Intelligent optimization method of wine rhizoma polygonati processing technology

Abstract

The invention discloses an intelligent optimization method of a wine rhizoma polygonati processing technology, and belongs to the field of traditional Chinese medicine processing. The processing technology of the wine rhizoma polygonati is optimized by utilizing a single factor test, a response surface method and a sparrow search method-neural network (SSA-BP). According to the method, six key components including polysaccharide, reducing sugar, flavone, saponin, 5-hydroxymethylfurfural content and ABTS + free radical scavenging rate in wine-processed polygonatum kingianum are taken as quality evaluation indexes, sensory characters of color, smell and taste are taken as auxiliary evaluation indexes, and a composite weighting strategy combining an entropy weight method and an analytic hierarchy process (AHP) is adopted to determine a weight coefficient of each index; and constructing a multi-dimensional quality comprehensive evaluation system, and taking the comprehensive score as a response index. And taking response surface data as neural network learning data, and determining an optimal process by combining an SSA-BP model with a response surface method: the steaming time is 3.6 hours, the drying temperature is 75 DEG C, the drying time is 11.4 hours, and the steaming cycle is 8 times. According to the method, the problem of parameter optimization of a traditional process is solved, and the product quality is remarkably improved.

Description

Technical Field

[0001] The invention belongs to the field of traditional Chinese medicine processing, and particularly relates to an intelligent optimization method for the processing technology of Polygonatum odoratum wine. Background Art

[0002] Polygonatum sibiricum is a perennial herb in the lily family. Modern pharmacological research has shown that it has anti-aging effects, lowers blood lipids, enhances immune function, regulates blood sugar, treats obesity, and alleviates inflammation. Consequently, it is widely used in food and medicine. However, since raw Polygonatum sibiricum can cause a numbing sensation on the tongue, it must be processed to eliminate this effect.

[0003] There are abundant records on the processing methods of Polygonatum odoratum. Currently, there are mainly steaming, steaming with wine, steaming with black bean juice and other processing techniques, among which nine steaming and nine drying is the most common. After nine steaming and nine drying, the medicinal properties change. Repeated steaming and drying promotes the transformation of medicinal properties, promotes the accumulation and activation of effective ingredients such as polysaccharides and saponins, thereby enhancing its effects of nourishing the spleen and moistening the lungs, and replenishing the kidneys and replenishing essence; at the same time, it effectively decomposes the irritating mucus in raw Polygonatum odoratum, and significantly reduces the irritation to the throat mucosa. Driven by its clinical value and market demand, most scholars have conducted a series of processing technology studies around the number of steaming and drying times of wine Polygonatum odoratum, the ratio of auxiliary materials, etc. The determination of the endpoint of traditional Polygonatum odoratum processing mainly relies on empirical sensory indicators such as dark color and sweet taste without numbness. There are significant subjective limitations, which can easily lead to problems such as poor or excessive processing. Zheng Xiaoqian mixed the rice wine thoroughly and kept a sample. The remaining material was steamed for 3 hours and dried at 60℃ for 9 cycles. The appearance, polysaccharides, 5-hydroxymaltol, 5-hydroxymethylfurfural (5-HMF), and polygonatine A were used to evaluate the final processing (Zheng Xiaoqian, Xu Chao, Jin Chuanshan, et al. Study on the correlation between the processing temperature and internal and external quality of "nine steaming and nine drying" of polygonatum based on color change [J]. Chinese Herbal Medicine, 2022, 53(06): 1719-1729.). Zhang Jie et al. used the appearance and polygonatum polysaccharides as indicators and used the orthogonal test method to optimize the optimal pressure wine steaming process of Jiuhua Mountain Polygonatum: 100g of polygonatum, 10% rice wine, steamed at 120℃ for 60min (Zhang Jie, Jin Chuanshan, Wu Deling, et al. Study on the optimization of pressure wine steaming process of polygonatum by orthogonal test method [J]. Journal of Anhui University of Traditional Chinese Medicine, 2014, 33(1): 72-74.). Current research is mostly based on the polysaccharide and extract content specified in the "Chinese Pharmacopoeia" as process optimization indicators, with a small amount of research on the characteristic components of the processing process, such as steroidal saponins, polyphenols, flavonoids, unsaturated fatty acids and 5-HMF, and subjective evaluation accounts for a large proportion.

[0004] During the heat processing stage, Polygonatum sibiricum promotes protein denaturation and caramelization, and the Maillard reaction may produce 5-HMF, posing a potential risk to human health. China's "National Standard for Honey" (GHT18796-2012) stipulates that the 5-HMF content must not exceed 40mg / kg. Although corresponding standards have been established for 5-HMF in the food and baking industries, it is difficult to cope with the nonlinear cumulative effects of hazards in the complex processing scenarios of traditional Chinese medicine. Therefore, it is urgent to design an intelligent optimization method for the preparation of Polygonatum sibiricum in wine to achieve a process that "increases efficacy and reduces toxicity."

