Drug solution pretreatment and sampling equipment for traditional Chinese medicine detection, traditional Chinese medicine detection system and method

Abstract

The invention belongs to the technical field of quality control and detection in the process of drug production, and proposes a drug liquid pretreatment and sampling equipment for traditional Chinese medicine detection, a traditional Chinese medicine detection system and a method. Through the specific structure of a second filter and an oil bath temperature control mechanism and the arrangement of a reflux pipe and a drug liquid input pipe, etc., the interference of temperature, impurities and bubbles on the quality of subsequent spectral detection can be eliminated; by opening a metering pump, a second solenoid valve and a third solenoid valve arranged on a pipeline, continuous circulation of the drug liquid between an extraction tank and the pipeline can be achieved; after the metering pump, the second solenoid valve and the third solenoid valve are closed, the drug liquid in the pipeline can be heated and temperature-controlled by the oil bath temperature control mechanism, and after controlling the second solenoid valve and the third solenoid valve to be closed and the first solenoid valve to be opened, offline sampling can be achieved at the sampling valve with the cooperation of a breathing valve, so as to provide reliable data for subsequent detection.

Description

Technical Field

[0001] The invention belongs to the technical field of quality control and detection in a drug production process, and in particular relates to a drug solution pretreatment and sampling device for traditional Chinese medicine detection, a traditional Chinese medicine detection system and a method. Background Art

[0002] The production process of traditional Chinese medicine preparations is complex, involving multiple links such as extraction and concentration. The composition and quality of different batches of medicinal materials vary, which poses a great challenge to the consistency control of the critical quality attributes (CQA) of drugs during the production process. The online detection and analysis system of the critical quality attributes of traditional Chinese medicine based on near infrared spectroscopy (NIR) can provide data basis for the consistency control of the critical quality attributes of drugs during the production process.

[0003] Pretreatment before drug liquid detection is particularly important. Factors such as temperature, impurities and liquid flow will seriously affect the quality of the spectrum, and then affect the accuracy of the final online detection. However, the pretreatment device before drug liquid detection in the current traditional Chinese medicine detection system cannot perform sufficient pretreatment, especially the insufficient renewal of drug liquid, the poor heating effect of the temperature control mechanism and the inability to perform constant temperature control, which cannot fully guarantee the accuracy of the detection and affect the reliability of the key quality attributes of drugs in the production process. At the same time, flammable gases such as ethanol are often present in the production environment of traditional Chinese medicine. The current traditional Chinese medicine detection system does not consider the explosion problem after flammable gases such as ethanol come into contact with the spectrum acquisition unit, data analysis unit and control unit, and has poor safety; and the current traditional Chinese medicine online detection system has problems such as low integration and poor real-time performance in data analysis and display. Summary of the invention

[0004] In order to solve the above problems, the present invention proposes a liquid medicine pretreatment and sampling device for traditional Chinese medicine detection, a traditional Chinese medicine detection system and method. Through the specific structure of the second filter, the oil bath temperature control mechanism, and the arrangement of the reflux pipe and the liquid medicine input pipe, the interference of temperature, impurities and bubbles on the quality of subsequent spectral detection can be eliminated; by opening the metering pump, the second solenoid valve and the third solenoid valve arranged on the pipeline, the continuous circulation of the liquid medicine between the extraction tank and the pipeline can be achieved, and the liquid medicine in the pipeline can be fully updated; the liquid medicine in the pipeline can be heated and temperature-controlled by the oil bath temperature control mechanism, and with the cooperation of the breathing valve, offline sampling can be achieved at the sampling valve, providing reliable data for subsequent detection.

[0005] In order to achieve the above-mentioned purpose, in the first aspect, the present invention provides a liquid pretreatment and sampling device for traditional Chinese medicine detection, which adopts the following technical scheme:

[0006] A liquid pretreatment and sampling device for Chinese medicine detection, comprising a second filter, a second solenoid valve, a metering pump, an oil bath temperature control mechanism, a circulation pool and a third solenoid valve connected in sequence through a pipeline; a breathing valve is arranged between the second solenoid valve and the metering pump through the first solenoid valve; a sampling valve is arranged between the oil bath temperature control mechanism and the circulation pool;

[0007] The two ends of the pipeline are connected to the extraction tank through a reflux pipe and a liquid medicine input pipe respectively; the oil bath temperature control mechanism includes a box body, a special-shaped conduit arranged in the box body, and a heating rod arranged on the box body; the box body is filled with silicone oil, and a temperature sensor is arranged on the box body.

[0008] Furthermore, a window is provided on the box.

