Ginseng extraction system and method

Abstract

According to the ginseng extraction system and method provided by the invention, the dual goals of high purity and high yield of ginsenoside extraction are achieved through raw material adaptive parameter pre-optimization, self-adaptive cascade purification and activity protection, the enrichment effect of high-value monomer saponin is remarkable, and the loss of active ingredients is accurately controlled. According to the application, a detection-decision-regulation full-process closed-loop intelligent control system is constructed, near infrared spectrum rapid screening and HPLC-MS / MS accurate quantitative dual-mode online detection are fused, real-time and accurate monitoring of the saponin content is realized, and off-line detection and artificial experience parameters are not needed; the multi-target central control unit can automatically decide a purification path and process parameters according to detection data, and designs such as membrane pollution on-line monitoring, automatic backwashing and solvent circulating preheating are combined, so that continuous operation of the process is ensured, and production interruption caused by equipment failure and process fluctuation is avoided.

Description

A ginseng extraction system and method Technical Field

[0001] This application relates to the field of extraction technology, and in particular to a ginseng extraction system and method. Background Technology

[0002] Ginsenosides are the core active ingredients of ginseng, possessing various pharmacological effects such as anti-fatigue, immune regulation, anti-tumor, and neuroprotection. They represent the core of ginseng's medicinal and commercial value and are now widely used in pharmaceuticals, health products, cosmetics, and other fields. With the modernization of traditional Chinese medicine and the increasing market demand for high-purity ginsenoside products, achieving efficient extraction and purification of ginsenosides has become a key research focus and a core requirement for industrial production in the field of extracting effective components from traditional Chinese medicine. This has also driven the research and application of various ginsenoside extraction methods.

[0003] Currently, the main industrial extraction methods for ginsenosides include three processes: alcohol extraction, ultrasound / microwave-assisted extraction, and supercritical fluid extraction. Some improved methods also attempt to optimize traditional processes. Alcohol extraction is the most widely used basic method due to its simplicity and low cost, using ethanol or methanol reflux to dissolve saponins, but it suffers from poor selectivity. Ultrasound / microwave-assisted extraction utilizes physical effects to accelerate saponin dissolution, increasing the extraction rate, but it suffers from localized high temperatures. Supercritical fluid extraction offers advantages such as low temperature and environmental friendliness, making it more suitable for extracting fat-soluble components, but its extraction efficiency for polar ginsenosides is extremely low. Furthermore, some improved methods only introduce single purification units such as membrane separation and column chromatography to attempt to improve saponin purity, without adapting the process to the characteristics of the raw materials. They also rely on offline detection and empirical parameters, resulting in a disconnect between detection and control, failing to achieve dynamic adjustment of the extraction process and thus failing to fundamentally address the shortcomings of traditional processes.

[0004] Existing ginsenoside extraction technologies generally suffer from numerous core problems, becoming key bottlenecks restricting the industry's development: First, there is an inherent contradiction between purity and yield; traditional processes often sacrifice yield to improve ginsenoside purity, with total yield typically below 60%. Second, raw material adaptability is poor; fixed process parameters cannot match ginseng raw materials from different origins, ages, and processing methods, leading to significant fluctuations in product quality. Third, there is hidden loss of active ingredients; high temperatures, oxidation, and strong solvents in the extraction environment can easily cause isomerization, hydrolysis, or oxidation of heat-sensitive and easily oxidized saponins, reducing product activity. Fourth, the process is not economically or environmentally friendly enough; solvent recovery efficiency is low, wastewater discharge is large, and energy consumption is concentrated in the purification stage, resulting in high production costs. Fifth, quality control methods are simplistic; existing processes only focus on total saponin purity, neglecting the targeted enrichment of high-value monomeric saponins such as Rg3 and Rh2, and lacking precise control over impurities such as heavy metals and microorganisms, making it difficult to meet the quality requirements of high-end products. Summary of the Invention

[0005] This application provides a ginseng extraction system and method to solve the problem of the contradiction between purity and yield in the existing ginsenoside extraction technology.

