Full-process automatic control method and system for high-purity methyl ester esterification reaction

Abstract

The application discloses a full-process automatic control method and system for high-purity methyl ester esterification reaction, and relates to the technical field of intelligent control. The method comprises the following steps: dividing a full process of the high-purity methyl ester esterification reaction into process stages, and obtaining a plurality of process stages; defining a set of adjustable control parameters and product indexes; performing esterification reaction purity influence analysis, and marking key control parameters and non-key control parameters according to purity influence indexes calculated for each process stage; obtaining key control parameter values; establishing a mapping relationship between the set of key control parameters and the set of non-key control parameters, and obtaining non-key control parameter values according to the mapping relationship; and sending the key control parameter values and the non-key control parameter values to a control system. The technical problem that, in the prior art, control parameters continuously expand during the high-purity methyl ester esterification reaction, the coupling relationship between the parameters is complex, and the control complexity is high is solved, and the technical effect of reducing the full-process control complexity is achieved.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, specifically to a fully automated control method and system for the esterification reaction of high-purity methyl ester. Background Technology

[0002] The esterification reaction of high-purity methyl esters is widely used in fine chemicals, biodiesel, and the preparation of chemical intermediates. The reaction process typically involves multiple control parameters, including raw material ratios, catalyst addition, reaction temperature, pressure, stirring rate, feed flow rate, and residence time. Furthermore, there is a strong parameter coupling relationship between different stages of the process. Because the esterification reaction is sensitive to temperature fluctuations, changes in material concentration, and reaction time, deviations in control at any stage can lead to an increase in side reactions, thus affecting the purity of the final methyl ester product. In existing technologies, most esterification reaction control methods still rely on empirical settings or single-stage local adjustments. As the complexity of the process increases and the number of control parameters continues to rise, problems such as overlapping parameter adjustment ranges, unclear control correlations, and difficulty in quickly identifying key control variables easily arise, leading to high overall process control complexity. Summary of the Invention

[0003] This application provides a fully automated control method and system for the esterification reaction of high-purity methyl ester, which solves the technical problem of high control complexity caused by the continuous expansion of control parameters and the complex coupling relationship between parameters in the esterification reaction of high-purity methyl ester in the prior art.

[0004] The first aspect of this application provides a fully automated control method for the esterification reaction of high-purity methyl esters, the method comprising: The entire process of high-purity methyl ester esterification reaction is divided into multiple stages. A set of adjustable control parameters and product indicators for each stage are defined. An impact analysis on the purity of the esterification reaction is performed on the adjustable control parameter set and product indicators for each stage. Key and non-key control parameters are marked according to the purity impact indicators calculated for each stage. A global coupling relationship model based on the key control parameter set as model input is established, and key control parameter values ​​based on the response of the global coupling relationship model are obtained. A mapping relationship between the key and non-key control parameter sets is established, and non-key control parameter values ​​are obtained according to the mapping relationship. The key and non-key control parameter values ​​for each stage are sent to the control system.

[0005] A second aspect of this application provides a fully automated control system for the esterification reaction of high-purity methyl esters, the system comprising: The process segmentation module divides the entire high-purity methyl ester esterification reaction into multiple process stages; the parameter definition module defines the set of adjustable control parameters for each process stage and the product indicators for each process stage; the analysis module performs an esterification reaction purity impact analysis on the set of adjustable control parameters and product indicators for each process stage, and marks key and non-key control parameters according to the purity impact indicators calculated for each process stage; the parameter solving module establishes a global coupling relationship model based on the key control parameter set as the model input, and obtains the key control parameter values ​​based on the response of the global coupling relationship model; the parameter mapping module establishes a mapping relationship between the key control parameter set and the non-key control parameter set, and obtains the non-key control parameter values ​​according to the mapping relationship; and the instruction issuance module sends the key control parameter values ​​and the non-key control parameter values ​​corresponding to each process stage to the control system.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: First, the entire process of high-purity methyl ester esterification reaction is divided into multiple stages. Next, a set of adjustable control parameters and product indicators for each stage are defined. Then, an impact analysis on the purity of the esterification reaction is performed on the adjustable control parameter set and product indicators for each stage, and key and non-key control parameters are marked according to the purity impact indicators calculated for each stage. Next, a global coupling relationship model based on the key control parameter set as model input is established, and the key control parameter values ​​based on the global coupling relationship model response are obtained. Further, a mapping relationship between the key and non-key control parameter sets is established, and the non-key control parameter values ​​are obtained based on the mapping relationship. Finally, the key and non-key control parameter values ​​for each stage are sent to the control system. This solves the technical problem of high control complexity caused by the continuous expansion of control parameters and complex coupling relationships between parameters in the high-purity methyl ester esterification reaction process in existing technologies, achieving the technical effect of reducing the overall process control complexity. Attached Figure Description

[0007] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 A schematic diagram of the fully automated control method for the high-purity methyl ester esterification reaction provided in the embodiments of this application; Figure 2 This is a schematic diagram of the fully automated control system for the high-purity methyl ester esterification reaction provided in the embodiments of this application.

