A crystallization-resistant fructose syrup processing control system based on data analysis

Through the anti-crystalline fructose syrup processing control system based on data analysis, the correlation analysis model and characteristic parameter matrix optimize control parameters are used to solve the crystallization problem in fructose syrup processing, improve product quality and processing efficiency, and achieve real-time and accurate Data analysis and control.

CN118584924BActive Publication Date: 2024-11-15ANHUI PINQING FOOD IND CO LTD
View PDF 11 Cites 0 Cited by

Patent Information

Application Number
CN202411055659.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-02
Publication Date
2024-11-15
Estimated Expiration
2044-08-02

AI Technical Summary

Technical Problem

Crystallization during fructose syrup processing leads to product quality decline, poor stability, reduced processing efficiency, economic losses and limited application. It is difficult for the existing technology to effectively predict crystallization behavior and adjust control parameters in real time.

Method used

Design an anti-crystalline fruit glucose syrup processing control system based on data analysis, including data acquisition module, association fitting module and control instruction generation module. By constructing an association analysis model and characteristic parameter matrix, the main factors affecting crystallization are extracted in real time and optimized. Control parameters to reduce waste of irrelevant data resources and improve data processing efficiency and response speed.

Benefits of technology

Real-time and precise anti-crystalline control of fructose syrup processing is realized, product quality and processing efficiency are improved, production costs are reduced, and the system's intelligent analysis capabilities and data processing flexibility are enhanced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118584924B_ABST
    Figure CN118584924B_ABST
Patent Text Reader

Abstract

The invention discloses an anti-crystallization fructose-glucose syrup processing control system based on data analysis, which relates to the field of data analysis. The technical scheme thereof mainly includes: constructing a target matrix and a sample matrix, and constructing an association analysis model, using the target matrix and the sample matrix to train the association analysis model, and obtaining an optimal association analysis model; collecting real-time data vectors and target variables at the current time point, importing the target variables and the real-time data vectors into the optimal association analysis model, and selecting data category values ​​corresponding to the first k regression coefficients; constructing a characteristic parameter matrix, traversing k data categories to form a numerical combination, calculating a path score of each numerical combination, and selecting a numerical combination with the smallest path score as an optimal control parameter set; adjusting the fructose-glucose syrup processing according to the optimal control parameter set as a processing instruction, and realizing anti-crystallization regulation and control of the fructose-glucose syrup processing.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of data analysis, and more specifically, to an anti-crystallization fructose syrup processing control system based on data analysis. Background Art

[0002] Crystallization of fructose corn syrup during processing will bring about a series of adverse effects. Crystallization will cause the taste of fructose corn syrup to deteriorate and affect its appearance. At the same time, it will lead to a decrease in the stability of fructose corn syrup, and it will be prone to stratification or further crystallization, affecting the storage and transportation of the product. In addition, it will increase the difficulty of the processing process, requiring additional steps to dissolve the crystals, consuming more time and energy, affecting production efficiency, and requiring additional manpower, material and financial resources, affecting the economic benefits of the enterprise. Therefore, anti-crystallization settings are required during the processing of fructose corn syrup.

[0003] In the process of realizing the anti-crystallization processing control of fructose-glucose solution, the crystallization behavior of glucose syrup is affected by many factors, such as temperature, concentration, pH value, impurity content, etc. The diversity and complexity of data increase the difficulty of crystallization prediction. In addition, too many collection factors lead to a significant increase in the workload of data processing and mining, and may not be able to give full play to the advantages of data analysis, resulting in untimely response speed or reduced efficiency, and failure to output prediction results in time, which in turn leads to the inability to adjust processing parameters in time, affecting product quality and reducing the real-time performance of the control system. Therefore, it is necessary to design the anti-crystallization processing control of fructose-glucose syrup. Summary of the invention

[0004] In view of the problems existing in the prior art, the object of the present invention is to provide an anti-crystallization high fructose syrup processing control system based on data analysis, so as to realize the anti-crystallization regulation of high fructose syrup processing.

[0005] To achieve the above object, the present invention provides the following technical solution: a data analysis-based anti-crystallization high fructose syrup processing control system, the data analysis-based anti-crystallization high fructose syrup processing control system comprising:

[0006] Data acquisition module: used to collect data during the processing of high fructose corn syrup, and perform preliminary processing to obtain the crystallization influencing parameter set and equipment control parameter set;

[0007] Correlation fitting module: used to select the value of the data categories in the crystallization influencing parameter set and the equipment control parameter set, and obtain k data categories that affect the degree of crystallization at each time point;

[0008] Control instruction generation module: used to construct the function expression of the characteristic parameter matrix and the path score, and take the numerical combination with the minimum path score as the optimal control parameter set; adjust the fructose corn syrup processing according to the optimal control parameter set as the processing instruction;

[0009] The modules are connected to each other via wired and / or wireless means.

[0010] Preferably, the data categories of the crystallization influencing parameter set include temperature, concentration, pH value, impurity content, microbial count, solute composition, additive data and pressure data of the fructose-glucose solution;

[0011] The data categories of the equipment control parameter set include stirring data, cooling data, gas phase data and filtration data.

