Intelligent puer tea blending method based on physicochemical component characteristics of raw materials

Abstract

The invention discloses an intelligent Pu'er tea blending method based on physicochemical component characteristics of raw materials. The method comprises the following steps: determining and collecting blending parameter information and control parameters; establishing a Pu'er tea raw material basic information and corresponding raw material aroma and taste physicochemical component characteristic library; data preprocessing is carried out according to physical and chemical component characteristics, and a high-quality data set is provided for modeling; collecting quantitative target values of users on physicochemical component characteristics of the blended Pu'er tea, and comprehensively calculating aroma and taste scores based on the physicochemical component characteristics of the raw materials; constructing a Pu'er tea intelligent blending model based on the random forest according to the physicochemical component characteristics of the raw materials and the blending target value; establishing a matching fitness function selection knowledge base according to the matching target value and the fitness function; through a genetic algorithm, a fitness function is selected to carry out cross validation optimization on the blending model, and an optimal blending scheme is formed; and confirming optimal blending formula parameter output and performing blending control. Scientific and efficient tea blending is achieved, and the quality and consistency of final products are improved.

Description

Technical Field

[0001] The invention relates to the technical field of Pu'er tea, and in particular to an intelligent Pu'er tea blending method based on the physical and chemical composition characteristics of raw materials. Background Art

[0002] As a unique local tea, Pu'er tea enjoys a high reputation in the domestic and international markets. The production process of Pu'er tea involves multiple links such as spreading, killing, fermenting, drying, and storage of raw materials. Among them, the blending of raw materials is an important factor in determining the quality of the final product. The traditional Pu'er tea blending method mainly relies on experienced tea experts to formulate blending plans through their sensory evaluation and experience accumulation. This method has certain limitations, which are mainly manifested in the following aspects:

[0003] 1. Highly subjective: Although the experience and sensory evaluation of tea experts can guarantee the quality of blending to a certain extent, due to individual differences and subjective judgment, the blending results of different experts may vary greatly, making it difficult to ensure the consistency of each blend.

[0004] 2. Low efficiency: Traditional assembly methods mainly rely on manual operation, and multiple tests and adjustments are required to determine the final assembly plan. This is not only time-consuming and labor-intensive, but also difficult to meet the needs of large-scale production.

[0005] 3. Poor stability: The quality of tea raw materials is affected by many factors such as climate, soil, and picking time. Even the quality of the same batch of tea may fluctuate to a certain extent. Traditional methods are difficult to effectively cope with these changes, resulting in unstable quality of the final product.

[0006] 4. Difficult to quantify: Traditional blending methods mainly rely on the experience and feelings of experts. These indicators are difficult to quantify and standardize, and it is impossible to form systematic blending standards and specifications, which is not conducive to the control and optimization of the production process.

[0007] With the development of big data technology and artificial intelligence technology, data-driven intelligent blending methods have gradually become a research hotspot. Through comprehensive data collection of the physical and chemical properties of Pu'er tea raw materials and the use of advanced machine learning algorithms, scientific analysis and optimization of the Pu'er tea blending process can be achieved, thereby improving the efficiency and quality of blending. Summary of the invention

[0008] The purpose of the present invention is to provide a Pu'er tea intelligent blending method based on the physical and chemical composition characteristics of raw materials to address the above-mentioned problems. By collecting, processing, analyzing and predicting Pu'er tea raw material information, scientific and efficient tea blending can be achieved to improve the quality and consistency of the final product.

[0009] The technical solution of the present invention is as follows:

[0010] A Pu'er tea intelligent blending method based on the physical and chemical composition characteristics of raw materials, comprising the following steps:

[0011] Determine the acquisition of assembly parameter information and control parameters;

[0012] Establish the basic information of Pu'er tea raw materials and the characteristic database of aroma, flavor and physical and chemical components of the corresponding raw materials;

[0013] Data preprocessing is performed based on the data of the aroma, flavor and physical and chemical component feature library of raw materials to provide high-quality data sets for modeling;

[0014] Collect users' quantitative target values ​​for the physical and chemical composition characteristics of blended Pu'er tea, and comprehensively calculate the aroma and taste scores based on the physical and chemical composition characteristics of the raw materials;

[0015] A Pu'er tea intelligent blending model based on random forest is constructed through the physical and chemical composition characteristics of raw materials and blending target values;

[0016] According to the matching target value and fitness function, a matching fitness function selection knowledge base is established;

[0017] Through genetic algorithm, the fitness function is selected to perform cross-validation optimization of the assembly model to form the best assembly plan;

[0018] Optimal blending formula parameter output confirmation and blending control.

