A method for evaluating quality of three-leaf gromwell based on electronic sensory technology

Abstract

The application discloses a three-leaf greenbrier quality evaluation method based on electronic sensory technology, which comprises the following steps: 1) data acquisition: using an electronic nose and an electronic tongue to collect olfactory and gustatory data of three-leaf greenbrier samples respectively; 2) model establishment: performing data fusion on the olfactory and gustatory data of the three-leaf greenbrier samples to establish a drying mode discrimination model; and 3) sample detection: inputting the olfactory and gustatory data of a three-leaf greenbrier sample to be detected into the drying mode discrimination model, determining the drying mode to which the sample belongs according to a prediction result of the discrimination model, and determining a quality grade according to the drying mode. The application establishes an integrated evaluation system of "sensory characteristics-intrinsic activity-comprehensive quality", realizes objective, rapid, nondestructive and accurate evaluation of three-leaf greenbrier in different drying modes, can efficiently screen the optimal drying process of three-leaf greenbrier in different producing areas, and provides scientific support for processing standardization and quality control of three-leaf greenbrier in different producing areas.

Description

Technical Field

[0001] This invention relates to a method for evaluating the quality of Trifoliate Oryza sativa based on electronic sensory technology. Background Technology

[0002] *Tetrastigma hemsleyanum* Diels et Gilg, a plant belonging to the genus *Tetrastigma* in the family Vitaceae, is a rare and precious medicinal plant unique to my country and is listed in the *Pharmacopoeia of the People's Republic of China*. It has a slightly bitter and pungent taste, and is cool in nature, possessing the core effects of clearing heat and detoxifying, reducing swelling and relieving pain, and resolving phlegm and nodules. Modern pharmacological research has confirmed that the flavonoids, phenolic acids, and other active ingredients in *Tetrastigma hemsleyanum* have significant anti-inflammatory, anti-tumor, immunomodulatory, and hepatoprotective pharmacological effects, forming the core material basis for its clinical efficacy and indicating broad market application prospects.

[0003] Drying is a core step in the initial processing of Chinese medicinal herbs at their place of origin. It directly determines the appearance, volatile flavor compounds, taste characteristics, and retention rate of key active ingredients in *Tripterygium wilfordii*, thus affecting its medicinal quality, clinical efficacy, and economic value. Currently, commonly used drying methods in *Tripterygium wilfordii* processing include natural sun drying, hot air drying, vacuum drying, freeze drying, microwave drying, and infrared drying. Different drying processes exhibit significant differences in temperature and humidity, drying time, and mass and heat transfer methods, leading to marked variations in the sensory and intrinsic quality of *Tripterygium wilfordii*. Therefore, establishing a scientific, comprehensive, and quantifiable quality evaluation method is a crucial prerequisite for selecting the optimal drying process for *Tripterygium wilfordii* and standardizing its processing at the place of origin.

[0004] The existing quality evaluation technology for Tripterygium wilfordii has the following significant drawbacks: Traditional human sensory evaluation is highly subjective and has poor repeatability: the industry still generally relies on pharmacists to evaluate the sensory quality of Tripterygium wilfordii, such as appearance, smell, taste, etc., by "seeing, touching, smelling, and tasting". The evaluation results are greatly affected by personnel experience, physiological state, and environmental factors, and cannot be quantified and standardized, let alone achieve consistent evaluation of large batches of samples.

[0005] Conventional physicochemical testing methods have significant limitations: current pharmacopoeias and existing studies mostly use high-performance liquid chromatography (HPLC) and ultra-high-performance liquid chromatography-tandem mass spectrometry (UPLC-MS / MS) to determine the content of one or two indicator components in *Tripterygium wilfordii* to evaluate the quality of the medicinal material. These methods involve complex pretreatment, long testing cycles, and high costs, relying on large, precision instruments and specialized operators, making rapid on-site testing impossible. Furthermore, traditional Chinese medicine materials are complex systems with synergistic effects of multiple components; the content of a single indicator component cannot comprehensively reflect the overall quality of the medicinal material, nor can it correlate with its sensory characteristics and clinical efficacy.

