonized biochar is scanned by Fourier transform infrared spectroscopy, and the molecular structure of the biochar is analyzed by infrared spectral data. The infrared spectral data is baseline corrected and normalized to obtain standardized spectral data to eliminate the influence of instrument noise and sample inhomogeneity. The standardized spectral data is feature extracted to identify representative spectral feature vectors, which can accurately describe the molecular structure characteristics of biochar. At the same time, the carbonized biochar is analyzed by X-ray fluorescence spectroscopy to obtain the elemental composition data of the biochar, and the content and distribution of each element are recorded. According to the elemental composition data, the proportion and content index of each element are calculated, and the elemental feature vector is generated to reflect the chemical composition characteristics of the biochar. The spectral feature vector and the element feature vector were fused to form a comprehensive feature matrix. The comprehensive feature matrix was subjected to partial least squares regression analysis, and a regression model between the feature matrix and the biochar quality index was established to obtain a biochar quality data set, which reflects the quality information of the carbonized biochar in terms of molecular structure, chemical composition, etc.

[0038] Step S106: Optimize the first preparation parameter combination according to the biochar quality data set and the carbonization process parameter set to obtain a second preparation parameter combination.

[0039] Specifically, the biochar quality data set and the carbonization process parameter set were analyzed to generate a parameter correlation matrix, revealing the degree of correlation between different preparation parameters and biochar quality indicators, helping to identify which parameters have a significant impact on the quality of the final product. An initial prediction model was constructed based on the parameter correlation matrix, and a mapping relationship between parameters and quality indicators was preliminarily established. In order to improve the accuracy and reliability of the model, the initial prediction model was cross-validated and trained. Through repeated training and verification processes, the model parameters were gradually optimized to obtain a neural network model with better performance. The optimized neural network model was used to perform sensitivity analysis on the first preparation parameter combination, and the parameter sensitivity ranking was obtained by analyzing the degree of influence of each parameter on the model output. The parameter sensitivity ranking was analyzed hierarchically, and the sensitivity weight vector of each parameter was calculated to reflect the importance of each parameter in the entire preparation process. Based on the sensitivity weight vector, an optimization objective function was constructed. The objective function combines the importance of each parameter and its impact on quality, providing a clear goal and direction for the subsequent optimization process. Particle swarm optimization was performed on the optimization objective function to find the optimal solution by simulating the movement of particles in the search space, and obtain a candidate parameter set. Each candidate parameter set represents a possible combination of optimization parameters. In order to evaluate the reliability and stability of these candidate parameter sets, the corresponding parameter probability distribution was calculated according to the candidate parameter sets, and the distribution characteristics of the parameter values ​​were understood through statistical analysis. The confidence interval analysis of the parameter probability distribution was performed to determine the credible range of each parameter under a certain confidence level. According to the parameter credible range, the first preparation parameter combination was adjusted and optimized, and the sensitivity and credibility of each parameter were comprehensively considered. The value of each parameter was gradually adjusted to obtain the second preparation parameter combination. This parameter combination optimizes the efficiency and stability of the preparation process while ensuring the quality of biochar.

[0040] In the embodiments of the present application, by digitally encoding the raw material types, discretizing the pyrolysis temperature, and mathematically modeling the cooling rate, the accurate numerical representation of the preparation process parameters is achieved. The method of combining local optimization and global optimization is adopted to improve the efficiency and accuracy of parameter optimization, and the optimal preparation parameter combination can be found more quickly. Through matrixization and standardization, combined with technologies such as singular value decomposition and non-negative matrix decomposition, the comprehensive extraction of preparation process characteristics is achieved. Through feature vector centrality processing, the relationship between features is considered, and the generated target parameter execution mode has strong adaptability. Combined with Fourier transform infrared spectroscopy and X-ray fluorescence spectroscopy analysis, the quality of biochar is evaluated in multiple dimensions, which improves the accuracy and reliability of quality data. Through neural network models and sensitivity analysis, dynamic optimization of preparation parameters is achieved, and the preparation plan can be continuously adjusted and improved according to actual production conditions, providing intelligent decision support for the preparation process of biochar-based fertilizers, which helps to improve production efficiency and product quality.

[0041] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0042] (1) Digitally encode the raw material types of biochar-based fertilizer to obtain a raw material type code set, and numerically process the raw material ratio according to the raw material type code set to obtain a raw material ratio vector;

[0043] (2) The pyrolysis temperature of the biochar-based fertilizer is discretized to obtain a discrete value set of the pyrolysis temperature, and the pyrolysis time of the biochar-based fertilizer is mapped by piecewise function to obtain a pyrolysis time vector;

[0044] (3) The cooling rate of biochar-based fertilizer was mathematically modeled to obtain the cooling rate function, and the types of additives in biochar-based fertilizer were parameterized to obtain the additive characteristic matrix;

[0045] (4) Normalizing the amount of additives in the biochar-based fertilizer to obtain a normalized vector of the amount of additives, and feature encoding the production equipment parameters of the biochar-based fertilizer to obtain an equipment parameter vector;

[0046] (5) Quantitatively evaluate the environmental factors of biochar-based fertilizers and obtain the environmental impact index vector;

[0047] (6) Based on the raw material ratio vector, the pyrolysis temperature discrete value set, the pyrolysis time vector, the cooling rate function, the additive characteristic matrix, the additive dosage normalization vector, the equipment parameter vector and the environmental impact index vector, multi-dimensional vector splicing is performed to obtain a parameter vector set.