[0009] Figure labeling: Process division module 11, parameter definition module 12, analysis module 13, parameter solving module 14, parameter mapping module 15, instruction issuance module 16. Detailed Implementation

[0010] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0011] Example 1, as Figure 1 As shown, this application provides a fully automated control method for the esterification reaction of high-purity methyl ester, wherein the method includes: The entire process of high-purity methyl ester esterification reaction is divided into multiple process stages.

[0012] First, process operation data of the methyl ester esterification reactor is acquired throughout the entire production cycle. This data includes at least raw material flow rate, catalyst dosage, reaction temperature, reaction pressure, stirring speed, liquid level, circulation flow rate, reaction residence time, and online purity detection data. Then, following the reaction process sequence, process association mapping is performed on each processing unit in the entire esterification process to establish a process flow path. Based on the change characteristics of the process operation data corresponding to each processing unit, stage boundaries are identified for the parameter change rate, material composition change amplitude, and product purity change trend during continuous production. When adjacent time intervals... When the rate of change of parameters exceeds the preset stage switching threshold, or when the product purity change trend shows a stable state transition, the corresponding time interval is divided into different process stages. Each process stage includes at least the raw material premixing stage, catalytic reaction stage, isothermal esterification stage, reaction stabilization stage, and post-processing stage. Finally, stage identification information is established for each process stage, and the start time, end time, participating equipment information, and main process control objectives within the stage are recorded. This generates multiple process stages corresponding to the entire high-purity methyl ester esterification reaction process, which are used for subsequent execution of control parameter definitions and purity impact analysis for each process stage.

[0013] Define the set of adjustable control parameters corresponding to each of the multiple process stages, and the product index of each process stage.

[0014] After dividing the entire process of high-purity methyl ester esterification reaction into stages, the process operation data corresponding to each stage were read, and an adjustable set of control parameters was established based on the process control objectives of each stage. Specifically, for the raw material premixing stage, the raw material feed flow rate, raw material ratio coefficient, preheating temperature, and stirring speed were defined as adjustable control parameters; for the catalytic reaction stage, the catalyst addition acceleration rate, reaction pressure, reaction temperature, and circulation flow rate were defined as adjustable control parameters; for the isothermal esterification stage, the isothermal holding time, jacket heat exchange power, stirring uniformity parameter, and reaction liquid level were defined as adjustable control parameters; and for the post-treatment stage, the separation temperature, discharge rate, and reflux ratio were defined as adjustable control parameters. Subsequently, parameter index tables were established for each process stage, and corresponding upper and lower limits and adjustment steps were configured for each adjustable control parameter for subsequent parameter adjustment and model solving. Simultaneously, product indicators were established based on the product quality control requirements for each process stage. These product indicators include at least methyl ester purity, by-product content, acid value, moisture content, and conversion rate. A stage-related relationship was established between the adjustable control parameter set for each process stage and the corresponding product indicators, generating a process stage parameter-product indicator mapping table for subsequent analysis of the impact of esterification reaction purity and screening of key control parameters.

[0015] An analysis of the impact of esterification reaction on the set of adjustable control parameters and product indicators corresponding to each process stage was conducted, and key control parameters and non-key control parameters were marked according to the purity impact indicators calculated for each process stage.

[0016] First, historical reaction data samples for the corresponding process stages are retrieved, and a parameter-product correlation data sequence is constructed in chronological order. This sequence includes the parameter values ​​of multiple adjustable control parameters at different sampling times, as well as the corresponding methyl ester purity, byproduct content, acid value, and conversion rate indicators. Then, parameter perturbation analysis is performed on each adjustable control parameter to calculate the response of the parameter changes to changes in methyl ester purity. Based on the correlation between parameter changes and product indicator changes, a purity sensitivity value is calculated. Simultaneously, a side reaction impact analysis is conducted on the correlation between each adjustable control parameter and byproduct content, and a byproduct sensitivity value is calculated. Finally, the purity sensitivity value and byproduct sensitivity value are compared. The values ​​are normalized and weighted in conjunction with the target purity requirements of the corresponding process stage to obtain the purity impact index for each adjustable control parameter. A higher purity impact index for a given adjustable control parameter indicates a greater influence of that parameter on the purity stability of methyl ester. Subsequently, a preset purity impact threshold is set, and the purity impact index for each adjustable control parameter is compared with this threshold. Adjustable control parameters greater than or equal to the preset purity impact threshold are marked as critical control parameters, while those less than the threshold are marked as non-critical control parameters. Finally, a set of critical control parameters and a set of non-critical control parameters are generated for each process stage.

