Traditional Chinese medicine multi-component pharmacy extracting and filtering device system
By designing a traditional Chinese medicine pharmaceutical device system that integrates multi-stage extraction and adaptive dynamic filtration, the problem of difficulty in efficiently extracting and isolating Chinese medicine multi-components is solved, and efficient and accurate extraction and filtration is achieved, ensuring the quality and efficacy of Chinese medicine preparations.
Patent Information
- Application Number
- CN202510196338.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-21
- Publication Date
- 2025-05-13
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional Chinese medicine extraction and filtration devices are difficult to efficiently extract and separate multi-components, with low efficiency and insufficient accuracy, resulting in loss of active ingredients and residues of impurities, affecting the quality and efficacy of Chinese medicine preparations.
A multi-component pharmaceutical extraction and filtration device system of traditional Chinese medicine is designed, including a multi-stage extraction module, a filtration module and a pressure sensing module. It adopts an adaptive dynamic filtration algorithm (ADFA) and partial least squares regression (PLS) model, combined with fuzzy logic control, dynamically adjusts the filter parameters to achieve real-time monitoring and precise control.
Through this device system, it is possible to efficiently extract and separate Chinese medicine multi-components, improve the extraction efficiency and filtration accuracy, reduce the loss of active ingredients and the residue of impurities, and ensure the quality and efficacy of Chinese medicine preparations.
Smart Images

Figure CN119989301A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of Chinese medicine pharmaceutical equipment, and in particular to a device system capable of efficiently extracting and filtering multiple components of Chinese medicine. Background Art
[0002] In the process of traditional Chinese medicine preparation, Chinese herbal medicines often contain multiple active ingredients. Traditional extraction and filtration devices can usually only process a single component or a limited number of components, which is difficult to meet the needs of simultaneous extraction and effective separation of multiple components. In addition, the existing extraction and filtration methods are inefficient and lack precision, which easily causes the loss of active ingredients and the residue of impurities, affecting the quality and efficacy of traditional Chinese medicine preparations. Therefore, it is of great significance to develop a new type of traditional Chinese medicine multi-component pharmaceutical extraction and filtration device system. Summary of the invention
[0003] A Chinese medicine multi-component pharmaceutical extraction and filtration device system comprises a multi-stage extraction module, a filtration module, and a pressure sensing module. The multi-stage extraction module is composed of 3-5 extraction tanks connected in series, and a stirring device is arranged in the extraction tank. The system adopts an adaptive dynamic filtration algorithm (ADFA), based on real-time monitoring of changes in solution components, and uses spectral analysis to establish a mathematical model of spectral data and solution information, and dynamically adjusts the filtration parameters accordingly.
[0004] For the mathematical model established by spectral analysis, partial least squares regression (PLS) is used. Let the spectral data matrix X (N×p, N is the number of samples, p is the number of spectral wavelength points), the concentration matrix C (N×q, q is the number of components), and the particle size distribution matrix D (N×r, r is the number of particle size intervals).
[0005] Furthermore, the traditional Chinese medicine multi-component pharmaceutical extraction and filtration device system,
[0006] Through the PLS algorithm, X is decomposed into X = TP T +E, where T is the score matrix, P is the load matrix, and E is the residual matrix. At the same time, the regression model of C and D with T is established C = TB C +F C and D = TB D +F D , where B C and B D is the regression coefficient matrix, F C and F D is the prediction residual matrix.
[0007] Furthermore, the traditional Chinese medicine multi-component pharmaceutical extraction and filtration device system,
[0008] In fuzzy logic control, the concentration of macromolecular impurities is set as I, the target active ingredient concentration is A, the pore size of the filter membrane is P, and the filtration pressure is F. A fuzzy rule base is established, such as "if I is high and A is low, then P is large and F is high", and accurate P and F values are obtained through fuzzy reasoning and defuzzification.
[0009] Furthermore, in the described traditional Chinese medicine multi-component pharmaceutical extraction and filtration device system, the extraction tank has a volume of 50-200 liters, the extraction tank is a double-layer jacket structure, the inner layer is made of 316L stainless steel, the outer layer is a heat-insulating material with a thickness of 5-10 cm, the stirring blade is a double-layer inclined blade structure, the rotation speed is adjustable within the range of 50-500 rpm, and a feed port and a solvent inlet are provided on the top of the extraction tank, the error of the automatic metering device of the feed port is within ±1%, and the flow regulating valve of the solvent inlet has a flow rate adjustable within the range of 0.5-5 liters / minute.