[0048] Specifically, different types of biochar-based fertilizer raw materials are digitally encoded, and each raw material is assigned a unique digital identifier to form a raw material type code set. For example, suppose there are three raw materials, which can be coded as 1, 2, and 3 respectively. According to the raw material type code set, the ratio of each raw material is numerically processed. Assuming that in a certain biochar-based fertilizer formula, rice husk accounts for 50%, biomass sawdust accounts for 30%, and livestock and poultry manure accounts for 20%, a raw material ratio vector R=[0.5, 0.3, 0.2] can be formed. Among them, each element in the vector represents the proportion of the corresponding raw material in the formula, ensuring the numerical representation of the ratio. The pyrolysis temperature of the biochar-based fertilizer is discretized. The continuous temperature range is divided into several discrete temperature intervals. For example, the pyrolysis temperature from 200°C to 800°C is divided into 6 discrete intervals: 200-300°C, 301-400°C, 401-500°C, 501-600°C, 601-700°C, 701-800°C. Each interval can be assigned a corresponding discrete value to form a pyrolysis temperature discrete value set T = [250, 350, 450, 550, 650, 750], where each value represents the center temperature of the corresponding interval. For the pyrolysis time, it is processed by piecewise function mapping. Assuming that the pyrolysis time is from 0 to 120 minutes, it is divided into 4 sections: 0-30 minutes, 31-60 minutes, 61-90 minutes, and 91-120 minutes. Each section corresponds to a function mapping value, forming a pyrolysis time vector t = [15, 45, 75, 105], where each value represents the middle value of the corresponding time period. The cooling rate of biochar-based fertilizer is mathematically modeled. The cooling rate can be expressed by the temperature change rate during the cooling process. The cooling rate model is assumed to be:

[0049] ;

[0050] where v(t) is the cooling rate, K and is the unknown parameter, and t is the time. By fitting the experimental data, K and The value of is used to form the cooling rate function. The types of additives in biochar-based fertilizers are parameterized, and each additive is assigned a set of parameters to form an additive feature matrix. Assume that there are three additives A, B, and C, and their parameters are , , , then the additive characteristic matrix is:

[0051] ;

[0052] The dosage of additives is normalized to obtain the normalized vector of additive dosage. For example, assuming that the dosage of additives A, B, and C is 10g, 20g, and 30g respectively, and the total amount is 60g, then after normalization, it is . The production equipment parameters of biochar-based fertilizer are feature-encoded, and each equipment parameter is assigned a specific encoding value. For example, equipment parameters include temperature control accuracy, stirring speed, and pressure range, which can be encoded as E=[accuracy, speed, pressure] to form an equipment parameter vector. The environmental factors of biochar-based fertilizer are quantitatively evaluated, such as ambient temperature, humidity, air quality, etc. Each factor is quantified as a numerical value to form an environmental impact index vector F=[temperature index, humidity index, air quality index]. All vectors are multi-dimensionally concatenated to obtain a complete parameter vector set. Assuming that the vectors are the raw material ratio vector R, the pyrolysis temperature discrete value set T, the pyrolysis time vector t, the cooling rate function v(t), the additive feature matrix M, the additive dosage normalization vector U, the equipment parameter vector E, and the environmental impact index vector F, the parameter vector set can be expressed as P=[R,T,t,v(t),M,U,E,F]. In this way, through multi-dimensional vector concatenation, a comprehensive parameter vector set is obtained, which can comprehensively represent all the key parameters involved in the preparation process of biochar-based fertilizer.

[0053] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0054] (1) Perform principal component analysis on the parameter vector set to obtain the parameter space after dimensionality reduction, and perform cluster analysis on the parameter space after dimensionality reduction to obtain the initial local optimization region;

[0055] (2) Perform gradient descent processing on the initial local optimization area to obtain the local optimal solution set, and construct the global optimization initial population based on the local optimal solution set;

[0056] (3) Perform fitness evaluation on the global optimization initial population to obtain a fitness matrix, and perform selection, crossover and mutation operations on the global optimization initial population based on the fitness matrix to obtain an evolved population;

[0057] (4) Conduct local search on the evolved population to obtain a hybrid optimization solution set, and construct a parameter response surface based on the hybrid optimization solution set;

[0058] (5) The parameter response surface is optimized to obtain the Pareto optimal solution set, and a decision analysis is performed based on the Pareto optimal solution set to obtain the first preparation parameter combination.