[0017] Furthermore, the method also includes analyzing the impact of the adjustable control parameters and product indicators on the purity of the esterification reaction for each process stage. Historical reaction data samples of the entire methyl ester esterification reaction process were extracted, including adjustable control parameter value samples and product index samples for each process stage. The product index samples included label samples characterizing the purity of the prepared product and label samples characterizing the content of by-products. Differential calculations were performed based on the adjustable control parameter value samples and the label samples characterizing the purity of the prepared product to obtain the purity-sensitive index for each adjustable control parameter item. Differential calculations were also performed based on the adjustable control parameter value samples and the label samples characterizing the content of by-products to obtain the by-product-sensitive index for each adjustable control parameter item. The purity-sensitive index and the by-product-sensitive index were then calculated to obtain the purity-impact index for each process stage.

[0018] Preferably, historical reaction data samples of the entire methyl ester esterification reaction process are extracted from the production database, distributed control system, and online detection system, and stored in a categorized manner according to process stage, sampling time, and batch number. The historical reaction data samples include adjustable control parameter value samples corresponding to each process stage and product index samples for each process stage. The adjustable control parameter value samples include at least reaction temperature, reaction pressure, catalyst dosage, raw material ratio, stirring speed, circulation flow rate, and reaction residence time. The product index samples include labels characterizing the purity of the prepared product and labels characterizing the content of by-products. The labels characterizing the purity of the prepared product include esterification conversion rate, methyl ester content, main product selectivity, and reaction rate. The labels characterizing the content of by-products include moisture content, acid value, saponification value, glycerol content, and free fatty acid content. Subsequently, using each adjustable control parameter as the independent variable and the corresponding product purity label as the dependent variable, local perturbation differential calculations are performed on the historical reaction data samples to obtain the value corresponding to the unit parameter change. The purity change rate serves as a purity-sensitive index for the corresponding adjustable control parameter. This purity-sensitive index characterizes the impact of changes in the corresponding adjustable control parameter on esterification conversion rate, methyl ester content, main product selectivity, and reaction rate. Simultaneously, using each adjustable control parameter as the independent variable and the corresponding byproduct content label sample as the dependent variable, local perturbation differential calculations are performed on historical reaction data samples to obtain the byproduct growth rate corresponding to a unit parameter change, serving as a byproduct-sensitive index for the corresponding adjustable control parameter. This byproduct-sensitive index characterizes the impact of changes in the corresponding adjustable control parameter on moisture content, acid value, saponification value, glycerol content, and free fatty acid content. Finally, the purity-sensitive index and the byproduct-sensitive index are dimensionless and weighted fused according to the positive contribution relationship of purity and the negative influence relationship of byproducts to obtain the purity impact index corresponding to each adjustable control parameter in each process stage. A larger purity impact index indicates a more significant impact of the corresponding adjustable control parameter on the purity stability of the final high-purity methyl ester product.

[0019] Furthermore, weights are calculated for the purity-sensitive index and the by-product-sensitive index. The purity-sensitive index is positively correlated with the purity-influencing index, and the by-product-sensitive index is negatively correlated with the purity-influencing index. The weights include a purity-sensitive weight calculated based on the ratio of the standard deviation to the mean of the purity label samples, and a by-product-sensitive weight calculated based on the ratio of the standard deviation to the mean of the purity label samples.

[0020] Specifically, purity label samples and by-product content label samples corresponding to each process stage are extracted, and statistical analysis is performed on each label sample based on a preset statistical time window. Specifically, for the purity label samples, the average and standard deviation of esterification conversion rate, methyl ester content, main product selectivity, and reaction rate within the statistical time window are calculated, and the purity fluctuation coefficient is calculated based on the ratio of the corresponding standard deviation to the average. This purity fluctuation coefficient is used as the purity sensitivity weight for the corresponding process stage. A larger standard deviation and a smaller average value for the purity label samples indicate a higher degree of purity fluctuation in the corresponding process stage, and thus a higher purity sensitivity weight. Simultaneously, for the by-product content label samples, the average and standard deviation of moisture content, acid value, saponification value, glycerol content, and free fatty acid content within the statistical time window are calculated, and the corresponding standard deviation and average value are used as the purity sensitivity weight for the process stage. The by-product fluctuation coefficient is calculated based on the ratio of the values, and this by-product fluctuation coefficient is used as the by-product sensitivity weight for the corresponding process stage. Subsequently, a weighted calculation is performed on the purity sensitivity index and purity sensitivity weight corresponding to each adjustable control parameter to obtain a positive purity contribution value, and a weighted calculation is performed on the by-product sensitivity index and by-product sensitivity weight corresponding to each adjustable control parameter to obtain a negative by-product impact value. The purity sensitivity index and purity impact index are positively correlated, meaning the larger the purity sensitivity index, the larger the corresponding purity impact index. The by-product sensitivity index and purity impact index are negatively correlated, meaning the larger the by-product sensitivity index, the smaller the corresponding purity impact index. Then, the positive purity contribution value and the negative by-product impact value are fused together to generate a purity impact index for the corresponding adjustable control parameter, which is used for subsequent screening of key and non-key control parameters.

