Traditional Chinese medicine alcohol precipitation regulation and control system and dynamic regulation and control method based on real-time component detection

Abstract

The invention belongs to the technical field of online detection of an alcohol precipitation process. The traditional Chinese medicine alcohol precipitation regulation and control system comprises an alcohol precipitation tank, a stirring motor, a detection liquid medicine input pipeline, a detection liquid medicine return pipeline and a buoy sampling mechanism, an output shaft of the stirring motor is connected with a stirring paddle in the alcohol precipitation tank, and a cooling pipeline is arranged at the bottom of the alcohol precipitation tank; the buoy sampling mechanism comprises a sampling buoy, a sampling suction pipe and a positioning element, the positioning element is fixedly connected with the inner wall of the alcohol precipitation tank, the sampling buoy is arranged on the positioning element in a sleeving manner and can move up and down along the positioning element along with the change of the liquid level of the liquid medicine, the bottom of an internal channel of the sampling buoy is connected with the sampling suction pipe, and a plurality of sampling holes are formed in the sampling suction pipe; and the top of the internal channel of the sampling buoy is connected with the sampling hose. Errors caused by traditional experience judgment are avoided, and the production efficiency of the alcohol precipitation process of the traditional Chinese medicine oral preparation is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of online detection of alcohol precipitation process, and in particular to a traditional Chinese medicine alcohol precipitation regulation system and a dynamic regulation method based on real-time detection of components. BACKGROUND

[0002] The statements in this section merely provide background technology related to the present application and do not necessarily constitute prior art.

[0003] The principle of alcohol precipitation of traditional Chinese medicine preparation is to use the difference in solubility of various components in the traditional Chinese medicine extract in ethanol, add ethanol to the traditional Chinese medicine extract, precipitate impurities, and separate and purify effective components. The main purpose is to effectively remove starch, protein, mucilage, gum, polysaccharide and other ineffective components in the traditional Chinese medicine extract, improve the purity of the preparation, reduce the turbidity, precipitation and other phenomena that may occur during storage of the preparation, and prolong the shelf life of the drug.

[0004] In the detection of alcohol precipitation liquor of traditional Chinese medicine preparation, the sampling process of the liquor directly affects whether the sample fully reflects the components of the liquor in the alcohol precipitation tank, and also affects the real-time and stability of the alcohol precipitation liquor sampling. During the alcohol precipitation process, the liquid level changes and is disturbed by the stirring paddle. The current alcohol precipitation liquor sampling scheme is to extract liquor at a fixed position. However, the fixed alcohol precipitation sampling position cannot fully reflect the overall quality of the alcohol precipitation liquor, the interference factors (interference of liquid level change and stirring paddle) affect the quality of the extracted alcohol precipitation liquor, and thus affect the detection accuracy. Moreover, the current online detection method has the problem of poor real-time data analysis. The traditional alcohol precipitation near-infrared spectrum detection system only simply collects and records spectral information, cannot perform real-time analysis on the collected spectral data, and when the raw material batch is changed, the original detection model is disturbed, resulting in a decrease in detection accuracy. The existing scheme cannot quickly adjust the process parameter mechanism, resulting in low alcohol precipitation efficiency. SUMMARY

[0005] In order to solve the problems of the prior art, the present application provides a traditional Chinese medicine alcohol precipitation regulation system and a dynamic regulation method based on real-time detection of components, which realizes real-time detection and quality control of the alcohol precipitation process, avoids errors caused by traditional experience judgment, improves the production efficiency of the alcohol precipitation process of traditional Chinese medicine oral preparation, effectively reduces the impurity residues of the alcohol precipitation liquor, and ensures the consistency of the drug quality.

[0006] In order to achieve the above purpose, the present application adopts the following technical scheme: In a first aspect, the present application provides a traditional Chinese medicine alcohol precipitation regulation system.

[0007] The traditional Chinese medicine alcohol precipitation regulation system comprises an alcohol precipitation tank, a stirring motor, a detection liquid input pipeline, a detection liquid return pipeline and a floating sampling mechanism, the output shaft of the stirring motor is connected with a stirring paddle in the alcohol precipitation tank, and a cooling pipeline is arranged at the bottom of the alcohol precipitation tank; The floating sampling mechanism comprises a sampling float, a sampling suction pipe and a positioning element, the positioning element is fixedly connected with the inner wall of the alcohol precipitation tank, the sampling float is sleeved on the positioning element and can move up and down along the positioning element with the change of the liquid level of the liquid, the bottom of the internal passage of the sampling float is connected with the sampling suction pipe, a plurality of sampling holes are formed in the sampling suction pipe, the top of the internal passage of the sampling float is connected with a sampling hose, and one end of the sampling hose is connected with the detection liquid input pipeline. The other end of the detection liquid input pipeline is connected with a first electromagnetic valve, a peristaltic pump, a manual sampling valve, a flow cell and a second electromagnetic valve in sequence through pipelines, one end of the detection liquid return pipeline is connected with the second electromagnetic valve through a pipeline, the other end of the detection liquid return pipeline is connected with the alcohol precipitation tank, and the position of the flow cell is arranged with an optical fiber connected with a spectrometer for detection.

