Prediction method for index component content in traditional Chinese medicine extraction process, storage medium and electronic device

Abstract

The invention discloses a method for predicting the content of index components in the traditional Chinese medicine extraction process, a storage medium and an electronic device, and belongs to the technical field of traditional Chinese medicine extraction.The method comprises the steps that 1, parameters of traditional Chinese medicine raw materials are collected, parameters of the traditional Chinese medicine extraction process are changed, traditional Chinese medicine extracting solution samples obtained under different extraction conditions and at different extraction time are collected, and information is recorded; determining the content of index components in the traditional Chinese medicine extracting solution sample, and performing data cleaning to obtain a data set; (2) establishing an extraction process heat transfer model and an extraction process mass transfer model, screening the extraction process mass transfer model, training the extraction process heat transfer model and the extraction process mass transfer model by using samples in the data set, and optimizing prediction model parameters; and (3) predicting the content of the index components in the traditional Chinese medicine extraction process by using the extraction process heat transfer model and the extraction process mass transfer model after parameter optimization. The method is suitable for various traditional Chinese medicinal materials and can be used for predicting the extraction process in actual production.

Description

Technical Field

[0001] The invention relates to the technical field of traditional Chinese medicine extraction, and in particular to a method for predicting the content of index components in a traditional Chinese medicine extraction process, a storage medium and an electronic device. Background Art

[0002] The extraction of effective ingredients of traditional Chinese medicine is an important link in the production process of traditional Chinese medicine. If there are abnormalities in the extraction process, the quality of traditional Chinese medicine extracts may fluctuate, thus affecting the consistency of the quality of traditional Chinese medicine intermediates and finished products. The extraction process of traditional Chinese medicine is complex and difficult to control. There are often quality differences between different batches of medicinal materials, resulting in inconsistent quality of extracts from different batches. Even under the same medicinal materials, the temperature rise curves of the extraction process under different working conditions are inconsistent, which will also lead to large differences in the quality of extracts from different batches of medicinal materials. Traditional Chinese medicine industrial extraction tanks often have large time lag characteristics, so it is necessary to adopt a suitable feedforward control strategy to set a suitable heating temperature before extraction. For some traditional Chinese medicines with degradation of index components, it is also necessary to control the appropriate extraction time to reduce the loss of index components on the basis of ensuring sufficient extraction. Therefore, in order to strengthen the control level of the traditional Chinese medicine extraction process, it is necessary to adopt mathematical modeling to study the kinetic curve of the traditional Chinese medicine extraction process under different parameter changes, and then help predict the extraction amount and extraction end point, so as to better control the traditional Chinese medicine extraction process.

[0003] Some modeling research methods for Chinese medicine extraction process have also been reported in the prior art. For example, the Chinese patent document with publication number CN111665716A discloses a mathematical modeling method for Chinese medicine / natural plant extraction. Based on the existing steam heating method, this method first establishes a system model from the initial temperature to the equilibrium temperature in the tank of the extraction device, then establishes a system model for the temperature reduction process in the tank, and then establishes a system model for the pressure change until stability during the temperature increase in the tank, a system model for the pressure change process during the temperature decrease in the tank, and a mathematical model for the dynamic characteristics of input and output in the Chinese medicine / natural plant extraction process. Finally, the mathematical model of the dynamic characteristics of the input and output is decoupled, and the multivariable coupling control is converted into a single input and single output control. This invention reduces the errors caused by human factors and the problem of reduced quality of the extract by optimizing the steam heating extraction method.

[0004] A Chinese patent document with publication number CN112964690A discloses a method for real-time monitoring of the extraction process of Chinese medicine formula granules based on Raman spectroscopy. The invention takes licorice formula granules as an example, firstly samples the extraction process of licorice formula granules to obtain samples, then collects the Raman spectrum of the samples, determines the actual values ​​of the glycyrrhizin and glycyrrhizic acid contents in the samples, and finally establishes a content prediction model, and predicts the contents of glycyrrhizin and glycyrrhizic acid in the actual extraction process according to the content prediction model.

[0005] However, some studies including the above invention still have the following problems: (1) Existing research reports on the kinetics of Chinese medicine extraction process are mostly concentrated in the heat preservation extraction stage, and the established models are mostly ideal models. In industrial production, most of them only have requirements for the extraction time and extraction temperature in the heat preservation stage. However, the actual Chinese medicine extraction process includes four stages: soaking, heating, heat preservation extraction and liquid release. In these four stages, chemical components in Chinese medicinal materials are dissolved, which leads to deviations between the ideal extraction process and the actual extraction process. (2) For modeling research on the extraction process, its kinetic models are mostly described by equations such as first-order kinetic models and second-order kinetic models. Although these equations can fit relatively simple kinetic curves, they are often poorly fitted in industrial production processes; and most models use a single kinetic curve fitting, which has poor prediction ability for different parameter changes (temperature changes, medicinal material batch changes, etc.), and is difficult to use for feedforward control of the actual extraction process. (3) There are significant differences in the extraction processes of different types of Chinese medicine. For example, some are just simple component dissolution processes, while some components undergo chemical transformation (such as degradation or generation). (4) Some studies have used spectroscopy and other technologies to implement online monitoring and use spectroscopy to calculate the content of index components in real time, thereby modeling the content of index components in the extraction process. The advantage of this method is that it can better determine the endpoint. Its disadvantage is that the equipment cost is high and the constructed model is often difficult to adapt to the extraction process under changes in materials or parameters.

[0006] In summary, it is necessary to develop a universal method for fitting and predicting the extraction process of traditional Chinese medicine as the basis for implementing advanced control of industrial extraction processes. Summary of the invention

[0007] The present invention provides a method for predicting the content of index components in the extraction process of traditional Chinese medicine. The method is applicable to a variety of traditional Chinese medicines and can be used for predicting the extraction process in actual production, thus overcoming the defect that the prior art is difficult to simulate and predict the complex extraction process of traditional Chinese medicine.