[0021] Furthermore, key and non-key control parameters are marked according to the purity impact index calculated for each process stage. The methods include: Set a preset purity impact threshold, and select adjustable control parameters from the set of adjustable control parameters in each process stage that are greater than or equal to the preset purity impact threshold and mark them as critical control parameters; select adjustable control parameters that are less than the preset purity impact threshold and mark them as non-critical control parameters.

[0022] Preferably, the purity impact indicators corresponding to all adjustable control parameters in each process stage are read, and a purity impact indicator sequence is constructed according to each process stage. Subsequently, statistical analysis is performed on the purity impact indicator sequences corresponding to each process stage to calculate the mean, standard deviation, and distribution range of the purity impact indicators for each process stage. A preset purity impact threshold is then generated based on the purity control target for the corresponding process stage. The preset purity impact threshold is determined using a dynamic threshold method, including at least an adaptive threshold generated based on the combination of the mean and standard deviation of the purity impact indicators. Afterward, the purity of each adjustable control parameter in each process stage is... The impact indicators are compared with the preset purity impact threshold. Adjustable control parameters that are greater than or equal to the preset purity impact threshold are selected from the set of adjustable control parameters for each process stage and marked as key control parameters. The key control parameters are used to characterize control variables that have a significant impact on the purity stability of methyl ester products, the trend of by-product formation, and the conversion efficiency of esterification reaction. At the same time, adjustable control parameters that are less than the preset purity impact threshold are selected and marked as non-key control parameters. The non-key control parameters are used to characterize auxiliary control variables that have a low degree of impact on the purity of methyl ester products but still participate in the coordinated regulation of the process.

[0023] Establish a global coupling relationship model based on a set of key control parameters as model input, and obtain the key control parameter values ​​based on the response of the global coupling relationship model.

[0024] Furthermore, a global coupling relationship model based on a set of key control parameters as model inputs is established, including the following methods: Based on the historical reaction data samples, key reaction data samples are extracted for each process stage, consisting of corresponding key control parameter sets, to generate multiple sets of key training sample sets. These sets of key training sample sets are then used as inputs to simulate and train regression sub-models, obtaining predicted product index samples for each process stage. This process continues until the prediction accuracy of the predicted product index samples exceeds a preset threshold, resulting in multiple convergent coupling relationship sub-models. Each coupling relationship sub-model includes parameter coupling relationships based on the corresponding key control parameter sets for each process stage. These convergent coupling relationship sub-models are then globally fused and optimized to obtain a global coupling relationship model.

[0025] First, based on historical reaction data samples, the key control parameter operation data corresponding to each process stage are extracted, and key reaction data samples are constructed by combining them with product index samples of the corresponding process stage. The key reaction data samples include at least a sequence of key control parameter values, methyl ester purity index, by-product content index, conversion rate index, and corresponding process status label within the corresponding time window. Then, outlier removal, normalization, and time series alignment are performed on the corresponding key reaction data samples according to each process stage, ensuring a one-to-one correspondence between key control parameter values ​​at different sampling times and corresponding product indicators, generating multiple key training sample sets corresponding to each process stage. Next, the key training sample sets corresponding to each process stage are input into the corresponding regression sub-models for simulation training. The regression sub-models establish a nonlinear response relationship between key control parameters and product indicators, and output predicted product indicator samples for the corresponding process stage. The predicted product indicator samples include at least predicted methyl ester purity, predicted by-product content, predicted conversion rate, and predicted main product selectivity. Finally, the predicted product indicator samples output by each regression sub-model are compared with the actual product indicator labels in the historical reaction data samples to determine the error. The feedback results continuously adjust the model parameters in the corresponding regression sub-models, enabling the regression sub-models to iteratively learn the coupling relationship between the changing trends of key control parameters and product indicators. The error feedback includes at least a comprehensive prediction error generated based on mean squared error, mean absolute error, and coefficient of determination. As the model parameters are continuously updated, the predicted product indicator samples output by each regression sub-model gradually approach the actual product indicator labels. When the prediction accuracy of the corresponding regression sub-model exceeds a preset threshold, the corresponding regression sub-model is considered to have reached convergence, resulting in multiple convergent coupling relationship sub-models. Each coupling relationship sub-model includes parameter coupling relationships between multiple key control parameters within the corresponding process stage and dynamic correlation relationships between key control parameters and product indicators. After multiple coupling relationship sub-models converge, stage-related constraints are further established based on material flow relationships, heat transfer relationships, and reaction state continuity relationships between adjacent process stages. The multiple coupling relationship sub-models are then globally fused and optimized. By jointly adjusting the model parameters in the multiple coupling relationship sub-models, the overall prediction product indicator error between multiple process stages is minimized, ultimately resulting in a global coupling relationship model covering the entire process of high-purity methyl ester esterification.