[0059] Specifically, the original parameter vector set P=[R,T,t,v(t),M,U,E,F] is standardized to ensure that all parameters have the same scale. After standardization, principal component analysis is applied to extract the main features in the data and reduce the high-dimensional data to a lower-dimensional parameter space. Through principal component analysis, a set of new variables (principal components) are obtained. These principal components are linear combinations of the original variables and retain the most variance information in the data. Assuming that the data is reduced to a two-dimensional space, the parameter space after dimensionality reduction can be expressed as , where PC1 and PC2 are the first two principal components. Perform cluster analysis in the parameter space after dimensionality reduction, and select, for example, the K-means clustering algorithm. Through cluster analysis, the data is divided into several initial local optimization regions, each of which contains a parameter combination with similar characteristics. Assuming that the parameter space is divided into three clusters, three initial local optimization regions are obtained , , For each initial local optimization area, the gradient descent algorithm is applied for optimization. The gradient descent algorithm finds the local optimal solution of each area by iteratively adjusting the parameters and minimizing the loss function. Assuming the loss function is used ,in Represents the parameter vector, by calculating the gradient , gradually update the parameters ,in is the learning rate. After several iterations, the local optimal solution set is found in each region. Based on the local optimal solution set, the global optimization initial population is constructed. Global optimization is usually processed using a genetic algorithm. The fitness of the global optimization initial population is evaluated. The fitness function is used to measure the quality of each individual and obtain a fitness matrix, in which each element represents the fitness value of the corresponding individual. According to the fitness matrix, the initial population is subjected to selection, crossover and mutation operations to generate a new generation of populations by simulating the biological evolution process. The selection operation retains individuals with higher fitness, the crossover operation generates new individuals by exchanging some genes between individuals, and the mutation operation randomly changes some genes to increase diversity. After completing the selection, crossover and mutation, the evolved population is obtained. Local search is performed on the evolved population, and by optimizing the parameters of each individual, a mixed optimization solution set is found, which includes the results of global optimization and local optimization, which helps to improve the optimization effect. Based on the mixed optimization solution set, a parameter response surface is constructed. The parameter response surface is a mathematical model that describes the effect of parameter changes on the objective function, usually expressed in the form of polynomials or other functions. For example, a quadratic polynomial is used to fit the response surface:

[0060] ;

[0061] where x and y are parameters, are the unknown coefficients. By fitting these coefficients, we get the parameter response surface. We optimize the parameter response surface and apply the Pareto optimization method to get the Pareto optimal solution set. Pareto optimization focuses on multi-objective optimization problems, and the goal is to find a solution set that is not inferior to other solutions in all objectives. Assume there are two objective functions and , the solution in the Pareto optimal solution set satisfies and The above cannot be dominated by other solutions. According to the Pareto optimal solution set, decision analysis is performed, the weights of each objective function and actual needs are comprehensively considered, the optimal parameter combination is selected, and the first preparation parameter combination is obtained.

[0062] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0063] (1) Sorting the first preparation parameter combination to obtain an ordered parameter set, and constructing an original parameter matrix based on the ordered parameter set;

[0064] (2) The original parameter matrix is ​​processed by multiple interpolation method to obtain a complete parameter matrix, and the descriptive statistics of each preparation parameter are calculated based on the complete parameter matrix to obtain a parameter statistical feature set;

[0065] (3) Transform the parameter statistical feature set to obtain normalized parameter data, and standardize the normalized parameter data to obtain a standardized parameter matrix;

[0066] (4) Perform singular value decomposition on the standardized parameter matrix to obtain a reduced dimension parameter matrix, and calculate the information gain ratio of the reduced dimension parameter matrix to obtain a parameter weight vector;

[0067] (5) Perform matrix multiplication on the parameter weight vector and the dimension reduction parameter matrix to obtain a weighted parameter matrix, and perform non-negative matrix decomposition on the weighted parameter matrix to obtain the preparation process characteristic matrix.

[0068] Specifically, the importance of parameters is evaluated and ranked through sensitivity analysis or expert scoring. The original parameter matrix X is constructed based on the ordered parameter set, where each row represents an experiment or production process and each column represents a parameter. For example, suppose there are three main parameters: temperature, time, and pressure, and their importance rankings are 1, 2, and 3, respectively. Assuming there are data from five experiments, the original parameter matrix may be:

[0069] ;

[0070] The original parameter matrix is ​​processed by multiple imputation to solve the problem of missing values ​​that may exist in the data. Multiple imputation is a statistical technique that improves the reliability of the analysis results by performing multiple interpolations on missing values ​​and generating multiple complete data sets. Assume that some data points are missing in the parameter matrix. Use multiple imputation to fill in these missing values ​​and obtain a complete parameter matrix. . For example, suppose some data points in the original matrix are missing:

[0071] ;

[0072] After multiple interpolation processing, the complete parameter matrix is ​​obtained:

[0073] ;