[0009] In order to achieve the above-mentioned purpose, in a second aspect, the present invention also provides a Chinese medicine detection system, which adopts the following technical scheme:

[0010] A method for pretreatment and sampling of a liquid medicine for Chinese medicine detection, using the liquid medicine pretreatment and sampling device for Chinese medicine detection as described in the first aspect, comprising:

[0011] Start the metering pump, the second solenoid valve and the third solenoid valve, and the liquid medicine enters the pretreatment pipeline from the pipeline of the original Chinese medicine production equipment through the liquid medicine input pipe, and circulates continuously to update the liquid medicine in the pipeline; close the metering pump, the second solenoid valve and the third solenoid valve, and the liquid medicine stops flowing; start the oil bath temperature control mechanism to perform constant temperature control on the liquid medicine until it reaches a preset specified temperature; start the metering pump, the second solenoid valve and the third solenoid valve again, and circulate within a preset time so that the constant temperature liquid fully covers the periphery of the circulation pool, and then close the metering pump, the second solenoid valve and the third solenoid valve, and the liquid medicine stops to eliminate bubbles and flow;

[0012] When the second solenoid valve and the third solenoid valve are closed, the first solenoid valve is opened to connect the pipeline to the atmosphere; at this time, the sampling valve is opened to perform offline sampling.

[0013] In order to achieve the above-mentioned purpose, in a third aspect, the present invention also provides a method for pretreatment and sampling of liquid medicine for Chinese medicine detection, which adopts the following technical scheme:

[0014] A Chinese medicine detection system comprises an extraction tank, a liquid medicine pretreatment and sampling device connected to the extraction tank, and an explosion-proof positive pressure cabinet;

[0015] The liquid medicine pretreatment and sampling equipment comprises a second filter, a second solenoid valve, a metering pump, an oil bath temperature control mechanism, a circulation pool and a third solenoid valve which are sequentially connected through pipelines; a breathing valve is arranged between the second solenoid valve and the metering pump through the first solenoid valve; a sampling valve is arranged between the oil bath temperature control mechanism and the circulation pool;

[0016] The two ends of the pipeline are connected to the extraction tank through a reflux pipe and a liquid medicine input pipe respectively; the oil bath temperature control mechanism includes a box body, a special-shaped conduit arranged in the box body, and a heating rod arranged on the box body; the box body is filled with silicone oil, and a temperature sensor is arranged on the box body; the sampling valve is connected to a spectrometer, and the spectrometer is arranged in the explosion-proof positive pressure cabinet.

[0017] Furthermore, a window is provided on the box.

[0018] Furthermore, optical fiber probes are provided at both ends of the circulation pool, and a sealing ring is provided between the optical fiber probe and the sampling valve.

[0019] Furthermore, the optical fiber probe is connected to an interface for connecting to the spectrometer via an optical fiber; a sealing bushing is provided between the optical fiber and the explosion-proof positive pressure cabinet, and a sealant is provided between the sealing bushing and the rubber sleeve of the optical fiber.

[0020] Furthermore, the spectrometer is connected to a processor, and the processor is connected to a display screen; the processor is arranged in the explosion-proof positive pressure cabinet.

[0021] In order to achieve the above-mentioned purpose, in a fourth aspect, the present invention also provides a method for detecting traditional Chinese medicine, which adopts the following technical scheme:

[0022] A method for detecting traditional Chinese medicine, using the traditional Chinese medicine detection system as described in the third aspect, comprising:

[0023] Start the metering pump, the second solenoid valve and the third solenoid valve, and the liquid medicine enters the pretreatment pipeline from the pipeline of the original Chinese medicine production equipment through the liquid medicine input pipe, and circulates continuously to update the liquid medicine in the pipeline; close the metering pump, the second solenoid valve and the third solenoid valve, and the liquid medicine stops flowing; start the oil bath temperature control mechanism to perform constant temperature control on the liquid medicine until it reaches a preset specified temperature; start the metering pump, the second solenoid valve and the third solenoid valve again, and circulate within a preset time so that the constant temperature liquid fully covers the periphery of the circulation pool, and then close the metering pump, the second solenoid valve and the third solenoid valve, and the liquid medicine stops to eliminate bubbles and flow;

[0024] The spectrometer collects spectral features and stores the spectral features; when the second solenoid valve and the third solenoid valve are closed, the first solenoid valve is opened to connect the pipeline to the atmosphere; at this time, the sampling valve is opened to take samples offline and store them in a test tube with a label; the liquid medicine samples in the test tubes are analyzed by high performance liquid chromatography one by one to calibrate the substance content, and the results are used as substance content labels;

[0025] The spectral features and substance content labels are screened for outliers, and then the spectral data is subjected to different preprocessing steps to optimize the data quality; the quantitative analysis model is modeled using partial least squares, decision tree regression algorithm, and gradient boosting decision tree algorithm, and the quantitative analysis model is selected based on its performance; the determined quantitative analysis model is saved in the processor to realize the prediction of drug content in online mode.