[0006] In a first aspect, this application provides a ginseng extraction system, comprising: a raw material pretreatment and analysis unit, an intelligent dynamic extraction unit, and a dual-mode online detection unit connected in sequence; further comprising a multi-target central control unit signal-connected to the dual-mode online detection unit, and an adaptive cascade purification unit and a green solvent recovery unit signal-connected to the multi-target central control unit and sequentially connected to the dual-mode online detection unit, and also comprising a multi-dimensional quality verification unit connected to the adaptive cascade purification unit; the raw material pretreatment and analysis unit includes a raw material pulverization and grading module and a near-infrared spectroscopy rapid analysis module; the raw material pulverization and grading module is a low-temperature cryogenic pulverization structure, equipped with an 80-120 mesh grading sieve, and the near-infrared spectroscopy rapid analysis module can detect the initial content of total saponins, the proportion of major monomeric saponins, and the types of major impurities in ginseng raw materials and generate a raw material characteristic spectrum; the intelligent dynamic extraction unit includes a temperature-pressure- The extraction tank is equipped with a solvent concentration co-regulation system, an inert gas protection module, and a solvent circulation preheating module. The inert gas protection module is a nitrogen replacement structure, and the solvent circulation preheating module can maintain a stable solvent temperature during the extraction process, with a solvent circulation rate of 1~2L / min. The multi-objective central control unit has a built-in neural network and genetic algorithm model, receives detection data from the dual-mode online detection unit, and makes decisions and regulates the extraction and purification path and process parameters.

[0007] In some possible implementations, the adaptive cascade purification unit includes a pretreatment membrane module, a composite adsorption chromatography module, and a smart crystallization module; the pretreatment membrane module is a two-stage membrane separation structure, and the composite adsorption chromatography module achieves smart gradient elution by changing the type of chromatography column.

[0008] In some possible implementations, the pretreatment membrane module includes a ceramic microfiltration membrane and a nanofiltration membrane; the ceramic microfiltration membrane has a pore size cutoff of 0.1 μm, and the nanofiltration membrane has an initial molecular weight cutoff of 500~1000 Da; the pretreatment membrane module is equipped with a pressure differential sensor and an automatic backwashing module, and the automatic backwashing module triggers the backwashing procedure when the pressure difference across the membrane is ≥0.1 MPa.

[0009] In some possible implementations, the chromatography column of the composite adsorption chromatography module includes a macroporous adsorption resin-cation exchange resin composite column and a macroporous adsorption resin-molecularly imprinted polymer composite column.

[0010] In some possible implementations, the green solvent recovery unit is a membrane distillation-distillation combined unit, including a membrane distillation module and a distillation column; the impurity rejection rate of the membrane distillation module is ≥98%, and the theoretical plate number of the distillation column is 30~50.

[0011] In some possible implementations, the multidimensional quality verification unit integrates an HPLC-MS / MS monomer quantification module, an inductively coupled plasma mass spectrometry heavy metal detection module, a microbial limit detection module, and a crystal morphology analysis module; the crystal morphology analysis module includes a laser particle size analyzer and a polarizing microscope, used to verify the crystal particle size distribution and morphology.

[0012] In some possible implementations, the automatic backwashing module uses a mixture of 70% ethanol and 0.05% neutral detergent for backwashing for 5 to 10 minutes, wherein the neutral detergent is polysorbate 80.

[0013] In some possible implementations, the concentration conditions of the reduced pressure concentration structure are a temperature of 40°C to 45°C and a pressure of 0.05 MPa, so as to concentrate the liquid to 1 / 5 of its original volume.

[0014] Secondly, this application provides a ginseng extraction method, which is applied to the aforementioned ginseng extraction system. The method includes: performing low-temperature freeze-crushing and grading of ginseng raw materials; obtaining raw material characteristic spectra through rapid NIR analysis; automatically optimizing initial extraction parameters by a multi-objective central control unit based on a neural network model; introducing inert gas to replace air in the extraction tank; adding optimized solvent and ginseng powder to initiate dynamic cyclic extraction; periodically collecting extract samples through a dual-mode online detection unit, simultaneously monitoring the total saponin purity and monomeric saponin content and feeding them back to the multi-objective central control unit; setting dual judgment thresholds to determine whether the initial extract meets the standards; if the initial extract does not meet the standards, the multi-objective central control unit dynamically adjusts the purification path and parameters according to the detection results, sequentially performing one or more purification operations from pretreatment membrane separation, composite adsorption chromatography, and intelligent crystallization; and detecting again after purification; sending the solvent waste generated during extraction and purification to a green solvent recovery unit for membrane distillation. The combined distillation and purification process enables high-purity solvent recovery and recycling; the product is subjected to monomer quantification, heavy metal detection, microbial limit detection and crystal morphology analysis through a multi-dimensional quality verification unit, thus completing the extraction and purification of ginsenosides.

[0015] In some possible implementation methods, the standards for multidimensional quality verification are lead ≤ 0.5 mg / kg, cadmium ≤ 0.3 mg / kg, total bacterial count ≤ 100 CFU / g, mold and yeast ≤ 10 CFU / g, crystal particle size distribution 10~50 μm and morphology regular prismatic.