[0008] In an implementation form of the first aspect of the present application, the positioning element is a positioning wire, the positioning wire is welded and fixed on the inner wall of the alcohol precipitation tank, the sampling float is sleeved on the positioning wire, the top of the internal passage of the sampling float is connected with the sampling hose through a vice, and the bottom of the internal passage of the sampling float is threadedly connected with the top of the sampling suction pipe.

[0009] In an implementation form of the first aspect of the present application, a plurality of sampling holes are arranged in the upper part, the middle part and the lower part of the sampling suction pipe in the circumferential direction, and the inner diameters of the sampling holes increase from top to bottom.

[0010] In an implementation form of the first aspect of the present application, the spectrometer is arranged in a detection explosion-proof cabinet, a processor connected with the spectrometer is further arranged in the detection explosion-proof cabinet, and the processor is in communication connection with the peristaltic pump, the stirring motor, the first electromagnetic valve and the second electromagnetic valve. The first electromagnetic valve, the peristaltic pump, the manual sampling valve, the flow cell and the second electromagnetic valve are arranged in a detection circulation pipeline work cabinet, and the spectrometer is inserted into the flow cell through an optical fiber to realize liquid detection.

[0011] In an implementation form of the first aspect of the present application, a membrane filter is connected on the pipeline between the other end of the detection liquid input pipeline and the first electromagnetic valve.

[0012] Secondly, the present application provides a dynamic regulation method based on real-time detection of components.

[0013] A dynamic regulation method based on real-time detection of components, which utilizes the traditional Chinese medicine alcohol precipitation regulation system of the first aspect of the present application, comprises the following processes: The peristaltic pump is opened, the first electromagnetic valve and the second electromagnetic valve are powered on and opened, the sampling float and the sampling suction tube suck the medicinal liquid in the alcohol precipitation tank into the detection medicinal liquid input pipeline, and the medicinal liquid covers the flow cell; The peristaltic pump is closed, the first electromagnetic valve and the second electromagnetic valve are powered off and closed, the medicinal liquid in the detection medicinal liquid input pipeline and the detection medicinal liquid return pipeline is static, and the spectrometer collects spectral data of the medicinal liquid in the flow cell through an optical fiber; The first electromagnetic valve is powered on and opened, the second electromagnetic valve is powered off and closed, the manual sampling valve is opened for offline sampling, the medicinal liquid is stored in a test tube corresponding to a label, and after offline sampling is completed, the manual sampling valve is closed. The first electromagnetic valve and the second electromagnetic valve are powered on and opened, and the medicinal liquid returns to the alcohol precipitation tank through the detection medicinal liquid return pipeline. The sample is taken cyclically at a specified time interval, liquid chromatography analysis is performed on the sample, a spectral analysis model is established according to a result of the liquid chromatography analysis, the spectral analysis model is used for online detection of the content of the medicine, process parameters are dynamically adjusted according to a result of the online detection of the content of the medicine, and end point judgment is performed.

[0014] In an implementation form of the second aspect of the present application, liquid chromatography analysis is performed on the sample, and a spectral analysis model is established according to a result of the liquid chromatography analysis, including: The medicinal liquid in the test tube is analyzed by liquid chromatography to obtain spectral data, and a spectral data preprocessing algorithm is used to preprocess the spectral data; The spectral data after preprocessing is combined to train a convolutional neural network, so that the convolutional neural network can output a detection quality marker, an ethanol concentration, a precipitate particle size and a water concentration according to the spectral data, and the convolutional neural network after training is used as the spectral analysis model.

[0015] As a further limitation of the second aspect of the present application, when the batch of raw materials changes in the alcohol precipitation process of traditional Chinese medicine, the spectral data of the alcohol precipitation medicinal liquid of the new batch of traditional Chinese medicine preparation is input into the shared convolutional layer of the convolutional neural network, a feature vector output by the shared convolutional layer is extracted, and a maximum mean difference between the feature vector of the new batch of medicinal liquid and the feature vector of the historical batch of medicinal liquid is calculated. When the maximum mean difference exceeds a set threshold, the shared convolutional layer of the convolutional neural network is fixed, the fully connected layer of the convolutional neural network is fine-tuned according to label data of the new batch of raw materials, a mapping relationship from the feature space of the new batch of medicinal liquid to the quality index is learned, and a spectral analysis model after transfer learning is obtained.