[0026] Furthermore, the quantitative analysis model was a gradient boosting decision tree algorithm model, the pre-prediction data was analyzed for outlier detection using principal component analysis, the Mahalanobis distance was calculated, points beyond the confidence interval were deleted as outliers, and the spectral features were processed using standard normal variable transformation.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] 1. The liquid medicine pretreatment and sampling equipment of the present invention can eliminate the interference of temperature, impurities and bubbles on the quality of subsequent spectral detection through the specific structure of the second filter and the oil bath temperature control mechanism and the arrangement of the reflux pipe and the liquid medicine input pipe. By opening the metering pump, the second solenoid valve and the third solenoid valve arranged on the pipeline, the continuous circulation of the liquid medicine between the extraction tank and the pipeline can be realized, and the liquid medicine in the pipeline can be fully updated. After the metering pump, the second solenoid valve and the third solenoid valve are closed, the liquid medicine in the pipeline can be heated and temperature-controlled by the oil bath temperature control mechanism, and after controlling the second solenoid valve and the third solenoid valve to be closed and the first solenoid valve to be opened, with the cooperation of the breathing valve, offline sampling can be realized at the sampling valve to provide reliable data for subsequent detection.

[0029] 2. In the present invention, electrical components such as the spectrometer and the processor are placed in an explosion-proof positive pressure cabinet. At the same time, a sealing bushing is provided between the optical fiber and the explosion-proof positive pressure cabinet, and a sealant is provided between the sealing bushing and the rubber sleeve of the optical fiber. Explosion-proof sealing treatment is performed to avoid explosion problems caused by the contact of flammable gases such as ethanol with the spectrometer and the processor.

[0030] 3. The present invention processes spectral data in real time through a determined quantitative analysis model, predicts the content of specified quality markers, and displays them in real time through a display screen, which not only improves the integration of data processing and display, but also enhances the real-time monitoring and optimization capabilities of the production process, ensuring that key quality attributes in the production process can be monitored and adjusted in real time, thereby improving the quality consistency and stability of oral Chinese medicine preparations.

[0031] 4. The present invention determines a feature processing method and a prediction model algorithm with higher prediction accuracy without affecting the data processing speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The drawings in the specification that constitute a part of this embodiment are used to provide a further understanding of this embodiment. The schematic embodiments of this embodiment and their descriptions are used to explain this embodiment and do not constitute improper limitations on this embodiment.

[0033] Figure 1 It is a schematic diagram of the overall structure of the traditional Chinese medicine detection system according to an embodiment of the present invention;

[0034] Figure 2 This is a schematic diagram of the structure of a liquid medicine pretreatment and sampling device according to an embodiment of the present invention;

[0035] Figure 3 This is a working principle diagram of the liquid medicine pretreatment and sampling equipment according to an embodiment of the present invention;

[0036] Figure 4 This is a schematic diagram of an oil bath temperature control mechanism according to an embodiment of the present invention;

[0037] Figure 5 A schematic diagram of a flow cell according to an embodiment of the present invention;

[0038] Figure 6 A schematic diagram of an optical fiber according to an embodiment of the present invention;

[0039] Figure 7 A schematic diagram of a sealing bushing according to an embodiment of the present invention;

[0040] Figure 8 The control architecture and data flow of an embodiment of the present invention;

[0041] Fig. 9 The modeling mode workflow of the embodiment of the present invention;

[0042] Fig.10 It is the principal component analysis outlier detection of the embodiment of the present invention;

[0043] Fig.11 It is a spectral image processed by the standard normal variable transformation method according to an embodiment of the present invention;

[0044] Fig.12 The spectral image processed by the standardization method according to the embodiment of the present invention;

[0045] Fig.13 It is a spectral image processed by the mean centering method adopted in an embodiment of the present invention;

[0046] Fig.14 This is a spectral image processed by the Savitzky-Golay smoothing method according to an embodiment of the present invention;

[0047] Fig.15 A spectral image using the original spectrum according to an embodiment of the present invention;

[0048] Fig.16 The spectral image processed by the first-order derivative method according to the embodiment of the present invention;

[0049] Fig.17 This is the fitting effect after being processed by the gradient boosting decision tree method in the embodiment of the present invention;

[0050] Fig.18 This is the fitting effect after processing using the decision tree regression method in the embodiment of the present invention;

[0051] Fig.19 This is the fitting effect after the partial least squares method is used in the embodiment of the present invention;

[0052] Fig. 20 A scatter plot showing the actual value and predicted value of the content of the present invention changing over time;

[0053] Among them, 1. extraction tank; 2. first filter; 3. circulation pump; 4. drug liquid input pipe; 5. drug liquid pretreatment and sampling equipment; 501. temperature sensor; 502. first solenoid valve; 503. circulation pool; 5031. sealing ring; 5032. optical fiber probe; 5033. sealing bushing; 50331. sealant; 50332. rubber sleeve; 50333. optical fiber core; 5034. interface; 504. sampling valve; 505. second solenoid valve; 506. breathing valve; 507. third solenoid valve; 508. oil bath temperature control mechanism; 5081. box; 5082. window; 5083. special-shaped catheter; 5084. heating rod; 509. metering pump; 510. second filter; 6. optical fiber; 7. explosion-proof positive pressure cabinet; 8. reflux pipe; 9. spectrometer. DETAILED DESCRIPTION

[0054] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0055] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0056] Near-infrared spectroscopy is based on molecular vibration spectrum. It has been rapidly developed and has become an extremely important analytical tool in modern analytical chemistry by analyzing the interaction between light waves and substances in the near-infrared region. Its advantages such as non-destructiveness, no need for complex sample pretreatment, rapidity, and the ability to perform online measurements have made it widely used in many fields, especially in the pharmaceutical industry. In the production process of traditional Chinese medicine, the application of near-infrared spectroscopy technology helps to achieve real-time monitoring of key quality attributes of drugs and improve product consistency and reliability.

