Intelligent regulation and control method and system for neutralizing organic acid by using sodium methoxide solution

Abstract

The invention provides an intelligent regulation and control method and system for neutralizing organic acid by using a sodium methoxide solution, and relates to the technical field of neutralization regulation and control, and the method comprises the following steps: acquiring methanol solution flow data, organic acid flow data and organic acid reaction data through a sensor module, carrying out key feature extraction and purity prediction fitting, and calculating the purity of the organic acid; building a product purity prediction channel; performing reaction prediction by adopting a product purity prediction channel to obtain a product purity prediction parameter; and constructing an organic acid segmented regulation and control strategy, performing segmented regulation and control analysis on the product purity prediction parameters based on the organic acid segmented regulation and control strategy, determining target methanol regulation and control strategy parameters, and performing organic acid neutralization regulation and control based on the target methanol regulation and control strategy parameters. The technical problem that in the prior art, the adjustment precision is insufficient in the methanol adding process can be solved, the technical target of dynamic methanol regulation and control strategy construction is achieved, and the technical effect of enhancing the regulation and control flexibility of all stages is achieved.

Description

Technical Field

[0001] This application relates to the field of neutralization and regulation technology, and in particular to an intelligent regulation method and system for neutralizing organic acids using sodium methoxide solution. Background Technology

[0002] In the field of modern fine chemical and pharmaceutical synthesis, the neutralization process of organic acids is a key reaction step to achieve high-purity target products. In particular, when methanol is used as a neutralizing agent, the precise control of the reaction process is especially important.

[0003] Currently, neutralization reaction control mainly relies on traditional single-variable or simple proportional feedback mechanisms, such as automatic feeding by setting a fixed methanol addition rate or using a threshold control based on the product pH value. However, these control methods generally suffer from problems such as response lag, strong adjustment rigidity, and poor stage adaptability, failing to adequately address the complexity and dynamic changes at different stages of the reaction. Especially in the initial, middle, and late stages of the reaction, due to significant differences in factors such as organic acid concentration, temperature, and reaction rate, fixed strategies often lead to insufficient neutralization in the initial stage, slow regulation in the middle stage, or excessive methanol in the later stage, resulting in fluctuations in product purity, increased energy consumption, and even the formation of side reactions, making it difficult to meet the industrial requirements for efficient, high-purity, and highly consistent processes.

[0004] In summary, existing technologies suffer from technical problems due to the lack of identification of reaction stage characteristics and segmented adaptive strategy design in the control methods. This results in lag in the response of the methanol dosing process, insufficient adjustment precision, and further affects the purity and stability of the product, energy consumption control, and the overall process efficiency of the neutralization process. Summary of the Invention

[0005] The purpose of this application is to provide an intelligent control method and system for neutralizing organic acids using sodium methoxide solution, in order to solve the technical problems in the prior art where the control method lacks the identification of the characteristics of the reaction stage and the design of segmented adaptive strategies, resulting in lag in the response of the methanol addition process, insufficient control precision, and further affecting the purity and stability of the product, energy consumption control and the overall process efficiency of the neutralization process.

[0006] In view of the above problems, this application provides a smart control method and system for neutralizing organic acids using sodium methoxide solution.

[0007] In a first aspect, this application provides an intelligent control method for neutralizing organic acids using sodium methoxide solution. This method is implemented through an intelligent control system for neutralizing organic acids using sodium methoxide solution, comprising: deploying a sensor module within a reaction vessel, the sensor module integrating a pH sensor, a temperature sensor, and a liquid chromatograph; acquiring methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data through the sensor module; extracting key features and performing purity prediction fitting on the methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data to establish a product purity prediction channel; using the product purity prediction channel to predict the reaction of the methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data to obtain product purity prediction parameters; constructing a segmented control strategy for organic acids; performing segmented control analysis on the product purity prediction parameters based on the segmented control strategy for organic acids to determine target methanol control strategy parameters; and performing organic acid neutralization control based on the target methanol control strategy parameters.

[0008] Preferably, the intelligent control method for neutralizing organic acids using sodium methoxide solution further includes: determining a preset window size based on the data acquisition frequency and reaction dynamics; preprocessing the methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data based on the preset window size to obtain standard methanol solution flow rate data, standard organic acid flow rate data, and standard organic acid reaction data; extracting key features from the standard methanol solution flow rate data, standard organic acid flow rate data, and standard organic acid reaction data to obtain a key feature set of reactants; and performing historical data mining and purity prediction fitting based on the key feature set of reactants to generate a product purity prediction channel.

[0009] Preferably, the intelligent control method for neutralizing organic acids using sodium methoxide solution further includes: arranging the methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data in chronological order to obtain methanol solution flow rate sequence data, organic acid flow rate sequence data, and organic acid reaction sequence data; calculating the average value and standard deviation of the methanol solution flow rate sequence data, organic acid flow rate sequence data, and organic acid reaction sequence data according to the preset window size to determine a window data anomaly judgment threshold; performing anomaly judgment and average value replacement on the methanol solution flow rate sequence data, organic acid flow rate sequence data, and organic acid reaction sequence data based on the window data anomaly judgment threshold to obtain usable methanol solution flow rate data, usable organic acid flow rate data, and usable organic acid reaction data; and normalizing the usable methanol solution flow rate data, usable organic acid flow rate data, and usable organic acid reaction data to obtain standard methanol solution flow rate data, standard organic acid flow rate data, and standard organic acid reaction data.