[0016] In an implementation form of the second aspect of the present application, when the liquor circulates in the detection liquor input pipeline and the detection liquor return pipeline, the spectral information of the liquor in the flow cell is collected, the built-in spectral analysis model is used to detect the quality marker, the ethanol concentration, the precipitate particle size and the water concentration, and the output of the spectral analysis model is used to control the flow of the alcohol metering pump, the rotating speed of the stirring motor and the flow of the cooling water.

[0017] As a further limitation of the second aspect of the present application, when the liquor enters the static precipitation stage, the liquor temperature is detected, the liquor temperature is controlled by controlling the flow of the cooling water in the cooling pipeline, the end point is judged according to the detected effective substance content and impurity content, and the alcohol precipitation process is ended when the end point is reached. The end point judgment includes that each quality marker concentration reaches a preset value in N consecutive sampling periods, and the impurity content is less than a preset value, then the end point is reached, and the alcohol precipitation is determined to be ended, wherein N is a positive integer greater than 2.

[0018] Compared with the prior art, the present application has the following advantages: The present application innovatively designs a sampling float mechanism, which has the function of automatically adjusting the position with the liquid level, ensuring that the alcohol precipitation sample at a specified height from the liquid level can be extracted. The sampling tube in the sampling float mechanism adopts a multi-hole structure design, and the hole diameter of the sampling tube increases from top to bottom, which can comprehensively collect the liquor sample information at different positions in the alcohol precipitation tank, better reflect the internal quality of the alcohol precipitation liquid in the whole alcohol precipitation tank, and at the same time, a positioning steel wire 10 is installed in the alcohol precipitation tank to position the sampling float, avoiding the interference of the stirring paddle and improving the precision of online detection.

[0019] The present application integrates the sampling detection circulation pipeline design, online near-infrared detection function and explosion-proof design, and the sampling detection circulation pipeline is composed of components such as sampling float mechanism, filter, peristaltic pump and electromagnetic valve, which ensures the continuous circulation and update of the alcohol precipitation liquor, the spectrometer collects the liquor spectral information of the flow cell with the help of optical fiber, and the offline sampling function is also provided on the circulation pipeline to realize offline sampling analysis of the liquor, providing data support for the near-infrared spectral analysis model. The spectrometer collects the liquor spectrum in real time, the processor processes the spectral data in real time to predict the quality marker content, and the content change trend is displayed through the display. At the same time, due to the particularity of the alcohol precipitation environment, the online detection related components of the alcohol precipitation liquor are designed to be explosion-proof, and the spectrometer, processor and other components are separated from the sampling detection circulation pipeline in the explosion-proof cabinet, realizing modular design and meeting the operation safety in special environment.

[0020] The present application dynamically controls three process parameters of alcohol addition amount, stirring speed and cooling water temperature in alcohol precipitation process according to the online detected effective component content of alcohol precipitation medicine liquid and collected process parameter data, reduces resource loss, improves production efficiency, judges the end point when the effective component content in alcohol precipitation medicine liquid changes, overcomes the deficiency of traditional experience judgment, and improves consistency and reliability of multiple batches of alcohol precipitation medicine liquid Advantages of the additional aspects of the application will be partly given in the following description, partly will become obvious from the following description, or will be known by the practice of the application. BRIEF DESCRIPTION OF DRAWINGS

[0021] The drawings accompanying the specification of this application form a part thereof, serve to provide further understanding of the application, and together with the description of the exemplary embodiments of the application given below, explain the application, and do not constitute an improper limitation of the application.

[0022] Figure 1 The overall schematic diagram of the traditional Chinese medicine alcohol precipitation control system provided for an exemplary embodiment of the present application is shown in the figure; Figure 2 The working principle schematic diagram of the traditional Chinese medicine alcohol precipitation control system provided for an exemplary embodiment of the present application is shown in the figure; Figure 3 The overall schematic diagram of the sampling float mechanism provided for an exemplary embodiment of the present application is shown in the figure; Figure 4 The connection schematic diagram of the sampling float mechanism provided for an exemplary embodiment of the present application is shown in the figure; Figure 5 The schematic diagram of the sampling suction tube provided for an exemplary embodiment of the present application is shown in the figure; Figure 6 The end point judgment flow chart provided for an exemplary embodiment of the present application is shown in the figure.

[0023] 1, stirring motor; 2, platform; 3, alcohol precipitation tank; 4, detection circulation pipeline work cabinet; 5, optical cable; 6, detection explosion-proof cabinet; 7, detection medicine liquid backflow pipeline; 8, detection medicine liquid input pipeline; 9, sampling float; 10, positioning steel wire; 11, sampling suction tube; 12, first stirring paddle; 13, second stirring paddle; 14, membrane filter; 15, first electromagnetic valve; 16, peristaltic pump; 17, manual sampling valve; 18, flow cell; 19, spectrometer; 20, optical fiber; 21, second electromagnetic valve; 22, caliper; 23, hose; 24, upper sampling hole; 25, middle sampling hole; 26, lower sampling hole; 27, positioning steel wire hole; 28, cooling pipeline. DETAILED DESCRIPTION

[0024] The application will be further described below in combination with the drawings and embodiments.