[0026] Furthermore, a global fusion optimization is performed on multiple converged coupling sub-models to obtain a global coupling model. The methods include: Obtain the predicted product indicators output by multiple converged coupling sub-models; optimize the model parameters of the multiple coupling sub-models with the global optimization objective of minimizing the weighted sum of errors between the predicted product indicators and the actual product indicator labels of each process stage, and obtain a global coupling model, which is used to output the key control parameter values ​​of the whole process and the corresponding predicted product indicators of the whole process.

[0027] Preferably, the predicted product indicators output by multiple coupled sub-models are obtained. These predicted product indicators include at least the predicted methyl ester purity, predicted by-product content, predicted conversion rate, predicted main product selectivity, and predicted reaction rate for each process stage. Subsequently, error calculations are performed between the predicted product indicators for each process stage and the actual product indicator labels in historical reaction data samples to generate multiple stage error values. Each stage error value includes at least the purity error between the predicted methyl ester purity and the actual methyl ester purity, the by-product error between the predicted by-product content and the actual by-product content, and the conversion rate error between the predicted conversion rate and the actual conversion rate. Then, based on the influence of each process stage in the overall esterification reaction, corresponding error weights are assigned to each stage error value. Specifically, high weight coefficients are assigned to the isothermal esterification stage and the catalytic reaction stage, and high weight coefficients are assigned to the raw material premixing stage and the post-processing stage. The system first configures low-weight coefficients for each stage. Then, with the goal of minimizing the weighted sum of errors between the predicted product indicators and the actual product indicator labels of each process stage, a global loss function is established. Based on this global loss function, the model parameters of multiple coupled sub-models are jointly optimized. Simultaneously, based on the material transfer and heat transfer relationships between adjacent process stages, the stage prediction results output by multiple coupled sub-models are continuously corrected to ensure that the overall prediction results meet the process continuity requirements. When the convergence change of the global loss function is less than a preset convergence threshold, the multiple coupled sub-models are determined to have completed global fusion optimization, resulting in a global coupled model. This global coupled model is used to output the key control parameter values ​​and corresponding predicted product indicators for the entire process based on the real-time process operation status, thereby achieving collaborative optimization control of the high-purity methyl ester esterification reaction process.

[0028] Establish a mapping relationship between the set of critical control parameters and the set of non-critical control parameters, and obtain the values ​​of non-critical control parameters based on the mapping relationship.

[0029] Furthermore, the method for establishing the mapping relationship between the set of critical control parameters and the set of non-critical control parameters includes: Based on the historical reaction data samples, non-critical reaction data samples composed of corresponding key control parameter sets are extracted to generate multiple sets of non-critical training samples; non-linear mapping is performed on the non-critical reaction data samples and the multiple sets of non-critical training samples to obtain multiple mapping relationships corresponding to the multiple process stages.

[0030] Preferably, historical reaction data samples corresponding to each process stage are obtained, and historical operating data corresponding to the key control parameter set and non-key control parameter set are extracted according to the process stage. The key control parameter set includes at least the reaction temperature, reaction pressure, catalyst dosage, and raw material ratio, which have a significant impact on the purity and stability of methyl ester. The non-key control parameter set includes at least the circulation flow rate, liquid level, stirring speed, heat exchange flow rate, and reflux ratio. Using the key control parameter values ​​in each process stage as input features and the corresponding non-key control parameter values ​​as output targets, non-key reaction data samples are constructed. Each set of non-key reaction data samples includes at least a sequence of key control parameter values, a sequence of non-key control parameter values, and corresponding timestamp information. Outlier removal, missing value compensation, and time series synchronization processing are performed on the non-key reaction data samples corresponding to each process stage to generate multiple sets of non-key training samples. A nonlinear mapping model is established based on the non-critical reaction data samples and the multiple sets of non-critical training samples. This model characterizes the dynamic correlation between changes in critical control parameters and changes in non-critical control parameters, and establishes a nonlinear response mapping between critical and non-critical control parameters through iterative training. During model training, minimizing the error between the mapped output value and the actual non-critical control parameter value is used as the optimization objective, and the mapping model parameters are continuously updated. When the predicted non-critical control parameters output by the mapping model meet a preset mapping error threshold, the mapping model for the corresponding process stage is considered to have reached convergence, and multiple mapping relationships corresponding to each process stage are output. These multiple mapping relationships are used to automatically derive the corresponding non-critical control parameter values ​​after the critical control parameter values ​​are determined, thereby achieving coordinated control based on critical control parameters during the high-purity methyl ester esterification reaction.

[0031] Furthermore, the method for obtaining non-critical control parameter values ​​based on the mapping relationship includes: Obtain the key control parameter values ​​for the real-time process stage; according to the corresponding mapping relationship of the real-time process stage, perform mapping and solution on the key control parameter values ​​to obtain the corresponding non-key control parameter values.