[0016] As described above, this application provides a ginseng extraction system and method. Through raw material-adaptive parameter pre-optimization, adaptive cascade purification, and activity protection, this application achieves the dual goals of high purity and high yield in ginsenoside extraction, with significant enrichment of high-value monomeric saponins and precise control over the loss of active ingredients. The total saponin purity consistently reaches 95%~98%, the total yield is ≥85%, and the active ingredient loss rate is ≤5%. Simultaneously, the intelligent crystallization process precisely controls the crystal size and morphology, resulting in a product crystal size distribution of 10~50μm with a regular prismatic shape, significantly improving the solubility and stability of the ginsenoside product. This application constructs a closed-loop intelligent control system covering the entire process of "detection-decision-control," integrating dual-mode online detection of near-infrared spectroscopy for rapid screening and HPLC-MS / MS for precise quantification. This enables real-time and accurate monitoring of saponin content without relying on offline detection or manual experience parameters. The multi-objective central control unit can automatically determine the purification path and process parameters based on the detection data. Combined with online membrane fouling monitoring, automatic backwashing, and solvent circulation preheating, this ensures continuous operation of the process and avoids production interruptions caused by equipment failures and process fluctuations. Attached Figure Description

[0017] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 is a schematic diagram of the ginseng extraction system provided in an embodiment of this application. Detailed Implementation

[0019] The embodiments described in the following examples do not represent all embodiments consistent with this application. They are merely examples of systems and methods consistent with some aspects of this application as detailed in the claims.

[0020] Currently, the main industrial extraction methods for ginsenosides include three processes: alcohol extraction, ultrasound / microwave-assisted extraction, and supercritical fluid extraction. Some improved methods also attempt to optimize traditional processes. Alcohol extraction is the most widely used basic method due to its simplicity and low cost, using ethanol or methanol reflux to dissolve saponins, but it suffers from poor selectivity. Ultrasound / microwave-assisted extraction utilizes physical effects to accelerate saponin dissolution, increasing the extraction rate, but it suffers from localized high temperatures. Supercritical fluid extraction offers advantages such as low temperature and environmental friendliness, making it more suitable for extracting fat-soluble components, but its extraction efficiency for polar ginsenosides is extremely low. Furthermore, some improved methods only introduce single purification units such as membrane separation and column chromatography to attempt to improve saponin purity, without adapting the process to the characteristics of the raw materials. They also rely on offline detection and empirical parameters, resulting in a disconnect between detection and control, failing to achieve dynamic adjustment of the extraction process and thus failing to fundamentally address the shortcomings of traditional processes.

[0021] Existing ginsenoside extraction technologies generally suffer from numerous core problems, becoming key bottlenecks restricting the industry's development: First, there is an inherent contradiction between purity and yield; traditional processes often sacrifice yield to improve ginsenoside purity, with total yield typically below 60%. Second, raw material adaptability is poor; fixed process parameters cannot match ginseng raw materials from different origins, ages, and processing methods, leading to significant fluctuations in product quality. Third, there is hidden loss of active ingredients; high temperatures, oxidation, and strong solvents in the extraction environment can easily cause isomerization, hydrolysis, or oxidation of heat-sensitive and easily oxidized saponins, reducing product activity. Fourth, the process is not economically or environmentally friendly enough; solvent recovery efficiency is low, wastewater discharge is large, and energy consumption is concentrated in the purification stage, resulting in high production costs. Fifth, quality control methods are simplistic; existing processes only focus on total saponin purity, neglecting the targeted enrichment of high-value monomeric saponins such as Rg3 and Rh2, and lacking precise control over impurities such as heavy metals and microorganisms, making it difficult to meet the quality requirements of high-end products.

[0022] Based on this, this application provides a ginseng extraction system and method. Through raw material adaptation parameter pre-optimization, adaptive cascade purification, and activity protection, this application achieves the dual goals of high purity and high yield in ginsenoside extraction, with significant enrichment of high-value monomeric saponins and precise control over the loss of active ingredients. The total saponin purity consistently reaches 95%~98%, the total yield is ≥85%, and the active ingredient loss rate is ≤5%. Simultaneously, the intelligent crystallization process precisely controls the crystal size and morphology, resulting in a product crystal size distribution of 10~50μm with a regular prismatic shape, significantly improving the solubility and stability of the ginsenoside product.