[0010] Preferably, the intelligent control method for neutralizing organic acids using sodium methoxide solution further includes: extracting correlation features from the standard methanol solution flow data, standard organic acid flow data, and standard organic acid reaction data to obtain a reactant correlation feature set; calculating a correlation coefficient set between each correlation feature in the reactant correlation feature set and the product purity; setting a correlation coefficient threshold according to purity prediction requirements; and performing key feature screening on the reactant correlation feature set based on the correlation coefficient set according to the correlation coefficient threshold to obtain the reactant key feature set.

[0011] Preferably, the intelligent control method for neutralizing organic acids using sodium methoxide solution further includes: performing historical data mining based on the key feature set of the reactants to obtain a reactant feature dataset and a corresponding product purity dataset; using the reactant feature dataset as an input variable and the product purity dataset as an output variable to perform purity regression fitting and construct an initial purity prediction channel; introducing a penalty term to perform performance cross-validation and parameter iterative optimization on the initial purity prediction channel to generate the product purity prediction channel.

[0012] Preferably, the intelligent control method for neutralizing organic acids using sodium methoxide solution further includes: dividing the reaction stage into the reactant feature dataset and the corresponding product purity dataset to obtain the initial reaction stage, the middle reaction stage, and the later reaction stage; setting initial product purity targets, middle product purity targets, and later product purity targets for the initial reaction stage, the middle reaction stage, and the later reaction stage; and performing control strategy analysis based on the initial product purity targets, the middle product purity targets, and the later product purity targets to obtain a segmented control strategy for organic acids.

[0013] Preferably, the intelligent control method for neutralizing organic acids using sodium methoxide solution further includes: setting the number of clusters K, where K=3; performing K-means clustering on the reactant feature dataset and the corresponding product purity dataset according to the number of clusters K to obtain clustering results; evaluating, optimizing, and classifying the clustering results to determine the initial reaction stage, the middle reaction stage, and the later reaction stage.

[0014] Preferably, the intelligent control method for neutralizing organic acids using sodium methoxide solution further includes: initializing PID parameters according to the initial product purity target, the intermediate product purity target, and the final product purity target to obtain an initial-stage methanol controller, an intermediate-stage methanol controller, and a final-stage methanol controller; and configuring a segmented strategy based on the initial-stage methanol controller, the intermediate-stage methanol controller, and the final-stage methanol controller to obtain the segmented control strategy for the organic acid.

[0015] Preferably, the intelligent control method for neutralizing organic acids using sodium methoxide solution further includes: performing staged strategy matching on the product purity prediction parameters based on the organic acid segmented control strategy, triggering a target segment methanol controller; and using the target segment methanol controller to perform methanol control calculations on the product purity prediction parameters to determine the target methanol control strategy parameters.

[0016] Secondly, this application also provides an intelligent control system for neutralizing organic acids using sodium methoxide solution, used to execute the intelligent control method for neutralizing organic acids using sodium methoxide solution as described in the first aspect, comprising: a sensor acquisition module for deploying sensor modules in a reaction vessel, the sensor modules integrating a pH sensor, a temperature sensor, and a liquid chromatograph, for acquiring methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data through the sensor modules; a channel construction module for extracting key features and fitting purity predictions to the methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data, and constructing a product purity prediction channel; a reaction prediction module for using the product purity prediction channel to predict the reaction of the methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data, and obtaining product purity prediction parameters; and a neutralization control module for constructing a segmented control strategy for organic acids, performing segmented control analysis on the product purity prediction parameters based on the segmented control strategy for organic acids, determining target methanol control strategy parameters, and performing organic acid neutralization control based on the target methanol control strategy parameters.

[0017] The technical solution provided in this application has at least the following technical effects or advantages: by realizing the technical objective of constructing a dynamic methanol control strategy based on product purity prediction and reaction stage division, it achieves the technical effects of improving the neutralization control response speed, enhancing the control flexibility of each stage, stabilizing product purity, and reducing energy consumption.

[0018] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in 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 merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the intelligent control method for neutralizing organic acids using sodium methoxide solution, as described in this application.

[0021] Figure 2 This is a schematic diagram of the intelligent control system for neutralizing organic acids using sodium methoxide solution, as described in this application.

[0022] Figure labeling: 11 sensor acquisition module, 12 channel construction module, 13 reaction prediction module, 14 neutralization and regulation module. Detailed Implementation

[0023] This application provides an intelligent control method and system for neutralizing organic acids using sodium methoxide solution. This solves the technical problems in existing technologies where the lack of identification of reaction stage characteristics and segmented adaptive strategy design in the control methods leads to lag in methanol addition response, insufficient control precision, and further affects product purity stability, energy consumption control, and the overall process efficiency of the neutralization process. The application achieves the technical goal of constructing a dynamic methanol control strategy based on product purity prediction and reaction stage division, thereby improving the neutralization control response speed, enhancing the flexibility of control at each stage, stabilizing product purity, and reducing energy consumption.

[0024] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0025] Example 1, please refer to the appendix. Figure 1 This application provides a smart control method for neutralizing organic acids using sodium methoxide solution, which is applied to a smart control system for neutralizing organic acids using sodium methoxide solution. The method specifically includes the following steps: S1: A sensor module is installed inside the reactor. The sensor module integrates a pH sensor, a temperature sensor, and a liquid chromatograph. The flow rate data of methanol solution, organic acid flow rate data, and organic acid reaction data are collected through the sensor module.

[0026] Specifically, sensor modules are installed in the reaction vessel, meaning devices with monitoring capabilities are installed inside the container where the chemical reaction takes place to acquire various key data during the reaction process in real time. A sensor module is a hardware assembly containing multiple sensor devices for collecting different types of physical or chemical parameters.