[0006] Current technologies fail to achieve integrated evaluation of sensory characteristics and intrinsic quality: Electronic noses, electronic tongues, and other electronic sensory technologies, as biomimetic detection technologies, can simulate the human olfactory and gustatory systems to quickly, objectively, and non-destructively acquire the overall odor and taste characteristics of samples, thus digitizing sensory information. While this technology is widely used in the food and tobacco industries, its application in the quality evaluation of *Tripterygium wilfordii* is still lacking. Existing research only uses electronic noses or electronic tongues for simple qualitative differentiation, failing to achieve fusion analysis of multi-source data from electronic noses and electronic tongues, and further failing to deeply correlate sensory characteristic data with the content of key active ingredients using chemometric methods. This makes it impossible to establish a comprehensive evaluation system of "sensory characteristics - intrinsic activity - overall quality," and makes it difficult to accurately quantify the impact of different drying methods on the quality of *Tripterygium wilfordii*.

[0007] There is a lack of a dedicated evaluation system for screening drying processes of Tripterygium wilfordii: existing methods cannot simultaneously achieve rapid classification, quality grading, and process superiority / inferiority ranking of Tripterygium wilfordii dried using different methods, and cannot provide a feasible technical solution for the standardization of processing in the production areas of Tripterygium wilfordii, which seriously restricts the high-quality development of the Tripterygium wilfordii industry.

[0008] In view of the many shortcomings of the existing technologies, there is an urgent need to develop an objective, rapid, non-destructive and comprehensive method for evaluating the quality of *Tripterygium wilfordii*, to overcome the subjectivity and limitations of traditional methods in physicochemical testing, and to achieve comprehensive quality evaluation and optimal process selection for *Tripterygium wilfordii* dried by different methods. Summary of the Invention

[0009] To address the shortcomings of existing methods for evaluating the quality of Tripterygium wilfordii, such as strong subjectivity, long testing cycles, complex operations, inability to consider both sensory characteristics and intrinsic active ingredients, and difficulty in achieving rapid and comprehensive evaluation of overall quality, this invention provides a method for evaluating the quality of Tripterygium wilfordii based on electronic sensory technology.

[0010] The method for evaluating the quality of Clematis armandii based on electronic sensory technology described in this invention includes the following steps: 1) Data acquisition: Olfactory and gustatory data of the *Tripterygium wilfordii* samples were collected using an electronic nose and electronic tongue, respectively; 2) Model establishment: The olfactory and gustatory data of the above-mentioned *Tripterygium wilfordii* samples were fused to establish a drying method discrimination model; 3) Sample testing: Input the olfactory and gustatory data of the sample to be tested into the drying method discrimination model, determine the drying method according to the prediction results of the discrimination model, and determine the quality grade according to the drying method.

[0011] Furthermore, the above drying methods are natural sun drying, cool and ventilated air drying, 55℃ medium-temperature hot air drying, 50℃ constant-temperature vacuum drying, and vacuum freeze drying; the correspondence between drying methods and quality grades is as follows: vacuum freeze drying is Grade I, 50℃ constant-temperature vacuum drying is Grade II, natural sun drying and cool and ventilated air drying are Grade III, and 55℃ medium-temperature hot air drying is Grade IV.

[0012] Furthermore, natural sun drying involves evenly spreading the slices of *Trifolium repens* on a clean drying tray and placing them in the open sun to dry naturally. The trays are turned 2-3 times daily, sun-dried during the day and brought back at night to prevent moisture buildup. The ambient temperature is 20℃-30℃, and the product is dried until the moisture content is ≤12.0%. The method of drying in a cool, ventilated, and shaded environment involves evenly spreading the slices of *Trifolium repens* on a clean tray and placing them in a cool, ventilated, and dark indoor environment with an ambient temperature of 20℃-25℃ and a relative humidity of ≤60%. The tray should be turned over twice daily until the moisture content is ≤12.0%. The 55℃ medium-temperature hot air drying method involves evenly spreading the *Trifolium repens* slices on a tray in a hot air drying oven, with a layer thickness of ≤2cm. The hot air temperature is 55℃, the air velocity is 1.0m / s, and the slices are turned over once every 3 hours until the moisture content is ≤12.0%. The 50℃ constant temperature vacuum drying process involves evenly spreading the sliced ​​*Trifolium repens* (a type of medicinal herb) on a tray in a vacuum drying oven, with a layer thickness ≤2cm, a vacuum degree ≥0.08MPa, a heating temperature of 50℃, and drying until the moisture content is ≤12.0%. Vacuum freeze drying involves first pre-freezing the *Trifolium repens* slices in a -40℃ pre-freezing chamber for 4 hours, then transferring them to a vacuum freeze dryer. The cold trap temperature is ≤-50℃, the vacuum degree is ≤10Pa, the partition heating temperature is 20℃, and the process involves sublimation drying for 24-36 hours, followed by desorption drying for 4 hours, until the moisture content is ≤12.0%.