[0023] As shown in Figure 1, in some embodiments, this application provides a ginseng extraction system, including: a raw material pretreatment and analysis unit, an intelligent dynamic extraction unit, and a dual-mode online detection unit connected in sequence; it also includes a multi-target central control unit signal-connected to the dual-mode online detection unit, and an adaptive cascade purification unit and a green solvent recovery unit signal-connected to the multi-target central control unit and sequentially connected to the dual-mode online detection unit, and a multi-dimensional quality verification unit connected to the adaptive cascade purification unit; the raw material pretreatment and analysis unit includes a raw material pulverization and grading module and a near-infrared spectroscopy rapid analysis module; the raw material pulverization and grading module is a low-temperature cryogenic pulverization structure, equipped with an 80-120 mesh grading sieve, and the near-infrared spectroscopy rapid analysis module can detect the initial content of total saponins, the proportion of major monomeric saponins, and the types of major impurities in ginseng raw materials and generate a raw material characteristic spectrum; the intelligent dynamic extraction unit includes a temperature-pressure- The extraction tank is equipped with a solvent concentration co-regulation system, an inert gas protection module, and a solvent circulation preheating module. The inert gas protection module is a nitrogen replacement structure, and the solvent circulation preheating module can maintain a stable solvent temperature during the extraction process, with a solvent circulation rate of 1~2L / min. The multi-objective central control unit has a built-in neural network and genetic algorithm model, receives detection data from the dual-mode online detection unit, and makes decisions and regulates the extraction and purification path and process parameters.

[0024] In some embodiments, the adaptive cascade purification unit includes a pretreatment membrane module, a composite adsorption chromatography module, and an intelligent crystallization module; the pretreatment membrane module is a two-stage membrane separation structure, and the composite adsorption chromatography module achieves intelligent gradient elution by changing the type of chromatography column.

[0025] In some embodiments, the pretreatment membrane module includes a ceramic microfiltration membrane and a nanofiltration membrane; the ceramic microfiltration membrane has a pore size of 0.1 μm, and the nanofiltration membrane has an initial molecular weight cutoff of 500~1000 Da; the pretreatment membrane module is equipped with a pressure differential sensor and an automatic backwashing module, and the automatic backwashing module triggers the backwashing procedure when the pressure difference across the membrane is ≥0.1 MPa.

[0026] In some embodiments, the chromatography column of the composite adsorption chromatography module includes a macroporous adsorption resin-cation exchange resin composite column and a macroporous adsorption resin-molecularly imprinted polymer composite column.

[0027] In some embodiments, the green solvent recovery unit is a membrane distillation-distillation combined unit, including a membrane distillation module and a distillation column; the impurity rejection rate of the membrane distillation module is ≥98%, and the theoretical plate number of the distillation column is 30~50.

[0028] In some embodiments, the multidimensional quality verification unit integrates an HPLC-MS / MS monomer quantification module, an inductively coupled plasma mass spectrometry heavy metal detection module, a microbial limit detection module, and a crystal morphology analysis module; the crystal morphology analysis module includes a laser particle size analyzer and a polarizing microscope, used to verify the crystal particle size distribution and morphology.

[0029] In some embodiments, the automatic backwashing module uses a mixture of 70% ethanol and 0.05% neutral detergent for backwashing, with a rinsing time of 5-10 minutes. The neutral detergent is polysorbate 80.

[0030] In some embodiments, the concentration conditions of the reduced pressure concentration structure are a temperature of 40℃~45℃ and a pressure of 0.05MPa, so as to concentrate the liquid to 1 / 5 of its original volume; the natural antioxidant added by the antioxidant addition module is glutathione, and the addition amount is 0.01%~0.05%; the seed crystal is ginsenoside seed crystal with a particle size of 10~20μm, and the addition amount is 0.1%~0.3%.

[0031] Secondly, this application provides a ginseng extraction method, which is applied to the aforementioned ginseng extraction system. The method includes: performing low-temperature freeze-crushing and grading of ginseng raw materials; obtaining raw material characteristic spectra through rapid NIR analysis; automatically optimizing initial extraction parameters by a multi-objective central control unit based on a neural network model; introducing inert gas to replace air in the extraction tank; adding optimized solvent and ginseng powder to initiate dynamic cyclic extraction; periodically collecting extract samples through a dual-mode online detection unit, simultaneously monitoring the total saponin purity and monomeric saponin content and feeding them back to the multi-objective central control unit; setting dual judgment thresholds to determine whether the initial extract meets the standards; if the initial extract does not meet the standards, the multi-objective central control unit dynamically adjusts the purification path and parameters according to the detection results, sequentially performing one or more purification operations from pretreatment membrane separation, composite adsorption chromatography, and intelligent crystallization; and detecting again after purification; sending the solvent waste generated during extraction and purification to a green solvent recovery unit for membrane distillation. The combined distillation and purification process enables high-purity solvent recovery and recycling; the product is subjected to monomer quantification, heavy metal detection, microbial limit detection and crystal morphology analysis through a multi-dimensional quality verification unit, thus completing the extraction and purification of ginsenosides.

