Pressure adjusting method for tea component separation

Abstract

The invention discloses a pressure adjusting method for tea component separation, and relates to the technical field of tea processing. Comprising the steps of obtaining a first extraction adjustment coefficient according to a relation between extraction pressure and extraction target quantity; measuring the concentration of the extract through a sensor to determine a second extraction adjustment coefficient; comparing the two extracted adjustment coefficients with a preset target value, and judging whether pressure adjustment is needed or not; if the extraction adjustment coefficient is not within the preset range, automatically adjusting the extraction pressure through a pressure adjustment algorithm in combination with a multi-index comprehensive scoring model; the device can automatically adjust the extraction pressure to meet the requirements of different types of tea leaves and extraction targets, so that the extraction efficiency and the product quality of target components are improved, the limitation of traditional fixed pressure setting is solved, the extraction efficiency and the product quality are improved through real-time monitoring and automatic adjustment of the extraction pressure, and the product quality is improved. And the process parameters can be automatically adjusted according to the characteristics and extraction targets of different tea leaves in the extraction process.

Description

Technical Field

[0001] The invention relates to the technical field of tea processing, in particular to a pressure regulating method for separating tea components. Background Art

[0002] Tea component separation is an important part of tea deep processing, which involves extracting specific active ingredients from tea, such as tea polyphenols, caffeine, etc. In this process, pressure is a key control parameter, which directly affects the extraction efficiency and product quality. Traditional methods often use fixed pressure settings, but this method is difficult to adapt to the requirements of different types of tea and different extraction targets. Summary of the invention

[0003] The purpose of the present invention is to provide a pressure regulation method for separating tea components, which brings technological progress and economic benefits to the field of deep processing of tea through precise process control, intelligent automatic regulation and the application of a comprehensive scoring model, and realizes efficient, intelligent and high-quality production of tea component separation.

[0004] The purpose of the present invention can be achieved through the following technical solutions: The present application provides a pressure regulating method for separating tea components, comprising the following steps: A pressure regulating method for separating tea components comprises the following steps: In the tea component separation device, the extraction target is determined according to the required tea components such as phenols or alkalis, and the process parameters of the extraction process are set; Wherein, the process parameters include any one of extraction temperature, extraction pressure, extraction time, and solvent concentration; A first extraction adjustment coefficient is calculated based on the relationship between the extraction pressure and the extraction target amount; Measuring the concentration of the extract by means of a sensor, determining a second extraction adjustment factor, and quantifying the deviation of the extract concentration from a desired target; Among them, after obtaining the second extraction adjustment coefficient, combined with the first extraction adjustment coefficient, a multi-index comprehensive scoring model is established to evaluate the contribution of tea quality and efficacy in combination with different extraction targets; Real-time detection of pressure changes during extraction in a tea component separation device; Compare the two extraction adjustment coefficients with the preset target values ​​to determine whether pressure adjustment is needed; if the extraction adjustment coefficient is not within the preset range, the extraction pressure is automatically adjusted through the pressure adjustment algorithm in combination with the multi-indicator comprehensive scoring model to optimize the extraction effect.

[0005] Furthermore, according to the extraction target, the response surface methodology was used to design the experiment and optimize the extraction process parameters, including: According to the type of tea and the intended use, determine the main active ingredients to be extracted and select the factors for extraction efficiency; The experimental data were used to construct a multivariate quadratic regression model to describe the relationship between each extraction component and extraction efficiency; Specifically, the multivariate quadratic regression model is expressed as: ; in, Expressed as the extraction efficiency, and It is an extracted ingredient. i Used to represent each independent extraction process parameter in the model, j Represents the interaction between the parameters in the model. In the multivariate quadratic regression model, j and i Used to represent the interaction between two different parameters. is the intercept, and is the regression coefficient, is the error term; The significance of the model was tested by analysis of variance, and the main effect, quadratic effect and interaction of each factor were analyzed; Using the optimization function of statistical software, set the objective function to maximize the content of specific components, and conduct verification experiments after finding the optimal process conditions; Among them, the objective function is expressed as: , where Y represents the content of a specific component, T, t, C and P represent the extraction temperature, time, solvent concentration and pressure, respectively, and are used to reflect the component content maximized during the extraction process.