[0032] The control system collects real-time process operation data during the high-purity methyl ester esterification reaction and identifies real-time process stages based on the current production time node, process operation status, and product index change trends. It reads the key control parameter values ​​corresponding to each real-time process stage, which include at least real-time reaction temperature, real-time reaction pressure, real-time catalyst dosage, real-time raw material ratio, and real-time reaction residence time. The system then calls the corresponding mapping relationship according to each real-time process stage and inputs the key control parameter values ​​into the corresponding nonlinear mapping model for mapping and solving. The nonlinear mapping model automatically derives the change trends of corresponding non-key control parameters based on the synergistic change relationship between key control parameters, obtaining the corresponding non-key control parameter values, which include at least circulation flow rate, liquid level, stirring speed, heat exchange flow rate, and reflux ratio.

[0033] The key control parameter values ​​and non-key control parameter values ​​corresponding to each process stage are sent to the control system.

[0034] Furthermore, the key control parameter values ​​and non-key control parameter values ​​corresponding to each process stage are sent to the control system. The control system includes multiple actuators, including a raw material metering and feeding actuator, a temperature control actuator, a pressure control actuator, a stirring and mixing actuator, and a liquid level actuator.

[0035] Specifically, the central control unit reads the key and non-key control parameter values ​​for the corresponding process stage and generates a corresponding parameter control instruction set according to the process stage. The parameter control instruction set includes at least the parameter target value, parameter adjustment rate, execution priority, and corresponding actuator identification information. The central control unit sends the parameter control instruction set to the corresponding control system via an industrial communication bus. The control system includes multiple actuators, including at least a raw material metering and feeding actuator, a temperature control actuator, a pressure control actuator, a stirring and mixing actuator, and a liquid level actuator. The raw material metering and feeding actuator adjusts the opening of the corresponding metering pump and feeding valve according to the target raw material ratio and target feed flow rate to control the raw material input rate. The temperature control actuator adjusts the heating power or cooling medium flow rate of the jacketed heat exchange system according to the target reaction temperature. To maintain a stable temperature in the reaction system, the pressure control actuator adjusts the opening of the pressure valve and the gas release rate according to the target reaction pressure to control the internal pressure of the reactor. The stirring and mixing actuator adjusts the output speed of the drive motor according to the target stirring speed to improve the mixing uniformity of the reaction system. The liquid level actuator adjusts the opening of the drain valve and reflux valve according to the target liquid level height to maintain a stable reaction liquid level. After each actuator completes the parameter adjustment, it collects the corresponding process status data in real time and feeds it back to the central control unit. The central control unit judges the deviation between the current process status and the target control status based on the feedback results. When the corresponding deviation exceeds the preset control deviation threshold, it calls back the global coupling relationship model and the corresponding mapping relationship to dynamically correct the key control parameter values ​​and non-key control parameter values, and regenerates the parameter control instruction set to achieve closed-loop collaborative automatic control of the entire process of high-purity methyl ester esterification reaction.

[0036] In summary, the embodiments of this application have at least the following technical effects: First, the entire process of high-purity methyl ester esterification reaction is divided into multiple stages. Next, a set of adjustable control parameters and product indicators for each stage are defined. Then, an impact analysis on the purity of the esterification reaction is performed on the adjustable control parameter set and product indicators for each stage, and key and non-key control parameters are marked according to the purity impact indicators calculated for each stage. Next, a global coupling relationship model based on the key control parameter set as model input is established, and the key control parameter values ​​based on the global coupling relationship model response are obtained. Further, a mapping relationship between the key and non-key control parameter sets is established, and the non-key control parameter values ​​are obtained based on the mapping relationship. Finally, the key and non-key control parameter values ​​for each stage are sent to the control system. This solves the technical problem of high control complexity caused by the continuous expansion of control parameters and complex coupling relationships between parameters in the high-purity methyl ester esterification reaction process in existing technologies, achieving the technical effect of reducing the overall process control complexity.

[0037] Example 2, based on the same inventive concept as the fully automated control method for the high-purity methyl ester esterification reaction in the foregoing examples, such as... Figure 2 As shown, this application provides a fully automated control system for the esterification reaction of high-purity methyl esters, wherein the system includes: Process Division Module 11: Divides the entire process of high-purity methyl ester esterification reaction into multiple process stages; Parameter Definition Module 12: Defines the set of adjustable control parameters corresponding to each process stage, as well as the product index of each process stage; Analysis Module 13: Performs an esterification reaction purity impact analysis on the set of adjustable control parameters and product index corresponding to each process stage, and marks key and non-key control parameters according to the purity impact index calculated for each process stage; Parameter Solving Module 14: Establishes a global coupling relationship model based on the set of key control parameters as model input, and obtains the key control parameter values ​​based on the response of the global coupling relationship model; Parameter Mapping Module 15: Establishes a mapping relationship between the set of key control parameters and the set of non-key control parameters, and obtains the non-key control parameter values ​​according to the mapping relationship; Command Issuance Module 16: Sends the key control parameter values ​​and the non-key control parameter values ​​corresponding to each process stage to the control system.

