Reaction kinetics model driven lithium carbonate preparation process optimization method and system

Abstract

The embodiment of the present disclosure provides a reaction kinetics model driven lithium carbonate preparation process optimization method and system. The method comprises the following steps: obtaining experimental data, the experimental data comprising chemical reaction data, experimental state parameters and experimental control data; performing chemical reaction characteristic analysis on the chemical reaction data and the experimental state parameters to construct a reaction dynamic model for indicating the reaction kinetics behavior in the lithium carbonate preparation process; constructing a reaction control model for indicating the control of the lithium carbonate preparation process based on the experimental state parameters and the experimental control data; obtaining preparation state parameters in the current lithium carbonate preparation process, so that under the joint control of the reaction dynamic model and the reaction control model on the preparation state parameters, an action decision is determined to maximize the reward value determined by the production performance index in the lithium carbonate preparation process, so as to optimize the lithium carbonate preparation process. The present disclosure improves flexibility and can face dynamic changes in the preparation production process.

Description

Technical Field

[0001] This disclosure relates to the field of lithium carbonate preparation technology, and more specifically, to a method and system for optimizing the lithium carbonate preparation process driven by a reaction kinetic model. Background Technology

[0002] Lithium carbonate has a wide range of applications, especially in the new energy industry. As a raw material for lithium batteries, lithium carbonate has extremely important economic value and social benefits. Currently, the preparation methods of lithium carbonate are often limited by fixed operating conditions, lacking flexibility and making it difficult to cope with dynamic changes in the preparation and production process. Summary of the Invention

[0003] This disclosure provides a reaction kinetics model-driven optimization method and system for lithium carbonate preparation, which addresses the technical problem of limited flexibility due to fixed operating conditions in existing technologies, and is able to adapt to dynamic changes in the preparation and production process.

[0004] According to one aspect of the present disclosure, a reaction kinetics model-driven optimization method for lithium carbonate preparation is provided, comprising:

[0005] Obtain experimental data from the lithium carbonate preparation experiment, including chemical reaction data, experimental state parameters, and experimental control data;

[0006] Chemical reaction characteristics were analyzed based on the chemical reaction data and experimental state parameters to construct a reaction kinetic model to indicate the reaction kinetics behavior in the lithium carbonate preparation process;

[0007] Based on the experimental state parameters and the experimental control data, a reaction control model for indicating and controlling the lithium carbonate preparation process is constructed.

[0008] The preparation state parameters of the current lithium carbonate preparation process are obtained, and under the joint control of the reaction kinetic model and the reaction control model with respect to the preparation state parameters, action decisions are made to maximize the reward value determined by the production performance indicators in the lithium carbonate preparation process, so as to optimize the lithium carbonate preparation process. The action decisions include adjustments to the preparation process parameters.

[0009] In one possible implementation, the chemical reaction characteristic analysis of the chemical reaction data and the experimental state parameters to construct a reaction kinetic model for indicating the reaction kinetics behavior in the lithium carbonate preparation process includes:

[0010] Chemical reaction characteristics were analyzed based on the chemical reaction data and experimental state parameters to determine the reaction mechanism related to the lithium carbonate preparation process;

[0011] Based on the experimental state parameters, the model parameters are determined, including the reaction rate constant and activation energy parameters.

[0012] The reaction mechanism and model parameters are applied to a preset kinetic model, and the kinetic model is verified to obtain a reaction kinetic model.

[0013] In one possible implementation, the chemical reaction characteristic analysis of the chemical reaction data and the experimental state parameters to determine the reaction mechanism related to the lithium carbonate preparation process includes:

[0014] Based on the aforementioned chemical reaction data, the reaction mechanism related to the preparation of lithium carbonate was determined;

[0015] Based on the reaction mechanism and the experimental state parameters, the chemical structure of the chemical reaction data is analyzed to determine the corresponding reaction sites and the influencing factors that affect the reaction sites.

[0016] Based on the reaction mechanism, the reaction sites, the influencing factors, the experimental state parameters, and the chemical reaction data, the reaction pathways involved are determined.

[0017] The correctness of the reaction pathway is verified in order to verify and correct the reaction mechanism.

[0018] In one possible implementation, determining the reaction pathway based on the reaction mechanism, the reaction site, the influencing factors, the experimental state parameters, and the chemical reaction data includes:

[0019] Based on the chemical reaction data, the reactants and products involved in the lithium carbonate preparation process are determined, and at least one intermediate and byproduct formed by the reactants and products are identified to determine the reaction participants related to the lithium carbonate preparation process.

[0020] Based on the reaction mechanism, the reaction site, and the influencing factors, a reaction pathway containing at least one reaction step related to the reaction participants is determined, and the reaction pathway records the reaction participants and their interconversion relationships.

[0021] Based on the experimental state parameters, the reaction correlation between any reaction condition and the reaction path and / or the reaction participants is determined, so that based on the reaction correlation, a target reaction path and its corresponding target reaction step are screened in the reaction path and / or the reaction step, wherein the reaction conditions include at least one of temperature, pressure, pH value and solvent.

[0022] The feasibility of each of the target reaction steps is evaluated using a reaction evaluation strategy, which is based on preset thermodynamic and kinetic principles, in order to verify the target reaction pathway.

[0023] In one possible implementation, determining the model parameters based on the experimental state parameters includes:

[0024] The initial values ​​of the model parameters are obtained by using the pre-established parameter estimation model and the experimental state parameters.

[0025] The estimated values ​​of the model parameters are optimized using a preset optimization model and the initial values ​​of the model parameters.

[0026] The optimized estimates are applied to the kinetic model to predict the reaction results under any reaction condition. The reaction results are then compared and verified with preset experimental verification data to adjust the estimated values ​​of the model parameters.

[0027] In one possible implementation, the method further includes:

[0028] Analytical parameters for sensitivity analysis are selected from the model parameters of the reaction kinetic model, and the range of variation for each analytical parameter is determined.

[0029] A preset sensitivity analysis strategy is used to simulate the analysis parameters of the reaction dynamics model within the range of variation, so as to output the parameter analysis results. The sensitivity analysis strategy is used to indicate whether to perform local sensitivity analysis or global sensitivity analysis. The parameter analysis results include the degree of influence of the parameter changes of the analysis parameters on the output of the reaction dynamics model.

[0030] Based on the parameter analysis results, the analytical parameters corresponding to the maximization of the influence degree are selected, and the influence reasons of the selected analytical parameters are determined based on the reaction mechanism, so that the sensitivity analysis results are composed of the analytical parameters and their influence reasons;

[0031] The reaction dynamics model was optimized based on the sensitivity analysis results.

[0032] In one possible implementation, the method further includes:

[0033] Based on the optimized reaction kinetics model, a preparation simulation model was established;

[0034] Determine the initial simulation conditions, which include the initial material concentration, initial temperature, and initial pressure;

[0035] The preparation simulation model is run using the initial simulation conditions, and simulation results are obtained. The simulation results include parameter changes for the simulation parameters obtained by screening the model parameters, and the simulation parameters include reactant concentration parameters and product concentration parameters.

[0036] The simulation results are evaluated to determine the simulation evaluation results, so that the optimized reaction kinetic model can be adjusted based on the simulation evaluation results. The simulation evaluation results include the simulated reaction progress and the simulated product yield.

[0037] Risk assessment and safety analysis are conducted based on the simulation results to obtain control optimization data for the lithium carbonate preparation process. The control optimization data includes operation adjustment data and material ratio adjustment data.

[0038] In one possible implementation, constructing a reaction control model for indicative control of the lithium carbonate preparation process based on the experimental state parameters and the experimental control data includes:

[0039] The experimental control data is used as training data, which includes operating condition data, process parameters, and product quality data.