[0006] Further, the first extraction adjustment coefficient is calculated according to the relationship between the extraction pressure and the extraction target amount. Specifically, the first extraction adjustment coefficient is obtained by a gradient ascent method, including: Initialize parameters, collect experimental data, record the extraction target volume obtained under different extraction pressures, and select the initial extraction pressure; Calculate the gradient, describe the relationship between extraction pressure and extraction target quantity through the multivariate quadratic regression model, and calculate the gradient of the objective function Y with respect to the extraction pressure P. Update the parameters and update the extracted pressure along the positive direction of the gradient, expressed as: in, is the learning rate, which controls the step size of each update; Repeat the above calculation of gradient and update of parameters until the gradient reaches the preset maximum number of iterations; The first extraction adjustment coefficient is calculated according to the preset maximum number of iterations, which is specifically expressed as: in, is the optimal extraction pressure obtained by the gradient ascent method.

[0007] Specifically, the gradient of the objective function Y with respect to the extraction pressure P is expressed as: in, is the regression coefficient of the model, Represents the sensitivity of the objective function with respect to the extraction pressure.

[0008] Further, determining the second extraction adjustment coefficient specifically includes: The concentration of the extract is measured in real time by a sensor, and the target concentration is set to calculate the concentration deviation value. According to the measured concentration deviation value, it is input into the fuzzy logic controller for analysis to generate control output for adjusting the parameters of the extraction process; Use reinforcement learning algorithms to adjust extraction parameters based on current extraction performance and feedback; Among them, the goal of reinforcement learning is to maximize the cumulative reward R, which is expressed as: ;in It is the instant reward at the current moment. is a discount factor that takes into account future rewards; The control outputs of fuzzy logic control and reinforcement learning are weighted and integrated to obtain the comprehensive control output; the comprehensive control output is expressed as: ,in is the control output from the fuzzy logic control, is the control output from reinforcement learning, A weight coefficient between 0 and 1.

[0009] According to the comprehensive control output, the second extraction adjustment coefficient is calculated, where the second extraction adjustment coefficient is expressed as: ;in Indicates the actual measured concentration, represents the target concentration, Indicates the comprehensive control output.

[0010] Furthermore, a multi-index comprehensive scoring model is established, and the specific contents include: A weight coefficient is determined for each extraction target to reflect the relative importance of each extraction target in tea quality and efficacy; For each extraction target, the extraction score is calculated based on the actual performance and target value during the extraction process. The extraction score is expressed as: ,in, It is i The weight coefficient of each extracted target indicates the relative importance of the target in the overall evaluation; The scores of all extracted targets are weighted and combined to obtain the final comprehensive score. The specific comprehensive score is expressed as: ;in m is the total number of extracted targets.

[0011] Furthermore, the extraction pressure is automatically adjusted through a pressure regulation algorithm, including: Measure the extraction pressure value in real time and set a preset target range for each extraction adjustment coefficient; Calculate the pressure adjustment amount that needs to be adjusted based on the currently calculated extraction adjustment coefficient, the multi-index comprehensive score and the preset target range; According to the calculated pressure adjustment amount, the extraction pressure is automatically adjusted so that the extraction process gradually approaches the optimal state; Among them, the current extraction pressure is updated according to the calculated pressure adjustment amount, which is expressed as: ,in Indicates the minimum safe value of extraction pressure, Indicates the maximum safe value of extraction pressure.

[0012] The process of real-time measurement, calculation and adjustment is repeated continuously to automatically adjust the extraction pressure.

[0013] Further, the pressure adjustment amount to be adjusted is calculated based on and The deviation direction determines the direction of pressure adjustment, and then uses the PID controller to calculate the pressure adjustment amount. The pressure adjustment is expressed as: , is the error term, are the proportional, integral and differential coefficients respectively.

[0014] Furthermore, the BP neural network-genetic algorithm is used to optimize the parameters of the multi-index comprehensive scoring model to improve the prediction accuracy of the model and the efficiency of the adjustment algorithm, which is specifically expressed as follows: , in, represents the comprehensive score of the optimized tea extract, T, t, C, and P represent the extraction temperature, time, solvent concentration, and pressure, respectively, and are the input variables of the model. Represents the parameter set of the BP neural network, which is adjusted by the BP algorithm. BPNN is a BP neural network model used to predict the initial score of the extract based on the input variables. GA is a genetic algorithm model used to optimize the parameters of the BP neural network. .