[0032] In some embodiments, the standards for multidimensional quality verification are lead ≤ 0.5 mg / kg, cadmium ≤ 0.3 mg / kg, total bacterial count ≤ 100 CFU / g, mold and yeast ≤ 10 CFU / g, crystal particle size distribution of 10~50 μm and regular prismatic morphology.

[0033] In some embodiments, the cryogenic pulverization temperature is -20℃ to -10℃, and the grading and screening adopts an 80-120 mesh sieve; the optimized initial extraction parameters include: ethanol concentration 65%-75%, material-liquid ratio 1:8-1:12, extraction temperature 55℃-65℃, extraction pressure 0.2-0.4MPa, and extraction time 1.5-2.5 hours.

[0034] In some embodiments, the dual-mode online detection cycle is to collect the extract sample every 8 minutes, and to perform HPLC-MS / MS precise quantification after every 2 NIR rapid screenings; the dual judgment thresholds are total saponin purity ≥90% and high-value monomer saponin (Rg3+Rh2) content ≥5%.

[0035] In some embodiments, pretreatment membrane separation is a mandatory step, employing a two-stage membrane separation process of "ceramic microfiltration membrane + nanofiltration membrane". The ceramic microfiltration membrane removes large particulate impurities and cellulose, while the nanofiltration membrane removes large molecular impurities such as polysaccharides and proteins. When the pressure difference across the membrane is ≥0.1MPa, a backwashing procedure is automatically triggered, and membrane separation operation resumes after backwashing is completed.

[0036] In some embodiments, the composite adsorption chromatography selects the column type based on the monomer saponin ratio detected by HPLC-MS / MS and adopts an intelligent gradient elution process. When the Rg3 elution peak is detected, the ethanol concentration increase rate is reduced from 5% / min to 2% / min. If the total saponin purity after purification is 90%~94% and the monomers meet the standards, the process directly proceeds to the intelligent crystallization step.

[0037] In some embodiments, the intelligent crystallization process includes: concentrating the chromatography eluent under reduced pressure, adding natural antioxidants, and then feeding it into an intelligent crystallizer. The crystal state is monitored in real time by an online monitoring module. The cooling rate and cooling range are adaptively adjusted by a multi-target central control unit. Ginsenoside seed crystals are added if necessary. After crystallization, the crystals are separated by centrifugation at 8000-10000 r / min and washed 2-3 times with 99.5% high-purity ethanol.

[0038] In some embodiments, the membrane distillation module of the green solvent recovery unit has a rejection rate of ≥98%, and the purified ethanol obtained after distillation in the distillation column has a purity of ≥99.5%. After the purity is confirmed by NIR detection, it is recycled for extraction, elution or washing processes, and the overall solvent recovery rate is ≥95%.

[0039] This application achieves the dual goals of high purity and high yield in ginsenoside extraction through raw material adaptation parameter pre-optimization, adaptive cascade purification, and activity protection. It also demonstrates significant enrichment of high-value monomeric saponins and precise control over the loss of active ingredients. The total saponin purity consistently reaches 95%–98%, with a total yield ≥85% and an active ingredient loss rate ≤5%. Simultaneously, the intelligent crystallization process precisely controls crystal size and morphology, resulting in a product crystal size distribution of 10–50 μm with a regular prismatic shape, significantly improving the solubility and stability of the ginsenoside product.

[0040] By using near-infrared spectroscopy rapid analysis of raw material pretreatment and intelligent parameter pre-optimization of neural network models, the optimal extraction temperature and pressure, solvent concentration, material-liquid ratio and other process parameters can be automatically matched according to the characteristics of ginseng raw materials of different origins, ages and processing methods (sun-dried ginseng, red ginseng, forest-grown ginseng), which completely solves the problem that the fixed parameters of traditional processes cannot be adapted to different raw materials.

[0041] This application constructs a closed-loop intelligent control system covering the entire process of "detection-decision-control," integrating dual-mode online detection of near-infrared spectroscopy for rapid screening and HPLC-MS / MS for precise quantification. This enables real-time and accurate monitoring of saponin content without relying on offline detection or manual experience parameters. The multi-objective central control unit can automatically determine the purification path and process parameters based on the detection data. Combined with online membrane fouling monitoring, automatic backwashing, and solvent circulation preheating, this ensures continuous operation of the process and avoids production interruptions caused by equipment failures and process fluctuations.