[0005] The field of wine-cured Polygonatum processing technology optimization currently relies mainly on single-factor methods, orthogonal experimental methods, and response surface methodology (RSM). The single-factor method analyzes process parameters by isolating variables, but cannot characterize the interaction effects of multiple factors; although the orthogonal experimental method can partially solve the multivariate problem, its uniform dispersion design easily leads to an exponential increase in the experimental scale during refined optimization; RSM constructs a process response model based on a second-order polynomial, but its ability to fit high-dimensional and strongly nonlinear relationships is insufficient, making it difficult to meet the analytical requirements for the multi-parameter coupling mechanism in modern processing technology. In order to overcome the limitations of the above methods, existing technologies have attempted to combine genetic algorithms with response surface methodology to optimize the wine-cured white peony root process (CN119049583 A) and the honeysuckle rapid drying process (CN118940611A). Although it has certain global optimization potential, it still has common problems such as premature convergence and high parameter sensitivity, and has failed to effectively resolve the contradiction between high-dimensional nonlinear modeling and dynamic process constraints. Currently, there is an urgent need for an innovative method that can take into account low-dimensional and high-dimensional nonlinear modeling, dynamic constraint processing and optimization efficiency improvement, so as to break through the technical bottlenecks in traditional process optimization and provide theoretical support for the digital upgrade of Chinese medicine preparation technology.

[0006] To address these issues, the present invention, based on existing technologies, provides an intelligent optimization method for the processing of Polygonatum odoratum (Swine-Fermented Polygonatum). This method effectively avoids premature convergence and obtains optimal processing parameters through single-factor initial screening, the use of a subjective-objective composite weighting method, and the combination of a response surface methodology and an SSA-BP neural network. These process parameters maximize the retention of the active ingredients of Polygonatum odoratum and minimize the formation of 5-HMF, while also ensuring that the appearance and taste of the product meet the requirements of the Chinese Pharmacopoeia. Summary of the Invention

[0007] The present invention provides an intelligent optimization method for the processing of Polygonatum glutinosa L. (Shenghuangjing) wine. This method is characterized by the deep integration of traditional processing techniques and intelligent algorithms to achieve a transition from fuzzy to precise process parameters and from extensive to detailed quality control. This method effectively overcomes the limitations of the response surface methodology in high-dimensional, nonlinear, dynamic constraints, and complex interactive scenarios, while also resolving the long-standing reliance on manual experience in the processing of traditional Chinese medicine. The specific experimental plan is as follows:

[0008] (1) Select raw materials of Polygonatum sibiricum, clean them, remove impurities and roots, wash them with clean water, steam them for 40 min to soften them, cut them into 0.5 ± 0.1 cm thick homogeneous slices, and dry them at a constant temperature of 60 ° C for 12 h to prepare Polygonatum sibiricum slices;

[0009] (2) Take the slices of Polygonatum sibiricum and add yellow rice wine in a ratio of 100g:20g. Add appropriate amount of water to the yellow rice wine so that the Polygonatum sibiricum can be completely stirred evenly. The wine-water ratio is dynamically adjusted to 1:2n+3 (steaming times n≤9). The ratio of Polygonatum sibiricum to water is 1:1 (g:g) for the first steaming. The remaining water is added in batches for the subsequent steaming. The total amount of yellow rice wine is added evenly according to the number of steaming times and stirred evenly. The first round of steaming time is 2.5±0.5h, and the subsequent steaming time is shortened to 1±0.2h;

[0010] (3) Through single factor experiments, it was determined that the steaming time was 3 h, the drying temperature was 70 °C, the drying time was 11 h, and 8 cycles of steaming were completed;

[0011] (4) The Box-Behnken response surface method was used to design a four-factor three-level experiment. Steaming time (A), drying temperature (B), drying time (C), and steaming times (D) were used as independent variables. A comprehensive evaluation model based on entropy weight-hierarchy analysis composite empowerment was established. The theoretical optimal parameter combination was obtained: steaming time 3.16 h, drying temperature 74.18 °C, drying time 11.02 h, and steaming times 8.27 times, with a comprehensive score of 68.19 points.

[0012] (5) An SSA-BP neural network model was constructed, with 4 nodes in the input layer corresponding to process parameters, 8 nodes in the hidden layer, and 1 node in the output layer corresponding to the comprehensive score. After 50 iterations of sparrow search algorithm optimization, the global optimal parameters were determined: 3.58 h, drying temperature of 75.29 °C, drying time of 11.42 h, steaming times of 8.24 times, and the comprehensive score increased to 72.75 points;

[0013] (6) Based on the response surface and neural network optimization results and the operational feasibility, the final process parameters were determined as steaming time 3.6 h, drying temperature 75 °C, drying time 11.4 h, and steaming times 8 times;

[0014] (7) The contents of the active ingredient polysaccharide, reducing sugar, flavonoids, saponin and 5-HMF in the Huangjing slices prepared by the optimized process were 480.8 mg / g, 395.7 mg / g, 7.0 mg / g, 28.9 mg / g and 0.21±0.03 mg / g, respectively; ABTS + IC of free radical scavenging rate 50 The value is 2.3mg / mL, and the properties of the finished product meet the standards of the 2020 edition of the "Chinese Pharmacopoeia": black horny, slightly burnt aroma, slightly sweet taste and sticky.