[0008] The specific technical solutions adopted are as follows:

[0009] A method for predicting the content of an index component in a traditional Chinese medicine extraction process comprises the following steps:

[0010] (1) First, the parameters of the Chinese medicine raw materials are collected, and then by changing the parameters of the Chinese medicine extraction process, samples of Chinese medicine extracts obtained under different extraction conditions and at different extraction times are collected, and the information is recorded. The content of the index components in the Chinese medicine extract samples is determined, and the data set is obtained after data cleaning;

[0011] The Chinese medicine extraction process includes soaking, heating and heat preservation processes, and the recorded information includes the extraction liquid temperature data and the heating medium temperature data;

[0012] (2) establishing a heat transfer model and a mass transfer model for the extraction process, screening the mass transfer model for the extraction process, and using samples in the data set to train the heat transfer model and the mass transfer model for the extraction process to optimize the prediction model parameters; the mass transfer model for the extraction process is a dissolution model or a chemical reaction model, the dissolution model includes a single solvent extraction model or a mixed solvent extraction model, and the chemical reaction model includes a degradation extraction model or a generation extraction model;

[0013] (3) Use the heat transfer model of the extraction process and the mass transfer model of the extraction process after parameter optimization to predict the content of index components in the traditional Chinese medicine extraction process.

[0014] The method of the present invention is based on Fick's mass transfer law and Fourier's heat transfer law, and establishes a complex functional relationship between the temperature of the heating medium and the temperature of the extract, the content of the index component and multiple extraction process parameters. After screening out a suitable model, the optimization algorithm and known experimental data are used to perform global parameter fitting and parameter tuning on the model, so as to achieve the purpose of training the model. After the model training is completed, the heat transfer model of the extraction process after parameter optimization and the mass transfer model of the extraction process are used to predict the content of the index component in the traditional Chinese medicine extraction process.

[0015] The parameters of Chinese medicine raw materials include the particle size of medicinal material particles, the specific surface area of ​​medicinal material particles, the content of index components, the volume of wet medicinal materials or the volume of extract. The specific surface area of ​​medicinal material particles is obtained in the following way: for Chinese medicinal materials that have been crushed, they are regarded as spherical particles, and the particle size is calculated according to the average of the sieve aperture that can retain the powder after screening and the sieve aperture that is one size larger than the sieve aperture (for example, if the powder is retained on the No. 3 sieve, then the particle size of the powder is the average of the apertures of the No. 3 sieve and the No. 2 sieve), and then the specific surface area is calculated based on the particle size; for medicinal materials that have not been crushed, an approximate conversion can be made depending on the situation. The content of index components is detected by the method recorded in the literature or pharmacopoeia. The volume of wet medicinal materials and the volume of extract are specifically determined by filtering and measuring the volume of wet medicinal materials and the volume of extract after soaking under different conditions, which are determined by soaking experiments or obtained by the following formula:

[0016] V M =αM1+κM1 Formula (1)

[0017]

[0018] κ=β r+ γ Formula (3)

[0019] Among them, V M Represents the volume of wet medicinal materials, V w represents the volume of the extract, M1 represents the mass of the raw materials, represents the initial volume of the extraction solvent, r is the particle size of the medicinal material, and α, β, γ and κ are all parameters of the soaking process, which are determined by the soaking experiment. In the modeling process, it is assumed that the volume before and after extraction remains unchanged, the medicinal material absorbs part of the solvent and the volume of the extract decreases, and the volume of the absorbed solvent is related to the mass of the medicinal material and the particle size of the medicinal material.

[0020] Specifically, the recorded information also includes the particle size of the medicinal material, the batch of the medicinal material, the extraction solvent, the solid-liquid ratio or the extraction time, etc.; the content of the index components in the Chinese medicine extract sample is determined by liquid chromatography.

[0021] Specifically, the data cleaning includes abnormal data removal or data format conversion.

[0022] In step (2), heat balance is performed on the Chinese medicine extraction process, and a heat transfer model of the extraction process is established to characterize the heat transfer of the heating medium, the extraction container and the extract. The heat transfer model of the extraction process is expressed as:

[0023]

[0024] Where H is the overall heat transfer coefficient, A w is the heat exchange area between the heater and the extraction container, T w is the temperature of the heating medium in the heater, ∑MC p =M1C p,1 +M2C p,2 +M3C p,3 , M1 is the weight of the raw materials, C p,1 is the specific heat capacity of the medicinal material particles, M2 is the mass of the extraction container, C p,2 is the specific heat capacity of the extraction container, M3 is the mass of the extraction solvent, C p,3 is the specific heat capacity of the extraction solvent, A is the heat dissipation area in contact with the air, h is the heat transfer coefficient between the extraction solution and the outside world, T is the extraction solution temperature, T0 is the ambient temperature, and t is the time.

[0025] In the above equation, the overall heat transfer coefficient H and heat exchange coefficient h are the parameters to be fitted. w It is not a constant value and changes with time during the actual extraction process. Therefore, the differential equation is a variable coefficient differential equation, which is solved by the variable coefficient Runge-Kutta numerical solution method.

[0026] In step (2), the extraction process mass transfer model is used to characterize the change of the index component content under different extraction conditions and different extraction times. The corresponding extraction process mass transfer model is selected according to the properties of the index component and different extraction solvents.

[0027] The mass transfer model of the screening extraction process is specifically as follows: when the index component extraction process is a component dissolution process and the extraction solvent is a pure solvent, a single solvent extraction model is selected; when the index component extraction process is a component dissolution process and the extraction solvent is a mixed solvent, a mixed solvent extraction model is selected; when the index component extraction process involves transformation or degradation of the index component, a degradation extraction model is selected for the index component that undergoes degradation, and a generation extraction model is selected for the index component generated by transformation.

[0028] The single solvent extraction model is expressed as:

[0029]

[0030]

[0031] Among them, M1 is the mass of the raw materials, is the initial content of the index component in the medicinal material, t is the time, k is the mass transfer coefficient, the mass transfer coefficient is related to the temperature T of the extract, and the relationship is shown in formula (S4), a is the specific surface area of ​​the medicinal material particles, C M is the index component content in the medicinal materials during the extraction process, C W is the content of the index component in the extract, K is the ratio coefficient of the index component in the medicinal material to the index component in the extract, and its relationship with temperature is shown in formula (S5), V M and V W The meaning is the same as formula (1) and (2). x1, x2, x3, and x4 are all parameters to be fitted (x1 and x2 are related to the mass transfer coefficient, and x3 and x4 are related to the distribution coefficient). If the experimental results show that the content of the index component is independent of temperature, x4 can be set to 0 to simplify the model.