[0040] The preset neural network architecture is trained using the training data, and then the model is validated and optimized to obtain a reaction control model for indicative control of the lithium carbonate preparation process.

[0041] In one possible implementation, obtaining the preparation state parameters of the current lithium carbonate preparation process, such that under the joint control of the reaction kinetic model and the reaction control model with respect to the preparation state parameters, determines the action decision that maximizes the reward value determined by the production performance indicators in the lithium carbonate preparation process, in order to optimize the lithium carbonate preparation process, includes:

[0042] The current preparation state parameters in the lithium carbonate preparation process are used as the current state of a preset reinforcement learning model. The adjustment actions for each of the preparation process parameters are iterated. The reaction prediction results determined by the reaction kinetic model and the control prediction results determined by the reaction control model are combined to predict the next state of the preparation state parameters in the subsequent lithium carbonate preparation process. The production performance indicators after the adjustment actions are executed are also predicted, wherein the production performance indicators include yield, purity and energy consumption.

[0043] Based on the achievement status obtained by comparing the production performance indicators with the preset achievement threshold, the reward value of the reinforcement learning model is determined, and the action decision that maximizes the reward value is determined.

[0044] The learning strategy of the reinforcement learning model is updated based on the determined action decision to optimize the lithium carbonate preparation process according to the learning strategy.

[0045] According to another aspect of the present disclosure, a reaction kinetics model-driven optimization system for lithium carbonate preparation is provided, comprising:

[0046] The data acquisition module is used to acquire experimental data in the lithium carbonate preparation experiment, including chemical reaction data, experimental state parameters, and experimental control data.

[0047] The reaction kinetic model construction module is used to analyze the chemical reaction characteristics of the chemical reaction data and the experimental state parameters in order to construct a reaction kinetic model to indicate the reaction kinetic behavior in the lithium carbonate preparation process.

[0048] The reaction control model construction module is used to construct a reaction control model for indicating and controlling the lithium carbonate preparation process based on the experimental state parameters and the experimental control data.

[0049] An optimization module is used to acquire the preparation state parameters in the current lithium carbonate preparation process, so that under the joint control of the reaction kinetic model and the reaction control model with respect to the preparation state parameters, an action decision is determined to maximize the reward value determined by the production performance indicators in the lithium carbonate preparation process, so as to optimize the lithium carbonate preparation process. The action decision includes adjustment actions for the preparation process parameters.

[0050] The beneficial effects of the technical solutions provided in this disclosure are:

[0051] The reaction kinetics model-driven optimization method for lithium carbonate preparation provided in this disclosure acquires experimental data from lithium carbonate preparation experiments, including chemical reaction data, experimental state parameters, and experimental control data. The method then performs chemical reaction characteristic analysis on the chemical reaction data and experimental state parameters to construct a reaction kinetics model to indicate the reaction kinetics behavior in the lithium carbonate preparation process. Subsequently, based on the experimental state parameters and experimental control data, a reaction control model is constructed to indicate and control the lithium carbonate preparation process. This allows for the acquisition of the current preparation state parameters in the lithium carbonate preparation process, ensuring that the reaction kinetics model and the reaction control model are optimized with respect to the chemical reaction data. Under the joint control of the preparation state parameters, action decisions are made to maximize the reward value determined by the production performance indicators in the lithium carbonate preparation process, thereby optimizing the lithium carbonate preparation process. The action decisions include adjustments to the preparation process parameters. In this way, by combining reaction kinetics models, reaction control models, and reinforcement learning algorithms to optimize the lithium carbonate preparation process, the complex reaction kinetics behavior in the lithium carbonate preparation process can be fully considered. It can cope with the dynamic changes in the preparation production process, achieve reasonable and efficient control, and ensure the stable and efficient operation of the preparation process. This solves the technical problem of the lack of flexibility caused by the limitation of fixed operating conditions in the prior art, thereby greatly improving production efficiency and flexibility.

[0052] The above description is merely an overview of the technical solution disclosed herein. In order to better understand the technical means of this disclosure and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this disclosure more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0053] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 A schematic flowchart illustrating a reaction kinetics model-driven optimization method for lithium carbonate preparation, provided in this embodiment of the disclosure;

[0055] Figure 2 This is a schematic diagram of a reaction kinetics model-driven optimization system for lithium carbonate preparation, provided as an embodiment of the present disclosure. Detailed Implementation

[0056] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0057] It should be understood that the following specific examples illustrate the implementation of this disclosure, and those skilled in the art can easily understand other advantages and effects of this disclosure from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. This disclosure can also be implemented or applied through other different specific implementation methods, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this disclosure. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0058] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this disclosure, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0059] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this disclosure. The drawings only show the components related to this disclosure and are not drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0060] Furthermore, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the described aspects can be practiced without these specific details.

[0061] Example 1

[0062] Figure 1 This is a schematic flowchart of a reaction kinetics model-driven optimization method for lithium carbonate preparation, which includes steps S101 to S104.

[0063] S101. Obtain experimental data from the lithium carbonate preparation experiment, including chemical reaction data, experimental state parameters, and experimental control data.

[0064] In this disclosure, chemical reaction data is used to characterize the chemical substances involved in the lithium carbonate preparation process, such as reactants, products, intermediates, byproducts, and catalysts. Experimental state parameters are used to characterize operational state data related to the lithium carbonate preparation process, such as temperature, pressure, pH value, reactant concentration, product concentration, and solvent. Experimental control data are used to characterize operational control data during the lithium carbonate preparation process, such as operating conditions, process parameters, and product quality data. Preferably, the experimental data is cleaned to remove outliers and noise, and the data is standardized to facilitate analysis.

[0065] S102. Perform chemical reaction characteristic analysis on the chemical reaction data and the experimental state parameters to construct a reaction kinetic model to indicate the reaction kinetic behavior in the lithium carbonate preparation process.

[0066] In this disclosure, based on the experimental data and the chemical reaction characteristics of lithium carbonate preparation, the relevant reaction pathways and mechanisms are determined, thereby establishing a reaction kinetic model. Therefore, by analyzing the chemical reaction characteristics to construct a reaction kinetic model, this disclosure can fully consider the complex reaction kinetics during lithium carbonate preparation, address the dynamic changes in the preparation process, improve the accuracy and applicability of predicting the lithium carbonate preparation process, and achieve flexible optimization of the lithium carbonate preparation process.

[0067] S103. Based on the experimental state parameters and the experimental control data, construct a reaction control model for indicating and controlling the lithium carbonate preparation process.

[0068] In this disclosure, a reaction control model is constructed to predict the control operations in the lithium carbonate preparation process, thereby optimizing the control operations, ensuring high production performance, and improving the efficiency of lithium carbonate preparation.

[0069] S104. Obtain the preparation state parameters in the current lithium carbonate preparation process, so that under the joint control of the reaction kinetic model and the reaction control model with respect to the preparation state parameters, determine the action decision that maximizes the reward value determined by the production performance index in the lithium carbonate preparation process, so as to optimize the lithium carbonate preparation process. The action decision includes adjustment actions for the preparation process parameters.