[0015] Compared with the prior art, the present invention has the following beneficial effects:

[0016] 1. Multidimensional quality evaluation system: By integrating the entropy weight method and the hierarchical analysis method, a multidimensional evaluation system with subjective and objective linkage is constructed, and the weight coupling algorithm w=0.7w is adopted. 熵权 +0.3w 层次分析 , breaking through the limitations of traditional technology that relies on a single indicator or subjective experience, and achieving 6 core evaluation indicators: polysaccharides, reducing sugars, flavonoids, saponins, 5-HMF and ABTS + The scientific empowerment of free radical scavenging rate constructs a multidimensional quality evaluation system.

[0017] 2. The alcohol-water ratio is dynamically increased at a ratio of 1:2n+3 (steaming times n≤9), so that the amount of alcohol-water increases step by step with the process progress, which not only ensures sufficient penetration in the initial stage, but also meets the reaction intensity requirements in the final stage, forming a progressive moistening and control mechanism. The interaction between Polygonatum sibiricum and alcohol-water during the processing process can be more accurately regulated. The polysaccharides, flavonoids, saponins and other active ingredients contained in Polygonatum sibiricum itself are retained to the greatest extent, avoiding the loss or destruction of these active ingredients during the processing process, thereby ensuring the medicinal value of the final product of Polygonatum sibiricum. It also plays a key role in improving the taste of Polygonatum sibiricum, effectively neutralizing the dryness, bitterness and other unpleasant tastes that may exist in Polygonatum sibiricum, making it taste more mellow and sweet, greatly improving the palatability of the processed Polygonatum sibiricum product.

[0018] 3. Based on the experimental data set of response surface methodology, the present invention constructs an SSA-BP neural network prediction model, and deeply optimizes the process parameters through the global optimization capability of the sparrow search algorithm. Compared with the response surface methodology, the SSA-BP model significantly improves the quality of Polygonatum odoratum, among which the polysaccharides, flavonoids, and saponins are increased by 9.1%, 7.7%, and 8.9% respectively compared with the response surface methodology; ABTS + The free radical scavenging rate increased from 75.66% to 78.04%, an increase of 3.14% (p<0.05); the 5-HMF content increased from 0.16 mg / g to 0.21 mg / g, with no significant difference (p=0.102>0.05), indicating that the process can effectively inhibit the formation of Maillard reaction by-products; the comprehensive score increased from 68.32 points to 70.65 points, a relative increase of 3.41%, verifying the advantages of intelligent algorithms in high-dimensional nonlinear relationship modeling. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 ABTS free radical scavenging rate of raw Polygonatum sibiricum and its optimal process and VC;

[0020] Figure 2 Normal probability distribution plot of residuals;

[0021] Figure 3Scatter plot of residuals and predicted values;

[0022] Figure 4 Scatter plot of predicted and actual values;

[0023] Figure 5 is the response surface of the effects of steaming time and drying time on the comprehensive score;

[0024] Figure 6 is the response surface of the effects of steaming times and drying time on the comprehensive score;

[0025] Figure 7 is the response surface of the effects of drying temperature and drying time on the comprehensive score;

[0026] Figure 8 SSA-BP optimal performance verification;

[0027] Figure 9 Training regression coefficients;

[0028] Figure 10 Comparison chart of predicted output and actual output of BP neural network test set;

[0029] Figure 11 SSA convergence curve; DETAILED DESCRIPTION

[0030] Example 1 Preparation process of Polygonatum odoratum slices

[0031] Take raw Polygonatum sibiricum, wash it, remove the roots, steam it for 40 minutes, cut it into 0.4-0.6 cm thick slices, and dry it at 60℃ for 12 hours to prepare Polygonatum sibiricum slices.

[0032] Example 2 Extraction and Content Determination Methods of Polysaccharides, Reducing Sugars, Flavonoids, and Saponins

[0033] Accurately weigh 3 g of sample powder, put it into a 100 mL round-bottom flask, add 45 mL of water, place it in a 75 ° C water bath, reflux extraction for 2 h, extract twice, filter, and use the filtrate as the test solution. The volume is adjusted to 100 mL and the polysaccharide content is determined by the phenol-sulfuric acid method.

[0034] Accurately weigh 3 g of sample powder, put it into a 100 mL round-bottom flask, add 50 mL of water, place it in a 50°C water bath, reflux and extract for 2 h, extract twice, filter, and use the filtrate as the test solution. Dilute to 100 mL and use the 3,5-dinitrosalicylic acid method to determine the reducing sugar content.

[0035] Accurately weigh 5 g of sample powder, place it in a 150 mL round-bottom flask, add 100 mL of 50% ethanol, place it in a 60 ° C water bath, reflux extraction for 1.5 h, filter, and use the filtrate as the test solution. The volume is adjusted to 100 mL, and the flavonoid content is determined by the reaction of sodium nitrite-aluminum nitrate and sodium hydroxide.

[0036] Accurately weigh 5 g of sample powder, place it in a 150 mL round-bottom flask, add 100 mL of 80% ethanol, place it in a 55°C water bath, and ultrasonically extract for 50 min. Extract twice, filter, and use the filtrate as the test solution. Dilute to 100 mL and determine the saponin content by reacting vanillin-glacial acetic acid with perchloric acid.

