Nondestructive testing and proportioning optimization method based on dough components

Abstract

The invention relates to the technical field of food processing, and particularly discloses a dough component-based nondestructive testing and proportioning optimization method, which comprises the following steps: acquiring a finished product target of a to-be-detected dough, and determining a finished product category, finished product proportioning data, finished product evaluation data, dough detection data and detection component data of the to-be-detected dough; acquiring historical batch data of multiple times of historical production, calculating a parameter optimization factor of each evaluation parameter of each time of historical production, and calculating a historical reference value of each time of historical production of the to-be-detected dough; determining a reference production set and component optimization data of the to-be-detected dough; and determining a detection label, an optimization label and optimization component data of the to-be-detected dough, and carrying out batch production. According to the method, closed-loop optimization from component detection to production landing can be realized, detection compliance and matching suitability are taken into consideration, the market suitability of finished products is improved, the scientificity, pertinence and efficiency of matching optimization are improved, and the quality stability and market competitiveness of the produced finished products are ensured.

Description

Technical Field

[0001] This invention relates to the field of food processing technology, and in particular to a method for non-destructive detection and ratio optimization of dough components. Background Technology

[0002] In the baking food processing industry, the proportions of dough ingredients directly determine the taste, stability, and market acceptance of the finished product. With the advancement of industrialized large-scale production and the increasing demands of consumers for consistent quality, dough ingredient testing and proportion optimization have become core technological requirements in the industry. Traditional dough ingredient testing often relies on destructive methods such as the Kjeldahl method and manual gluten washing, which not only damage the original processing properties of the dough, making test samples unusable, but also have long testing cycles and large errors, making it difficult to adapt to the needs of real-time production quality control. Although non-destructive testing technologies such as near-infrared spectroscopy and ultrasonic testing equipment are gradually being applied, they mostly focus on the detection of single basic parameters, and the data is disconnected from the proportion optimization process, failing to directly support adjustments. Moreover, these technologies do not utilize intelligent sensing systems, relying only on single types of testing equipment, and do not fully leverage the advantages of multi-parameter collaborative detection by intelligent sensors.

[0003] In terms of ingredient ratio optimization, early adjustments relied on manual experience, which was highly subjective and resulted in significant batch-to-batch quality fluctuations. Subsequent introductions of intelligent dough mixing equipment and constant-temperature fermentation chambers only achieved precise control of single parameters, failing to consider the interactions of multiple components. Furthermore, existing data-driven optimization methods are largely limited to production process parameters, failing to effectively integrate end-to-end information such as intelligent sensor detection data, historical production records, finished product sales data, and consumer feedback, leading to a disconnect between optimization direction and market demand. Simultaneously, issues such as the separation of detection and optimization processes, the lack of intelligent sensing system applications, and the fragmented storage and low utilization rate of historical data result in high raw material loss rates and insufficient stability in finished product quality, making it difficult to meet the standardized and precise production requirements of the modern food industry.

[0004] Therefore, this invention proposes a method based on non-destructive detection and proportion optimization of dough components. Summary of the Invention

[0005] This invention provides a method for non-destructive testing and proportion optimization of dough components. By acquiring the target finished product data and historical batch data from multiple production runs of the dough to be tested, the method determines the finished product category, proportion data, evaluation data, dough testing data, and component data. It calculates the parameter optimization factor for each evaluation parameter from each historical production run and calculates historical reference values ​​for each production run of the dough to be tested. This determines the reference production set, component optimization data, testing labels, optimization labels, and optimized component data for the dough to be tested. Based on the optimized component data, batch production of the dough to be tested is then carried out. This method achieves closed-loop optimization from component testing to production implementation, identifying core influencing components and clarifying the optimization direction. It balances testing compliance and proportion suitability, improves the market adaptability of the finished product, enhances the scientific rigor, relevance, and efficiency of proportion optimization, and ensures the quality stability and market competitiveness of the finished product.