[0032] In step (2), the mixed solvent extraction model is expressed as:

[0033]

[0034] in, is the solvent ratio of one component in the mixed solvent, and the meanings of the other parameters are the same as those in formulas (S1)-(S5). x1, x2, x3, x4, x5, x6, x7, x8, and x9 are all parameters to be fitted (x1, x2, x3, and x4 are related to the mass transfer coefficient, and x5, x6, x7, x8, and x9 are related to the distribution coefficient). If the content of the index component at the extraction endpoint is independent of some parameters, the relevant parameters can be set to 0 to simplify the model. If it is independent of temperature, that is, when the temperature changes, there is no significant difference in the extraction amount, x6 can be set to 0; if it is independent of the solvent ratio, that is, when the solvent ratio changes, there is no significant difference in the extraction amount, x7 can be set to 0; among them, x9 can be expressed as That is, assuming that it is the specific surface area of ​​spherical particles, r is the particle size of the medicinal material particles. If it is independent of the particle size of the medicinal material particles, that is, there is no obvious difference in the extraction amount under different medicinal material particle sizes, then x9 can be set to 1.

[0035] In step (2), for the extraction model with chemical reaction transformation, assuming that both the index component A and the index component B exist in the medicinal material, part of the index component A will be transformed into the index component B during the extraction process, then for the index component A, it belongs to the degradation extraction model;

[0036] The degradation extraction model is expressed as:

[0037]

[0038] Among them, k d,A is the degradation rate constant of the index component A, k A is the mass transfer coefficient of the index component A, K A The ratio coefficient of the index component A in the medicinal material to the index component A in the extract, C M,A is the content of the index component A in the medicinal material during the extraction process, C W,A is the content of index component A in the extract, the meanings of the other parameters are the same as those in formulas (S1)-(S5), x1, x2, x3, x4, x5, x6 are all parameters to be fitted (x1, x2 are related to the mass transfer coefficient, x3, x4 are related to the degradation coefficient, x5, x6 are related to the distribution coefficient);

[0039] For the index component B, it belongs to the generation and extraction model:

[0040]

[0041] Among them, k d,B is the generation rate constant of the index component B, k B is the mass transfer coefficient of the index component B, K B The ratio coefficient of the index component B in the medicinal material to the index component B in the extract, C M,B is the content of the index component B in the medicinal material, C W,B is the content of index component B in the extract, the meanings of other parameters are the same as in formulas (S11)-(S15), x1, x2, x3, x4, x5, x6 are all parameters to be fitted (x1, x2 are related to the mass transfer coefficient, x3, x4 are related to the generation coefficient, x5, x6 are related to the distribution coefficient).

[0042] Specifically, different samples from multiple groups of experiments in the data set are used as training sets, and the heat transfer model of the extraction process and the mass transfer model of the extraction process are trained using the training sets to fit appropriate parameters while satisfying the extraction process data obtained under different extraction conditions and at different extraction times. In order to improve the fitting effect of the fitting parameter values, the objective function is defined, and the optimization algorithm is used to optimize the parameters to be fitted so that the deviation between the actual data and the calculated data is minimized.

[0043] The optimization algorithm is set to be one of a genetic algorithm, a gray wolf optimization algorithm, and a Markov-Monte Carlo optimization algorithm. Further, the optimization algorithm is set to be a genetic algorithm; the objective function is a determination coefficient, a mean square error, a root mean square error, or a mean absolute error. Further, the objective function is a sum of the determination coefficients, the sum of the mean square errors, the sum of the root mean square errors, or the sum of the mean absolute errors of the training set. Most preferably, the objective function is the sum of the mean square errors of the training set.

[0044] Furthermore, the objective function F is defined as the mean square error:

[0045]

[0046] Among them, m is the number of experimental groups, n is the number of experimental data points, and y i,j is the actual measured value of the data point at the jth point in the experimental group i, is the predicted value of the extraction process corresponding to the time point calculated by the j-th point model under the experimental group i under this parameter combination. The parameter combination is X = [x1, x2…x m ], where m is the number of parameters to be fitted.

[0047] The samples outside the training set in the data set are used as the prediction set. The heat transfer model of the extraction process and the mass transfer model of the extraction process after parameter optimization are used to predict the samples in the prediction set. The determination coefficient (R 2 ), mean square error (MSE), root mean square error (RMSE) and mean absolute error (MAE) to evaluate the prediction effect of the model.

[0048] The present invention also provides a storage medium, in which a program is stored, wherein when the program is run, the method for predicting the content of the index component in the traditional Chinese medicine extraction process is executed.

[0049] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the method for predicting the content of index components in the traditional Chinese medicine extraction process through the computer program.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] (1) The method of the present invention establishes a process prediction method for the index components in traditional Chinese medicine under complex extraction conditions, which can make full use of the known data in the experimental stage for model screening and model training. The model after parameter optimization can be used for the prediction of the extraction process kinetic curve (the curve of content change over time) in actual production, thus solving the defect that the existing technology is difficult to simulate and predict the complex extraction process of traditional Chinese medicine.

[0052] (2) The method of the present invention can predict the content of index components and extraction rate under different extraction parameter changes according to the temperature rise curve in the industrial extraction process, and judge the extraction endpoint, so as to achieve the purpose of improving the quality and efficiency of the traditional Chinese medicine industrial extraction process.

[0053] (3) The method of the present invention has good applicability and can be applied to the prediction of the content of multiple index components in the extraction process of various traditional Chinese medicines including Ginkgo biloba, Panax notoginseng, Scutellaria baicalensis, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 The present invention is a flow chart of a method for predicting the content of index components in the traditional Chinese medicine extraction process.

[0055] Figure 2 Flowchart of the mass transfer model for the screening extraction process.

[0056] Figure 3 This is the temperature curve of the ginkgo leaf extraction process under different conditions. The scattered points are experimental values, and the curve is the predicted curve of the heat transfer model of the extraction process.

[0057] Figure 4 This is the content curve of the ginkgo leaf extraction process under different conditions, where the scattered points are experimental values ​​and the curve is the predicted curve of the heat transfer model of the extraction process.

[0058] Figure 5 This is a diagram showing the effect of using the method of the present invention to predict the content of total flavonol glycosides in the process of extracting Ginkgo biloba leaves.

[0059] Figure 6 This is a diagram showing the effect of using the method of the present invention to predict the contents of notoginsenoside R1, ginsenoside Rg1 and ginsenoside Rb1 in the process of notoginseng extraction.

[0060] Figure 7 The present invention is a diagram showing the effect of using the method of the present invention to predict the baicalin content during the extraction process of Scutellaria baicalensis.

[0061] Figure 8The present invention is a graph showing the effect of using the method of the present invention to predict the contents of chlorogenic acid, cryptochlorogenic acid, neochlorogenic acid, caffeic acid, isochlorogenic acid A, isochlorogenic acid B and isochlorogenic acid C in the process of honeysuckle extraction. DETAILED DESCRIPTION

[0062] The present invention will be further described below in conjunction with the embodiments and drawings. It should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention.