[0037] The contents of polysaccharide, reducing sugar, flavonoids and saponin in raw Polygonatum sibiricum are 340.0 mg / g, 32.8 mg / g, 0.3 mg / g and 12.1 mg / g, respectively.

[0038] Example 35 - HMF extraction and content determination method

[0039] Accurately weigh 0.1 g of sample powder and place it in a 100 mL round-bottom flask. Add 25 mL of methanol and ultrasonically extract for 30 minutes. Filter and dilute to a 25 mL volumetric flask with methanol to obtain the test solution. 5-HMF content was determined using HPLC on a Phenomenex Luna C18 (250 mm × 4.6 mm, 5 μm) column with a mobile phase of 0.1% formic acid in water and methanol (80:20) at a flow rate of 1.0 mL / min, a detection wavelength of 284 nm, a column temperature of 30°C, and an injection volume of 20 μl. No 5-HMF was observed in the raw sample.

[0040] Example 4 ABTS + Determination of free radical scavenging rate

[0041] Mix 2.0 mL of the test solution with 9 times the volume of ABTS + The solution was mixed and reacted in the dark for 5 minutes. Then, the absorbance A1 was measured at a wavelength of 734 nm, with water as the blank group. In the control group, 2 mL of distilled water was accurately pipetted into a test tube and 9 times the volume of ABTS was added. + The solution was shaken for 10 seconds, kept in the dark for 5 minutes, and the absorbance A2 was measured at 734 nm to calculate the clearance rate. VC was used as a positive control.

[0042]

[0043] Test solution for antioxidant experiment: Accurately weigh 4.0g of sample powder, add 60mL of water to a 100mL round-bottom flask, place in a 75℃ water bath, heat reflux extraction for 2h, extract twice, filter, and the filtrate is the test solution, and the volume is made up to 100mL. On this basis, for single-factor antioxidant experiment, take 2.5mL and make up to 10mL, that is, dilute 4 times for testing. For other antioxidants (response surface and comparative literature 1 and 2), take 1.25mL of test solution and make up to 10mL, that is, dilute 8 times for testing. The test solution concentration formula is:

[0044]

[0045] Prepare five test solutions of raw Polygonatum sibiricum antioxidant test with concentrations ranging from 6.25 mg / mL to 50 mg / mL, and obtain the clearance rate as follows: Figure 1 As shown, IC 50 The value is 29.4mg / mL. The IC value of VC 50 The value is 0.02mg / mL.

[0046] Example 5 Scoring criteria for wine-scented Polygonatum odoratum slices

[0047] In order to more fully meet the requirements of the 2020 edition of the Chinese Pharmacopoeia for the properties of Huangjing (Polygonatum odoriferum) as "black as lacquer, slightly burnt aroma, and slightly sweet taste", color, texture, and smell were used as auxiliary evaluation indicators. On the basis of the above three evaluation indicators, the content of polysaccharides, reducing sugars, flavonoids, saponins, 5-HMF and ABTS were also examined. + Six indicators of antioxidant activity were used to construct a multi-dimensional quality control system. The specific scoring criteria are shown in Table 1.

[0048] Table 1 Rating criteria for wine-brewed Polygonatum slices

[0049]

[0050] Example 6 Dynamic Gradient Wine Addition Process in the Preparation of Polygonatum Wine

[0051] Take the polygonatum slices and add yellow rice wine at a ratio of 100g:20g. Add appropriate amount of water to the yellow rice wine so that the polygonatum can be completely stirred evenly. The wine-water ratio is 1:2n+3 (g:g) and dynamically adjusted (steaming times n≤9). The first steaming uses a ratio of polygonatum to water of 1:1 (g:g). The remaining water is added in batches in subsequent steamings. The total amount of yellow rice wine is added evenly according to the number of steamings and stirred evenly. The first round of stewing time is 2.5±0.5h, and the subsequent stewing time is shortened to 1±0.2h;

[0052] Example 7 Intelligent Optimization Method for the Processing Process of Polygonatum odoratum

[0053] (1) Single-factor experiment

[0054] In order to determine the experimental parameters of response surface, the contents of polysaccharides, reducing sugars, flavonoids, saponins, 5-HMF and ABTS were used. + Free Radical Scavenging Rate A single-factor experiment was conducted on steaming time, drying temperature, drying time, and steaming times. The ABTS free radical scavenging rate was determined in the same manner as in Example 4, with the test solutions diluted 4 times.

[0055] 1. Steaming time

[0056] Weigh 5 portions of 50g Polygonatum slices, add 10g rice wine and 50g water, stir evenly, seal and moisten for 2h. Fix the drying temperature at 60℃ and the drying time for 12h, steam once, and change the steaming time to conduct a single factor experiment, steaming for 1, 2, 3, 4, and 5h respectively, and drying at 60℃ for 12h. The steamed sample was crushed and sieved. Since the steaming time has a great influence on the chemical composition, it is necessary to measure the content of polysaccharides, reducing sugars, flavonoids, saponins, 5-HMF and ABTS. + The results of the free radical scavenging rate determination experiment are shown in Table 2.