[0027] The sensor module integrates a pH sensor, a temperature sensor, and a liquid chromatograph. The pH sensor is mainly used to detect the acidity or alkalinity of the solution, reflecting whether the neutralization reaction between methanol and organic acids is nearing its endpoint. The temperature sensor is used to monitor the temperature changes of the solution in real time during the reaction, as temperature directly affects the reaction rate and product distribution. The liquid chromatograph is a high-precision analytical instrument that can detect the concentration changes of various substances during the reaction, especially the content of organic acids and products. Its strong separation capability can be used to determine whether the reaction is complete.

[0028] Furthermore, the data acquired through the sensor module includes methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data. Methanol solution flow rate data refers to the volume of methanol solution entering the reactor per unit time, measured in milliliters per minute; organic acid flow rate data indicates the feed rate of the organic acid; and organic acid reaction data includes trends in organic acid concentration, dynamic changes in pH value, etc., reflecting the actual progress.

[0029] S2: Extract key features and fit purity prediction data from the methanol solution flow data, organic acid flow data, and organic acid reaction data to build a product purity prediction channel.

[0030] Specifically, key features are extracted from methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data to identify the most representative parameters that influence reaction behavior or product purity. The purpose of key feature extraction is to remove redundant information and retain only variables closely related to product changes. Purity prediction fitting involves establishing a mathematical model based on the extracted key features to simulate the relationship between input variables and reaction product purity. Regression analysis methods, such as linear regression, support vector regression, or neural networks, are used to estimate the corresponding purity output given a set of input parameters. During the fitting process, an error function evaluates the difference between the model's predicted and actual values, and the model parameters are continuously optimized to improve prediction accuracy. Ultimately, a product purity prediction channel is built, capable of quickly outputting purity estimates based on real-time reaction characteristics.

[0031] S3: The product purity prediction channel is used to predict the reaction of the methanol solution flow data, organic acid flow data, and organic acid reaction data to obtain the product purity prediction parameters.

[0032] Specifically, a product purity prediction channel is used to predict reactions based on methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data. The channel inputs and processes the collected flow rate and reaction-related data, and outputs predicted reaction results based on changes in the input variables. In other words, by inputting methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data into the product purity prediction channel, it automatically calculates and predicts the purity of the final product under the current raw material input and reaction conditions based on a pre-trained correlation model. Reaction prediction is a process of deducing reaction results from input conditions, achieved through regression analysis, neural networks, or other machine learning methods. The product purity prediction parameter is the output of the prediction process, given as a percentage, describing the proportion of pure substances in the target product. For example, predicting a product purity of 95% under current conditions means that 95% of the components in all reaction products are the target compound.

[0033] S4: Construct an organic acid segmented regulation strategy, analyze the product purity prediction parameters based on the organic acid segmented regulation strategy, determine the target methanol regulation strategy parameters, and perform organic acid neutralization regulation based on the target methanol regulation strategy parameters.

[0034] Specifically, the entire organic acid neutralization reaction process is divided into several stages, such as the initial stage, the middle stage, and the later stage. Corresponding control methods are formulated for the reaction characteristics and purity control requirements of each stage. That is, a continuous reaction process is decomposed into multiple representative segments so that different control schemes can be adopted at different stages, thereby improving the flexibility and precision of the overall control.

[0035] Next, the segmented control analysis of the product purity prediction parameters based on the organic acid segmented control strategy involves substituting the predicted product purity values ​​into the control strategy models for different stages to determine the current reaction stage and further analyze the corresponding control requirements. For example, if the predicted purity is 90% and the intermediate target is 88% to 92%, then the current stage should be intermediate. Analysis shows that a moderate methanol rate should be maintained to stabilize the reaction equilibrium. Subsequently, determining the target methanol control strategy parameters means combining the analysis results to calculate a set of numerical parameters that can be used for actual control, such as target flow rate, adjustment step size, and response time, as direct inputs to the control system to guide the actual methanol solution dosing behavior.

[0036] Finally, the neutralization of organic acids is regulated based on the target methanol control strategy parameters, thereby achieving automated control in the actual reaction. This ensures that the neutralization reaction remains within the optimal range at each stage, demonstrating a complete closed-loop chain of prediction-matching-analysis-control, and improving the timeliness and adaptability of the control.

[0037] Furthermore, this application also includes: determining a preset window size based on the data acquisition frequency and reaction dynamic characteristics; preprocessing the methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data based on the preset window size to obtain standard methanol solution flow rate data, standard organic acid flow rate data, and standard organic acid reaction data; extracting key features from the standard methanol solution flow rate data, standard organic acid flow rate data, and standard organic acid reaction data to obtain a key feature set of reactants; and performing historical data mining and purity prediction fitting based on the key feature set of reactants to generate a product purity prediction channel.

[0038] Specifically, the preset window size is determined based on the data acquisition frequency and reaction dynamics. This means selecting an appropriate time period or data segment size as the basic unit for subsequent analysis, based on the time interval between sensor data acquisitions and the rate of change of variables during the reaction. Data acquisition frequency refers to how often the sensor records data; reaction dynamics refer to the changing trends of physical or chemical parameters during a chemical reaction, such as a rapid increase in temperature or a rapid decrease in pH. By evaluating the data acquisition frequency and reaction dynamics, a window size can be set that covers key changes in the reaction while avoiding redundant noise interference; for example, a time period of 10 seconds or 30 seconds can be selected as the sliding analysis window.

[0039] Subsequently, based on the preset window size, the methanol solution flow data, organic acid flow data, and organic acid reaction data are preprocessed. The average value and standard deviation of the raw collected data are calculated within each time window. The calculation results are used as the window data anomaly judgment threshold. Then, outliers are judged and replaced by the window data anomaly judgment threshold to eliminate possible instantaneous outliers, drift, or sampling errors in the data, making the obtained data more stable and reliable.