[0063] The flowchart of the method for predicting the content of index components in the traditional Chinese medicine extraction process is as follows: Figure 1 As shown;

[0064] (1) Specifically, in the following embodiments, the heat transfer model of the extraction process is shown in formula (4).

[0065] (2) Specifically, in the following embodiments, the mass transfer model of the extraction process is established, and the mass transfer model of the extraction process is screened. Figure 2 The method is as follows: if it is a simple component dissolution process, that is, the index component is contained in the medicinal material and the extract also contains the index component, and there is no literature research report recording that the component undergoes component degradation or component transformation during the extraction process, then it is classified as a dissolution component modeling, and a single solvent extraction model or a mixed solvent extraction model is selected according to the type of extraction solvent. The single solvent extraction model is shown in formulas (S1)-(S5), and the mixed solvent extraction model is shown in formulas (S6)-(S10); if the component undergoes chemical component transformation and degradation during the extraction process, it is classified as an extraction model with chemical reaction. For the extraction model with chemical reaction transformation, it is assumed that both the index component A and the index component B exist in the medicinal material, and part of the index component A will be transformed into the index component B during the extraction process. For the index component A, it belongs to the degradation extraction model, and the degradation extraction model is shown in formulas (S11)-(S15); for the component B, it belongs to the generation extraction model, and the generation extraction model is shown in formulas (S16)-(S20). After the model is selected, the numerical solution is performed according to the improved variable coefficient Runge-Kutta method.

[0066] (3) In the parameter fitting process, the objective function is defined as the mean square error, see formula (5) for details.

[0067] Example 1

[0068] (1.1) Collecting parameters of Chinese medicine raw materials

[0069] Ginkgo leaves were crushed and sieved to obtain the coarsest powder, coarse powder and medium powder. The particle size of the coarsest powder was the average of the apertures of the No. 1 and No. 2 sieves, which was 1.425 mm. The particle size of the coarse powder was the average of the apertures of the No. 2 and No. 3 sieves, which was 0.6025 mm. The particle size of the medium powder was the average of the apertures of the No. 3 and No. 4 sieves, which was 0.3025 mm. According to the specific surface area formula of spherical particles, their specific surface areas are 2.105 mm and 1.425 mm, respectively. -1 、4.979mm -1 and 9.917mm -1 .

[0070] The formula for determining the volume of Ginkgo biloba medicinal material and the volume of the extract through different soaking experiments is:

[0071] V M =0.2M1+1.8M1 formula (1.1)

[0072]

[0073] The unit of mass is (g), the unit of volume is (mL), V M Represents the volume of wet medicinal materials, V w represents the volume of the extract, M1 represents the mass of the raw materials, Represents the initial volume of extraction solvent. The initial content of total flavonol glycosides in three batches of Ginkgo biloba leaves was determined according to the literature method.

[0074] (1.2) Dataset establishment

[0075] Different groups of Chinese medicine extraction experiments were designed. The factors investigated in the extraction process included the temperature of the heating medium, the proportion of ethanol in the extracted ethanol-water mixed solvent (v / v), the solid-liquid ratio, the particle size of the medicinal material, and the batch of the medicinal material. The extraction experimental conditions of different groups are shown in Table 1. The extraction process is to soak the medicinal material for 1 hour, and then turn on the heating tank to heat for 3 hours. The Chinese medicine extraction process includes soaking, heating and insulation processes. Samples at different time points in the extraction process are collected and the information is recorded. The recorded information includes the particle size of the medicinal material, the batch of the medicinal material, the extraction solvent, the solid-liquid ratio, the temperature data of the extraction liquid, and the temperature data of the heating medium;

[0076] Table 1 Ginkgo leaf extraction experimental table

[0077]

[0078] The collected samples were subjected to liquid chromatography analysis to determine the content of the index components in the Chinese medicine extract samples, and the data set was obtained after data cleaning.

[0079] (1.3) Extraction model screening

[0080] Considering that the indicative component flavonol glycosides do not degrade, the extraction solvent is a mixed solvent, and the extraction solvent ratio parameter is used as an investigation factor, the mixed solvent extraction model was selected.

[0081] (1.4) Establishing heat transfer model of extraction process

[0082] The temperature data of the extract and the temperature data of the heating tank at different times obtained in each extraction experiment, where temperature T = [T1, T2…T n ], heating tank temperature data T w =[T w,1 ,T w,2 …T w,n ], satisfying formula (4), specifically, the extraction container is an extraction tank, wherein the heating area A w is the jacket heating area, A is the heat dissipation area in contact with air, and the extraction tank can be regarded as a cylinder and its surface area minus the jacket heating area to obtain, ∑MC p All data in can be obtained through literature search or actual extraction process. That is, the parameters to be solved are h and H. In solving the equation, firstly, the temperature T vector and the heating tank temperature data T w An interpolation algorithm is used to generate a larger data matrix, and then the differential equation is differentiated using a numerical calculation method and solved using the variable coefficient Runge-Kutta method. The different temperatures obtained by interpolation are used to iterate the solution during each time advancement process.

[0083] (1.5) Establishing the mass transfer model of the extraction process

[0084] After the model is obtained by screening in step (1.3), it is examined whether different parameters have a significant impact on the extraction results, so as to consider reducing the parameters in the model. In the same batch of medicinal materials, the result that affects the final extraction content in the model is parameter K, where parameter Corresponding to the mixed solvent extraction model formula (S6)-(S10), the calculation results are: t = 0, C W =0, there are 8 parameters that need to be fitted. In solving the equation, the temperature T vector is first interpolated using the interpolation algorithm to generate a larger data matrix, and then the differential equation is differentiated using the numerical calculation method and solved using the variable coefficient Runge-Kutta method. In each time advancement process, the different temperatures obtained by interpolation are used to iterate and solve to finally obtain the extraction process curve.

[0085] (1.6) Optimization algorithm to solve temperature parameters

[0086] Different samples from multiple groups of experiments in the data set are used as training sets. The heat transfer model of the extraction process is trained using the training sets to fit the appropriate parameters and satisfy the extraction process data obtained under different extraction conditions and different extraction times. In order to improve the fitting effect of the fitting parameter values, the objective function is defined as the mean square error of all groups of experimental values ​​and calculated values. The genetic algorithm is used for parameter optimization. The temperature parameter is substituted into equation (1.4) to solve the time curve of the internal temperature of the extraction tank of the experimental group, see Figure 3 The determination coefficient R of the fitting of the temperature change curve inside the extraction tank during the extraction of 10 groups of ginkgo leaves 2 The results are 0.9953, 0.9952, 0.9940, 0.9977, 0.9941, 0.9971, 0.9965, 0.9977, 0.9940 and 0.9933 respectively. The average determination coefficient is 0.9955. Overall, the fitting effect is good, which can reflect the temperature change law inside the extraction tank during the ginkgo leaf extraction process.