[0057] Table 2 Results of single factor experiment on steaming time

[0058]

[0059] 2. Drying temperature

[0060] Weigh 5 portions of 50g of Polygonatum sibiricum slices, add 10g of rice wine and 50g of water, let it sit for 2h, steam it for 2h, steam it once, and dry it at 40, 50, 60, 70, and 80℃ for 12h. Determine the content of polysaccharides, reducing sugars, flavonoids, saponins, 5-HMF and ABTS + The free radical scavenging rate was determined by the same method as above. The experimental results are shown in Table 3.

[0061] Table 3 Results of single factor experiment on drying temperature

[0062]

[0063] 3. Drying time

[0064] Weigh 5 portions of 50g Polygonatum sibiricum slices, add 10g yellow rice wine and 50g water, let it sit for 2h, steam it for 2h, dry it at 60°C for 10, 11, 12, 13, and 14h, and steam it once. The determination method is the same as above, and the results are shown in Table 4.

[0065] Table 4 Results of single factor experiment on drying time

[0066]

[0067] 4. Number of steaming times

[0068] Steaming was performed for a fixed time of 2 hours, a drying temperature of 60°C, and a drying time of 12 hours, with steaming cycles of 1-9. The amount of rice wine added for each soaking was 1:2n+3 (steaming cycles n≤9). For the first steaming, the ratio of Huangjing to water was 1:1 (g:g). The remaining water was added in batches, with the amount of rice wine distributed evenly according to the number of steaming cycles. Stirring was carried out evenly. The first soaking was performed for 2-3 hours, and the remaining steaming cycles were performed for 1 hour. The measurement method was the same as above, and the results are shown in Table 5.

[0069] Table 5 Results of single factor test on steaming times

[0070]

[0071] (2) Composite weighting method

[0072] In order to comprehensively and fully evaluate the quality of Polygonatum sibiricum, the present invention uses the hierarchical analysis method combined with the entropy weight method to determine the content of polysaccharides, reducing sugars, flavonoids, saponins, 5-HMF, ABTS + The free radical scavenging rate weight is used to calculate the comprehensive score value, including the following steps:

[0073] 1. Calculate polysaccharides, reducing sugars, flavonoids, saponins, 5-HMF, and ABTS by entropy weight method (EWM) + The six indicators of free radical scavenging rate were objectively weighted. First, the original data matrix was constructed. Since 5-HMF production is related to the content of reducing sugars and 5-HMF has potential safety hazards, reducing sugars and 5-HMF were used as negative indicators, polysaccharides, flavonoids, saponins and ABTS were used as negative indicators. + The free radical scavenging rate was calculated as a positive indicator.

[0074] The weight values ​​obtained by entropy weight method are: polysaccharides, reducing sugars, flavonoids, saponins, 5-HMF, ABTS + The weight values ​​of each indicator of free radical scavenging rate are: 11.09%, 9.23%, 38.14%, 7.84%, 4.73% and 28.97% respectively.

[0075] 2. Using the analytic hierarchy process (AHP) to analyze the polysaccharides, reducing sugars, flavonoids, saponins, 5-HMF, ABTS + The six indicators of free radical scavenging rate were subjectively weighted. Combined with the experimental content, the priority of the six indicators was preliminarily determined to be ABTS. + Free radical scavenging rate>flavonoids>saponins>polysaccharides>reducing sugars>5-HMF, thereby constructing a priority judgment matrix for pairwise comparison of targets and assigning relative scores to each indicator.

[0076] Table 6 Indicator layer comparison and judgment optimization matrix

[0077]

[0078] The AHC analysis software was used to calculate the content of polysaccharides, reducing sugars, flavonoids, saponins, 5-HMF, and ABTS. + The weight values ​​of the free radical scavenging rate indicators are: 13.51%, 8.85%, 20.4%, 16.6%, 5.94%, and 34.70%. The experimental consistency test result CR = 0.078 < 0.1, indicating that the matrix consistency result passed and the weight calculation was correct and reasonable.

[0079] 3. Determination of the combined weight of AHP-EWM: w j =αw′ j +βw″ j , where α+β=1, 0.6≤α≤0.8, in order to effectively reduce the influence of subjective factors on the evaluation results. For better objective evaluation, the weight coefficient α of the entropy weight method is set to 0.7, and the weight coefficient β of the hierarchical analysis method is set to 0.3. Polysaccharides, reducing sugars, flavonoids, saponins, 5-HMF, ABTS + The weight values ​​of the free radical scavenging rate indicators were 11.81%, 9.12%, 32.82%, 10.47%, 5.09%, and 30.69%, respectively. 100 points were allocated according to the weight coefficient ratio to determine the comprehensive score Y = (polysaccharide / maximum polysaccharide × 11.81 + flavonoid / maximum flavonoid × 32.82 + saponin / maximum saponin)

[0080] ×10.47+ABTS + Free radical scavenging rate / ABTS + Maximum free radical scavenging rate × 30.69 - reducing sugar / maximum reducing sugar × 9.12 - 5-HMF / maximum 5-HMF × 5.09).

[0081] (III) Single factor results

[0082] According to the above single factor experimental results and the final weight score, the comprehensive score of the single factor experiment was calculated, and the results are shown in Table 7. The optimal process was screened out as steaming time 3h, drying temperature 70℃, drying time 11h, and completing 8 cycles of steaming.