[0010] Furthermore, the traditional Chinese medicine multi-component pharmaceutical extraction and filtration device system includes a filtration module immediately following the multi-stage extraction module, wherein the filtration module adopts a filtration method combining a ceramic membrane and an ultrafiltration membrane, wherein the pore size of the ceramic membrane is 0.1-0.5 microns, and the molecular weight cutoff of the ultrafiltration membrane is 5000-50000 Daltons.
[0011] Furthermore, in the traditional Chinese medicine multi-component pharmaceutical extraction and filtration device system, a pressure sensor with a pressure range of 0.1-1 MPa is provided in the filtration module. When the pressure exceeds the set value, a gas-liquid mixed backwashing program can be automatically started. The gas is compressed air of 0.2-0.5 MPa, and the liquid is a pure solution identical to the extraction solvent. The flushing time is 5-10 minutes.
[0012] Furthermore, in the described traditional Chinese medicine multi-component pharmaceutical extraction and filtration device system, the extraction tank volume of the integrated extraction system is 100-300 liters, the tank body is made of a glass-carbon fiber composite material, and the tank is provided with a spiral agitator with an adjustable speed within the range of 100-600 rpm and a surface coated with polytetrafluoroethylene. The bottom is provided with a heating plate with a temperature control accuracy of ±0.5°C and a temperature range of 30-90°C, and the top is provided with a solvent and raw material inlet of a flow metering device with an error within ±0.5%.
[0013] Furthermore, in the traditional Chinese medicine multi-component pharmaceutical extraction and filtration device system, the adaptive dynamic filtration module is equipped with an intelligent filtration membrane with adjustable pore size, which is composed of a polymer-inorganic nanomaterial composite, with an initial pore size of 0.2-0.8 microns, which can be dynamically adjusted within the range of 0.1-1 microns, and is equipped with a pressure sensor with a measuring range of 0-1.5 MPa and an accuracy of ±0.01 MPa and a flow sensor with a measuring range of 0-10 liters / minute and an accuracy of ±0.1 liters / minute.
[0014] Furthermore, in the Chinese medicine multi-component pharmaceutical extraction and filtration device system, the adaptive dynamic filtration module is provided with an automatic cleaning device. When the filtration efficiency is lower than 80% of the initial efficiency, a cleaning program with an ultrasonic vibration frequency of 20-50 kHz and a mild surfactant solution is automatically started, and the cleaning time is 3-8 minutes.
[0015] Furthermore, the present invention also provides a method for extracting and filtering a multi-component Chinese medicine, comprising feeding a Chinese medicine raw material into a first-stage extraction tank through a feed inlet, injecting a solvent determined by an MCOA algorithm from a solvent inlet, and starting a stirring device and a temperature control system so that the temperature and stirring speed reach the optimal values determined by the algorithm, wherein the MCOA algorithm is:
[0016] Assume that the extraction parameter vector is x=(x1, x2, ..., x n ), where x1 represents temperature, x2 represents pressure, x3 represents solvent flow, etc. The target vector is y = (y1, y2, ..., y m ), where y1 is the extraction rate of a certain component, y2 is the purity of the component, y3 is the total processing time, etc.
[0017] The mathematical model established is in the form of y=f(x), where f is a complex nonlinear function approximated by a neural network. The structure of the neural network can be expressed as where υ ik is the weight from the input layer to the hidden layer, w ji is the weight from the hidden layer to the output layer, b i and c j are the biases of the hidden layer and the output layer respectively, σ is the activation function (such as Sigmoid function or ReLU function), is the activation function of the output layer (selected according to the properties of the target value, such as a linear function for real-number targets such as extraction rate), and l is the number of hidden layer nodes.
[0018] Furthermore, in the method described above, the solution flows by gravity into the next-stage extraction tank after multi-stage extraction, and enters the filtration module after completing the multi-stage extraction, and is first filtered through a ceramic membrane and then through an ultrafiltration membrane to obtain a filtrate containing target multiple components, and the filtrate is used in subsequent pharmaceutical processes.
