A multi-objective optimization method for bubble column carbon capture process based on surrogate model

Abstract

The present invention provides a multi-objective optimization method for a bubbling column carbon capture process based on a surrogate model, comprising the following steps: constructing a physical simulation model of a packed bubbling column reactor through COMSOL Multiphysics software; sampling in the design domain by the optimal Latin hypercube method according to preselected optimization objectives and design variables; constructing an adaptive hybrid surrogate model, training the model using simulation results, and evaluating the accuracy and stability of the model using R<supgt;2< / supgt> and RMSE, etc.; performing Sobol global sensitivity analysis on different design variables; solving the multi-objective optimization of the selected surrogate model using the non-dominated genetic algorithm II; drawing a Pareto front plot based on the solution set obtained by NSGA-II to obtain the optimal solution of the multi-objective optimization; outputting a multi-objective optimization scheme that meets the pre-specified process requirements, obtaining the best trade-off solution between the energy consumption required for the process and the carbon dioxide capture rate and the corresponding optimal operating conditions, and these data have guiding significance for the subsequent industrial application of the process.

Description

Technical Field

[0001] The present invention belongs to the technical field of carbon dioxide capture process optimization, and in particular relates to a multi-objective optimization method for a bubbling tower carbon capture process based on an agent model. Background Art

[0002] The demand for the "net zero" goal has prompted the development of carbon capture, utilization, and storage (CCUS) technology, in which enhanced weathering (EW) of mineral particles may absorb hundreds of billions of tons of atmospheric carbon dioxide each year and store it in the ocean in the form of bicarbonates and carbonates. However, weathering under natural conditions is very slow, and chemical reactors are needed to accelerate this process under controllable conditions. Therefore, finding a suitable chemical reactor and suitable operating conditions for the reactor is the key to whether this method can be implemented on a large scale in an industrial manner. Studies in recent years have shown that packed bubble columns (PBCs) can be used as reactors for capturing carbon dioxide based on mineral EW. Before the packed bubble column is used for carbon dioxide capture, its performance indicators such as carbon dioxide capture rate and energy consumption need to be optimized to obtain the best design variables that meet different goals.

[0003] Generally speaking, for the optimization of industrial processes with long experiments and high costs, a physical simulation model of the process is often constructed first to conduct computational research. However, for multi-variable and multi-objective optimization, thousands or even millions of simulation tasks are often required. It is not feasible to directly solve the physical simulation model in the entire design domain to obtain the optimal solution. Summary of the invention

[0004] In response to the above technical problems, one of the purposes of one embodiment of the present invention is to provide a multi-objective optimization method for a bubble tower carbon capture process based on a proxy model, so that it can be used to solve the optimization problem of a packed bubble tower carbon dioxide capture process and obtain the required multi-objective optimization solution.

[0005] The present invention is a multi-objective optimization method for capturing carbon dioxide through mineral particle enhanced weathering in a packed bubble tower based on a proxy model, and a hybrid proxy model is used to simplify the physical model to obtain a multi-objective optimization solution. The proxy model is a simplified model based on a statistical basis, with a small amount of calculation and a fast solution, but its calculation results are close to those of the physical model and have a higher accuracy. The present invention applies the hybrid proxy model to the optimization of the carbon dioxide capture process, greatly reducing the amount of calculation of multivariable optimization, while improving the accuracy of the simulation optimization results.