[0006] This invention provides a method for non-destructive detection and proportion optimization of dough components, including: S1: Obtain the target finished product of the dough to be tested, determine the finished product category, finished product ratio data, finished product evaluation data and dough test data of the dough to be tested, and determine the test component data of the dough based on the dough test data and finished product ratio data of the dough to be tested; S2: Obtain historical batch data of the dough to be tested from multiple historical productions, analyze the historical batch data to determine the set of associated components for each evaluation parameter, calculate the parameter optimization factor for each evaluation parameter in each historical production, and calculate the historical reference value for each historical production of the dough to be tested. S3: Based on the historical reference values ​​of all historical productions of the dough to be tested, determine the reference production set of the dough to be tested. Based on the reference production set of the dough to be tested and the associated component set of each evaluation parameter, determine the component optimization data of the dough to be tested. S4: Based on the test component data, finished product ratio data, and component optimization data of the dough to be tested, determine the test label, optimization label, and optimized component data of the dough to be tested, and carry out batch production of the dough to be tested based on the optimized component data.

[0007] Preferably, a method based on non-destructive testing and proportion optimization of dough components obtains the target finished product of the dough to be tested, and determines the finished product category, finished product proportion data, finished product evaluation data, and dough testing data of the dough to be tested, including: Obtain the target finished product of the dough to be tested, and determine the finished product category, finished product ratio data, and finished product evaluation data based on the finished product target. The finished product ratio data includes multiple component parameters and the target parameter range for each component parameter. The finished product evaluation data includes multiple evaluation parameters and the evaluation weight for each evaluation parameter. The dough testing data is determined based on the finished product category of the dough to be tested. The dough testing data includes multiple testing parameters and the testing equipment label for each testing parameter.

[0008] Preferably, a method for non-destructive testing and proportion optimization of dough components determines the component data of the dough to be tested based on the dough testing data and the finished product proportion data, including: Based on all the testing device tags in the dough testing data, non-destructive testing is performed on the dough to be tested to determine the testing parameter data of the dough to be tested. The testing parameter data includes multiple testing parameters and the testing parameter value of each testing parameter. Map all detection parameters in the detection parameter data and all component parameters in the finished product ratio data to determine the mapping detection parameter or mapping detection parameter set for each component parameter in the finished product ratio data. The mapping detection parameter set includes at least two or more detection parameters. The detection parameter value of the mapping detection parameter of each component parameter in the finished product ratio data or the detection parameter value of all detection parameters in the mapping detection parameter set is input into the parameter mapping model, and the component value of each component parameter in the finished product ratio data is determined based on the output of the parameter mapping model. Based on the component values ​​of all component parameters in the finished product formulation data, the component data for the dough to be tested are determined.

[0009] Preferably, a method based on non-destructive testing and proportion optimization of dough components includes historical batch data such as batch production data, batch component data, finished product sales data, and finished product feedback data. The batch production data includes production time, production cost, production volume, and rework volume. The batch component data includes multiple component parameters and historical parameter values ​​for each component parameter. The finished product sales data includes sales time intervals, sales prices, actual daily sales prices and sales volumes within the sales time intervals, and finished product feedback data includes the consumption times of multiple consumers and evaluation descriptions based on each evaluation parameter in the finished product evaluation data.