[0057] There are still many challenges and shortcomings in the current production process of traditional Chinese medicine. First, the production process of traditional Chinese medicine preparations is complex, involving multiple links such as extraction and concentration. There are differences in the composition and quality of different batches of medicinal materials, which poses a great challenge to the consistency control of key quality attributes of drugs during the production process. Traditional quality control methods are mostly offline detection, which is not only time-consuming and labor-intensive, but also difficult to promptly feedback abnormal situations in the production process, resulting in poor consistency of product quality batches.

[0058] As recorded in the background technology, first, in the application of near-infrared spectroscopy technology in the production process of traditional Chinese medicine, the pretreatment before the liquid medicine detection is particularly important; factors such as temperature, impurities and liquid flow will seriously affect the spectral quality, and then affect the accuracy of the final online detection; therefore, the liquid medicine needs to be fully pretreated before detection to eliminate these interfering factors; however, the existing pretreatment technology still has room for improvement in efficiency and effect, and the accuracy of the detection cannot be fully guaranteed. Secondly, the current online detection system has the problems of low integration and poor real-time performance in data analysis and display; traditional systems can often only realize simple spectral data acquisition and recording, and cannot analyze and display key quality indicators in real time, lack practical application value, and cannot be combined with production systems for intelligent control and process optimization. Finally, flammable gases such as ethanol are often present in the production environment of traditional Chinese medicine, which puts forward explosion-proof requirements for production equipment; the existing online detection system has deficiencies in explosion-proof design, and it is difficult to operate safely in an environment with explosion-proof requirements; in order to ensure production safety, explosion-proof design becomes an important part of the online detection system.

[0059] Based on the above problems, in order to solve at least one of them, in some embodiments, a traditional Chinese medicine detection system is provided, which can detect the key quality attributes of traditional Chinese medicine in the production process online with high precision, and provide producers with immediate and accurate quality feedback through real-time data analysis and display. At the same time, the system also fully considers safety factors, and builds an overall architecture based on explosion-proof safety requirements to ensure stable operation even in specific environments. It can provide strong assistance for improving the production efficiency and product quality of oral traditional Chinese medicine preparations, and inject new vitality and impetus into the high-quality development of the traditional Chinese medicine industry. Figure 1 As shown, the provided traditional Chinese medicine detection system includes an extraction tank 1, a first filter 2, a circulation pump 3, a liquid medicine input pipe 4, a liquid medicine pretreatment and sampling device 5, an optical fiber 6, an explosion-proof positive pressure cabinet 7 and a reflux pipe 8, etc.

[0060] In some embodiments, Figure 1 As shown, an extraction tank 1 is provided, wherein the extraction tank 1 contains a Chinese medicine production liquid, and the effective ingredients in the Chinese medicinal materials are dissolved by a solvent at high temperature. The extraction time can vary from tens of minutes to several hours. A large number of bubbles are usually generated during the heating process, which will cause uneven heating. Therefore, the circulation pump 3 is usually used to circulate periodically to ensure uniform internal heating. As the extraction proceeds, the solid content in the solution increases, which may cause the viscosity to gradually increase.

[0061] In some embodiments, Figure 2 As shown, a liquid medicine pretreatment and sampling device 5 is provided, and the liquid medicine pretreatment and sampling device 5 comprises a second filter 510, a second solenoid valve 505, a metering pump 509, an oil bath temperature control mechanism 508, a circulation pool 503 and a third solenoid valve 507 which are sequentially connected through pipelines; a breathing valve 506 is provided between the second solenoid valve 505 and the metering pump 509 through the first solenoid valve 502; a sampling valve 504 is provided between the oil bath temperature control mechanism 508 and the circulation pool 503;

[0062] The two ends of the pipeline are connected to the extraction tank 1 through a reflux pipe 8 and a liquid medicine input pipe 4 respectively; the oil bath temperature control mechanism 508 includes a box body 5081, a special-shaped catheter 5083 arranged in the box body 5081, and a heating rod 5084 arranged on the box body 5081; the box body 5081 is filled with silicone oil, and a temperature sensor 501 is arranged on the box body 5081; the sampling valve 504 is connected to a spectrometer 9, and the spectrometer 9 is arranged in the explosion-proof positive pressure cabinet 7.