[0074] According to the complete parameter matrix, the descriptive statistics of each preparation parameter, such as mean, variance, standard deviation, etc., are calculated to obtain the parameter statistical feature set S. These statistics can reflect the distribution characteristics of each parameter. The parameter statistical feature set is transformed to convert it into parameter data that conforms to the normal distribution. For example, the Box-Cox transformation is used to obtain normalized parameter data. The normalized parameter data is standardized to eliminate the dimensional differences between different parameters and obtain a standardized parameter matrix. The standardized parameter matrix is ​​subjected to singular value decomposition to reduce the dimensionality of the high-dimensional data and obtain a reduced-dimensional parameter matrix D. The formula for singular value decomposition is:

[0075] ;

[0076] Among them, U and V are orthogonal matrices, It is a diagonal matrix whose diagonal elements are singular values. After dimensionality reduction, the first k singular values ​​and corresponding singular vectors are retained to obtain the reduced dimension parameter matrix D. The information gain ratio is calculated for the reduced dimension parameter matrix to obtain the parameter weight vector W. The information gain ratio can measure the contribution of each parameter to the target variable, and the formula is:

[0077] ;

[0078] Among them, information gain is a parameter Information gain, entropy of the target variable is a parameter The entropy of . Perform matrix multiplication on the parameter weight vector W and the dimension reduction parameter matrix D to obtain the weighted parameter matrix A. The formula for matrix multiplication is:

[0079] ;

[0080] The weighted parameter matrix A is decomposed into the product of two non-negative matrices to obtain the preparation process feature matrix H. The formula for non-negative matrix decomposition is:

[0081] ;

[0082] Among them, W and H are both non-negative matrices. Through this process, the characteristics of each parameter in the preparation of biochar-based fertilizer are extracted, optimized and analyzed.

[0083] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0084] (1) Generate a feature vector set based on the preparation process feature matrix, and calculate the cosine similarity of each vector based on the feature vector set to obtain a similarity matrix;

[0085] (2) Calculate the centrality score of the similarity matrix to obtain the initial centrality score, and construct the feature relationship network based on the initial centrality score;

[0086] (3) Perform feature clustering on the feature relationship network to obtain feature clustering results, and calculate the inter-class and intra-class centrality based on the feature clustering results to obtain a multi-scale centrality index;

[0087] (4) Comprehensively evaluate the multi-scale centrality indicators to obtain a comprehensive centrality score, and sort the feature vector set according to the comprehensive centrality score to obtain the key feature sequence;

[0088] (5) Perform feature weight analysis on the key feature sequence to obtain the feature importance weight, and generate the target parameter execution mode based on the feature importance weight.

[0089] Specifically, a feature vector set V is generated according to the preparation process feature matrix H. Each vector of the feature vector set is a column or a row of the feature matrix, that is, ,in is a feature vector. Calculate the cosine similarity of each pair of feature vectors. Cosine similarity measures the similarity between two vectors, and its formula is:

[0090] ;

[0091] in represents the dot product of vectors, and Represents the modulus of the vector. By calculating the cosine similarity between all feature vectors, we get the similarity matrix S, where express and After obtaining the similarity matrix, the centrality score is calculated. The centrality score is used to evaluate the importance of each feature in the entire network. Common centrality indicators include degree centrality, closeness centrality, and betweenness centrality. Assuming that degree centrality is used, the formula is:

[0092] ;

[0093] in, Represents the feature vector Calculate the centrality score of each eigenvector and get the initial centrality score set . According to the initial centrality score, a feature relationship network G is constructed. In this network, each node represents a feature vector, and the edge weights between nodes are determined by the elements of the similarity matrix. Feature clustering is performed on the feature relationship network. Common clustering algorithms include K-means clustering, hierarchical clustering, and spectral clustering. Assuming that the spectral clustering algorithm is used, its basic steps include calculating the Laplace matrix, performing eigenvalue decomposition on the Laplace matrix, and performing cluster analysis on the eigenvectors. Through cluster analysis, the feature clustering results are obtained, and the feature vectors are divided into several clusters. According to the feature clustering results, the inter-class and intra-class centralities are calculated. The intra-class centrality is used to evaluate the importance of the feature vector within the cluster to which it belongs, while the inter-class centrality is used to evaluate the association between different clusters. The calculation formula of the intra-class centrality is similar to the initial centrality score, while the inter-class centrality can be achieved by calculating the similarity between the cluster center and the centers of other clusters. The multi-scale centrality indicators are comprehensively evaluated to obtain a comprehensive centrality score. The comprehensive centrality score combines the intra-class and inter-class centralities to provide a more comprehensive feature importance assessment. The calculation formula for the comprehensive centrality score is:

[0094] ;

[0095] in and is the weight coefficient, and Represents the eigenvectors The intra-class centrality and inter-class centrality of the feature vector set are calculated. The feature vector set is sorted according to the comprehensive centrality score to obtain the key feature sequence. The key feature sequence is arranged from high to low according to the comprehensive centrality score, reflecting the importance of each feature in the preparation process. The key feature sequence is subjected to feature weight analysis to obtain the feature importance weight. The feature weight can be determined by regression analysis, information gain and other methods. Assuming a linear regression model is used, the feature weight can be expressed as a regression coefficient:

[0096] ;

[0097] Where y is the target variable, X is the feature matrix, and w is the feature weight vector. is the error term. Through regression analysis, the feature importance weight w is obtained. The target parameter execution mode is generated according to the feature importance weight. The target parameter execution mode combines the importance of each feature and provides guidance for practical operations. For example, assuming that the importance weights of temperature, time, and pressure are 0.5, 0.3, and 0.2 respectively, the temperature parameter can be adjusted first to ensure that it is within the optimal range, and then the time and pressure parameters can be optimized.