[0006] The present invention provides a multi-objective optimization method for a packed bubble tower carbon dioxide capture process based on a proxy model, comprising the following steps: S1, construction of a multi-physics field simulation model: using COMSOL Multiphysics software, constructing a physical simulation model of a packed bubble tower reactor, which includes the reaction kinetics of carbon dioxide capture and the mass transfer process of carbon dioxide at the gas-liquid interface in the reactor; S2, sampling of a proxy model data set: sampling in a design domain by an optimal Latin hypercube sampling method (OLHS) according to preselected optimization objectives and design variables; S3, construction and evaluation of a proxy model: constructing an adaptive hybrid proxy model, using simulation results to train the model, and using R 2 and RMSE to evaluate the accuracy and stability of the model; S4, global sensitivity analysis: Sobol global sensitivity analysis is performed on different design variables; S5, proxy model optimization solution: non-dominated sorting genetic algorithm II (NSGA-II) is used to perform multi-objective optimization solution on the selected proxy model; S6, multi-objective trade-off solution diagram drawing: Pareto frontier diagram is drawn according to the solution set obtained by NSGA-II to obtain the optimal solution for multi-objective optimization; S7, optimization scheme output: output a multi-objective optimization scheme that meets the pre-specified process requirements. The multi-objective optimization method of the packed bubble tower carbon dioxide capture process based on the proxy model proposed in the present invention can obtain the optimal trade-off solution of the energy consumption and carbon dioxide capture rate required by the process and the corresponding optimal operating conditions (design variables), and these data have guiding significance for the subsequent industrial application of the process.

[0007] The present invention achieves the above technical objectives through the following technical means.

[0008] A multi-objective optimization method for a bubble column carbon capture process based on a surrogate model comprises the following steps:

[0009] Step S1, construction of a multi-physics field simulation model: constructing a physical simulation model of a series-packed bubble column reactor for capturing CO2 based on enhanced weathering of mineral particles, wherein the physical simulation model includes the reaction kinetics of carbon dioxide capture and the mass transfer process of carbon dioxide at the gas-liquid interface in the reactor;

[0010] Step S2, sampling of the proxy model data set: sampling in the design domain by an optimal Latin hypercube sampling method according to the preselected optimization objective function and design variables of the cascade packed bubble column reactor process for capturing CO2;

[0011] Step S3: Construction and evaluation of the proxy model: Based on the simulation results of the sampling points in step S2, an adaptive hybrid proxy model is constructed, and the model is trained using the simulation results. 2 and RMSE to evaluate the accuracy and stability of the model;

[0012] Step S4, global sensitivity analysis: after the accuracy of the proxy model in step S3 reaches the standard, Sobol global sensitivity analysis is performed on different design variables;

[0013] Step S5, surrogate model optimization solution: after the sensitivity analysis in step S4, a non-dominated genetic algorithm is used to perform multi-objective optimization solution on the selected surrogate model;

[0014] Step S6, drawing a multi-objective trade-off solution diagram: drawing a Pareto frontier diagram based on the solution set obtained by the non-dominated genetic algorithm in step S5 to obtain the optimal solution;

[0015] Step S7, output of optimization scheme: by comparing the optimal solution obtained in step S6, find out the CO2 capture rate with the maximum value and the process energy consumption with the corresponding objective function value of the optimal design variable X corresponding to the optimal solution in the non-dominated genetic algorithm, and output a multi-objective optimization scheme that meets the pre-specified process requirements.

[0016] In the above scheme, the physical simulation model of the filled bubble tower reactor in step S1 is a physical simulation model of a series of filled bubble tower reactors for capturing CO2 based on the enhanced weathering of mineral particles. The series of filled bubble tower reactors includes at least two filled bubble tower reactors, the first reactor is a filled bubble tower based on seawater, and the second reactor is a filled bubble tower based on fresh water.

[0017] Furthermore, the construction of the multi-physics field simulation model in step S1 is based on the finite element method in COMSOL Multiphysics software.

[0018] In the above scheme, the objective function in step S2 is:

[0019] f(X)=[f1(X),f2(X)] T

[0020] Among them, f1(X) represents the maximum CO2 capture rate, and f2(X) represents the minimum process energy consumption;

[0021] The design variable X:

[0022] X=[x1,x2,x3,x4,x5,x6,x7,x8] T ,X∈K

[0023] Among them, x1, x2, x3 and x4 are the mineral particle size, bed height, gas flow rate and liquid flow rate in the seawater-filled bubble column reactor respectively; x5, x6, x7 and x8 correspond to the mineral particle size, bed height, gas flow rate and liquid flow rate in the freshwater-filled bubble column reactor, and K is the design domain of the process to be optimized, which is expressed as:

[0024] K=[[1,10],[1,10],[0.0001,0.001],[0.0001,0.001],[1,10],[1,10],[0.0001,0.001],[0.001,0.01]] T .

