A method for optimizing microplastic degradation rate and CO2 fixation rate

Abstract

The present invention relates to a method for optimizing microplastic degradation rate and CO2 fixation rate, comprising: obtaining an initial population, wherein the individuals of the initial population are target data for different soil types; obtaining the microplastic degradation rate and CO2 fixation rate of each individual through a machine learning model, wherein the machine learning model is trained with a training set, wherein the training set includes historical target influencing factors of different soil types and corresponding microplastic degradation rates and CO2 fixation rates; and optimizing the microplastic degradation efficiency and carbon dioxide fixation efficiency using a genetic algorithm to obtain the optimal combination of microplastic degradation rate and soil carbon fixation rate. The present invention improves soil remediation efficiency while optimizing the synchronous process of microplastic degradation and CO2 fixation.

Description

Technical Field

[0001] The present invention relates to the technical field of soil remediation, and in particular to a method for optimizing microplastic degradation rate and CO2 fixation rate. Background Art

[0002] With the global population growth and the intensification of human activities, soil, as one of the largest carbon reservoirs in the Earth system, plays a vital role in fixing carbon dioxide and is of great significance for mitigating climate change. The fixation of carbon dioxide in the soil not only helps to reduce the concentration of greenhouse gases in the atmosphere, but also improves soil structure and enhances its fertility and productivity. However, on the other hand, the extensive use of plastic products has led to the widespread release of microplastics (i.e., plastic particles with a diameter of less than 5 mm), which gradually accumulate in the soil, especially in agricultural soils. Microplastic pollution has become a serious problem. In particular, in farmland soils where plastic mulch has been used for a long time, the content of microplastics can reach hundreds to thousands of particles per kilogram, which poses a serious threat to soil health and ecological functions.

[0003] Although traditional physical and chemical improvement methods have certain applications in soil remediation and microplastic treatment, they have obvious limitations. These methods are often prone to secondary pollution. For example, toxic intermediates may be produced during the chemical decomposition of microplastics. In addition, these methods are usually costly and complex to operate, making them difficult to promote and apply on a large scale. More importantly, existing methods fail to effectively combine the needs of soil remediation and carbon dioxide fixation, and fail to simultaneously achieve microplastic degradation and CO2 storage. Therefore, they cannot achieve multiple environmental protection goals in practical applications.

[0004] In the quest to simultaneously increase microplastic degradation and CO2 fixation, existing technologies face significant challenges. In particular, existing methods lack effective optimization of microbial metabolic pathways, which are considered key to achieving simultaneous increases in microplastic degradation and CO2 fixation. Therefore, to overcome these limitations, there is an urgent need to develop new technological approaches that can simultaneously enhance microplastic degradation and CO2 fixation by optimizing microbial metabolic pathways, thereby achieving the dual goals of soil remediation and environmental protection.

[0005] Taiyuan University of Technology's "A Biological Control Method for Geologically Sequestered CO2 Leakage" aims to provide a biological control method for geologically sealed CO2 leaks. This method uses the biochemical reaction between microorganisms, calcium sources, and CO2 to form carbonate minerals to cement the pore and fissure structure of the formation, blocking the leakage channel of CO2, thereby providing an effective leakage prevention solution. However, the reaction efficiency of this method is greatly affected by environmental factors (such as temperature, pH value, calcium source concentration, etc.), making it difficult to accurately and dynamically predict the CO2 fixation efficiency. In addition, this method requires external force to achieve its implementation, and the instability of the external force supply may affect the CO2 fixation efficiency.

[0006] The Nanjing Institute of Soil Science, Chinese Academy of Sciences, published a paper titled "A Synthetic Microbial Community and Composite Agent for Degrading Microplastics and Their Application in the Remediation of Composite Pollutants." These synthetic bacterial communities and composite agents are capable of effectively degrading microplastics, herbicides, and polycyclic aromatic hydrocarbons (PAHs), as well as removing heavy metals. However, the degradation efficiency of this method is affected by the type of microplastic, environmental conditions (such as temperature, pH, and humidity), and composite pollutants (such as heavy metals), making it difficult to accurately predict and optimize.