[0063] The liquid pretreatment and sampling device 5 eliminates the interference of temperature, impurities, and bubbles on the spectral quality through various internal functional modules; through the control of the diaphragm pump and the electromagnetic valve, the control of multiple processes such as liquid renewal, liquid online detection, and liquid offline sampling is realized, thereby realizing the overall function of online detection. The Chinese medicine extraction liquid enters the liquid pretreatment and sampling device 5 through the liquid input pipe 4, and the liquid after treatment and detection flows back to the extraction tank 1 through the reflux pipe 8. The metering pump 509 provides power to pump the liquid to be tested into the liquid pretreatment pipeline to be tested; the impurities in the liquid to be tested are filtered through the first filter 2 to eliminate the interference of impurities; the oil bath temperature control mechanism 508 is used to ensure the constant temperature of the liquid to be tested and eliminate the interference of temperature changes; the designed circulation pool 503 is used to ensure the stability of spectral acquisition, and the spectrometer 9 obtains spectral data; the sampling valve 504, combined with the opening and closing of the electromagnetic valve and the atmospheric connection of the breathing valve 506, realizes offline sampling; the first filter 2 can be optionally a Y-type filter.

[0064] In some embodiments, based on the liquid pretreatment and sampling device 5, such as Figure 3 As shown, a method for pretreatment and sampling of liquid medicine for Chinese medicine detection is also provided, comprising:

[0065] Start the metering pump 509, the second solenoid valve 505 and the third solenoid valve 507, and the liquid medicine enters the pretreatment pipeline from the pipeline of the original Chinese medicine production equipment through the liquid medicine input pipe 4, and circulates continuously to update the liquid medicine in the pipeline; close the metering pump 509, the second solenoid valve 505 and the third solenoid valve 507, and the liquid medicine stops flowing; start the oil bath temperature control mechanism 508 to perform constant temperature control on the liquid medicine until it reaches a preset specified temperature; start the metering pump 509, the second solenoid valve 505 and the third solenoid valve 507 again, and circulate within a preset time so that the constant temperature liquid fully covers the circulation pool 503, and then close the metering pump 509, the second solenoid valve 505 and the third solenoid valve 507, and the liquid medicine stops to eliminate bubbles and flow;

[0066] When the second solenoid valve 505 and the third solenoid valve 507 are closed, the first solenoid valve 502 is opened to connect the pipeline with the atmosphere; at this time, the sampling valve 504 is opened to perform offline sampling.

[0067] In some embodiments, Figure 4As shown, an oil bath temperature control mechanism 508 is provided, the oil bath temperature control mechanism 508 comprises a box 5081, a special-shaped conduit 5083 arranged in the box 5081, and a heating rod 5084 arranged on the box 5081; the box 5081 is filled with silicone oil, the box 5081 is provided with a temperature sensor 501; the box 5081 is provided with a window 5082. The box 5081 is filled with silicone oil, which is heated by the heating rod 5084 and closed-loop controlled by the temperature sensor 501. The oil bath generates less vibration and can reach a higher temperature than the water bath heating process; the internal silicone oil amount can be observed through the window 5082 to ensure the internal oil level; the special-shaped conduit 5083 can increase the contact area between the liquid and the silicone oil.

[0068] In some embodiments, Figure 5 As shown, a flow cell 503 is provided, and optical fiber probes 5032 are arranged at both ends of the flow cell 503, and a sealing ring 5031 is arranged between the optical fiber probe 5032 and the sampling valve 504. The flow cell 503 is connected to the optical fiber probe 5032 through a thread, and the optical path of the two probe end faces is ensured to be 2mm through the end face; the sealing is ensured by the sealing ring 5031, and the sealing ring 5031 can be a rubber sealing ring.

[0069] In some embodiments, Figure 6 and Figure 7 As shown, an optical fiber 6 is provided, including a rubber sleeve 50332 and an optical fiber core 50333; specifically, the optical fiber probe 5032 is connected to an interface 5034 for connecting to the spectrometer 9 through an optical fiber; a sealing sleeve 5033 is provided between the optical fiber 6 and the explosion-proof positive pressure cabinet 7, and a sealant 50331 is provided between the sealing sleeve 5033 and the rubber sleeve 50332 of the optical fiber 6.

[0070] like Figure 6 and Figure 7 As shown, one side of the optical fiber 6 is an optical fiber probe 5032, which is connected to the circulation pool 503 through a thread; the other side of the optical fiber 6 is an interface 5034, which can be an SMA905 interface, and the interface 5034 is connected to the interface of the spectrometer 9 in the explosion-proof positive pressure cabinet 7 through a nut; the two wires on each side correspond to the output and input of the light source, respectively, and complete the collection of spectral data together with the spectrometer 9. The sealing bushing 5033 is designed in the middle of the optical fiber 6, and two optical fiber cores 50333 pass through the sealing bushing 5033 from the middle, and the sealing effect is achieved inside the sealing bushing 5033 by the sealant 50331, and the sealing bushing 5033 is tightly sealed with the explosion-proof gland on the explosion-proof positive pressure cabinet 7, and dangerous gases cannot enter the explosion-proof positive pressure cabinet 7.

