Self-repairing absorbent formulation optimization method and system, device, medium
By using a self-healing absorbent formulation optimization method and a support vector regression model, the chemical degradation problem of organic amine solution absorbents was solved, achieving self-healing and degradation inhibition of the absorbent, reducing operating costs and secondary pollution, and improving the industrial feasibility of the optimized formulation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HUANENG CLEAN ENERGY RES INST
- Filing Date
- 2026-03-19
- Publication Date
- 2026-07-24
AI Technical Summary
Existing organic amine solution absorbents suffer from chemical degradation in flue gas carbon capture in industries such as thermal power and steel. This leads to a decrease in absorbent concentration, reduction in absorption capacity, generation of corrosive degradation products, and increased operating costs. Existing optimization methods are difficult to accurately establish nonlinear relationships between formulation components and process parameters in small sample scenarios, and cannot simultaneously achieve degradation inhibition, reduction of secondary pollution, and energy consumption control.
A self-healing absorbent formulation optimization method was adopted. Through self-healing absorbent formulation optimization experiments and reversible degradation-equilibrium inhibition chemical kinetic calculations, a support vector regression model was constructed to screen candidate formulations that meet engineering constraints. The optimized formulation was verified through experiments, and a repair adjuvant was added to establish chemical equilibrium to inhibit the degradation of the main absorbent.
It achieves the self-healing ability of the absorbent, reduces chemical degradation and secondary pollution, extends the stable service life, reduces the cost of long-term supplementation, and improves the industrial feasibility and small-sample predictive ability of the optimized formulation.
Smart Images
Figure CN122455151A_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of absorbent formulation optimization technology, and specifically relates to methods, systems, equipment and media for optimizing self-healing absorbent formulations. Background Technology
[0002] Chemical CO2 absorption methods, represented by organic amine solutions, are widely used for carbon capture in high-emission industries such as thermal power and steel due to their mature technology and high absorption efficiency. However, chemical degradation of the absorbent remains a core bottleneck for long-term operation, cost reduction, and environmental impact mitigation. This manifests as decreased absorbent concentration, reduced absorption capacity, corrosion and secondary pollution caused by degradation products, and increased operating costs due to frequent replenishment / replacement. Traditional strategies tend to passively defend against degradation by improving process operation or adding inhibitors, with limited effectiveness.
[0003] Current methods for improving the anti-degradation performance of organic amine absorbent formulations remain within a "passive protection" framework. These methods only slow down absorbent degradation by screening degradation inhibitors and adjusting process parameters (such as temperature and oxygen partial pressure), failing to break through the traditional logic of "inhibiting degradation" and unable to achieve "active regeneration" of the absorbent. This results in absorbents still struggling to overcome the industry bottleneck of continuous activity decay and shortened stabilization time during long-term use. Furthermore, existing optimization methods (such as orthogonal experiments and single linear regression) have poor adaptability to small sample scenarios. They struggle to accurately establish nonlinear relationships between formulation components (such as amine ratios and additive concentrations), process parameters, and multiple performance indicators of the absorbent (absorption capacity, degradation rate, and desorption energy consumption) with limited experimental data. This leads to insufficient generalization of optimization results, failing to simultaneously address multiple engineering objectives such as degradation inhibition, secondary pollution reduction, and energy consumption control. Ultimately, optimized formulations struggle to balance stability and economy during industrial scale-up, failing to fundamentally solve the long-term cost and environmental impact issues of absorbents. Summary of the Invention
[0004] To address the aforementioned issues, this application provides a method, system, equipment, and medium for optimizing the formulation of self-healing absorbents.
[0005] The first objective of this application is to provide a method for optimizing the formulation of a self-healing absorbent, including: Experimental datasets were obtained based on self-healing absorbent formulation optimization experiments and reversible degradation-equilibrium inhibition chemical kinetic calculations. Based on the experimental dataset, construct the model dataset; A formula-performance mapping model is established using support vector regression. Based on the model dataset, a recipe-performance mapping model for support vector regression is trained. The recipe space is simulated and sampled on the trained relational model to select several candidate recipes that meet the engineering constraints. Based on the laboratory and pilot-scale verification results of several candidate formulations, an optimized self-healing absorbent formulation was obtained.