[0012] Preferably, the preliminary processing of data in the fructose-glucose syrup processing process comprises:

[0013] Identify non-numerical information in solute composition, additive data and gas control, perform binary conversion on the solute category, additive category and gas category respectively through unique hot encoding, use binary strings to represent them, extract the numerical values ​​corresponding to the solute proportion, additive addition amount and gas flow value, and fill them into the corresponding binary strings, and set the representation of solute composition, additive data and gas control as binary string + numerical value;

[0014] The data are standardized and normalized to make them have the same dimension, forming a crystallization influencing parameter set and a device control parameter set.

[0015] Preferably, the value selection of the data categories in the crystallization influencing parameter set and the equipment control parameter set comprises:

[0016] At the current time point, the real-time data vector I is constructed according to a certain sequence using the values ​​corresponding to the data categories in the crystallization influence parameter set and the device control parameter set. t =[x1,…,xo,…,xm], xo represents the sample value of the oth data category, m represents the number of data categories, and o=1,…,m;

[0017] Collect the crystallization degree of fructose syrup at the current time point, obtain the target variable yt, import the target variable yt and the real-time data vector into the best association analysis model, obtain the regression coefficient corresponding to each data category, arrange the absolute values ​​of the regression coefficients in descending order, select the data category features corresponding to the first k regression coefficients as the main factors affecting the crystallization degree, and complete the feature extraction.

[0018] Preferably, the method for obtaining the value of k comprises: taking the current time point as an anchor point, tracing back s time points to collect values ​​of m data categories, and extracting the degree of crystallization corresponding to the fructose-glucose syrup at each time point;

[0019] At each time point c, by calculating the value coefficient The number of data categories selected when the collection value coefficient is the highest, and then the number of data categories lc with the highest correlation with the degree of crystallization among the m data categories is obtained.

[0020] Extract the number of data categories obtained at each time point and calculate the average Then the value of k at the current time point is

[0021] Preferably, the specific calculation process of the value coefficient includes:

[0022] Among the m data categories, randomly select l 1 The values ​​corresponding to the data categories are the remaining ml 1 The data categories are numerically standardized, the corresponding degree of crystallization of high fructose corn syrup under different selections are collected, the value coefficients of different quantities and different data categories are traversed and calculated, the number of data categories corresponding to the highest value coefficient value is extracted, and then the number of data categories with the highest correlation with the degree of crystallization at each time point is obtained.

[0023] Preferably, the method for obtaining the optimal association analysis model specifically includes:

[0024] Collect q groups of crystallization influencing parameter sets and equipment control parameter sets, and traverse the numbers according to the sequence of data categories in the real-time data vector, mark the crystallization degree of fructose syrup in each group corresponding to each number, yi represents the crystallization degree of fructose syrup corresponding to the i-th group, and the crystallization degree of fructose syrup is used as the target variable to construct the target matrix Constructing the sample matrix All data categories of the crystallization influencing parameter set and the equipment control parameter set are taken as the rows of the matrix, and the numbers of each group are taken as the columns, wherein xij represents the sample value of the jth data category of the i-th group, m represents the number of data categories, and i=1, ..., m, j=1, ..., n;

[0025] Build a correlation analysis model, Among them, L(β) refers to the regression coefficient of the model output, β0 is the intercept term, βj is the regression coefficient of the jth parameter, and λ is the regularization parameter used to control the strength of the regularization term;

[0026] Initialize the parameters in the association analysis model, import w groups of target matrices and corresponding sample matrices into the association analysis model in batches, train the association analysis model, and set the constraints as The best association analysis model is obtained until the value of the constraint condition is minimized and no longer changes.

[0027] Preferably, the construction of the characteristic parameter matrix includes: defining a characteristic parameter matrix and a crystallization target; the crystallization target is to find a set of control parameter sets with the minimum degree of crystallization;

[0028] Define the value z of the pth row of the oth data category in the feature parameter matrix op Start from, select a value in each data category, traverse k data categories to form a value combination, and calculate the path score FG of each value combination;

[0029] Set the limit function to W FG v≤Wmax, where W FG v represents the economic value consumed when processing control is performed according to the vth path score FG, Wmax represents the maximum economic value of consumption limited by the enterprise, and when the restriction function is satisfied, the numerical combination with the smallest path score FG is selected as the optimal control parameter set.

[0030] Preferably, the method for constructing the characteristic parameter matrix includes: obtaining the data category values ​​corresponding to the first k regression coefficients, discretizing the parameter processing control range according to the parameter processing control range of each data category and taking the adjustment accuracy as the interval to obtain the selected parameters, taking the value at the current time point as the middle value, and selecting the selected parameters in turn by offsetting d distances to the left and right, and then there are a total of 2d+1 values ​​selected for each data category, arranging the values ​​in descending order, with the data categories as rows and the number of values ​​as columns, to construct a k×(2d+1) characteristic parameter matrix.