[0038] Furthermore, the analysis module 13 is used to perform the following methods: Historical reaction data samples of the entire methyl ester esterification reaction process were extracted, including adjustable control parameter value samples and product index samples for each process stage. The product index samples included label samples characterizing the purity of the prepared product and label samples characterizing the content of by-products. Differential calculations were performed based on the adjustable control parameter value samples and the label samples characterizing the purity of the prepared product to obtain the purity-sensitive index for each adjustable control parameter item. Differential calculations were also performed based on the adjustable control parameter value samples and the label samples characterizing the content of by-products to obtain the by-product-sensitive index for each adjustable control parameter item. The purity-sensitive index and the by-product-sensitive index were then calculated to obtain the purity-impact index for each process stage.

[0039] Furthermore, the analysis module 13 is used to perform the following methods: The purity-sensitive index and the by-product-sensitive index are weighted. The purity-sensitive index is positively correlated with the purity-influencing index, and the by-product-sensitive index is negatively correlated with the purity-influencing index. The weights include the purity-sensitive weight calculated based on the ratio of the standard deviation to the mean of the purity label samples, and the by-product-sensitive weight calculated based on the ratio of the standard deviation to the mean of the purity label samples.

[0040] Furthermore, the analysis module 13 is used to perform the following methods: Set a preset purity impact threshold, and select adjustable control parameters from the set of adjustable control parameters in each process stage that are greater than or equal to the preset purity impact threshold and mark them as critical control parameters; select adjustable control parameters that are less than the preset purity impact threshold and mark them as non-critical control parameters.

[0041] Furthermore, the parameter solving module 14 is used to perform the following method: Based on the historical reaction data samples, key reaction data samples are extracted for each process stage, consisting of corresponding key control parameter sets, to generate multiple sets of key training sample sets. These sets of key training sample sets are then used as inputs to simulate and train regression sub-models, obtaining predicted product index samples for each process stage. This process continues until the prediction accuracy of the predicted product index samples exceeds a preset threshold, resulting in multiple convergent coupling relationship sub-models. Each coupling relationship sub-model includes parameter coupling relationships based on the corresponding key control parameter sets for each process stage. These convergent coupling relationship sub-models are then globally fused and optimized to obtain a global coupling relationship model.

[0042] Furthermore, the parameter solving module 14 is used to perform the following method: Obtain the predicted product indicators output by multiple converged coupling sub-models; optimize the model parameters of the multiple coupling sub-models with the global optimization objective of minimizing the weighted sum of errors between the predicted product indicators and the actual product indicator labels of each process stage, and obtain a global coupling model, which is used to output the key control parameter values ​​of the whole process and the corresponding predicted product indicators of the whole process.

[0043] Furthermore, the parameter mapping module 15 is used to perform the following method: Based on the historical reaction data samples, non-critical reaction data samples composed of corresponding key control parameter sets are extracted to generate multiple sets of non-critical training samples; non-linear mapping is performed on the non-critical reaction data samples and the multiple sets of non-critical training samples to obtain multiple mapping relationships corresponding to the multiple process stages.

[0044] Furthermore, the parameter mapping module 15 is used to perform the following method: Obtain the key control parameter values ​​for the real-time process stage; according to the corresponding mapping relationship of the real-time process stage, perform mapping and solution on the key control parameter values ​​to obtain the corresponding non-key control parameter values.

[0045] Furthermore, the instruction issuing module 16 is used to execute the following method: The key control parameter values ​​and non-key control parameter values ​​corresponding to each process stage are sent to the control system. The control system includes multiple actuators, including a raw material metering and feeding actuator, a temperature control actuator, a pressure control actuator, a stirring and mixing actuator, and a liquid level actuator.

[0046] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A fully automated control method for the esterification reaction of high-purity methyl ester, characterized in that, The method includes: The entire process of high-purity methyl ester esterification reaction is divided into multiple process stages. Define the set of adjustable control parameters corresponding to each of the multiple process stages, and the product index of each process stage; An analysis of the impact of esterification reaction on the set of adjustable control parameters and product indicators corresponding to each process stage was conducted, and key control parameters and non-key control parameters were marked according to the purity impact indicators calculated for each process stage. Establish a global coupling relationship model based on a set of key control parameters as model input, and obtain the key control parameter values ​​based on the response of the global coupling relationship model; Establish a mapping relationship between the set of critical control parameters and the set of non-critical control parameters, and obtain the values ​​of non-critical control parameters according to the mapping relationship; Send the key control parameter values ​​and the non-key control parameter values ​​corresponding to each process stage to the control system; The method also includes analyzing the impact of the adjustable control parameters and product indicators on the purity of the esterification reaction for each process stage. Historical reaction data samples of the entire methyl ester esterification reaction process were extracted, including adjustable control parameter value samples corresponding to each process stage and product index samples for each process stage. The product index samples included label samples characterizing the purity of the prepared product and label samples characterizing the content of by-products. Based on the adjustable control parameter value sample and the product purity label sample, differential calculation is performed to obtain the purity sensitivity index for each adjustable control parameter item; Differential calculations are performed based on the adjustable control parameter value samples and the label samples characterizing by-product content to obtain the by-product sensitivity index for each adjustable control parameter item. The purity-sensitive index and the by-product-sensitive index are calculated to obtain the purity impact index calculated for each process stage; The purity-sensitive index and the by-product-sensitive index are weighted and calculated. The purity-sensitive index is positively correlated with the purity-affecting index, and the by-product-sensitive index is negatively correlated with the purity-affecting index. The weights include purity-sensitive weights calculated based on the ratio of the standard deviation to the mean of the purity label samples of the prepared products, and by-product-sensitive weights calculated based on the ratio of the standard deviation to the mean of the purity label samples of the prepared products.