[0006] In a specific embodiment of this application, the experimental dataset is constructed based on the characteristic quantity data and performance variable data corresponding to different self-healing absorbent formulations; The performance variables include degradation rate constant, main absorbent stabilization time, absorption capacity, absorption rate, degradation product concentration, and desorption energy consumption. The characteristic quantities include the main absorbent type code, the proportion of the main absorbent in the formulation, the proportion of the remaining components in the formulation, the molecular weight of the main absorbent, temperature, and oxygen partial pressure in the flue gas.
[0007] In a specific embodiment of this application, the step of constructing a model dataset based on the experimental dataset includes: Dimensionality reduction is performed on the feature quantities and performance variables in the experimental dataset; The dimensionality-reduced feature values and performance variable data are used to construct the model dataset.
[0008] In a specific embodiment of this application, training a recipe-performance mapping model based on a model dataset using support vector regression includes: On the model dataset, a two-step grid search is performed on the relation model parameters, and the optimal relation model parameters are selected by combining K-fold cross-validation.
[0009] In a specific embodiment of this application, the engineering constraint is the range of values for one or more performance variables.
[0010] In a specific embodiment of this application, the self-healing absorbent formulation optimization method further includes: updating the formulation-performance mapping relationship model established by support vector regression; And / or, the update of the recipe-performance mapping model for support vector regression includes: The experimental validation results of several candidate formulations are incorporated into the model dataset and the relational model is iteratively updated.
[0011] A second objective of this application is to provide a self-healing absorbent formulation optimization system, comprising: Database module: used to obtain experimental datasets based on self-healing absorbent formulation optimization experiments and reversible degradation-equilibrium inhibition chemical kinetic calculations; also used to construct model datasets based on experimental datasets; Model module: used to establish a recipe-performance mapping model using support vector regression; also used to train the recipe-performance mapping model based on the model dataset; simulate sampling of the recipe space on the trained relationship model to select several candidate recipes that meet the engineering constraints; Output module: Used to obtain an optimized self-healing absorbent formulation based on laboratory and pilot-scale verification results of several candidate formulations.
[0012] In a specific embodiment of this application, an iteration module is also included, which is used to update the recipe-performance mapping relationship model established for support vector regression; And / or, the update of the recipe-performance mapping model for support vector regression includes: The experimental validation results of several candidate formulations are incorporated into the model dataset and the relational model is iteratively updated.
[0013] A third objective of this application is to provide an electronic device comprising: a processor coupled to a memory; The memory is used to store computer programs; The processor is configured to execute the computer program stored in the memory, so that the electronic device performs the method.
[0014] A fourth objective of this application is to provide a computer-readable storage medium storing a program or instructions that, when run on a computer, cause the computer to perform the method as described.
[0015] Compared with the prior art, this application has the following advantages: This application discloses a method, system, equipment, and medium for optimizing the formulation of a self-healing absorbent. In the process of optimizing the formulation of the self-healing absorbent, this application constructs and trains a formulation-performance mapping relationship model. Through self-healing absorbent formulation optimization experiments and reversible degradation-equilibrium inhibition chemical kinetic calculations, the trained relationship model can simulate the process of inhibiting the forward degradation reaction and promoting the reverse repair in the actual optimization process of the self-healing absorbent formulation from a thermodynamic and kinetic perspective. This makes the optimized formulation after training more consistent with the actual process, improving the self-healing ability and feasibility of the optimized formulation. Meanwhile, the optimized formulation obtained by the method of this application, after cross-validation by the trained relational model and laboratory and pilot-scale verification, can not only reduce the chemical degradation of the absorbent and inhibit the secondary pollution caused by degradation, but also achieve in-situ active self-repair of the absorbent, and is also rapidly industrially feasible.
[0016] Furthermore, in the process of optimizing the formula output, this application incorporates experimental verification and Chinese verification results into the training set and iteratively updates the model, thereby further improving the industrial feasibility of the optimized formula obtained by the optimized model.