[0031] Preferably, the construction of the path score FG includes:

[0032] The function expression of the path score is set as FG = ω0 + ω1z 1f +......+ω k z ke +ε, where ω0 represents the initial weight value, ω vc represents the weight coefficient corresponding to the vcth data category, ε represents the error term, and vc=1,……,k,

[0033] At the current time point, collect the data category values ​​and crystallinity corresponding to k regression coefficients, and define ω vc The value range of ω0 and ε can be selected according to the specific situation. The grid search method is used to traverse within the given value range. For each set of ω vc , ω0 and ε, calculate the absolute difference between the degree of crystallization predicted by the function expression and the degree of crystallization at the current time point, and use cross-validation to train the model of the function expression.

[0034] Choose the ω with the smallest absolute difference vcThe values ​​of , ω0 and ε are taken as the optimal solution, and the optimal solution is imported into the function expression of the path score as the final function expression.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] In the present invention, by constructing an association analysis model, the data category features corresponding to the first k regression coefficients at the current time point are selected as the main factors affecting the degree of crystallization, the main factors affecting the crystallization of fructose syrup are extracted in real time, and targeted analysis and deep mining are carried out on the main data category factors to reduce the waste of resources on irrelevant data, concentrate system resources on high-value associated data, and achieve data optimization. By constructing a feature parameter matrix, k data categories are traversed to form a numerical combination, the path score of each numerical combination is calculated, and the numerical combination with the smallest path score is selected as the optimal control parameter set. The fructose syrup is adjusted according to the optimal control parameter set as a processing instruction. Syrup processing can achieve the best control parameter output under the condition of economic benefit constraints. It not only performs high-quality decomposition and mining of feature data, improves the effect of data processing and model training, as well as the retrieval and response speed, and provides an accurate data basis for subsequent model optimization; it also injects new impetus into the agility and flexibility of data processing. The system implements path scoring based on the current time point, predicts the control parameters at the next time point, and has an adaptive regulation and optimization mechanism, which continuously improves the overall efficiency of intelligent analysis. Multiple modules work together to realize comprehensive data collection, processing and analysis, and provide real-time and accurate data analysis in fructose syrup anti-crystallization control. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 The present invention proposes a structural schematic diagram of an anti-crystallization fructose syrup processing control system based on data analysis;

[0038] Figure 2 The figure is a schematic diagram of the method steps for the anti-crystallization high fructose syrup processing control system based on data analysis in the present invention. DETAILED DESCRIPTION

[0039] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0040] Embodiment 1

[0041] Reference Figure 1 and Figure 2 Embodiment 1 further illustrates an anti-crystallization high fructose syrup processing control system based on data analysis proposed by the present invention.

[0042] Crystallization of fructose syrup during processing will bring a series of adverse effects, mainly including:

[0043] 1. Product quality declines

[0044] Taste and appearance: Crystallization will cause the taste of high fructose corn syrup to deteriorate, and its appearance will also be affected. It will no longer be a uniform transparent liquid, but granular crystals will appear, affecting the user experience.

[0045] Poor stability: Crystallization may lead to decreased stability of high fructose corn syrup, making it prone to stratification or further crystallization, affecting the storage and transportation of the product.

[0046] 2. Reduced processing efficiency

[0047] Increased processing difficulty: Crystallization will increase the difficulty of the processing, requiring additional steps to dissolve the crystals, consuming more time and energy.

[0048] Equipment blockage: Crystallization may cause blockage of pipes, filters and other components of processing equipment, affecting production efficiency and increasing equipment maintenance costs.

[0049] 3. Economic losses

[0050] Reduced output: Reduced processing efficiency and equipment blockage caused by crystallization may reduce output and affect the economic benefits of the enterprise.

[0051] Increased costs: In order to solve the crystallization problem, additional manpower, material and financial resources may be required, such as purchasing anti-crystallization additives, adding heating and stirring steps, thereby increasing production costs.

[0052] 4. Limited product application

[0053] Narrowed application scope: Crystallized fructose syrup may not be suitable for certain foods and beverages that have strict requirements on transparency and stability, limiting the application scope of the product.

[0054] Therefore, the technical problems that need to be solved in the process of realizing anti-crystallization processing control mainly include the following aspects:

[0055] 1. How to accurately predict the crystallization behavior of fructose syrup under different conditions;

[0056] The crystallization behavior of fructose syrup is affected by many factors, such as temperature, concentration, pH value, impurity content, etc. The diversity and complexity of the data increase the difficulty of prediction. In order to ensure the accuracy of the prediction, it is necessary to conduct in-depth correlation mining of the data. However, if all the collected data are directly invested in all computing power for mining and analysis with equal resource allocation, it will cause waste of resources, reduce operating efficiency and prolong processing time. Because under different production conditions, the factors affecting crystallization are different. Some factors will dominate the production of crystallization, while some factors have little effect on crystallization. For these data, they can be discarded when analyzing crystallization to achieve the effect of streamlining and selecting data, balance the allocation of system computing resources, and improve data analysis speed.

[0057] If all data are analyzed, it may also lead to inaccurate data indexing, making it impossible to adjust parameters in time during processing, leading to crystallization problems. Crystallization problems will affect the quality and stability of products and reduce market competitiveness.

[0058] 2. How to associate and respond to control parameters during processing:

[0059] When mining and analyzing the parameters in the crystallization-affecting parameter set, it is necessary to associate them with the current equipment control parameter set in order to accurately output each control parameter based on the relevant data. However, how to establish a correlation mapping between the two and output the optimal control parameter adjustment under the constraint of optimal energy use to improve the performance and stability of the anti-crystallization fructose syrup processing control system and ensure product quality and production efficiency.