[0015] The beneficial effects of the present invention are: The present invention realizes precise control of the extraction process by real-time monitoring of key parameters in the extraction process, such as temperature, pressure, time, and solvent concentration, and establishing a multivariate quadratic regression model. By using the response surface method and the gradient rise method, the present invention can automatically adjust the extraction pressure to meet the requirements of different types of tea leaves and extraction targets, thereby significantly improving the extraction efficiency and product quality of target components (such as tea polyphenols and caffeine). The present invention solves the limitations of traditional fixed pressure settings, improves the extraction efficiency and product quality by real-time monitoring and automatic adjustment of the extraction pressure, makes the extraction process more flexible and efficient, and can automatically adjust the process parameters according to the characteristics of different tea leaves and extraction targets; By combining the fuzzy logic controller and the reinforcement learning algorithm, the present invention realizes the automatic adjustment of the extraction pressure. The fuzzy logic controller generates the control output according to the concentration deviation value, while the reinforcement learning algorithm dynamically adjusts the extraction parameters based on the current extraction performance and feedback to maximize the cumulative reward. This intelligent control strategy not only improves the automation level of the extraction process, but also ensures that the optimal extraction effect can be achieved under different extraction conditions. By establishing a multi-index comprehensive scoring model, a weight coefficient is assigned to each target component in the tea extraction process to reflect its contribution to the quality and efficacy of tea. A weight coefficient is then assigned to each target component through weighted comprehensive scoring to ensure efficient extraction of the target component. The model parameters are optimized through BP neural network and genetic algorithm, which improves the prediction accuracy of the model and the efficiency of the adjustment algorithm, thereby significantly improving the consistency and market competitiveness of tea products. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] For better understanding and implementation, the technical solution of the present application is described in detail below with reference to the accompanying drawings.

[0017] Figure 1 A schematic flow chart of a pressure regulating method for separating tea components provided in an embodiment of the present application; Figure 2 A schematic diagram of a process for designing experiments and optimizing extraction process parameters by response surface methodology for a pressure regulation method for separating tea components provided in an embodiment of the present application; Figure 3 A schematic flow chart of a pressure regulation method for separating tea components provided in an embodiment of the present application, for obtaining a first extraction regulation coefficient by a gradient rise method; Figure 4 A schematic diagram of a flow chart of determining a second extraction adjustment coefficient for a pressure adjustment method for separating tea components provided in an embodiment of the present application; Figure 5A schematic flow chart of a pressure regulation method for separating tea components provided in an embodiment of the present application, which automatically adjusts the extraction pressure through a pressure regulation algorithm. DETAILED DESCRIPTION

[0018] In order to further explain the technical means and effects taken by the present invention to achieve the predetermined invention purpose, exemplary embodiments will be described in detail here, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are only examples of methods and systems consistent with some aspects of the present application as detailed in the attached claims.

[0019] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The singular forms of "a", "said" and "the" used in this application and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0020] The specific implementation methods, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0021] See also Figure 1-Figure 5 This embodiment provides a pressure regulation method for separating tea components. Through precise process control, intelligent automatic regulation and the application of a comprehensive scoring model, it brings technological progress and economic benefits to the field of deep processing of tea, and realizes efficient, intelligent and high-quality production of tea component separation.

[0022] The present invention provides a pressure regulating method for separating tea components, comprising the following steps: S1. In the tea component separation device, the extraction target is determined according to the required tea components such as phenols or alkalis, and the process parameters such as temperature, pressure and time of the extraction process are set; Among them, the extraction targets usually include various bioactive ingredients in tea, such as tea polyphenols, caffeine, amino acids, vitamins and minerals, etc. These ingredients are highly valued for their potential health benefits, such as antioxidant, refreshing, anti-inflammatory and antibacterial effects. When determining the extraction target, a specific group of ingredients will be selected based on the intended use of the final product and market demand, so as to maximize the extraction efficiency and yield of these ingredients by optimizing the extraction process.

[0023] The process parameters include any one of extraction temperature, extraction pressure, extraction time, and solvent concentration.

