Process parameter optimization method, device, and program product

Abstract

The application provides a process parameter optimization method, device and program product. The method comprises the following steps: constructing a yield prediction model according to a pre-acquired experimental data set; performing an explainability analysis on the yield prediction model to obtain key process parameters; establishing a kinetic model and a computational fluid dynamics model according to kinetic parameters of microbial growth, and coupling the kinetic model to the computational fluid dynamics model; generating a plurality of process parameter combinations according to the key process parameters by using a response surface method; performing numerical simulation on the plurality of process parameter combinations by using the computational fluid dynamics model to obtain a plurality of corresponding methane yields; and determining an optimal process parameter combination according to the plurality of process parameter combinations and the plurality of corresponding methane yields. The scientificity and efficiency of process parameter optimization can be improved, a reliable basis can be provided for industrial scaling of a biological methanation trickle bed, and thus the conversion efficiency of the trickle bed can be improved and the methane yield can be improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a method, apparatus and program product for optimizing process parameters. Background Technology

[0002] To achieve long-term storage of electrical energy and recycling of carbon resources, carbon dioxide hydrogenation biomethane technology has attracted widespread attention. This technology electrolyzes surplus electricity to produce hydrogen, which is then used to produce methane (…). ) and carbon dioxide ( ) synthesize renewable methane under the action of microorganisms ( ).

[0003] Currently, traditional trickle beds are commonly used to achieve biomethanation processes. In this reactor, the gas phase ( A mixed gas and liquid nutrient medium flow in parallel or countercurrent through a bed filled with a biofilm carrier. The reaction depends on a methanogenic microbial community attached to the carrier surface, which catalyzes the reaction of carbon dioxide and hydrogen to produce methane at the gas-liquid-solid three-phase interface.

[0004] However, existing trickle beds suffer from a decrease in overall conversion efficiency and methane production as the scale increases, making it difficult to meet the needs of stable and efficient industrial production. Summary of the Invention

[0005] This application provides a method, apparatus, and procedure for optimizing process parameters to improve the conversion efficiency of trickle beds and increase methane production.

[0006] In a first aspect, this application provides a method for optimizing process parameters, including:

[0007] A production prediction model is constructed based on a pre-collected experimental dataset, wherein the input variables of the production prediction model include process parameters that affect methane production, and the output variable includes methane production.

[0008] An interpretability analysis was performed on the production prediction model to obtain key process parameters;

[0009] A kinetic model was established based on the kinetic parameters of microbial growth;

[0010] A computational fluid dynamics model for a biomethanation ectopic trickle bed is established, and the dynamics model is coupled to the computational fluid dynamics model.

[0011] Multiple sets of process parameter combinations are generated based on the key process parameters using the response surface methodology.

[0012] The computational fluid dynamics model is used to numerically simulate the multiple combinations of process parameters to obtain the corresponding multiple methane yields;

[0013] The optimal combination of process parameters is determined based on the multiple combinations of process parameters and the corresponding multiple methane yields.

[0014] In one possible implementation, the step of constructing a yield prediction model based on a pre-collected experimental dataset includes: collecting experimental data of a bio-methanation trickle bed process and constructing an experimental dataset, the experimental dataset including input variables and output variables, the input variables including process parameters affecting methane yield, and the output variables including methane yield; performing outlier removal and missing value imputation on the experimental dataset to obtain a regularized dataset; dividing the regularized dataset into a training set and a test set according to a preset ratio; training a preset machine learning model using the training set to construct an initial yield prediction model; optimizing the initial yield prediction model using a strategy of cross-validation combined with hyperparameter optimization to determine the model parameter combination; and evaluating the performance of the optimized initial yield prediction model using preset performance evaluation metrics based on the test set to determine the yield prediction model.

[0015] In one possible implementation, the step of performing interpretability analysis on the production forecasting model to obtain key process parameters includes: calculating the contribution of the input variables to the output variables of the production forecasting model using a preset interpretability analysis method; and selecting key process parameters from the input variables based on the contribution.

[0016] In one possible implementation, establishing a kinetic model based on the kinetic parameters of microbial growth includes: constructing a system of differential equations based on the Monod equation; solving the system of differential equations using a preset numerical integration method to obtain kinetic parameters, wherein the kinetic parameters include the maximum specific growth rate, half-saturation constant, microbial stem cell concentration, and yield coefficient; fitting the kinetic parameters to pre-collected dissolved hydrogen concentration data to obtain a biological reaction rate expression, wherein the dissolved hydrogen concentration data is collected through headspace equilibrium experiments; and establishing a kinetic model based on the biological reaction rate expression.

[0017] In one possible implementation, establishing a computational fluid dynamics model of the biomethanation ex-situ trickle bed and coupling the dynamics model to the computational fluid dynamics model includes: establishing a computational fluid dynamics model based on preset size data of the biomethanation ex-situ trickle bed; meshing the computational fluid dynamics model to obtain a computational mesh; configuring model parameters and boundary conditions on the computational mesh; writing the dynamics model as a user-defined function; and coupling the user-defined function to the computational fluid dynamics model.

[0018] In one possible implementation, the step of numerically simulating the multiple combinations of process parameters using the computational fluid dynamics model to obtain multiple corresponding methane yields includes: configuring convergence control parameters; using the computational fluid dynamics model, based on the convergence control parameters, performing steady-state coupled solution for the multiple combinations of process parameters, and iteratively processing them using an adaptive time step to obtain multiple corresponding simulation results; and post-processing the multiple simulation results to obtain multiple corresponding methane yields.

[0019] In one possible implementation, determining the optimal process parameter combination based on the multiple sets of process parameter combinations and the corresponding multiple methane production yields includes: performing regression analysis on the multiple sets of process parameter combinations and the corresponding multiple methane production yields using the response surface methodology to obtain a second-order regression prediction model; and solving the second-order regression prediction model with the objective of maximizing methane production yield to obtain the optimal process parameter combination.

[0020] In one possible implementation, after determining the optimal process parameter combination based on the multiple sets of process parameter combinations and the corresponding multiple methane production rates, the method further includes: obtaining the maximum methane production rate corresponding to the optimal process parameter combination in the second-order regression prediction model; inputting the optimal process parameter combination into the computational fluid dynamics model for verification simulation to obtain the verification methane production rate; comparing the verification methane production rate with the maximum methane production rate; and confirming the effectiveness of the optimal process parameter combination if the deviation between the verification methane production rate and the maximum methane production rate is less than a preset deviation threshold.

[0021] Secondly, this application provides a process parameter optimization device, comprising:

[0022] The module is used to build a production prediction model based on a pre-collected experimental dataset, wherein the input variables of the production prediction model include process parameters that affect methane production, and the output variable includes methane production.

[0023] The analysis module is used to perform interpretability analysis on the production prediction model to obtain key process parameters;

[0024] A module is established to create a kinetic model based on the kinetic parameters of microbial growth;

[0025] The establishment module is also used to establish a computational fluid dynamics model of the biomethanation ex-situ trickle bed and couple the dynamics model to the computational fluid dynamics model;

[0026] The generation module is used to generate multiple sets of process parameter combinations based on the key process parameters using the response surface methodology.