[0025] It should be noted that the following detailed description is exemplary in nature and is intended to provide further description of the application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0026] There are still many challenges in the current preparation process of traditional Chinese medicine products. The production process of traditional Chinese medicine preparation involves drug extraction, concentration, and alcohol precipitation, etc. The process of traditional Chinese medicine preparation is complex, with multiple batches and strict inspection of ingredients. Traditional detection of traditional Chinese medicine preparation is mostly offline detection, which is time-consuming and labor-intensive, and it is difficult to timely feedback the content of traditional Chinese medicine ingredients in the production process to control abnormal situations in the production process. In the detection process of alcohol precipitation liquor of traditional Chinese medicine preparation, the liquor sampling process is extremely important, which directly affects whether the sample can fully reflect the composition of the liquor in the alcohol precipitation tank 3, and also affects the real-time and stability of the alcohol precipitation liquor sampling. The liquid level will change during the alcohol precipitation process, and there is interference of the stirring paddle. The current alcohol precipitation liquor sampling scheme is to extract liquor at a fixed position, but the fixed alcohol precipitation sampling position cannot fully reflect the overall quality of the alcohol precipitation liquor. These interference factors affect the quality of the extracted alcohol precipitation liquor, and further affect the detection feasibility. The current online detection method of traditional Chinese medicine alcohol precipitation process has the problem of poor real-time data analysis. The traditional alcohol precipitation NIR (Near Infrared System, i.e. near infrared system) only simply collects and records spectral information, and cannot perform real-time analysis on the collected spectral data. When the batch of raw materials changes, it will interfere with the original detection model, resulting in a decrease in detection accuracy. Therefore, it is also extremely important to solve the problem of model disturbance when the batch changes. Moreover, in the actual application scene, there are flammable substances such as ethanol, so the production equipment system needs to meet the explosion-proof requirement.

[0027] In view of the above problems, the present implementation proposes a traditional Chinese medicine alcohol precipitation control system, which designs a specific sampling float mechanism to extract alcohol precipitation liquor, performs real-time online detection on the extracted liquor, obtains the content of key quality markers in the production process, and provides timely quality feedback of the alcohol precipitation liquor through analysis and display of spectral data. According to the key quality markers, the end point of alcohol precipitation is judged and the process parameters of alcohol precipitation are controlled, and an intelligent control system with real-time closed loop of alcohol precipitation process (i.e. the traditional Chinese medicine alcohol precipitation control system of the present application) is constructed. The control system of the present application considers the safety of the alcohol precipitation process environment, and performs overall explosion-proof design on the detection equipment, which can work normally in a specific environment, improves the quality of liquor products in the alcohol precipitation process of traditional Chinese medicine preparation, improves the production efficiency of liquor alcohol precipitation, and promotes the high-quality development of traditional Chinese medicine industry.

[0028] More specifically, as Figure 1As shown, the alcohol precipitation tank 3 contains traditional Chinese medicine concentrate, and ethanol is added at room temperature or low temperature to dissolve the active ingredients in the concentrate and precipitate the impurities in the extract concentrate. To ensure that the concentrate and ethanol are fully contacted and improve the consistency and reliability of the liquid, the stirring motor 1 is usually operated at an appropriate stirring speed during the alcohol precipitation process, and the alcohol precipitation time is usually more than 24 hours. The extracted liquid passes through the detection liquid input pipeline 8 into the detection circulation pipeline work cabinet 4. In the detection circulation pipeline work cabinet 4, the peristaltic pump 16 and the electromagnetic valve (including the first electromagnetic valve 15 and the second electromagnetic valve 21) are controlled to realize the control of the traditional Chinese medicine liquid update and offline sampling process in the detection circulation pipeline (i.e. the pipeline between the detection liquid input pipeline 8 and the detection liquid return pipeline 7 and the components on the pipeline). The spectral data is collected by the optical cable 5 into the detection explosion-proof cabinet 6, and the online detection function is realized by the spectrometer 19 and the processor. The detected liquid returns to the alcohol precipitation tank 3 through the detection liquid return pipeline 7. The alcohol precipitation tank 3 is arranged on the platform 2, and the staff performs related operations through the platform 2.