[0024] Furthermore, according to the extraction target, the response surface methodology was used to design the experiment and optimize the extraction process parameters, including: S11. According to the type of tea and the intended use, determine the main active ingredients to be extracted, such as tea polyphenols, caffeine, etc., and select factors that affect the extraction efficiency, such as extraction temperature (T), time (t), solvent concentration (C), etc.; S12. Use the experimental data to construct a multivariate quadratic regression model to describe the relationship between each extraction component and extraction efficiency; Specifically, the multivariate quadratic regression model is expressed as: ; in, Expressed as the extraction efficiency, and It is an extracted ingredient. i It is used to represent the independent extraction process parameters in the model, such as extraction temperature (T), time (t), solvent concentration (C) and pressure (P). i It is used to traverse all extraction parameters to consider the individual effect of each parameter on the extraction efficiency (linear term) and the effect of the square of each parameter on the extraction efficiency. It represents the interaction between the parameters in the model. In the multivariate quadratic regression model, the interaction term represents the effect of two different parameters acting together on the extraction efficiency. j and i are used together to represent the interaction between two different parameters. is the intercept, and is the regression coefficient, is the error term; S13. Test the significance of the model through variance analysis and analyze the main effect, quadratic effect and interaction of each factor; S14. Use the optimization function of statistical software to set the objective function to maximize the content of specific components. After finding the optimal process conditions, conduct verification experiments to ensure that the model's predictions are consistent with the actual results.

[0025] Among them, the objective function is expressed as: , where Y represents the content of a specific component, and T, t, C, and P represent the extraction temperature, time, solvent concentration, and pressure, respectively, which is used to reflect the component content that is maximized during the extraction process.

[0026] Among them, response surface methodology (RSM) is a statistical method for optimizing the extraction process by establishing a mathematical model. It is used to study the effects of multiple variables on one or more response variables. In the process of extracting tea components, the response surface methodology can be used to determine the optimal process parameters such as temperature, time, and pressure to achieve the maximum extraction efficiency of specific components. First, the key factors affecting the extraction efficiency, such as temperature, time, and solvent concentration, are selected based on single-factor experiments; then, a series of experiments are systematically arranged through experimental design methods such as central composite design, and data are collected to construct a multivariate quadratic regression model, which reveals the relationship between each factor and the extraction efficiency; then, variance analysis is used to verify the accuracy of the model, and an optimization algorithm is used to find the combination of factors that can maximize the extraction efficiency; finally, the optimal conditions are confirmed through verification experiments and applied to actual production to improve the extraction efficiency and product quality of the target components in tea.

[0027] S2. Calculating a first extraction adjustment coefficient according to the relationship between the extraction pressure and the extraction target amount; Further, the first extraction adjustment coefficient is calculated according to the relationship between the extraction pressure and the extraction target amount. Specifically, the first extraction adjustment coefficient is obtained by a gradient ascent method, including: S21, initializing parameters, collecting experimental data, recording the extraction target amount obtained under different extraction pressures, and selecting the initial extraction pressure; S22, calculate the gradient, describe the relationship between the extraction pressure and the extraction target amount through the multivariate quadratic regression model, and calculate the gradient of the objective function Y with respect to the extraction pressure P, Specifically, the gradient of the objective function Y with respect to the extraction pressure P is expressed as: in, is the regression coefficient of the model, Represents the sensitivity of the objective function with respect to the extraction pressure.

[0028] S23, update the parameters, and update the extraction pressure along the positive direction of the gradient, expressed as: in, is the learning rate, which controls the step size of each update; S24, repeat the above calculation of gradient and update of parameters until the gradient reaches a preset maximum number of iterations; S25, calculating the first extraction adjustment coefficient according to the preset maximum number of iterations, which is specifically expressed as: in, is the optimal extraction pressure obtained by the gradient ascent method.

[0029] S3, measuring the concentration of the extract by a sensor, determining a second extraction adjustment coefficient, and quantifying the deviation of the extract concentration from the expected target; After obtaining the second extraction adjustment coefficient, a multi-index comprehensive scoring model is established in combination with the first extraction adjustment coefficient. The multi-index comprehensive scoring model combines different extraction targets and their contributions to tea quality and efficacy. Further, determining the second extraction adjustment coefficient specifically includes: S31, measuring the concentration of the extract in real time through a sensor, setting a target concentration, calculating a concentration deviation value, inputting the measured concentration deviation value into a fuzzy logic controller for analysis, generating a control output for adjusting the parameters of the extraction process; S32. Use reinforcement learning algorithms (such as Q-learning or deep Q-networks) to adjust extraction parameters based on current extraction performance and feedback; Among them, the goal of reinforcement learning is to maximize the cumulative reward R, which is expressed as: ;in It is the instant reward at the current moment. is a discount factor that takes into account future rewards; S33, weighted synthesis is performed according to the control outputs of fuzzy logic control and reinforcement learning to obtain a comprehensive control output; the comprehensive control output is expressed as: ,in is the control output from the fuzzy logic control, is the control output from reinforcement learning, A weight coefficient between 0 and 1.