[0040] Next, key features were extracted from the standard methanol solution flow data, standard organic acid flow data, and standard organic acid reaction data, such as maximum flow rate, flow rate change rate, inflection point position in reaction time, and pH equilibrium point occurrence time, to obtain a key feature set of reactants, which can effectively summarize the role mode and transformation law of methanol and organic acid in the reaction process.

[0041] Next, historical data mining and purity prediction fitting are performed using the key feature set of reactants. By calling existing historical reaction records, the current feature combination is compared and trend analyzed, and machine learning or fitting algorithms are used to predict the purity of the final product. The product purity prediction channel is a prediction model or computational path generated in this process, which can automatically estimate the purity value of the product based on the current feature set. The product purity prediction channel may be trained using methods such as regression models, neural networks, or decision trees to establish a mapping relationship between features and purity.

[0042] Furthermore, this application also includes: arranging the methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data in chronological order to obtain methanol solution flow rate sequence data, organic acid flow rate sequence data, and organic acid reaction sequence data; calculating the average value and standard deviation of the methanol solution flow rate sequence data, organic acid flow rate sequence data, and organic acid reaction sequence data according to the preset window size to determine a window data anomaly judgment threshold; performing anomaly judgment and average value replacement on the methanol solution flow rate sequence data, organic acid flow rate sequence data, and organic acid reaction sequence data based on the window data anomaly judgment threshold to obtain usable methanol solution flow rate data, usable organic acid flow rate data, and usable organic acid reaction data; and normalizing the usable methanol solution flow rate data, usable organic acid flow rate data, and usable organic acid reaction data to obtain standard methanol solution flow rate data, standard organic acid flow rate data, and standard organic acid reaction data.

[0043] Specifically, methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data are arranged in chronological order, sequentially from the starting point of recording, to construct a continuous, time-driven data sequence. This yields methanol solution flow rate sequence data, organic acid flow rate sequence data, and organic acid reaction sequence data, which helps to capture the changing trends of methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data over time and provides a basis for subsequent statistical analysis.

[0044] Next, the mean and standard deviation of the methanol solution flow rate sequence data, organic acid flow rate sequence data, and organic acid reaction sequence data are calculated according to a preset window size. Within each set time interval (e.g., 30 seconds), the mathematical mean and fluctuation degree of the data segments falling within that interval are calculated separately. The mean reflects the central level of the variable within the preset window size, while the standard deviation measures the degree of dispersion of the data around the mean. Through the analysis of mean and standard deviation, it is possible to determine whether each data segment is stable and whether sudden fluctuations have occurred.

[0045] Therefore, the anomaly detection threshold calculated based on the mean and standard deviation of the window data is used as the standard for judging whether subsequent data is abnormal. Anomaly detection is performed on the methanol solution flow rate sequence data, organic acid flow rate sequence data, and organic acid reaction sequence data to obtain the anomaly detection results. These results are then replaced with the mean of the current window to obtain a more stable and reliable usable dataset, outputting usable methanol solution flow rate data, usable organic acid flow rate data, and usable organic acid reaction data.

[0046] Finally, the available methanol solution flow rate data, available organic acid flow rate data, and available organic acid reaction data are normalized to map data with different dimensions and value ranges to a unified standard scale (e.g., between 0 and 1), obtaining standard methanol solution flow rate data, standard organic acid flow rate data, and standard organic acid reaction data. This ensures the comparability of variables in subsequent modeling or comparison processes. Normalization methods include min-max normalization or z-score standardization.

[0047] Furthermore, this application also includes: extracting correlation features from the standard methanol solution flow data, standard organic acid flow data, and standard organic acid reaction data to obtain a reactant correlation feature set; calculating a correlation coefficient set between each correlation feature in the reactant correlation feature set and the product purity; setting a correlation coefficient threshold according to purity prediction requirements; and performing key feature screening on the reactant correlation feature set based on the correlation coefficient set according to the correlation coefficient threshold to obtain the reactant key feature set.

[0048] Specifically, correlation features are extracted from standard methanol solution flow rate data, standard organic acid flow rate data, and standard organic acid reaction data. By analyzing the interactions between different variables, representative or explanatory combined features are extracted to obtain a reactant correlation feature set. Correlation features refer to characteristics formed by combining multiple variables in a mathematical or statistical manner, such as the ratio of methanol flow rate to organic acid flow rate, the interaction effect of their time delays, or their product terms with the reaction data. Correlation features describe the coupling relationships between variables in the reaction system from multiple dimensions, thus forming a reactant correlation feature set.

[0049] Next, the correlation coefficient set between the associated features in the reactant correlation feature set and product purity is calculated to measure the correlation strength between each associated feature and product purity. The correlation coefficient is a numerical value between -1 and 1, used to represent the degree of linear correlation between two variables. Positive values ​​indicate positive correlation, negative values ​​indicate negative correlation, and the larger the absolute value, the stronger the correlation.

[0050] Then, a correlation coefficient threshold is set according to the purity prediction requirements to filter out features that have a significant impact on product purity. Features that meet the conditions are selected from the correlation coefficient set of related features according to the correlation coefficient threshold, and low-relevance or redundant features are removed from the correlation coefficient set of related features. Finally, a simplified and accurate set of key reactant features is obtained, which is used for subsequent purity prediction model or control logic construction.