[0029] As shown in Figure 2 , the system first updates the detection liquid in the detection circulation pipeline. The peristaltic pump 16 is opened, the first electromagnetic valve 15 is powered on and opened, and the second electromagnetic valve 21 is powered on and opened. The sampling float 9 and the sampling suction tube 11 suck the alcohol precipitation liquid into the detection circulation pipeline. The liquid in the detection circulation pipeline is fully updated, so that the detection liquid fully covers the flow cell 18. The peristaltic pump 16 is closed, the first electromagnetic valve 15 and the second electromagnetic valve 21 are de-energized and closed, and the liquid in the detection circulation pipeline is static. The spectral data is collected by the spectrometer 19 and the optical fiber probe in the optical fiber 20, and the spectral data is transmitted to the processor for analysis and storage. At the same time, the first electromagnetic valve 15 is powered on and opened, the second electromagnetic valve 21 is de-energized and closed, the manual sampling valve 17 is opened, the detection liquid is sampled offline, and the detection liquid is stored in the test tube corresponding to the label. The first electromagnetic valve 15 and the second electromagnetic valve 21 are powered on and opened, the detection liquid returns to the alcohol precipitation tank 3 through the detection liquid return pipeline 7, and the above process is repeated at specified time intervals to obtain a large amount of data. The offline sampling samples are analyzed by high performance liquid chromatography, a spectral analysis model is established according to the results of high performance liquid chromatography analysis, the analysis model is used for online detection of drug content, and the processor is configured to dynamically adjust process parameters and determine the end point. The other end of the detection liquid input pipeline 8 is connected with the first electromagnetic valve 15 through a pipeline, and a membrane filter 14 is connected on the pipeline.

[0030] As shown in Figure 3 , Figure 4 , and Figure 5As shown, the float sampling mechanism includes a sampling suction tube 11, positioning steel wires 10 (three or more can also be used), a sampling float 9, a hose 23, and a clamp 22. The float sampling mechanism extracts the liquid medicine to be detected to the detection circulation pipeline, and completes sampling. In the pharmaceutical equipment, the surface treatment of each component of the float sampling mechanism meets the GMP standard, and the material also meets the GMP standard. Preferably, the material of the sampling suction tube 11, the positioning steel wires 10, and the sampling float 9 in the float sampling mechanism is stainless steel, which meets the pharmaceutical standard. The surface of the part is polished to a degree of 0.4, ensuring that each part meets the GMP standard.

[0031] More specifically, the specific working process of the float sampling mechanism includes the following steps. Two positioning steel wires 10 are welded on the inner wall of the alcohol precipitation tank 3 to fix the sampling float 9, reduce the influence of the stirring vortex, and prevent the stirring paddle from interfering with the sampling float 9. The sampling suction tube 11 has a sampling hole (upper sampling hole 24, middle sampling hole 25, and lower sampling hole 26) at its upper end, middle, and lower end. The hole diameter increases from top to bottom, reducing the influence of pressure and achieving multi-hole sampling of the liquid medicine. The comprehensive multi-position sample liquid can better reflect the quality attributes of the internal liquid in the alcohol precipitation tank 3. The sampling float 9 can move up and down on the positioning steel wires 10 during the alcohol precipitation process, solving the problem of fixed-point sampling and reflecting the quality of the alcohol precipitation liquid. The sampling float 9 has a positioning steel wire hole 27 for the positioning steel wires 10 to pass through. The sampling float 9 is connected to the sampling suction tube 11 by screwing. The top of the sampling float 9 is connected to the hose 23 by the clamp 22. The entire float sampling mechanism has good airtightness, solving the problems of internal stirring paddle (including the first stirring paddle 12 and the second stirring paddle 13) disturbance and single-point sampling.

[0032] In this implementation, the detection explosion-proof cabinet 6 connected to the detection circulation pipeline work cabinet 4 by the optical cable 5 and the electric cable is the core of data analysis and control of the entire online detection, which includes a spectrometer 19, a processor, and an input and output device. The detection explosion-proof cabinet 6 is a core production device to ensure the safety of personnel and equipment production. The processor is the control core of the entire online detection, which communicates with the peristaltic pump 16, the stirring motor 1, the electromagnetic valve (i.e., the first electromagnetic valve 15 and the second electromagnetic valve 21), and the metering pump (i.e., the metering pump on the pipeline for adding ethanol) through the I / O interface. At the same time, the processor sends instructions to the spectrometer 19, collects the spectral data of the alcohol precipitation liquid through the optical fiber 20, and transmits the spectral data back to the processor in real time.

[0033] When detecting the drug liquid to be detected in the detection circulation pipeline, the manual sampling valve 17 is opened, the drug liquid to be detected is collected into the test tube, and the time label corresponding to the spectral data collected by the spectrometer 19 is pasted, and the rotation speed of the stirring motor 1, the flow of the metering pump and the temperature of the drug liquid are recorded. After updating the drug liquid in the detection circulation pipeline, the above steps are repeated. The drug liquid in the test tube is analyzed by high performance liquid chromatography (HPLC), the content of the substance is accurately determined, a variety of spectral data preprocessing algorithms (such as Savitzky-Golay filtering, standard normal variable (SNV) or multivariate scatter correction (MSC)) are fused to smooth noise, eliminate light path and particle size interference and correct sample non-uniformity, and reliable spectral data is provided for establishing an analysis model.