[0071] Optionally, the explosion-proof positive pressure cabinet 7 may include a control system, a data analysis system, a spectrometer 9, etc. The explosion-proof positive pressure cabinet 7 is a special type of pressure vessel used to protect the safety of personnel and equipment. The explosion-proof positive pressure cabinet 7 is designed to maintain positive pressure inside the container. The optional gas source pressure is 0.2Mpa~0.8Mpa, and prevents explosive gas from entering the container. The optical fiber 6 is used as a channel connecting the liquid pretreatment and sampling equipment 5 and the explosion-proof positive pressure cabinet 7. It is not only responsible for the control and power distribution tasks of electrical components, but also serves as an efficient transmission medium for the spectrometer 9 to collect signals. Specifically, the cable is responsible for transmitting control signals and electricity to ensure that various functional modules in the work cabinet and electrical components in the explosion-proof positive pressure cabinet 7 can operate normally. The various functional modules in the work cabinet include filters, heating modules and metering pumps, etc.; while the optical fiber 6 is used for lossless transmission of spectral data to ensure that the spectrometer 9 can collect and analyze the spectral information of the processed liquid in the work cabinet in real time and accurately.

[0072] like Figure 8 As shown, the keyboard, button and mouse provide input instructions for the processor to specify the corresponding functional operations, such as starting the equipment, adjusting the equipment parameters, etc.; the processor communicates with the electrical components through the I / O interface, and realizes the precise control of the electrical components in combination with the preset control logic to ensure the normal operation of the external circulation liquid pretreatment and offline sampling system; the I / O interface is an input and output interface, and the electrical components include filters, heating modules and metering pumps. The processor sends a spectrum acquisition instruction to the spectrometer 9 and receives spectrum data; the processor sends a spectrum acquisition instruction to the spectrometer 9, and the spectrometer 9 starts to collect the spectrum data of the liquid after receiving the instruction, and transmits the collected spectrum data to the processor in real time; the processor has a built-in computing unit, which combines a variety of spectrum preprocessing algorithms and an established quantitative analysis model to process the received spectrum data in real time to predict the content of the specified quality marker. And the display screen displays historical data, content change trends and end stop suggestions in real time; the processor displays the processed data to the user in real time through the display screen, providing intuitive production process monitoring and quality feedback, and the data includes historical data, content change trends, end stop suggestions and other information. Information on changes in substance content during the Chinese medicine production process is saved in memory for subsequent production data analysis, quality traceability and archiving needs.

[0073] In some embodiments, Figure 3 and Fig. 9 As shown, based on the proposed Chinese medicine detection system, a Chinese medicine detection method is also provided, including:

[0074] S1. In modeling mode, update the liquid to be tested in the pipeline:

[0075] The metering pump 509, the second solenoid valve 505 and the third solenoid valve 507 are started, and the medicinal liquid enters the pretreatment pipeline from the pipeline of the original Chinese medicine production equipment through the medicinal liquid input pipe 4 and continues to circulate until the medicinal liquid in the pipeline is fully updated.

[0076] S2. Control the liquid to the specified temperature:

[0077] The metering pump 509, the second solenoid valve 505 and the third solenoid valve 507 are closed to stop the flow of the liquid medicine; the oil bath temperature control mechanism 508 is started to perform constant temperature control on the liquid medicine until the preset designated temperature is reached.

[0078] S3, push the constant temperature liquid medicine to the circulation pool 503:

[0079] The metering pump 509, the second solenoid valve 505 and the third solenoid valve 507 are started again to circulate within a preset time so that the constant temperature liquid fully covers the circulation pool 503, and then the metering pump 509, the second solenoid valve 505 and the third solenoid valve 507 are closed, and the liquid medicine is static to eliminate the interference of bubbles and flow on the detection accuracy.

[0080] S4, spectrum characteristics collected by the beam probe:

[0081] The spectrometer 9 collects spectral features and transmits the collected spectral features to the processor for storage as a basis for subsequent analysis.

[0082] S5. Offline sampling and testing of liquid medicine:

[0083] When the second solenoid valve 505 and the third solenoid valve 507 are closed, the first solenoid valve 502 is opened to connect the pipeline to the atmosphere; at this time, the sampling valve 504 is manually opened to perform offline sampling and store the samples in labeled test tubes; ensuring that the test tube labels of each batch accurately correspond to the corresponding spectral data.

[0084] Through steps S1 to S5, according to the process flow, multiple batches of tests are performed at specified time intervals, and about five batches and 200 data are sufficient.

[0085] S6. Labeling of substance content determined by high performance liquid chromatography:

[0086] The drug solutions sampled offline in the test tubes were analyzed one by one by HPLC to accurately calibrate the substance content, and the results were used as predictive labels.

[0087] S7. Establish a quantitative analysis model for online detection:

[0088] The spectral data is stored in the processor, and the obtained material content label data is transmitted to the processor through the keyboard or memory. The processor will automatically build a model using the internal algorithm module and automatically determine the optimal model based on the evaluation index. The collected spectral features and the material content labels determined by high-performance liquid chromatography are subjected to necessary data preprocessing. First, the entire data is screened for outliers, and then different preprocessing steps are performed on the spectral data to optimize the data quality. Subsequently, the preprocessed data is used to model the quantitative analysis model using a variety of algorithms such as partial least squares (PLS), decision tree regression (DTR) and gradient boosting decision tree (GBDT), and the model is selected based on the performance of the model. Finally, the determined model is saved in the processor to achieve fast and accurate content prediction of the specified drug variety in online mode.