[0027] The simulation module is used to perform numerical simulations of the multiple sets of process parameter combinations using the computational fluid dynamics model to obtain the corresponding multiple methane yields;

[0028] The determination module is used to determine the optimal combination of process parameters based on the multiple combinations of process parameters and the corresponding multiple methane yields.

[0029] Thirdly, this application provides a process parameter optimization device, including: a memory and a processor;

[0030] The memory stores computer-executed instructions;

[0031] The processor executes computer execution instructions stored in the memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0032] Fourthly, this application provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible embodiments of the first aspect.

[0033] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0034] The process parameter optimization method, apparatus, and program products provided in this application quickly screen out key process parameters that significantly affect methane production through interpretability analysis of the production prediction model, couple the kinetic model to the computational fluid dynamics model, and perform numerical simulation through the computational fluid dynamics model; combined with the response surface methodology, they quickly locate the optimal combination of process parameters, which can improve the scientific nature and efficiency of process parameter optimization, provide a reliable basis for the industrial scale-up of biomethanation trickle beds, thereby improving the conversion efficiency of trickle beds and increasing methane production. Attached Figure Description

[0035] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0036] Figure 1 A schematic diagram illustrating a scenario for the process parameter optimization method provided in an embodiment of this application;

[0037] Figure 2 A schematic flowchart of a process parameter optimization method provided in one embodiment of this application;

[0038] Figure 3 A schematic diagram of the theoretical framework provided for the embodiments of this application;

[0039] Figure 4 A schematic diagram illustrating the descriptive statistical analysis of the input and output variables of the experimental dataset provided in this application embodiment;

[0040] Figure 5 This is a schematic diagram showing the distribution of the experimental dataset before and after imputation, provided in an embodiment of this application.

[0041] Figure 6 A schematic diagram illustrating the predictive performance of the yield prediction model provided in the embodiments of this application;

[0042] Figure 7 A schematic diagram illustrating the interpretability analysis results provided for embodiments of this application;

[0043] Figure 8 This is a schematic diagram of a partial dependency graph of input variables provided in an embodiment of this application;

[0044] Figure 9 A schematic diagram of the dynamic model curves provided in the embodiments of this application;

[0045] Figure 10 A schematic diagram of a computational fluid dynamics model provided in an embodiment of this application;

[0046] Figure 11 This is a schematic diagram of mesh division provided for an embodiment of this application;

[0047] Figure 12 A schematic diagram of temperature, pressure, and flow rate variation cloud maps and streamline diagrams provided in the computational fluid dynamics model for embodiments of this application;

[0048] Figure 13 Different reaction temperatures provided for embodiments of this application , and A schematic diagram showing the change in volume fraction as a function of inlet distance;

[0049] Figure 14 This is a schematic diagram illustrating the variation of component concentrations with inlet pipe distance under different reaction pressures, as provided in the embodiments of this application.

[0050] Figure 15 Different feed gas loads are provided for the embodiments of this application. and A schematic diagram showing the change in volume fraction with intake manifold distance;

[0051] Figure 16 A schematic diagram illustrating the results of response surface methodology analysis of the optimal combination of process parameters provided in an embodiment of this application;

[0052] Figure 17This is a schematic diagram of simulation results provided for an embodiment of this application;

[0053] Figure 18 This is a schematic diagram of the process parameter optimization device provided in the embodiments of this application;

[0054] Figure 19 This is a schematic diagram of the process parameter optimization equipment provided in the embodiments of this application.

[0055] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concepts of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0056] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0057] The process parameter optimization method provided in this application uses a machine learning model to quickly screen out key process parameters that significantly affect methane production. It couples a kinetic model to a computational fluid dynamics model and uses the computational fluid dynamics model for numerical simulation to analyze the flow field, mass transfer, and bioreaction dynamics within the trickle bed. Combined with response surface methodology, it quickly locates the optimal combination of process parameters, which can improve the scientific nature and efficiency of process parameter optimization. This provides a reliable basis for the industrial scale-up of biomethanation trickle beds, thereby improving the conversion efficiency of the trickle bed and increasing methane production.

[0058] Figure 1 This is a schematic diagram illustrating a scenario of the process parameter optimization method provided in the embodiments of this application, such as... Figure 1 As shown, the scene is a computer device, including: a receiving device 101, a processor 102 and a display device 103.

[0059] It is understood that the structures illustrated in the embodiments of this application do not constitute a specific limitation on the process parameter optimization method. In other feasible embodiments of this application, the above architecture may include more or fewer components than illustrated, or combine some components, or split some components, or arrange different components, which can be determined according to the actual application scenario and is not limited here. Figure 1 The components shown can be implemented in hardware, software, or a combination of both.

[0060] In the specific implementation process, the receiving device 101 can be an input / output interface or a communication interface, which can acquire the pre-collected experimental dataset.

[0061] The processor 102 can process the pre-collected experimental dataset to determine the optimal combination of process parameters.

[0062] The display device 103 can be used to display the above-mentioned optimal process parameter combination, etc.

[0063] The display device can also be a touch screen, used to receive user commands while displaying the above content, so as to achieve interaction with the user.

[0064] It should be understood that the aforementioned processor can be implemented by reading instructions from memory and executing those instructions, or it can be implemented through chip circuitry.

[0065] Furthermore, the network architecture and business scenarios described in the embodiments of this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided in the embodiments of this application. As those skilled in the art will know, with the evolution of network architecture and the emergence of new business scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0066] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0067] Figure 2 This is a flowchart illustrating a process parameter optimization method provided in one embodiment of this application. The executing entity of this embodiment can be... Figure 1 The computer equipment shown is not specifically limited in this embodiment. Figure 2 As shown, the method includes:

[0068] S201: Construct a production prediction model based on the pre-collected experimental dataset. The input variables of the production prediction model include process parameters that affect methane production, and the output variable includes methane production.

[0069] Optionally, methane production may include volumetric methane yield. (L / L / d) and methane content (%).

[0070] Optionally, a yield prediction model is constructed based on a pre-collected experimental dataset, including: collecting experimental data of the bio-methanation trickle bed process and constructing an experimental dataset, which includes input variables and output variables. The input variables include process parameters affecting methane yield, and the output variables include methane yield. Outlier removal and missing value imputation are performed on the experimental dataset to obtain a regularized dataset. The regularized dataset is divided into a training set and a test set according to a preset ratio. A preset machine learning model is trained using the training set to construct an initial yield prediction model. The initial yield prediction model is optimized using a strategy combining cross-validation and hyperparameter optimization to determine the model parameter combination. A preset performance evaluation metric is used to evaluate the performance of the optimized initial yield prediction model based on the test set to determine the yield prediction model.

[0071] Optionally, process parameters affecting methane production include: substrate type, total solids (TS) (%), volatile solids (VS) (%), volume (L), height-to-diameter ratio, feed gas loading rate (L / L / d), and packing specific surface area (m²). 2 / m 3 The parameters include: hydrogen ion concentration index (pH), temperature (°C), pressure (MPa), operating time (d), H2 / CO2 ratio, gas residence time (h), and spray rate (L / L / d).