[0083] Table 7 Comprehensive score of single factor experiment

[0084]

[0085] (IV) Response surface experiment

[0086] 1. Response surface experimental design and results

[0087] The single factor combined with response surface experiment method was used to evaluate and rank each sample according to the comprehensive score. The more ideal processing parameters in the above single factor experiment were selected, and 29 groups of response surface analysis experiments with four factors and three levels were designed. The response surface experiment used steaming time (A), steaming times (B), drying temperature (C), and drying time (D) as independent variables, and the polysaccharide, reducing sugar, flavonoids, saponin, 5-HMF content and ABTS + The comprehensive score of free radical scavenging rate was used as the response value. The Box-Behnken experimental design in the response surface software Design-Expert was used to perform a central composite experimental design. The resulting four-factor three-level response surface analysis experimental plan and results are shown in Table 8.

[0088] Table 8 Response surface analysis experimental results

[0089]

[0090]

[0091] The binary regression equation of the predicted value of the comprehensive score on the A, B, C, and D coding values ​​is: Y = 65.59 + 5.60 × A + 5.53 × B + 5.29 × C + 0.61 × D - 2.91 × AB - 0.35 × AC + 0.14 × AD - 3.91 × BC - 1.66 × BD - 0.125 × CD - 8.04 × A 2 -6.54×B 2 -4.17×C 2 -7.16×D 2

[0092] The variance analysis is shown in Table 9. According to Table 9, the P value of the regression equation model is extremely significant, indicating that the model has a good fit and the experimental method is feasible; the P value of the lack of fit term is not significant, indicating that the model error is small; the larger the F value of the factor and the smaller the P value, the greater the impact of the factor on the result.

[0093] Table 9 Variance results of response surface experiment

[0094]

[0095]

[0096] 2. Response surface plot analysis

[0097] Figure 2 is the normal probability distribution diagram of the residual in this experiment, Figure 3 is a scatter plot of residuals and predicted values, Figure 4 is a scatter plot of predicted and actual values, Figure 2-4 It can be seen that the linear relationship of the data distribution is good. The response surface experiment surface diagram and contour lines are as follows Figure 5-7 As shown, it can be seen that the curves of interaction terms AD, BD, and CD are steeper and have a greater impact on the results.

[0098] Through single factor and response surface analysis, the optimal conditions were obtained: steaming time 3.16h, drying temperature 74.18℃, drying time 11.02h, and steaming times 8.27 times. Based on operational feasibility, steaming time 3.2h, drying temperature 74℃, drying time 11.0h, and steaming times 8 times were selected.

[0099] (5) SSA-BP neural network training

[0100] 80% of the response surface data was used as the training set and 20% of the response surface data was used as the test set. A three-layer BP neural network model was established using MATLAB R2024a software to explore the relationship between input and output. These data were processed by the hidden layer tansig activation function and the output layer activation function. The input layer contains four neurons for steaming time, drying temperature, drying time, and number of steaming times, while the output layer has one neuron (Y) to predict the comprehensive score of the processing process. At the same time, the search range of the number of hidden layer nodes was set to 3-15 through a dynamic search mechanism. The MSE (mean square error) corresponding to all candidate node numbers was compared, and the optimal number of hidden layer nodes was finally determined to be 8, with an MSE of 0.02182. The training times were 1000, the network training rate was 0.01, and the training target minimum error was 0.0001. Figure 8 The performance of the training, validation, and test datasets under different training rounds is shown. The MSE of the yield reaches the optimal value in round 35. The average relative error of the predictions of the five groups of data is 1.13%, and the training regression coefficient R is greater than 0.9. Figure 9 This shows that the model has good prediction accuracy. To verify the accuracy of the prediction model, the prediction results are compared with the actual data ( Figure 10 ), indicating that the constructed model has stable prediction performance and can be used to optimize the processing parameters of Huangjing wine preparation.

[0101] In order to improve the generalization ability of the model, SSA is added for optimization. Initialize the BP neural network model established above. Through continuous trial and comparison, the maximum number of iterations of SSA is 50, the population size is 30, the safety value is 0.8, the proportion of discoverers is 0.2, and the proportion of sparrows that realize there is danger is 0.2. The curve of fitness changing with the number of iterations is obtained ( Figure 11 ). The fitness curve shows that as the number of iterations increases, the curve gradually tends to a straight line, at which point the population has reached the optimal state. When the iteration reaches 5 times, the comprehensive score reaches 72.54 points. When the iteration reaches 23 times, it reaches the maximum value, and the population has screened out the optimal individual. The best optimization result is a steaming time of 3.58h, a drying temperature of 75℃, a drying time of 11.42h, 8.24 steaming times, and a comprehensive score of 72.75 points. Based on operational feasibility, the optimal process selection is a steaming time of 3.6h, a drying temperature of 75℃, a drying time of 11.4h, and 8 steaming times.