[0070] In this disclosure, a reinforcement learning algorithm (such as a reinforcement learning model) is used in conjunction with a reaction kinetics model and a reaction control model to optimize the lithium carbonate preparation process and achieve intelligent process control. Specifically, the current preparation state parameters in the lithium carbonate preparation process are used as the current state of the reinforcement learning model, and adjustments to the preparation process parameters are used as action decisions. The agent of the reinforcement learning model iterates through each action decision based on the current state to predict the next state, which characterizes the preparation state parameters in the subsequent lithium carbonate preparation process, and to obtain the reward value after executing the action decision. More specifically, for the prediction of the next state, the current preparation state parameters in the lithium carbonate preparation process are input into the reaction kinetics model to perform chemical reaction prediction, obtaining reaction prediction results. This takes into account the complex reaction kinetics in the lithium carbonate preparation process, which is beneficial to improving production efficiency and yield. At the same time, the current preparation state parameters in the lithium carbonate preparation process are input into the reaction control model to perform process control prediction, obtaining control prediction results, thereby achieving process optimization and quality control. Then, based on the reaction prediction results and the control prediction results, the preparation state parameters in the lithium carbonate preparation process obtained after responding to any action decision are determined, i.e., the next state. Therefore, in the decision-making process for entering the next state, the action decision that maximizes the reward value is selected, and the lithium carbonate preparation process is optimized based on the selected action decision. Thus, this disclosure combines reaction kinetics model, reaction control model, and reinforcement learning model for process prediction and optimization, ensuring that the lithium carbonate preparation process meets the predetermined functional and performance requirements, and guaranteeing the efficiency and correctness of the lithium carbonate preparation process.

[0071] This embodiment provides a reaction kinetics model-driven optimization method for lithium carbonate preparation. By acquiring experimental data from lithium carbonate preparation experiments, including chemical reaction data, experimental state parameters, and experimental control data, the method analyzes the chemical reaction characteristics of the chemical reaction data and experimental state parameters to construct a reaction kinetics model to indicate the reaction kinetics behavior during lithium carbonate preparation. Then, based on the experimental state parameters and experimental control data, a reaction control model is constructed to indicate and control the lithium carbonate preparation process. This obtains the current preparation state parameters during the lithium carbonate preparation process. Under the joint control of the reaction kinetics model and the reaction control model regarding these preparation state parameters, the method determines action decisions that maximize the reward value determined by the production performance indicators during the lithium carbonate preparation process, thereby optimizing the lithium carbonate preparation process. These action decisions include adjustments to the preparation process parameters. By combining the reaction kinetics model, the reaction control model, and reinforcement learning algorithms to optimize the lithium carbonate preparation process, this method fully considers the complex reaction kinetics behavior during lithium carbonate preparation, addresses dynamic changes during the preparation process, and achieves reasonable and efficient control. This ensures stable and efficient operation of the preparation process, solving the technical problem of limited flexibility due to fixed operating conditions in existing technologies, thus greatly improving production efficiency and flexibility.

[0072] In some embodiments, the step of performing chemical reaction characteristic analysis on the chemical reaction data and the experimental state parameters to construct a reaction kinetic model for indicating the reaction kinetic behavior in the lithium carbonate preparation process includes:

[0073] Chemical reaction characteristics were analyzed based on the chemical reaction data and experimental state parameters to determine the reaction mechanism related to the lithium carbonate preparation process;

[0074] Based on the experimental state parameters, the model parameters are determined, including the reaction rate constant and activation energy parameters.

[0075] The reaction mechanism and model parameters are applied to a preset kinetic model, and the kinetic model is verified to obtain a reaction kinetic model.

[0076] In this embodiment, based on information regarding the reactants, products, intermediates, and catalysts in the lithium carbonate preparation process, and in conjunction with relevant literature and research findings, the basic reaction principle and mechanism of the lithium carbonate preparation process are determined, and the corresponding reaction pathway is determined based on this mechanism. Subsequently, based on the experimental state parameters of the lithium carbonate preparation process, preset parameter estimation methods, such as least squares method and maximum likelihood estimation, are selected to determine the estimated values ​​of the model parameters. Furthermore, the reaction kinetic model is validated using additional experimental data to ensure its accuracy and reliability.

[0077] Therefore, this embodiment analyzes the chemical reaction characteristics of the lithium carbonate preparation process, optimizes the model parameters to fit the dynamic changes in the lithium carbonate preparation process, and takes into account the complex reaction kinetics behavior in the lithium carbonate preparation process, thereby improving the accuracy of the reaction kinetic model.

[0078] In some embodiments, the chemical reaction characteristic analysis of the chemical reaction data and the experimental state parameters to determine the reaction mechanism related to the lithium carbonate preparation process includes:

[0079] Based on the aforementioned chemical reaction data, the reaction mechanism related to the preparation of lithium carbonate was determined;

[0080] Based on the reaction mechanism and the experimental state parameters, the chemical structure of the chemical reaction data is analyzed to determine the corresponding reaction sites and the influencing factors that affect the reaction sites.

[0081] Based on the reaction mechanism, the reaction sites, the influencing factors, the experimental state parameters, and the chemical reaction data, the reaction pathways involved are determined.

[0082] The correctness of the reaction pathway is verified in order to verify and correct the reaction mechanism.

[0083] In this embodiment, the chemical reaction characteristics are analyzed by examining the chemical structures of reactants and products to determine possible reaction sites. Factors influencing the reaction, such as temperature, pressure, pH, solvent, and catalyst, are considered. Then, based on the reaction mechanism, reaction sites, influencing factors, experimental parameters, and chemical reaction data, possible reaction pathways are constructed. Furthermore, the feasibility of the reaction pathways is evaluated, specifically by assessing the thermodynamic and kinetic characteristics of each pathway, such as free energy change and activation energy, to determine the rate-limiting step and possible side reactions. Further, the reaction mechanism is verified through experimental validation, data analysis, and mechanism correction. For experimental validation, a small-scale reaction is conducted under laboratory conditions to collect data verifying the correctness of the reaction pathway. Advanced analytical techniques, such as mass spectrometry and nuclear magnetic resonance (NMR), are used to identify reaction intermediates and products. For data analysis and reaction mechanism correction, the experimental data is analyzed in detail and compared with experimental results. The reaction mechanism is then corrected based on the experimental results to ensure it conforms to the actual reaction process. Therefore, this embodiment analyzes the reaction mechanism and reaction path of the lithium carbonate preparation process, which is consistent with the actual reaction process and can reflect the dynamic changes in the lithium carbonate preparation process, thus improving the accuracy of the reaction kinetic model.

[0084] In some embodiments, determining the reaction pathway involved based on the reaction mechanism, the reaction site, the influencing factors, the experimental state parameters, and the chemical reaction data includes:

[0085] Based on the chemical reaction data, the reactants and products involved in the lithium carbonate preparation process are determined, and at least one intermediate and byproduct formed by the reactants and products are identified to determine the reaction participants related to the lithium carbonate preparation process.

[0086] Based on the reaction mechanism, the reaction site, and the influencing factors, a reaction pathway containing at least one reaction step related to the reaction participants is determined, and the reaction pathway records the reaction participants and their interconversion relationships.

[0087] Based on the experimental state parameters, the reaction correlation between any reaction condition and the reaction path and / or the reaction participants is determined, so that based on the reaction correlation, a target reaction path and its corresponding target reaction step are screened in the reaction path and / or the reaction step, wherein the reaction conditions include at least one of temperature, pressure, pH value and solvent.

[0088] The feasibility of each of the target reaction steps is evaluated using a reaction evaluation strategy, which is based on preset thermodynamic and kinetic principles, in order to verify the target reaction pathway.

[0089] In this embodiment, the reactants and products involved in the lithium carbonate preparation process are determined. For example, the main reactants in lithium carbonate preparation may include a lithium source (such as lithium salt) and a carbonate source. Potential intermediates and byproducts are also identified, which have a certain impact on the purity and yield of the final product. This determines the reaction participants in the lithium carbonate preparation process. Subsequently, based on the reaction mechanism, reaction sites, influencing factors, and experimental data, possible reaction steps are designed. These steps include all reaction participants and their interconversion relationships, organizing them into a reaction pathway. Furthermore, the effects of different reaction conditions (such as temperature, pressure, pH, solvent, etc.) on the reaction pathway and / or reaction participants are analyzed to determine the corresponding reaction correlations. Based on these correlations, the target reaction pathway and its reaction steps are determined. Further, a reaction evaluation strategy incorporating thermodynamic and kinetic principles is used to assess the feasibility of each reaction step. This strategy includes the spontaneity and rate of the reaction, enabling thermodynamic and kinetic analysis of the reaction pathway. This simulates the actual preparation process and improves the accuracy of the reaction kinetic model. Finally, chemical simulation software is used to simulate the reaction pathway to further verify the feasibility and efficiency of the reaction. Therefore, by constructing a reaction pathway and verifying its feasibility, this embodiment can adapt to the dynamic changes in the preparation process, thereby improving the accuracy and reliability of the reaction kinetic model.