[0087] (1.7) Optimization algorithm to solve mass transfer parameters

[0088] Different samples from multiple groups of experiments in the data set are used as training sets. The mass transfer model of the extraction process is trained using the training sets to fit the appropriate parameters and simultaneously satisfy the extraction process data obtained under different extraction conditions and at different extraction times. In order to improve the fitting effect of the fitting parameter values, the optimization algorithm is used to fit the parameters and the 8 parameters in (1.5) are fitted. The fitting results are brought into the equation to solve the kinetic curves of the experimental groups. Figure 4 The determination coefficient R of the fitting curves of 10 groups of ginkgo leaf extraction processes 2 They are 0.9587, 0.9344, 0.9716, 0.9533, 0.9547, 0.9920, 0.9798, 0.9285, 0.9822 and 0.9496 respectively. The average determination coefficient is 0.9605. Overall, the fitting effect is good, which can reflect the composition change law of ginkgo leaf extraction process.

[0089] (1.8) Using the model for prediction

[0090] The 11th group was used as the prediction group to predict the heat transfer model of the extraction process and the mass transfer model of the extraction process after parameter optimization. After knowing the temperature change of the heating medium, the temperature change curve of the extraction tank over time was predicted, as shown in Figure 5 , the prediction R of the prediction group 2 is 0.9950. Then the temperature change law is used to predict the change of total flavonoid glycoside content in Ginkgo biloba leaves, and the R 2The value is 0.9862, which indicates that the model developed by the present invention can better fit and predict the changes of temperature and index components under different parameter conditions and different extraction processes (immersion extraction, temperature-raising extraction, and heat preservation extraction).

[0091] Example 2

[0092] (2.1) Collecting parameters of Chinese medicine raw materials

[0093] The Panax notoginseng medicinal material was crushed and sieved to obtain the coarsest powder, coarse powder and medium powder respectively. The volume of Panax notoginseng medicinal material and the volume of extract were determined by different soaking experiments, and the initial contents of three saponins (notoginseng saponin R1, ginsenoside Rg1 and ginsenoside Rb1) in three batches of Panax notoginseng medicinal material were determined according to the literature method.

[0094] (2.2) Dataset establishment

[0095] Different groups of Chinese medicine extraction experiments were designed. The factors investigated in the extraction process included the temperature of the heating medium, the proportion of ethanol in the extracted ethanol-water mixed solvent (v / v), the solid-liquid ratio, the particle size of the medicinal material, and the batch of the medicinal material. The extraction experiments of different groups are shown in Table 2. The extraction process is that the medicinal material is first soaked for 1 hour, and then the heating tank is turned on for heating for 3 hours. The Chinese medicine extraction process includes soaking, heating and insulation processes. Samples at different time points in the extraction process are collected and the information is recorded. The recorded information includes the particle size of the medicinal material, the batch of the medicinal material, the extraction solvent, the solid-liquid ratio, the extraction liquid temperature data, and the heating medium temperature data;

[0096] Table 2 Panax notoginseng extraction experiment table

[0097]

[0098] The collected samples were subjected to liquid chromatography analysis to determine the content of the index components in the Chinese medicine extract samples, and the data set was obtained after data cleaning.

[0099] (2.3) Extraction model screening

[0100] Considering that the index components do not degrade, the extraction solvent is a mixed solvent, and the extraction solvent ratio parameter is used as an investigation factor, the mixed solvent extraction model is selected.

[0101] (2.4) Establishing heat transfer model of extraction process

[0102] Same as Example 1.

[0103] (2.5) Establishment of mass transfer model for extraction process

[0104] After the model is screened, the expression form of parameter K is specifically considered. In the same batch of medicinal materials, the result that affects the final extraction content in the model is parameter K, where parameter The extraction content results are related to temperature, extraction solvent ratio and medicinal material particle size. Examining whether the actual extraction results are related to these parameters can reduce the complexity of the model.

[0105] Specifically, three index components are involved: notoginsenoside R1, ginsenoside Rg1 and ginsenoside Rb1. For notoginsenoside R1, the corresponding mixed solvent extraction model formula (S6)-(S10) has little difference under different conditions. Therefore, K can be considered as a constant, K=x5 simplified model, the initial solution of the equation group is: t=0, C W =0, there are 5 parameters that need to be fitted.

[0106] For ginsenoside Rg1, the corresponding mixed solvent extraction model formula (S6)-(S10) has little difference under different conditions. Therefore, K can be considered as a constant, K = x5 simplified model, the initial solution of the equation system is: t = 0, C W =0, there are 5 parameters that need to be fitted.

[0107] For ginsenoside Rb1, the corresponding mixed solvent extraction model formula (S6)-(S10) has little difference under different conditions. Therefore, K can be considered as a constant, K = x5 simplified model, the initial solution of the equation system is: t = 0, C W =0, there are 5 parameters that need to be fitted.

[0108] In solving the equation, the temperature T vector is first interpolated using an interpolation algorithm to generate a larger data matrix. Then the differential equation is differentiated using a numerical calculation method and solved using the variable coefficient Runge-Kutta method. In each time advancement process, the different temperatures obtained by interpolation are used to iterate and solve the problem, finally obtaining the extraction process curve.

[0109] (2.6) Optimization algorithm to solve temperature parameters

[0110] The objective function was defined as the mean square error of the experimental values ​​and the calculated values ​​of all groups. The genetic algorithm was used to optimize the parameters. The temperature parameters were brought into the equation to solve the curve of the internal temperature change of the extraction tank of the experimental group over time. The determination coefficient R of the fitting of the internal temperature change curve of the extraction tank during the extraction of 12 groups of Panax notoginseng 2 The results were 0.9965, 0.9982, 0.9974, 0.9955, 0.9950, 0.9975, 0.9967, 0.9916, 0.9978, 0.9966 and 0.9979 respectively. The average determination coefficient was 0.9964. Overall, the fitting effect was good, which could reflect the temperature change law inside the extraction tank during the extraction of Panax notoginseng.