[0060] The solution is as follows:

[0061] By taking these measures, the adverse effects of crystallization during fructose syrup processing can be effectively reduced, ensuring product quality and processing efficiency.

[0062] Data acquisition module: used to collect data during the processing of high fructose corn syrup, and perform preliminary processing to obtain the crystallization influencing parameter set and equipment control parameter set;

[0063] The data categories of the crystallization-affecting parameter set include the temperature, concentration, pH value, impurity content, microbial count, solute composition, additive data, and pressure data of the fructose-glucose solution;

[0064] Temperature: Temperature is an important factor affecting the crystallization of fructose syrup. Lower temperature is conducive to the occurrence and control of crystallization, and in the present invention, the display value of the temperature sensor is read and collected.

[0065] Concentration: The concentration of fructose solution has a significant effect on crystallization behavior. A higher concentration increases the possibility of crystallization. In the present invention, the concentration is read and collected by a concentration sensing device or an electronic production data center, and the obtained value is expressed as a percentage.

[0066] pH value: The pH value of the fructose solution will also affect crystallization. Crystallization behavior may change under different pH conditions. The pH value is read numerically by a pH monitoring device.

[0067] Impurity content: The impurity content in the fructose-glucose solution will affect the occurrence and rate of crystallization. The impurity content is read and collected by the laboratory's data analysis center or electronic production data center, and the proportion of impurities in the fructose-glucose solution raw material is expressed as a percentage.

[0068] Number of microorganisms: In a standard processing and production container, the number of microorganisms will interact with the substances in the fructose-glucose solution, affecting the generation and rate of crystallization. The number of microorganisms in a standard unit is obtained through microscopic observation or experimental data analysis, and the number of microorganisms in the present invention is numerically expressed.

[0069] Solute composition: In addition to the total concentration, the composition of the solute will also affect the crystallization behavior. The interaction between different solutes and the difference in solubility between solutes will affect the occurrence and rate of crystallization. The solute here refers to the fructose-glucose solution formed by mixing different types of fruits or different varieties of the same fruit. In the present invention, the solute composition is read and collected according to experimental analysis or electronic production data center, and the expression is (solute 1 + proportion value) + (solute 2 + proportion value) + ...

[0070] Additive data: Adding an appropriate amount of anti-crystallization agent, such as sucrose, maltose, etc., during the processing can inhibit the formation of crystals. Selecting appropriate additives and adjusting the amount of additives added can change the solubility of solutes and solvent properties in fructose syrup, thereby affecting the crystallization behavior. Additive data is read and collected based on experimental analysis or the electronic production data center, and is expressed in the form of (additive 1 + amount added) + (additive 2 + amount added) +...

[0071] Pressure data: In some cases, adjusting the pressure during the process can affect the gas phase conditions in the fructose syrup and thus the crystallization behavior. The pressure data is read and collected through the display value of the pressure equipment.

[0072] The data categories of the equipment control parameter set include stirring data, cooling data, gas phase data and filtration data.

[0073] Stirring speed: The stirring speed affects the distribution and mixing degree of the solute in the fructose-glucose syrup by the output of the equipment controlling the power, thereby affecting the occurrence and rate of crystallization. In the present invention, the stirring speed is numerically read through the control center (display device) of the stirring equipment.

[0074] Cooling rate: The cooling rate is achieved by adjusting the control parameters of the cooling device, which can affect the crystallization process of the solute in the fructose-glucose syrup. In the present invention, the cooling rate is numerically read by the control center (display device) of the cooling device.

[0075] Gas control: By controlling the gas flow rate and gas composition during the processing, the gas phase conditions in the fructose syrup can be adjusted, thereby affecting the formation and characteristics of crystallization. In the present invention, the gas control is read by the control center (display device) of the gas control equipment, and the collection form is (gas 1+flow value)+(gas 2+flow value)+…….

[0076] Filtration control: The operating parameters of the filtration equipment, such as filtration speed, frequency and gap control, can separate and process the crystals in the fructose-glucose syrup. In the present invention, the filtration control is numerically read by the control center (display device) of the filtration control equipment.

[0077] By collecting and analyzing a large amount of fructose syrup processing data, an accurate crystallization prediction model is constructed to predict crystallization behavior under different conditions. Advanced sensors, data acquisition equipment, processing and production data storage centers, and experimental data centers are used to collect key parameters in real time, and efficient data processing algorithms are used for real-time analysis and adjustment. Control algorithms are designed and optimized, and processing parameters are dynamically adjusted in combination with real-time data to prevent crystallization. The use of machine learning and optimization algorithms can improve the performance of control algorithms.

[0078] Reasonably plan the system architecture to ensure smooth data transmission and communication of each module, and adopt standardized interfaces and protocols to improve the stability and compatibility of the system. Use technical means such as encrypted transmission and access control to ensure the security and privacy protection of data collected and processed during the processing. Adopt modular design and open interfaces to facilitate system expansion and upgrades, and carry out standardized documentation and maintenance work to improve the maintainability of the system.