[0019] Furthermore, the Pu'er tea intelligent blending model based on the physical and chemical composition characteristics of raw materials includes the following steps:

[0020] Establish objective function: blending scheme with minimum finished product grade, maximum finished product flavor characteristics, and maximum finished product aroma characteristics as the goals;

[0021] Establish rules and constraints: physical and chemical composition characteristics of raw materials and user blending target values ​​are input conditions;

[0022] Model building: Use Python to build an intelligent matching prediction model based on the random forest algorithm;

[0023] Adaptation function: Use the trained random forest model to calculate the aroma and flavor scores of each blending solution and compare them with the user's target value. Use the genetic algorithm to further optimize the Pu'er tea blending and construct a fitness function to calculate and evaluate the degree of proximity between the random forest model and the user's expected aroma, flavor and grade targets for each blending solution.

[0024] Assembly model optimization: fitness function calculation, assembly fitness function coefficient feature library as the cross-validation input of the optimization model, and model parameter optimization through genetic algorithm.

[0025] Furthermore, the fitness function is:

[0026]

[0027] Among them, S P 、F P and G P is the aroma, taste and grade score of the blending solution predicted by the model, T x , T y and T z is the target rating provided by the user, w x 、w y and w z are the weight coefficients of aroma, taste and grade, and satisfy w x +w y +w z =1,|S P -T x | is the absolute difference between the aroma score of the blending solution and the target aroma score, |F P -T y | is the absolute difference between the flavor score of the blending solution and the target flavor score, |G P -T z | is the absolute difference between the tea grade score of the blending scheme and the target grade score, and the fitness value F fitness The value range is from 0 to 1. The closer the value is to 1, the closer the matching scheme is to the target.

[0028] Furthermore, the flavor characteristics are scored by a scoring model, which uses nonlinear weighting, dynamic adjustment and synergistic effect; the scoring model is:

[0029] T comprehensive =w b f(T b )+w k (100-g(T k ))+w s f(T s )+w h f(T h )+λh(T s , T k ),

[0030] Among them, f(x) is a nonlinear transformation of thickness, sweetness and aftertaste, g(T k ) is the nonlinear processing of bitterness, λh(T s , T k ) is to introduce synergistic effect, h(T s , T k ) represents the interaction between sweetness and bitterness, λ is the coefficient used to adjust the intensity of the synergistic effect, and wb is the thickness weight, w k is the bitterness weight, w s is the bitterness weight, w h It is the weight of sweetness.

[0031] Furthermore, the fitness function feature library predicts each blending scheme according to the intelligent blending model, combines the user target value with the expert knowledge, and establishes a blending fitness function feature library for optimizing the blending model and seeking the best formula solution.

[0032] Furthermore, the blending fitness function characteristics include a blended tea taste score, a blended tea aroma score, and a blended tea grade score.

[0033] Furthermore, the main chemical components of the Pu'er tea aroma are quantitatively scored according to a 100-point system, and a weighted average method is used to assign different weights w to each component according to its contribution to the overall aroma, and its concentration is standardized to 0-100 points. Finally, the overall aroma score is obtained by combining the scores of each component;

[0034] For each component, the normalized concentration score was calculated as:

[0035]

[0036] Where X is the actual concentration (in mg / L), Min represents the lower limit of the ideal concentration, Max represents the upper limit of the ideal concentration, and i represents the chemical composition; if the actual concentration is lower than Min, the score is 0; if it is higher than Max, the score is 100;

[0037] The calculation formula for the comprehensive aroma score is:

[0038]

[0039] Furthermore, the confirmed collection of blending parameter information includes: blending electronic scale information, blending material information, blending equipment information, and blending material weight information; control parameters involve blending formula parameters, including: blending material weight, blending scale flow, blending equipment control, and blending material ratio.