[0090] For example, in the preparation of lithium carbonate, a possible reaction pathway includes: Step 1, lithium salt dissolution: Assuming lithium chloride is used as the lithium source, lithium chloride first dissolves in water, generating lithium ions and chloride ions; Step 2, carbonation reaction: Next, the dissolved lithium ions react with a carbonic acid source (such as carbon dioxide) to form lithium carbonate. This step may include the dissolution of CO2, its reaction with water to form carbonic acid, and further reactions between carbonic acid and lithium ions; Step 3, crystallization and separation: Under appropriate temperature and pressure, lithium carbonate crystallizes from the solution and is then separated from the mother liquor by methods such as filtration or centrifugation; Step 4, drying and purification: The separated lithium carbonate is dried and, if necessary, further purified by methods such as recrystallization. Therefore, in this example, each reaction step needs to be considered from a thermodynamic and kinetic perspective to ensure the efficiency and economy of the entire process. Through detailed study and optimization of the above reaction pathway, the efficiency and product purity of the lithium carbonate preparation process can be improved.

[0091] For example, in the preparation of lithium carbonate, it is assumed that the reaction mainly involves the reaction between lithium salt and carbonate source. One reaction pathway is that the lithium salt first forms an intermediate under the action of carbonate source, and then is converted into lithium carbonate. The specific reaction steps include: Step 1, lithium salt dissolution: The lithium salt dissolves in a solvent to form lithium ions and corresponding anions; Step 2, carbonate source reaction: The carbonate source reacts with the dissolved lithium ions to form a precursor of lithium carbonate; Step 3, crystallization and precipitation: Under appropriate conditions, the lithium carbonate precursor is further converted into lithium carbonate, and pure lithium carbonate is obtained through crystallization and precipitation. Each reaction step needs to be studied in detail through experiments and theoretical analysis to ensure the accuracy of the reaction mechanism and the optimization of the reaction process.

[0092] In some embodiments, determining the model parameters based on the experimental state parameters includes:

[0093] The initial values ​​of the model parameters are obtained by using the pre-established parameter estimation model and the experimental state parameters.

[0094] The estimated values ​​of the model parameters are optimized using a preset optimization model and the initial values ​​of the model parameters.

[0095] The optimized estimates are applied to the kinetic model to predict the reaction results under any reaction condition. The reaction results are then compared and verified with preset experimental verification data to adjust the estimated values ​​of the model parameters.

[0096] In this embodiment, experimental state parameters (such as temperature, pressure, concentration, reaction time, etc.) are cleaned to remove abnormal and inconsistent records, and the data is formatted to suit model requirements. Then, based on the data type and the model requirements of the reaction kinetic model, a parameter estimation model (i.e., parameter estimation method, such as least squares method, maximum likelihood estimation, etc.) is selected to perform preliminary estimation of the model parameters of the reaction kinetic model, obtaining initial values ​​for model parameters such as the reaction rate constant and activation energy. Further, an optimization model (i.e., optimization algorithm, such as gradient descent, genetic algorithm, etc.) is used to refine the parameter estimation to improve the accuracy and reliability of the reaction kinetic model. Furthermore, cross-validation is used to evaluate the accuracy of the parameter estimation and prevent overfitting. Even further, the optimized model parameter estimates are applied to the kinetic model for simulation and prediction. Based on the comparison between the model predictions and the experimental state parameters, the model parameters of the reaction kinetic model are further adjusted and optimized. Therefore, this embodiment estimates the model parameters of the reaction kinetic model to fit the actual situation of the preparation process, and continuously adjusts and optimizes the model parameters, so that the reaction kinetic model has a wide range of applicability and improves the stability and flexibility of the model.

[0097] For example, in the lithium carbonate preparation process, assuming the key reaction step is the reaction between lithium salt and carbonic acid source, the following steps are taken: First, experimental data is collected: a laboratory-scale lithium carbonate preparation experiment is conducted, recording data such as reactant concentration, temperature, and pressure. Second, preliminary parameter estimation is performed: the reaction rate constant and activation energy are estimated based on the experimental data using the least squares method. Third, parameter optimization is performed: the estimated value of the reaction rate constant is refined using advanced optimization methods such as genetic algorithms to more accurately simulate the reaction process. Fourth, model validation is performed: the optimized parameters are applied to the kinetic model to predict the reaction results under different conditions, and the results are compared and validated with additional experimental data. Through these steps, the model parameters are accurately estimated for the reaction kinetic model in the lithium carbonate preparation process, thereby improving the model's predictive ability and its practicality in guiding production.

[0098] In one alternative embodiment, experimental validation data for model validation is determined from experimental data, such that the reaction dynamics model is validated using the experimental validation data to ensure its stability and reliability.

[0099] In some embodiments, the method further includes:

[0100] Analytical parameters for sensitivity analysis are selected from the model parameters of the reaction kinetic model, and the range of variation for each analytical parameter is determined.

[0101] A preset sensitivity analysis strategy is used to simulate the analysis parameters of the reaction dynamics model within the range of variation, so as to output the parameter analysis results. The sensitivity analysis strategy is used to indicate whether to perform local sensitivity analysis or global sensitivity analysis. The parameter analysis results include the degree of influence of the parameter changes of the analysis parameters on the output of the reaction dynamics model.

[0102] Based on the parameter analysis results, the analytical parameters corresponding to the maximization of the influence degree are selected, and the influence reasons of the selected analytical parameters are determined based on the reaction mechanism, so that the sensitivity analysis results are composed of the analytical parameters and their influence reasons;

[0103] The reaction dynamics model was optimized based on the sensitivity analysis results.

[0104] In this embodiment, analytical parameters used for sensitivity analysis, such as reaction rate constant, activation energy, and concentration, are determined in the model parameters. Based on experimental data and the reaction mechanism, the range of variation for each analytical parameter is determined. Then, the sensitivity analysis strategy, such as local sensitivity analysis or global sensitivity analysis, is determined. Simulations are performed within the range of parameter variation according to this strategy to analyze the impact of parameter changes on the output of the reaction kinetic model, obtaining parameter analysis results that include the degree of influence. Furthermore, based on the parameter analysis results, the analytical parameters with the greatest impact on the output of the reaction kinetic model are identified, and the reasons for the influence of the identified analytical parameters on the output of the reaction kinetic model are analyzed, and the designed reaction mechanism is determined. Further, the reaction kinetic model is adjusted based on the sensitivity analysis results to improve its accuracy and reliability. The results of the sensitivity analysis are also used to guide future experimental designs, such as the focus of parameter control and measurement. Therefore, this embodiment, by performing sensitivity analysis on the reaction kinetic model, facilitates the identification of key parameters in the lithium carbonate preparation process, thereby achieving efficient parameter control, optimizing the preparation process, and effectively improving yield and efficiency.