[0072] By constructing an experimental dataset containing process parameters affecting methane production and training a production prediction model using a machine learning model, a comprehensive coverage of biomethanation process parameters and accurate capture of nonlinear relationships were achieved. This provides a reliable data source for the subsequent screening of key process parameters, improves the scientific rigor and efficiency of process parameter optimization, and provides a reliable basis for the industrial scale-up of biomethanation trickle beds, thereby improving the conversion efficiency of trickle beds and increasing methane production.

[0073] Optionally, in another embodiment of this application, a yield prediction model is constructed based on a pre-collected experimental dataset, including: collecting multiple experimental data of the ex-situ hydrogenation bio-methanation trickle bed process using a data retrieval algorithm to establish an experimental dataset with volumetric methane yield and methane content as output variables and process parameters of methane production as input variables. Outliers are removed from the experimental dataset; a missing value imputation system is built using an iterative imputation method, and a random forest imputation model is used to iteratively fit and impute the missing data in the experimental dataset to obtain a regularized dataset. The regularized dataset is randomly divided into a training set and a test set according to a preset ratio; an initial yield prediction model is constructed based on the gradient boosting tree algorithm, and the initial yield prediction model is globally optimized by combining cross-validation and hyperparameter optimization; the performance of the optimized initial yield prediction model is evaluated based on the coefficient of determination and root mean square error to determine the yield prediction model.

[0074] Optionally, the preset ratio can be 7:3.

[0075] Optionally, the training set is used to build the yield prediction model, and the test set is used to validate the performance of the yield prediction model.

[0076] Optionally, an initial yield prediction model is constructed based on the gradient boosting tree algorithm, and the initial yield prediction model is globally optimized by combining cross-validation and hyperparameter optimization, including: using the gradient boosting tree algorithm, using an artificial neural network (ANN) to optimize hyperparameters, and using a five-fold cross-validation method to prevent overfitting.

[0077] S202: Perform interpretability analysis on the production forecasting model to obtain key process parameters.

[0078] Optionally, interpretability analysis is performed on the production forecasting model to obtain key process parameters, including: calculating the contribution of the input variables of the production forecasting model to the output variables using a preset interpretability analysis method; and selecting key process parameters from the input variables based on the contribution.

[0079] By using interpretability analysis to accurately quantify the contribution of input variables, key process parameters are screened out, improving the scientific rigor and efficiency of parameter selection. This provides a reliable basis for the industrial scale-up of bio-methanation trickle beds, thereby improving the conversion efficiency of the trickle bed and increasing methane production.

[0080] Optionally, interpretability analysis can be performed on the production forecasting model to obtain key process parameters, including: using the Shapley Additive Explanations (SHAP) method to perform interpretability analysis on the production forecasting model to determine key process parameters.

[0081] S203: Establish a kinetic model based on the kinetic parameters of microbial growth.

[0082] Optionally, a kinetic model is established based on the kinetic parameters of microbial growth, including: constructing a system of differential equations based on the Monod equation; solving the system of differential equations using a pre-defined numerical integration method to obtain kinetic parameters, which include the maximum specific growth rate, half-saturation constant, microbial stem cell concentration, and yield coefficient; fitting the kinetic parameters to pre-collected dissolved hydrogen concentration data to obtain a biological reaction rate expression, wherein the dissolved hydrogen concentration data is collected through headspace equilibrium experiments; and establishing a kinetic model based on the biological reaction rate expression.

[0083] Solving the differential equations by numerical integration and fitting them with dissolved hydrogen concentration data to determine the kinetic model can improve the physical reliability of the kinetic model and ensure the accuracy of subsequent numerical simulations coupled to the computational fluid dynamics model, thereby improving the conversion efficiency of the trickle bed and increasing methane production.

[0084] Optionally, a kinetic model is established based on the kinetic parameters of microbial growth, including: constructing a system of differential equations based on the Monod equation; solving the system of differential equations using the explicit Runge-Kutta formula and the adaptive step-size method to obtain the kinetic parameters. Dissolved hydrogen concentration data from headspace equilibrium experiments are collected, and the kinetic parameters are fitted using the dissolved hydrogen concentration data to obtain the biological reaction rate expression; a kinetic model is then established based on the biological reaction rate expression.

[0085] It should be noted that the headspace balance experiment was conducted at 37°C using the headspace balance method, with a H2 / CO2 mixture (volume ratio 4:1) as the gas phase and filtered biogas slurry as the liquid phase. Online sampling was performed and the dissolved hydrogen concentration was measured using gas chromatography.

[0086] S204: Establish a computational fluid dynamics model for the bio-methanation ex-situ trickle bed and couple the dynamics model to the computational fluid dynamics model.

[0087] Optionally, a computational fluid dynamics model of the bio-methanation ex-situ trickle bed is established, and the dynamics model is coupled to the computational fluid dynamics model, including: establishing a computational fluid dynamics model based on the preset size data of the bio-methanation ex-situ trickle bed; meshing the computational fluid dynamics model to obtain a computational mesh; configuring model parameters and boundary conditions on the computational mesh; writing the dynamics model as a user-defined function; and coupling the user-defined function to the computational fluid dynamics model.

[0088] By configuring model parameters and boundary conditions on the computational grid, the physical benchmark of the numerical simulation is anchored, effectively reducing the deviation between the simulation results and the actual biomethanation process. By coupling a custom function to the computational fluid dynamics model, the flow field mass transfer process and the microbial metabolic reaction process within the reactor are integrated, overcoming the limitations of single flow field simulation or single reaction analysis. This allows for the quantification of the intrinsic relationship between mass transfer efficiency and methane yield, improving the accuracy of the numerical simulation, thereby enhancing the scientific rigor and efficiency of parameter selection. This provides a reliable basis for the industrial scale-up of biomethanation trickle beds, ultimately improving the conversion efficiency of the trickle bed and increasing methane production.

[0089] Optionally, a computational fluid dynamics model of the bio-methanation ex-situ trickle bed is established, including: using numerical simulation geometric modeling methods to establish a computational fluid dynamics model based on the preset size data of the bio-methanation ex-situ trickle bed, wherein the computational fluid dynamics model includes a bottom air inlet channel and a top product gas outlet.

[0090] Optionally, coupling the dynamic model to the computational fluid dynamics model includes: using a mesh generation method to mesh the computational fluid dynamics model according to a preset meshing strategy to obtain a computational mesh, wherein the preset meshing strategy includes using a hexahedral mesh and refining it at irregular interfaces; configuring model parameters and boundary conditions on the computational mesh, configuring the incompressible water-air system and laminar flow model, configuring the boundary type of the velocity inlet-pressure outlet, and configuring the hydrogen mole fraction and carbon dioxide mole fraction at the inlet; writing the dynamic model as a user-defined function (UDF) based on preset custom macros; compiling the user-defined function and loading it into the computational fluid dynamics solver corresponding to the computational fluid dynamics model to achieve coupling between the user-defined function and the computational fluid dynamics model.

[0091] Optionally, the molar fraction of hydrogen can be 0.8; the molar fraction of carbon dioxide can be 0.2.

[0092] S205: The response surface methodology is used to generate multiple combinations of process parameters based on key process parameters.