[0010] Preferably, a method based on non-destructive testing and proportion optimization of dough components analyzes historical batch data to determine the associated component set for each evaluation parameter, calculates the parameter optimization factor for each evaluation parameter in each historical production, and calculates the historical reference value for each historical production of the dough to be tested, including: In the feedback data of finished products from each historical production of the dough to be tested, word sense recognition is performed on the evaluation sub-description of each evaluation parameter in the finished product evaluation data for each consumer in the finished product evaluation data. The evaluation labels are determined for each consumer in the feedback data of finished products from each historical production of the dough to be tested, based on each evaluation parameter in the finished product evaluation data. The evaluation labels include very weak, weak, moderate, strong, and very strong. The evaluation labels of each consumer based on each evaluation parameter in the finished product evaluation data in each historical production batch are quantified to determine the evaluation quantification value of each consumer based on each evaluation parameter in the finished product feedback data in each historical production batch. Pearson correlation analysis was performed on the quantitative evaluation values ​​of all consumers based on all evaluation parameters in the finished product feedback data of all historical productions and the historical parameter values ​​of all component parameters in the batch component data to determine the Pearson correlation value of each evaluation parameter and each component parameter. Based on the Pearson correlation values ​​of each evaluation parameter and all component parameters, as well as the preset correlation threshold, the set of associated components for each evaluation parameter is determined, wherein the set of associated components includes at least one or more component parameters. Based on the finished product evaluation data of the dough to be tested, the batch production data of each historical production, the finished product sales data, and the finished product feedback data, the quantitative evaluation values ​​of all consumers based on all evaluation parameters are calculated in the finished product feedback data of each historical production of the dough to be tested, and the parameter optimization factor of all consumers based on each evaluation parameter is calculated. The historical reference value of each historical production of the dough to be tested is also calculated.

[0011] Preferably, a method for non-destructive testing and proportion optimization of dough components involves determining a reference production set for the dough based on historical reference values ​​from all previous production runs, and determining component optimization data for the dough based on the reference production set and the associated component set for each evaluation parameter. This includes: Sort the historical reference values ​​of all historical productions of the dough to be tested from largest to smallest, and select the historical productions corresponding to the first specified number of historical reference values ​​after sorting as the reference production set of the dough to be tested. Based on the batch composition data of each historical production in the reference production set, historical reference values, parameter optimization factors of all evaluation parameters, the associated composition set of each evaluation parameter, and the composition optimization model, the optimization parameter range and optimal parameter value of each composition parameter are determined. Based on the optimized parameter range and optimal parameter values ​​of all component parameters, the optimized component data of the dough to be tested are determined.

[0012] Preferably, a method for non-destructive testing and proportion optimization of dough components, based on the detected component data, finished product proportion data, and component optimization data of the dough to be tested, determines the test label, optimization label, and optimized component data of the dough to be tested, and performs batch production of the dough to be tested based on the optimized component data, including: The component values ​​of each component parameter in the component data of the dough to be tested and the target parameter range of each component parameter in the component optimization data are judged. If the component values ​​of all component parameters are within the corresponding target parameter range, the test label of the dough to be tested is determined to be qualified. If the component value of any component parameter is not within the corresponding target parameter range, the test label of the dough to be tested is determined to be unqualified. If the test label of the dough to be tested is unqualified, the dough to be tested will be optimized to meet the qualified ratio based on all the component parameters in the test component data that are not within the target parameter range; If the test label of the dough to be tested is qualified, based on the test component data and component optimization data, the sales ratio of the dough to be tested is optimized to determine the optimized component data of the dough to be tested; Based on the optimized composition data of the dough to be tested, batch production of the dough to be tested was carried out.

[0013] Preferably, a method for non-destructive testing and proportion optimization of dough components involves, if the test label of the dough to be tested is qualified, determining the optimized component data of the dough to be tested by optimizing the sales proportion based on the tested component data and component optimization data, including: The component values ​​of each component parameter in the component data of the dough to be tested and the optimized parameter range of each component parameter in the finished product ratio data are judged. If the component values ​​of all component parameters are within the corresponding target parameter range, the optimization label of the dough to be tested is determined to be no optimization required. If the component value of any component parameter is not within the corresponding optimized parameter range, the optimization label of the dough to be tested is determined to be optimization required. If the optimization label of the dough to be tested is that it needs to be optimized, the selling ratio of the dough to be tested is optimized based on the component parameters corresponding to all component values ​​that are not within the optimization parameter range and the optimal parameter value of each component parameter. Based on all optimized component values ​​that are outside the optimized parameter range, and all component values ​​that are within the optimized parameter range, the optimized component data of the dough to be tested is determined.