[0034] In the present embodiment, high-quality data provided by HPLC (High Performance Liquid Chromatography) is combined, a CNN (Convolutional Neural Network) and a batch transfer learning algorithm are used for model establishment, the CNN has better detection performance than traditional methods, can accurately detect a variety of quality markers, ethanol concentration, precipitate particle size and water concentration, and the transfer learning algorithm can quickly adapt the model when the batch of raw materials changes in the traditional Chinese medicine alcohol precipitation process, and reduce the influence of the batch change of raw materials on the quality of the drug liquid. The spectral data of the alcohol precipitation drug liquid of the new batch of traditional Chinese medicine preparation is input into the shared convolutional layer of the CNN model, then high-order features output by the shared convolutional layer are extracted, the maximum mean difference between the feature vector of the new batch of drug liquid and the feature vector of the historical batch is calculated, a maximum mean difference threshold is set, when the new batch of drug liquid exceeds the threshold, the shared convolutional layer in the CNN model is fixed, the model full connection layer is fine-tuned according to the label data collected in the new batch, and the mapping relationship from the feature space of the new batch of drug liquid to the quality index is quickly learned, so that the model transfer learning is achieved.

[0035] According to the constructed online detection model, when the drug liquid circulates in the detection circulation pipeline, the spectrometer 19 collects the spectral information of the drug liquid in the flow cell 18, the model detects the quality markers, ethanol concentration, precipitate particle size and water concentration of the drug liquid, and the processor controls the alcohol metering pump flow, the stirring motor 1 rotation speed and the cooling water flow according to the model output, adjusts and controls the content of the drug liquid in the alcohol precipitation process, and maintains the consistency and reliability of the drug liquid.

[0036] In the static precipitation stage, the processor controls the cooling water flow according to the temperature of the drug liquid detected by the temperature sensor, the model detects the content of the effective substance and the content of the impurity to judge the end point, the alcohol precipitation process is ended, the loss of the effective component caused by excessive alcohol precipitation is avoided, and then the processor saves the change information of the quality marker content in the alcohol precipitation process to the memory for subsequent data analysis.

[0037] More specifically, the traditional Chinese medicine alcohol precipitation regulation system provided by the present embodiment includes a sampling layer, a detection layer, a control layer, and an output layer.

[0038] The sampling layer specifically includes: Floating sampling, sampling operation is performed through floating; liquid pipeline circulation, liquid circulates in the pipeline; spectral information of liquid is collected by the spectrometer 19, and spectral information of liquid is collected by the spectrometer 19; Off-line sampling is performed to construct a model, and the process is described in detail in the right dashed box: sampling, sampling is performed manually; recording process parameters, recording related process parameters; spectral data labeling, labeling the spectral data; spectral data preprocessing, preprocessing the spectral data; HPLC analysis, performing high-performance liquid chromatography analysis; using neural network and transfer learning algorithm to construct a detection model; The detection layer specifically includes: spectral data preprocessing, preprocessing the collected spectral data; detection model, using the detection model to detect the following parameters: quality marker X1,···, quality marker Xn, ethanol concentration, water concentration, and precipitate particles; The control layer specifically includes: model decoupling parameters to processor, transmitting the decoupled parameters of the detection model to the processor; processor controls process parameters, the processor controls the following process parameters: alcohol metering pump flow, stirring motor 1 speed, cooling water flow; static precipitation stage, entering the static precipitation stage, performing temperature control, then performing end point judgment, and finally alcohol precipitation is completed; The output layer specifically displays historical data, content change trend, process parameters and spectral information: outputs and displays historical data, content change trend, process parameters and spectral information.

[0039] The entire process described above realizes accurate control of liquid quality and process parameters through the three main layers of the sampling layer, the detection layer and the control layer, combined with off-line modeling and online detection, and outputs related data and information.

[0040] In the present embodiment, after the alcohol precipitation liquid is pretreated, ethanol is added to the alcohol precipitation tank 3 according to a preset ethanol gradient concentration. The model established above is used to process the spectral information of the liquid in the flow cell 18 to detect the deposition of the effective component caused by denaturation. The processor adjusts the ethanol metering pump to slow down the ethanol addition speed when the content of impurities such as protein is high. In order to improve the contact area between the liquid and ethanol during the ethanol addition process according to the traditional Chinese medicine preparation process, the stirring motor 1 is started at the same time. When the ethanol concentration is too high, the effective component is wrapped and precipitated by the precipitate. The processor controls the motor to increase the stirring speed of the stirring paddle. When the size of the precipitate particles is too small and the energy consumption of the stirring motor 1 is too large, the stirring speed of the stirring paddle is reduced. When the heat generated during the ethanol addition and stirring process causes the temperature of the alcohol precipitation liquid to rise, the processor uses the MPC (Model Predictive Control, Model Predictive Control) prediction algorithm to increase the flow of cooling water in the cooling pipeline 28 in advance. By controlling the flow, the alcohol precipitation liquid is maintained at a certain temperature. Enter the static precipitation stage. The cooling water flow is controlled according to the liquid temperature to cool the liquid.