[0030] S34. According to the comprehensive control output, a second extraction adjustment coefficient is calculated, wherein the second extraction adjustment coefficient is expressed as: ;in Indicates the actual measured concentration, represents the target concentration, Indicates the comprehensive control output.

[0031] Specifically, in the process of combining the fuzzy logic controller and the reinforcement learning algorithm, the deviation between the real-time measured extract concentration and the target concentration is first analyzed by the FLC to generate a preliminary adjustment strategy as the control output. This control output is then used as part of the input of the reinforcement learning algorithm, which updates and optimizes the adjustment strategy of the extraction parameters by evaluating the extraction effect (immediate reward) under the current strategy and considering the long-term effect (through the discount factor). In this way, the fuzzy logic controller provides an immediate adjustment method based on the current deviation, while reinforcement learning optimizes these adjustments in the long term to achieve higher extraction efficiency and extracts closer to the target concentration. Ultimately, these adjustment strategies are converted into specific operations, such as changing the extraction pressure, temperature or time, to achieve precise control of the extraction process.

[0032] Furthermore, a multi-index comprehensive scoring model is established, and the specific contents include: A weight coefficient is determined for each extraction target to reflect the relative importance of each extraction target in tea quality and efficacy; For each extraction target, the extraction score is calculated based on the actual performance and target value during the extraction process. The extraction score is expressed as: ,in, It is i The weight coefficient of each extracted target indicates the relative importance of the target in the overall evaluation; The scores of all extracted targets are weighted and combined to obtain the final comprehensive score. The specific comprehensive score is expressed as: ;in m is the total number of extracted targets.

[0033] Specifically, by establishing a multi-index comprehensive scoring model, a weight coefficient is assigned to each target component in the tea extraction process to reflect its contribution to the quality and efficacy of tea, thus optimizing the tea extraction process and ensuring the efficient extraction of target components such as tea polyphenols and caffeine; then, the score of each component is calculated based on its actual extraction effect and expected target; these scores are weighted and combined to form a comprehensive score, thus providing a quantitative evaluation indicator for the tea extraction process. This method not only optimizes the extraction process and ensures the efficient extraction of target components, but also ultimately improves the consistency and market competitiveness of tea products.

[0034] S4, detecting the pressure change during the extraction process in the tea component separation device in real time; Furthermore, high-precision pressure sensors are first used to collect data in real time, which is key information about pressure changes during the extraction process. Subsequently, the raw signal is conditioned through signal processing steps, including amplification to increase signal strength, filtering to eliminate noise interference, and digitization to convert analog signals into digital signals for subsequent analysis. Next, data analysis techniques such as moving average or exponential smoothing methods are used to conduct in-depth analysis of the pressure data to identify and track trends and patterns of pressure changes, thereby understanding the dynamic characteristics of the extraction process. Finally, anomaly detection algorithms are used to scan outliers or mutations in the data, which may be signs of problems in the extraction process, such as equipment failure or operational errors, so as to achieve timely diagnosis and adjustment to ensure the stability of the extraction process and maximize the extraction efficiency.

[0035] S5. Compare the two extraction adjustment coefficients with the preset target values ​​to determine whether pressure adjustment is needed; if the extraction adjustment coefficient is not within the preset range, the extraction pressure is automatically adjusted through the pressure adjustment algorithm in combination with the multi-indicator comprehensive scoring model to optimize the extraction effect and make the pressure closer to the preset target value.

[0036] Specifically, the multi-index comprehensive scoring model helps to more accurately evaluate the effect of the extraction process and guide pressure regulation to achieve more efficient and high-quality tea component extraction. This method can automatically adjust the extraction pressure according to the requirements of different teas and extraction targets to improve the extraction efficiency.

[0037] Furthermore, the extraction pressure is automatically adjusted through a pressure regulation algorithm, including: S51, measuring the extraction pressure value in real time, and setting a preset target range for each extraction adjustment coefficient; S52, calculating the pressure adjustment amount to be adjusted according to the currently calculated extraction adjustment coefficient, the multi-index comprehensive score and the preset target range; S53, automatically adjusting the extraction pressure according to the calculated pressure adjustment amount, so that the extraction process gradually approaches the optimal state; Among them, the current extraction pressure is updated according to the calculated pressure adjustment amount, which is expressed as: ,in Indicates the minimum safe value of extraction pressure, Indicates the maximum safe value of extraction pressure.

