Machine learning optimization method and system for low-temperature high-activity CO2 methanation catalyst

Abstract

The invention provides a machine learning optimization method and system for a low-temperature high-activity CO2 methanation catalyst, and relates to the technical field of machine learning. The optimization method comprises the following steps: collecting CO2 methanation reaction data, building a machine learning model for hyper-parameter optimization, selecting an optimal model, carrying out interpretable analysis on the optimal model, and screening the catalyst by combining the optimal model with a multi-objective optimization algorithm, so that the machine learning optimization of the CO2 methanation catalyst is realized, and the machine learning can analyze a large amount of data so as to realize the optimization of the CO2 methanation catalyst. The catalyst with high activity and selectivity is rapidly predicted and screened out, and the research and development efficiency is greatly improved; machine learning can reduce experiment times, optimize experiment design, significantly reduce research and development cost, and improve research and development success rate; machine learning can identify and utilize complex chemical and physical data, a catalyst with better performance is designed in a high-precision manner, and the conversion rate and selectivity of converting CO2 into CH4 are improved.

Description

Technical Field

[0001] The present invention relates to the field of machine learning technology, and in particular to a machine learning optimization method and system for a low-temperature and high-activity CO2 methanation catalyst. Background Art

[0002] CO2 methanation technology converts CO2 and H2 into CH4 and H2O. The exothermic nature of the reaction gives it a thermodynamic advantage in the temperature range of ≤400°C, resulting in high selectivity for CH4.

[0003] However, due to the inherent stability of CO2 molecules and the higher temperatures usually required for their activation, when the reaction temperature is below 250 °C, CO2 activation is not complete, resulting in reduced CO2 conversion and CH4 production.

[0004] Therefore, CO2 to CH4 is a challenging process that requires optimal operating conditions and appropriate catalysts to achieve high CH4 yield and selectivity. Given the endothermic side reactions, the activity of the catalyst at lower temperatures must be improved to mitigate the large increase in byproduct CO yield observed at high temperatures.

[0005] In the research and development of CO2 methanation catalysts, traditional trial and error methods often rely on the experience and intuition of researchers, which is time-consuming, costly, and difficult to systematically explore all possibilities. For example, Chinese patent CN118287098A discloses a cobalt-based catalyst for low-temperature carbon dioxide hydrogenation methanation reaction and a preparation method thereof, and prepares a low-temperature and highly active cobalt-based catalyst. However, the preparation process of the catalyst is complicated, involving a variety of raw materials and precise ratio control, which not only increases the material cost, but also puts forward higher requirements for experimental operations, while prolonging the research and development cycle. This complexity highlights the limitations of traditional optimization methods in terms of efficiency and resource utilization. Summary of the invention

[0006] In view of the shortcomings of the prior art, the present invention provides a machine learning optimization method and system for a low-temperature and high-activity CO2 methanation catalyst, which solves the problems of low R&D efficiency, high cost and low precision of CO2 methanation catalysts.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0008] A machine learning optimization method for a low-temperature, high-activity CO2 methanation catalyst, the optimization method comprising the following steps:

[0009] S1. Collecting CO2 methanation reaction catalyst descriptors, preparation conditions and reaction condition parameters;

[0010] S2, preprocessing the CO2 methanation reaction data set collected by S1;

[0011] S3, import the data processed by S2 into multiple machine learning models respectively to perform hyperparameter tuning;

[0012] S4, select the best model according to the prediction accuracy of the model after S3 tuning, and use model interpretation technology to perform interpretability analysis on the best model;

[0013] S5. Based on the insights gained from the interpretation of the S4 model, a low-temperature, high-activity CO2 methanation reaction catalyst was screened out based on a multi-objective optimization algorithm combined with the best model optimization.

[0014] Preferably, in said S1, the catalyst descriptor includes: catalyst composition and content;

[0015] Preparation conditions include: preparation method, calcination temperature, calcination time, reduction hydrogen content, reduction temperature, reduction time and reduction pressure;

[0016] The reaction condition parameters include: reaction temperature, pressure, weight hourly space velocity, inert component content in the raw materials and H2 / CO2 ratio.

[0017] Preferably, in S1, the output characteristics predicted by the collected input characteristics include: CO2 conversion rate, CH4 selectivity and CH4 yield.