[0040] Furthermore, the data preprocessing is used for standardization of characteristic parameters of the raw material physical and chemical component characteristic library and preprocessing and feature engineering of the collected data to extract numerical representations of the physical and chemical component characteristics of the tea raw materials.

[0041] Furthermore, the Pu'er tea raw material physical and chemical component characteristic library contains the raw material origin, year, grade basic raw material information and raw material physical and chemical component characteristic detection data, the raw material physical and chemical component detection data includes: taste characteristics and aroma characteristics, taste characteristics include: caffeine, tea polyphenols, catechins, water-soluble sugars, amino acids, water extract material components; aroma characteristics include: volatile aromatic substances, polyphenol compounds, amino acids, aldoses, sulfur-containing compounds, fatty acid oxidation products, carotenoid derivatives and microbial metabolite material components;

[0042] The taste characteristics of Pu'er tea raw materials include four dimensions: "bitterness", "sweetness", "aftertaste" and "thickness", and each dimension is quantified into a user target value on a 100-point scale; the aroma characteristics of Pu'er tea are based on a comprehensive scoring result of the chemical component concentration of the water extract of the raw material, and the user target value is quantified on a 100-point scale.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] The present invention provides an intelligent blending method for Pu'er tea based on the physical and chemical composition characteristics of raw materials. An intelligent blending model is established by quantifying raw material characteristic indicators and blended tea indicators. The fitness of the blending model is optimized according to user target values ​​and the characteristics of the blended Pu'er tea. The minimum finished product grade, maximum finished product flavor characteristics and maximum finished product aroma characteristics of the blended Pu'er tea are taken as goals. The optimal raw material composition is intelligently calculated from the basic information of the Pu'er tea raw materials, and finally the optimal blended Pu'er tea formula is obtained.

[0045] The present invention can not only accurately blend Pu'er tea that meets the customer's personalized product customization target value, but also utilizes big data technology to drive Pu'er tea blending based on objective data, thereby reducing the influence of human subjective factors; at the same time, automated data processing and optimization algorithms improve blending efficiency; and through a scientific blending plan, the consistency of the quality and characteristics of each batch of tea is guaranteed. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of the process of the present invention.

[0047] Figure 2 The present invention is a flowchart of the intelligent blending of Pu'er tea.

[0048] Figure 3 The figure is a schematic diagram of the implementation process of the Pu'er tea intelligent blending algorithm of the present invention. DETAILED DESCRIPTION

[0049] It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0050] The features and performance of the present invention are further described in detail below in conjunction with the embodiments.

[0051] See also Figure 1-2 , an intelligent blending method of Pu'er tea based on the physical and chemical composition characteristics of raw materials, such as Figure 1 , Figure 2 and Figure 3 As shown, the following steps are included:

[0052] Determine the acquisition of assembly parameter information and control parameters;

[0053] Establish the basic information of Pu'er tea raw materials and the characteristic database of aroma, flavor and physical and chemical components of the corresponding raw materials;

[0054] Data preprocessing is performed based on the data of the aroma, flavor and physical and chemical component feature library of raw materials to provide high-quality data sets for modeling;

[0055] Collect users' quantitative target values ​​for the physical and chemical composition characteristics of blended Pu'er tea, and comprehensively calculate the aroma and taste scores based on the physical and chemical composition characteristics of the raw materials;

[0056] A Pu'er tea intelligent blending model based on random forest is constructed through the physical and chemical composition characteristics of raw materials and blending target values;

[0057] According to the matching target value and fitness function, a matching fitness function selection knowledge base is established;

[0058] Through genetic algorithm, the fitness function is selected to perform cross-validation optimization of the assembly model to form the best assembly plan;

[0059] Optimal blending formula parameter output confirmation and blending control.

[0060] The Pu'er tea intelligent blending model based on the physical and chemical composition characteristics of raw materials includes the following steps:

[0061] Establish objective function: blending scheme with minimum finished product grade, maximum finished product flavor characteristics, and maximum finished product aroma characteristics as the goals;

[0062] Establish rules and constraints: physical and chemical composition characteristics of raw materials and user blending target values ​​are input conditions;

[0063] Model building: Use Python to build an intelligent matching prediction model based on the random forest algorithm;

[0064] Adaptation function: Use the trained random forest model to calculate the aroma and flavor scores of each blending solution and compare them with the user's target value. Use the genetic algorithm to further optimize the Pu'er tea blending and construct a fitness function to calculate and evaluate the degree of proximity between the random forest model and the user's expected aroma, flavor and grade targets for each blending solution.