[0041] More specifically, during the pretreatment of the liquid, ethanol is added according to the preset ethanol gradient, and the stirring motor 1 is started. Then the model is used to process the spectral information of the liquid in the flow cell 18. If the effective component is denatured and precipitated, the processor controls the metering pump to slow down the ethanol addition speed. Then it is judged whether the energy consumption and vibration of the stirring motor 1 exceed the threshold value. If they exceed, the speed of the motor is reduced. If they do not exceed, the current speed of the motor is maintained, and the speed of the stirring motor 1 is increased. If not, it is judged whether the content of impurities such as protein is high. If it is high, the processor controls the metering pump to speed up the ethanol addition speed. Then it is judged whether the energy consumption and vibration of the stirring motor 1 exceed the threshold value. Depending on the result, the speed of the motor is reduced or maintained. If the content of impurities such as protein is not high, the current ethanol addition speed is maintained. Then it is judged whether the size of the precipitate particles is too small. If it is too small, the speed of the stirring motor 1 is reduced. If it is not too small, the current speed of the motor is maintained. At the same time, the temperature sensor detects the temperature of the liquid. If the temperature of the liquid rises, the cooling state is predicted by the MPC algorithm, and then the flow of cooling water is adjusted (may be increased). If the temperature of the liquid does not rise, the current cooling state is maintained. The whole process is repeated at specified times.

[0042] In the static precipitation stage, the content of the effective component is detected online. The concentration of each quality marker in three consecutive sampling periods (or more sampling periods) meets the preset value, and the content of impurities is less than the preset value. If the two content test conditions are met, the system determines whether the alcohol precipitation is complete, and generates a process parameter record to facilitate subsequent optimization.

[0043] More specifically, as Figure 6As shown, in the static alcohol precipitation stage, the system carries out real-time sampling analysis, and then judges whether the effective component content reaches the preset value for three consecutive periods. If not, the system continues to carry out real-time sampling analysis; if yes, it further judges whether the impurity content is lower than the preset value; if not, it returns to continue to carry out real-time sampling analysis; if yes, it judges that the alcohol precipitation ends, and then generates a process parameter record, and the whole process is cyclically carried out at a specified time.

[0044] In the present implementation, the component content of the drug solution is detected in real time through online detection, and the alcohol precipitation process parameters are dynamically controlled to optimize the production process, realizing the whole process control from data acquisition to model analysis to process control. Compared with the traditional offline detection mode, high-quality and high-efficiency production can be realized, and dynamic regulation and control can reduce energy consumption and improve the consistency and reliability of multiple batches of drug solutions.

[0045] The above is only a preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A Chinese medicine alcohol precipitation regulating system, characterized in that, it comprises: an alcohol precipitation tank, a stirring motor, a detection liquid input pipeline, a detection liquid return pipeline, and a floating sampling mechanism, the output shaft of the stirring motor is connected with a stirring paddle in the alcohol precipitation tank, and a cooling pipeline is arranged at the bottom of the alcohol precipitation tank; the floating sampling mechanism comprises a sampling float, a sampling suction tube, and a positioning element, the positioning element is fixedly connected with the inner wall of the alcohol precipitation tank, the sampling float is sleeved on the positioning element and can move up and down along the positioning element with the change of the liquid level of the liquid, the bottom of the internal passage of the sampling float is connected with the sampling suction tube, a plurality of sampling holes are formed in the sampling suction tube, the top of the internal passage of the sampling float is connected with a sampling hose, and one end of the sampling hose is connected with the detection liquid input pipeline; the other end of the detection liquid input pipeline is connected with a first electromagnetic valve, a peristaltic pump, a manual sampling valve, a flow cell, and a second electromagnetic valve in sequence through pipelines, one end of the detection liquid return pipeline is connected with the second electromagnetic valve through a pipeline, the other end of the detection liquid return pipeline is connected with the alcohol precipitation tank, and the flow cell is arranged with an optical fiber connected with a spectrometer for detection. 2.The Chinese medicine alcohol precipitation regulating system according to claim 1, characterized in that, the positioning element is a positioning wire, the positioning wire is welded and fixed on the inner wall of the alcohol precipitation tank, the sampling float is sleeved on the positioning wire, the top of the internal passage of the sampling float is connected with the sampling hose through a clamp, and the bottom of the internal passage of the sampling float is threadedly connected with the top of the sampling suction tube. 3.The Chinese medicine alcohol precipitation regulating system according to claim 1, characterized in that, a plurality of sampling holes are arranged in the upper part, the middle part, and the lower part of the sampling suction tube in the circumferential direction, and the inner diameters of the sampling holes increase from top to bottom. 4.The Chinese medicine alcohol precipitation regulating system according to any one of claims 1-3, characterized in that, the spectrometer is arranged in a detection explosion-proof cabinet, a processor connected with the spectrometer is further arranged in the detection explosion-proof cabinet, and the processor is in communication connection with the peristaltic pump, the stirring motor, the first electromagnetic valve, and the second electromagnetic valve; the first electromagnetic valve, the peristaltic pump, the manual sampling valve, the flow cell, and the second electromagnetic valve are arranged in a detection circulation pipeline work cabinet, and the spectrometer is inserted into the flow cell through the optical fiber to realize liquid detection. 5.The Chinese medicine alcohol precipitation regulating system according to any one of claims 1-3, characterized in that, a membrane filter is connected on the pipeline between the other end of the detection liquid input pipeline and the first electromagnetic valve.