[0013] Furthermore, the detection method for the above-mentioned electronic nose is as follows: accurately weigh 2.0g of sample, equilibrate in a 50℃ constant temperature water bath for 20min, use a carrier gas flow rate of 200mL / min, clean for 40s, inject for 10s, collect data for 90s, and take the stable response value of 30-80s as valid data.

[0014] Furthermore, the detection method for the above-mentioned electronic tongue is as follows: accurately weigh 3.0g of sample, with a material-to-liquid ratio of 1:30, ultrasonic extraction at 300W for 30min, centrifugation at 4000r / min for 8min, and test each sample for 5 cycles, taking the average response value of the last 3 cycles as valid data.

[0015] Furthermore, the above data fusion method is feature layer fusion, the unsupervised analysis adopts a combination of PCA and HCA, and the discriminant model adopts the SVM model.

[0016] Beneficial effects: This invention comprehensively acquires the digital characteristics of the aroma and taste of *Tripterygium wilfordii* using electronic nose-electronic tongue technology. Combined with chemometric algorithms, it constructs a highly accurate drying method discrimination model and establishes an integrated evaluation system encompassing "sensory characteristics, intrinsic activity, and comprehensive quality." This enables objective, rapid, non-destructive, and precise evaluation of *Tripterygium wilfordii* dried using different methods. It allows for the detection of *Tripterygium wilfordii*, rapid determination of the drying method, and assessment of quality grade based on the drying method, achieving rapid classification of medicinal material quality. Furthermore, this invention can efficiently screen the optimal drying process for *Tripterygium wilfordii* from its production areas, providing scientific support for the standardization and quality control of *Tripterygium wilfordii* processing at its production sites. Attached Figure Description

[0017] Figure 1 The PCA score chart for the three drying methods of the *Trifolium repens* electronic nose in this invention is shown. Figure 2 The radar diagrams showing the responses of the electronic nose sensor to the five drying methods of *Trifolium repens* in this invention are shown. Figure 3 The PCA score diagram of the electronic tongue for the five drying methods of *Trifolium repens* in this invention; Figure 4 The bar chart shows the taste response of *Trifolium repens* via electronic tongue under five drying methods in this invention. Figure 5 This invention uses the PCA score map of the pre-trained features of the electronic nose-electronic tongue fusion feature of *Trifolium repens* for five drying methods to predict the grade of the submitted sample. Detailed Implementation

[0018] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and embodiments. However, the scope of protection of the present invention is not limited to the embodiments described.

[0019] This invention is a quality evaluation method for *Trifolium repens* based on electronic sensory technology. Using fresh *Trifolium repens* tubers from the same origin (Yaowang Valley, Suichang County, Lishui City, Zhejiang Province), the same harvest period (November 2025), and the same growth age (three years old) as raw materials, five groups of samples were prepared using different drying methods (natural sun-drying, cool and ventilated air-drying, 55℃ medium-temperature hot air drying, 50℃ constant-temperature vacuum drying, and vacuum freeze-drying). Electronic sensory data acquisition, determination of key active ingredients, construction of chemometric models, and comprehensive quality evaluation were conducted.

[0020] Basic moisture determination of samples Natural sun-drying group (Group 1): Cut fresh San Ye Qing into 3mm thick slices (San Ye Qing medicinal slices) and spread them evenly on a clean drying tray. Place them in the open sun to dry naturally, turning them over 2-3 times a day. Dry during the day and bring them back at night to prevent moisture. The ambient temperature is 20℃-30℃. Dry until the moisture content is 9.86%.