[0038] S54, continuously repeating the process of real-time measurement, calculation and adjustment, is used to automatically adjust the extraction pressure to achieve the optimized extraction effect.

[0039] Further, the pressure adjustment amount to be adjusted is calculated based on and The deviation direction determines the direction of pressure adjustment, and then uses the PID controller to calculate the pressure adjustment amount. The pressure adjustment is expressed as: , is the error term, are the proportional, integral and differential coefficients respectively.

[0040] Specifically, in the process of tea extraction, precise control of extraction pressure is achieved through a PID controller: first, a target pressure value is set, and then the current pressure is measured in real time using a sensor. The controller calculates the error between the target and the current pressure, and adjusts the error through three links: proportional (P), integral (I), and differential (D). The proportional link directly amplifies the error, the integral link eliminates long-term deviations, and the differential link predicts the error trend to reduce fluctuations. The outputs of these three links are added together to obtain the pressure adjustment amount. The controller automatically adjusts the pressure according to this adjustment amount until the error is reduced to an acceptable range. This process is dynamic, and the controller will continuously monitor and adjust the pressure to ensure that the extraction process is always in the best state. By automatically adjusting the extraction pressure, this method can adapt to different types of tea and diverse extraction targets, and achieve precise control of the extraction process. This precise pressure regulation significantly improves the extraction efficiency of the target components, ensures the quality and consistency of the extract, and also enhances the flexibility of the production process and the ability to respond to market changes, thereby improving overall production efficiency and product quality.

[0041] Furthermore, the BP neural network-genetic algorithm is used to optimize the parameters of the multi-index comprehensive scoring model to improve the prediction accuracy of the model and the efficiency of the adjustment algorithm, which is specifically expressed as follows: , in, represents the comprehensive score of the optimized tea extract, T, t, C, and P represent the extraction temperature, time, solvent concentration, and pressure, respectively, and are the input variables of the model. Represents the parameter set of the BP neural network, which is adjusted by the BP algorithm. BPNN is a BP neural network model used to predict the initial score of the extract based on the input variables. GA is a genetic algorithm model used to optimize the parameters of the BP neural network. , in order to improve the accuracy of the prediction.

[0042] Specifically, the BP neural network-genetic algorithm is used to optimize the parameters of the multi-index comprehensive scoring model. The principle process includes: first, define the BP neural network (BPNN) model, which accepts extraction temperature (T), time (t), solvent concentration (C) and pressure (P) as input variables and outputs the initial score of tea extract. Then, adjust the parameter set of the BPNN model through the BP algorithm to ensure that the model can accurately predict the score of the extract based on the input variables; then, through the intervention of the genetic algorithm (GA), use the prediction error of the BPNN model as the fitness function, and search for the optimal solution in the parameter space through selection, crossover and mutation operations to optimize the parameters of the BPNN model. This process aims to find a parameter combination that can maximize the prediction accuracy; through this strategy that combines the learning and prediction capabilities of the BP neural network and the global search capabilities of the genetic algorithm, the model can more accurately predict the score of tea extract under different process parameters, thereby improving the extraction efficiency and product quality. Finally, this optimization process improves the comprehensive score of tea extract and realizes the precise control and optimization of the extraction process. Among them, the BP neural network first predicts the score of the extract based on the input process parameters, and then the genetic algorithm optimizes the parameters of the BP neural network to find the best parameter combination, thereby improving the prediction performance of the model and the comprehensive score of the extract. It combines the learning and prediction capabilities of the BP neural network and the global search capabilities of the genetic algorithm to optimize the process parameters in the tea extraction process to achieve maximum extraction efficiency of specific components.