[0025] In the above solution, the mixed proxy model in step S3 includes at least two single proxy models.

[0026] Furthermore, the hybrid proxy model in step S3 includes polynomial response surface, Kriging, radial basis function and support vector machine.

[0027] In the above scheme, in step S3, R 2 And RMSE is obtained by the following formula:

[0028]

[0029]

[0030] Where N is the number of test sets, is the predicted value of the mixed surrogate model, y i is the test set obtained by simulating the physical model, is the average value of the test set obtained by simulating the physical model.

[0031] In the above scheme, the process of the non-dominated genetic algorithm in the agent model optimization solution in step S5 is:

[0032] Step S1), population initialization;

[0033] Step S2), obtaining the initial population P through fast non-dominated sorting, selection, crossover and mutation operations;

[0034] Step S3), merge the parent population and the child population into 2P, and then calculate the next generation population individuals through fast non-dominated sorting and crowding degree;

[0035] Step S4), continue to generate the next generation according to the genetic operation until the maximum evolutionary generation Gen is reached max stop;

[0036] Step S5), output the optimal solution set.

[0037] Furthermore, the crossover distribution index and the mutation distribution index of the crossover and mutation operations in step S2) are both 20.

[0038] In the above scheme, the multi-objective optimization scheme is an optimization scheme that uses minimum energy consumption to achieve the maximum capture rate.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] The present invention simplifies the physical model by constructing a hybrid proxy model to obtain a multi-objective optimization solution, fully considers the coupling relationship between design variables and the correlation characteristics between objective functions, and adopts a fast non-dominated multi-objective optimization algorithm (NSGA-II) with an elite retention strategy to achieve multi-variable and multi-objective optimization, and finally obtains an optimization solution that simultaneously satisfies the minimum energy consumption and the maximum capture rate, providing more accurate supporting data for subsequent process use. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a flow chart of a multi-objective optimization method for a packed bubble tower carbon dioxide capture process based on a surrogate model according to one embodiment of the present invention;

[0042] Figure 2 for Figure 1 Flowchart of the NSGA-II optimization algorithm in step (S5);

[0043] Figure 3 This is a schematic diagram of a series of packed bubble column reactors in an example of one embodiment of the present invention;

[0044] Figure 4 is the R of the hybrid agent model for the two performance indicators in an embodiment of an embodiment of the present invention 2 And RMSE analysis chart;

[0045] Figure 5 A global Sobol sensitivity analysis diagram of different variables according to an embodiment of the present invention;

[0046] Figure 6 It is a dual-objective optimization Pareto front diagram obtained in an example of an implementation manner of the present invention. DETAILED DESCRIPTION

[0047] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0048] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "front", "back", "left", "right", "up", "down", "axial", "radial", "vertical", "horizontal", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0049] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0050] Figure 1 The figure shows a preferred embodiment of the multi-objective optimization method of the bubbling tower carbon capture process based on the surrogate model. The multi-objective optimization method of the bubbling tower carbon capture process based on the surrogate model comprises the following steps:

[0051] Step S1, construction of a multi-physics field simulation model: constructing a physical simulation model of a series-packed bubble column reactor for capturing CO2 based on enhanced weathering of mineral particles, wherein the physical simulation model includes the reaction kinetics of carbon dioxide capture and the mass transfer process of carbon dioxide at the gas-liquid interface in the reactor;

[0052] Step S2, sampling of the proxy model data set: sampling in the design domain by an optimal Latin hypercube sampling method according to the preselected optimization objective function and design variables of the cascade packed bubble column reactor process for capturing CO2;

[0053] Step S3: Construction and evaluation of the proxy model: Based on the simulation results of the sampling points in step S2, an adaptive hybrid proxy model is constructed, and the model is trained using the simulation results. 2 and RMSE to evaluate the accuracy and stability of the model;