[0042] Example Step 1: Raw Material Pretreatment and Extraction Parameter Pre-optimization 1) Ginseng raw materials (sun-dried ginseng / red ginseng / forest-grown ginseng) are cryogenically pulverized at low temperature (-20℃~-10℃) and screened through a grading sieve (80~120 mesh) to ensure uniform powder particle size and avoid uneven local concentration during extraction; 2) A small amount of pulverized sample is taken, and the initial content of total saponins, the proportion of major monomeric saponins (Rg1, Re, Rb1, Rg3, Rh2), and the types of major impurities (polysaccharides, proteins, tannins) in the raw materials are detected by the NIR rapid analysis module to generate a raw material characteristic spectrum; 3) The central control unit calls the preset neural network model, inputs the raw material characteristic spectrum, and automatically optimizes the initial extraction parameters: Ethanol concentration: adjusted to 65%~75% according to the saponin polarity distribution (e.g., 65% ethanol is used if the Rg3 content is high in red ginseng; 72% ethanol is used if the Re content is high in sun-dried ginseng). Ethanol); Material-to-liquid ratio: Adjust to 1:8~1:12 according to the saponin content of the raw material (1:8 for high saponin content, 1:12 for low content); Extraction conditions: Temperature 55℃~65℃, pressure 0.2~0.4MPa, extraction time 1.5~2.5 hours (lower temperature and shorter time for high content of heat-sensitive saponins); 4) Inert gas (nitrogen) is introduced into the extraction tank to replace the air, and the optimized ethanol solvent and ginseng powder are added. Dynamic circulation extraction is started (solvent circulation rate 1~2L / min). During the extraction process, the solvent temperature is kept stable by the solvent circulation preheating module to avoid saponin degradation caused by temperature fluctuations.

[0043] Step 2: Dual-mode online detection and initial extract purity determination 1) During the extraction process, the dual-mode online detection unit collects an extract sample every 8 minutes: the NIR rapid screening probe measures the total saponin purity in real time (detection time ≤ 30 seconds) to achieve continuous monitoring; after every 2 NIR detections, a small amount of sample is automatically diverted to the HPLC-MS / MS precise quantification module to determine the total saponin purity and the content of monomers such as Rg1, Re, Rb1, Rg3, and Rh2 (detection time ≤ 10 minutes), and the data is fed back to the central control unit to calibrate the NIR detection results and ensure accuracy; 2) Set dual judgment thresholds: the total saponin purity ≥ 90% and the content of high-value monomer saponins (Rg3+Rh2) ≥ 5% are considered acceptable. If the initial extract detection results meet the standards, proceed directly to step 5; if not, the adaptive cascade purification unit is activated.

[0044] Step 3: Adaptive Cascade Purification (Dynamically adjust the path based on detection results) Step 3.1: Pretreatment Membrane Separation (Required step, reduces subsequent purification pressure) 1) The initial extract first enters the pretreatment membrane module, using a two-stage membrane separation process: "ceramic microfiltration membrane (0.1μm pore size cutoff) + nanofiltration membrane (500~1000Da molecular weight cutoff)". The ceramic microfiltration membrane removes large particulate impurities, cellulose, etc., avoiding nanofiltration membrane fouling; the nanofiltration membrane removes large molecular impurities such as polysaccharides and proteins. 2) During membrane separation, the membrane fouling level is monitored in real time by a pressure differential sensor (when the pressure difference across the membrane is ≥0.1MPa), the central control unit automatically triggers the backwashing program: backwashing with 70% ethanol + 0.05% neutral cleaning agent (such as polysorbate 80) for 5~10 minutes to restore membrane flux. 3) The membrane filtrate is detected by dual-mode. If the total saponin purity is ≥90%, If the monomer meets the standard, proceed to step 5; if the purity is 80%~90%, proceed to sub-step 3.2; if the purity is <80%, adjust the molecular weight cutoff of the nanofiltration membrane (reduce by 50~100 Da) and filter again.

[0045] Step 3.2: Composite Adsorption Chromatography (Targeted Enrichment) 1) Select the column type based on the monomeric saponin ratio detected by HPLC-MS / MS: If the total saponin purity is insufficient but the monomer ratio is balanced, use a composite column of "macroporous adsorption resin (upper layer) + cation exchange resin (lower layer)". The upper layer enriches the total saponins, and the lower layer removes residual proteins, alkaloids, and other alkaline impurities. If the content of high-value monomeric saponins (Rg3, Rh2) is insufficient, use a composite column of "macroporous adsorption resin (upper layer) + molecularly imprinted polymer (MIP, lower layer, specifically recognizing Rg3 / Rh2)". 2) Start intelligent gradient elution: The central control unit dynamically adjusts the rate of change of ethanol concentration (30%→70%) based on the real-time elution peak spectrum. If the Rg3 elution peak is detected, the concentration increase rate is reduced from 5% / min to 2% / min to ensure sufficient elution of high-value monomers. 3) The eluent is detected by dual-mode analysis. If the total saponin purity is ≥90%, If the monomers meet the standard, proceed to step 5; if the purity is 90%~94% and the monomers meet the standard, proceed to sub-step 3.3; if they still do not meet the standard, perform cyclic chromatography 1~2 times (determine whether to cyclically based on the energy consumption model).