[0051] Furthermore, this application also includes: performing historical data mining based on the key feature set of the reactants to obtain a reactant feature dataset and a corresponding product purity dataset; using the reactant feature dataset as an input variable and the product purity dataset as an output variable to perform purity regression fitting and construct an initial purity prediction channel; introducing a penalty term to perform performance cross-validation and parameter iterative tuning on the initial purity prediction channel to generate the product purity prediction channel.

[0052] Specifically, historical data mining is performed based on the key feature set of reactants. This involves filtering and analyzing historical reaction data using identified key input parameters that significantly affect product purity. Historical data refers to actual records accumulated over multiple past batches of reactions, including flow rates, reaction conditions, and final purity. The goal of data mining is to extract useful information from historical data, specifically including reactant feature datasets (numerical combinations of key feature parameters describing each reaction) and corresponding product purity datasets (the product purity results corresponding to each set of reaction conditions). For example, reactant features might include a methanol flow rate of 5 liters per hour and an organic acid concentration of 0.8 mol per liter, with a corresponding purity of 93%.

[0053] Next, using the reactant feature dataset as the input variable and the product purity dataset as the output variable, purity regression fitting is performed. Through mathematical modeling, the model can predict the purity of the output product based on the input features. Regression fitting is a process of establishing the relationship between input and output using functional forms. Methods include linear regression, multinomial regression, support vector regression, or neural networks, aiming to minimize the difference between the model's predicted value and the true purity value. The resulting model constitutes the initial purity prediction channel.

[0054] Furthermore, to improve the accuracy and stability of the initial purity prediction channel, a penalty term is introduced for performance cross-validation and parameter iterative tuning. The penalty term is a mathematical technique to prevent overfitting, suppressing the influence of unnecessary complex features in the initial purity prediction channel. Performance cross-validation is a model validation method that divides data into training and validation sets, alternating between training and validation to test its generalization ability. Parameter iterative tuning involves continuously adjusting model parameters, such as the learning rate, regularization coefficient, or number of neurons, over multiple training rounds to optimize the performance of the initial purity prediction channel. The resulting product purity prediction channel exhibits stronger stability and generalization ability than the initial model, enabling more accurate prediction of product purity under various operating conditions.

[0055] Furthermore, this application also includes: dividing the reactant feature dataset and the corresponding product purity dataset into reaction stages to obtain an initial reaction stage, a middle reaction stage, and a later reaction stage; setting initial product purity targets, middle product purity targets, and later product purity targets for the initial reaction stage, the middle reaction stage, and the later reaction stage; and performing control strategy analysis based on the initial product purity targets, the middle product purity targets, and the later product purity targets to obtain a segmented control strategy for organic acids.

[0056] Specifically, the reaction stage is divided by the reactant characteristic dataset and the corresponding product purity dataset. Based on the time sequence or characteristic change trend, the entire reaction process is divided into multiple stages with different characteristic evolution laws, resulting in the initial reaction stage, the middle reaction stage, and the later reaction stage. This helps to identify the process characteristics in different stages. For example, the initial stage is the reaction start-up stage with a low reaction rate; the middle stage is the main reaction stage with a faster reaction rate; and the later stage is the equilibrium stage with a slower reaction rate or an increase in side reactions.

[0057] Next, product purity targets are set for the initial, middle, and final stages of the reaction, with desired purity control values ​​for each stage. For example, a lower target of 85% can be set for the initial stage, allowing for gradual increases after the reaction stabilizes; the middle stage might be set at 92%, as this is the main product formation stage; and the final stage might be set at 95%, reflecting the required precision of the final product. This phased target setting approach considers the balance between reaction kinetics and control costs, and also facilitates process optimization.

[0058] Then, based on the initial, intermediate, and final product purity targets, a control strategy analysis was conducted to determine the appropriate feedstock addition pace, temperature control scheme, or pH adjustment method for each stage, in order to more effectively achieve the targets at each stage. The final organic acid segmented control strategy is a specific adjustment method for the organic acid flow rate or addition method at each reaction stage. For example, a slowly increasing flow rate strategy can be chosen in the initial stage; a constant flow rate can be used in the intermediate stage to ensure a stable reaction rate; while in the later stage, intermittent or pulsed addition methods may be adopted to reduce the formation of by-products.

[0059] Furthermore, this application also includes: setting the number of clusters K, where K=3; performing K-means clustering on the reactant feature dataset and the corresponding product purity dataset according to the number of clusters K to obtain clustering results; evaluating, optimizing and classifying the clustering results into stages to determine the initial reaction stage, the middle reaction stage and the later reaction stage.

[0060] Specifically, setting the number of clusters K predetermines how many subsets or categories the data will be divided into in the clustering algorithm. A cluster is a group of data points that are close to each other in the feature space, and the number of clusters K determines how many classes the data will be divided into. Setting K to 3 means dividing the entire reaction process into 3 stages, so that the reaction characteristics of different stages can be analyzed and controlled separately.

[0061] Next, K-means clustering was performed on the reactant feature dataset and product purity dataset according to the set number of clusters K. The K-means clustering algorithm was applied to automatically classify the data. K-means is an unsupervised learning algorithm that first randomly selects K initial centroids, and then iteratively adjusts the centroids to ensure that each data point is assigned to the cluster closest to its feature distance. The reactant feature dataset includes input parameters such as methanol solution flow rate, organic acid flow rate, and reaction data, while the product purity dataset is the corresponding output result. Clustering can group data points with similar characteristics in different reaction stages into one class, thereby revealing the stage-specific patterns in the process.