[0017] Furthermore, the optimized self-healing absorbent formulation of this application features efficient formulation screening with small samples: it employs support vector regression, which has superior generalization ability for small-sample, nonlinear problems, and can efficiently predict formulation performance and screen high-potential candidate formulations even with limited experimental data, significantly saving R&D time and experimental costs. Unlike traditional technologies that add antioxidants such as phenols, inorganic salts, and organic amines, relying on passive protection logic such as "consuming themselves to block the degradation chain reaction" or "directly reacting with O2," this application adds a key product or its stable analogue to the initial formulation. Based on the Le Chatelier principle, this product establishes a reversible chemical equilibrium with the main absorbent, kinetically reducing the net degradation rate. Throughout the process, the adjuvant is not consumed, maintaining a long-term equilibrium and effectively extending the stable service life of the absorbent, significantly reducing the long-term replenishment costs caused by frequent absorbent degradation.
[0018] In traditional technologies, antioxidants are easily degraded after consumption, leading to the continued degradation of the main absorbent and the generation of harmful byproducts such as aldehydes, ketones, and organic acids, resulting in secondary pollution. This application establishes a chemical balance through additives, inhibiting the degradation reaction of the main amine at its source and directly reducing the amount of harmful degradation products generated. Simultaneously, the additives used are harmless and do not participate in side reactions, avoiding the introduction of new pollutants and significantly reducing secondary pollution.
[0019] Other features and advantages of this application will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description
[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0021] Figure 1 A flowchart of a method for optimizing a self-healing absorbent formulation according to an embodiment of this application is shown; Figure 2 This invention illustrates the framework of a self-healing absorbent formulation optimization system according to an embodiment of this application; Figure 3 A frame diagram of an electronic device according to an embodiment of this application is shown; In the diagram: 10, Database module; 20, Model module; 30, Output module; 300, Electronic device; 301, Processor; 302, Memory. Detailed Implementation
[0022] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0023] like Figure 1 As shown, a method for optimizing a self-healing absorbent formulation according to certain embodiments of this application includes: S1. Based on the self-healing absorbent formulation optimization experiment and the reversible degradation-equilibrium inhibition chemical kinetic calculation, the experimental dataset was obtained; S2. Based on the experimental dataset, construct the model dataset; S3. Use support vector regression to establish a formula-performance mapping model; S4. Based on the model dataset, train a recipe-performance mapping relationship model for support vector regression. S5. Simulate sampling of the recipe space on the trained relational model to select several candidate recipes that meet the engineering constraints. S6. Based on the laboratory and pilot-scale verification results of several candidate formulations, the optimized self-healing absorbent formulation is obtained.
[0024] In some embodiments of this application, the self-healing absorbent includes a primary absorbent and a repair aid.
[0025] The primary absorbent is, for example, one or a mixture of several organic amines such as monoethanolamine (MEA), diethanolamine (DEA), methyldiethanolamine (MDEA), and piperazine (PZ) in any proportion; The repair aid is a substance added according to the type of main absorbent, and its function is to inhibit the degradation of the main absorbent.
[0026] In some embodiments of this application, the self-healing absorbent further includes a stabilizer or corrosion inhibitor, and a solvent; The stabilizer / corrosion inhibitor is used to adjust the redox potential and inhibit metal corrosion and side reactions; The stabilizer is, for example, sodium bisulfite (NaHSO3) or sodium metabisulfite (Na2S2O5). The corrosion inhibitors are, for example, benzotriazole (BTA) and methylbenzotriazole (MBT). The solvent is water or an organic solvent.
[0027] In some embodiments of this application, if the main absorbent is MEA, the repair aid corresponds to N-(2-hydroxyethyl)ethylenediamine (HEEDA); If the primary absorbent is DEA, then the repair aid is triethanolamine (TEA). If the primary absorbent is PZ, then the repair aid is 1-formylpiperazine (FPZ). If the primary absorbent is MDEA, then the repair aid is dimethylaminoethanol (DMAE).
[0028] In some embodiments of this application, the self-healing absorbent comprises, by weight percentage (wt.%): Main absorbent: 10-40 wt.%; Repairing agent: 0.01-5 wt.%; Stabilizer / Corrosion Inhibitor: 0-1 wt.%; The remainder is solvent.