[0079] The preliminary processing of data during the fructose syrup processing includes:

[0080] Identify non-numerical information in solute composition, additive data and gas control, perform binary conversion on the solute category, additive category and gas category respectively through unique hot encoding, use binary strings to represent them, extract the numerical values ​​corresponding to the solute proportion, additive addition amount and gas flow value, and fill them into the corresponding binary strings, and set the representation of solute composition, additive data and gas control as binary string + numerical value;

[0081] The data are standardized and normalized to make them have the same dimension, forming a crystallization influencing parameter set and a device control parameter set.

[0082] In the present invention, assuming that the solute composition includes apples and grapes, and the proportion of apples is 70%, the proportion of grapes is 30%, the binary string of apples encoded by one-hot encoding is 00011, and the binary string of grapes encoded by one-hot encoding is 01010, then the data of the solute composition is: 00011.7 and 01010.3;

[0083] The only additive is sucrose, and the amount of sucrose added is 1 mg / L. The binary string of sucrose encoded by one-hot encoding is 11000, so the additive data is: 11001;

[0084] The gas types are nitrogen and carbon dioxide, and the flow rate of nitrogen is 5m 3 / h, the flow rate of carbon dioxide is 7m 3 / h, the binary string of nitrogen encoded by one-hot encoding is 10000, and the binary string of carbon dioxide encoded by one-hot encoding is 10100, then the gas control data is: 10005 and 10107;

[0085] By converting all data into numerical representations and unifying the data representation, we can better conduct in-depth data mining and correlation analysis, and reduce the time used for conversion between different formats.

[0086] Standardization makes the distribution of data have zero mean and unit variance by subtracting the mean of the data and dividing by the standard deviation; normalization makes the features have similar scale by scaling the values ​​of the data to the range between 0 and 1.

[0087] Correlation fitting module: used to select the value of the data categories in the crystallization influencing parameter set and the equipment control parameter set, and obtain k data categories that affect the degree of crystallization at each time point;

[0088] The value selection for the data categories in the crystallization influencing parameter set and the equipment control parameter set includes:

[0089] At the current time point, the real-time data vector I is constructed according to a certain sequence using the values ​​corresponding to the data categories in the crystallization influence parameter set and the device control parameter set.t =[x1,...,xo,...,xm], xo represents the sample value of the oth data category, m represents the number of data categories, and o=1,...,m;

[0090] Collect the crystallization degree of fructose syrup at the current time point, obtain the target variable yt, import the target variable yt and the real-time data vector into the best association analysis model, obtain the regression coefficient corresponding to each data category, arrange the absolute values ​​of the regression coefficients in descending order, select the data category features corresponding to the first k regression coefficients as the main factors affecting the crystallization degree, and complete the feature extraction.

[0091] Use the optimal λ value to fit the association analysis model and obtain the regression coefficient of each data category. The importance of the feature (data category) is judged according to the size of the regression coefficient. The larger the absolute value of the regression coefficient, the greater the influence of the feature (data category) on the crystallization behavior. Sort and select k data categories with larger absolute values ​​of regression coefficients as important features. Suppose we have collected parameter data (temperature, concentration, pH value, impurity content, number of microorganisms, solute composition, solvent addition, pressure, stirring speed, cooling speed, gas phase flow rate, filtration rate), standardized these data, and fitted them through the association analysis model to obtain the following regression coefficients:

[0092]

[0093] According to the absolute value sorting of the regression coefficients, we can select the top four features that have the greatest impact on crystallization behavior, as follows: concentration (1.2), gas flow rate (1.0), solute composition (0.9), temperature (0.8), and pressure (0.7), thereby extracting data categories from the crystallization influencing parameter set and the equipment control parameter set, allocating more resources for the processing and deep mining of high-value features, reducing the interference of unnecessary features, and improving the accuracy and practicality of the analysis.

[0094] The method for obtaining the value of k includes: taking the current time point as an anchor point, tracing back s time points to collect the values ​​of m data categories, and extracting the crystallization degree corresponding to the fructose-glucose syrup at each time point;

[0095] At each time point c, by calculating the value coefficient The number of data categories selected when the collection value coefficient is the highest, and then the number of data categories lc with the highest correlation with the degree of crystallization among the m data categories is obtained.

[0096] Extract the number of data categories obtained at each time point and calculate the average Then the value of k at the current time point is

[0097] The specific calculation process of the value coefficient includes:

[0098] Among the m data categories, randomly select l 1 The values ​​corresponding to the data categories are the remaining ml 1 The data categories are numerically standardized, the corresponding degree of crystallization of high fructose corn syrup under different selections are collected, the value coefficients of different quantities and different data categories are traversed and calculated, the number of data categories corresponding to the highest value coefficient value is extracted, and then the number of data categories with the highest correlation with the degree of crystallization at each time point is obtained.