[0007] Although the above-mentioned prior arts involve independent aspects of CO2 fixation and microplastic degradation, they do not provide an effective strategy to simultaneously optimize the microplastic degradation rate and CO2 fixation rate. Summary of the Invention

[0008] The purpose of the present invention is to provide a method for optimizing the microplastic degradation rate and CO2 fixation rate, so as to achieve accurate prediction and optimization of microplastic degradation and CO2 fixation efficiency, thereby improving soil remediation efficiency.

[0009] To achieve the above object, the present invention provides the following solutions:

[0010] A method for optimizing microplastic degradation rate and CO2 fixation rate, comprising:

[0011] Obtain an initial population, wherein the individuals of the initial population are target data of different soil types, and the microplastic degradation rate and CO2 fixation rate of each individual are obtained through a machine learning model, and the machine learning model is obtained by training a training set, and the training set includes historical target influencing factors of different soil types and corresponding microplastic degradation rates and CO2 fixation rates;

[0012] A genetic algorithm is used to optimize the microplastic degradation efficiency and carbon dioxide fixation efficiency to obtain the optimal combination of microplastic degradation rate and soil carbon fixation rate.

[0013] Optionally, obtaining the initial population includes:

[0014] Obtaining a set of high-quality solutions that meet preset conditions from experimental data, performing non-dominated sorting on the high-quality solutions, and selecting a core population;

[0015] A supplementary solution set is generated in the input parameter space through Latin hypercube sampling, and the supplementary solution set and the core population are merged to obtain the initial population.

[0016] Optionally, obtaining the target data of different soil types includes:

[0017] Obtain initial target data of different soil types, wherein the initial target data include physical and chemical properties of different soil types, types of inoculated microorganisms, concentration characteristics, and types of microplastics, particle morphology, initial concentration, degradation rate, carbon dioxide concentration, and carbon dioxide fixation. Perform principal component analysis, dimensionality reduction, and processing on the target data of different soil types to obtain the target data of different soil types.

[0018] Optionally, the physical and chemical property data of different soil types include: soil pH value, temperature, moisture, porosity and particle size distribution.

[0019] Optionally, the machine learning model is obtained by Bayesian optimization of model parameters, L1 regularization and L2 regularization, wherein the machine learning is constructed using at least one of linear regression, support vector machine, decision tree, K nearest neighbor algorithm, random forest algorithm, gradient boosting tree algorithm, XGBoost, AdaBoost, LightGBM, and neural network.

[0020] Optionally, a genetic algorithm is used to optimize the microplastic degradation efficiency and carbon dioxide fixation efficiency, including:

[0021] Determine the objective function and its constraints;

[0022] The initial population is selected, crossed, mutated, and elitist according to the objective function and its constraints to obtain the next generation population, and the process is repeated to a preset number of iterations to obtain the optimal combination of the microplastic degradation rate and the soil carbon sequestration rate, wherein a new solution generated by LHS is added to the population when the preset number of iterations is reached.

[0023] Optionally, the objective function is:

[0024] F(D,C)=w D ·D+w C ·C

[0025] The constraints of the objective function are:

[0026] D≧D {min}

[0027] C≧C {min} )

[0028] Among them, D is the degradation rate of microplastics, C is the soil carbon sequestration rate, F is the objective function, and w D and w C is the weight coefficient, which determines the relative importance of microplastic degradation rate and soil carbon sequestration rate in the objective function. {min} is the minimum threshold of microplastic degradation rate, C {min} is the minimum threshold of soil carbon sequestration rate.