[0054] Step S4, global sensitivity analysis: after the accuracy of the proxy model in step S3 reaches the standard, Sobol global sensitivity analysis is performed on different design variables;

[0055] Step S5, surrogate model optimization solution: after the sensitivity analysis in step S4, a non-dominated genetic algorithm is used to perform multi-objective optimization solution on the selected surrogate model;

[0056] Step S6, drawing a multi-objective trade-off solution diagram: drawing a Pareto frontier diagram based on the solution set obtained by the non-dominated genetic algorithm in step S5 to obtain the optimal solution;

[0057] Step S7, output of optimization scheme: by comparing the optimal solution obtained in step S6, find out the CO2 capture rate with the maximum value and the process energy consumption with the corresponding objective function value of the optimal design variable X corresponding to the optimal solution in the non-dominated genetic algorithm, and output a multi-objective optimization scheme that meets the pre-specified process requirements.

[0058] According to this embodiment, preferably, the construction of the multi-physics field simulation model in step S1 is: Figure 3 A schematic diagram of a series of packed bubble tower reactors for capturing CO2 based on enhanced weathering (EW) of mineral particles is shown, wherein the first reactor is a seawater-based packed bubble tower, and the second reactor is a freshwater-based packed bubble tower. The mineral particles used in both are calcite, and the inlet gas is assumed to be power plant exhaust gas containing 15% CO2.

[0059] The waste gas from the power plant to be treated is first passed into the first reactor, a seawater-filled bubbling tower, for preliminary capture. As the reaction proceeds, the content of CO2 gradually decreases, because the enhanced weathering in the seawater-filled bubbling tower can only capture mixed gases with a high CO2 content, and the uncaptured CO2 is discharged with the waste gas. The exhaust gas is then passed into the second reactor, a freshwater reactor, a seawater-based filled bubbling tower for secondary capture. The main chemical reaction equation involved in the whole process is:

[0060]

[0061]

[0062]

[0063]

[0064]

[0065]

[0066]

[0067] Based on the process flow, the corresponding physical simulation model was constructed by the finite element method (FEM) in COMSOL Multiphysics software. The model includes the reaction kinetics of carbon dioxide capture and the mass transfer process of carbon dioxide at the gas-liquid interface in the reactor. The one-dimensional geometric space and time-related variables are obtained by solving the control equations composed of multiple partial differential equations (PDEs). The calculation domain is treated as a uniform porous medium, in which the actual geometric shape of individual particles is not considered. The MUMPS time-dependent solver is used, the default parameter settings, the calculation tolerance is set to physical control, and the relative tolerance is 0.001.

[0068] The step S2 is to sample the proxy model data set. Preferably, the step first determines the optimization target and design variables of the process. In this embodiment, there are two objective functions in total:

[0069] f(X)=[f1(X),f2(X)] T

[0070] Where f1(X) represents the maximum CO2 capture rate (mass of CO2 captured per hour), and f2(X) represents the minimum process energy consumption (energy required per unit mass of CO2 captured).

[0071] There are eight design variables X studied:

[0072] X=[x1,x2,x3,x4,x5,x6,x7,x8] T ,X∈K

[0073] Where x1, x2, x3 and x4 are the mineral particle size, bed height, gas flow rate and liquid flow rate in the seawater-filled bubble column reactor; x5, x6, x7 and x8 correspond to the mineral particle size, bed height, gas flow rate and liquid flow rate in the fresh water reactor. K is the design domain of the process to be optimized, which is expressed as:

[0074] K=[[1,10],[1,10],[0.0001,0.001],[0.0001,0.001],[1,10],[1,10],[0.0001,0.001],[0.001,0.01]] TBecause the numerical differences between different variables are large, all variables are normalized to make their range [0,1]. After determining the design variables and research objectives, the optimal Latin hypercube sampling (OLHS) method is used to sample in the design domain, and the ratio of the training set (M) to the test set (N) in the obtained data set is 4:1.