[0018] Preferably, in S2, the preprocessing of the data set includes the following steps:

[0019] Label encoding is used to convert categorical data into numerical data, and then Z-Score standardization is used to make features of different dimensions have the same scale. Then, the Pearson correlation coefficient is used to analyze the correlation between input features.

[0020] Preferably, in S3, the machine learning models include: a random forest model, an extreme gradient boosting model, a lightweight gradient boosting machine model and a multi-layer perceptron model.

[0021] Preferably, in S3, grid search is used to tune the hyperparameters of the model, using R 2 , RMSE and MAE evaluate the model accuracy, the formula is as follows:

[0022]

[0023]

[0024] Among them, y i represents the true value;

[0025] represents the predicted value;

[0026] represents the average of the true values;

[0027] n represents the sample size.

[0028] Preferably, in S4, the optimal model is explained using Shapley added value interpretation and partial dependence diagram.

[0029] Preferably, in S5, the optimal model lightweight gradient boosting machine model is combined with the multi-objective optimization algorithm non-dominated sorting genetic algorithm III to maximize the CO2 conversion rate and CH4 selectivity, and optimization screening is performed under low temperature conditions of ≤250°C to obtain a low-temperature and high-activity CO2 methanation reaction catalyst.

[0030] A machine learning optimization system for low-temperature and high-activity CO2 methanation catalysts, characterized in that the optimization system comprises: a data collection module, a data preprocessing module, a hyperparameter tuning module, an optimal model interpretation module, and a catalyst screening module;

[0031] The data collection module is used to collect CO2 methanation reaction catalyst descriptors, preparation conditions and reaction condition parameters;

[0032] The data preprocessing module is used to preprocess the CO2 methanation reaction data set;

[0033] The hyperparameter tuning module is used to import the preprocessed data set into the machine learning model to perform hyperparameter tuning;

[0034] The best model explanation module is used to select the best model and use model explanation technology to perform interpretability analysis on the best model;

[0035] The catalyst screening module is used to combine a multi-objective optimization algorithm with an optimal model to optimize and screen low-temperature and high-activity CO2 methanation reaction catalysts.

[0036] The present invention provides a machine learning optimization method and system for a low-temperature and high-activity CO2 methanation catalyst. Compared with the prior art, it has the following beneficial effects:

[0037] In the present invention, the optimization method realizes machine learning optimization of CO2 methanation catalyst by collecting CO2 methanation reaction data, building machine learning model hyperparameter tuning, selecting the best model, performing interpretability analysis on the best model, and combining the best model with a multi-objective optimization algorithm to screen catalysts. Machine learning can quickly predict and screen catalysts with high activity and selectivity by analyzing a large amount of data, greatly improving research and development efficiency; machine learning can reduce the number of experiments, optimize experimental design, significantly reduce research and development costs, and improve research and development success rate; machine learning can identify and utilize complex chemical and physical data, and design catalysts with better performance with high precision, thereby improving the conversion rate and selectivity of CO2 to CH4. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 creative work.

[0039] Figure 1 is a flow chart of an optimization method in an embodiment of the present invention;

[0040] Figure 2 This is a table of input and output characteristics describing the CO2 methanation reaction in an embodiment of the present invention;

[0041] Figure 3 A Pearson correlation coefficient graph of input features in an embodiment of the present invention;

[0042] Figure 4 This is a comparison chart of the accuracy of various machine learning models in the embodiments of the present invention;

[0043] Figure 5 It is a Shapley value explanation diagram of the lightweight gradient boosting machine model in an embodiment of the present invention;

[0044] Figure 6 It is a partial dependency explanation diagram of the lightweight gradient boosting machine model in an embodiment of the present invention;

[0045] Figure 7 This is a table of low-temperature, high-activity CO2 methanation reaction catalysts optimized and screened in the embodiments of the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are 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 creative work are within the scope of protection of the present invention.

[0047] The embodiments of the present application solve the problems of low efficiency, high cost and low precision in the research and development of CO2 methanation catalysts by providing a machine learning optimization method and system for low-temperature and high-activity CO2 methanation catalysts.