[0029] In some embodiments of this application, the reversible degradation-equilibrium inhibition chemical kinetic calculation is performed using a reversible degradation-equilibrium inhibition chemical framework, which is constructed based on the degradation reaction of the primary absorbent in a self-healing absorbent. For example (this is only an example using the case of two primary absorbents; the case of one or more primary absorbents is analogous): The degradation process of the main absorbent is simplified as a reversible reaction using the following equation (1): Main absorbent A + Main absorbent B Product C+ Product D (1) In formula (1): A and B are the initial primary absorbents (undegraded state), unit: mol·L -1 ; C and D are degradation products or intermediates, in mol·L⁻¹. -1 ; The forward (degradation) rate constant (if it is a second-order reaction), unit: L·mol -1 ·min -1 or min -1 (Depending on the reaction order); The reverse (reconstruction / recombination) rate constant, in units of .
[0030] The expression for the forward velocity is shown in equation (2): (2) The expression for the reverse rate is shown in equation (3): (3) The expression for the net rate is shown in equation (4): (4) The definition of the equilibrium constant is given in equation (5): (5) The expression for the chemical potential of a substance is shown in equation (6): (6) The expression for the total chemical potential difference in the above reaction is shown in equation (7): (7) Actively adding a small amount of harmless analogues of the key product C or intermediates (i.e., the aforementioned remediation adjuvants) to the initial absorbent formulation can increase the chemical potential on the product side, thereby inhibiting the forward degradation reaction.
[0031] By actively adding small amounts of harmless remediation adjuvants to the initial absorbent formulation, the concentration on the product side can be increased. Thermodynamically reduce the driving force of positive degradation (making This shifts the equilibrium towards zero or the positive direction, making the equilibrium more biased towards the left (reactants), thus inhibiting degradation. Simultaneously, it kinetically improves... It helps to increase the reverse speed This, in turn, reduces the net degradation rate. .
[0032] During engineering implementation, the dosage of remediation adjuvants should be controlled within a range that is "sufficient to increase the chemical potential of the product side to inhibit net degradation, but without affecting absorption / regeneration performance and without producing harmful side effects." An initial dosage range of 0.01-1 wt.% can be used, and an industrial-scale dosage strategy can be established through experimental / pilot-scale optimization (multiple methods are available, such as direct addition of C, addition of stable analogues, or addition of precursors and online release).
[0033] In some embodiments of this application, the experimental dataset is constructed based on characteristic quantity data and performance variable data corresponding to different self-healing absorbent formulations; The performance variables include: Degradation rate constant Unit: min -1 ; Main absorbent stabilization time (Time to maintain the main absorbent mass concentration >90% under given conditions (e.g., fixed temperature), unit: min; Carbon dioxide absorption capacity Unit: mol CO2·mol -1 amine; Absorption rate Unit: mol·L -1 ·min -1 ; Degradation product concentration Unit: mol·L -1 ; Desorption energy consumption index Unit: GJ / t CO2; The characteristic parameters include: the primary absorbent type code, the proportion of the primary absorbent in the formulation, the proportion of the remaining components in the formulation, the molecular weight of the primary absorbent, and the temperature. (°C), partial pressure of oxygen in flue gas (kPa).
[0034] In some embodiments of this application, the experimental dataset is expressed as ,in: The feature vector for the i-th sample (features include the main absorber type code, main absorber mole fraction, and temperature). (°C), partial pressure of oxygen in flue gas (kPa, etc.), feature quantities are uniformly normalized or standardized before modeling; For the corresponding performance variables ( , E, β, q (etc.). Sample size Generally within the range of 10-200 (small sample scenarios), feature dimensionality. It depends on the number of molecular descriptors; The characteristic quantity is obtained based on the initial self-healing absorbent formulation and experimental conditions; the performance variable is calculated based on the characteristic quantity and equations (1)-(7).
[0035] In some embodiments of this application, step S2 includes: S1-1, Reduce the dimensionality of the features and performance variables in the experimental dataset; S1-2. Construct the model dataset from the dimensionality-reduced feature quantities and performance variable data.