[0089] Optionally, the spectrum of the extraction process of Danhong injection is used as an example to introduce the quantitative analysis model. Four batches were repeated in four days, and a total of 203 samples were obtained. The spectral data and content label data were obtained by the method in this embodiment. The spectral data is stored in the processor, and the obtained content label data is transmitted to the processor through the keyboard or memory. The processor will automatically model using the internal algorithm module and automatically determine the optimal model according to the evaluation index.

[0090] Screening outliers: Compared with offline data, online spectral data collection will have more interference factors, such as residual bubbles, so the proportion of outliers will increase. It is necessary to strictly screen outliers to prevent them from affecting the accuracy of the entire quantitative analysis model. In this embodiment, principal component analysis (PCA) outlier detection is used to calculate the Mahalanobis distance. Points outside the confidence interval are outliers, such as Fig.10 As shown, there are 13 outlier points among the 203 data in this group, which are not used as the basis for modeling.

[0091] At the same time, from the online detection mode, if the new data is not within the confidence interval, it may be because there are bubbles in the liquid to be tested. The processor will re-instruct the spectrometer 9 to collect data to eliminate interference factors. As shown in Table 1, the determination coefficient (R-Square, R2), root mean square error (Root Mean Squared Error, RMSE) and R square prediction deviation (Relative Prediction Deviation, RPD) are used as the main evaluation indicators, and the algorithm effect before and after removing outliers based on the partial least squares model is compared:

[0092] Table 1 Comparison of algorithm effects before and after removing outliers based on the partial least squares model

[0093]

[0094] The spectral features are processed using a variety of preprocessing methods, including standardization (AU), standard normal variable transformation (SNV), first-order derivative (FD), Savitzky-Golay smoothing (SG) and mean centering (MS), to obtain different processed spectral data, such as Figure 11~Figure 16 Based on these different spectral data, a variety of modeling algorithms are used, including partial least squares method used in statistical learning model, decision tree regression used in machine learning model and gradient boosting decision tree used in ensemble learning model, to determine the optimal preprocessing method of each algorithm as the best effect of each model. The modeling effect is shown in Table 2, and the fitting effect is compared. Fig.17 , Fig.18 and Fig.19 As shown. Taking the root mean square error (RMSE) as the main evaluation index, the gradient boosting decision tree algorithm after standard normal variable transformation preprocessing is determined as the optimal model and stored in the processor. In the online detection mode, this model is called to predict the content of the collected online data in real time, and the verification effect is as follows Fig. 20 shown.

[0095] Table 2 Comparison of modeling effects of different preprocessing and algorithms

[0096]

[0097] In summary, the best effect is achieved when the gradient boosting decision tree (GBDT) algorithm model and standard normal variable transformation (SNV) preprocessing are used.

[0098] S8, online detection mode:

[0099] S8.1. The start of the online mode is the same as the offline mode, including updating the drug solution to be tested in the pipeline, controlling the drug solution to a specified temperature, pushing the constant temperature drug solution to the circulation pool 503, and collecting spectral characteristics by the irradiation probe. These steps ensure the stability of the drug solution and the accuracy of the spectral data during the online detection process.

[0100] S8.2. Predicting the content by the quantitative model in the processor:

[0101] The spectral data collected online is transmitted to a high-performance processor in real time. The processor uses the data processing method and quantitative analysis model pre-established in the modeling mode to quickly process the real-time spectral data to predict the content of the specified quality marker. This process realizes real-time monitoring and rapid feedback of key quality attributes in the production process of traditional Chinese medicine.

[0102] S8.3, the display screen shows the substance content and historical trend in real time:

[0103] The processor displays the predicted material content information, content change trend, possible production suggestions, etc. in real time on the display screen. This step not only provides producers with intuitive and immediate production data feedback, but also helps them better understand the stability of the production process and the consistency of product quality, so as to make corresponding adjustments and optimization decisions.

[0104] Through the above-mentioned embodiment, the filter, oil bath heating module and metering pump are integrated, which can effectively filter impurities, keep the liquid temperature constant, and realize the circulation and renewal of the liquid. During the detection, the liquid is kept stationary, external interference is reduced, and the accuracy of online detection is improved. At the same time, the system has an offline sampling function, which can realize the offline analysis of samples without affecting the pressure in the tank, and provide high-quality data support for the establishment of near-infrared spectroscopy quantitative analysis models. The system processes the spectral data in real time through the processor, predicts the content of the specified quality marker, and displays historical data, content change trends, and endpoint stop suggestions in real time through the display screen; it not only improves the integration of data processing and display, but also enhances the real-time monitoring and optimization capabilities of the production process, ensuring that the key quality attributes of the drugs in the production process can be monitored and adjusted in real time, thereby improving the quality consistency and stability of oral Chinese medicine preparations. In view of the presence of flammable gases such as ethanol in the traditional Chinese medicine production workshop, a series of explosion-proof designs and safety measures are adopted. All or part of the external electrical components use explosion-proof electrical components to ensure safety and reliability in flammable environments. In addition, the spectral acquisition and data analysis system and the control system are integrated in an explosion-proof positive pressure cabinet, and the optical fiber is also explosion-proof and sealed, which further improves the system's operational safety in complex environments and greatly enhances the system's application range and safety assurance under harsh production conditions.