[0093] Optionally, the response surface methodology is used to generate multiple combinations of process parameters based on key process parameters, including: using a response surface model, with key process parameters as influencing factors, to construct a three-factor, three-level experimental scheme, wherein the experimental scheme includes multiple combinations of process parameters.

[0094] Optionally, the key process parameters can be reaction temperature, reaction pressure, and feed gas load. The reaction temperature can range from 37 to 55°C, the reaction pressure can range from 0.1 to 1.1 MPa, and the feed gas load can range from 5 to 65 L / L / d.

[0095] S206: Numerical simulation of multiple combinations of process parameters is performed using a computational fluid dynamics model to obtain the corresponding methane production rates.

[0096] Optionally, multiple sets of process parameter combinations are numerically simulated using a computational fluid dynamics model to obtain multiple corresponding methane yields, including: configuring convergence control parameters; using the computational fluid dynamics model, based on the convergence control parameters, performing steady-state coupling solutions on multiple sets of process parameter combinations, and iterating through an adaptive time step to obtain multiple corresponding simulation results; and post-processing the multiple simulation results to obtain multiple corresponding methane yields.

[0097] Optionally, convergence control parameters include relaxation factors, residual thresholds, and local initial conditions.

[0098] Optionally, a steady-state coupled solution is performed on multiple sets of process parameter combinations, and iterative processing is carried out using an adaptive time step to obtain multiple corresponding simulation results. This includes: using a steady-state coupled solution and iterating with an adaptive time step until the residual of each phase is less than or equal to a preset residual threshold to obtain multiple corresponding simulation results.

[0099] Optionally, the preset residual threshold can be 10. -5 .

[0100] Optionally, multiple simulation results are post-processed to obtain multiple corresponding methane yields, including: extracting and processing multiple simulation results to obtain velocity, pressure contour maps and axial velocity curves, statistically analyzing the methane, hydrogen and carbon dioxide concentrations at the top product gas outlet, and calculating the hydrogen conversion rate and methane yield to obtain the corresponding methane yield.

[0101] S207: Determine the optimal combination of process parameters based on multiple combinations of process parameters and corresponding methane yields.

[0102] Optionally, the optimal process parameter combination is determined based on multiple combinations of process parameters and corresponding multiple methane production rates, including: using the response surface methodology to perform regression analysis on multiple combinations of process parameters and corresponding multiple methane production rates to obtain a second-order regression prediction model; and solving the second-order regression prediction model with the objective of maximizing methane production to obtain the optimal process parameter combination.

[0103] By analyzing the parameter interaction effects using a second-order regression model, the optimal combination of process parameters can be determined, significantly improving reactor performance and providing a clear optimization direction for industrial scale-up.

[0104] The process parameter optimization method provided in this application quickly identifies key process parameters that significantly affect methane production through interpretability analysis of the production prediction model. It couples the kinetic model to a computational fluid dynamics model and performs numerical simulations using the computational fluid dynamics model. Combined with the response surface methodology, it quickly locates the optimal combination of process parameters, thereby improving the scientific rigor and efficiency of process parameter optimization. This provides a reliable basis for the industrial scale-up of biomethanation trickle beds, thereby increasing the conversion efficiency of the trickle bed and improving methane production.

[0105] In one embodiment of this application, based on the above embodiment, after step S207, a parameter validity verification process is further included, which is detailed below:

[0106] Obtain the maximum methane production corresponding to the optimal combination of process parameters in the second-order regression prediction model. Input the optimal combination of process parameters into the computational fluid dynamics model for verification simulation to obtain the verification methane production. Compare the verification methane production with the maximum methane production. If the deviation between the verification methane production and the maximum methane production is less than a preset deviation threshold, the optimal combination of process parameters is confirmed to be effective.

[0107] Optionally, the preset deviation threshold can be 5%.

[0108] The process parameter optimization method provided in this application improves the reliability and accuracy of parameter optimization and increases the efficiency of parameter verification by verifying the optimal parameters by substituting them back into the computational fluid dynamics model, thus forming a closed-loop "data-mechanism" verification. This provides a reliable basis for the industrial scale-up of biomethanation trickle beds, thereby improving the conversion efficiency of the trickle bed and increasing methane production.

[0109] This application's embodiments illustrate a process parameter optimization through specific examples. Figure 3 A schematic diagram of the theoretical framework provided for the embodiments of this application, such as Figure 3 As shown, the specific steps of this method are as follows:

[0110] Step 1: Use machine learning methods to screen key process parameters. The reference indicators for screening are volumetric methane yield and methane content.

[0111] Optionally, step one specifically includes the following steps:

[0112] A search using the keywords "biomethanation ex-situ trickle bed" yielded multiple experimental data articles on ex-situ hydrogenation biomethanation trickle bed processes. An experimental dataset was constructed based on this data, categorized by volumetric methane yield. The output variables are (L / L / d) and methane content (%), where methane content can be either average methane content or maximum methane production; the output variables are substrate type, substrate TS (%), substrate VS (%), reactor volume (L), height-to-diameter ratio, feed gas loading (L / L / d), and packing specific surface area (m²). 2 / m 3 The input variables are pH, temperature (°C), pressure (MPa), running time (d), H2 / CO2 ratio, gas residence time (h), and spray rate (L / L / d). Figure 4 This diagram illustrates the descriptive statistical analysis of the input and output variables of the experimental dataset provided in this embodiment of the application. For example... Figure 4 As shown, the input variables, including reactor volume, height-to-diameter ratio, packing specific surface area, temperature, pressure, operating time, H2 / CO2 ratio, gas residence time, and spray rate, are presented in a descriptive statistical analysis, including volumetric methane yield and maximum methane content. Figure 4 The vertical axis of each bar represents the numerical range (%) of each input and output parameter.

[0113] Some input variables in the experimental dataset have missing values. A missing value imputation system was built using the Iterative Imputer class from the sklearn library. A random forest imputation model was selected to iteratively fit and impute the missing data, resulting in a regularized dataset. This improved the accuracy of missing value imputation and the prediction accuracy of the subsequent yield prediction model. Figure 5 This is a schematic diagram illustrating the distribution of the experimental dataset before and after imputation provided in the embodiments of this application, as shown below. Figure 5 As shown, the distribution of sample numbers before (yellow portion) and after (purple portion) missing value imputation is illustrated. Each bar represents a variable, and the vertical axis of the bar represents the number of samples. The input variables for the experimental dataset include substrate TS, substrate VS, reactor volume, height-to-diameter ratio, feed gas loading, packing specific surface area, pH, reaction temperature, reaction pressure, run time, H2 / CO2 ratio, gas retention time (GRT), spray rate, and substrate type (anaerobic wastewater, hydrogen nutrients, digestate, anaerobic sludge, and cow manure, etc.). The output variable includes volumetric methane yield. Preprocessing the experimental dataset significantly improves the quality and reliability of subsequent machine learning decisions.