[0098] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0099] (1) Generate a carbonization process parameter set according to the target parameter execution mode, and perform pyrolysis and carbonization treatment of the raw material according to the carbonization process parameter set to obtain carbonized biochar;

[0100] (2) Real-time monitoring of the temperature, pressure and gas composition during the carbonization process to obtain a carbonization process parameter set;

[0101] (3) Performing Fourier transform infrared spectroscopy scanning on the carbonized biochar to obtain infrared spectral data, and performing baseline correction and normalization on the infrared spectral data to obtain standardized spectral data;

[0102] (4) Extract features from the standardized spectral data to obtain spectral feature vectors, and perform X-ray fluorescence spectroscopy analysis on the carbonized biochar to obtain elemental composition data;

[0103] (5) Calculate the element ratio and content index according to the element composition data to obtain the element feature vector, and perform feature fusion on the spectral feature vector and the element feature vector to obtain a comprehensive feature matrix;

[0104] (6) Partial least squares regression analysis was performed on the comprehensive feature matrix to obtain the biochar quality data set.

[0105] Specifically, the carbonization process parameter set is generated according to the target parameter execution mode. The target parameter execution mode contains the optimal range of key parameters in the biochar preparation process, such as temperature, pressure, time, etc. Assume that carbonization needs to be carried out in the temperature range of 300°C to 700°C, the pressure is controlled between 1 and 5 bar, and the time is controlled between 30 and 120 minutes. Based on these target parameters, the carbonization process parameter set is generated. The raw material is pyrolyzed and carbonized according to the carbonization process parameter set. In the carbonization furnace, the raw material is placed under the set temperature, pressure and time conditions for pyrolysis to obtain carbonized biochar. In this process, the temperature, pressure and gas composition of the carbonization process are monitored in real time by sensors, and the parameter values ​​at each moment are recorded to obtain the carbonization process parameter set. ,in Indicates temperature, Indicates pressure, Indicates time, Represents the gas composition. The carbonized biochar is scanned by Fourier transform infrared spectroscopy (FTIR), and the infrared spectrum data of the biochar is obtained by FTIR spectrometer. Fourier transform infrared spectroscopy can provide information on the molecular structure of biochar. The acquired infrared spectrum data is baseline corrected to eliminate the influence of baseline drift, and normalized to ensure that the spectrum data of different samples are on the same scale. The spectrum data after baseline correction and normalization can be expressed as ,in Indicated in wavelength The spectral intensity at the point. Feature extraction is performed on the standardized spectral data. The main features in the spectral data are extracted using methods such as principal component analysis to obtain the spectral feature vector ,in Represents the i-th eigenvector. Principal component analysis reduces the dimension of high-dimensional data by maximizing the variance and retains the most important characteristic information. X-ray fluorescence spectroscopy (XRF) analysis is performed on the carbonized biochar, and the elemental composition data of the biochar is obtained by XRF spectrometer. XRF spectrum can provide quantitative information of each element in biochar. The proportion and content index of each element are calculated based on the obtained elemental composition data to obtain the elemental characteristic vector ,in Represents the feature vector of the i-th element. The spectral feature vector and the element feature vector are fused, and the comprehensive feature matrix is ​​obtained through concatenation or other feature fusion methods. . The comprehensive feature matrix contains the spectral characteristics and elemental composition characteristics of biochar. Partial least squares regression analysis is performed on the comprehensive feature matrix. Partial least squares regression is a multivariate statistical analysis method that finds the best linear regression model by decomposing the covariance matrix of the predictor variable and the response variable. Assuming that the comprehensive feature matrix is ​​F and the quality data of biochar is Y, the partial least squares regression model can be expressed as:

[0106] ;

[0107] Where B is the regression coefficient matrix and E is the error term. By training the partial least squares regression model, the biochar quality data set Y is obtained, in which each data point contains the quality indicators of biochar, such as specific surface area, porosity, carbon content, etc.