6. A dynamic regulation method based on real-time detection of ingredients, characterized in that, The Chinese medicine alcohol precipitation regulating system according to any one of claims 1-5 comprises the following processes: the peristaltic pump is opened, the first electromagnetic valve and the second electromagnetic valve are powered on and opened, the sampling float and the sampling suction tube suck the liquid in the alcohol precipitation tank into the detection liquid input pipeline, and the liquid covers the flow cell; the peristaltic pump is closed, the first electromagnetic valve and the second electromagnetic valve are powered off and closed, the liquid in the detection liquid input pipeline and the detection liquid return pipeline is static, and the spectrometer collects the spectral data of the liquid in the flow cell through the optical fiber; the first electromagnetic valve is powered on and opened, the second electromagnetic valve is powered off and closed, the manual sampling valve is opened for offline sampling, the liquid is stored in a test tube with a corresponding label, and the manual sampling valve is closed after the offline sampling is completed. The first electromagnetic valve and the second electromagnetic valve are powered on to open, and the liquid medicine returns to the alcohol precipitation tank through the detection liquid return pipeline; Cycle sampling at a specified time interval, perform liquid chromatography analysis on the offline sampled samples, establish a spectral analysis model based on the results of the liquid chromatography analysis for online detection of drug content, dynamically control process parameters based on the results of the online detection of drug content, and perform end point judgment.

7. The dynamic control method based on real-time detection of ingredients according to claim 6, wherein, performing liquid chromatography analysis on the offline sampled samples, and establishing a spectral analysis model based on the results of the liquid chromatography analysis, comprises: performing liquid chromatography analysis on the liquid medicine in the test tube to obtain spectral data, and performing preprocessing on the spectral data using a spectral data preprocessing algorithm; combined with the preprocessed spectral data, training of the convolutional neural network is performed to enable the convolutional neural network to output a detection quality marker, ethanol concentration, precipitate particle size, and water concentration based on the spectral data, and the trained convolutional neural network is used as the spectral analysis model.

8. The dynamic control method based on real-time detection of ingredients according to claim 7, wherein, when the batch of raw materials is changed during the traditional Chinese medicine alcohol precipitation process, the spectral data of the new batch of traditional Chinese medicine preparation alcohol precipitation liquid is input into the shared convolutional layer of the convolutional neural network, the feature vectors output by the shared convolutional layer are extracted, and the maximum mean difference between the feature vectors of the new batch of liquid and the feature vectors of the historical batch of liquid is calculated; when the maximum mean difference exceeds a set threshold, the shared convolutional layer of the convolutional neural network is fixed, the fully connected layer of the convolutional neural network is fine-tuned based on the label data of the new batch of raw materials to learn the mapping relationship from the feature space of the new batch of liquid to the quality indicators, and a spectral analysis model after transfer learning is obtained.

9. The dynamic control method based on real-time detection of ingredients according to claim 6, wherein, when the liquid medicine circulates in the detection liquid input pipeline and the detection liquid return pipeline, the spectral information of the liquid medicine in the flow cell is collected, the spectral analysis model is used to detect the quality marker, ethanol concentration, precipitate particle size, and water concentration of the liquid medicine, and the flow of the alcohol metering pump, the rotation speed of the stirring motor, and the flow of the cooling water are controlled based on the output of the spectral analysis model.

10. The dynamic control method based on real-time detection of ingredients according to claim 9, wherein, when the liquid medicine enters the static precipitation stage, the temperature of the detection liquid is detected, the flow of the cooling water in the cooling pipeline is controlled to control the temperature of the liquid medicine, and the end point is judged based on the detected content of the active substance and the content of impurities, and the alcohol precipitation process is ended when the end point is reached; the end point judgment comprises: the concentration of each quality marker reaches a preset value in consecutive N sampling periods, and the content of impurities is less than a preset value, then the end point is reached, and the alcohol precipitation is ended, wherein N is a positive integer greater than 2.

Images