[0046] Step 3.3: Intelligent Crystallization (Optimizing Crystal Morphology and Yield) 1) Concentrate the chromatography eluent under reduced pressure (temperature 40℃~45℃, pressure 0.05MPa) to 1 / 5 of its original volume, and add 0.01%~0.05% of a natural antioxidant (such as glutathione) to prevent saponin oxidation during crystallization; 2) Send the concentrated solution into the intelligent crystallizer and start online monitoring of the crystallization process: monitor the crystal particle size in real time with a laser particle size analyzer, and observe the crystal morphology with a polarizing microscope; 3) The central control unit adaptively adjusts the crystallization parameters based on the monitoring data: Initial cooling rate: 0.3~0.7℃ / min (slower if the particle size is too small, faster if the particle size is too large); Cooling range: adjusted according to the crystallization characteristics of the monomer saponins (e.g., the optimal crystallization range for Rg3 is 35℃~5℃, and the optimal range for Rh2 is 40℃~4℃); Seed crystal addition: If no obvious crystals precipitate after 1 hour of crystallization, automatically add 0.1%~0.3% of the seed crystal. 4) After crystallization, centrifuge (8000~10000r / min) and wash with 99.5% high-purity ethanol 2~3 times to remove residual impurities on the surface.

[0047] Step 4: Solvent Recovery and Recycling 1) Ethanol wastewater generated during extraction, washing, and backwashing is centrally sent to the green solvent recovery unit: First, water and small molecule impurities in the wastewater are removed through a membrane distillation system (retention rate ≥98%) to obtain crude recovered ethanol; the crude recovered ethanol is then distilled through a distillation column (theoretical plate number 30~50) to obtain refined ethanol with a purity ≥99.5%; 2) After the purity of the refined ethanol is confirmed by NIR detection, it is recycled for extraction, elution, or washing processes, with a solvent recovery rate ≥95%, reducing production costs and wastewater discharge.

[0048] Step 5: Multidimensional Quality Validation 1) Take the crystallized product and determine the total saponin purity and the content of monomers such as Rg1, Re, Rb1, Rg3, and Rh2 by HPLC-MS / MS; 2) Use ICP-MS to detect heavy metal (lead, cadmium, mercury, arsenic) residues to ensure compliance with the standards of the Pharmacopoeia of the People's Republic of China (lead ≤ 0.5 mg / kg, cadmium ≤ 0.3 mg / kg); 3) Perform microbial limit tests (total bacterial count ≤ 100 CFU / g, molds and yeasts ≤ 10 CFU / g); 4) Use a laser particle size analyzer and polarizing microscope to verify the crystal particle size distribution (10~50 μm) and morphology (regular prismatic shape) to ensure the solubility and stability of the product.

[0049] This application provides a ginseng extraction system and method. Through raw material-adaptive parameter pre-optimization, adaptive cascade purification, and activity protection, it achieves the dual goals of high purity and high yield in ginsenoside extraction, with significant enrichment of high-value monomeric saponins and precise control of active ingredient loss. This application constructs a closed-loop intelligent control system encompassing detection, decision-making, and regulation. It integrates dual-mode online detection—near-infrared spectroscopy for rapid screening and HPLC-MS / MS for precise quantification—to achieve real-time and accurate monitoring of saponin content without relying on offline detection or manual experience parameters. The multi-objective central control unit automatically determines the purification path and process parameters based on detection data. Combined with online membrane fouling monitoring, automatic backwashing, and solvent circulation preheating, it ensures continuous process operation and avoids production interruptions caused by equipment failures and process fluctuations.

Claims

1. A ginseng extraction system, characterized in that, include: The system comprises, in sequence, a raw material pretreatment and analysis unit, an intelligent dynamic extraction unit, and a dual-mode online detection unit; it also includes a multi-target central control unit connected to the dual-mode online detection unit, and an adaptive cascade purification unit connected to both the multi-target central control unit and the dual-mode online detection unit; a green solvent recovery unit connected to both the multi-target central control unit and the adaptive cascade purification unit; and a multi-dimensional quality verification unit connected to the adaptive cascade purification unit. The raw material pretreatment and analysis unit includes a raw material pulverization and grading module and a near-infrared spectroscopy rapid analysis module. The raw material pulverization and grading module is a low-temperature cryogenic pulverization structure equipped with an 80-120 mesh grading sieve. The near-infrared spectroscopy rapid analysis module can detect the initial total saponin content, the proportion of major monomeric saponins, and the types of major impurities in ginseng raw materials and generate a raw material characteristic spectrum. The intelligent dynamic extraction unit includes a temperature-pressure extraction unit... The extraction tank is equipped with a solvent concentration co-regulation system, an inert gas protection module, and a solvent circulation preheating module. The inert gas protection module is a nitrogen replacement structure, and the solvent circulation preheating module can maintain a stable solvent temperature during the extraction process, with a solvent circulation rate of 1~2L / min. The multi-objective central control unit has a built-in neural network and genetic algorithm model, receives detection data from the dual-mode online detection unit, and makes decisions and regulates the extraction and purification path and process parameters.