[0019] Beneficial effects of the invention:
[0020] The multiple key objectives (extraction rate, purity, processing time) of multi-component extraction of traditional Chinese medicine are integrated, and the optimal extraction parameters are solved by combining complex mathematical models and intelligent algorithms. The algorithm fully considers the physical and chemical properties and interactions of each component of traditional Chinese medicine, and can automatically adjust the parameters according to different traditional Chinese medicine formulas and goals to achieve personalized and precise extraction control. The spectral analysis technology is innovatively integrated with the fuzzy logic control algorithm. The solution component information model is established through real-time spectral scanning, and the concentration and particle size distribution changes of each component in the solution can be dynamically tracked. Based on fuzzy logic, the filtration parameters are accurately adjusted according to the real-time state of the solution, avoiding the loss of active ingredients or impurity residues caused by traditional fixed parameter filtration, and greatly improving the adaptability and accuracy of filtration. It ensures efficient, stable and precise control of the extraction process, and can effectively improve the extraction efficiency of components of different properties. The filtration method of the combination of ceramic membrane and ultrafiltration membrane realizes staged and efficient filtration according to the characteristics of impurities and active ingredients with different particle sizes and molecular weights. At the same time, the backwashing program controlled by the pressure sensor can effectively prevent membrane clogging, extend the service life of the membrane, and ensure long-term stable filtration. DETAILED DESCRIPTION
[0021] Figure 1 The figure is a flow chart of the algorithm principle of the system of the present invention.
[0022] Example 1
[0023] In the initialization stage, the initial extraction parameter range is set according to the type of Chinese medicine and the expected components to be extracted. For example, the temperature range is set at [20-80]°C, the pressure range is [normal pressure-5 atmospheres], and the solvent type is selected from a variety of common solvents such as ethanol, water, and their mixtures.
[0024] In each iteration, the neural network model in machine learning is used to predict the target values under the current extraction parameters. The input layer nodes of the neural network are the extraction parameters, and the output layer nodes are the target values. The neural network is trained with a large amount of pre-experimental data so that it can accurately reflect the relationship between parameters and targets.
[0025] According to the prediction results, the selection, crossover and mutation operations in the genetic algorithm are used to generate new extraction parameter combinations. The selection operation selects excellent parameter combinations according to the fitness function (comprehensively considering the pros and cons of each target value), the crossover operation exchanges some parameters of the excellent combination, and the mutation operation makes small-scale random changes to some parameters to increase the diversity of the search space.
[0026] The above steps are repeated until the preset convergence condition is met (such as the target value change is less than a certain threshold after multiple consecutive iterations) to obtain the optimal extraction parameter combination.
[0027] The device system structure and parameters, multi-stage extraction module: It consists of [3-5] extraction tanks connected in series, and the volume of each extraction tank is [50-200] liters. The extraction tank adopts a double-layer jacket structure, the inner layer is [316L stainless steel], and the outer layer is a heat-insulating material (such as polyurethane foam, with a thickness of [5-10] cm), which can achieve precise temperature control. In each extraction tank, there is a stirring device, and the stirring blade adopts a [double-layer inclined blade structure], and the speed can be adjusted within the range of [50-500] rpm to ensure that the solvent is in full contact with the Chinese medicine raw materials. There are a feed port and a solvent inlet on the top of the extraction tank. The feed port is equipped with an automatic metering device, which can accurately control the input amount of Chinese medicine raw materials, and the error range is within ± [1]%. The solvent inlet is connected to the solvent storage tank and is equipped with a flow control valve, and the flow rate can be adjusted within the range of [0.5-5] liters / minute.
[0028] Filtration module: Immediately after the last stage of extraction tank, a [combination of ceramic membrane and ultrafiltration membrane filtration method] is used. The pore size of the ceramic membrane is [0.1-0.5] microns, which is used to preliminarily remove larger particles of impurities and insoluble substances. The molecular weight cutoff of the ultrafiltration membrane is [5000-50000] Daltons, which can effectively intercept large molecular impurities and undissolved Chinese medicine components, while allowing the target small molecule active ingredients to pass through. A pressure sensor is installed inside the filtration module to monitor the pressure changes during the filtration process in real time. The pressure range can be between [0.1-1] MPa. When the pressure exceeds the set value, the system automatically starts the backwashing program to prevent the membrane from clogging. Backwashing adopts [gas-liquid mixed flushing method], the gas is [compressed air, pressure of 0.2-0.5 MPa], the liquid is [pure solution of the same extraction solvent], and the flushing time is [5-10] minutes.