[0043] The present invention aims at optimizing the extraction efficiency and product quality in the process of separating tea components. By real-time monitoring of key parameters in the extraction process, such as temperature, pressure, time and solvent concentration, a multivariate quadratic regression model is constructed using these data. The model can describe the relationship between different extraction parameters and extraction efficiency. Then, the response surface methodology (RSM) is used for experimental design. Experiments are systematically arranged through methods such as central composite design, data are collected, and extraction process parameters are further optimized to achieve the maximum extraction efficiency of specific components. In addition, the present invention also uses a fuzzy logic controller and a reinforcement learning algorithm to adjust the extraction pressure in real time. The fuzzy logic controller generates a control output based on the concentration deviation value, while the reinforcement learning algorithm dynamically adjusts the extraction parameters based on the current extraction performance and feedback by maximizing the cumulative reward. The combination of these two methods enables the extraction process to respond to different extraction conditions and goals more flexibly and accurately, and realizes the automatic adjustment of the extraction pressure; Through the multi-index comprehensive scoring model, the present invention can comprehensively consider the content of different components in tea and their contribution to the quality and efficacy of tea, and provide a quantitative evaluation index for the tea extraction process. This comprehensive scoring model not only optimizes the extraction process and ensures the efficient extraction of target components, but also improves the consistency and market competitiveness of tea products.

[0044] The present invention introduces a BP neural network to process and learn the complex nonlinear relationship between input parameters and extraction efficiency. Through the adaptive learning of the network, the prediction accuracy of the model for extraction efficiency is improved. Then, the parameters of the BP neural network are optimized by a genetic algorithm. The genetic algorithm simulates the process of natural selection, continuously iterates and evolves, and searches for the optimal parameter combination that can maximize the prediction accuracy, thereby further improving the prediction performance of the model.

[0045] The present invention successfully solves the limitation of fixed pressure setting in traditional extraction methods, realizes precise control of the extraction process, significantly improves the extraction efficiency and product quality of target components, and brings significant technological progress and economic benefits to the field of deep processing of tea.

[0046] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A pressure regulating method for separating tea components, characterized in that: In the tea component separation device, the extraction target is determined according to the required tea components, and the process parameters of the extraction process are set; Wherein, the process parameters include any one of extraction temperature, extraction pressure, extraction time, and solvent concentration; A first extraction adjustment coefficient is calculated based on the relationship between the extraction pressure and the extraction target amount; Measuring the concentration of the extract by means of a sensor, determining a second extraction adjustment factor, and quantifying the deviation of the extract concentration from a desired target; Among them, after obtaining the second extraction adjustment coefficient, combined with the first extraction adjustment coefficient, a multi-index comprehensive scoring model is established to evaluate the contribution of tea quality and efficacy in combination with different extraction targets; Real-time detection of pressure changes during extraction in a tea component separation device; The two extraction adjustment coefficients are compared with the preset target values ​​to determine whether pressure adjustment is needed; if the extraction adjustment coefficient is not within the preset range, the extraction pressure is automatically adjusted through the pressure adjustment algorithm in combination with the multi-indicator comprehensive scoring model.

2. A pressure regulating method for separating tea components according to claim 1, characterized in that: According to the extraction target, the response surface methodology is used to design experiments and optimize the extraction process parameters. The specific contents include: According to the type of tea and the intended use, determine the active ingredients to be extracted and select the factors for extraction efficiency; The experimental data were used to construct a multivariate quadratic regression model to describe the relationship between each extraction component and extraction efficiency; Specifically, the multivariate quadratic regression model is expressed as: ； in, Expressed as the extraction efficiency, and It is an extracted ingredient. i Used to represent each independent extraction process parameter in the model, j Represents the interaction between the parameters in the model. In the multivariate quadratic regression model, j and i Used to represent the interaction between two different parameters. is the intercept, and is the regression coefficient, is the error term; The significance of the model was tested by analysis of variance, and the main effect, quadratic effect and interaction of each factor were analyzed; Using the optimization function of statistical software, set the objective function to maximize the content of specific components, and conduct verification experiments after finding the optimal process conditions; Among them, the objective function is expressed as: , where Y represents the content of a specific component, and T, t, C, and P represent the extraction temperature, time, solvent concentration, and pressure, respectively, which is used to reflect the component content that is maximized during the extraction process.

3. The pressure regulating method for separating tea components according to claim 1, characterized in that: The first extraction adjustment coefficient is calculated according to the relationship between the extraction pressure and the extraction target amount. Specifically, the first extraction adjustment coefficient is obtained by a gradient rise method, including: Initialize parameters, collect experimental data, record the extraction target volume obtained under different extraction pressures, and select the initial extraction pressure; Calculate the gradient, describe the relationship between extraction pressure and extraction target quantity through the multivariate quadratic regression model, and calculate the gradient of the objective function Y with respect to the extraction pressure P. Update the parameters and update the extracted pressure along the positive direction of the gradient, expressed as: in, is the learning rate, which controls the step size of each update; Repeat the above calculation of gradient and update of parameters until the gradient reaches the preset maximum number of iterations; The first extraction adjustment coefficient is calculated according to the preset maximum number of iterations, which is specifically expressed as: in, is the optimal extraction pressure obtained by the gradient ascent method.