[0062] Subsequently, the cluster division results are evaluated, optimized, and categorized into stages. After the initial clustering is completed, the quality of the clusters (such as intra-cluster distance, inter-cluster distance, and silhouette coefficient) is evaluated to further determine the rationality of the clustering effect. Based on the characteristic distribution of the clusters, they are categorized into the initial, middle, and late stages of the reaction. Evaluation and optimization can be carried out using methods such as the elbow method and silhouette analysis to ensure the scientific validity and practical significance of the division. Finally, each cluster is matched with its corresponding reaction stage to form a staged process understanding.

[0063] Furthermore, this application also includes: initializing PID parameters according to the initial product purity target, the intermediate product purity target, and the late product purity target to obtain an initial segment methanol controller, an intermediate segment methanol controller, and a late segment methanol controller; and configuring a segmented strategy based on the initial segment methanol controller, the intermediate segment methanol controller, and the late segment methanol controller to obtain the organic acid segmented control strategy.

[0064] Specifically, based on the purity targets for the initial, intermediate, and final products, PID parameters are initialized separately, and appropriate proportional (P), integral (I), and derivative (D) parameters are configured for each of the three stages, thus constructing corresponding PID controllers. The PID controller is a closed-loop control method that dynamically adjusts the control output by weighting and combining three adjustment methods based on the error (the difference between the actual and target values). By regulating the flow rate of the methanol solution using PID, the reaction process at each stage is made closer to the expected purity level. This results in initial-stage methanol controllers, intermediate-stage methanol controllers, and final-stage methanol controllers, representing three independent PID controllers constructed for the initial, intermediate, and final stages of the reaction, respectively. Each controller has its parameters initialized and functions set according to the control requirements and dynamic characteristics of its stage. For example, in the initial stage of the reaction, more emphasis may be placed on regulating the speed, so the P value is larger, and the I and D parameters are smaller; while in the intermediate stage of the reaction, more stable flow control is needed to ensure a continuously improving purity level, so the I parameter may be appropriately increased to eliminate long-term errors.

[0065] Next, a segmented strategy configuration is implemented based on the initial, intermediate, and final methanol controllers. The three independent controllers are integrated according to time or reaction stage logic to form a complete segmented control scheme. Each controller is responsible for adjusting the methanol flow rate within a specified stage, thereby adapting to the reaction kinetics characteristics of different stages. The configuration process includes the controller's start-stop switching logic, parameter transition mechanisms, and anomaly handling strategies.

[0066] Ultimately, the segmented control strategy for organic acids was obtained, forming a multi-stage control system with methanol control as the core and matching the characteristics of different reaction stages. This system is used to precisely adjust key parameters in the organic acid reaction process and improve the stability and consistency of product purity.

[0067] Furthermore, this application also includes: performing staged strategy matching on the product purity prediction parameters based on the organic acid segmented regulation strategy, triggering a target segment methanol regulator; and using the target segment methanol regulator to perform methanol regulation calculation on the product purity prediction parameters to determine the target methanol regulation strategy parameters.

[0068] Specifically, based on the organic acid segmented control strategy, the product purity prediction parameters are matched with stage strategies. The currently predicted product purity parameters are analyzed, and the reaction stage is determined, such as whether it belongs to the initial stage, middle stage, or late stage of the reaction. Then, the target section methanol controller is triggered to ensure that the controller used in different stages is targeted, thereby improving the control efficiency and the control accuracy of product purity.

[0069] Subsequently, a target-section methanol controller was used to calculate methanol control parameters based on the predicted product purity, determining the target methanol control strategy parameters. The selected PID controller for this stage was then combined with the current purity prediction value to calculate the required methanol flow adjustment strategy. The calculation process includes proportional control, integral error correction, and derivative prediction within the controller, outputting a set of parameters that can be directly applied to the methanol control execution system. Strategy parameters typically include target flow rate, adjustment rate, and control frequency, enabling rapid response to the current reaction state. For example, when the purity prediction value is lower than the stage target, the target-section methanol controller might output a strategy to increase the methanol flow rate by 10 ml / min; while when it approaches the target value, it might only adjust by 2 ml / min to avoid overshoot. Table 1 shows a partial record of the most recent matching between product purity prediction and methanol control parameters.

[0070] Table 1: Partial Records of the Most Recent Matching of Product Purity Prediction and Methanol Control Parameters

[0071] In summary, the intelligent control method for neutralizing organic acids using sodium methoxide solution provided in this application has the following technical effects: by achieving the technical goal of constructing a dynamic methanol control strategy based on product purity prediction and reaction stage division, it achieves the technical effects of improving the neutralization control response speed, enhancing the control flexibility of each stage, stabilizing product purity, and reducing energy consumption.

[0072] Example 2: Based on the same inventive concept as the intelligent control method for neutralizing organic acids using sodium methoxide solution in the foregoing examples, this application also provides an intelligent control system for neutralizing organic acids using sodium methoxide solution. Please refer to the appendix. Figure 2The system includes: a sensor acquisition module 11, used to deploy sensor modules inside the reaction vessel, the sensor modules integrating a pH sensor, a temperature sensor, and a liquid chromatograph, to acquire methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data through the sensor modules; a channel construction module 12, used to extract key features and perform purity prediction fitting on the methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data to build a product purity prediction channel; a reaction prediction module 13, used to perform reaction prediction on the methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data using the product purity prediction channel to obtain product purity prediction parameters; and a neutralization control module 14, used to construct an organic acid segmented control strategy, perform segmented control analysis on the product purity prediction parameters based on the organic acid segmented control strategy, determine the target methanol control strategy parameters, and perform organic acid neutralization control based on the target methanol control strategy parameters.