2. The method as described in claim 1, characterized in that, Critical and non-critical control parameters are marked based on the purity impact indicators calculated for each process stage. The methods include: Set a preset purity impact threshold, and select adjustable control parameters that are greater than or equal to the preset purity impact threshold from the set of adjustable control parameters for each process stage and mark them as key control parameters; Adjustable control parameters that are less than the preset purity impact threshold are marked as non-critical control parameters.

3. The method as described in claim 1, characterized in that, A global coupling relationship model is established based on a set of key control parameters as model inputs. The methods include: Based on the historical reaction data samples, key reaction data samples are extracted for each process stage based on the corresponding key control parameter set, and multiple sets of key training sample sets are generated. Using the multiple sets of key training samples as input, the regression sub-models are simulated and trained to obtain the predicted product index samples for each process stage. Until the prediction accuracy of the predicted product index sample is greater than a preset threshold, multiple convergent coupling relationship sub-models are obtained, wherein each coupling relationship sub-model includes parameter coupling relationships based on the set of key control parameters corresponding to each process stage; A global coupling relationship model is obtained by performing global fusion optimization on multiple converged coupling relationship sub-models.

4. The method as described in claim 3, characterized in that, A global coupling relationship model is obtained by globally fusing and optimizing multiple converged coupling relationship sub-models. The methods include: Obtain the prediction product indexes output by multiple converged coupled sub-models; With the goal of minimizing the weighted sum of errors between the predicted product indicators and the actual product indicator labels at each process stage, the model parameters of the multiple coupled sub-models are optimized to obtain a global coupled model, which is used to output the key control parameter values ​​and corresponding predicted product indicators for the entire process.

5. The method as described in claim 3, characterized in that, The method for establishing the mapping relationship between the set of critical control parameters and the set of non-critical control parameters includes: Based on the historical reaction data samples, non-critical reaction data samples composed of corresponding key control parameter sets are extracted to generate multiple sets of non-critical training samples. Based on the non-critical reaction data samples and the multiple sets of non-critical training samples, a non-linear mapping is performed to obtain multiple mapping relationships corresponding to the multiple process stages.

6. The method as described in claim 5, characterized in that, The method for obtaining non-critical control parameter values ​​based on the mapping relationship includes: Obtain the key control parameter values ​​for real-time process stages; The key control parameter values ​​are mapped and solved according to the corresponding mapping relationship called in the real-time process stage to obtain the corresponding non-key control parameter values.

7. The method as described in claim 1, characterized in that, The key control parameter values ​​and non-key control parameter values ​​corresponding to each process stage are sent to the control system. The control system includes multiple actuators, including a raw material metering and feeding actuator, a temperature control actuator, a pressure control actuator, a stirring and mixing actuator, and a liquid level actuator.

8. A fully automated control system for the esterification reaction of high-purity methyl ester, characterized in that, The fully automated control method for implementing the high-purity methyl ester esterification reaction according to any one of claims 1-7, the system comprising: Process segmentation module: Divides the entire process of high-purity methyl ester esterification reaction into process stages, obtaining multiple process stages; Parameter definition module: Defines the set of adjustable control parameters corresponding to each process stage in the multiple process stages, as well as the product indicators of each process stage; Analysis module: Performs esterification reaction purity impact analysis on the set of adjustable control parameters and product indicators corresponding to each process stage, and marks key control parameters and non-key control parameters according to the purity impact indicators calculated for each process stage; Parameter solving module: Establishes a global coupling relationship model based on the set of key control parameters as model input, and obtains the key control parameter values ​​based on the response of the global coupling relationship model; Parameter mapping module: Establishes a mapping relationship between the set of critical control parameters and the set of non-critical control parameters, and obtains the values ​​of non-critical control parameters according to the mapping relationship; Command issuance module: Sends the key control parameter values ​​and non-key control parameter values ​​corresponding to each process stage to the control system.

Images