[0029] The beneficial effects of the present invention are as follows: The purpose of the present invention is to provide an innovative soil remediation method that achieves efficient degradation of microplastics in soil and precise fixation of carbon dioxide by combining machine learning technology and microbially induced carbon dioxide mineralization. This method aims to overcome the limitations of traditional soil remediation technology and achieve simultaneous optimization of microplastic degradation and CO2 fixation efficiency through intelligent prediction and dynamic regulation to achieve the dual goals of soil remediation and environmental protection, significantly improving the efficiency and adaptability of soil remediation, while improving soil properties and restoring ecological functions. It provides an environmentally friendly, efficient and sustainable technical solution for solving microplastic pollution and promoting carbon recycling, and has broad application potential and important environmental benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0031] Figure 1 This is a flow chart of a method for optimizing microplastic degradation rate and CO2 fixation rate according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0033] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0034] Example 1:

[0035] This embodiment provides a method for optimizing microplastic degradation rate and CO2 fixation rate, including:

[0036] An initial population is obtained, where the individuals in the initial population are target data for different soil types. The microplastic degradation rate and CO2 fixation rate of each individual are obtained through a machine learning model. The machine learning model is trained with a training set that includes historical target influencing factors for different soil types and the corresponding microplastic degradation rate and CO2 fixation rate.

[0037] A genetic algorithm is used to optimize the microplastic degradation efficiency and carbon dioxide fixation efficiency to obtain the optimal combination of microplastic degradation rate and soil carbon fixation rate.

[0038] Furthermore, target data for different soil types are obtained including:

[0039] Obtain initial target data for different soil types, where the initial target data include physical and chemical properties of different soil types, types of inoculated microorganisms, concentration characteristics, and types, particle morphology, initial concentration, degradation rate, carbon dioxide concentration, and carbon dioxide fixation of microplastics. Perform principal component analysis, dimensionality reduction, and processing on the target data of different soil types to obtain target data of different soil types.

[0040] Specifically, obtaining a training set involves designing an experimental plan and collecting data on the physical and chemical properties of different soil types in the experiment, the types of microorganisms inoculated, concentration characteristics, and the type, particle morphology, initial concentration, degradation rate, carbon dioxide concentration, and carbon dioxide fixation of microplastics. The various variables in the experimental data have different dimensions, so data standardization and normalization are required to preprocess the data. More meaningful feature variables such as microplastic degradation efficiency, CO2 fixation efficiency, soil pH, temperature, and other features are extracted, and redundant or irrelevant features are removed. The dataset is divided into a training set and a test set, and predictions are made using 10 machine learning algorithms, including traditional algorithms and ensemble algorithms.

[0041] In this embodiment, a normalization method is used to scale the data to the interval [0, 1]. When using the standardization method, the mean and standard deviation of each variable are calculated and normalized using the Z-score.

[0042] In this embodiment, the data set is divided into a training set and a test set: 70% of the data is used for training and 30% of the data is used for testing. The 10 machine learning algorithms are: Linear Regression, Support Vector Machine (SVM), Decision Tree, K-Nearest Neighbor Algorithm (KNN), Random Forest (RF), Gradient Boosting Tree (GBM), XGBoost, AdaBoost, LightGBM, and Neural Network (ANN).

[0043] Furthermore, the machine learning model is obtained by Bayesian optimization of model parameters, L1 regularization and L2 regularization, wherein the machine learning is constructed using at least one of linear regression, support vector machine, decision tree, K nearest neighbor algorithm, random forest algorithm, gradient boosting tree algorithm, XGBoost, AdaBoost, LightGBM, and neural network.

[0044] Specifically, this embodiment uses common mean square error (MSE), root mean square error (RMSE), determination coefficient (R2), mean absolute error (MAE), and mean absolute percentage error (MAPE) for evaluation, and adjusts the model's hyperparameters. In order to find the optimal microplastic degradation rate and soil carbon sequestration rate for different soil types, it is necessary to add an optimization algorithm. Among them, the model hyperparameter adjustment adopts the Bayesian optimization method. Compared with grid and random search, Bayesian optimization can find the optimal hyperparameters with fewer experiments, and the optimization algorithm adopts a genetic algorithm. L1 regularization and L2 regularization are adopted to reduce model complexity, prevent overfitting, and reduce the variance of the model. In the neural network model, a learning rate scheduler is used to dynamically adjust the learning rate, and an early stopping technique is used to avoid overfitting.