[0048] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0049] Example:

[0050] like Figure 1 to Figure 7 As shown, the present invention provides a machine learning optimization method for a low-temperature and high-activity CO2 methanation catalyst, and the optimization method comprises the following steps:

[0051] S1. Collecting CO2 methanation reaction catalyst descriptors, preparation conditions and reaction condition parameters;

[0052] Catalyst descriptors include: catalyst composition and content;

[0053] Preparation conditions include: preparation method (PPM), calcination temperature (T C ) Calcination time (t C ), reducing hydrogen content (C H ), reduction temperature (T Rd ), restoration time (t Rd ) and reduction pressure (P Rd );

[0054] The reaction conditions of CO2 methanation include: reaction temperature (T), pressure (P), weight hourly space velocity (WHSV), content of inert components in the raw materials (C Inert ) and H2 / CO2 ratio (H / C);

[0055] Collect the above input features, and the output features used for prediction include: CO2 conversion rate (X CO2 ), CH4 selectivity (S CH4 ) and CH4 yield (Y CH4 ), specific features such as Figure 2 shown.

[0056] S2, preprocess the CO2 methanation reaction data set collected by S1 to facilitate the subsequent machine learning model construction;

[0057] Label encoding is used to convert categorical data into numerical data, and then Z-Score standardization is used to make features of different dimensions have the same scale. Then the Pearson correlation coefficient is used to analyze the correlation between input features, such as Figure 3 shown.

[0058] S3, import the data processed by S2 into multiple machine learning models respectively to perform hyperparameter tuning;

[0059] The machine learning models include: random forest model (RF), extreme gradient boosting model (XGB), lightweight gradient boosting machine model (LGBM) and multi-layer perceptron model (MLP);

[0060] Use grid search to tune model hyperparameters using R 2 , RMSE and MAE evaluate the model accuracy, the formula is as follows:

[0061]

[0062] Among them, y i represents the true value;

[0063] represents the predicted value;

[0064] represents the average of the true values;

[0065] n represents the sample size;

[0066] After adjusting the parameters, the model accuracy is as follows Figure 4 shown.

[0067] S4, select the best model according to the prediction accuracy of the model after S3 tuning, and use model interpretation technology to perform interpretability analysis on the best model;

[0068] according to Figure 4 In the comparison of model prediction performance, the lightweight gradient boosting machine model (LGBM) has the best performance. The Shapley plus interpretation (SHAP) and partial dependence plot (PDP) are used to explain the best model, such as Figure 5 , Figure 6 shown.

[0069] S5. Based on the insights gained from the interpretation of the S4 model, a low-temperature and high-activity CO2 methanation catalyst was screened based on a multi-objective optimization algorithm combined with the best model optimization;

[0070] According to the insights gained from the interpretation of the S4 model, the best model lightweight gradient boosting machine model (LGBM) was combined with the multi-objective optimization algorithm non-dominated sorting genetic algorithm III (NSGA-III) to maximize the CO2 conversion rate and CH4 selectivity. The optimization screening was carried out under low temperature conditions of ≤250℃, and 6 low-temperature and high-activity CO2 methanation reaction catalysts were obtained, such as Figure 7 shown.

[0071] The present invention also provides a machine learning optimization system for a low-temperature and high-activity CO2 methanation catalyst, the optimization system comprising: a data collection module, a data preprocessing module, a hyperparameter tuning module, an optimal model interpretation module and a catalyst screening module;

[0072] The data collection module is used to collect CO2 methanation reaction catalyst descriptors, preparation conditions and reaction condition parameters;

[0073] The data preprocessing module is used to preprocess the CO2 methanation reaction data set;

[0074] The hyperparameter tuning module is used to import the preprocessed data set into the machine learning model to perform hyperparameter tuning;

[0075] The best model explanation module is used to select the best model and use model explanation technology to perform interpretability analysis on the best model;

[0076] The catalyst screening module is used to combine a multi-objective optimization algorithm with an optimal model to optimize and screen low-temperature and high-activity CO2 methanation reaction catalysts.