[0014] The beneficial effects of this invention compared to existing technologies are as follows: By acquiring the target finished product of the dough to be tested and historical batch data from multiple historical productions, the finished product category, finished product ratio data, finished product evaluation data, dough testing data, and tested component data of the dough to be tested are determined. The parameter optimization factor for each evaluation parameter in each historical production is calculated, and the historical reference value for each historical production of the dough to be tested is calculated. This determines the reference production set, component optimization data, testing label, optimization label, and optimized component data for the dough to be tested. Based on the optimized component data, batch production of the dough to be tested is carried out. This enables closed-loop optimization from component testing to production implementation. It can also link multiple intelligent sensors such as near-infrared spectrometers and ultrasonic testing equipment to achieve non-destructive collaborative detection of dough components, breaking through the limitations of traditional single-detection equipment. It accurately acquires multi-dimensional detection parameter data, providing comprehensive and reliable basic data support for subsequent ratio optimization. Combining the advantages of multi-parameter collaborative detection of intelligent sensing systems with the integrated analysis of historical batch data, it further identifies core influencing components, clarifies optimization directions, balances testing compliance and ratio adaptability, improves the market adaptability of the finished product, enhances the scientific, targeted, and efficient nature of ratio optimization, and ensures the quality stability and market competitiveness of the produced product.

[0015] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0016] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a method for non-destructive detection and proportion optimization of dough components in an embodiment of the present invention. Detailed Implementation

[0018] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention. Example 1:

[0019] This invention provides a method for non-destructive detection and proportion optimization of dough components, referencing... Figure 1 ,include: S1: Obtain the target finished product of the dough to be tested, determine the finished product category, finished product ratio data, finished product evaluation data and dough test data of the dough to be tested, and determine the test component data of the dough based on the dough test data and finished product ratio data of the dough to be tested; S2: Obtain historical batch data of the dough to be tested from multiple historical productions, analyze the historical batch data to determine the set of associated components for each evaluation parameter, calculate the parameter optimization factor for each evaluation parameter in each historical production, and calculate the historical reference value for each historical production of the dough to be tested. S3: Based on the historical reference values ​​of all historical productions of the dough to be tested, determine the reference production set of the dough to be tested. Based on the reference production set of the dough to be tested and the associated component set of each evaluation parameter, determine the component optimization data of the dough to be tested. S4: Based on the test component data, finished product ratio data, and component optimization data of the dough to be tested, determine the test label, optimization label, and optimized component data of the dough to be tested, and carry out batch production of the dough to be tested based on the optimized component data.

[0020] In this embodiment, the target finished product of the dough to be tested is first obtained, and the corresponding finished product category, finished product ratio data, finished product evaluation data, and dough test data are determined based on the target. Then, based on the non-destructive testing parameters in the dough test data and the component parameter mapping relationship in the finished product ratio data, the actual values ​​of each component of the dough to be tested are obtained through parameter mapping model transformation, and finally the test component data of the dough to be tested is determined, providing an actual component benchmark for subsequent optimization.

[0021] In this embodiment, historical batch data from multiple production runs of the dough to be tested are first acquired. This data covers multi-dimensional information such as production process, market sales, and consumer feedback. By analyzing the quantitative values ​​of finished product feedback and batch composition data in the historical batch data, the set of associated components corresponding to each evaluation parameter is determined, identifying the core components affecting the quality of the finished product. Simultaneously, the parameter optimization factor for each evaluation parameter in each historical production run is calculated, indicating the direction and magnitude of component adjustments. Combined with multi-dimensional data, historical reference values ​​for each historical production run are calculated to quantify the overall performance of that batch.

[0022] In this embodiment, the historical reference values ​​of all historical productions are first sorted, and a specified number of historical productions with the best performance are selected as a reference production set to ensure the reliability of the optimization basis. Then, combining the batch component data, historical reference values, parameter optimization factors, and the associated component set of each evaluation parameter in this reference production set, the optimized parameter range and optimal parameter value of each component parameter are determined through component optimization model analysis and calculation, forming the component optimization data of the dough to be tested, and clarifying the optimization target and boundary.