4. A method for regulating pressure for separating tea components according to claim 3, characterized in that: The gradient of the objective function Y with respect to the extraction pressure P is expressed as: Wherein, T represents the extraction temperature, t represents the extraction time, C represents the extraction solvent concentration, and P represents the extraction pressure. is the regression coefficient of the model, Represents the sensitivity of the objective function with respect to the extraction pressure.

5. The pressure regulating method for separating tea components according to claim 1, characterized in that: Determining the second extraction adjustment coefficient specifically includes: The concentration of the extract is measured in real time by a sensor, and the target concentration is set to calculate the concentration deviation value. The measured concentration deviation value is input into the fuzzy logic controller for analysis to generate a control output for adjusting the parameters of the extraction process; Use reinforcement learning algorithms to adjust extraction parameters based on current extraction performance and feedback; Among them, the goal of reinforcement learning is to maximize the cumulative reward R, which is expressed as: ;in It is the instant reward at the current moment. is a discount factor that takes into account future rewards; The control outputs of fuzzy logic control and reinforcement learning are weighted and integrated to obtain the comprehensive control output; the comprehensive control output is expressed as: ,in is the control output from the fuzzy logic control, is the control output from reinforcement learning, is a weight coefficient between 0 and 1; According to the comprehensive control output, a second extraction adjustment coefficient is calculated.

6. A method for regulating pressure for separating tea components according to claim 5, characterized in that: The second extraction adjustment coefficient is expressed as: ;in Indicates the actual measured concentration, represents the target concentration, Indicates the comprehensive control output.

7. The pressure regulating method for separating tea components according to claim 1, characterized in that: Establish a multi-index comprehensive scoring model, the specific contents include: A weight coefficient is determined for each extraction target to reflect the relative importance of each extraction target in tea quality and efficacy; For each extraction target, the extraction score is calculated based on the actual performance and target value during the extraction process. The extraction score is expressed as: ,in, It is i The weight coefficient of each extracted target indicates the relative importance of the target in the overall evaluation; The scores of all extracted targets are weighted and combined to obtain the final comprehensive score. The specific comprehensive score is expressed as: ;in m is the total number of extracted targets.

8. The pressure regulating method for separating tea components according to claim 1, characterized in that: The extraction pressure is automatically adjusted through the pressure regulation algorithm, including: Measure the extraction pressure value in real time and set a preset target range for each extraction adjustment coefficient; Calculate the pressure adjustment amount that needs to be adjusted based on the currently calculated extraction adjustment coefficient, the multi-index comprehensive score and the preset target range; According to the calculated pressure adjustment amount, the extraction pressure is automatically adjusted so that the extraction process gradually approaches the optimal state; Among them, the current extraction pressure is updated according to the calculated pressure adjustment amount, which is expressed as: ,in Indicates the minimum safe value of extraction pressure, Indicates the maximum safe value of extraction pressure; The process of real-time measurement, calculation and adjustment is repeated continuously to automatically adjust the extraction pressure.

9. A method for regulating pressure for separating tea components according to claim 8, characterized in that: Calculate the pressure adjustment required, based on and The deviation direction determines the direction of pressure adjustment, and then uses the PID controller to calculate the pressure adjustment amount. The pressure adjustment is expressed as: , is the error term, are the proportional, integral and differential coefficients respectively.

10. The pressure regulating method for separating tea components according to claim 7, characterized in that: The BP neural network-genetic algorithm is used to optimize the parameters of the multi-index comprehensive scoring model, which is specifically expressed as follows: ， in, represents the comprehensive score of the optimized tea extract, T, t, C, and P represent the extraction temperature, time, solvent concentration, and pressure, respectively, and are the input variables of the model. Represents the parameter set of the BP neural network, which is adjusted by the BP algorithm. BPNN is a BP neural network model used to predict the initial score of the extract based on the input variables. GA is a genetic algorithm model used to optimize the parameters of the BP neural network. .

Images