[0108] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0109] (1) Correlation analysis was performed on the biochar quality data set and the carbonization process parameter set to obtain the parameter correlation matrix;

[0110] (2) Construct an initial prediction model based on the parameter correlation matrix, and perform cross-validation training on the initial prediction model to obtain an optimized neural network model;

[0111] (3) Performing sensitivity analysis on the first preparation parameter combination according to the optimized neural network model to obtain parameter sensitivity ranking;

[0112] (4) Perform hierarchical analysis on the parameter sensitivity ranking to obtain the sensitivity weight vector, and construct the optimization objective function based on the sensitivity weight vector;

[0113] (5) Perform particle swarm optimization on the optimization objective function to obtain a candidate parameter set, and calculate the corresponding parameter probability distribution based on the candidate parameter set;

[0114] (6) Perform confidence interval analysis on the parameter probability distribution to obtain the parameter credible range, and adjust and optimize the first preparation parameter combination based on the parameter credible range to obtain the second preparation parameter combination.

[0115] Specifically, the correlation analysis is performed on the biochar quality data set and the carbonization process parameter set. Assume that there is a biochar quality data set, which contains multiple quality indicators, such as specific surface area, porosity, carbon content, etc. At the same time, there is a carbonization process parameter set, which contains parameters such as temperature, pressure, time, gas composition, etc. in the carbonization process. By calculating the correlation coefficient between each parameter and the quality indicator, the parameter correlation matrix R is obtained, whose elements are Represents the correlation between the i-th quality indicator and the j-th process parameter. Construct the initial prediction model according to the parameter correlation matrix. Suppose a multilayer perceptron neural network model is used to measure the biochar quality indicators. The input layer of the neural network model contains the carbonization process parameters, and the output layer contains the biochar quality indicators. In order to evaluate the performance of the model, cross-validation training is performed. By dividing the data set into a training set and a validation set, the model is alternately trained and validated to obtain the optimized neural network model. The purpose of cross-validation is to reduce the overfitting of the model and improve the generalization ability of the model. After obtaining the optimized neural network model, a sensitivity analysis is performed on the first preparation parameter combination to evaluate the degree of influence of each process parameter on the model output. Suppose there is a neural network model , where W represents the weight parameter of the model and Q represents the carbonization process parameter. By calculating the partial derivative of each parameter with respect to the output, the sensitivity ranking of the parameters is obtained. A hierarchical analysis is performed on the parameter sensitivity ranking to obtain the sensitivity weight vector. The analytic hierarchy process is a multi-criteria decision-making method that decomposes complex problems into multiple levels by constructing a hierarchical structure, and compares the elements of each level pairwise to finally obtain the weight of each element. Assuming there are three parameters , , , through the hierarchical analysis method, the sensitivity weight vector is obtained The optimization objective function is constructed based on the sensitivity weight vector. The purpose of the optimization objective function is to maximize or minimize a target value, usually a quality indicator of biochar. Assume that the goal is to maximize the specific surface area , then the optimization objective function can be expressed as:

[0116] ;

[0117] Particle swarm optimization is performed on the optimization objective function. Particle swarm optimization is an optimization algorithm based on swarm intelligence. It finds the global optimal solution by simulating the movement of particles in the search space. The position of each particle represents a parameter combination, and the speed represents the direction and amplitude of the parameter adjustment. The algorithm gradually approaches the optimal solution by iteratively updating the position and speed of the particles. Through particle swarm optimization, a candidate parameter set C is obtained. The corresponding parameter probability distribution is calculated based on the candidate parameter set. Assume that there are several candidate parameter combinations , by counting the values ​​of each parameter, we can get the probability distribution of the parameter . Perform confidence interval analysis on the parameter probability distribution to obtain the credible range of the parameter. The confidence interval is an estimate of the range of the true value of the parameter. Assuming a 95% confidence level is selected, the confidence interval of the parameter is obtained by calculating the sample mean and standard error of the parameter. ,in is the standard error of the parameter, and z is the quantile of the standard normal distribution. The first preparation parameter combination is adjusted and optimized according to the parameter credible range to obtain the second preparation parameter combination. The optimized parameter combination is obtained by fine-tuning the parameters to ensure that they are within the confidence interval and meet the optimization goal at the same time.

[0118] The above describes the information management method for preparing biochar-based fertilizer in the embodiment of the present application. The following describes the information management system for preparing biochar-based fertilizer in the embodiment of the present application. Figure 2 In one embodiment of the present application, an information management system for preparing biochar-based fertilizer includes:

[0119] A digitization module 201 is used to digitize the parameters of the biochar-based fertilizer preparation process to obtain a parameter vector set;

[0120] A processing module 202 is used to perform local optimization and global optimization processing on the parameter vector set to obtain a first preparation parameter combination;

[0121] A matrixing module 203 is used to perform matrixing and standardization processing on the first preparation parameter combination to obtain a preparation process characteristic matrix;

[0122] An execution module 204 is used to perform characteristic vector centrality processing on the characteristic matrix of the preparation process to obtain a target parameter execution mode;

[0123] The analysis module 205 is used to obtain the carbonized biochar and the carbonization process parameter set according to the target parameter execution mode, and perform spectral analysis on the carbonized biochar to obtain a biochar quality data set;

[0124] The parameter optimization module 206 is used to optimize the first preparation parameter combination according to the biochar quality data set and the carbonization process parameter set to obtain a second preparation parameter combination.