[0099] For example, suppose that at a certain time point, there are three data categories, namely temperature, concentration, and pH value, and the corresponding values ​​are 57, 78%, and 8, respectively. The standardized values ​​of temperature, concentration, and pH value are set to 50, 70%, and 6. Then, suppose the acquired data is:

[0100] temperature concentration pH Degree of crystallinity Value coefficient 57 78% 8 0.61 0.67 57 70% 8 0.56 0.68 50 70% 8 0.57 0.57 57 78% 6 0.72 0.67 50 78% 6 0.58 0.7 57 70% 6 0.4 0.87 50 78% 8 0.68 0.62

[0101] Then, according to the above table, we can obtain the highest value coefficient of 0.87 at this time point, and the corresponding data category is temperature. Then we can obtain the number 1 with the highest correlation with the degree of crystallization among the three data categories, so k=1 at this time point.

[0102] Numerical standardization can also be understood as numerical initialization. The intention of the present invention is to perform numerical standardization for m data categories. If a data type is selected, the corresponding standardized value is replaced by the value of the selected data type at a time point, thereby obtaining the value of different quantitative variables (the value of the selected data type at a time point) for the degree of crystallinity.

[0103] The method for obtaining the best association analysis model specifically includes:

[0104] Collect q groups of crystallization influencing parameter sets and equipment control parameter sets, and traverse the numbers according to the sequence of data categories in the real-time data vector, mark the crystallization degree of fructose syrup in each group corresponding to each number, yi represents the crystallization degree of fructose syrup corresponding to the i-th group, and the crystallization degree of fructose syrup is used as the target variable to construct the target matrix Constructing the sample matrix All data categories of the crystallization influencing parameter set and the equipment control parameter set are taken as the rows of the matrix, and the numbers of each group are taken as the columns, wherein xij represents the sample value of the jth data category of the i-th group, m represents the number of data categories, and i=1, ..., m, j=1, ..., n;

[0105] Build a correlation analysis model, Among them, β0 is the intercept term, βj is the regression coefficient of the jth parameter, and λ is the regularization parameter used to control the strength of the regularization term;

[0106] Initialize the parameters in the association analysis model, import w groups of target matrices and corresponding sample matrices into the association analysis model in batches, train the association analysis model, and set the constraints as The best association analysis model is obtained until the value of the constraint condition is minimized and no longer changes.

[0107] Control instruction generation module: used to construct the function expression of the characteristic parameter matrix and the path score, and take the numerical combination with the minimum path score as the optimal control parameter set; adjust the fructose corn syrup processing according to the optimal control parameter set as the processing instruction;

[0108] The modules are connected to each other via wired and / or wireless means.

[0109] Define a characteristic parameter matrix and a crystallization target; the crystallization target is to find a set of control parameters with the minimum degree of crystallization;

[0110] Define the value z of the pth row of the oth data category in the feature parameter matrix op Start from, select a value in each data category, traverse k data categories to form a value combination, and calculate the path score FG of each value combination;

[0111] Set the limit function to W FG v≤Wmax, where W FG v represents the economic value consumed when processing control is performed according to the vth path score FG, Wmax represents the maximum economic value of consumption limited by the enterprise, and when the restriction function is satisfied, the numerical combination with the smallest path score FG is selected as the optimal control parameter set.

[0112] The method for constructing a characteristic parameter matrix includes: obtaining the data category values ​​corresponding to the first k regression coefficients, discretizing the parameter processing control range according to the parameter processing control range of each data category and taking the adjustment accuracy as the interval to obtain the candidate parameters, taking the value at the current time point as the middle value, and selecting the candidate parameters in turn by offsetting d distances to the left and right, and then there are a total of 2d+1 values ​​selected for each data category, arranging the values ​​in descending order, with the data categories as rows and the number of values ​​as columns, to construct a k×(2d+1) characteristic parameter matrix.

[0113] The construction of the path score FG includes:

[0114] The function expression of the path score is set as FG = ω0 + ω1z 1f +......+ω k z ke +ε, where ω0 represents the initial weight value, ω vcrepresents the weight coefficient corresponding to the vcth data category, ε represents the error term, and vc=1,……,k,

[0115] At the current time point, collect the data category values ​​and crystallinity corresponding to k regression coefficients, and define ω vc The value range of ω0 and ε can be selected according to the specific situation. The grid search method is used to traverse within the given value range. For each set of ω vc , ω0 and ε, calculate the absolute difference between the degree of crystallization predicted by the function expression and the degree of crystallization at the current time point, and use cross-validation to train the model of the function expression.

[0116] Choose the ω with the smallest absolute difference vc The values ​​of , ω0 and ε are taken as the optimal solution, and the optimal solution is imported into the function expression of the path score as the final function expression.

[0117] In this embodiment, the present invention constructs an association analysis model, selects the data category features corresponding to the first k regression coefficients at the current time point as the main factors affecting the degree of crystallization, extracts the main factors affecting the crystallization of fructose syrup in real time, conducts targeted analysis and in-depth mining on the main data category factors, reduces the waste of resources on irrelevant data, concentrates system resources on high-value associated data, and optimizes data. By constructing a feature parameter matrix, traversing k data categories to form a numerical combination, calculating the path score of each numerical combination, selecting the numerical combination with the smallest path score as the optimal control parameter set, and adjusting the processing instructions according to the optimal control parameter set. In the processing of high fructose corn syrup, the best control parameter output is achieved under the condition of economic benefit constraints. It not only performs high-quality decomposition and mining of characteristic data, improves the effect of data processing and model training, as well as the retrieval and response speed, and provides an accurate data basis for subsequent model optimization; it also injects new impetus into the agility and flexibility of data processing. The system implements path scoring based on the current time point, predicts the control parameters at the next time point, and has an adaptive regulation and optimization mechanism, which continuously improves the overall efficiency of intelligent analysis. Multiple modules work together to realize comprehensive data collection, processing and analysis, and provide real-time and accurate data analysis in anti-crystallization control of high fructose corn syrup.