[0029] Operation process: First, according to the Chinese medicine formula and the multi-component information to be extracted, the Chinese medicine raw materials are put into the first-stage extraction tank through the feed port, and at the same time, a suitable solvent (determined by the MCOA algorithm) is injected from the solvent inlet according to the set flow rate. Start the stirring device and temperature control system to make the temperature and stirring speed in the extraction tank reach the optimal value determined by the algorithm. After the first-stage extraction, the solution flows into the second-stage extraction tank by gravity, and the above extraction process is repeated, and so on, to complete the multi-stage extraction. The solution after extraction enters the filtration module, first filtered through a ceramic membrane, and then filtered through an ultrafiltration membrane to obtain a clear filtrate containing the target multi-components. The filtrate can be further subjected to subsequent pharmaceutical processes such as concentration and drying.
[0030] Embodiment 2,
[0031] (II) Adaptive dynamic filtration integrated extraction system algorithm principle: ADFA is designed based on real-time monitoring of changes in solution composition during the extraction process of traditional Chinese medicine. It uses spectral analysis technology (such as near-infrared spectroscopy, ultraviolet-visible spectroscopy, etc.) to scan the solution in real time during the extraction process. By establishing a mathematical model of spectral data and information such as the concentration of each component in the solution and particle size distribution (using methods such as partial least squares regression), the state of the solution can be obtained in real time. Then, the filtration parameters (such as the pore size of the filter membrane, filtration pressure, etc.) are dynamically adjusted according to the state of the solution.
[0032] Algorithm steps: Before the extraction begins, the Chinese medicine raw materials and the initial solvent are spectrally scanned to establish an initial spectral database. At the same time, the initial filtration parameters are set according to experience, such as the initial pore size of the filter membrane is [0.2-0.8] microns and the initial filtration pressure is [0.15-0.3] MPa. During the extraction process, the solution is spectrally scanned every [1-5] minutes, and the spectral data obtained by the scan is input into the pre-established mathematical model to obtain the concentration change and particle size distribution information of each component in the solution. According to the change of the solution state, the fuzzy logic control algorithm is used to adjust the filtration parameters. For example, if the concentration of macromolecular impurities in the solution increases, the fuzzy logic controller will increase the pore size of the filter membrane or increase the filtration pressure according to the preset rules (established based on expert experience and experimental data) to ensure the filtration effect. At the same time, in order to avoid the loss of active ingredients due to excessive filtration, when the concentration of the target active ingredient is detected to be close to the critical value, the filtration intensity is appropriately reduced. The filtration parameters are continuously monitored and adjusted until the extraction and filtration process is completed.
[0033] The device system structure and parameters integrated extraction module: adopts an integrated extraction tank design with a volume of [100-300] liters. The tank body is made of [glass-carbon fiber composite material], which has good corrosion resistance and allows light of a certain wavelength range to pass through, which is convenient for spectral scanning. A spiral agitator is installed in the extraction tank. The agitator speed can be adjusted within the range of [100-600] rpm. The agitator surface is coated with [polytetrafluoroethylene coating] to prevent the adsorption of traditional Chinese medicine ingredients. A heating plate is provided at the bottom of the extraction tank, which can achieve precise control of the temperature within the range of [30-90]℃, and the temperature control accuracy is ±[0.5]℃. The top of the extraction tank is provided with a solvent and raw material inlet, and a flow metering device is provided at the inlet to accurately control the amount of solvent and raw material added, with an error within ±[0.5]%.
[0034] Adaptive dynamic filtration module: The filtration module is directly connected to the extraction tank and is equipped with an intelligent filtration membrane with adjustable pore size. The filtration membrane is made of [polymer-inorganic nanomaterial composite], and its initial pore size can be in the range of [0.2-0.8] microns. By applying different voltage signals (controlled by the ADFA algorithm), the pore size can be dynamically adjusted in the range of [0.1-1] microns. The filtration module is equipped with a high-precision pressure sensor and a flow sensor to monitor the filtration pressure and filtrate flow in real time. The pressure sensor has a measurement range of [0-1.5] MPa and an accuracy of ± [0.01] MPa; the flow sensor has a measurement range of [0-10] liters / minute and an accuracy of ± [0.1] liters / minute. According to the data fed back by the sensor, the ADFA algorithm adjusts the filtration parameters. The filtration module is also equipped with an automatic cleaning device. When the filtration efficiency drops to a certain threshold (such as less than
[80] % of the initial efficiency), the cleaning program is automatically started. The cleaning method adopts [a combination of ultrasonic vibration and chemical cleaning agent], the chemical cleaning agent is [mild surfactant solution], the ultrasonic vibration frequency is [20-50] kilohertz, and the cleaning time is [3-8] minutes.