[0125] Through the synergy of the above components, the precise numerical representation of the parameters of the preparation process is achieved by digitally encoding the types of raw materials, discretizing the pyrolysis temperature, and mathematical modeling of the cooling rate. The efficiency and accuracy of parameter optimization are improved by combining local optimization with global optimization, and the optimal preparation parameter combination can be found more quickly. Through matrixization and standardization, combined with technologies such as singular value decomposition and non-negative matrix decomposition, the comprehensive extraction of preparation process characteristics is achieved. Through feature vector centrality processing, the relationship between features is considered, and the generated target parameter execution mode has strong adaptability. Combined with Fourier transform infrared spectroscopy and X-ray fluorescence spectroscopy analysis, the quality of biochar is evaluated in multiple dimensions, which improves the accuracy and reliability of quality data. Through neural network models and sensitivity analysis, dynamic optimization of preparation parameters is achieved, and the preparation plan can be continuously adjusted and improved according to actual production conditions, providing intelligent decision support for the preparation process of biochar-based fertilizers, which helps to improve production efficiency and product quality.

[0126] The present application also provides an information management device for the preparation of biochar-based fertilizers, the information management device for the preparation of biochar-based fertilizers comprising a memory and a processor, the memory storing computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes the steps of the information management method for the preparation of biochar-based fertilizers in the above-mentioned embodiments.

[0127] The present application also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein instructions are stored in the computer-readable storage medium, and when the instructions are executed on a computer, the computer executes the steps of the information management method for preparing biochar-based fertilizer.

[0128] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0129] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0130] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An information management method for preparing biochar-based fertilizer, characterized in that: The method comprises: The preparation process parameters of the biochar-based fertilizer are numerically processed to obtain a parameter vector set; specifically, the process includes: digitally encoding the raw material types of the biochar-based fertilizer to obtain a raw material type code set, and numerically processing the raw material ratio according to the raw material type code set to obtain a raw material ratio vector; discretizing the pyrolysis temperature of the biochar-based fertilizer to obtain a pyrolysis temperature discrete value set, and performing piecewise function mapping on the pyrolysis time of the biochar-based fertilizer to obtain a pyrolysis time vector; mathematically modeling the cooling rate of the biochar-based fertilizer to obtain a cooling rate function, and parameterizing the types of additives of the biochar-based fertilizer Representation, obtain an additive feature matrix; normalize the additive dosage of the biochar-based fertilizer to obtain an additive dosage normalized vector, and feature encode the production equipment parameters of the biochar-based fertilizer to obtain an equipment parameter vector; quantitatively evaluate the environmental factors of the biochar-based fertilizer to obtain an environmental impact index vector; perform multi-dimensional vector splicing according to the raw material ratio vector, the pyrolysis temperature discrete value set, the pyrolysis time vector, the cooling rate function, the additive feature matrix, the additive dosage normalized vector, the equipment parameter vector and the environmental impact index vector to obtain a parameter vector set; The parameter vector set is locally optimized and globally optimized to obtain a first preparation parameter combination; specifically comprising: performing principal component analysis on the parameter vector set to obtain a parameter space after dimension reduction, and performing cluster analysis based on the parameter space after dimension reduction to obtain an initial local optimization region; performing gradient descent processing on the initial local optimization region to obtain a local optimal solution set, and constructing a global optimization initial population based on the local optimal solution set; performing fitness evaluation on the global optimization initial population to obtain a fitness matrix, and performing selection, crossover and mutation operations on the global optimization initial population based on the fitness matrix to obtain an evolved population; performing local search on the evolved population to obtain a mixed optimization solution set, and constructing a parameter response surface based on the mixed optimization solution set; optimizing the parameter response surface to obtain a Pareto optimal solution set, and performing decision analysis based on the Pareto optimal solution set to obtain a first preparation parameter combination; The first preparation parameter combination is matrixed and standardized to obtain a preparation process characteristic matrix; specifically comprising: performing parameter sorting on the first preparation parameter combination to obtain an ordered parameter set, and constructing an original parameter matrix based on the ordered parameter set; performing multiple interpolation processing on the original parameter matrix to obtain a complete parameter matrix, and calculating the descriptive statistics of each preparation parameter based on the complete parameter matrix to obtain a parameter statistical feature set; transforming the parameter statistical feature set to obtain normalized parameter data, and standardizing based on the normalized parameter data to obtain a standardized parameter matrix; performing singular value decomposition on the standardized parameter matrix to obtain a reduced dimension parameter matrix, and calculating the information gain ratio of the reduced dimension parameter matrix to obtain a parameter weight vector; performing matrix multiplication on the parameter weight vector and the reduced dimension parameter matrix to obtain a weighted parameter matrix, and performing non-negative matrix decomposition on the weighted parameter matrix to obtain a preparation process characteristic matrix; Performing eigenvector centrality processing on the characteristic matrix of the preparation process to obtain a target parameter execution mode; Acquire the carbonized biochar and a carbonization process parameter set according to the target parameter execution mode, and perform spectral analysis on the carbonized biochar to obtain a biochar quality data set; The first preparation parameter combination is optimized according to the biochar quality data set and the carbonization process parameter set to obtain a second preparation parameter combination.