[0045] Furthermore, obtaining the initial population includes:

[0046] Obtain a set of high-quality solutions that meet preset conditions from experimental data, perform non-dominated sorting on the high-quality solutions, and select the core population;

[0047] A supplementary solution set is generated in the input parameter space through Latin hypercube sampling, and the supplementary solution set is merged with the core population to obtain the initial population.

[0048] Furthermore, the genetic algorithm is used to optimize the microplastic degradation efficiency and carbon dioxide fixation efficiency, including:

[0049] Determine the objective function and its constraints;

[0050] According to the objective function and its constraints, the initial population is selected, crossed, mutated, and elitist to obtain the next generation population, and the process is repeated to a preset number of iterations to obtain the optimal combination of microplastic degradation rate and soil carbon sequestration rate. When the number of iterations reaches the preset number, the new solution generated by LHS is added to the population.

[0051] Furthermore, the objective function is:

[0052] F(D,C)=w D ·D+w C ·C

[0053] The constraints of the objective function are:

[0054] D≧D {min}

[0055] C≧C {min} )

[0056] Among them, D is the degradation rate of microplastics, C is the soil carbon sequestration rate, F is the objective function, and w D and w C is the weight coefficient, which determines the relative importance of microplastic degradation rate and soil carbon sequestration rate in the objective function. {min} is the minimum threshold of microplastic degradation rate, C {min} is the minimum threshold of soil carbon sequestration rate.

[0057] Example 2:

[0058] like Figure 1 As shown, this embodiment provides a method for optimizing microplastic degradation rate and CO2 fixation rate, including:

[0059] (1) The experiment will explore three different types of soil: sandy soil, saline-alkali soil, and clay soil. These three types of soil have significant differences in physical and chemical properties, which may have different effects on microbial activity and microplastic degradation. Each soil will be used in a microbial mineralization cell experiment to study its efficiency in microplastic degradation and carbon dioxide sequestration.

[0060] (2) The pH value, temperature, humidity, porosity and particle size distribution of each soil type in the collection experiment will be measured separately. These parameters will affect microbial activity and the physical properties of the soil. For each soil, at least 5 physical and chemical data items × 3 soil types × 5 sampling points = 75 raw data items will be recorded.

[0061] (3) Collect the inoculated microbial species and record 2 parameters × 5 concentration points × 3 soil types = 30 raw data points. Concentration characteristics and microplastic type, particle morphology, initial concentration, and degradation rate were recorded, and 3 types × 3 morphologies × 3 initial concentrations × 3 soil types = 81 raw data points were recorded. Data on carbon dioxide concentration and carbon dioxide fixation were also recorded. Each experiment was repeated 3 times, and 90 raw data points were collected (2 data types × 5 sampling times × 3 repetitions) × 3 soil types = 90 raw data points.

[0062] (4) There are 67 original input variables. After dimensionality reduction by principal component analysis (PCA), 18 input variables are retained. The variables in the experimental data have different dimensions. Duplicate data points and data that are displayed as zero due to unsuccessful collection are removed. The other data are scaled to the interval [0,1] and normalized. The mean and standard deviation of each variable are calculated and Z-score standardization is performed. By subtracting the mean and dividing by the standard deviation, the data has zero mean and unit variance, eliminating the dimensional differences between the data and improving the training effect of the model.

[0063] (5) The dataset was divided into a 70% training set and a 30% test set. Ten machine learning algorithms were used: linear regression, support vector machine (SVM), decision tree, K-nearest neighbor (KNN), random forest (RF), gradient boosting tree (GBM), XGBoost, AdaBoost, LightGBM, and neural network (ANN). The ten machine learning algorithms, including traditional algorithms and ensemble algorithms, were used to predict the microplastic degradation efficiency and carbon dioxide fixation efficiency.