[0023] In this embodiment, the detected component data is first compared with the optimized component data of the finished product formulation to determine whether the test label of the dough to be tested is qualified and whether the optimized label needs optimization. If the test label is qualified and the optimized label needs optimization, then the component parameters that are outside the optimization range are adjusted based on the optimal parameter values ​​to finally determine the optimized component data. Finally, the optimized component data is used as the basis for production to carry out batch production of the dough to be tested, ensuring that the produced finished product meets both quality requirements and market demand.

[0024] The beneficial effects of the above technology are as follows: By acquiring the target finished product data of the dough to be tested and historical batch data from multiple production runs, the finished product category, finished product ratio data, finished product evaluation data, dough testing data, and tested component data of the dough to be tested can be determined. The parameter optimization factor for each evaluation parameter in each historical production run can be calculated, and the historical reference value for each historical production run of the dough to be tested can be calculated. This determines the reference production set, component optimization data, testing labels, optimization labels, and optimized component data for the dough to be tested. Based on the optimized component data, batch production of the dough to be tested can be carried out. This enables closed-loop optimization from component testing to production implementation, identifying core influencing components and clarifying the optimization direction, balancing testing compliance and ratio suitability, improving the market adaptability of the finished product, enhancing the scientific, targeted, and efficient nature of ratio optimization, and ensuring the quality stability and market competitiveness of the produced product. Example 2:

[0025] Based on Example 1, a method for non-destructive testing and proportion optimization of dough components is provided to obtain the target finished product of the dough to be tested, and to determine the finished product category, finished product proportion data, finished product evaluation data, and dough testing data of the dough to be tested, including: Obtain the target finished product of the dough to be tested, and determine the finished product category, finished product ratio data, and finished product evaluation data based on the finished product target. The finished product ratio data includes multiple component parameters and the target parameter range for each component parameter. The finished product evaluation data includes multiple evaluation parameters and the evaluation weight for each evaluation parameter. The dough testing data is determined based on the finished product category of the dough to be tested. The dough testing data includes multiple testing parameters and the testing equipment label for each testing parameter.

[0026] In this embodiment, the target finished product of the dough to be tested is first obtained. Based on this target, the product category of the dough to be tested is determined, which is a clear classification of the product type. Simultaneously, the finished product formulation data is determined. This data includes multiple component parameters, which are the indicators of various raw materials that constitute the dough. Each component parameter corresponds to a target parameter range, which serves as the baseline for the amount of raw materials used to ensure the finished product meets the target. Furthermore, finished product evaluation data is determined. This data includes multiple evaluation parameters, which are specific dimensions for measuring whether the finished product meets the standards. Each evaluation parameter has an evaluation weight, reflecting the importance of different evaluation dimensions in the finished product evaluation.

[0027] In this embodiment, the ingredient parameters include flour, water, yeast, salt, sugar, fat, eggs, and sodium bicarbonate residue. Evaluation parameters include hardness, sweetness, saltiness, chewiness, elasticity, and crispness.

[0028] In this embodiment, different product categories have different core characteristics and quality control points, resulting in different dough testing data. The dough testing data includes multiple testing parameters, which are indicators that require non-destructive measurement to control dough quality. For example, bread requires testing for protein content, fermentation gas production rate, and water activity; cakes require testing for water activity, emulsification degree, and flour particle size; and whole wheat steamed buns require testing for whole wheat flour percentage, moisture content, and starch gelatinization degree. Each testing parameter corresponds to a non-destructive testing equipment label, which indicates the non-destructive instrument used to perform the parameter test. For example, protein content is measured using a near-infrared spectrometer, fermentation gas production rate using a closed-loop non-destructive fermentation gas generator, water activity using a non-contact water activity meter, emulsification degree using a near-infrared spectrometer, and flour particle size using a laser particle size analyzer.