2. The information management method for preparing biochar-based fertilizer according to claim 1, characterized in that: The step of performing characteristic vector centrality processing on the characteristic matrix of the preparation process to obtain a target parameter execution mode includes: Generating a feature vector set according to the preparation process feature matrix, and calculating the cosine similarity of each vector according to the feature vector set to obtain a similarity matrix; Performing centrality score calculation on the similarity matrix to obtain an initial centrality score, and constructing a feature relationship network according to the initial centrality score; Performing feature clustering on the feature relationship network to obtain feature clustering results, and calculating inter-class and intra-class centrality based on the feature clustering results to obtain a multi-scale centrality index; Comprehensively evaluating the multi-scale centrality index to obtain a comprehensive centrality score, and sorting the feature vector set according to the comprehensive centrality score to obtain a key feature sequence; A feature weight analysis is performed on the key feature sequence to obtain feature importance weights, and a target parameter execution mode is generated according to the feature importance weights.

3. The information management method for preparing biochar-based fertilizer according to claim 1, characterized in that: The step of acquiring the carbonized biochar and the carbonization process parameter set according to the target parameter execution mode, and performing spectral analysis on the carbonized biochar to obtain a biochar quality data set includes: Generate a carbonization process parameter set according to the target parameter execution mode, and perform pyrolysis and carbonization of the raw material according to the carbonization process parameter set to obtain carbonized biochar; The temperature, pressure and gas composition of the carbonization process are monitored in real time to obtain a carbonization process parameter set; Performing Fourier transform infrared spectroscopy scanning on the carbonized biochar to obtain infrared spectral data, and performing baseline correction and normalization processing on the infrared spectral data to obtain standardized spectral data; Performing feature extraction on the standardized spectral data to obtain a spectral feature vector, and performing X-ray fluorescence spectroscopy analysis on the carbonized biochar to obtain elemental composition data; Calculating element ratios and content indices according to the element composition data to obtain element feature vectors, and performing feature fusion on the spectral feature vectors and the element feature vectors to obtain a comprehensive feature matrix; The comprehensive feature matrix was subjected to partial least squares regression analysis to obtain a biochar quality data set.

4. The information management method for preparing biochar-based fertilizer according to claim 1, characterized in that: The step of optimizing the first preparation parameter combination according to the biochar quality data set and the carbonization process parameter set to obtain a second preparation parameter combination includes: Performing correlation analysis on the biochar quality data set and the carbonization process parameter set to obtain a parameter correlation matrix; Constructing an initial prediction model according to the parameter correlation matrix, and performing cross-validation training on the initial prediction model to obtain an optimized neural network model; Performing a sensitivity analysis on the first preparation parameter combination according to the optimized neural network model to obtain a parameter sensitivity ranking; Performing a hierarchical analysis on the parameter sensitivity ranking to obtain a sensitivity weight vector, and constructing an optimization objective function according to the sensitivity weight vector; Performing particle swarm optimization on the optimization objective function to obtain a candidate parameter set, and calculating corresponding parameter probability distribution according to the candidate parameter set; A confidence interval analysis is performed on the parameter probability distribution to obtain a parameter credible range, and the first preparation parameter combination is adjusted and optimized according to the parameter credible range to obtain a second preparation parameter combination.

5. An information management system for the preparation of biochar-based fertilizer, characterized in that: The information management method for executing the preparation of biochar-based fertilizer according to any one of claims 1 to 4, the system comprising: A numerical module is used to perform numerical processing on the parameters of the preparation process of biochar-based fertilizer to obtain a parameter vector set; A processing module, used for performing local optimization and global optimization processing on the parameter vector set to obtain a first preparation parameter combination; A matrixing module, used for performing matrixing and standardization processing on the first preparation parameter combination to obtain a preparation process characteristic matrix; An execution module, used for performing characteristic vector centrality processing on the characteristic matrix of the preparation process to obtain a target parameter execution mode; An analysis module, used for acquiring the carbonized biochar and a carbonization process parameter set according to the target parameter execution mode, and performing spectral analysis on the carbonized biochar to obtain a biochar quality data set; A parameter optimization module is used to optimize the first preparation parameter combination according to the biochar quality data set and the carbonization process parameter set to obtain a second preparation parameter combination.

6. An information management device for preparing biochar-based fertilizer, characterized in that: The information management device for preparing biochar-based fertilizer comprises: a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instructions in the memory so that the information management device for biochar-based fertilizer preparation executes the information management method for biochar-based fertilizer preparation as described in any one of claims 1-4.

7. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the information management method for preparing biochar-based fertilizer according to any one of claims 1 to 4 is implemented.

Images