[0077] In summary, compared with the prior art, the present invention has the following beneficial effects:

[0078] In an embodiment of the present invention, the optimization method realizes machine learning optimization of CO2 methanation catalyst by collecting CO2 methanation reaction data, building a machine learning model for hyperparameter tuning, selecting the best model, performing interpretability analysis on the best model, and combining the best model with a multi-objective optimization algorithm to screen catalysts. Machine learning can quickly predict and screen catalysts with high activity and selectivity by analyzing a large amount of data, thereby greatly improving research and development efficiency; machine learning can reduce the number of experiments, optimize experimental design, significantly reduce research and development costs, and improve research and development success rate; machine learning can identify and utilize complex chemical and physical data, and design catalysts with better performance with high precision, thereby improving the conversion rate and selectivity of CO2 to CH4.

[0079] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0080] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. 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 invention.

Claims

1. A machine learning optimization method for low-temperature and high-activity CO2 methanation catalyst, characterized in that: The optimization method comprises the following steps: S1. Collecting CO2 methanation reaction catalyst descriptors, preparation conditions and reaction condition parameters; S2, preprocessing the CO2 methanation reaction data set collected by S1; S3, import the data processed by S2 into multiple machine learning models respectively to perform hyperparameter tuning; S4, select the best model according to the prediction accuracy of the model after S3 tuning, and use model interpretation technology to perform interpretability analysis on the best model; S5. Based on the insights gained from the interpretation of the S4 model, a low-temperature, high-activity CO2 methanation reaction catalyst was screened out based on a multi-objective optimization algorithm combined with the best model optimization.

2. The machine learning optimization method for the low temperature and high activity CO2 methanation catalyst according to claim 1, characterized in that: In said S1, the catalyst descriptor includes: catalyst composition and content; Preparation conditions include: preparation method, calcination temperature, calcination time, reduction hydrogen content, reduction temperature, reduction time and reduction pressure; The reaction condition parameters include: reaction temperature, pressure, weight hourly space velocity, inert component content in the raw materials and H2 / CO2 ratio.

3. The machine learning optimization method for the low temperature and high activity CO2 methanation catalyst according to claim 2, characterized in that: In S1, the output features predicted by the collected input features include: CO2 conversion rate, CH4 selectivity and CH4 yield.

4. The machine learning optimization method for the low temperature and high activity CO2 methanation catalyst according to claim 1, characterized in that: In S2, the preprocessing of the data set includes the following steps: Label encoding is used to convert categorical data into numerical data, and then Z-Score standardization is used to make features of different dimensions have the same scale. Then, the Pearson correlation coefficient is used to analyze the correlation between input features.

5. The machine learning optimization method for the low temperature and high activity CO2 methanation catalyst according to claim 1, characterized in that: In S3, the machine learning models include: random forest model, extreme gradient boosting model, lightweight gradient boosting machine model and multi-layer perceptron model.

6. The machine learning optimization method for low temperature and high activity CO2 methanation catalyst according to claim 1, characterized in that: In S3, grid search is used to tune the hyperparameters of the model, and R 2 , RMSE and MAE evaluate the model accuracy, the formula is as follows: Among them, y i represents the true value; represents the predicted value; represents the average of the true values; n represents the sample size.

7. The machine learning optimization method for the low temperature and high activity CO2 methanation catalyst according to claim 1, characterized in that: In S4, the best model was explained using Shapley's added value interpretation and partial dependence plots.

8. The machine learning optimization method for low temperature and high activity CO2 methanation catalyst according to claim 1, characterized in that: In S5, the optimal model lightweight gradient boosting machine model is combined with the multi-objective optimization algorithm non-dominated sorting genetic algorithm III to maximize the CO2 conversion rate and CH4 selectivity, and optimization screening is performed under low temperature conditions of ≤250°C to obtain a low-temperature and high-activity CO2 methanation reaction catalyst.

9. A machine learning optimization system for low-temperature and high-activity CO2 methanation catalyst, characterized in that: The optimization system includes: a data collection module, a data preprocessing module, a hyperparameter tuning module, an optimal model interpretation module and a catalyst screening module; The data collection module is used to collect CO2 methanation reaction catalyst descriptors, preparation conditions and reaction condition parameters; The data preprocessing module is used to preprocess the CO2 methanation reaction data set; The hyperparameter tuning module is used to import the preprocessed data set into the machine learning model to perform hyperparameter tuning; The best model explanation module is used to select the best model and use model explanation technology to perform interpretability analysis on the best model; The catalyst screening module is used to combine a multi-objective optimization algorithm with an optimal model to optimize and screen low-temperature and high-activity CO2 methanation reaction catalysts.

Images