[0064] (6) Use common mean square error (MSE), root mean square error (RMSE), coefficient of determination (R 2 ), mean absolute error (MAE), and mean absolute percentage error (MAPE) were used for evaluation. Bayesian optimization was used to adjust model hyperparameters. Compared with grid search and random search, Bayesian optimization can find the optimal hyperparameters with fewer experiments. The optimal model is XGboost. The specific parameter adjustment values ​​are max_depth = 3-10; min_child_weight = 1-10; gamma = 0-5; eta = 0.01-0.3; n_estimators = 100-1000;

[0065] (7) L1 regularization and L2 regularization are used to reduce model complexity, prevent overfitting, and reduce model variance. In the neural network model, a learning rate scheduler is used to dynamically adjust the learning rate, and early stopping technology is used to avoid overfitting. The specific parameters are reg_alpha = 0-1 (L1 regularization); reg_lambda = 0-1 (L2 regularization); learning_rate = 0.01-0.1;

[0066] (8) In order to find the optimal combination of microplastic degradation rate and soil carbon sequestration rate under different soil types, a genetic algorithm will be used for optimization. The genetic algorithm can explore the global optimal solution and help find the appropriate parameter combination in complex multidimensional problems. The objective function is to maximize the comprehensive benefits of microplastic degradation rate and soil carbon sequestration rate. Let (D) be the microplastic degradation rate and (C) be the soil carbon sequestration rate. The objective function (F) can be defined as:

[0067] F(D,C)=w D ·D+w C ·C

[0068] Among them, w D and w C are weight coefficients, which determine the relative importance of microplastic degradation rate and soil carbon sequestration rate in the objective function. The sum of the weight coefficients should be 1, that is, w D +wC =1.

[0069] In addition, some constraints are set for the objective function, and the microplastic degradation rate D must be greater than or equal to the minimum threshold (D≧D {min} ), D {min} will be set to 0.3, and the soil carbon sequestration rate C must be greater than or equal to the minimum threshold (C≧C {min} ), C {min} will be set to 0.05.

[0070] (9) During the optimization process of the genetic algorithm, considering that historical data may not cover all potential optimization spaces, there is a risk of falling into local optimality, and some parameter combinations may not be covered by the experiment, it is necessary to use a mixed population initialization method driven by historical data to further explore the algorithm. First, a high-quality solution set that satisfies D≥0.3 and C≥0.05 is extracted from the experimental data, and the high-quality solution set is non-dominated sorted, and 20% of the Pareto frontier is selected as the core population. Latin hypercube sampling (LHS) is used to generate a supplementary solution set in the input parameter space. The new solution generated by LHS is merged with the core population to form an initial population (100 individuals). The microplastic degradation rate and carbon dioxide fixation rate are predicted by two XGboosts, and the D and C values ​​of each individual in the initial population are predicted. For individuals that do not meet the constraints, a penalty function is used to reduce their fitness value; the tournament selection method is used to select parent individuals for reproduction (size = 3); a single-point crossover operation is performed at a crossover rate of 0.9 to generate a new generation of individuals; random mutations are performed on the newly generated individuals at a mutation rate of 0.01; the best individuals in the previous generation are retained and directly enter the next generation; the above steps of selection, crossover, mutation, and elitism are repeated until the genetic generation reaches 500, and 5 new solutions generated by LHS are injected every 20 generations to prevent the population from converging prematurely. After 500 generations, the individual with the highest fitness value is found as the optimal solution, that is, the optimal combination of microplastic degradation rate and soil carbon sequestration rate.

[0071] (9) In order to gain a deeper understanding of the prediction mechanism of the optimal model XGBoost, the SHAP (SHapley Additive exPlanations) feature importance analysis method based on cooperative game theory will be used. The SHAP method can quantify the contribution of each feature to the model prediction results, thereby revealing the importance of each feature in the prediction of microplastic degradation and carbon dioxide fixation efficiency. In addition, the LIME (Local Interpretable Model-agnostic Explanations) method is used to further explain the local prediction effect of the optimal model on specific data points. LIME can provide a simple, interpretable model for each prediction, which approximates the behavior of the original complex model in local areas, thereby helping to understand the prediction basis of the optimal model in a specific situation.