[0102] (V) Verification experiment

[0103] According to the test results of the response surface methodology and the experimental process of SSA-BP, three groups of tests were conducted in parallel to calculate the comprehensive score. The properties of the finished products of the two methods strictly comply with the provisions of the 2020 edition of the "Chinese Pharmacopoeia". The appearance is uniformly dark and keratinous, the surface gloss is significantly improved, the gas is slightly burnt, the taste is sweet, and the numbness of the tongue is eliminated. The specific scores are shown in Table 10, among which the average comprehensive sensory score of the response surface is 84 points, and the average comprehensive sensory score of the SSA-BP test is 91 points (p < 0.05). The average comprehensive score of the response surface methodology test is 68.32 points, with an RSD value of 0.96%, and the average test score of SSA-BP is 70.65 points, with an RSD value of 0.85%, which verifies that the process is stable and feasible. The comprehensive score of the experimental process of SSA-BP is higher than that of the response surface methodology test (p < 0.05). Therefore, the final process was determined to be a steaming time of 3.6h, a drying temperature of 75℃, a drying time of 11.4h, and 8 steaming times. Under this process, the contents of polysaccharide, reducing sugar, flavonoids, saponin and 5-HMF were 480.8 mg / g, 395.7 mg / g, 7.0 mg / g, 28.9 mg / g and 0.21±0.03 mg / g respectively; the IC value of the optimal process for the antioxidant experiment of Polygonatum odoratum with a concentration of 0.31-5.0 mg / mL was 50 The value is 2.3mg / mL. Figure 1 shown.

[0104] Table 10 Sensory evaluation results of Huangjing (Polygonatum odoratum)

[0105]

[0106] Comparative test

[0107] Patent application method: Take polygonatum slices and add yellow rice wine at a ratio of 100g:20g. Add an appropriate amount of water to the yellow rice wine to ensure that the polygonatum is completely stirred. The wine-water ratio is dynamically adjusted to 1:2n+3 (steaming times n≤9). The first steaming uses a polygonatum to water ratio of 1:1 (g:g). The remaining water is added in batches for subsequent steaming. The total amount of yellow rice wine is added evenly according to the number of steaming times and stirred evenly. The first steaming time is 2.5±0.5h, and the subsequent steaming time is shortened to 1±0.2h. The steaming time is 3.6h, the drying temperature is 75℃, and the drying time is 11.4h, completing 8 steaming cycles.

[0108] Literature comparison 1 Method: For every 100kg of Polygonatum, use 20kg of yellow rice wine and steam until the wine is completely absorbed by the medicinal material. A small amount of Polygonatum sample soaked in yellow rice wine is retained, which is the sample S1 with 0 steaming and drying times. The remaining Polygonatum samples are steamed in a steamer for 3 hours, taken out, placed in a drying oven, and dried at a constant temperature of 60℃ until they are no longer sticky. Repeat this steaming for a total of 5 times (Zheng Xiaoqian, Xu Chao, Jin Chuanshan, et al. Study on the correlation between the processing temperature of Polygonatum and its internal and external quality based on the "nine steaming and nine drying" method based on color change [J]. Chinese Herbal Medicine, 2022, 53(06): 1719-1729.).

[0109] Literature comparison 2 method

[0110] Weigh 1 kg of fresh Polygonatum sibiricum. Place all ingredients in a 60°C oven until the skin is slightly dry. Take a portion as the raw Polygonatum sibiricum. Place the remaining ingredients in a steamer and steam with water for 3 hours after the water boils. Turn off the heat and simmer overnight. Dry in an oven until the skin is slightly dry. This constitutes one steaming and drying process. Each time, keep a portion of the ingredients and repeat the steaming and drying process with the remaining ingredients until the Polygonatum sibiricum preparation has been completed five times (Ding Rui. Study on the Material Basis and Mechanism of Nine-Steaming and Nine-Drying of Polygonatum sibiricum for Enhanced Anti-ulcerative Colitis Effects [D]. Zhejiang University of Traditional Chinese Medicine, 2024).

[0111] The contents of polysaccharides, reducing sugars, flavonoids, saponins, 5-HMF and ABTS were compared between the two methods of patent application response surface and SSA-BP neural network and the methods described in references 1 and 2. + The free radical scavenging rates were compared and analyzed, and the results are shown in Table 11. The data showed that the four had the following contents: total polysaccharides, saponins, reducing sugars, flavonoids and ABTS + Core indicators such as free radical scavenging rate showed similar levels. However, after the optimization of key process control parameters, the 5-HMF production in this patented method was significantly reduced to 0.16-0.21 mg / g, compared with 0.75 mg / g in Reference 1 and 0.61 mg / g in Reference 2, and the comprehensive score was also better than that of References 1 and 2 (p < 0.05). The comprehensive score of the process after SSA-BP neural network optimization was 2.58% higher than that of the process after response surface optimization (p < 0.05). While ensuring the efficient extraction of active ingredients, this patented method effectively inhibits the formation of Maillard reaction by-products through process innovation, and the overall technical solution has significant advantages.

[0112] Table 11 Comparative experimental results

[0113]

[0114] Note: * indicates p<0.05 (significant), ** indicates p<0.01 (very significant), *** indicates p<0.001 (extremely significant)

[0115] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and the concept of the invention, should be covered by the scope of protection of the present invention.