[0118] Embodiment 2

[0119] Reference Figure 1 and Figure 2 , Example 2 further illustrates an anti-crystallization fructose syrup processing control system based on data analysis proposed by the present invention.

[0120] The connector intelligent monitoring and control method is applied to a crystallization-resistant fructose-glucose syrup processing control system based on data analysis, and comprises the following steps:

[0121] Step S1: collecting the values ​​corresponding to each data category in q groups of crystallization influencing parameter sets and equipment control parameter sets, constructing a target matrix and a sample matrix, and constructing an association analysis model, and using the target matrix and the sample matrix to train the association analysis model to obtain the best association analysis model;

[0122] Step S2: collecting the real-time data vector and target variable at the current time point, importing the target variable and the real-time data vector into the best association analysis model, and selecting the data category values ​​corresponding to the first k regression coefficients;

[0123] Step S3: construct a feature parameter matrix, traverse k data categories to form a numerical combination, calculate the path score of each numerical combination, and select the numerical combination with the smallest path score as the optimal control parameter set;

[0124] Step S4: adjusting the fructose-glucose syrup processing according to the optimal control parameter set as the processing instruction.

[0125] The anti-crystallization high fructose syrup processing control system based on data analysis includes:

[0126] Data acquisition module: used to collect data during the processing of high fructose corn syrup, and perform preliminary processing to obtain the crystallization influencing parameter set and equipment control parameter set;

[0127] Correlation fitting module: used to select the value of the data categories in the crystallization influencing parameter set and the equipment control parameter set, and obtain k data categories that affect the degree of crystallization at each time point;

[0128] Control instruction generation module: used to construct the function expression of the characteristic parameter matrix and the path score, and use the numerical combination with the minimum path score as the optimal control parameter set; adjust the fructose syrup processing according to the optimal control parameter set as the processing instruction.

[0129] The modules are connected to each other via wired and / or wireless means.

[0130] In addition, according to the embodiment of the present application, the process described in the attached figure of a control system for processing anti-crystallization high fructose syrup based on data analysis can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided in the present application. Of course, the architecture shown in the attached figure of a control system for processing anti-crystallization high fructose syrup based on data analysis is only exemplary. When implementing different devices, adaptive selection or adjustment can be performed according to actual needs.

[0131] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0132] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A crystallization-resistant fructose syrup processing control system based on data analysis, characterized in that: The anti-crystallization fructose syrup processing control system based on data analysis includes: Data acquisition module: used to collect data during the processing of high fructose corn syrup, and perform preliminary processing to obtain the crystallization influencing parameter set and equipment control parameter set; The data categories of the crystallization-affecting parameter set include the temperature, concentration, pH value, impurity content, microbial count, solute composition, additive data and pressure data of the fructose-glucose solution; The data categories of the equipment control parameter set include stirring data, cooling data, gas phase data and filtration data; Initial treatment includes: Identify non-numerical information in solute composition, additive data and gas control, perform binary conversion on the solute category, additive category and gas category respectively through unique hot encoding, use binary strings to represent them, extract the numerical values ​​corresponding to the solute proportion, additive addition amount and gas flow value, and fill them into the corresponding binary strings, and set the representation of solute composition, additive data and gas control as binary string + numerical value; Standardize and normalize the data to make them have the same dimension, and form a crystallization influencing parameter set and a device control parameter set; Correlation fitting module: used to select the value of the data categories in the crystallization influencing parameter set and the equipment control parameter set, and obtain k data categories that affect the degree of crystallization at each time point; Control instruction generation module: used to construct the function expression of the characteristic parameter matrix and the path score, and take the numerical combination with the minimum path score as the optimal control parameter set; adjust the fructose corn syrup processing according to the optimal control parameter set as the processing instruction; The modules are connected to each other via wired and / or wireless means.

2. The anti-crystallization high fructose syrup processing control system based on data analysis according to claim 1, characterized in that: The value selection of the data categories in the crystallization influencing parameter set and the equipment control parameter set includes: At the current time point, the real-time data vector I is constructed according to a certain sequence using the values ​​corresponding to the data categories in the crystallization influence parameter set and the device control parameter set. t =[x1,…,xo,…,xm], xo represents the sample value of the oth data category, m represents the number of data categories, and o=1,…,m; Collect the crystallization degree of fructose syrup at the current time point, obtain the target variable yt, import the target variable yt and the real-time data vector into the best association analysis model, obtain the regression coefficient corresponding to each data category, arrange the absolute values ​​of the regression coefficients in descending order, select the data category features corresponding to the first k regression coefficients as the main factors affecting the crystallization degree, and complete the feature extraction.