[0035] Operation process: Add the Chinese medicine raw materials and solvents into the integrated extraction tank through the inlet according to the set ratio, start the agitator and heating plate, and extract the solution at a suitable temperature and stirring speed. During the extraction process, the adaptive dynamic filtration module filters the solution in real time. The ADFA algorithm dynamically adjusts the filter membrane pore size and filtration pressure according to the spectral analysis results and sensor feedback data to ensure the filtration effect. After the extraction and filtration is completed, the filtrate can be used in the subsequent Chinese medicine pharmaceutical process.
[0036] Example 3
[0037] By comparing the traditional device with the two solution devices of the present invention, a detailed explanation is given from the aspects of algorithm model, extraction unit structure, filtration unit structure, multi-component extraction efficiency, filtration effect, active ingredient purity and operation flexibility.
[0038]
[0039]
[0040]
[0041] It can be seen that the multiple key objectives (extraction rate, purity, processing time) of multi-component extraction of traditional Chinese medicine are integrated, and the optimal extraction parameters are solved by combining complex mathematical models and intelligent algorithms. The nonlinear relationship between parameters and targets is accurately fitted by neural networks, and the optimal solution is efficiently searched by genetic algorithms, breaking through the limitations of traditional single parameter optimization or empirical extraction. The algorithm fully considers the physicochemical properties and interactions of each component of traditional Chinese medicine, and can automatically adjust parameters according to different traditional Chinese medicine formulas and goals to achieve personalized and precise extraction control. The spectral analysis technology is innovatively integrated with the fuzzy logic control algorithm. The solution component information model is established by real-time spectral scanning, and the changes in the concentration and particle size distribution of each component in the solution can be dynamically tracked. Based on fuzzy logic, the filtration parameters are accurately adjusted according to the real-time state of the solution, avoiding the loss of active ingredients or impurity residues caused by traditional fixed parameter filtration, greatly improving the adaptability and accuracy of filtration. The design of 3-5 extraction tanks in series is adopted, and the volume of each extraction tank, the double-layer jacket structure (316L stainless steel inner layer and insulation outer layer), the double-layer inclined blade stirring blade with adjustable speed, the high-precision automatic metering feed port and the adjustable flow solvent inlet ensure the efficient, stable and precise control of the extraction process, which can effectively improve the extraction efficiency of components with different properties. The filtration method of the combination of ceramic membrane and ultrafiltration membrane realizes phased and efficient filtration according to the characteristics of impurities and active ingredients with different particle sizes and molecular weights. At the same time, the backwashing program controlled by the pressure sensor can effectively prevent membrane clogging, extend the service life of the membrane, and ensure long-term stable filtration. The integrated glass-carbon fiber composite material extraction tank is both corrosion-resistant and light-transmissive, meeting the needs of spectral scanning. The high-precision control of the spiral agitator and heating plate and the inlet flow metering device ensure the accuracy of the extraction conditions, providing a good foundation for subsequent adaptive filtration. The intelligent filter membrane with adjustable pore size is a highlight. It dynamically changes the pore size through voltage signals. Combined with high-precision pressure and flow sensors and automatic cleaning devices, it can accurately filter according to real-time changes in solution composition, ensuring filtering effect and efficiency while reducing manual intervention and maintenance costs.
Claims
1. A Chinese medicine multi-component pharmaceutical extraction and filtration device system, characterized in that: The system comprises a multi-stage extraction module, a filtration module and a pressure sensing module. The multi-stage extraction module is composed of 3-5 extraction tanks connected in series. A stirring device is arranged inside the extraction tank. The system adopts an adaptive dynamic filtration algorithm (ADFA). Based on the real-time monitoring of the changes in the solution components, a mathematical model of spectral data and solution information is established by spectral analysis, and the filtration parameters are dynamically adjusted accordingly. For the mathematical model established by spectral analysis, partial least squares regression (PLS) is used. Let the spectral data matrix X (N×p, N is the number of samples, p is the number of spectral wavelength points), the concentration matrix C (N×q, q is the number of components), and the particle size distribution matrix D (N×r, r is the number of particle size intervals).
2. The Chinese medicine multi-component pharmaceutical extraction and filtration device system according to claim 1, characterized in that: Through the PLS algorithm, X is decomposed into X = TP T +E, where T is the score matrix, P is the loading matrix, and E is the residual matrix. Simultaneously establish the regression model of C and D with T C = TB C +F C and D = TB D +F D , where B C and B D is the regression coefficient matrix, F C and F D is the prediction residual matrix.