[0021] Cool and ventilated air-drying group (Group 2): Cut fresh San Ye Qing into 3mm thick slices (San Ye Qing medicinal slices) and spread them evenly on a clean tray. Place them in a cool, ventilated, and dark indoor environment with an ambient temperature of 20℃-25℃ and a relative humidity of ≤60%. Turn them over twice a day and dry them until the moisture content is 10.52%.

[0022] 55℃ medium-temperature hot air drying group (Group 3): Cut fresh San Ye Qing into thin slices with a thickness of 3mm (San Ye Qing medicinal slices) and spread them evenly on the tray of the hot air drying box. The thickness of the spread is ≤2cm. The hot air temperature is 55℃ and the wind speed is 1.0m / s. Turn it over once every 3 hours and dry it until the moisture content is 7.65%.

[0023] 50℃ constant temperature vacuum drying group (group 4): Cut fresh San Ye Qing into thin slices with a thickness of 3mm (San Ye Qing medicinal slices) and spread them evenly on the tray of the vacuum drying oven. The thickness of the spread is ≤2cm, the vacuum degree is ≥0.08MPa, the heating temperature is 50℃, and the moisture content is dried to 7.91%.

[0024] Vacuum freeze-drying group (group 5): Fresh *Trifolium repens* was cut into 3mm thick slices (*Trifolium repens* medicinal slices) and pre-frozen in a -40℃ pre-freezing chamber for 4 hours. Then it was transferred to a vacuum freeze dryer with a cold trap temperature ≤ -50℃, a vacuum degree ≤ 10Pa, a partition heating temperature of 20℃, sublimation drying for 24-36 hours, and desorption drying for 4 hours until the moisture content was 3.24%.

[0025] The final moisture content of the five groups of *Tripterygium wilfordii* samples dried by different methods all met the requirement of "moisture ≤ 12.0%" under the *Pharmacopoeia of the People's Republic of China*. All samples were successfully dried without any abnormalities such as charring, mold, or browning. The specific moisture content is shown in Table 1.

[0026] Table 1 Moisture Content

[0027] Electronic nose detection The detection method using the electronic nose (HERACLES NEO, Alpha MOS, France) was as follows: 2.0 g of each sample was precisely weighed, equilibrated in a 50°C water bath for 20 min, with a carrier gas flow rate of 200 mL / min, a washing time of 40 s, an injection time of 10 s, and a data acquisition time of 90 s. Stable response values ​​between 30 and 80 s were taken as valid data. The electronic nose used 10 sensors to detect 10 different odors. The detection results are shown in Table 2.

[0028] Table 2 Electronic nose detection results

[0029] Principal component analysis (PCA) was performed on the electronic nose dataset. The first two principal components (Principal Component 1, whose core contributing sensors are S2 (terpenes), S3 (aromatic compounds), and S8 (organic acids); and Principal Component 2, whose core contributing sensors are S1 (sulfides), S7 (nitrogen oxides), and S9 (amines)) had a cumulative contribution rate of 91.84%, representing the vast majority of the olfactory features of the samples. The results are as follows: Figure 1 As shown, in Figure 1 In the PCA score diagram, the parallel samples within each of the five groups were highly clustered, and the boundaries between groups were clear with no overlap, enabling preliminary and complete differentiation of samples from the five drying methods. Among them, the vacuum freeze-dried group was the furthest from the other groups, indicating that its olfactory characteristics differed most significantly from those of samples from other drying methods. Hierarchical clustering analysis (HCA) results showed that at a Euclidean distance threshold of 10, the vacuum freeze-dried group formed a separate class, further validating the uniqueness of its olfactory characteristics. Figure 2 The radar diagrams showing the response of the electronic nose sensor to five different drying methods for *Trifolium repens* are provided by [the relevant authority / organism]. Figure 2 It can be seen that the overall retention effect differs significantly: the response values ​​of all 10 sensors in the vacuum freeze-drying group (group 5) are the highest among the five groups of samples, and the total area covered by the radar map is significantly larger than that of the other four groups. P <0.05), indicating that the low temperature, vacuum, and oxygen-free environment of vacuum freeze drying can minimize the volatilization loss, thermal degradation, and oxidative deterioration of the volatile components of Tripterygium wilfordii, and completely preserve the inherent olfactory characteristics of Tripterygium wilfordii, which is completely consistent with the experimental results of the group with the highest content of active ingredients (see Table 5 below).