[0114] The regularized dataset was randomly divided into training and test sets in a 7:3 ratio. The training set was used specifically for model building, while the test set was used to validate model performance. A gradient boosting tree algorithm was selected, and an artificial neural network was used for hyperparameter tuning to achieve global optimization of model performance. To ensure model reliability, a grid search technique and a five-fold cross-validation method were combined to effectively prevent overfitting. Model evaluation used two metrics: the coefficient of determination (R²) and the root mean square error (RMSE). Through these steps, a yield prediction model was obtained, which can be a stacked ensemble. Figure 6 This is a schematic diagram illustrating the predictive performance of the yield prediction model provided in the embodiments of this application, as shown below. Figure 6 As shown, a scatter plot comparing actual and predicted volumetric methane yields is presented, with the horizontal axis representing actual volumetric methane yield (L / L / d) and the vertical axis representing predicted volumetric methane yield (L / L / d). From Figure 6 It can be seen that the yield prediction model's predictions for both the training set and the test set are within a very small error range, R0 2 The accuracy was 0.91, the RMSE was 1.23, and the mean square error (MSE) was 1.52. Furthermore, the training set prediction line and the test set prediction line showed good overlap. This indicates that the established yield prediction model has high prediction accuracy and generalization ability.

[0115] SHAP was used to perform interpretability analysis on the yield forecasting model. Figure 7 This is a schematic diagram illustrating the interpretability analysis results provided for embodiments of this application, such as... Figure 7 As shown, the horizontal axis represents the SHAP value, indicating the contribution of each input variable to the output variable. The color bars on the right represent the values ​​of the input variables themselves, with colors from top to bottom indicating values ​​from highest to lowest. Input variables include: TS, VS, reactor volume, height-to-diameter ratio, feed gas load, packing specific surface area, pH, reaction temperature, reaction pressure, running time, H2 / CO2 ratio, GRT, spray rate, and anaerobic wastewater, hydrogen nutrients, digestate, anaerobic sludge, and cow manure. Figure 7 It is evident that the feed gas loading rate is the parameter with the highest characteristic value. The feed gas loading rate is the amount of gas that can be processed per liter of reactor per day; theoretically, the higher the loading rate, the more CH4 is generated. Next are substrate type, reactor temperature, and pressure. Based on these factors, temperature, pressure, and feed gas loading rate are identified as the key process parameters.

[0116] Univariate partial dependency plots (PDPs) can visually display the impact of a single input variable on the model's prediction results, such as... Figure 8 As shown. Figure 8This is a schematic diagram of a partial dependency graph of input variables provided in an embodiment of this application. Figure 8 In Figure (a), the relationship between feed gas load and volumetric methane yield is shown, with the x-axis representing feed gas load and the y-axis representing volumetric methane yield (L / L / d). Figure (b) shows the relationship between reaction temperature and volumetric methane yield, with the x-axis representing reaction temperature and the y-axis representing volumetric methane yield (L / L / d). Figure (c) shows the relationship between reaction pressure and volumetric methane yield, with the x-axis representing reaction pressure and the y-axis representing volumetric methane yield (L / L / d). Figure 8 It can be seen that a larger feed gas load, a reaction temperature greater than 50℃, and a reaction pressure greater than 0.8MPa are conducive to increasing the volumetric methane yield.

[0117] Step 2: Establish a kinetic model based on the kinetic parameters of the biomethanation reaction. The kinetic parameters are determined using a microbial growth model.

[0118] It should be noted that the first step in the biomethanation reaction is the dissolution of gaseous H2 and CO2. Subsequently, microorganisms use specialized enzymes to convert the dissolved H2 and CO2 into CH4 and H2O. The low partition coefficient of H2 in aqueous solution limits its solubility; therefore, the main limitations on reaction kinetics come from H2 uptake and interfacial mass transfer. Headspace equilibrium experiments were conducted to determine the dissolved hydrogen concentration data. The specific experimental steps are listed below:

[0119] A: A 250mL fermentation flask was used for the experiment, containing 100mL of working volume biogas slurry. The anaerobic environment was maintained by purging with 99.99% nitrogen for 5 minutes. After the initial purging, a mixed gas (H2: 80%, CO2: 20%) was introduced from the top of the fermentation flask and purged for 5 minutes to ensure that the flask was filled with the mixed gas.

[0120] B: Determine the saturation concentration of the gas in a 250 mL serum bottle containing 100 mL of working volume distilled water and 150 mL of mixed gas headspace. Incubate the bottle overnight at 37°C. Then, take two 5 mL aliquots of the liquid sample and store them in a 10 mL sealed serum bottle.

[0121] C: Take 5 mL of liquid sample every 3 minutes after the start of the experiment, avoiding collecting air bubbles, and store it in a 10 mL sealed serum bottle.

[0122] D: Heat a 10 mL serum bottle to 100°C for 30 minutes, then measure the headspace components using gas chromatography.

[0123] Optionally, step two specifically includes the following steps:

[0124] Dissolved hydrogen concentration data were obtained from the headspace components in the headspace equilibrium experiment described above.

[0125] A system of differential equations was constructed based on the reaction kinetics of the Monod equations to simulate methane formation. The system of differential equations can be hyperbolic and can be used to estimate microbial growth and activity based on nutrient supply.

[0126] Alternatively, the formula for the system of differential equations is:

[0127]

[0128] In the formula, K represents the concentration of hydrogen in the liquid phase. S k represents the half-saturation constant. m Indicates the specific absorption rate relative to the substrate; t represents the concentration of microbial stem cells; t represents time.

[0129] Optionally, the formula for calculating the maximum absorption rate of microorganisms is:

[0130]

[0131] In the formula, Indicates the maximum growth rate of microorganisms; This indicates the biomass yield of carbon dioxide.

[0132] Numerical integration of the system of differential equations yields ordinary differential equations (ODEs). The ODEs are then solved using the explicit Runge-Kutta formula and an adaptive step-size method to obtain kinetic parameters, including the maximum specific growth rate, half-saturation constant, microbial stem cell concentration, and yield coefficient. The yield coefficient includes the biomass yield of carbon dioxide. ), biomass yield of hydrogen ( ) and the cellular yield of hydrogen ( ).

[0133] A nonlinear regression method was used to fit and optimize the kinetic parameters based on dissolved hydrogen concentration data. The maximum growth rate of microorganisms, the half-saturation constant, and the biomass yield of hydrogen were constrained to values ​​greater than 0 to avoid obtaining spurious solutions in the model. The correlation coefficient (R²) was used... 2 The goodness of fit between dissolved hydrogen concentration data and kinetic parameters was evaluated, and the final values ​​of kinetic parameters were obtained. Based on the values ​​of kinetic parameters and the system of differential equations, the expression for the biological reaction rate was obtained. A kinetic model was established based on the expression for the biological reaction rate. Figure 9 The schematic diagram of the kinetic model curve provided in the embodiments of this application is used to show the change of hydrogen (H2) concentration over time and to compare the model prediction value with the experimental data value. Figure 9 The x-axis represents time (h), and the y-axis represents H2 concentration (μmol / L). From Figure 9It can be seen that the kinetic model's prediction data on H2 concentration changes are in high agreement with the experimental data, indicating that the model can accurately describe the hydrogen consumption kinetics in this process and can be used for subsequent process simulation and optimization.

[0134] Optionally, in the numerical values ​​of dynamic parameters =0.5h -1 ; =2.123 mg stem cells / mmolH2; =15.234 mg / L; =0.25 mg stem cells / μmol H2; Ks = 1.286 μM.