3. The anti-crystallization high fructose syrup processing control system based on data analysis according to claim 2, characterized in that: The method for obtaining the value of k includes: taking the current time point as an anchor point, tracing back s time points to collect the values ​​of m data categories, and extracting the crystallization degree corresponding to the fructose-glucose syrup at each time point; At each time point c, by calculating the value coefficient The number of data categories selected when the acquisition value coefficient is the highest, and then the number of data categories lc with the highest correlation with the degree of crystallization among the m data categories is obtained; Extract the number of data categories obtained at each time point and calculate the average Then the value of k at the current time point is 4. The anti-crystallization high fructose syrup processing control system based on data analysis according to claim 3, characterized in that: The specific calculation process of the value coefficient includes: Among the m data categories, randomly select l 1 The values ​​corresponding to the data categories are the remaining ml 1 The data categories are numerically standardized, the corresponding degree of crystallization of high fructose corn syrup under different selections are collected, the value coefficients of different quantities and different data categories are traversed and calculated, the number of data categories corresponding to the highest value coefficient value is extracted, and then the number of data categories with the highest correlation with the degree of crystallization at each time point is obtained.

5. The anti-crystallization high fructose syrup processing control system based on data analysis according to claim 4, characterized in that: The method for obtaining the optimal association analysis model specifically includes: Collect q groups of crystallization influencing parameter sets and equipment control parameter sets, and traverse the numbers according to the sequence of data categories in the real-time data vector, mark the crystallization degree of fructose syrup in each group corresponding to each number, yi represents the crystallization degree of fructose syrup corresponding to the i-th group, and the crystallization degree of fructose syrup is used as the target variable to construct the target matrix Constructing the sample matrix All data categories of the crystallization influencing parameter set and the equipment control parameter set are taken as the rows of the matrix, and the numbers of each group are taken as the columns, wherein xij represents the sample value of the jth data category of the i-th group, m represents the number of data categories, and i=1, ..., n, j=1, ..., m; Build a correlation analysis model, Among them, L(β) refers to the regression coefficient of the model output, β0 is the intercept term, βj is the regression coefficient of the jth parameter, and λ is the regularization parameter used to control the strength of the regularization term; Initialize the parameters in the association analysis model, import w groups of target matrices and corresponding sample matrices into the association analysis model in batches, train the association analysis model, and set the constraints as The best association analysis model is obtained until the value of the constraint condition is minimized and no longer changes.

6. The anti-crystallization high fructose syrup processing control system based on data analysis according to claim 5, characterized in that: The construction of the characteristic parameter matrix includes: Define a characteristic parameter matrix and a crystallization target; the crystallization target is to find a set of control parameters with the minimum degree of crystallization; Define the value z of the pth row of the oth data category in the feature parameter matrix op Start from, select a value in each data category, traverse k data categories to form a value combination, and calculate the path score FG of each value combination; Set the limit function to W FG v≤W max, where W FG v represents the economic value consumed when processing control is performed according to the vth path score FG, W max represents the maximum economic value of consumption limited by the enterprise, and when the restriction function is satisfied, the numerical combination with the smallest path score FG is selected as the optimal control parameter set.

7. The anti-crystallization high fructose syrup processing control system based on data analysis according to claim 6, characterized in that: The method for constructing the characteristic parameter matrix includes: obtaining the data category values ​​corresponding to the first k regression coefficients, discretizing the parameter processing control range according to the parameter processing control range of each data category and taking the adjustment accuracy as the interval to obtain the candidate parameters, taking the value at the current time point as the middle value, and selecting the candidate parameters in turn by offsetting d distances to the left and right, and then there are a total of 2d+1 values ​​selected for each data category, arranging the values ​​in descending order, with the data categories as rows and the number of values ​​as columns, to construct a k×(2d+1) characteristic parameter matrix.

8. The anti-crystallization high fructose syrup processing control system based on data analysis according to claim 7, characterized in that: The construction of the path score FG includes: The function expression of the path score is set as FG = ω0 + ω1z 1f +…+ω vc z ve +...+ω k z ke +ε, where ω0 represents the initial weight value, ω vc represents the weight coefficient corresponding to the vc-th data category, ε represents the error term, and vc=1,...,k; At the current time point, collect the data category values ​​and crystallinity corresponding to k regression coefficients, and define ω vc The value range of ω0 and ε can be selected according to the specific situation. The grid search method is used to traverse within the given value range. For each set of ω vc , ω0 and ε, calculate the absolute difference between the degree of crystallization predicted by the function expression and the degree of crystallization at the current time point, and use cross-validation to train the model of the function expression; Choose the ω with the smallest absolute difference vc The values ​​of , ω0 and ε are taken as the optimal solution, and the optimal solution is imported into the function expression of the path score as the final function expression.

Citation Information

Patent Citations

  • Cane preparation unit and method of operation

    CN103562938A

  • Energy-saving production method based on sugar factory refining system

    CN103592930A

  • DCS control system for producing anhydrous crystalline fructose by using sugarcanes as raw materials

    CN105242652A

  • Purification method for HFCS (high fructose corn syrup) in fructose production process

    CN105256079A

  • Control system and method for inducing hexafluorophosphate to crystallize by ultrasonic waves

    CN116036639A