[0111] (2.7) Optimization algorithm to solve mass transfer parameters

[0112] For the index component Panax notoginseng saponin R1, the optimization algorithm was used to fit the parameters, and 5 of the parameters were fitted. The fitting results were brought into the equation to solve the kinetic curve of the experimental group. The determination coefficient R of the fitting curve of the first 11 groups of Panax notoginseng saponin R1 extraction process 2 The results were 0.9834, 0.9845, 0.9290, 0.9806, 0.9855, 0.9781, 0.9610, 0.9532, 0.8893, 0.9756 and 0.9325 respectively. The average determination coefficient was 0.9577. Overall, the fitting effect was good, which could reflect the composition change law of notoginseng saponin R1 extraction process.

[0113] For the index component ginsenoside Rg1, the optimization algorithm was used to fit the parameters, and 5 of the parameters were fitted. The fitting results were brought into the equation to solve the kinetic curve of the experimental group. The determination coefficient R of the fitting curve of the first 11 groups of ginsenoside Rg1 extraction process 2 They are 0.9779, 0.9800, 0.9863, 0.9847, 0.9760, 0.9712, 0.9582, 0.7860, 0.8871, 0.9496 and 0.9296 respectively. The average determination coefficient is 0.9422. Overall, the fitting effect is good, which can reflect the composition change law of ginsenoside Rg1 extraction process.

[0114] For the index component ginsenoside Rb1, the optimization algorithm was used to fit the parameters, and 5 of the parameters were fitted. The fitting results were substituted into the equation to solve the kinetic curve of the experimental group. The determination coefficient R of the fitting curve of the first 11 groups of ginsenoside Rb1 extraction process 2 They are 0.9695, 0.9741, 0.9686, 0.9486, 0.9649, 0.9646, 0.9423, 0.8094, 0.8820, 0.9503, and 0.9002 respectively. The average determination coefficient is 0.9340. Overall, the fitting effect is good, which can reflect the composition change law of ginsenoside Rb1 extraction process.

[0115] (2.8) Using the model for prediction

[0116] The 12th group is used as the prediction group for prediction. After knowing the temperature change of the heating medium, the temperature change curve of the extraction tank over time is predicted. Figure 6 . The prediction R of the prediction group 2 The temperature variation law was then used to predict the content changes of the three saponin components of the three medicinal materials of Panax notoginseng. For Panax notoginseng saponin R1, the predicted R 2is 0.9612; for ginsenoside Rg1, its predicted R 2 is 0.9341; for ginsenoside Rb1, its predicted R 2 It is 0.9638. This shows that the model developed by the present invention can better fit and predict the changes in temperature and index components under different parameter conditions and different extraction processes (immersion extraction, temperature-raising extraction, and heat preservation extraction).

[0117] Example 3

[0118] (3.1) Collecting parameters of Chinese medicine raw materials

[0119] The scutellariae baicalensis was crushed and sieved to obtain the coarsest powder, coarse powder and medium powder, and the volume of the extract and the volume of the medicinal material of scutellariae baicalensis were determined by soaking experiments. The content of baicalin in different batches of scutellariae baicalensis medicinal materials was determined according to the literature method.

[0120] (3.2) Dataset establishment

[0121] Different groups of Chinese medicine extraction experiments were designed. The factors investigated in the extraction process included heating medium temperature, solid-liquid ratio, medicinal material particle size and medicinal material batch. The extraction experiments of different groups are shown in Table 3. The extraction process is to soak the medicinal materials for 1 hour, and then turn on the heating tank for 3 hours. The Chinese medicine extraction process includes soaking, heating and insulation processes. Samples at different time points in the extraction process are collected and the information is recorded. The recorded information includes medicinal material particle size, medicinal material batch, solid-liquid ratio, extraction liquid temperature data and heating medium temperature data;

[0122] Table 3 Scutellaria baicalensis extraction experimental table

[0123]

[0124] The collected samples were subjected to liquid chromatography analysis to determine the content of the index components in the Chinese medicine extract samples, and the data set was obtained after data cleaning.

[0125] (3.3) Extraction model screening

[0126] Considering that the index components did not degrade and the extraction solvent was pure water, a single solvent extraction model was selected.

[0127] (3.4) Establishing heat transfer model of extraction process

[0128] Same as Example 1.

[0129] (3.5) Establishing the mass transfer model of the extraction process

[0130] In the same batch of medicinal materials, the parameter K in the model affects the final extraction content. The extraction content results are related to temperature, extraction solvent ratio and medicinal material particle size. The model complexity can be reduced by examining whether the actual extraction results are related to these parameters. For baicalin, the corresponding single solvent extraction model formula (S1)-(S5) can be appropriately simplified according to the experimental results. K = x3, and the initial solution of the equation group is: t = 0, C W =0, there are 3 parameters that need to be fitted.

[0131] (3.6) Optimization algorithm to solve temperature parameters

[0132] The objective function was defined as the mean square error of the experimental values ​​and the calculated values ​​of all groups. The genetic algorithm was used to optimize the parameters. The temperature parameters were brought into the equation to solve the curve of the internal temperature change of the extraction tank in the experimental group over time. The determination coefficient R of the fitting of the internal temperature change curve of the extraction tank during the extraction of 8 groups of Scutellaria baicalensis 2 The results are 0.9914, 0.9955, 0.9934, 0.9746, 0.9965, 0.9966, 0.9944 and 0.9979 respectively. The average determination coefficient is 0.9925. Overall, the fitting effect is good, which can reflect the temperature change law inside the extraction tank during the extraction of Scutellaria baicalensis.

[0133] (3.7) Optimization algorithm to solve mass transfer parameters

[0134] The optimization algorithm was used to fit the parameters, and three of the parameters were fitted. The fitting results were brought into the equation to solve the kinetic curves of the experimental groups. The determination coefficient R of the fitting curves of the 10 groups of Scutellaria baicalensis extraction process 2 They are 0.7221, 0.8984, 0.9437, 0.9897, 0.9613, 0.9934, 0.9206 and 0.9656 respectively. The average determination coefficient is 0.9243. Overall, the fitting effect is good, which can reflect the change law of components in the extraction process of Scutellaria baicalensis.

[0135] (3.8) Using the model for prediction

[0136] Take another group as the prediction group for prediction. After knowing the temperature change of the heating medium, predict the temperature change curve of the extraction tank over time. Figure 7 . The prediction R of the prediction group 2 is 0.9985. Then the temperature change law is used to predict the change of baicalin content in Scutellaria baicalensis, and the R 2 The value is 0.9645, which indicates that the model developed by the present invention can better fit and predict the changes of temperature and index components under different parameter conditions and different extraction processes (immersion extraction, temperature-raising extraction, and heat preservation extraction).