[0135] Alternatively, the formula for the dynamic model is:

[0136]

[0137] Steps one and two provide data support, which can provide key process parameters and kinetic support for step three.

[0138] Step 3: Establish a computational fluid dynamics model and optimize key process parameters. Optimization is performed through dual-scale analysis, which includes flow field analysis and analysis of hydrogen conversion rate and methane yield.

[0139] Response surface methodology was used to design experiments for the combination of three parameters in biomethanation: reaction temperature, reaction pressure, and feed gas loading, to obtain multiple sets of process parameter combinations. The reaction temperature was set between 37-55℃; the reaction pressure between 0.101325-1.101325 MPa; and the feed gas loading between 5-65 L / L / d. The actual and coded values ​​of the experimental variables are shown in Table 1.

[0140] Table 1

[0141]

[0142] Based on the three-factor, three-level experimental design method, a total of 17 experimental schemes were arranged. Among them, No. 1-12 were experimental groups with different parameter combinations, and No. 13-17 were repeated verification experiments at the center point. The specific experimental schemes of the response surface methodology experimental design are detailed in Table 2.

[0143] Table 2

[0144]

[0145] Figure 10 This is a schematic diagram of a computational fluid dynamics model provided in an embodiment of this application. Figure 10As shown, the main body of the computational fluid dynamics model is a cylindrical structure, consisting of a gas inlet 1001, a packing zone 1002, a gas outlet 1003, a liquid inlet 1004, and a liquid outlet 1005. The inner diameter of the main body of the computational fluid dynamics model is 39.69 mm, and the height is 316.67 mm. It has an internal air inlet channel with a diameter of 50 mm and a height of 50 mm, with the bottom of the channel 55 mm from the bottom surface of the computational fluid dynamics model.

[0146] Mesh the computational fluid dynamics model. Figure 11 This is a schematic diagram of mesh division provided in an embodiment of this application, such as... Figure 11 As shown, the partitioning method uses a hexahedral mesh, which is densified at irregular interfaces. Figure 11 (a) shows the overall grid division diagram, and (b) shows the grid division diagram at the irregular interface. A total of 505,166 grids are present, with a grid orthogonality quality of 0.53.

[0147] A non-compressible water-air system was used for fluid dynamics studies. In this system, the inlet is achieved through velocity inlet boundary conditions, and the outlet on the top surface region is a pressure outlet. Multiphase flow was configured with anti-slip walls and symmetrical boundaries. The simulation of biomethanation involves mass transfer, and its inlet molar composition was configured as xH2=0.8 and xCO2=0.2. The dynamic model was written as a custom function; this custom function was coupled to the computational fluid dynamics model. The physical properties and simulation parameters of the computational fluid dynamics model are shown in Table 3. In Table 3, the Ergun coefficient is the Ergun coefficient.

[0148] Table 3

[0149]

[0150] Transient simulations were employed, solved using a high-resolution scheme. The Navier-Stocks equations were solved using the SIMPLE algorithm within the split-method approach, with a root mean square residual convergence criterion of 1.0E⁻⁵ and an energy residual convergence criterion of 1.0E⁻⁶. A custom model was initialized with inlet molar compositions of xH₂ = 0.8 and xCO₂ = 0.2. After initialization, a step size of 0.05 s and a step count of 1000 were used for iterative solving. Simulation results were obtained after the residual convergence condition was met.

[0151] Step 4: Obtain the simulation results from Step 3, determine the optimal combination of process parameters and conduct effective verification, specifically including response surface methodology analysis and computational fluid dynamics model verification simulation.

[0152] Optionally, the simulation results of the computational fluid dynamics model are used to establish a second-order regression model using the response surface methodology. The second-order regression model is then analyzed to determine the optimal combination of process parameters, which are then substituted back into the computational fluid dynamics model for verification simulation.

[0153] Optionally, step four includes the following steps:

[0154] The simulation results are then visualized. A cross-section is created along the central axis of the computational fluid dynamics model, and the temperature, pressure, and flow rate variations, as well as streamline diagrams, are observed below this cross-section. Figure 12 These are schematic diagrams of temperature, pressure, and flow rate variations, as well as streamline diagrams, provided for embodiments of this application within a computational fluid dynamics model. Figure 12 As shown. Figure 12 (a) and (d) show the velocity distribution of the gas within the computational fluid dynamics model. The color of the bars represents the magnitude of the velocity (m / s). It can be observed that the gas exhibits a distinct spiral flow characteristic within the computational fluid dynamics model, with higher velocities in the top region. Furthermore, the contour plot shows an uneven velocity distribution at the inlet, with higher velocities near the top outlet center region, which is consistent with the phenomenon observed in the streamline plot. Figure 12 (b) and (e) show the temperature analysis within the computational fluid dynamics model. The color of the bar represents the temperature (K). It can be observed that the temperature is generally uniform within the computational fluid dynamics model, but the temperature is slightly higher at the inlet due to the faster fluid flow. Figure 12 (c) and (f) show the pressure distribution inside the computational fluid dynamics model, with the color of the bars representing the static pressure (Pa). The pressure is higher at the bottom of the computational fluid dynamics model and gradually decreases as the fluid rises, which is consistent with the flow characteristics of the fluid within the model.

[0155] Next, a computational fluid dynamics model is created with the centerline of the vertical section. Data on the H2 and CO2 conversion rates and the CH4 yield as a function of axial distance are extracted from the centerline of the vertical section. Figure 13 This is a schematic diagram showing the changes in the volume fractions of H2, CO2, and CH4 with the distance from the inlet at different reaction temperatures, as provided in the embodiments of this application. Figure 13The different colored curves represent different reaction temperatures: yellow (37℃), purple (46℃), and blue (55℃). In (a), the x-axis represents the horizontal distance (m), and the y-axis represents the H2 volume fraction; in (b), the x-axis represents the horizontal distance (m), and the y-axis represents the CO2 volume fraction; and in (c), the x-axis represents the horizontal distance (m), and the y-axis represents the CH4 volume fraction. The results show that the volume fractions of H2 and CO2 decreased most rapidly at 55℃, indicating that high temperature accelerated the consumption of H2 and CO2. The volume fraction of CH4 increased with increasing horizontal distance at all temperatures, with the fastest increase at 55℃, ultimately reaching 98.61%, indicating that high temperature significantly promoted CH4 formation.

[0156] Figure 14 This is a schematic diagram illustrating the variation of component concentrations with inlet pipe distance under different reaction pressures, as provided in the embodiments of this application. Figure 14 The different colored curves represent different reaction pressures: purple (0.101325 MPa), yellow (0.601325 MPa), and blue (1.101325 MPa). In (a), the x-axis represents the horizontal distance (m), and the y-axis represents the H2 volume fraction; in (b), the x-axis represents the horizontal distance (m), and the y-axis represents the CO2 volume fraction; and in (c), the x-axis represents the horizontal distance (m), and the y-axis represents the CH4 volume fraction. Figure 14 It can be seen that as the pressure increases from 0.101325 MPa to 1.101325 MPa, the volume fractions of H2 and CO2 decrease with the increase of lateral distance, while the volume fraction of CH4 increases accordingly. This indicates that the increase in pressure helps to improve the driving force of gas-liquid mass transfer, and the consumption rate of H2 and CO2 increases significantly, thereby enhancing the generation of CH4.