[0105] For example, in the reaction kinetic model for lithium carbonate preparation, assuming the goal of sensitivity analysis is to understand the influence of the reaction rate constant and activation energy on the product yield, the sensitivity analysis includes: First, selecting key parameters: determining the reaction rate constant and activation energy as analytical parameters. Second, performing sensitivity analysis: simulating the lithium carbonate yield within a set range of variation by changing the initial values ​​of the reaction rate constant and activation energy. Third, analyzing the results: identifying which analytical parameter has a greater impact on the yield, for example, finding that changes in the reaction rate constant have a more significant impact on the yield. Fourth, applying the analysis results: based on the above analysis results, optimizing the reaction kinetic model, and focusing on the precise control and measurement of the reaction rate constant in future experimental designs. Therefore, through the above sensitivity analysis, the lithium carbonate preparation process can be better understood and optimized, improving yield and efficiency.

[0106] In one optional embodiment, the model parameters of the reaction dynamics model are adjusted based on the results of sensitivity analysis to improve the model's prediction accuracy and generalization ability.

[0107] In some embodiments, the method further includes:

[0108] Based on the optimized reaction kinetics model, a preparation simulation model was established;

[0109] Determine the initial simulation conditions, which include the initial material concentration, initial temperature, and initial pressure;

[0110] The preparation simulation model is run using the initial simulation conditions, and simulation results are obtained. The simulation results include parameter changes for the simulation parameters obtained by screening the model parameters, and the simulation parameters include reactant concentration parameters and product concentration parameters.

[0111] The simulation results are evaluated to determine the simulation evaluation results, so that the optimized reaction kinetic model can be adjusted based on the simulation evaluation results. The simulation evaluation results include the simulated reaction progress and the simulated product yield.

[0112] Risk assessment and safety analysis are conducted based on the simulation results to obtain control optimization data for the lithium carbonate preparation process. The control optimization data includes operation adjustment data and material ratio adjustment data.

[0113] In this embodiment, a mathematical model for simulation, namely a preparation simulation model, is constructed based on the reaction kinetic model processed in the above steps. This model ensures that it accurately reflects key factors in the actual reaction process, such as reaction rate, temperature, pressure, and concentration. Furthermore, the model parameters of the preparation simulation model are set, including estimated reaction rate constants, activation energy, and other parameters. Considering actual operating conditions, such as equipment characteristics and operational limitations, the preparation simulation model is appropriately adjusted. Subsequently, initial simulation conditions, such as initial material concentration, temperature, and pressure, are determined to conform to the actual production environment. This allows the preparation simulation model to be run using computer software to simulate the entire lithium carbonate preparation process. During this process, the changes in simulation parameters, such as reactant concentration, product concentration, temperature, and pressure, are monitored and recorded. Further, the simulation results are analyzed to assess the reaction progress and product yield, and compared with experimental data or known results to verify the accuracy of the simulation. If the simulation results deviate significantly from expectations, the causes of the deviation are diagnosed, such as inaccurate model assumptions or errors in parameter estimation. Based on the simulation evaluation results, the model is adjusted and optimized to improve the accuracy and practicality of the simulation. Optionally, based on the simulation results, optimization suggestions for the lithium carbonate preparation process can be proposed, such as changing operating conditions and adjusting raw material ratios. Furthermore, the simulation results can be used for risk assessment and safety analysis to ensure the safety and stability of the production process. In addition, the simulation results can be used as technical support to guide decision-making and operations in actual production, providing theoretical basis and data support for future production expansion and technological upgrades. Moreover, the preparation state parameters in the preparation process can be adjusted based on the simulation results of the preparation simulation model to improve efficiency and yield.

[0114] Therefore, this embodiment simulates the lithium carbonate preparation process indicated by the reaction kinetic model by constructing a preparation simulation model, taking into account the actual situation of the lithium carbonate preparation process, and further improving the practicality and reliability of the reaction kinetic model.

[0115] In some embodiments, constructing a reaction control model for indicative control of the lithium carbonate preparation process based on the experimental state parameters and the experimental control data includes:

[0116] The experimental control data is used as training data, which includes operating condition data, process parameters, and product quality data.

[0117] The preset neural network architecture is trained using the training data, and then the model is validated and optimized to obtain a reaction control model for indicative control of the lithium carbonate preparation process.

[0118] In this embodiment, based on the characteristics of the lithium carbonate preparation process, a suitable neural network architecture is selected, such as a convolutional neural network (CNN) for processing image data and a recurrent neural network (RNN) for processing time series data. For example, a CNN is used to process image data from lithium carbonate production, analyzing the image data during the production process to achieve real-time monitoring of the reaction state; an RNN is used to predict time series data from lithium carbonate production, predicting changes in key parameters such as temperature and pressure. Furthermore, key characteristics of the lithium carbonate preparation process (such as temperature, pressure, and concentration) are identified, and appropriate data preprocessing, such as normalization, is performed. Next, experimental control data, including operating conditions, process parameters, and product quality data, is acquired, and this experimental control data is cleaned, standardized, and normalized to adapt to the input requirements of the neural network. The architecture of the neural network, such as the number of layers, neurons, and activation function, is selected, along with a loss function and optimizer. This allows the neural network to be trained using training data determined by the experimental control data. The loss function value and evaluation index are monitored during the training process, and the training parameters are adjusted to optimize model performance. Next, model validation involves obtaining validation data from the experimental control data. Cross-validation is then used to assess the model's generalization ability, and various evaluation metrics (such as accuracy, recall, and F1 score) are used to evaluate the model's performance on the validation data. Further, the model is adjusted and optimized based on the validation results, including modifying the network architecture, adjusting the learning rate, and increasing the number of training epochs. The adjusted model is then validated again to ensure performance improvement. Therefore, through the above steps, a reaction control model is obtained. Thus, this embodiment, by constructing a reaction control model, predicts the control operations, product quality, yield, and other production process control data in the lithium carbonate preparation process, improving the accuracy of control prediction and achieving process optimization and quality control.

[0119] In one optional embodiment, the reaction control model is optimized. Specifically, the network performance of the reaction control model is optimized by adjusting hyperparameters such as the learning rate and batch size, achieving parameter tuning. Furthermore, the accuracy and stability of the reaction control model in predicting the lithium carbonate preparation process are evaluated, achieving model assessment. Therefore, the reaction control model is applied to the preparation state parameters of the real-time lithium carbonate preparation process to monitor and predict the process, and the preparation state parameters (such as temperature, pressure, pH, concentration, etc.) are adjusted based on the model prediction results, achieving process optimization and quality control.

[0120] In some embodiments, obtaining the preparation state parameters in the current lithium carbonate preparation process, such that under the joint control of the reaction kinetic model and the reaction control model with respect to the preparation state parameters, determines the action decision that maximizes the reward value determined by the production performance indicators in the lithium carbonate preparation process, in order to optimize the lithium carbonate preparation process, includes:

[0121] The current preparation state parameters in the lithium carbonate preparation process are used as the current state of a preset reinforcement learning model. The adjustment actions for each of the preparation process parameters are iterated. The reaction prediction results determined by the reaction kinetic model and the control prediction results determined by the reaction control model are combined to predict the next state of the preparation state parameters in the subsequent lithium carbonate preparation process. The production performance indicators after the adjustment actions are executed are also predicted, wherein the production performance indicators include yield, purity and energy consumption.

[0122] Based on the achievement status obtained by comparing the production performance indicators with the preset achievement threshold, the reward value of the reinforcement learning model is determined, and the action decision that maximizes the reward value is determined.

[0123] The learning strategy of the reinforcement learning model is updated based on the determined action decision to optimize the lithium carbonate preparation process according to the learning strategy.