[0029] The beneficial effects of the above technologies are: obtaining the target finished product of the dough to be tested, determining the finished product category, finished product ratio data, finished product evaluation data, and dough test data of the dough to be tested, which can provide accurate data benchmarks for subsequent optimization. Example 3:

[0030] Based on Example 2, a method for non-destructive detection and proportion optimization of dough components is provided. This method determines the component data of the dough to be tested based on dough detection data and finished product proportion data, including: Based on all the testing device tags in the dough testing data, non-destructive testing is performed on the dough to be tested to determine the testing parameter data of the dough to be tested. The testing parameter data includes multiple testing parameters and the testing parameter value of each testing parameter. Map all detection parameters in the detection parameter data and all component parameters in the finished product ratio data to determine the mapping detection parameter or mapping detection parameter set for each component parameter in the finished product ratio data. The mapping detection parameter set includes at least two or more detection parameters. The detection parameter value of the mapping detection parameter of each component parameter in the finished product ratio data or the detection parameter value of all detection parameters in the mapping detection parameter set is input into the parameter mapping model, and the component value of each component parameter in the finished product ratio data is determined based on the output of the parameter mapping model. Based on the component values ​​of all component parameters in the finished product formulation data, the component data for the dough to be tested are determined.

[0031] In this embodiment, the dough to be tested is tested using the corresponding non-destructive testing instruments according to all the testing device labels in the dough testing data. This process does not damage the original structure and processing performance of the dough. After the test is completed, the test parameter data of the dough to be tested is obtained. The test parameter data includes multiple test parameters and the actual measured test parameter value of each parameter.

[0032] In this embodiment, all detection parameters in the detection parameter data and all component parameters in the finished product formulation data are correlated. This correlation process is called mapping. Mapping needs to match the core attributes of the component parameters with the characterization capabilities of the detection parameters: if the key features of a component parameter can be directly quantified by a single detection parameter, then it corresponds to a single mapped detection parameter; if multiple dimensions of detection parameters are needed to comprehensively reflect its function or state, then it corresponds to a set of mapped detection parameters. For example, the core quality of flour is directly related to its protein content, and its mapped detection parameter is protein content; the core function of yeast is fermentation and gas production, and its mapped detection parameter is the fermentation and gas production rate; the state of water in dough is reflected by water activity, and its mapped detection parameter is water activity; although there is no direct corresponding detection parameter for salt, a set of mapped detection parameters can be formed by combining protein content and fermentation and gas production rate; the role of sugar is related to water activity, fermentation and gas production rate, and gluten network strength, and the set of mapped detection parameters is water activity, fermentation and gas production rate, and gluten network strength.

[0033] In this embodiment, the detection parameter value of the mapping detection parameter corresponding to each component parameter in the finished product formulation data, or the detection parameter value of all detection parameters in the mapping detection parameter set, is input into a pre-constructed parameter mapping model. The parameter mapping model performs calculations based on the input detection parameter values, and the specific component value of each component parameter in the finished product formulation data can be determined based on the model's output.

[0034] In this embodiment, the parameter mapping model is based on a large amount of historical production data. Through machine learning algorithms, such as multiple linear regression, random forest, and neural networks, it mines the inherent correlation between detection parameters and component parameters. Whether it is a direct linear correspondence between a single detection parameter and a component parameter, or a nonlinear coupling relationship between a set of multiple detection parameters and a component parameter, the model will learn and solidify this mapping logic through training. When a new detection parameter value or set of detection parameters is input, the model will accurately output the corresponding component parameter quantification value according to the learned correlation rules, realizing the efficient conversion from non-destructive testing data to component parameters and adapting to the representation needs of different component parameters.

[0035] In this embodiment, the component values ​​of all component parameters in the finished product formulation data are collected, and these component values ​​are integrated to form the test component data of the dough to be tested. This data fully reflects the actual situation of each component of the dough to be tested.

[0036] The beneficial effects of the above technology are as follows: Based on the dough detection data and finished product ratio data of the dough to be tested, the detection component data of the dough to be tested can be determined, which can improve the comprehensiveness and accuracy of the detection component data and provide reliable data support for subsequent ratio optimization. Example 4:

[0037] Based on Example 1, a method for non-destructive testing and proportion optimization of dough components is proposed. Historical batch data includes: batch production data, batch component data, finished product sales data, and finished product feedback data. Among them, batch production data includes production time, production cost, production volume, and rework volume; batch component data includes multiple component parameters and historical parameter values ​​for each component parameter; finished product sales data includes sales time interval, sales price, actual daily sales price and sales volume within the sales time interval; and finished product feedback data includes the consumption time of multiple consumers and the evaluation description of each evaluation parameter based on the finished product evaluation data.