[0065] Assembly model optimization: fitness function calculation, assembly fitness function coefficient feature library as the cross-validation input of the optimization model, and model parameter optimization through genetic algorithm.

[0066] The fitness function is:

[0067]

[0068] Among them, S P 、F P and G P is the aroma, taste and grade score of the blending solution predicted by the model, T x , T y and T z is the target rating provided by the user, w x 、w y and w z are the weight coefficients of aroma, taste and grade, and satisfy w x +w y +w z =1,|S P -T x | is the absolute difference between the aroma score of the blending solution and the target aroma score, |F P -T y | is the absolute difference between the flavor score of the blending solution and the target flavor score, |G P -T z | is the absolute difference between the tea grade score of the blending scheme and the target grade score, and the fitness value F fitness The value range is from 0 to 1. The closer the value is to 1, the closer the matching scheme is to the target.

[0069] The flavor characteristics are scored using a scoring model that uses nonlinear weighting, dynamic adjustment, and synergy effects; the scoring model is:

[0070] T comprehensive =w b f(T b )+w k (100-g(T k ))+w s f(T s )+w h f(T h )+λh(T s , T k ),

[0071] Among them, f(x) is a nonlinear transformation of thickness, sweetness and aftertaste, g(T k ) is the nonlinear processing of bitterness, λh(T s , T k ) is to introduce synergistic effect, h(T s , T k ) represents the interaction between sweetness and bitterness, λ is the coefficient used to adjust the intensity of the synergistic effect, and w b is the thickness weight, w k is the bitterness weight, w s is the bitterness weight, w h It is the weight of sweetness.

[0072] The fitness function feature library predicts each blending scheme based on the intelligent blending model, combines the user's target value with expert knowledge, and establishes a blending fitness function feature library to optimize the blending model and seek the best formula solution.

[0073] The characteristics of the blending fitness function include the blended tea taste score, the blended tea aroma score, and the blended tea grade score.

[0074] The main chemical components of Pu'er tea aroma were quantitatively scored on a 100-point scale. A weighted average method was used to assign different weights w to each component according to its contribution to the overall aroma, and its concentration was standardized to 0-100 points. Finally, the overall aroma score was obtained by combining the scores of each component.

[0075] For each component, the normalized concentration score was calculated as:

[0076]

[0077] Where X is the actual concentration (in mg / L), Min represents the lower limit of the ideal concentration, Max represents the upper limit of the ideal concentration, and i represents the chemical composition; if the actual concentration is lower than Min, the score is 0; if it is higher than Max, the score is 100;

[0078] The comprehensive aroma score calculation formula is:

[0079]

[0080] Confirm that the collected blending parameter information includes: blending electronic scale information, blending material information, blending equipment information, and blending material weight information; control parameters involve blending formula parameters, including: blending material weight, blending scale flow, blending equipment control, and blending material ratio.

[0081] Data preprocessing is used for standardization of characteristic parameters of the raw material physical and chemical component characteristic library and preprocessing and feature engineering of the collected data to extract numerical representation of the physical and chemical component characteristics of the tea raw materials.

[0082] The Pu'er tea raw material physical and chemical component characteristic database contains the raw material origin, year, grade, basic raw material information and the test data of the raw material physical and chemical component characteristics. The test data of the raw material physical and chemical components include: taste characteristics and aroma characteristics. The taste characteristics include: caffeine, tea polyphenols, catechins, water-soluble sugars, amino acids, and water extract material components; the aroma characteristics include: volatile aromatic substances, polyphenol compounds, amino acids, aldoses, sulfur-containing compounds, fatty acid oxidation products, carotenoid derivatives and microbial metabolite material components;

[0083] The taste characteristics of Pu'er tea raw materials include four dimensions: "bitterness", "sweetness", "aftertaste" and "thickness", and each dimension is quantified into a user target value on a 100-point scale; the aroma characteristics of Pu'er tea are based on a comprehensive scoring result of the chemical component concentration of the water extract of the raw material, and the user target value is quantified on a 100-point scale.