[0036] In some embodiments of this application, the dimensionality reduction process employs correlation analysis, principal component analysis (PCA), or linear regression (LASSO).
[0037] In some embodiments of this application, step S3, for example, uses support vector regression (SVR), with radial basis function (RBF) as the kernel function, to construct a recipe-performance mapping relationship model. The expression of the relationship model is shown in equation (8): (8) In equation (8), x is the kernel width parameter. i and x j Both are vectors of the training samples in the feature space (the i-th and j-th samples). K(xi, xj) represents the similarity (or inner product) between two samples in the training set and is used to construct the kernel matrix during the training phase.
[0038] The training objective of SVR (ε-insensitive loss) is formalized as Equation (9): (9) In equation (9), For the weight vector, This is the bias value. For implicit mapping, For insensitive bandwidth, As a penalty factor, These are slack variables.
[0039] In some embodiments of this application, step S4 includes: On the model dataset, a two-step grid search and K-fold cross-validation were used to select the optimal relational model parameters. These parameters included the dimensionless penalty factor F for support vector regression and the kernel width parameter. and insensitive bandwidth K-fold cross-validation is a well-known verification method in this technical field, and will not be elaborated upon here. In some embodiments of this application, in step S5, the method of simulation sampling is, for example, Monte Carlo simulation or uniform design numerical simulation.
[0040] In some embodiments of this application, in step S5, Monte Carlo simulation sampling is performed on the formulation space (including the concentration, ratio, temperature, etc. of the repair adjuvant) within permissible limits to select candidates. use Predicting targets Filter to meet engineering constraints (e.g.) , < Energy consumption < Several candidate formulations (etc.).
[0041] In some embodiments of this application, in step S5, the expression of the obtained trained relational model is as shown in equation (10): (10) In equation (10), These are the Lagrange multipliers obtained through training.
[0042] In some embodiments of this application, the engineering constraint is a condition regarding the range of values of one or more performance variables. For example, the engineering constraint is... , , .
[0043] In some embodiments of this application, in step S6, the laboratory and pilot-scale verification of several candidate formulations is as follows: Laboratory and pilot-scale testing of candidate formulations were conducted to obtain... , E, and assess the risks and economics of industrial scaling up (such as solvent replenishment, regeneration energy consumption, solvent purification costs, etc.).
[0044] For example, indicators of the risks and economics of industrial scaling include degradation product profile, CO2 uptake capacity, and regeneration energy consumption.
[0045] The laboratory and pilot-scale verification of absorbent formulations, as well as the assessment of risks and economics for industrial scale-up, are well-known practices in this field and will not be elaborated upon here.
[0046] In some embodiments of this application, the method further includes: updating the recipe-performance mapping model for support vector regression, wherein the updating of the recipe-performance mapping model for support vector regression includes: The experimental validation results of several candidate formulations are incorporated into the model dataset, and the relational model is iteratively updated. The experimental validation results of the candidate formulations include laboratory and pilot-scale validation results.
[0047] In some embodiments of this application, during the update process of the relational model, the iteration stopping condition is: among the candidate recipes screened in the current iteration, there exists at least one set of recipes that fully satisfies the preset engineering constraints (such as...). > , < , E < ).
[0048] The above methods were used to optimize the actual self-healing absorbent formulation: Example 1: MEA Main Absorbent Self-Healing Absorbent Formulation Initial formulation conditions: main absorbent is monoethanolamine (MEA) with a mass fraction of 30 wt.%; temperature is 40℃; oxygen partial pressure in flue gas is 6 kPa; corrosion inhibitor is 0.3 wt.%; initial liquid phase CO2 loading is 0.25 mol CO2 / mol amine; the initial addition range of degradation products is 0.01–1 wt%.
[0049] Experimental data collection: Under the above conditions, small-scale parallel experiments were conducted through targeted formulation design (different proportions, concentrations of different repair adjuvants, and temperatures, etc.), obtaining 18 initial sample data (including formulation characteristics and corresponding stabilization times). (and other indicators), the target requirements are: >1200 min.