[0075] In step S3, the construction and evaluation of the proxy model is preferably to construct an extended adaptive hybrid proxy model (E-AHF, Extended adaptive hybrid functions) based on the simulation results of the OLHS sampling points. In this example, the hybrid proxy model includes four single proxy models (polynomial response surface, Kriging, radial basis function and support vector machine). After the model is trained, R is calculated. 2 The accuracy and stability of the model are evaluated by using the RMSE, and the calculation formula is as follows:

[0076]

[0077]

[0078] Where N is the number of test sets; is the predicted value of the mixed agent model; is the average value of the test set obtained by physical model simulation, R 2 The closer it is to 1, the smaller the RMSE is, and the higher the prediction accuracy and stability of the corresponding model are. If the performance of the obtained model is poor, it is necessary to further improve the model accuracy by increasing the data set.

[0079] In this embodiment, a total of 400 sets of data are used to construct the required proxy model. 2 The values ​​of RMSE are shown in the following table:

[0080] Table 1 R of the E-AHF model for PBC series reactors based on different data sets 2 and RMSE

[0081]

[0082] CR – CO2 capture rate, EC – energy consumption

[0083] Figure 4 More intuitively shows the R of the model after adding sample points 2 and RMSE value changes. Figure 4 The following conclusions can be drawn from the analysis:

[0084] 1) As the number of sample points increases, the R 2 will gradually increase, and RMSE will gradually decrease;

[0085] 2) When there are fewer sample points, the change in the number of sample points has an impact on R 2 The impact of RMSE is more obvious;

[0086] 3) When the number of samples is greater than 300, the R of the CO2 capture rate is increased by increasing the number of samples. 2 and RMSE have little impact;

[0087] 4) When the data set reaches 400, the CO2 capture rate and energy consumption R 2 All of them exceeded 0.95, indicating good prediction accuracy of the surface model. In the final E-AHF proxy model, the R of CO2 capture rate 2 and RMSE were 0.9686 and 0.0061 respectively, and the R 2 and RMSE are 0.9720 and 1.2955 respectively. This shows that both models have good prediction accuracy and stability, and also shows that the accuracy of the obtained hybrid surrogate model can meet the requirements of achieving multi-objective optimization.

[0088] In step S4, the global sensitivity analysis is preferably performed on different design variables after the accuracy of the proxy model reaches the standard. The obtained global sensitivity index S T As shown in Table 2 below:

[0089] Table 2 Sobol global sensitivity index values ​​for different design variables

[0090]

[0091] According to the results in Table 2, the Figure 5 By analyzing Table 2 and Figure 5 Different design variables have different effects on different objective functions. The design variable with the greatest impact on CO2 capture rate is the gas flow rate (x3) of the seawater-filled packed tower, which has a Sobol global sensitivity index of 0.55710 for CO2 capture rate; followed by the liquid flow rate (x8) of the freshwater-filled bubble tower, which has a Sobol global sensitivity index of 0.34641 for CO2 capture rate. As for energy consumption, the bed height (x6) of the freshwater-filled bubble tower and the bed height (x2) of the seawater-filled packed tower have much greater impact than other design variables, and their Sobol global sensitivity indexes for energy consumption are 0.50203 and 0.25873, respectively. The larger global sensitivity index indicates that after these design variables change, the results of the objective function will also change significantly.

[0092] In the step S5, the proxy model is optimized and solved. Preferably, after the sensitivity analysis, a non-dominated sorting genetic algorithm (NSGA-II) is used to perform multi-objective optimization and solution on the selected proxy model.

[0093] Figure 2 The main process of NSGA-II is shown:

[0094] (1) Population initialization;

[0095] (2) The initial population P is obtained through fast non-dominated sorting, selection, crossover and mutation operations;

[0096] (3) The parent population and the offspring population are merged into 2P, and then the individuals of the next generation population are obtained through fast non-dominated sorting and crowding calculation;

[0097] (4) Continue to generate the next generation according to genetic operations until the maximum number of generations (Gen max )stop;

[0098] (5) Output the optimal solution set;

[0099] In this example, preferably, the crossover distribution index and the variation distribution index are both 20, the population number P is 35, and the maximum evolutionary generation number Gen max The number of generations is 20. In this example, the two objective functions used for non-dominated sorting are f1(X) and f2(X) determined in step (S2), namely, the maximum CO2 capture rate and the minimum energy consumption.