[0030] Electronic tongue detection The detection method using the electronic tongue (SA402B, Insent, Japan) was as follows: 3.0 g of each sample was precisely weighed, with a material-to-liquid ratio of 1:30 (g / mL). Extraction was performed using ultrasonic extraction at 300W for 30 min, followed by centrifugation at 4000 rpm for 8 min using a centrifuge (CF15RN, HITACHI, Japan). Each sample was tested for 5 cycles, and the average response value of the last 3 cycles was taken as valid data. The electronic tongue used 6 sensors to detect 6 different tastes. The detection results are shown in Table 3.

[0031] Table 3 Electronic tongue detection results

[0032] This test was performed using a 6-channel taste sensor array. The results showed that there were highly significant differences in the sensor response values ​​for bitterness, astringency, and sourness among the five groups of *Tripterygium wilfordii* samples. P<0.01), among which the vacuum freeze-drying group (group 5) had the highest bitter and astringent response values, which highly matched the results of having the highest content of flavonoids and phenolic acids (bitterness and astringency are the core taste characteristics of flavonoids and phenolic acids in *Trifolium repens*) (see Table 5 below); the 55℃ medium-temperature hot air drying group (group 3) had the lowest bitter and astringent response values, which is consistent with the pattern of partial degradation of active ingredients.

[0033] PCA analysis of the electronic tongue dataset showed that the first two principal components (principal component 1, whose core contributing taste indicators are bitterness (CO0) and astringency (AE1); and principal component 2, whose core contributing taste indicators are sourness (CA0) and saltiness (CT0)) had a cumulative contribution rate of 89.72%, which can comprehensively represent the taste characteristics of the samples. The results are as follows: Figure 3 As shown, in Figure 3 The PCA score graph shows good intra-group repeatability and significant inter-group differentiation among the five groups of samples, enabling effective differentiation of samples with different drying methods. Figure 4 A bar chart showing the electronic tongue taste response of *Trifolium repens* under five drying methods, such as... Figure 4 It can be seen that there are highly significant differences in the bitterness and astringency response values ​​among the five groups of samples. P <0.01), among which the bitter and astringent response values ​​of the vacuum freeze-drying group (group 5) reached 3.42 and 3.15 respectively, which were significantly higher than the other four groups. This is completely consistent with the experimental results that the total flavonoid and total phenolic acid content of this group was the highest. Bitterness and astringency are characteristic tastes of the core medicinal active ingredients of flavonoids and phenolic acids in *Trifolium repens*. The two are highly significantly positively correlated. The higher the response value, the more completely the active ingredients are preserved, which directly confirms the quality advantages of the vacuum freeze-drying process.

[0034] PCA analysis was performed on the electronic nose and electronic tongue detection datasets. The results showed that the cumulative variance contribution rate of the first two principal components reached 99.96%, which can completely represent all the fused sensory feature information of the samples. The results are as follows: Figure 5 As shown, by Figure 5It can be seen that the parallel samples within the five groups are tightly clustered, and the boundaries between groups are clear and without overlap. The core differences between samples dried by different methods are entirely determined by the first principal component (PC1, with a variance contribution rate of 99.85%), which includes the S2 (terpenes), S3 (aromatic compounds), and S8 (organic acids) sensors of the electronic nose and the bitter and astringent taste sensors of the electronic tongue. The PC1 value is significantly positively correlated with the content of the core active ingredients of the samples and can directly characterize the overall quality of the samples. The second principal component (PC2, which includes the S1 (sulfides), S7 (nitrogen oxides), and S9 (amines) sensors of the electronic nose and the sour and salty taste sensors of the electronic tongue) can only explain 0.11% of the characteristic differences, reflecting only the small fluctuations of the parallel samples within the group. The vacuum freeze-dried group is located at the farthest end of the positive axis of PC1, which is far away from the other four groups of samples, indicating that its integrated sensory characteristics and intrinsic quality are significantly better than those of samples dried by other methods.