[0050] Data Modeling and Optimization: Radial Basis Function (RBF) Support Vector Regression (RBF-SVR) was used to establish the formulation-performance mapping relationship. A two-step grid search (coarse selection followed by refinement) was performed on the model parameters, combined with K-fold cross-validation to select the optimal parameters. Monte Carlo or uniform sampling predictions were then performed on the trained relational model to predict the formulation space (including the concentration and ratio of remedial adjuvants, temperature, etc.) and to screen for formulations that meet the requirements. Several candidate formulations that meet the conditions of >1200 min and other engineering constraints (such as energy consumption and degradation product limits).
[0051] Experimental verification: The actual performance of some candidate formulations was verified under laboratory / pilot-scale conditions. The optimized formulation meets the requirements in terms of degradation product spectrum, CO2 absorption capacity, and regeneration energy consumption.
[0052] Table 1 shows the optimized formulation of the MEA main absorbent self-healing absorbent.
[0053] Table 1
[0054] Example 2: Optimization of PZ main absorbent self-healing absorbent formulation: Initial formulation conditions: main absorbent is piperazine (PZ) with a mass fraction of 10 wt.%; temperature is 50°C; oxygen partial pressure in flue gas is 10 kPa; corrosion inhibitor is 0.1 wt.%; initial CO2 loading in the liquid phase is 0.15 mol CO2 / mol amine; the initial addition range of degradation products is 0.05–1.5 wt%.
[0055] Experimental data collection: Under the above conditions, small-scale parallel experiments were conducted through targeted formulation design (different proportions, concentrations of different repair adjuvants, and temperatures, etc.), obtaining 18 initial sample data (including formulation characteristics and corresponding stabilization times). (and other indicators), the target requirements are: >2000 min.
[0056] Data Modeling and Optimization: The RBF-SVR modeling process (same as in Example 1) was adopted. The optimal model parameters were determined by combining coarse and fine two-step grid search with K-fold cross-validation. After training, the optimal parameters were predicted and selected in the formulation space. Candidate formulations with a lifespan of >2000 min were selected, while energy consumption and degradation product concentration were kept within acceptable limits.
[0057] Experimental validation: Several candidate formulations were validated in the laboratory to evaluate their long-term stability, degradation product profile and operational economy. The validation data were incorporated into the model for iterative optimization, and the selected formulation met the requirements.
[0058] Table 2 shows the optimized formulation of the PZ main absorbent self-healing absorbent.
[0059] Table 2
[0060] like Figure 2 The self-healing absorbent formulation optimization system according to certain embodiments of this application includes: Database Module 10: Used to obtain experimental datasets based on self-healing absorbent formulation optimization experiments and reversible degradation-equilibrium inhibition chemical kinetic calculations; also used to construct model datasets based on experimental datasets; Model Module 20: It is used to establish a recipe-performance mapping relationship model using support vector regression; it is also used to train the recipe-performance mapping relationship model established by support vector regression based on the model dataset; and to simulate sampling the recipe space on the trained relationship model to select several candidate recipes that meet the engineering constraints. Output module 30: used to obtain an optimized self-healing absorbent formulation based on laboratory and pilot-scale verification results of several candidate formulations.
[0061] In some embodiments of this application, the model module 20 includes a training submodule and a screening submodule; The training submodule is used to train a recipe-performance mapping model for support vector regression based on the model dataset. The screening submodule is used to simulate sampling the formulation space on the trained relational model, screen out several candidate formulations that meet the engineering constraints, and output the optimized self-healing absorbent formulation based on the laboratory and pilot-scale verification results of several candidate formulations.
[0062] In some embodiments of this application, the system further includes an iterative module for updating the recipe-performance mapping model for support vector regression, wherein the updating of the recipe-performance mapping model for support vector regression includes: The experimental validation results of several candidate formulations are incorporated into the model dataset and the relational model is iteratively updated.
[0063] like Figure 3 As shown, in some embodiments of this application, an electronic device is provided, the electronic device 300 including: a processor 301, the processor 301 being coupled to a memory 302; The memory 302 is used to store computer programs; The processor 301 is configured to execute the computer program stored in the memory 302, so that the electronic device performs the method described in the above embodiments.