[0100] The step S6 is to draw a multi-objective trade-off solution diagram. Preferably, a Pareto solution set distribution diagram is drawn based on the solution set obtained by NSGA-II. Figure 6 The red part is the optimal solution set obtained from NSGA-II, and the black points are the data sets previously used for the hybrid agent model. Table 3 shows ten sets of Pareto optimal solutions obtained by the NSGA-II algorithm.

[0101] Table 3 Partial Pareto solution set obtained by NSGA-II

[0102]

[0103] pass Figure 6 It can be seen that the obtained Pareto frontier is better than the simulation data used previously.

[0104] The optimization solution output in step S7 is preferably obtained by using a Pareto frontier diagram to obtain a multi-objective Pareto optimal solution, and the corresponding objective function value is recorded. Then, the corresponding data of the optimal design variable X corresponding to the optimal solution in NSGA-II is found by comparing the given optimal solution. In this example, the optimal trade-off solution between energy consumption and CO2 capture rate is found, that is, the corresponding design variable value under the condition of capturing more CO2 while consuming as little energy as possible.

[0105] In this example, the optimal solution selected after comprehensive analysis of the Pareto distribution chart is the green point in the chart, that is, the fourth group of optimal solutions in Table 3. The corresponding optimal design variable values ​​are:

[0106] X=[x1,x2,x3,x4,x5,x6,x7,x8] T =[2.2315,4.8145,0.001,0.0007,8.3635,1,0.00047,0.00835] T Where x1, x2, x3 and x4 correspond to the mineral particle size, bed height, gas flow rate and liquid flow rate in the seawater-filled bubble column reactor; x5, x6, x7 and x8 correspond to the mineral particle size, bed height, gas flow rate and liquid flow rate in the freshwater reactor, and the corresponding energy consumption and CO2 capture rate are 5.58933 MJ kg -1 CO2 and 0.11756 kg h -1 Finally, the solution is output as the final multi-objective optimization solution.

[0107] The present invention relates the energy consumption and capture rate performance indicators of the carbon dioxide capture process in a packed bubble tower under ore weathering to design variables such as flow rate and bed height by constructing a hybrid proxy model, fully considering the coupling relationship between design variables and the correlation characteristics between performance indicators, and adopting a fast non-dominated multi-objective optimization algorithm (NSGA-II) with an elite retention strategy to achieve multi-variable and multi-objective optimization, and ultimately obtaining an optimization solution that can achieve the maximum capture rate with minimum energy consumption, providing relatively accurate supporting data for subsequent process use.

[0108] It should be understood that although this specification is described according to various embodiments, not every embodiment contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

[0109] The series of detailed descriptions listed above are only specific descriptions of feasible embodiments of the present invention. They are not intended to limit the scope of protection of the present invention. All equivalent embodiments or changes that do not deviate from the technical spirit of the present invention should be included in the scope of protection of the present invention.