[0124] In this embodiment, a reinforcement learning model is used to optimize the lithium carbonate production process. Specifically, the reinforcement learning environment in the model is determined, including states, action decisions, and a reward mechanism, enabling the agent in the model to learn and make decisions within this environment. The reinforcement learning algorithm used in the model can be Q-learning, Deep Q-Network (DQN), or a policy gradient method. Regarding the reward mechanism, production performance indicators in the lithium carbonate preparation process, such as yield, purity, and energy consumption, are determined. The logic for rewarding or penalizing (i.e., the reward value) is determined based on the achievement of these performance indicators; for example, a positive reward is given when the yield reaches a predetermined threshold, and a penalty is imposed when energy consumption is too high.

[0125] Optionally, the agent (i.e., DQN, Deep Q-Network) combines the advantages of deep learning and Q-learning, approximating the optimal Q-function through a neural network. The DQN neural network architecture is designed, including an input layer (representing the state of the reinforcement learning environment), hidden layers, and an output layer (representing the reward value corresponding to each action decision). Specifically, to train this DQN model, data on state, action decisions, and rewards are collected through simulation or real-world operations. This collected data is used to train the DQN model, and the neural network learns the optimal policy through repeated iterations. Subsequently, the performance of the DQN model is tested in simulated environments or real-world operations to evaluate its decision-making ability under different states. Based on the evaluation results, the DQN model is adjusted, such as modifying the network structure, adjusting the reward logic, or the learning rate.

[0126] Furthermore, the current state parameters of the lithium carbonate preparation process are input into the reinforcement learning model to iterate through the adjustment actions for these state parameters. The model then combines the reaction kinetics model and the reaction control model to predict the next state, thereby obtaining the production performance indicators and corresponding reward values ​​after executing any adjustment action. Specifically, the current state parameters of the lithium carbonate preparation process are input into the reaction kinetics model for chemical reaction prediction, obtaining reaction prediction results. This takes into account the complex reaction kinetics in the lithium carbonate preparation process, which is beneficial for improving production efficiency and yield. Simultaneously, the current state parameters of the lithium carbonate preparation process are input into the reaction control model for process control prediction, obtaining control prediction results, achieving process optimization and quality control. Thus, in the action decisions for entering the next state, the action decisions that maximize the reward value (i.e., the optimal action decision, such as adjusting reaction conditions) are selected. The learning strategy of the reinforcement learning model is optimized based on the selected action decisions, thereby optimizing the reinforcement learning model and the lithium carbonate preparation process. Optionally, the performance of the reinforcement learning model in the lithium carbonate production process can be periodically evaluated, such as for yield and quality improvements, to achieve performance assessment. Based on the performance assessment results, the reinforcement learning model can be adjusted, such as by improving the reward function or adjusting the learning rate. Subsequently, the reinforcement learning model can be integrated with the existing production control system to achieve automated and intelligent control. The reinforcement learning system can be run during long-term production, continuously monitoring its performance to ensure stable and efficient operation. For example, the production strategy for lithium carbonate production can be optimized using reinforcement learning algorithms, adjusting operating parameters such as raw material input and reaction time in real time to optimize yield and quality.

[0127] Therefore, this disclosure combines reaction kinetics models, reaction control models, and reinforcement learning models to predict and optimize the process, ensuring that the lithium carbonate preparation process meets the predetermined functional and performance requirements, and guaranteeing the efficiency and correctness of the lithium carbonate preparation process.

[0128] In an optional embodiment, the reaction dynamics model, reaction control model, and reinforcement learning model are integrated into a unified system. Specifically, a comprehensive system architecture incorporating the reaction dynamics model, reaction control model, and reinforcement learning model is designed, ensuring modular design for seamless integration of components. Then, interfaces between modules are developed to enable efficient data and command transmission, ensuring the correctness and efficiency of the system's data flow, control flow, and error handling mechanisms. Furthermore, the integrated system undergoes comprehensive testing, including unit testing, integration testing, and system testing, to verify its performance and ensure it meets predetermined functional and performance requirements.

[0129] In another alternative embodiment, the system is deployed in a factory environment for real-time monitoring and adjustment. Specifically, the hardware and software requirements of the deployment environment are determined, and the necessary hardware and software facilities are installed in the actual production environment. Then, the developed system is deployed to the factory's production environment, and system parameters are configured to adapt to actual production conditions and requirements. Furthermore, the system's operating status is monitored in real time, operational data is collected and analyzed, and the system is continuously optimized and adjusted based on its operating status and the collected data. Even further, the system is regularly maintained to ensure stable operation, and is upgraded and improved according to the latest technological advancements and production needs.

[0130] Example 2

[0131] Figure 2 This is a schematic diagram of a reaction kinetics model-driven optimization system for lithium carbonate preparation, provided in an embodiment of this disclosure. The reaction kinetics model-driven optimization system 200 includes:

[0132] Data acquisition module 201 is used to acquire experimental data in the lithium carbonate preparation experiment, including chemical reaction data, experimental state parameters and experimental control data;

[0133] The reaction kinetic model construction module 202 is used to analyze the chemical reaction characteristics of the chemical reaction data and the experimental state parameters in order to construct a reaction kinetic model to indicate the reaction kinetic behavior in the lithium carbonate preparation process.

[0134] The reaction control model construction module 203 is used to construct a reaction control model for indicating and controlling the lithium carbonate preparation process based on the experimental state parameters and the experimental control data.

[0135] The optimization module 204 is used to acquire the preparation state parameters in the current lithium carbonate preparation process, so that under the joint control of the reaction kinetic model and the reaction control model with respect to the preparation state parameters, it determines the action decision that maximizes the reward value determined by the production performance indicators in the lithium carbonate preparation process, so as to optimize the lithium carbonate preparation process. The action decision includes adjustment actions for the preparation process parameters.

[0136] In some embodiments, the reaction dynamics model building module 202 includes:

[0137] The reaction mechanism determination unit is used to analyze the chemical reaction characteristics of the chemical reaction data and the experimental state parameters in order to determine the reaction mechanism of the lithium carbonate preparation process.

[0138] The model parameter determination unit is used to determine model parameters based on the experimental state parameters, wherein the model parameters include reaction rate constant and activation energy parameters;

[0139] The reaction kinetic model building unit is used to apply the reaction mechanism and the model parameters to a preset kinetic model and to verify the kinetic model in order to obtain a reaction kinetic model.

[0140] In some embodiments, the reaction mechanism determination unit includes:

[0141] The reaction mechanism acquisition unit is used to determine the reaction mechanism related to the preparation of lithium carbonate based on the chemical reaction data.

[0142] The chemical reaction characteristic analysis unit is used to analyze the chemical structure of the chemical reaction data based on the reaction mechanism and the experimental state parameters, so as to determine the corresponding reaction sites and the influencing factors that affect the reaction sites.

[0143] The reaction pathway determination unit is used to determine the reaction pathways involved based on the reaction mechanism, the reaction sites, the influencing factors, the experimental state parameters, and the chemical reaction data.

[0144] The reaction mechanism verification unit is used to verify the correctness of the reaction pathway, so as to verify and correct the reaction mechanism.

[0145] In some embodiments, the reaction pathway determination unit includes:

[0146] A reaction participant identification unit is used to determine the reactants and products involved in the lithium carbonate preparation process based on the chemical reaction data, and to identify at least one intermediate and byproduct formed by the reactants and products, so as to determine the reaction participants related to the lithium carbonate preparation process.

[0147] A reaction pathway generation unit is used to determine, based on the reaction mechanism, the reaction site, and the influencing factors, a reaction pathway containing at least one reaction step related to the reaction participants, wherein the reaction pathway records the reaction participants and their interconversion relationships.

[0148] The target reaction path generation unit is used to determine the reaction correlation between any reaction condition and the reaction path and / or the reaction participants based on the experimental state parameters, so that a target reaction path and its corresponding target reaction step are screened out from the reaction path and / or the reaction steps based on the reaction correlation, wherein the reaction conditions include at least one of temperature, pressure, pH value and solvent.