[0064] In some embodiments of this application, a computer-readable storage medium is provided that stores a program or instructions that, when executed on a computer, cause the computer to perform the methods described in the above embodiments.
[0065] According to embodiments of this application, the computer-readable storage medium may be a non-volatile computer-readable storage medium, such as including but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, electronic device, or apparatus.
[0066] In some embodiments of this application, a computer program product is provided, including a computer program / instructions that, when executed by a processor, implement the methods described in the above embodiments.
[0067] Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.
Claims
1. A method for optimizing the formulation of a self-healing absorbent, characterized in that, include: Experimental datasets were obtained based on self-healing absorbent formulation optimization experiments and reversible degradation-equilibrium inhibition chemical kinetic calculations. Based on the experimental dataset, construct the model dataset; A formula-performance mapping model is established using support vector regression. Based on the model dataset, a recipe-performance mapping relationship model for support vector regression is trained. The recipe space is simulated and sampled on the trained relational model to select several candidate recipes that meet the engineering constraints. Based on the laboratory and pilot-scale verification results of several candidate formulations, an optimized self-healing absorbent formulation was obtained.
2. The method for optimizing the formulation of a self-healing absorbent according to claim 1, characterized in that, The experimental dataset was constructed based on the characteristic data and performance variable data corresponding to different self-healing absorbent formulations. The performance variables include degradation rate constant, main absorbent stabilization time, absorption capacity, absorption rate, degradation product concentration, and desorption energy consumption. The characteristic quantities include the main absorbent type code, the proportion of the main absorbent in the formulation, the proportion of the remaining components in the formulation, the molecular weight of the main absorbent, temperature, and oxygen partial pressure in the flue gas.
3. The method for optimizing the formulation of a self-healing absorbent according to claim 1, characterized in that, The construction of the model dataset based on the experimental dataset includes: Dimensionality reduction is performed on the feature quantities and performance variables in the experimental dataset; The reduced-dimensionality feature data and performance variable data are used to construct the model dataset.
4. The method for optimizing the formulation of a self-healing absorbent according to claim 1, characterized in that, The process of training a recipe-performance mapping model based on the model dataset for support vector regression includes: On the model dataset, a two-step grid search is performed on the relation model parameters, and the optimal relation model parameters are selected by combining K-fold cross-validation.
5. The method for optimizing the formulation of a self-healing absorbent according to claim 1, characterized in that, The engineering constraints are the conditions governing the range of values for one or more performance variables.
6. The method for optimizing the formulation of a self-healing absorbent according to claim 1, characterized in that, Also includes: Update the recipe-performance mapping model for support vector regression; And / or, the update of the recipe-performance mapping model for support vector regression includes: The experimental validation results of several candidate formulations are incorporated into the model dataset and the relational model is iteratively updated.
7. A self-healing absorbent formulation optimization system, characterized in that, include: Database module: used for experimental datasets based on self-healing absorbent formulation optimization experiments and reversible degradation-equilibrium inhibition chemical kinetic calculations; It is also used to build model datasets based on experimental datasets; Model module: Used to establish a recipe-performance mapping model using support vector regression; It is also used to train a recipe-performance mapping model for support vector regression based on the model dataset; The recipe space is simulated and sampled on the trained relational model to select several candidate recipes that meet the engineering constraints. Output module: Used to obtain an optimized self-healing absorbent formulation based on laboratory and pilot-scale verification results of several candidate formulations.
8. The self-healing absorbent formulation optimization system according to claim 7, characterized in that, It also includes an iteration module, which is used to update the recipe-performance mapping relationship model for support vector regression; And / or, the update of the recipe-performance mapping model for support vector regression includes: The experimental validation results of several candidate formulations are incorporated into the model dataset and the relational model is iteratively updated.
9. An electronic device, characterized in that, include: Processor, the processor being coupled to memory; The memory is used to store computer programs; The processor is configured to execute the computer program stored in the memory to cause the electronic device to perform the method as described in any one of claims 1 to 6.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program or instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1 to 6.