[0157] Figure 15 This is a schematic diagram illustrating the variation of H2 and CH4 volume fractions with inlet pipe distance under different feed gas loads, as provided in the embodiments of this application. Figure 15 The different colored curves represent different feed gas loads: purple (5 L / L / d), yellow (35 L / L / d), and blue (65 L / L / d). In (a), the x-axis represents the horizontal distance (m), and the y-axis represents the H2 volume fraction. In (b), the x-axis represents the horizontal distance (m), and the y-axis represents the CH4 volume fraction. Figure 15 It can be seen that increasing the feed gas load from 5 L / L / d to 65 L / L / d accelerates the H2 consumption rate, resulting in a faster increase in the CH4 volume fraction and an 18% increase in CH4 yield. This confirms that increasing the load can scale up the throughput while maintaining the conversion rate, significantly optimizing the reactor's methane production capacity per unit volume.

[0158] Response surface methodology was used to perform regression analysis on the simulation results to obtain a second-order regression prediction model. Analysis of variance was performed on the second-order regression prediction model, and the results are shown in Table 4. In Table 4, the F-value is the ratio of parameter influence to experimental error, and the P-value is the probability value.

[0159] Table 4

[0160]

[0161] The second-order regression prediction model had a significance level of 0.0151 < 0.05, indicating that the model was statistically significant. The significance levels for reaction temperature (A) and reaction pressure (B) were 0.0325 and 0.0341, respectively, indicating that reaction temperature and pressure significantly affected the volume fraction of CH4. The significance level for feed gas loading (C) was 0.1210, indicating that feed gas loading had no significant effect on the volume fraction of CH4. Based on the above analysis, the effects of the selected reaction temperature, reaction pressure, and feed gas loading on the volume fraction of CH4 are not completely linear, exhibiting quadratic and interaction effects.

[0162] The optimal combination of variables for methane yield was determined by comprehensive analysis using response surface methodology. Figure 16 This is a schematic diagram showing the results of response surface methodology analysis of the optimal combination of process parameters provided in an embodiment of this application. Figure 16 In the equation, A: reaction temperature (°C) = 55; B: reaction pressure (Pa) = 683859; C: feed gas load (L / L / d) = 37; D: methane volume fraction (%) = 99.2505. Figure 16 It can be seen that the optimal combination of process parameters is a reaction temperature of 55℃, a reaction pressure of 684kPa, and a feed gas load of 37L / L / d. Under these conditions, the consumption rate of H2 and CO2 is the fastest, the gas-liquid mixing efficiency is the highest, and the predicted CH4 volume fraction is 99.3%.

[0163] The optimal process parameters were combined and substituted back into the computational fluid dynamics model for verification, and the simulation results were obtained. Figure 17 This is a schematic diagram of the simulation results provided for the embodiments of this application, such as... Figure 17 As shown, (a) is the mole fraction distribution of hydrogen; (b) is the mole fraction distribution of carbon dioxide; (c) is the mole fraction distribution of methane; and (d) is the mole fraction distribution of water. The simulation yielded a CH4 volume fraction of 99.7%, which deviates from the predicted value by <0.4%, meeting the preset <5% criterion, thus confirming the effectiveness and repeatability of the optimized parameters.

[0164] The process parameter optimization method provided in this application quickly identifies key process parameters that significantly affect methane production through interpretability analysis of the yield prediction model, couples the kinetic model to a computational fluid dynamics model, and performs numerical simulations using the computational fluid dynamics model. By combining the response surface methodology to quickly locate the optimal combination of process parameters, the scientific rigor and efficiency of process parameter optimization are improved, thereby enhancing the overall efficiency and performance of the biomethanation process. This provides new ideas and solutions for the application of biochemical engineering and reactor optimization in the field of carbon dioxide hydrogenation to methane.

[0165] Figure 18 This is a schematic diagram of the process parameter optimization device provided in the embodiments of this application, as shown below. Figure 18 As shown, the process parameter optimization device provided in this embodiment includes: a construction module 1801, an analysis module 1802, a creation module 1803, a generation module 1804, a simulation module 1805, and a determination module 1806.

[0166] Module 1801 is used to build a production prediction model based on a pre-collected experimental dataset. The input variables of the production prediction model include process parameters that affect methane production, and the output variable includes methane production.

[0167] Analysis module 1802 is used to perform interpretability analysis on the production forecasting model in order to obtain key process parameters;

[0168] Module 1803 is established to create a kinetic model based on the kinetic parameters of microbial growth.

[0169] Module 1803 is also used to establish a computational fluid dynamics model for the bio-methanation ectopic trickle bed and to couple the dynamics model to the computational fluid dynamics model.

[0170] The generation module 1804 is used to generate multiple combinations of process parameters based on key process parameters using the response surface methodology.

[0171] Simulation module 1805 is used to perform numerical simulations of multiple combinations of process parameters using computational fluid dynamics models to obtain the corresponding multiple methane yields;

[0172] The determination module 1806 is used to determine the optimal combination of process parameters based on multiple combinations of process parameters and corresponding multiple methane yields.

[0173] In one possible implementation, module 1801 is specifically used for: collecting experimental data from the bio-methanation trickle bed process and constructing an experimental dataset, which includes input variables and output variables. The input variables include process parameters affecting methane production, and the output variables include methane production. The experimental dataset is then processed to remove outliers and impute missing values ​​to obtain a regularized dataset. The regularized dataset is divided into a training set and a test set according to a preset ratio. A preset machine learning model is trained using the training set to construct an initial production prediction model. The initial production prediction model is optimized using a strategy combining cross-validation and hyperparameter optimization to determine the model parameter combination. Finally, the optimized initial production prediction model is evaluated using preset performance evaluation metrics based on the test set to determine the production prediction model.

[0174] In one possible implementation, the analysis module 1802 is specifically used to: calculate the contribution of the input variables of the production forecast model to the output variables using a preset interpretability analysis method; and select key process parameters from the input variables based on the contribution.

[0175] In one possible implementation, module 1803 is established, specifically for: constructing a system of differential equations based on the Monod equation; solving the system of differential equations using a pre-defined numerical integration method to obtain kinetic parameters, including the maximum specific growth rate, half-saturation constant, microbial stem cell concentration, and yield coefficient; fitting the kinetic parameters to pre-collected dissolved hydrogen concentration data to obtain a biological reaction rate expression, wherein the dissolved hydrogen concentration data is collected through headspace equilibrium experiments; and establishing a kinetic model based on the biological reaction rate expression.

[0176] In one possible implementation, module 1803 is further specifically used for: establishing a computational fluid dynamics model based on the preset size data of the bio-methanation ectopic trickle bed; meshing the computational fluid dynamics model to obtain a computational mesh; configuring model parameters and boundary conditions on the computational mesh; writing the dynamics model as a user-defined function; and coupling the user-defined function to the computational fluid dynamics model.

[0177] In one possible implementation, the simulation module 1805 is specifically used for: configuring convergence control parameters; performing steady-state coupling solutions on multiple combinations of process parameters based on the convergence control parameters using a computational fluid dynamics model, and iterating through an adaptive time step to obtain multiple corresponding simulation results; and post-processing the multiple simulation results to obtain multiple corresponding methane yields.