[0035] Multi-source data fusion is used to construct a drying mode discrimination model (SVM classification model). The electronic nose and electronic tongue datasets were fused at the feature layer, and three supervised classification and discrimination models were constructed: PLS-DA (Partial Least Squares-Discriminant Analysis), LDA (Linear Discriminant Analysis), and SVM (Support Vector Machine). The model performance validation results are shown in Table 4. As can be seen from Table 4, the SVM classification and discrimination model is the best model for discriminating the drying method.

[0036] Table 4. Performance Comparison of Different Classification and Discriminant Models (Five Groups of Samples)

[0037] Determination of Key Active Ingredient Content The active components, total flavonoids, and total phenolic acids of three core monomers (rutin, chlorogenic acid, and caffeic acid) in five groups of *Trifolium repens* samples were determined by HPLC. The specific steps are as follows: Chromatographic conditions: A C18 column (4.6 mm × 250 mm, 5 μm) was used. The mobile phase was acetonitrile (phase A) - 0.1% phosphoric acid aqueous solution (phase B). The gradient elution program was: 0-10 min, 5%-15% A; 10-25 min, 15%-25% A; 25-40 min, 25%-40% A; 40-45 min, 40%-5% A. The detection wavelength was 280 nm, the column temperature was 30 ℃, the flow rate was 1.0 mL / min, the injection volume was 10 μL, and the theoretical plate number calculated based on the target component peak was not less than 3000.

[0038] Preparation of reference solution: Accurately weigh appropriate amounts of chlorogenic acid, caffeic acid, and rutin reference standards, dissolve them in methanol and make up to volume to prepare a mixed reference stock solution. Dilute stepwise to obtain a series of mixed reference working solutions of different concentrations and store at 4°C.

[0039] Preparation of test solution: Accurately weigh 1.0 g of each of the five groups of *Tripterygium wilfordii* samples dried by different methods, place them in a stoppered conical flask, add 25 mL of 70% methanol aqueous solution, weigh, and then extract by ultrasonication for 30 min. After cooling to room temperature, add weight, shake well, and filter. Take the subsequent filtrate and filter it through a 0.22 μm organic phase filter membrane to obtain the test solution.

[0040] Content determination: The series of mixed reference working solutions and test solutions were injected sequentially into HPLC, the chromatograms were recorded, and the content of each target active ingredient was calculated using the external standard method. At the same time, the total flavonoid and total phenolic acid contents were calculated to obtain the active ingredient datasets of five groups of *Trifolium repens* samples with different drying methods for later use.

[0041] The test results are shown in Table 5. As can be seen from Table 5, the content of all active ingredients in the vacuum freeze-drying group (group 5) was significantly higher than that in the other four groups. P <0.05), with total flavonoid content reaching 38.62 mg / g and total phenolic acid content reaching 26.35 mg / g, which are 153.15% and 143.08% higher than the relatively lowest 55℃ medium-temperature hot air drying group, respectively. This indicates that the low temperature, vacuum, and oxygen-free environment of vacuum freeze drying can maximize the preservation of heat-sensitive flavonoid and phenolic acid active ingredients in *Trifolium repens*, and reduce their oxidation and degradation. The active ingredient contents of the other four groups of samples showed a gradient difference in random order.

[0042] Table 5. Content of key active ingredients in five groups of *Tripterygium wilfordii* samples dried using different methods (unit: mg / g, dry weight)

[0043] Comprehensive quality evaluation of Trifoliate orange under different drying methods Using the analytic hierarchy process (AHP) combined with the entropy weight method, a comprehensive quality scoring system for *Tripterygium wilfordii* was established with “active ingredient content (70% weight), electronic nose sensory characteristics (15% weight), and electronic tongue sensory characteristics (15% weight)” as the core evaluation dimensions (100 points in total). The comprehensive scores and quality grading results of the five groups of samples are shown in Table 6.

[0044] The results showed that the vacuum freeze-drying group had the highest comprehensive score of 95.62, ranking first and belonging to Grade I (Excellent). This was the best drying process for *Trifolium repens* among the five drying methods. The remaining four groups of samples were randomly ranked according to their comprehensive scores, and were divided into three grades: Grade II (Good), Grade III (Medium), and Grade IV (Poor), which were completely consistent with the differences in active ingredient content and sensory characteristics. The scoring criteria were set as follows: Grade I (Excellent) ≥ 90 points, Grade II (Good) 75-89 points, Grade III (Medium) 65-74 points, Grade IV (Poor) 55-64 points, and Grade V (Unqualified) < 55 points.