2. The ginseng extraction system according to claim 1, characterized in that, The adaptive cascade purification unit includes a pretreatment membrane module, a composite adsorption chromatography module, and an intelligent crystallization module; the pretreatment membrane module is a two-stage membrane separation structure, and the composite adsorption chromatography module achieves intelligent gradient elution by changing the type of chromatography column.

3. The ginseng extraction system according to claim 2, characterized in that, The pretreatment membrane module includes a ceramic microfiltration membrane and a nanofiltration membrane; the ceramic microfiltration membrane has a pore size of 0.1 μm, and the nanofiltration membrane has an initial molecular weight cutoff of 500~1000 Da; the pretreatment membrane module is equipped with a pressure differential sensor and an automatic backwashing module, and the automatic backwashing module triggers the backwashing program when the pressure difference across the membrane is ≥0.1 MPa.

4. The ginseng extraction system according to claim 2, characterized in that, The chromatography columns of the composite adsorption chromatography module include macroporous adsorption resin-cation exchange resin composite columns and macroporous adsorption resin-molecularly imprinted polymer composite columns.

5. The ginseng extraction system according to claim 1, characterized in that, The green solvent recovery unit is a membrane distillation-distillation combined unit, including a membrane distillation module and a distillation column; the impurity rejection rate of the membrane distillation module is ≥98%, and the theoretical plate number of the distillation column is 30~50.

6. The ginseng extraction system according to claim 1, characterized in that, The multidimensional quality verification unit integrates an HPLC-MS / MS monomer quantification module, an inductively coupled plasma mass spectrometry heavy metal detection module, a microbial limit detection module, and a crystal morphology analysis module; the crystal morphology analysis module includes a laser particle size analyzer and a polarizing microscope, used to verify the crystal particle size distribution and morphology.

7. The ginseng extraction system according to claim 3, characterized in that, The automatic backwashing module uses a mixture of 70% ethanol and 0.05% neutral detergent for backwashing, with a rinsing time of 5-10 minutes. The neutral detergent is polysorbate 80.

8. The ginseng extraction system according to claim 1, characterized in that, The ginseng extraction system also includes a vacuum concentration structure, wherein the concentration conditions of the vacuum concentration structure are a temperature of 40℃~45℃ and a pressure of 0.05MPa, so as to concentrate the liquid to 1 / 5 of its original volume.

9. A method for ginseng extraction, characterized in that, The method is applied to the ginseng extraction system of claim 1. The method includes: cryogenically pulverizing and grading ginseng raw materials; obtaining raw material characteristic spectra through rapid NIR analysis; automatically optimizing initial extraction parameters by a multi-objective central control unit based on a neural network model; purging the extraction tank with inert gas to replace air; adding optimized solvent and ginseng powder to start dynamic cyclic extraction; periodically collecting extract samples through a dual-mode online detection unit, simultaneously monitoring the total saponin purity and monomer saponin content and feeding back to the multi-objective central control unit; setting dual judgment thresholds to determine whether the initial extract meets the standards; if the initial extract does not meet the standards, the multi-objective central control unit dynamically adjusts the purification path and parameters according to the detection results, sequentially performing one or more purification operations such as pretreatment membrane separation, composite adsorption chromatography, and intelligent crystallization; detecting again after purification; sending the solvent waste generated during extraction and purification to a green solvent recovery unit, achieving high-purity solvent recovery and recycling through a membrane distillation-rectification combined process; and performing monomer quantification, heavy metal detection, microbial limit detection, and crystal morphology analysis on the product through a multi-dimensional quality verification unit to complete the ginseng saponin extraction and purification.

10. The ginseng extraction method according to claim 9, characterized in that, The standards for multidimensional quality verification are: lead ≤0.5mg / kg, cadmium ≤0.3mg / kg, total bacterial count ≤100CFU / g, mold and yeast ≤10CFU / g, crystal particle size distribution 10~50μm and regular prismatic morphology.

Images