[0084] In another specific embodiment, a Pu'er tea intelligent blending method based on the physical and chemical composition characteristics of raw materials, such as Figure 1 shown.

[0085] Obtain basic information on Pu'er tea raw materials and characteristic data of physical and chemical components;

[0086] The raw material physical and chemical component characteristic library includes basic elements such as raw material origin, year, grade, and the collection or detection data of raw material physical and chemical components; in this example, large-leaf sun-dried green tea of ​​different grades from different tea mountains in the same origin in Yunnan Province is selected to establish the raw material physical and chemical component characteristic library. Specifically, in this example, the raw material physical and chemical component characteristic data are shown in Table 1:

[0087] Table 1

[0088]

[0089]

[0090] Specifically, the raw material quality and taste characteristic library data is shown in Table 2:

[0091] Table 2

[0092]

[0093]

[0094]

[0095] Calculate the characteristic score of aroma components of Pu'er tea raw materials;

[0096] Aroma characteristics are quantified according to a 100-point system after data preprocessing. The actual concentration data of the main chemical components of the aroma component characteristics of the raw materials are calculated based on the comprehensive aroma score Table 3 and the comprehensive aroma score Table 4:

[0097] The main chemical components that affect the aroma of Pu'er tea and their weights w (the sum is 1):

[0098] Volatile aromatic substances (VOCs): 0.3

[0099] Polyphenols: 0.2

[0100] Amino Acids: 0.2

[0101] Sugars: 0.1

[0102] Sulfur Compounds: 0.1

[0103] Others: 0.1

[0104] The ideal concentration range for each chemical component is as follows:

[0105] Volatile aromatic substances (VOCs): [10,50] mg / L

[0106] Polyphenols: [100,300] mg / L

[0107] Amino Acids: [20,100] mg / L

[0108] Sugars: [5,15] mg / L

[0109] Sulfur Compounds: [0.1,1] mg / L

[0110] Others: [0.5,2]mg / L

[0111] Table 3

[0112]

[0113]

[0114] Table 4

[0115]

[0116] Calculate the characteristic scores of flavor components of Pu'er tea raw materials;

[0117] The taste characteristics of Pu'er tea include four dimensions: "bitterness", "sweetness", "aftertaste", and "thickness". After data preprocessing, each dimension is quantified into a 100-point system to determine the user's target value.

[0118] Example of flavor profile scoring:

[0119] 1) Choose the nonlinear transformation as a square function (i.e. ), the synergistic effect is the square of the difference between sweetness and bitterness, and then the adaptive weight adjustment is brought in, and the calculation formula is as follows:

[0120]

[0121] 2) Rating calculation

[0122] Assume the weights are:

[0123] Thickness weight wb = 0.3;

[0124] Bitterness weight wk = 0.2;

[0125] Sweetness weight ws = 0.25;

[0126] Sweetness weight wh=0.25;

[0127] Synergistic effect strength λ = 0.1;

[0128] The ratings for a particular blended tea are as follows:

[0129] Thickness Tb = 95;

[0130] Bitterness Tk = 50;

[0131] Sweetness Ts = 85;

[0132] Aftertaste Th = 90;

[0133] Substituting these values ​​into the formula:

[0134]

[0135]

[0136] Calculation result output: T comprehensive =85.6125.

[0137] Collect the quantitative data of the quality component characteristics of blended Pu'er tea selected by users; Table 5 is the quantitative target value table of the blended Pu'er tea formula characteristics selected by users (the level adopts a 1 to 5 point system, 5 is the highest):

[0138] Table 5

[0139]

[0140] According to the user's target value and the physical and chemical composition characteristics of the raw materials, the raw materials are split into 1000 samples as the input for random forest model training, and the objective function is the user's target value.