Claims

1. An intelligent optimization method for the processing technology of Polygonatum glutinosa, characterized in that: The following steps are involved: Take the slices of Polygonatum sibiricum, add a certain amount of yellow wine and water, stir evenly, and steam. The key process parameters affecting the quality of the preparation of Polygonatum sibiricum were determined by single factor experiment. The key process parameters included steaming time, drying temperature, drying time and steaming times. The Box-Behnken response surface methodology was used to design a multi-factor interaction experimental scheme. The contents of polysaccharides, reducing sugars, flavonoids, saponins, 5-hydroxymethylfurfural and ABTS in Polygonatum sibiricum were analyzed. + The free radical scavenging rate is a multidimensional quality evaluation indicator. The weight coefficient of each indicator is determined through a composite weighting strategy to construct a comprehensive scoring response value system. The response surface experimental data were used as training sets, and a BP neural network prediction model based on the sparrow search algorithm (SSA) optimization was established. The nonlinear mapping model was performed with the process parameters as input variables and the comprehensive score as output variable. Through the synergistic optimization of SSA-BP model and response surface methodology, this process maximizes the retention of the effective ingredients of Polygonatum sibiricum and reduces the generation of 5-HMF.

2. The process for preparing the Huangjing (Shenghuangjing) wine extract according to claim 1, wherein: Add 20kg of rice wine to every 100kg of polygonatum, add appropriate amount of water to the rice wine so that the polygonatum can be completely stirred evenly, the wine-water ratio is 1:2n+3 (steaming times n≤9), and the ratio of polygonatum to water for the first steaming is 1g:1g. Add the remaining water in batches, and distribute the amount of rice wine evenly according to the number of steaming times. Stir evenly, and stew for 2-3 hours for the first steaming, and for 1±0.2 hours for the remaining steaming times.

3. The process for preparing the Huangjing (Shenghuangjing) of claim 1, wherein: The single factor experimental study was conducted to study the effects of steaming time, drying temperature, drying time and steaming times on the content of polysaccharides, reducing sugars, flavonoids, saponins, 5-hydroxymethylfurfural and ABTS in Rhizoma Polygonati. + The free radical scavenging rate was used as the quality evaluation index, and the composite weight method was used for comprehensive evaluation; color, smell, and taste sensory properties were used as auxiliary evaluation indicators, and the optimal conditions were preliminarily screened out based on the comprehensive score: steaming time 3h, drying temperature 70℃, drying time 11h, and steaming times 8 times.

4. The composite weighting strategy method according to claim 3, characterized in that: The subjective and objective composite weighting strategy combining entropy weight method and analytic hierarchy process was used to determine the content of polysaccharides, reducing sugars, flavonoids, saponins, 5-hydroxymethylfurfural and ABTS. + The weight of the free radical scavenging rate, where the two are fused by weighted average method, the weight fusion formula is: j =αw′ j +βw″ j , where α+β=1, 0.6≤α≤0.8, α is the weight coefficient of entropy weight method, which can effectively reduce the influence of subjective factors on the evaluation results, and determine the polysaccharides, reducing sugars, flavonoids, saponins, 5-hydroxymethylfurfural, ABTS + The weight values ​​of each indicator of free radical scavenging rate are: 11.81%, 9.12%, 32.82%, 10.47%, 5.09% and 30.69% respectively.

5. The process for preparing the Polygonatum glutinosa according to claim 3, wherein: The response surface methodology was used to conduct a four-factor three-level experimental design. 80% of the response surface experimental data were randomly selected as the training set of the BP neural network, and 20% as the test set. The process parameters steaming time, drying temperature, drying time and steaming times were used as input variables; the comprehensive score was used as the output variable to construct a BP neural network prediction model.

6. The method according to claim 5, characterized in that The BP neural network model structure includes four input layer nodes, namely steaming time, drying temperature, drying time and steaming times; one output layer node, namely a comprehensive score; a dynamic search mechanism is used to set the search range of the number of hidden layer nodes to 3-15, and by comparing the mean square errors corresponding to all candidate node numbers, the optimal number of hidden layer nodes is determined to be 8.

7. The method according to claim 6, characterized in that The sparrow search algorithm (SSA) was used to globally optimize the initial weight threshold of the BP neural network to form an SSA-BP hybrid model, in which the population size was set to 30 and the maximum number of iterations was 50. The optimal process parameters of the wine-scented Polygonatum odoratum preparation process were determined by combining the SSA-BP hybrid model with response surface analysis: steaming time 3.6 h, drying temperature 75 ° C, drying time 11.4 h, and steaming times 8 times.

8. The process for preparing the Huangjing (Shenghuangjing) of claim 7, wherein: The polysaccharide content, reducing sugar content, flavonoid content, and saponin content were 480.8 mg / g, 395.7 mg / g, 7.0 mg / g, and 28.9 mg / g, respectively; the 5-HMF content was 0.21 ± 0.03 mg / g; and the ABTS + Free radical scavenging rate IC 50 The value is 2.3mg / mL. The wine-cured Polygonatum odoratum slices obtained under this process are black in color, high in quality, and fragrant in taste, and comply with the requirements of the 2020 edition of the "Chinese Pharmacopoeia".

Images