3. The Chinese medicine multi-component pharmaceutical extraction and filtration device system according to claim 1, characterized in that: In fuzzy logic control, let the concentration of macromolecular impurities be I, the target active ingredient concentration be A, the pore size of the filter membrane be P, and the filtration pressure be F. A fuzzy rule base is established, for example, "if I is high and A is low, then P is large and F is high", and accurate P and F values are obtained through fuzzy reasoning and defuzzification.
4. The Chinese medicine multi-component pharmaceutical extraction and filtration device system according to claim 1, characterized in that: The extraction tank has a volume of 50-200 liters, a double-layer jacket structure, an inner layer of 316L stainless steel, an outer layer of a heat-insulating material with a thickness of 5-10 cm, a double-layer inclined-blade structure, and an adjustable rotation speed within the range of 50-500 rpm. A feed port and a solvent inlet are provided on the top of the extraction tank, the error of the automatic metering device of the feed port is within ±1%, and the flow rate of the flow regulating valve of the solvent inlet is adjustable within the range of 0.5-5 liters / minute.
5. The Chinese medicine multi-component pharmaceutical extraction and filtration device system according to claim 1, characterized in that: The invention comprises a filtration module immediately following the multi-stage extraction module, wherein the filtration module adopts a filtration mode of a combination of a ceramic membrane and an ultrafiltration membrane, wherein the pore size of the ceramic membrane is 0.1-0.5 micron, and the molecular weight cutoff of the ultrafiltration membrane is 5000-50000 Dalton.
6. The Chinese medicine multi-component pharmaceutical extraction and filtration device system according to claim 1, characterized in that: The filter module is provided with a pressure sensor with a pressure range of 0.1-1 MPa. When the pressure exceeds the set value, the gas-liquid mixed backwashing program can be automatically started. The gas is 0.2-0.5 MPa compressed air, and the liquid is a pure solution identical to the extraction solvent. The flushing time is 5-10 minutes.
7. The Chinese medicine multi-component pharmaceutical extraction and filtration device system according to claim 1, characterized in that: The extraction tank of the integrated extraction system has a volume of 100-300 liters, and the tank body is made of a glass-carbon fiber composite material. The tank is equipped with a spiral agitator with an adjustable speed within the range of 100-600 rpm and a surface coated with polytetrafluoroethylene. The bottom is equipped with a heating plate with a temperature control accuracy of ±0.5°C and a temperature range of 30-90°C, and the top is equipped with a flow metering device with an error within ±0.5% for solvent and raw material inlets.
8. The Chinese medicine multi-component pharmaceutical extraction and filtration device system according to claim 1, characterized in that: The adaptive dynamic filtration module is provided with an automatic cleaning device. When the filtration efficiency is lower than 80% of the initial efficiency, a cleaning procedure with an ultrasonic vibration frequency of 20-50 kHz and a mild surfactant solution is automatically started for 3-8 minutes.
9. A method for extracting and filtering multi-component pharmaceutical preparation of traditional Chinese medicine, characterized in that: The method comprises feeding the Chinese medicine raw material into the first-stage extraction tank through the feed port, injecting the solvent determined by the MCOA algorithm from the solvent inlet, starting the stirring device and the temperature control system, so that the temperature and stirring speed reach the optimal values determined by the algorithm, and the MCOA algorithm is: Assume that the extraction parameter vector is x=(x1, x2, ..., x n ), where x1 represents temperature, x2 represents pressure, x3 represents solvent flow, etc. The target vector is y = (y1, y2, ..., y m ), where y1 is the extraction rate of a certain component, y2 is the purity of the component, y3 is the total processing time, etc. The mathematical model established is in the form of y=f(x), where f is a complex nonlinear function approximated by a neural network. The structure of the neural network can be expressed as where v ik is the weight from the input layer to the hidden layer, w ji is the weight from the hidden layer to the output layer, b i and c j are the biases of the hidden layer and the output layer respectively, σ is the activation function (such as Sigmoid function or ReLU function), is the activation function of the output layer (selected according to the properties of the target value, such as a linear function for real-number targets such as extraction rate), and l is the number of hidden layer nodes.
10. The method according to claim 9, characterized in that After multi-stage extraction, the solution flows by gravity into the next-stage extraction tank. After completing multi-stage extraction, it enters the filtration module, is first filtered through a ceramic membrane and then filtered through an ultrafiltration membrane to obtain a filtrate containing target multiple components. The filtrate is used in subsequent pharmaceutical processes.