[0137] Example 4

[0138] (4.1) Collecting parameters of Chinese medicine raw materials

[0139] The volume of the extract and the volume of the honeysuckle were determined by soaking experiments. The contents of the index components in different batches of honeysuckle were determined according to the literature method. The index components were chlorogenic acid, cryptochlorogenic acid, neochlorogenic acid, caffeic acid, isochlorogenic acid A, isochlorogenic acid B, and isochlorogenic acid C.

[0140] (4.2) Dataset establishment

[0141] Different groups of Chinese medicine extraction experiments were designed. The factors investigated in the extraction process included heating medium temperature, solid-liquid ratio, soaking time and medicinal material batch. The extraction experiments of different groups are shown in Table 4. There were 7 groups of experiments in total, and the honeysuckle extraction experiment of another batch in the seventh group was used as the prediction set. The extraction process was to soak the medicinal materials for 1 hour, and then turn on the heating tank for 3 hours. The Chinese medicine extraction process includes soaking, heating and insulation processes. Samples at different time points in the extraction process were collected and the information was recorded. The recorded information included medicinal material batch, soaking time, solid-liquid ratio, extraction liquid temperature data and heating medium temperature data;

[0142] Table 4 Honeysuckle Extraction Experiment Table

[0143]

[0144] The collected samples were subjected to liquid chromatography analysis to determine the content of the index components in the Chinese medicine extract samples, and the data set was obtained after data cleaning.

[0145] (4.3) Extraction model screening

[0146] The transformation rules of the components in honeysuckle are complex, and appropriate assumptions are made by referring to relevant literature: chlorogenic acid is partially degraded during the extraction process, and the degradation products are cryptochlorogenic acid and neochlorogenic acid; isochlorogenic acid A is partially degraded during the extraction process, and the degradation products are caffeic acid, isochlorogenic acid B and isochlorogenic acid C. Therefore, chlorogenic acid and isochlorogenic acid A are degradation models, and cryptochlorogenic acid, neochlorogenic acid, caffeic acid, isochlorogenic acid B and isochlorogenic acid C are generation models.

[0147] (4.4) Establishing heat transfer model of extraction process

[0148] Same as Example 1.

[0149] (4.5) Establishing the mass transfer model of the extraction process

[0150] For chlorogenic acid and isochlorogenic acid A in honeysuckle, the corresponding degradation extraction model formulas (S11)-(S15) can be appropriately simplified according to the experimental results. K = x5, and the initial solution of the equation group is: t = 0, CW =0, there are 5 parameters that need to be fitted for each component.

[0151] For cryptochlorogenic acid and neochlorogenic acid, the corresponding extraction model formulas (S16)-(S20) were generated. According to the experimental results, the model parameters can be appropriately simplified, K = x5, C W,A The initial solution of the equations is: t = 0, C W =0, there are 5 parameters that need to be fitted for each component.

[0152] For caffeic acid, isochlorogenic acid B and isochlorogenic acid C, the corresponding extraction model formulas (S16)-(S20) are generated. According to the experimental results, the model parameters can be appropriately simplified, K = x5, C W,A is the isochlorogenic acid A content at different times, and the initial solution of the equation group is: t = 0, C W =0, there are 5 parameters that need to be fitted for each component.

[0153] (4.6) Optimization algorithm to solve temperature parameters

[0154] The objective function was defined as the mean square error of the experimental values ​​and the calculated values ​​of all groups. The genetic algorithm was used to optimize the parameters. The temperature parameters were brought into the equation to solve the curve of the internal temperature change of the extraction tank of the experimental group over time. The determination coefficient R of the fitting of the internal temperature change curve of the extraction tank during the extraction of 7 groups of honeysuckle 2 They are 0.9986, 0.9981, 0.9970, 0.9970, 0.9958, 0.9947, and 0.9970 respectively. The average determination coefficient is 0.9969. Overall, the fitting effect is good, which can reflect the temperature change law inside the extraction tank during the extraction process of honeysuckle.

[0155] (4.7) Optimization algorithm to solve mass transfer parameters

[0156] For the index component chlorogenic acid, the optimization algorithm was used to fit the parameters, and 5 of the parameters were fitted. The fitting results were brought into the equation to solve the kinetic curves of the experimental groups. The determination coefficient R of the chlorogenic acid fitting curves in the 7 groups of honeysuckle extraction 2 They are 0.9399, 0.9405, 0.9689, 0.9429, 0.9810, 0.9889 and 0.9831 respectively. The average determination coefficient is 0.9636. Overall, the fitting effect is good, which can reflect the composition change law of chlorogenic acid extraction process.

[0157] For the index component neochlorogenic acid, the optimization algorithm was used to fit the parameters, and 5 of the parameters were fitted. The fitting results were brought into the equation to solve the kinetic curves of the experimental groups. The determination coefficient R of the fitting curves of neochlorogenic acid in the extraction process of 7 groups of honeysuckle 2 They are 0.9522, 0.9474, 0.9799, 0.9436, 0.9891, 0.9885 and 0.9758 respectively. The average determination coefficient is 0.9681. Overall, the fitting effect is good, which can reflect the composition change law of the new chlorogenic acid extraction process.

[0158] For the index component chlorogenic acid, the optimization algorithm was used to fit the parameters, and 5 of the parameters were fitted. The fitting results were brought into the equation to solve the kinetic curves of the experimental groups. The determination coefficient R of the fitting curves of chlorogenic acid in the extraction process of 7 groups of honeysuckle 2 The results are 0.9326, 0.9750, 0.9697, 0.9778, 0.9837, 0.9767 and 0.9735 respectively. The average determination coefficient is 0.9699. Overall, the fitting effect is good, which can reflect the composition change law of the chlorogenic acid extraction process.

[0159] For the index component isochlorogenic acid A, the optimization algorithm was used to fit the parameters, and 5 of the parameters were fitted. The fitting results were brought into the equation to solve the kinetic curves of the experimental groups. The determination coefficient R of the fitting curve of isochlorogenic acid A in the process of honeysuckle extraction in 7 groups 2 The results are 0.9617, 0.9210, 0.9498, 0.8737, 0.9189, 0.9559 and 0.8930 respectively. The average determination coefficient is 0.9249. Overall, the fitting effect is good, which can reflect the composition change law of isochlorogenic acid A extraction process.

[0160] For the index component isochlorogenic acid B, the optimization algorithm was used to fit the parameters, and the parameters were fitted. The fitting results were brought into the equation to solve the kinetic curves of the experimental groups. The determination coefficient R of the isochlorogenic acid B fitting curves in the 7 groups of honeysuckle extraction 2 The results were 0.9901, 0.9976, 0.9920, 0.9839, 0.9913, 0.9946 and 0.9972 respectively. The average determination coefficient was 0.9924. Overall, the fitting effect was good, which could reflect the composition change law of isochlorogenic acid B extraction process.