[0072] The present invention is based on a machine learning approach. By establishing a prediction model, the metabolic activity of microorganisms is combined to trigger the mineralization of carbon dioxide, promote CO2 fixation, improve soil properties, and cause redox reactions. The generated electrical energy and chemical energy accelerate the degradation of microplastics through electron transfer, thereby achieving accurate prediction and optimization of microplastic degradation and CO2 fixation efficiency. The machine learning model can automatically learn and analyze multiple variables such as soil type, microbial activity, and environmental factors, dynamically adjust the remediation strategy, and significantly improve the efficiency and accuracy of the remediation process. While improving soil remediation efficiency, this method optimizes the synchronous process of microplastic degradation and CO2 fixation, and has good environmental adaptability and broad application prospects.

[0073] The embodiments described above are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by persons skilled in the art should fall within the scope of protection defined by the claims of the present invention.

Claims

1. A method for optimizing microplastic degradation rate and CO2 fixation rate, characterized in that: include: Obtain an initial population, wherein the individuals of the initial population are target data of different soil types, and the microplastic degradation rate and CO2 fixation rate of each individual are obtained through a machine learning model, and the machine learning model is obtained by training a training set, and the training set includes historical target influencing factors of different soil types and corresponding microplastic degradation rates and CO2 fixation rates; Acquiring target data of different soil types includes: Obtaining initial target data for different soil types, wherein the initial target data includes physical and chemical properties of different soil types, inoculated microbial species, concentration characteristics, and microplastic type, particle morphology, initial concentration, degradation rate, carbon dioxide concentration, and carbon dioxide fixation; performing principal component analysis dimensionality reduction and processing on the initial target data for different soil types to obtain target data for the different soil types; A genetic algorithm is used to optimize the microplastic degradation efficiency and carbon dioxide fixation efficiency to obtain the optimal combination of microplastic degradation rate and soil carbon fixation rate.

2. The method for optimizing microplastic degradation rate and CO2 fixation rate according to claim 1, characterized in that: Acquiring the initial population includes: Obtaining a set of high-quality solutions that meet preset conditions from experimental data, performing non-dominated sorting on the high-quality solutions, and selecting a core population; A supplementary solution set is generated in the input parameter space through Latin hypercube sampling, and the supplementary solution set and the core population are merged to obtain the initial population.

3. The method for optimizing microplastic degradation rate and CO2 fixation rate according to claim 1, characterized in that: The physical and chemical properties of different soil types include: soil pH, temperature, moisture, porosity, and particle size distribution.

4. The method for optimizing microplastic degradation rate and CO2 fixation rate according to claim 1, characterized in that: The machine learning model is obtained by Bayesian optimization of model parameters, L1 regularization and L2 regularization, wherein the machine learning is constructed using at least one of a support vector machine, a decision tree, and a K-nearest neighbor algorithm.

5. The method for optimizing microplastic degradation rate and CO2 fixation rate according to claim 1, characterized in that: The optimization of microplastic degradation efficiency and carbon dioxide fixation efficiency using genetic algorithms includes: Determine the objective function and its constraints; The initial population is selected, crossed, mutated, and elitist according to the objective function and its constraints to obtain the next generation population, and the process is repeated to a preset number of iterations to obtain the optimal combination of the microplastic degradation rate and the soil carbon sequestration rate, wherein a new solution generated by LHS is added to the population when the preset number of iterations is reached.

6. The method for optimizing microplastic degradation rate and CO2 fixation rate according to claim 5, characterized in that: The objective function is: F(D,C)=w D ·D+w C ·C The constraints of the objective function are: D≧D {min} C≧C {min} Among them, D is the degradation rate of microplastics, C is the soil carbon sequestration rate, F is the objective function, and w D and w C is the weight coefficient, which determines the relative importance of microplastic degradation rate and soil carbon sequestration rate in the objective function. {min} is the minimum threshold of microplastic degradation rate, C {min} is the minimum threshold of soil carbon sequestration rate.

Images