[0073] Furthermore, the intelligent control system for neutralizing organic acids using sodium methoxide solution is also used for: determining a preset window size based on the data acquisition frequency and reaction dynamics; preprocessing the methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data based on the preset window size to obtain standard methanol solution flow rate data, standard organic acid flow rate data, and standard organic acid reaction data; extracting key features from the standard methanol solution flow rate data, standard organic acid flow rate data, and standard organic acid reaction data to obtain a key feature set of reactants; and performing historical data mining and purity prediction fitting based on the key feature set of reactants to generate a product purity prediction channel.

[0074] Furthermore, the intelligent control system for neutralizing organic acids using sodium methoxide solution is also used for: arranging the methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data in chronological order to obtain methanol solution flow rate sequence data, organic acid flow rate sequence data, and organic acid reaction sequence data; calculating the average value and standard deviation of the methanol solution flow rate sequence data, organic acid flow rate sequence data, and organic acid reaction sequence data according to the preset window size to determine the window data anomaly judgment threshold; performing anomaly judgment and average value replacement on the methanol solution flow rate sequence data, organic acid flow rate sequence data, and organic acid reaction sequence data based on the window data anomaly judgment threshold to obtain usable methanol solution flow rate data, usable organic acid flow rate data, and usable organic acid reaction data; and normalizing the usable methanol solution flow rate data, usable organic acid flow rate data, and usable organic acid reaction data to obtain standard methanol solution flow rate data, standard organic acid flow rate data, and standard organic acid reaction data.

[0075] Furthermore, the intelligent control system for neutralizing organic acids using sodium methoxide solution is also used for: extracting correlation features from the standard methanol solution flow data, standard organic acid flow data, and standard organic acid reaction data to obtain a reactant correlation feature set; calculating the correlation coefficient set between each correlation feature in the reactant correlation feature set and the product purity; setting a correlation coefficient threshold according to purity prediction requirements; and performing key feature screening on the reactant correlation feature set based on the correlation coefficient set according to the correlation coefficient threshold to obtain the reactant key feature set.

[0076] Furthermore, the intelligent control system for neutralizing organic acids using sodium methoxide solution is also used for: performing historical data mining based on the key feature set of the reactants to obtain a reactant feature dataset and a corresponding product purity dataset; using the reactant feature dataset as an input variable and the product purity dataset as an output variable to perform purity regression fitting and construct an initial purity prediction channel; introducing a penalty term to perform performance cross-validation and parameter iterative optimization on the initial purity prediction channel to generate the product purity prediction channel.

[0077] Furthermore, the intelligent control system for neutralizing organic acids using sodium methoxide solution is also used for: dividing the reaction stage into the reactant characteristic dataset and the corresponding product purity dataset to obtain the initial reaction stage, the middle reaction stage, and the later reaction stage; setting the initial product purity target, the middle product purity target, and the later product purity target for the initial reaction stage, the middle reaction stage, and the later reaction stage; and performing control strategy analysis based on the initial product purity target, the middle product purity target, and the later product purity target to obtain a segmented control strategy for organic acids.

[0078] Furthermore, the intelligent control system for neutralizing organic acids using sodium methoxide solution is also used for: setting the number of clusters K, where K=3; performing K-means clustering on the reactant feature dataset and the corresponding product purity dataset according to the number of clusters K to obtain clustering results; evaluating, optimizing, and classifying the clustering results to determine the initial reaction stage, the middle reaction stage, and the later reaction stage.

[0079] Furthermore, the intelligent control system for neutralizing organic acids using sodium methoxide solution is also used to: initialize PID parameters according to the initial product purity target, the intermediate product purity target, and the final product purity target, respectively, to obtain an initial-stage methanol controller, an intermediate-stage methanol controller, and a final-stage methanol controller; and perform segmented strategy configuration based on the initial-stage methanol controller, the intermediate-stage methanol controller, and the final-stage methanol controller to obtain the segmented control strategy for the organic acid.

[0080] Furthermore, the intelligent control system for neutralizing organic acids using sodium methoxide solution is also used for: performing staged strategy matching on the product purity prediction parameters based on the organic acid segmented control strategy, triggering the target segment methanol controller; and using the target segment methanol controller to perform methanol control calculation on the product purity prediction parameters to determine the target methanol control strategy parameters.

[0081] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The intelligent control method and specific examples of neutralizing organic acids using sodium methoxide solution in the foregoing embodiment 1 are also applicable to the intelligent control system of neutralizing organic acids using sodium methoxide solution in this embodiment. Through the foregoing detailed description of the intelligent control method of neutralizing organic acids using sodium methoxide solution, those skilled in the art can clearly understand the intelligent control system of neutralizing organic acids using sodium methoxide solution in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0082] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0083] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. A smart control method for neutralizing organic acids using sodium methoxide solution, characterized in that, The method includes: A sensor module is installed inside the reaction vessel. The sensor module integrates a pH sensor, a temperature sensor, and a liquid chromatograph. The methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data are collected through the sensor module. Key features were extracted and purity prediction fitting was performed on the methanol solution flow data, organic acid flow data, and organic acid reaction data to build a product purity prediction channel. The product purity prediction channel is used to predict the reaction of the methanol solution flow data, organic acid flow data, and organic acid reaction data to obtain product purity prediction parameters. An organic acid segmented regulation strategy is constructed. Based on the organic acid segmented regulation strategy, the product purity prediction parameters are analyzed and regulated in segments to determine the target methanol regulation strategy parameters. Based on the target methanol regulation strategy parameters, organic acid neutralization regulation is performed.