[0141] In this example, the implementation process of the intelligent blending model based on the physical and chemical composition characteristics of raw materials

[0142] The entire process is divided into two core parts: the splicing prediction model training process and the splicing optimization improvement process. The splicing prediction model is trained through random forests, and the optimization is improved through genetic algorithms. The following is a detailed algorithm implementation process:

[0143] Data collection and preprocessing:

[0144] Input: Through the physical and chemical properties of Pu'er tea raw materials (such as tea polyphenol content, caffeine, aroma components, etc.), collect relevant historical blending data and the aroma, taste, and grade scores of the finished products.

[0145] Data preprocessing: fill missing values, normalize values, and ensure the consistency of input data. Divide the data set into training set and test set.

[0146] Model training:

[0147] Random forest algorithm: The processed spliced ​​data set is input into the random forest model for training. Random forest makes predictions through a collection of multiple decision trees, which can better handle high-dimensional data and avoid overfitting.

[0148] Input features: chemical composition characteristic data of blended raw materials and historical formulas.

[0149] Output labels: corresponding aroma score, taste score, and tea grade.

[0150] Model Evaluation:

[0151] Cross-validation: Use cross-validation techniques to evaluate the performance of the model, avoid overfitting of the model, and ensure generalization ability.

[0152] Performance indicators: The mean square error (MSE) indicator was used to evaluate the model's prediction ability for aroma, taste, and grade.

[0153] Adjust hyperparameters: Use random search to tune the hyperparameters of the random forest model, such as the number of decision trees, maximum depth, etc.

[0154] Save the model:

[0155] The trained and well-validated random forest model was saved as the baseline model for predicting Pu’er tea blending.

[0156] Blending optimization and improvement process:

[0157] This part uses genetic algorithms to optimize the blending scheme, aiming to find the best blending scheme so that the aroma, flavor, and grade scores meet the user's target target values.

[0158] Initialize the population:

[0159] Population initialization: Randomly generate several blending schemes, each of which consists of different proportions of raw materials to form an initial population.

[0160] Chromosome representation: Each combination scheme can be represented by a vector, and each element in the vector represents the proportion of a certain raw material.

[0161] Fitness function calculation:

[0162] Prediction score: Each blending scheme in the population is input into the random forest prediction model to obtain the corresponding aroma, taste and grade scores.

[0163] Fitness function: The fitness value of each blending solution is calculated based on the fitness function model (the difference between the comprehensive aroma, taste, grade and the user's target value). The higher the fitness, the closer the solution is to the user's target.

[0164] Select an action:

[0165] Selection operator: Through the tournament selection algorithm, the combination scheme with higher fitness is selected as the parent generation based on the fitness value and passed to the next generation.

[0166] Crossover operation:

[0167] Crossover operator: Perform a crossover operation on the selected parent solution to generate a new offspring solution. Use multivariate crossover to exchange some genes (raw material ratio) of the parent chromosome to generate a new combination solution.

[0168] Mutation operation:

[0169] Mutation operator: Perform mutation operations on the newly generated offspring solutions and randomly adjust the proportions of certain raw materials to increase the diversity of the population and avoid local optimal solutions.

[0170] Fitness evaluation:

[0171] Evaluate the new population: Input the new population after crossover and mutation into the prediction model again and recalculate its fitness value.

[0172] Retain the best individuals: retain the individuals (matching solutions) with the highest fitness to the next generation to ensure that the population evolves towards the optimal solution.

[0173] Termination conditions:

[0174] Convergence judgment: If the fitness of the best individual in the population does not improve significantly in several generations, or reaches the set algebraic limit, the algorithm is considered to have converged and the optimization is stopped.

[0175] Output the optimal solution: Get the assembly solution with the highest fitness and output it as the optimal assembly solution.

[0176] As shown in Table 6, it is used as the initial prediction solution formula of random forest.

[0177] Table 6

[0178]

[0179] The fitness function instance calculation sets the goal as follows:

[0180] Aroma Target T s =90 (using percentage system, 100 is the highest);

[0181] Taste Target T f =85 (using the percentage system, 100 is the highest);

[0182] Tea grade target T g =5 (the rating is based on a 1 to 5 point system, with 5 being the highest);

[0183] One of the initial predictions of the random forest-based matching solution is:

[0184] Aroma score S p =88;

[0185] Taste score: F p =82;

[0186] Rating G p =4;

[0187] Set the weight w s =0.4, w f =0.4, w g =0.2, that is, the weights of aroma and taste are higher, and the weight of grade is relatively low.