[0178] In one possible implementation, module 1806 is specifically used to: perform regression analysis on multiple combinations of process parameters and corresponding multiple methane production using the response surface methodology to obtain a second-order regression prediction model; and solve the second-order regression prediction model with the goal of maximizing methane production to obtain the optimal combination of process parameters.

[0179] In one possible implementation, the process parameter optimization device further includes:

[0180] The verification module is used to obtain the maximum methane production corresponding to the optimal process parameter combination in the second-order regression prediction model; input the optimal process parameter combination into the computational fluid dynamics model for verification simulation to obtain the verification methane production; compare the verification methane production with the maximum methane production; if the deviation between the verification methane production and the maximum methane production is less than the preset deviation threshold, the optimal process parameter combination is confirmed to be effective.

[0181] The process parameter optimization device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0182] Figure 19 This is a schematic diagram of the process parameter optimization equipment provided in an embodiment of this application. Figure 19 As shown, the process parameter optimization device provided in this embodiment includes at least one processor 1901 and a memory 1902. Optionally, the device further includes a communication component 1903. The processor 1901, memory 1902, and communication component 1903 are connected via a bus 1904.

[0183] In a specific implementation, at least one processor 1901 executes computer execution instructions stored in memory 1902, causing at least one processor 1901 to perform the above-described method.

[0184] The specific implementation process of processor 1901 can be found in the above method embodiments, and its implementation principle and technical effect are similar, so it will not be repeated here.

[0185] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0186] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0187] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0188] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0189] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0190] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0191] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0192] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0193] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0194] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0195] If a function is implemented as 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 this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0196] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0197] Finally, it should be noted that other embodiments of this application will readily conceive of by those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and alterations may be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for optimizing process parameters, characterized in that, include: A production prediction model is constructed based on a pre-collected experimental dataset, wherein the input variables of the production prediction model include process parameters that affect methane production, and the output variable includes methane production. An interpretability analysis was performed on the production prediction model to obtain key process parameters; A kinetic model was established based on the kinetic parameters of microbial growth. A computational fluid dynamics model for a biomethanation ex-situ trickle bed is established, and the dynamics model is coupled to the computational fluid dynamics model. Multiple combinations of process parameters are generated based on the key process parameters using the response surface methodology. The computational fluid dynamics model is used to numerically simulate the multiple combinations of process parameters to obtain the corresponding multiple methane yields; The optimal combination of process parameters is determined based on the multiple combinations of process parameters and the corresponding multiple methane yields.

2. The method according to claim 1, characterized in that, The process of constructing a yield prediction model based on pre-collected experimental datasets includes: Experimental data of the biomethanation trickle bed process were collected and an experimental dataset was constructed. The experimental dataset includes input variables and output variables. The input variables include process parameters that affect methane production, and the output variables include methane production. The experimental dataset was subjected to outlier removal and missing value imputation to obtain a regularized dataset; The regularized dataset is divided into a training set and a test set according to a preset ratio; The training set is used to train a pre-defined machine learning model to construct an initial yield prediction model; The initial yield prediction model is optimized by employing a strategy of cross-validation combined with hyperparameter optimization to determine the combination of model parameters. Using preset performance evaluation indicators, the optimized initial output prediction model is evaluated based on the test set to determine the output prediction model.

3. The method according to claim 1, characterized in that, The process of performing interpretability analysis on the production prediction model to obtain key process parameters includes: The contribution of the input variables to the output variables of the production forecasting model is calculated using a pre-defined interpretability analysis method. Key process parameters are obtained by filtering from the input variables based on the contribution level.

4. The method according to claim 1, characterized in that, The establishment of a kinetic model based on the kinetic parameters of microbial growth includes: Construct a system of differential equations based on Monod's equations; The system of differential equations is solved using a pre-defined numerical integration method to obtain kinetic parameters, which include the maximum specific growth rate, half-saturation constant, microbial stem cell concentration, and yield coefficient. The kinetic parameters are fitted based on pre-collected dissolved hydrogen concentration data to obtain an expression for the biological reaction rate, wherein the dissolved hydrogen concentration data is collected through headspace equilibrium experiments. A kinetic model is established based on the aforementioned biological reaction rate expression.

5. The method according to any one of claims 1 to 4, characterized in that, The establishment of a computational fluid dynamics model for the bio-methanation ex-situ trickle bed, and the coupling of the dynamics model to the computational fluid dynamics model, includes: A computational fluid dynamics model was established based on the pre-set size data of the bio-methanation ex-situ trickle bed. The computational fluid dynamics model is meshed to obtain a computational mesh; Configure model parameters and boundary conditions on the computational grid; The dynamic model is written as a user-defined function; Couple the custom function to the computational fluid dynamics model.

6. The method according to any one of claims 1 to 4, characterized in that, The process of numerically simulating the multiple combinations of process parameters using the computational fluid dynamics model to obtain corresponding methane yields includes: Configure convergence control parameters; Using the computational fluid dynamics model and the convergence control parameters, the multiple sets of process parameter combinations are solved in a steady-state coupled manner, and iterative processing is performed using an adaptive time step to obtain multiple corresponding simulation results. The simulation results are post-processed to obtain the corresponding methane yields.

7. The method according to any one of claims 1 to 4, characterized in that, The step of determining the optimal combination of process parameters based on the multiple combinations of process parameters and the corresponding multiple methane yields includes: The response surface methodology was used to perform regression analysis on the multiple combinations of process parameters and the corresponding multiple methane yields to obtain a second-order regression prediction model. With the goal of maximizing methane production, the second-order regression prediction model is solved to obtain the optimal combination of process parameters.

8. The method according to claim 7, characterized in that, After determining the optimal combination of process parameters based on the multiple combinations of process parameters and the corresponding multiple methane yields, the method further includes: Obtain the maximum methane production corresponding to the optimal combination of process parameters in the second-order regression prediction model; The optimal combination of process parameters is input into the computational fluid dynamics model for verification simulation to obtain the verified methane yield; Compare the verified methane production rate with the maximum methane production rate; If the deviation between the verified methane production and the maximum methane production is less than a preset deviation threshold, then the optimal process parameter combination is confirmed to be effective.

9. A process parameter optimization device, characterized in that, include: The module is used to build a production prediction model based on a pre-collected experimental dataset, wherein the input variables of the production prediction model include process parameters that affect methane production, and the output variable includes methane production. The analysis module is used to perform interpretability analysis on the production prediction model to obtain key process parameters; A module is established to create a kinetic model based on the kinetic parameters of microbial growth; The establishment module is also used to establish a computational fluid dynamics model of the biomethanation ex-situ trickle bed and couple the dynamics model to the computational fluid dynamics model; The generation module is used to generate multiple sets of process parameter combinations based on the key process parameters using the response surface methodology. The simulation module is used to perform numerical simulations of the multiple sets of process parameter combinations using the computational fluid dynamics model to obtain the corresponding multiple methane yields; The determination module is used to determine the optimal combination of process parameters based on the multiple combinations of process parameters and the corresponding multiple methane yields.

10. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method described in any one of claims 1-8.

Images