[0149] A reaction evaluation unit is used to evaluate the feasibility of each of the target reaction steps using a reaction evaluation strategy to verify the target reaction path. The reaction evaluation strategy is associated with preset thermodynamic and kinetic principles.

[0150] In some embodiments, the model parameter determination unit includes:

[0151] The model parameter initialization unit is used to perform preliminary estimation using a pre-established parameter estimation model and the experimental state parameters to obtain initial values ​​for the model parameters.

[0152] The model parameter optimization unit is used to optimize the estimated values ​​of the model parameters using a preset optimization model and the initial values ​​of the model parameters.

[0153] The model parameter adjustment unit is used to apply the optimized estimated values ​​to the kinetic model to predict the reaction results under any reaction conditions, so that the reaction results are compared and verified with preset experimental verification data to adjust the estimated values ​​of the model parameters.

[0154] In some embodiments, the reaction dynamics model building module 202 further includes:

[0155] The analytical parameter determination unit is used to select analytical parameters for sensitivity analysis from the model parameters of the reaction kinetic model and determine the range of variation of each analytical parameter.

[0156] A sensitivity analysis unit is used to simulate the analysis parameters of the reaction dynamics model within the range of variation using a preset sensitivity analysis strategy, so as to output parameter analysis results. The sensitivity analysis strategy is used to indicate whether to perform local sensitivity analysis or global sensitivity analysis. The parameter analysis results include the degree of influence of the parameter changes of the analysis parameters on the output of the reaction dynamics model.

[0157] An analysis unit is used to select the analysis parameters corresponding to the maximization of the influence degree based on the parameter analysis results, and to determine the influence reasons of the selected analysis parameters based on the reaction mechanism, so that the analysis parameters and their influence reasons constitute the sensitivity analysis results;

[0158] The reaction dynamics model optimization unit is used to optimize the reaction dynamics model based on the sensitivity analysis results.

[0159] In some embodiments, the reaction dynamics model building module 202 further includes:

[0160] The simulation model building unit is used to build a preparation simulation model based on the optimized reaction kinetic model;

[0161] An initial simulation condition determination unit is used to determine the initial simulation conditions, which include the initial material concentration, initial temperature, and initial pressure.

[0162] The simulation unit is used to run the preparation simulation model using the initial simulation conditions and obtain simulation results. The simulation results include parameter change results for the simulation parameters obtained by screening the model parameters. The simulation parameters include reactant concentration parameters and product concentration parameters.

[0163] The simulation result evaluation unit is used to evaluate the simulation results to determine the simulation evaluation results, so that the optimized reaction kinetic model can be adjusted based on the simulation evaluation results. The simulation evaluation results include the simulated reaction progress and the simulated product yield.

[0164] The control optimization unit is used to perform risk assessment and safety analysis based on the simulation results to obtain control optimization data on the lithium carbonate preparation process. The control optimization data includes operation adjustment data and material ratio adjustment data.

[0165] In some embodiments, the reaction control model construction module 203 includes:

[0166] The training data determination unit is used to use the experimental control data as training data, wherein the experimental control data includes operating condition data, process parameters, and product quality data.

[0167] The model training unit is used to train a preset neural network architecture using the training data, and then perform model verification and optimization to obtain a reaction control model for instructing and controlling the lithium carbonate preparation process.

[0168] In some embodiments, the optimization module 204 includes:

[0169] The reinforcement learning unit is used to take the preparation state parameters in the current lithium carbonate preparation process as the current state of the preset reinforcement learning model, traverse the adjustment actions for each of the preparation process parameters, and combine the reaction prediction results determined by the reaction kinetic model and the control prediction results determined by the reaction control model to predict the next state of the preparation state parameters in the subsequent lithium carbonate preparation process, and predict the production performance indicators after the adjustment actions are executed, wherein the production performance indicators include yield, purity and energy consumption.

[0170] The reward value determination unit is used to determine the reward value of the reinforcement learning model based on the achievement status obtained by comparing the production performance index with the preset achievement threshold, and to determine the action decision that maximizes the reward value.

[0171] The lithium carbonate preparation process optimization unit is used to update the learning strategy of the reinforcement learning model based on the determined action decisions, so as to optimize the lithium carbonate preparation process according to the learning strategy.

[0172] The system of this disclosure embodiment can execute the method provided in this disclosure embodiment, and the implementation principle is similar. The actions performed by each module in the system of each disclosure embodiment correspond to the steps in the method of each disclosure embodiment. For detailed functional descriptions of each module of the system, please refer to the descriptions in the corresponding methods shown above, which will not be repeated here.

[0173] The above description is only an optional implementation method for some implementation scenarios of this disclosure. It should be noted that for those skilled in the art, other similar implementation methods based on the technical concept of this disclosure without departing from the technical concept of this disclosure also fall within the protection scope of the embodiments of this disclosure.

Claims

1. A method for optimizing the lithium carbonate preparation process driven by a reaction kinetic model, characterized in that, include: Obtain experimental data from the lithium carbonate preparation experiment, including chemical reaction data, experimental state parameters, and experimental control data; Chemical reaction characteristics were analyzed based on the chemical reaction data and experimental state parameters to construct a reaction kinetic model to indicate the reaction kinetics behavior in the lithium carbonate preparation process; Based on the experimental state parameters and the experimental control data, a reaction control model for indicating and controlling the lithium carbonate preparation process is constructed. The preparation state parameters of the current lithium carbonate preparation process are obtained, and under the joint control of the reaction kinetic model and the reaction control model with respect to the preparation state parameters, action decisions are determined to maximize the reward value determined by the production performance indicators in the lithium carbonate preparation process, so as to optimize the lithium carbonate preparation process. The action decisions include adjustment actions for the preparation process parameters. The step of performing chemical reaction characteristic analysis on the chemical reaction data and experimental state parameters to construct a reaction kinetic model for indicating the reaction kinetics behavior in the lithium carbonate preparation process includes: Chemical reaction characteristics were analyzed based on the chemical reaction data and experimental state parameters to determine the reaction mechanism related to the lithium carbonate preparation process; Based on the experimental state parameters, the model parameters are determined, including the reaction rate constant and activation energy parameters. The reaction mechanism and the model parameters are applied to a preset kinetic model, and the kinetic model is verified to obtain a reaction kinetic model. The method further includes: Analytical parameters for sensitivity analysis are selected from the model parameters of the reaction kinetic model, and the range of variation for each analytical parameter is determined. A preset sensitivity analysis strategy is used to simulate the analysis parameters of the reaction dynamics model within the range of variation, so as to output the parameter analysis results. The sensitivity analysis strategy is used to indicate whether to perform local sensitivity analysis or global sensitivity analysis. The parameter analysis results include the degree of influence of the parameter changes of the analysis parameters on the output of the reaction dynamics model. Based on the parameter analysis results, the analytical parameters corresponding to the maximization of the influence degree are selected, and the influence reasons of the selected analytical parameters are determined based on the reaction mechanism, so that the sensitivity analysis results are composed of the analytical parameters and their influence reasons; The reaction kinetic model was optimized based on the sensitivity analysis results. The step of acquiring the preparation state parameters in the current lithium carbonate preparation process, and determining the action decision that maximizes the reward value determined by the production performance indicators in the lithium carbonate preparation process under the joint control of the reaction kinetic model and the reaction control model with respect to the preparation state parameters, in order to optimize the lithium carbonate preparation process, includes: The current preparation state parameters in the lithium carbonate preparation process are used as the current state of a preset reinforcement learning model. The adjustment actions for each of the preparation process parameters are iterated. The reaction prediction results determined by the reaction kinetic model and the control prediction results determined by the reaction control model are combined to predict the next state of the preparation state parameters in the subsequent lithium carbonate preparation process. The production performance indicators after the adjustment actions are executed are also predicted, wherein the production performance indicators include yield, purity and energy consumption. Based on the achievement status obtained by comparing the production performance indicators with the preset achievement threshold, the reward value of the reinforcement learning model is determined, and the action decision that maximizes the reward value is determined. The learning strategy of the reinforcement learning model is updated based on the determined action decision to optimize the lithium carbonate preparation process according to the learning strategy.