[0188] Substituting these values ​​into the fitness function:

[0189]

[0190] The initial fitness value of the matching scheme is 0.937, which means that this scheme is very close to the user's goal and is a good matching scheme.

[0191] In this example, after 20 generations of evolution through genetic algorithm optimization, the optimal blending solution was obtained, with an aroma score of 89.5, a taste score of 84.8, a grade score of 5, and a fitness value of 0.98, which is close to the user's expected goal. The system provides the following blending solutions:

[0192] -Raw material 1 accounts for 40%

[0193] -Raw material 2 accounts for 40%

[0194] -Raw material 4 accounts for 20%

[0195] The system also suggests that the proportion of raw material 4 can be appropriately increased to get closer to the user's target score.

[0196] The best blending solution is obtained through the Pu'er tea intelligent blending solution, as shown in Table 7:

[0197] Table 7

[0198]

[0199] Through the present invention, the control parameters of the blending formula are sent to the control system for blending production.

[0200] Summary: According to the results in Table 8, after expert review, the characteristic component contents corresponding to the conversion of intelligent blended Pu'er tea were compared, and the analysis showed that (the aroma characteristics considering volatility and oxidizability were finally scored 85) that the products blended by the Pu'er tea intelligent blending method based on the physical and chemical composition characteristics of the raw materials meet the user's target value, and the method of the present invention is effective and feasible.

[0201] Table 8

[0202]

[0203] The above-mentioned embodiments only express the specific implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the protection scope of the present application. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the technical solution concept of the present application, and these all belong to the protection scope of the present application.

Claims

1. A Pu'er tea intelligent blending method based on the physical and chemical composition characteristics of raw materials, characterized in that: The following steps are involved: Determine the acquisition of assembly parameter information and control parameters; Establish the basic information of Pu'er tea raw materials and the characteristic database of aroma, flavor and physical and chemical components of the corresponding raw materials; Data preprocessing is performed based on the data of the aroma, flavor and physical and chemical component feature library of raw materials to provide high-quality data sets for modeling; Collect users' quantitative target values ​​for the physical and chemical composition characteristics of blended Pu'er tea, and comprehensively calculate the aroma and taste scores based on the physical and chemical composition characteristics of the raw materials; A Pu'er tea intelligent blending model based on random forest is constructed through the physical and chemical composition characteristics of raw materials and blending target values; According to the matching target value and fitness function, a matching fitness function selection knowledge base is established; Through genetic algorithm, the fitness function is selected to perform cross-validation optimization of the assembly model to form the best assembly plan; Optimal blending formula parameter output confirmation and blending control.

2. The method for intelligently blending Pu'er tea based on the physical and chemical composition characteristics of raw materials according to claim 1, characterized in that: The Pu'er tea intelligent blending model based on the physical and chemical composition characteristics of raw materials comprises the following steps: Establish objective function: blending scheme with minimum finished product grade, maximum finished product flavor characteristics, and maximum finished product aroma characteristics as the goals; Establish rules and constraints: physical and chemical composition characteristics of raw materials and user blending target values ​​are input conditions; Model building: Use Python to build an intelligent matching prediction model based on the random forest algorithm; Adaptation function: Use the trained random forest model to calculate the aroma and flavor scores of each blending solution and compare them with the user's target value. Use the genetic algorithm to further optimize the Pu'er tea blending and construct a fitness function to calculate and evaluate the degree of proximity between the random forest model and the user's expected aroma, flavor and grade targets for each blending solution. Assembly model optimization: fitness function calculation, assembly fitness function coefficient feature library as the cross-validation input of the optimization model, and model parameter optimization through genetic algorithm.

3. The method for intelligently blending Pu'er tea based on the physical and chemical composition characteristics of raw materials according to claim 2, characterized in that: The fitness function is: Among them, S P 、F P and G P is the aroma, taste and grade score of the blending solution predicted by the model, T x , T y and T z is the target rating provided by the user, w x 、w y and w z are the weight coefficients of aroma, taste and grade, and satisfy w x +w y +w z =1,|S P -T x | is the absolute difference between the aroma score of the blending solution and the target aroma score, |F P -T y | is the absolute difference between the flavor score of the blending solution and the target flavor score, |G P -T z | is the absolute difference between the tea grade score of the blending scheme and the target grade score, and the fitness value F fitness The value range is from 0 to 1. The closer the value is to 1, the closer the matching scheme is to the target.