[0161] For the index component isochlorogenic acid C, the optimization algorithm was used to fit the parameters, and 5 of the parameters were fitted. The fitting results were brought into the equation to solve the kinetic curves of the experimental groups. The determination coefficient R of the isochlorogenic acid C fitting curves of the 7 groups of honeysuckle extraction processes 2They are 0.9808, 0.9806, 0.9856, 0.9462, 0.9882, 0.9738 and 0.9433 respectively. The average determination coefficient is 0.9712. Overall, the fitting effect is good, which can reflect the composition change law of isochlorogenic acid C extraction process.

[0162] For the index component caffeic acid, the optimization algorithm was used to fit the parameters, and the parameters were fitted. The fitting results were brought into the equation to solve the kinetic curves of the experimental groups. The determination coefficient R of the caffeic acid fitting curves in the 7 groups of honeysuckle extraction 2 They are 0.8834, 0.9525, 0.9328, 0.8294, 0.9304, 0.9656 and 0.9649 respectively. The average determination coefficient is 0.9227. Overall, the fitting effect is good, which can reflect the composition change law of caffeic acid extraction process.

[0163] (4.8) Using the model to make predictions

[0164] The extraction process of another batch of medicinal materials is used as the prediction group. After knowing the temperature change of the heating medium, the temperature change curve of the extraction tank over time is predicted. Figure 8 . The prediction R of the prediction group 2 The temperature change law was then used to predict the changes in the content of different components in honeysuckle. Among them, chlorogenic acid predicted R 2 The predicted R of chlorogenic acid was 0.9501. 2 The new chlorogenic acid prediction R 2 is 0.9404, isochlorogenic acid A predicts R 2 is 0.9325, isochlorogenic acid B predicts R 2 is 0.9831, and isochlorogenic acid C predicts R 2 The predicted R of caffeic acid is 0.9714. 2 It is 0.8996. This shows that the model developed by the present invention can better fit and predict the changes in temperature and index components under different parameter conditions and different extraction processes (immersion extraction, temperature-raising extraction, and heat preservation extraction).

[0165] The embodiments described above provide a detailed description of the technical solutions of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, supplements or similar substitutions made within the scope of the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predicting the content of index components in a traditional Chinese medicine extraction process, characterized in that: The following steps are involved: (1) First, the parameters of the Chinese medicine raw materials are collected, and then by changing the parameters of the Chinese medicine extraction process, samples of Chinese medicine extracts obtained under different extraction conditions and at different extraction times are collected, and the information is recorded. The content of the index components in the Chinese medicine extract samples is determined, and the data set is obtained after data cleaning; The Chinese medicine extraction process includes soaking, heating and heat preservation processes, and the recorded information includes the extraction liquid temperature data and the heating medium temperature data; (2) establishing a heat transfer model and a mass transfer model for the extraction process, screening the mass transfer model for the extraction process, and using samples in the data set to train the heat transfer model and the mass transfer model for the extraction process to optimize the prediction model parameters; the mass transfer model for the extraction process is a dissolution model or a chemical reaction model, the dissolution model includes a single solvent extraction model or a mixed solvent extraction model, and the chemical reaction model includes a degradation extraction model or a generation extraction model; (3) Use the heat transfer model of the extraction process and the mass transfer model of the extraction process after parameter optimization to predict the content of index components in the traditional Chinese medicine extraction process.

2. The method for predicting the content of indicative components in the traditional Chinese medicine extraction process according to claim 1, characterized in that: The parameters of the Chinese medicine raw materials include the particle size of the medicinal material particles, the specific surface area of ​​the medicinal material particles, the content of index components, the volume of the wet medicinal material or the volume of the extract.

3. The method for predicting the content of indicative components in the traditional Chinese medicine extraction process according to claim 1, characterized in that: The recorded information also includes the particle size of the medicinal material, the batch of medicinal material, the extraction solvent, the solid-liquid ratio or the extraction time; the content of the index components in the Chinese medicine extract sample is determined by liquid chromatography.

4. The method for predicting the content of indicative components in the traditional Chinese medicine extraction process according to claim 1, characterized in that: The heat balance of the traditional Chinese medicine extraction process was carried out, and a heat transfer model of the extraction process was established to characterize the heat transfer of the heating medium, extraction container and extraction liquid.

5. The method for predicting the content of indicative components in the traditional Chinese medicine extraction process according to claim 1, characterized in that: The mass transfer model of the extraction process is used to characterize the changes in the content of the index components under different extraction conditions and different extraction times. The corresponding mass transfer model of the extraction process is selected according to the properties of the index components and different extraction solvents.

6. The method for predicting the content of index components in the traditional Chinese medicine extraction process according to claim 1, characterized in that: The mass transfer model of the screening extraction process is: when the index component extraction process is a component dissolution process and the extraction solvent is a pure solvent, the single solvent extraction model is selected; when the index component extraction process is a component dissolution process and the extraction solvent is a mixed solvent, the mixed solvent extraction model is selected; when the index component extraction process involves transformation or degradation of the index component, the degradation extraction model is selected for the index component that undergoes degradation, and the generation extraction model is selected for the index component generated by transformation.

7. The method for predicting the content of index components in the traditional Chinese medicine extraction process according to claim 1, characterized in that: Different samples from multiple groups of experiments in the data set are used as training sets. The heat transfer model of the extraction process and the mass transfer model of the extraction process are trained using the training sets to fit the appropriate parameters while satisfying the extraction process data obtained under different extraction conditions and at different extraction times. In order to improve the fitting effect of the fitting parameter values, the objective function is defined, and the optimization algorithm is used to optimize the parameters to be fitted so that the deviation between the actual data and the calculated data is minimized.

8. The method for predicting the content of index components in the extraction process of traditional Chinese medicine according to claim 7, characterized in that: The optimization algorithm is set to one of a genetic algorithm, a gray wolf optimization algorithm, and a Markov-Monte Carlo optimization algorithm. Further, the optimization algorithm is set to a genetic algorithm. The objective function is a determination coefficient, a mean square error, a root mean square error, or a mean absolute error. Further, the objective function is a mean square error.

9. A storage medium, characterized in that: The storage medium stores a program, wherein when the program is run, the method for predicting the content of an indicator component in the traditional Chinese medicine extraction process described in any one of claims 1-8 is executed.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: The processor executes the method for predicting the content of indicative components in the traditional Chinese medicine extraction process described in any one of claims 1-8 through the computer program.

Images