[0045] Table 6. Comprehensive scores and quality grading of Trifoliate Orange under five different drying methods

[0046] Application examples The drying method of the sample (number SY-2026-001) (dried finished product slices without a clear drying method label) submitted by the processing enterprise of Tripterygium wilfordii (Zhejiang Wuyangtang Pharmaceutical Co., Ltd.) was determined, and the overall quality grade of the sample was evaluated.

[0047] The sample was pulverized using a high-speed universal pulverizer, passed through a 50-mesh standard sieve, sealed, and stored at 4°C. The moisture content was determined according to the pharmacopoeia method, and was found to be 3.31±0.05%, meeting the "moisture ≤12.0%" requirement under the *Tripterygium wilfordii* section of the *Pharmacopoeia of the People's Republic of China*, and thus eligible for subsequent testing.

[0048] Using the aforementioned electronic nose and electronic tongue, and following the parameters set above, the sensory characteristics of the submitted sample are obtained as a "digital fingerprint".

[0049] 1) Electronic nose detection Pretreatment: Accurately weigh 2.0g of sample powder, place it in a 20mL headspace vial, seal it, and equilibrate in a 50℃ constant temperature water bath for 20min.

[0050] Instrument parameters: high-purity nitrogen carrier gas flow rate 200 mL / min, cleaning time 40 s, injection time 10 s, data acquisition time 90 s, and the average response value of the 30-80 s steady-state phase is taken as valid data.

[0051] The results of the electronic nose test on the submitted samples (response values ​​of 10-channel sensors) are shown in Table 7.

[0052] Table 7. Results of Electronic Nose Testing on Submitted Samples

[0053] 2) Electronic tongue detection Pretreatment: Accurately weigh 3.0g of sample powder, add ultrapure water at a material-to-liquid ratio of 1:30 (g / mL), extract by ultrasonication at 300W for 30min, centrifuge at 4000r / min for 8min using a centrifuge (CF15RN, HITACHI Corporation, Japan), and take the supernatant as the test solution.

[0054] Instrument parameters: Each sample is tested for 5 cycles, and the average response value of the last 3 cycles is taken as valid data.

[0055] The results of the electronic tongue test on the submitted samples (response values ​​of the 6-channel taste sensor) are shown in Table 8.

[0056] Table 8. Detection results of the electronic tongue on the submitted samples.

[0057] The electronic nose and electronic tongue detection data of the submitted samples were preprocessed using the Z-score standardization method, which is completely consistent with the standard model training set, to eliminate dimensional and systematic errors and obtain a standardized dataset that can be input into the model.

[0058] The preprocessed electronic sensory fusion dataset (6.25, -0.10) was input into the SVM classification and discrimination model described above, and the model output results are as follows: Classification: The submitted sample was classified as belonging to the vacuum freeze-drying group; The discrimination confidence level is 99.87%, which is far higher than the 95% confidence threshold, and the discrimination result is highly significant. Auxiliary verification: The submitted sample data (6.25, -0.10) was projected onto the PCA score chart of the electronic nose and electronic tongue, as shown below. Figure 5 As shown, the scatter plots of the submitted sample fall entirely within the 95% confidence ellipse of the vacuum freeze-drying group, with clear boundaries and no overlap with the other four drying methods, further verifying the accuracy of the classification results. Furthermore, the processing company of the submitted sample confirmed that the actual drying method was indeed vacuum freeze-drying, consistent with the above judgment. Based on the correspondence between drying method and quality grade (Table 6), the medicinal material quality of the submitted sample is determined to be Grade I (Excellent).

[0059] The evaluation method established in this invention can effectively achieve accurate classification and comprehensive quality evaluation of *Tripterygium wilfordii* under five different drying methods; the vacuum freeze-drying process can retain the sensory characteristics and core active ingredients of *Tripterygium wilfordii* to the greatest extent, and is the best quality processing technology among the five drying methods; the method of this invention is rapid in detection, objective in results, and comprehensive in evaluation, and can be widely applied to the screening of drying processes for *Tripterygium wilfordii* and the whole-chain quality control of medicinal materials.