2. The intelligent control method for neutralizing organic acids using sodium methoxide solution as described in claim 1, characterized in that, The construction of the product purity prediction channel includes: The preset window size is determined based on the data acquisition frequency and dynamic response characteristics; Based on the preset window size, the methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data are preprocessed to obtain standard methanol solution flow rate data, standard organic acid flow rate data, and standard organic acid reaction data. Key features were extracted from the standard methanol solution flow data, standard organic acid flow data, and standard organic acid reaction data to obtain a set of key features of the reactants. Historical data mining and purity prediction fitting are performed based on the key feature set of the reactants to generate a product purity prediction channel.

3. The intelligent control method for neutralizing organic acids using sodium methoxide solution as described in claim 2, characterized in that, The acquisition of standard methanol solution flow rate data, standard organic acid flow rate data, and standard organic acid reaction data includes: The methanol solution flow rate data, organic acid flow rate data, and organic acid reaction data are arranged in chronological order to obtain methanol solution flow rate sequence data, organic acid flow rate sequence data, and organic acid reaction sequence data. The average value and standard deviation of the methanol solution flow rate sequence data, organic acid flow rate sequence data and organic acid reaction sequence data are calculated according to the preset window size to determine the window data anomaly judgment threshold. Based on the window data anomaly judgment threshold, the methanol solution flow rate sequence data, organic acid flow rate sequence data, and organic acid reaction sequence data are subjected to anomaly judgment and average value replacement to obtain usable methanol solution flow rate data, usable organic acid flow rate data, and usable organic acid reaction data; The available methanol solution flow rate data, available organic acid flow rate data, and available organic acid reaction data are normalized to obtain the standard methanol solution flow rate data, standard organic acid flow rate data, and standard organic acid reaction data.

4. The intelligent control method for neutralizing organic acids using sodium methoxide solution as described in claim 2, characterized in that, The obtained key feature set of reactants includes: Correlation feature extraction was performed on the standard methanol solution flow data, standard organic acid flow data, and standard organic acid reaction data to obtain a reactant correlation feature set; Calculate and obtain the correlation coefficient set between each associated feature and product purity in the reactant associated feature set; Based on the purity prediction requirements, a correlation coefficient threshold is set, and the key features of the reactant association feature set are screened according to the correlation coefficient threshold and the correlation coefficient set of the associated features to obtain the reactant key feature set.

5. The intelligent control method for neutralizing organic acids using sodium methoxide solution as described in claim 2, characterized in that, The product purity prediction channel includes: Historical data mining is performed based on the key feature set of the reactants to obtain a reactant feature dataset and a corresponding product purity dataset. Using the reactant feature dataset as input variables and the product purity dataset as output variables, purity regression fitting is performed to construct an initial purity prediction channel. A penalty term is introduced to perform performance cross-validation and parameter iterative tuning on the initial purity prediction channel to generate the product purity prediction channel.

6. The intelligent control method for neutralizing organic acids using sodium methoxide solution as described in claim 5, characterized in that, The constructed organic acid segmented regulation strategy includes: The reaction stage is divided into the reactant characteristic dataset and the corresponding product purity dataset to obtain the initial reaction stage, the middle reaction stage and the later reaction stage. Set the initial product purity target, the intermediate product purity target, and the final product purity target for the initial stage, the intermediate stage, and the final stage of the reaction. Based on the purity targets of the initial product, intermediate product, and late product, a control strategy analysis was conducted to obtain a segmented control strategy for organic acids.

7. The intelligent control method for neutralizing organic acids using sodium methoxide solution as described in claim 6, characterized in that, The process of obtaining the initial stage, middle stage, and later stage of the reaction includes: Set the number of clusters K, where K=3; K-means clustering was performed on the reactant feature dataset and the corresponding product purity dataset according to the number of clusters K, and the clustering results were obtained. The clustering results are evaluated, optimized, and classified into stages to determine the initial, middle, and late stages of the reaction.

8. The intelligent control method for neutralizing organic acids using sodium methoxide solution as described in claim 6, characterized in that, The strategy for obtaining segmented regulation of organic acids includes: Based on the initial product purity target, the intermediate product purity target, and the final product purity target, the PID parameters are initialized respectively to obtain the initial section methanol controller, the intermediate section methanol controller, and the final section methanol controller. Based on the initial methanol regulator, the mid-term methanol regulator, and the late-term methanol regulator, a segmented strategy is configured to obtain the segmented control strategy for organic acids.

9. The intelligent control method for neutralizing organic acids using sodium methoxide solution as described in claim 8, characterized in that, The determination of the target methanol control strategy parameters includes: Based on the organic acid segmented regulation strategy, the product purity prediction parameters are matched in stages to trigger the target segment methanol regulator. The target methanol control strategy parameters are determined by using the target section methanol controller to calculate the methanol control parameters based on the product purity prediction parameters.

10. A smart control system for neutralizing organic acids using sodium methoxide solution, characterized in that, The steps for implementing the intelligent control method for neutralizing organic acids using sodium methoxide solution according to any one of claims 1 to 9 include: The sensor acquisition module is used to deploy sensor modules inside the reaction vessel. The sensor module integrates a pH sensor, a temperature sensor, and a liquid chromatograph. The sensor module acquires methanol solution flow data, organic acid flow data, and organic acid reaction data. The channel construction module is used to extract key features and fit purity predictions to the methanol solution flow data, organic acid flow data, and organic acid reaction data, and to build a product purity prediction channel. The reaction prediction module is used to perform reaction prediction on the methanol solution flow data, organic acid flow data and organic acid reaction data using the product purity prediction channel, and obtain product purity prediction parameters. The neutralization and regulation module is used to construct a segmented regulation strategy for organic acids, analyze the predicted parameters of product purity based on the segmented regulation strategy, determine the target methanol regulation strategy parameters, and perform neutralization and regulation of organic acids based on the target methanol regulation strategy parameters.

Images