4. The method for intelligently blending Pu'er tea based on the physical and chemical composition characteristics of raw materials according to claim 1, characterized in that: The flavor characteristics are scored by a scoring model that uses nonlinear weighting, dynamic adjustment, and synergistic effects; the scoring model is: T comprehensive =w b f(T b )+w k (100-g(T k ))+w s f(T s )+w h f(T h )+λh(T s ,T k ), Among them, f(x) is a nonlinear transformation of thickness, sweetness and aftertaste, g(T k ) is the nonlinear processing of bitterness, λh(T s , T k ) is to introduce synergistic effect, h(T s , T k ) represents the interaction between sweetness and bitterness, λ is the coefficient used to adjust the intensity of the synergistic effect, and w b is the thickness weight, w k is the bitterness weight, w s is the bitterness weight, w h It is the weight of sweetness.

5. The intelligent Pu'er tea blending method based on the physical and chemical composition characteristics of raw materials according to claim 2 or 3, characterized in that: The fitness function feature library predicts each blending scheme according to the intelligent blending model, combines the user target value with expert knowledge, and establishes a blending fitness function feature library for optimizing the blending model and seeking the best formula solution.

6. The method for intelligently blending Pu'er tea based on the physical and chemical composition characteristics of raw materials according to claim 5, characterized in that: The blending fitness function features include a blended tea taste score, a blended tea aroma score, and a blended tea grade score.

7. The method for intelligently blending Pu'er tea based on the physical and chemical composition characteristics of raw materials according to claim 6, characterized in that: The main chemical components of the Pu'er tea aroma are quantitatively scored according to a 100-point system, and a weighted average method is used to assign different weights w to each component according to its contribution to the overall aroma, and its concentration is standardized to 0-100 points. Finally, the overall aroma score is obtained by combining the scores of each component; For each component, the normalized concentration score was calculated as: Where X is the actual concentration (in mg / L), Min represents the lower limit of the ideal concentration, Max represents the upper limit of the ideal concentration, and i represents the chemical composition; if the actual concentration is lower than Min, the score is 0; if it is higher than Max, the score is 100; The calculation formula for the comprehensive aroma score is:

8. The method for intelligently blending Pu'er tea based on the physical and chemical composition characteristics of raw materials according to claim 1, characterized in that: The confirmed collection of blending parameter information includes: blending electronic scale information, blending material information, blending equipment information, and blending material weight information; control parameters involve blending formula parameters, including: blending material weight, blending scale flow, blending equipment control, and blending material ratio.

9. The method for intelligently blending Pu'er tea based on the physical and chemical composition characteristics of raw materials according to claim 1, characterized in that: The data preprocessing is used for standardization of characteristic parameters of the raw material physical and chemical component characteristic library and preprocessing and characteristic engineering of the collected data to extract numerical representations of the physical and chemical component characteristics of the tea raw materials.

10. The method for intelligently blending Pu'er tea based on the physical and chemical composition characteristics of raw materials according to claim 1, characterized in that: The Pu'er tea raw material physical and chemical component characteristic library contains the raw material origin, year, grade, basic raw material information and the detection data of the raw material physical and chemical component characteristics. The detection data of the raw material physical and chemical components include: taste characteristics and aroma characteristics. The taste characteristics include: caffeine, tea polyphenols, catechins, water-soluble sugars, amino acids, and water extract material components; the aroma characteristics include: volatile aromatic substances, polyphenol compounds, amino acids, aldoses, sulfur-containing compounds, fatty acid oxidation products, carotenoid derivatives and microbial metabolite material components; The taste characteristics of Pu'er tea raw materials include four dimensions: "bitterness", "sweetness", "aftertaste", and "thickness", and each dimension is quantified into a user target value on a 100-point scale; the aroma characteristics of Pu'er tea are based on a comprehensive scoring result of the chemical component concentration of the water extract of the raw material, and the user target value is quantified on a 100-point scale.

Images