2. The method for optimizing the lithium carbonate preparation process driven by the reaction kinetics model according to claim 1, characterized in that, The analysis of chemical reaction characteristics of the chemical reaction data and experimental state parameters to determine the reaction mechanism of the lithium carbonate preparation process includes: Based on the aforementioned chemical reaction data, the reaction mechanism related to the preparation of lithium carbonate was determined; Based on the reaction mechanism and the experimental state parameters, the chemical structure of the chemical reaction data is analyzed to determine the corresponding reaction sites and the influencing factors that affect the reaction sites. Based on the reaction mechanism, the reaction sites, the influencing factors, the experimental state parameters, and the chemical reaction data, the reaction pathways involved are determined. The correctness of the reaction pathway is verified in order to verify and correct the reaction mechanism.

3. The method for optimizing the lithium carbonate preparation process driven by the reaction kinetics model according to claim 2, characterized in that, The determination of the relevant reaction pathways based on the reaction mechanism, reaction sites, influencing factors, experimental state parameters, and chemical reaction data includes: Based on the chemical reaction data, the reactants and products involved in the lithium carbonate preparation process are determined, and at least one intermediate and byproduct formed by the reactants and products are identified to determine the reaction participants related to the lithium carbonate preparation process. Based on the reaction mechanism, the reaction site, and the influencing factors, a reaction pathway containing at least one reaction step related to the reaction participants is determined, and the reaction pathway records the reaction participants and their interconversion relationships. Based on the experimental state parameters, the reaction correlation between any reaction condition and the reaction path and / or the reaction participants is determined, so that based on the reaction correlation, a target reaction path and its corresponding target reaction step are screened in the reaction path and / or the reaction step, wherein the reaction conditions include at least one of temperature, pressure, pH value and solvent. The feasibility of each of the target reaction steps is evaluated using a reaction evaluation strategy, which is based on preset thermodynamic and kinetic principles, in order to verify the target reaction pathway.

4. The method for optimizing the lithium carbonate preparation process driven by the reaction kinetics model according to claim 3, characterized in that, The process of determining model parameters based on the experimental state parameters includes: The initial values ​​of the model parameters are obtained by using the pre-established parameter estimation model and the experimental state parameters. The estimated values ​​of the model parameters are optimized using a preset optimization model and the initial values ​​of the model parameters. The optimized estimates are applied to the kinetic model to predict the reaction results under any reaction condition. The reaction results are then compared and verified with preset experimental verification data to adjust the estimated values ​​of the model parameters.

5. The method for optimizing the lithium carbonate preparation process driven by the reaction kinetics model according to claim 4, characterized in that, The method further includes: Based on the optimized reaction kinetics model, a preparation simulation model was established; Determine the initial simulation conditions, which include the initial material concentration, initial temperature, and initial pressure; The preparation simulation model is run using the initial simulation conditions, and simulation results are obtained. The simulation results include parameter changes for the simulation parameters obtained by screening the model parameters, and the simulation parameters include reactant concentration parameters and product concentration parameters. The simulation results are evaluated to determine the simulation evaluation results, so that the optimized reaction kinetic model can be adjusted based on the simulation evaluation results. The simulation evaluation results include the simulated reaction progress and the simulated product yield. Risk assessment and safety analysis are conducted based on the simulation results to obtain control optimization data for the lithium carbonate preparation process. The control optimization data includes operation adjustment data and material ratio adjustment data.

6. The method for optimizing the lithium carbonate preparation process driven by the reaction kinetics model according to claim 5, characterized in that, The process of constructing a reaction control model for indicative control of the lithium carbonate preparation process based on the experimental state parameters and the experimental control data includes: The experimental control data is used as training data, which includes operating condition data, process parameters, and product quality data. The preset neural network architecture is trained using the training data, and then the model is validated and optimized to obtain a reaction control model for indicative control of the lithium carbonate preparation process.

7. A reaction kinetics model-driven optimization system for lithium carbonate preparation, characterized in that, include: The data acquisition module is used to acquire experimental data in the lithium carbonate preparation experiment, including chemical reaction data, experimental state parameters, and experimental control data. The reaction kinetic model construction module is used to analyze the chemical reaction characteristics of the chemical reaction data and the experimental state parameters in order to construct a reaction kinetic model to indicate the reaction kinetic behavior in the lithium carbonate preparation process. The reaction control model construction module is used to construct a reaction control model for indicating and controlling the lithium carbonate preparation process based on the experimental state parameters and the experimental control data. An optimization module is used to acquire the preparation state parameters in the current lithium carbonate preparation process, so that under the joint control of the reaction kinetic model and the reaction control model with respect to the preparation state parameters, an action decision is determined to maximize the reward value determined by the production performance indicators in the lithium carbonate preparation process, so as to optimize the lithium carbonate preparation process. The action decision includes adjustment actions for the preparation process parameters. The step of performing chemical reaction characteristic analysis on the chemical reaction data and experimental state parameters to construct a reaction kinetic model for indicating the reaction kinetics behavior in the lithium carbonate preparation process includes: Chemical reaction characteristics were analyzed based on the chemical reaction data and experimental state parameters to determine the reaction mechanism related to the lithium carbonate preparation process; Based on the experimental state parameters, the model parameters are determined, including the reaction rate constant and activation energy parameters. The reaction mechanism and the model parameters are applied to a preset kinetic model, and the kinetic model is verified to obtain a reaction kinetic model. The system is also used for: Analytical parameters for sensitivity analysis are selected from the model parameters of the reaction kinetic model, and the range of variation for each analytical parameter is determined. A preset sensitivity analysis strategy is used to simulate the analysis parameters of the reaction dynamics model within the range of variation, so as to output the parameter analysis results. The sensitivity analysis strategy is used to indicate whether to perform local sensitivity analysis or global sensitivity analysis. The parameter analysis results include the degree of influence of the parameter changes of the analysis parameters on the output of the reaction dynamics model. Based on the parameter analysis results, the analytical parameters corresponding to the maximization of the influence degree are selected, and the influence reasons of the selected analytical parameters are determined based on the reaction mechanism, so that the sensitivity analysis results are composed of the analytical parameters and their influence reasons; The reaction kinetic model was optimized based on the sensitivity analysis results. The step of acquiring the preparation state parameters in the current lithium carbonate preparation process, and determining the action decision that maximizes the reward value determined by the production performance indicators in the lithium carbonate preparation process under the joint control of the reaction kinetic model and the reaction control model with respect to the preparation state parameters, in order to optimize the lithium carbonate preparation process, includes: The current preparation state parameters in the lithium carbonate preparation process are used as the current state of a preset reinforcement learning model. The adjustment actions for each of the preparation process parameters are iterated. The reaction prediction results determined by the reaction kinetic model and the control prediction results determined by the reaction control model are combined to predict the next state of the preparation state parameters in the subsequent lithium carbonate preparation process. The production performance indicators after the adjustment actions are executed are also predicted, wherein the production performance indicators include yield, purity and energy consumption. Based on the achievement status obtained by comparing the production performance indicators with the preset achievement threshold, the reward value of the reinforcement learning model is determined, and the action decision that maximizes the reward value is determined. The learning strategy of the reinforcement learning model is updated based on the determined action decision to optimize the lithium carbonate preparation process according to the learning strategy.

Images