Claims

1. A multi-objective optimization method for a bubble tower carbon capture process based on a proxy model, characterized in that: The following steps are involved: Step S1, construction of a multi-physical field simulation model: constructing a physical simulation model of a series of packed bubble tower reactors for capturing CO2 based on enhanced weathering of mineral particles; the physical simulation model of the packed bubble tower reactor is a physical simulation model of a series of packed bubble tower reactors for capturing CO2 based on enhanced weathering of mineral particles, the series of packed bubble tower reactors includes at least two packed bubble tower reactors, the packed bubble tower reactors include a packed bubble tower based on seawater and a packed bubble tower based on fresh water; the physical simulation model includes the reaction kinetics of carbon dioxide capture and the mass transfer process of carbon dioxide at the gas-liquid interface in the reactor; Step S2, sampling of proxy model data set: according to the process of capturing CO2 with a series-packed bubble column reactor, First, determine the optimization objectives and design variables of the process. Objective function: f(X)=[f1(X),f2(X)] T Among them, f1(X) represents the maximum CO2 capture rate, and f2(X) represents the minimum process energy consumption; The design variable X: X=[x1,x2,x3,x4,x5,x6,x7,x8] T ,X∈K Among them, x1, x2, x3 and x4 are the mineral particle size, bed height, gas flow rate and liquid flow rate in the seawater-filled bubble column reactor respectively; x5, x6, x7 and x8 correspond to the mineral particle size, bed height, gas flow rate and liquid flow rate in the freshwater-filled bubble column reactor, and K is the design domain of the process to be optimized, which is expressed as: K=[[1,10],[1,10],[0.0001,0.001],[0.0001,0.001],[1,10],[1,10],[0.0001,0.001],[0.001,0.01]] T ;exist After the design variables and objectives are determined, sampling is performed in the design domain by using the optimal Latin hypercube sampling method, and the sampling points are input into the physical simulation model constructed in step S1 to obtain simulation results of the corresponding sampling points; Step S3, construction and evaluation of the proxy model: construct an adaptive hybrid proxy model based on the simulation results of the sampling points in step S2, train the hybrid proxy model using the simulation results, and use R 2 and RMSE to evaluate the accuracy and stability of the hybrid surrogate model; Step S4, global sensitivity analysis: After the accuracy and stability of the hybrid proxy model in step S3 meet the standards, Sobol global sensitivity analysis is performed on different design variables X to obtain the global sensitivity index S T ; Step S5, surrogate model optimization solution: after the sensitivity analysis in step S4, a non-dominated genetic algorithm is used to perform multi-objective optimization solution on the selected hybrid surrogate model; Step S6, drawing a multi-objective trade-off solution diagram: drawing a Pareto frontier diagram based on the solution set obtained by the non-dominated genetic algorithm in step S5 to obtain the optimal solution; Step S7, output of optimization scheme: by comparing the optimal solution obtained in step S6, find the CO2 capture rate with the maximum and the process energy consumption with the minimum corresponding objective function value of the optimal design variable X corresponding to the optimal solution in the non-dominated genetic algorithm, and output a multi-objective optimization scheme that meets the pre-specified process requirements; the multi-objective optimization scheme includes the achievable minimum energy consumption and maximum capture rate and the corresponding value of the optimal design variable X.

2. The multi-objective optimization method for the bubble tower carbon capture process based on the proxy model according to claim 1, characterized in that: The construction of the multi-physics field simulation model in step S1 is based on the finite element method in COMSOL Multiphysics software.

3. The multi-objective optimization method for the bubble tower carbon capture process based on the proxy model according to claim 1, characterized in that: The mixed proxy model in step S3 includes at least two single proxy models.

4. The multi-objective optimization method for the bubble tower carbon capture process based on the proxy model according to claim 3, characterized in that: The hybrid proxy model in step S3 includes polynomial response surface, Kriging, radial basis function and support vector machine.

5. The multi-objective optimization method for bubble column carbon capture process based on agent model according to claim 1, characterized in that: In step S3, R 2 And RMSE is obtained by the following formula: Where N is the number of test sets, is the predicted value of the mixed surrogate model, y i is the test set obtained by simulating the physical model, is the average value of the test set obtained by simulating the physical model.

6. The multi-objective optimization method for a bubble column carbon capture process based on a proxy model according to claim 1, characterized in that: The process of the non-dominated genetic algorithm in the agent model optimization solution in step S5 is: Step S1), population initialization; Step S2), obtaining the initial population P through fast non-dominated sorting, selection, crossover and mutation operations; Step S3), merge the parent population and the child population into 2P, and then calculate the next generation population individuals through fast non-dominated sorting and crowding degree; Step S4), continue to generate the next generation according to the genetic operation until the maximum evolutionary generation Gen is reached max stop; Step S5), output the optimal solution set.

7. The multi-objective optimization method for a bubble tower carbon capture process based on a proxy model according to claim 6, characterized in that: The crossover distribution index and mutation distribution index of the crossover and mutation operations in step S2) are both 20.

Images