[0060] Unless otherwise specified, all technologies mentioned above refer to existing technologies.

[0061] Based on the above-described preferred embodiments of the present invention, and through the foregoing description, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification.

Claims

1. A method for evaluating the quality of *Trifolium repens* based on electronic sensory technology, characterized in that, Includes the following steps: 1) Data acquisition: Olfactory and gustatory data of the *Tripterygium wilfordii* samples were collected using an electronic nose and electronic tongue, respectively; 2) Model establishment: The olfactory and gustatory data of the *Tripterygium wilfordii* samples were fused to establish a drying method discrimination model; 3) Sample testing: Input the olfactory and gustatory data of the sample to be tested into the drying method discrimination model, determine the drying method according to the prediction results of the discrimination model, and determine the quality grade according to the drying method.

2. The method for evaluating the quality of *Trifolium repens* based on electronic sensory technology according to claim 1, characterized in that, The drying methods are natural sun drying, cool and ventilated air drying, 55℃ medium-temperature hot air drying, 50℃ constant-temperature vacuum drying, and vacuum freeze drying. The correspondence between the drying methods and the quality grades is as follows: vacuum freeze drying is Grade I, 50℃ constant-temperature vacuum drying is Grade II, natural sun drying and cool and ventilated air drying are Grade III, and 55℃ medium-temperature hot air drying is Grade IV.

3. The method for evaluating the quality of *Trifolium repens* based on electronic sensory technology according to claim 2, characterized in that, Natural sun-drying involves spreading the slices of *Trifolium repens* evenly on a clean drying tray and placing them in the open sun to dry naturally. The trays are turned 2-3 times daily, sun-dried during the day and brought back in at night to prevent moisture buildup. The ambient temperature is 20℃-30℃, and the product is dried until the moisture content is ≤12.0%. The method of drying in a cool, ventilated, and shaded environment involves evenly spreading the slices of *Trifolium repens* on a clean tray and placing them in a cool, ventilated, and dark indoor environment with an ambient temperature of 20℃-25℃ and a relative humidity of ≤60%. The tray should be turned over twice daily until the moisture content is ≤12.0%. The 55℃ medium-temperature hot air drying method involves evenly spreading the *Trifolium repens* slices on a tray in a hot air drying oven, with a layer thickness of ≤2cm. The hot air temperature is 55℃, the air velocity is 1.0m / s, and the slices are turned over once every 3 hours until the moisture content is ≤12.0%. The 50℃ constant temperature vacuum drying process involves evenly spreading the sliced ​​*Trifolium repens* (a type of medicinal herb) on a tray in a vacuum drying oven, with a layer thickness ≤2cm, a vacuum degree ≥0.08MPa, a heating temperature of 50℃, and drying until the moisture content is ≤12.0%. Vacuum freeze drying involves first pre-freezing the *Trifolium repens* slices in a -40℃ pre-freezing chamber for 4 hours, then transferring them to a vacuum freeze dryer. The cold trap temperature is ≤-50℃, the vacuum degree is ≤10Pa, the partition heating temperature is 20℃, and the process involves sublimation drying for 24-36 hours, followed by desorption drying for 4 hours, until the moisture content is ≤12.0%.

4. The method for evaluating the quality of *Trifolium repens* based on electronic sensory technology according to claim 1, characterized in that, The detection method for the electronic nose is as follows: accurately weigh 2.0g of sample, equilibrate in a 50℃ constant temperature water bath for 20min, use a carrier gas flow rate of 200mL / min, clean for 40s, inject for 10s, collect data for 90s, and take the stable response value between 30-80s as valid data.

5. The method for evaluating the quality of *Trifolium repens* based on electronic sensory technology according to claim 1, characterized in that, The detection method for the electronic tongue is as follows: accurately weigh 3.0g of sample, with a material-to-liquid ratio of 1:30, extract by ultrasonication at 300W for 30min, centrifuge at 4000r / min for 8min, test each sample for 5 cycles, and take the average response value of the last 3 cycles as the valid data.

6. The method for evaluating the quality of *Trifolium repens* based on electronic sensory technology according to claim 1, characterized in that, The data fusion method is feature layer fusion, the unsupervised analysis adopts the PCA+HCA combination, and the discriminant model adopts the SVM model.

Images