[0105] The above description is only a preferred embodiment of the present embodiment and is not intended to limit the present embodiment. For those skilled in the art, the present embodiment may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present embodiment shall be included in the protection scope of the present embodiment.

Claims

1. Traditional Chinese medicine detection system, characterized in that: It includes an extraction tank, a liquid pretreatment and sampling device connected to the extraction tank, and an explosion-proof positive pressure cabinet (7); The liquid medicine pretreatment and sampling device comprises a second filter (510), a second solenoid valve (505), a metering pump (509), an oil bath temperature control mechanism (508), a circulation pool (503) and a third solenoid valve (507) which are sequentially connected via pipelines; a breathing valve (506) is provided between the second solenoid valve (505) and the metering pump (509) via the first solenoid valve (502); a sampling valve (504) is provided between the oil bath temperature control mechanism (508) and the circulation pool (503); The two ends of the pipeline are connected to the extraction tank (1) via a reflux pipe (8) and a liquid medicine input pipe (4) respectively; the oil bath temperature control mechanism (508) comprises a box (5081), a special-shaped conduit (5083) arranged in the box (5081), and a heating rod (5084) arranged on the box (5081); the box (5081) is filled with silicone oil, and a temperature sensor (501) is arranged on the box (5081); the sampling valve (504) is connected to a spectrometer (9), and the spectrometer (9) is arranged in the explosion-proof positive pressure cabinet (7); Optical fiber probes (5032) are provided at both ends of the circulation pool (503), and a sealing ring (5031) is provided between the optical fiber probe (5032) and the circulation pool (503); The optical fiber probe (5032) is connected to an interface (5034) for connecting to the spectrometer (9) via an optical fiber; a sealing bushing (5033) is provided between the optical fiber and the explosion-proof positive pressure cabinet (7), and a sealant (50331) is provided between the sealing bushing (5033) and the rubber sleeve (50332) of the optical fiber.

2. The Chinese medicine detection system according to claim 1, characterized in that: The box body (5081) is provided with a window (5082).

3. The Chinese medicine detection system according to claim 1, characterized in that: The spectrometer (9) is connected to a processor, and the processor is connected to a display screen; the processor is arranged in the explosion-proof positive pressure cabinet (7).

4. A method for detecting traditional Chinese medicine, characterized in that: The Chinese medicine detection system according to any one of claims 1 to 3 is used, comprising: The metering pump (509), the second solenoid valve (505) and the third solenoid valve (507) are started, and the medicinal liquid enters the pretreatment pipeline from the pipeline of the original Chinese medicine production equipment through the medicinal liquid input pipe (4) and circulates continuously, thereby updating the medicinal liquid in the pipeline; the metering pump (509), the second solenoid valve (505) and the third solenoid valve (507) are closed, and the flow of the medicinal liquid stops; the oil bath temperature control mechanism (508) is started to control the temperature of the medicinal liquid until a preset specified temperature is reached; the metering pump (509), the second solenoid valve (505) and the third solenoid valve (507) are started again, and circulation is carried out within a preset time, so that the constant temperature liquid fully covers the periphery of the circulation pool (503); then the metering pump (509), the second solenoid valve (505) and the third solenoid valve (507) are closed, and the medicinal liquid is stationary to eliminate bubbles and flow; The spectrometer (9) collects spectral features and stores the spectral features; when the second solenoid valve (505) and the third solenoid valve (507) are closed, the first solenoid valve (502) is opened to connect the pipeline to the atmosphere; at this time, the sampling valve (504) is opened to take samples offline and store them in a test tube with a label attached; the liquid medicine samples in the test tubes are analyzed one by one by high performance liquid chromatography to calibrate the substance content, and the results are used as substance content labels; The spectral features and substance content labels are screened for outliers, and then the spectral data is subjected to different preprocessing steps to optimize the data quality; the quantitative analysis model is modeled using partial least squares, decision tree regression algorithm, and gradient boosting decision tree algorithm, and the quantitative analysis model is selected based on its performance; the determined quantitative analysis model is saved in the processor to realize the prediction of drug content in online mode.

5. The method for detecting traditional Chinese medicine according to claim 4, characterized in that: The quantitative analysis model is a gradient boosting decision tree algorithm model. The data before prediction uses principal component analysis to detect outliers, calculates the Mahalanobis distance, deletes points beyond the confidence interval as outliers, and processes the spectral features using standard normal variable transformation.

Images