[0038] In this embodiment, historical batch data is a systematic collection and integration of relevant information from the entire production and sales process of each batch of dough in the past. It mainly includes four core data categories: batch production data, batch ingredient data, finished product sales data, and finished product feedback data. Batch production data focuses on key information in the production process, including the specific time of production, the total cost of producing the batch, the total production quantity, and the number of products requiring rework due to quality issues. Batch ingredient data focuses on raw material composition information, including multiple ingredient parameters involved in the batch and the historical parameter values ​​actually used during the production of that batch. Finished product sales data focuses on market sales information, including the sales time range of the batch, the initially set sales price, the actual daily sales price during the entire sales time range, and the corresponding daily sales volume. Finished product feedback data focuses on consumer evaluation information, including the specific consumption times of multiple consumers who purchased the batch and the descriptive evaluations given by consumers for each evaluation parameter in the finished product evaluation data. These evaluations directly reflect the actual consumption experience of the product.

[0039] In this embodiment, the sales time range is determined by the production time and shelf life.

[0040] The beneficial effects of the above technology are: obtaining historical batch data of the dough to be tested from multiple historical productions can provide comprehensive and three-dimensional reference data for subsequent analysis. Example 5:

[0041] Based on Example 4, a method for non-destructive testing and proportion optimization of dough components is proposed. This method analyzes historical batch data to determine the associated component set for each evaluation parameter, calculates the parameter optimization factor for each evaluation parameter in each historical production, and calculates the historical reference value for each historical production of the dough to be tested, including: In the feedback data of finished products from each historical production of the dough to be tested, word sense recognition is performed on the evaluation sub-description of each evaluation parameter in the finished product evaluation data for each consumer in the finished product evaluation data. The evaluation labels are determined for each consumer in the feedback data of finished products from each historical production of the dough to be tested, based on each evaluation parameter in the finished product evaluation data. The evaluation labels include very weak, weak, moderate, strong, and very strong. The evaluation labels of each consumer based on each evaluation parameter in the finished product evaluation data in each historical production batch are quantified to determine the evaluation quantification value of each consumer based on each evaluation parameter in the finished product feedback data in each historical production batch. Pearson correlation analysis was performed on the quantitative evaluation values ​​of all consumers based on all evaluation parameters in the finished product feedback data of all historical productions and the historical parameter values ​​of all component parameters in the batch component data to determine the Pearson correlation value of each evaluation parameter and each component parameter. Based on the Pearson correlation values ​​of each evaluation parameter and all component parameters, as well as the preset correlation threshold, the set of associated components for each evaluation parameter is determined, wherein the set of associated components includes at least one or more component parameters. Based on the finished product evaluation data of the dough to be tested, the batch production data of each historical production, the finished product sales data, and the finished product feedback data, the quantitative evaluation values ​​of all consumers based on all evaluation parameters are calculated in the finished product feedback data of each historical production of the dough to be tested, and the parameter optimization factor of all consumers based on each evaluation parameter is calculated. The historical reference value of each historical production of the dough to be tested is also calculated.

[0042] In this embodiment, for each historical production feedback data of the dough to be tested, the evaluation sub-descriptions given by each consumer for each evaluation parameter in the finished product evaluation data are extracted. These descriptions are analyzed using semantic recognition technology to determine their intensity level, and then each consumer is assigned a corresponding evaluation label for each evaluation parameter. The evaluation labels are uniformly divided into five levels: very weak, weak, moderate, strong, and very strong, achieving standardized classification of qualitative descriptions. For example, if the evaluation parameter is sweetness, and the user's evaluation description includes "not sweet," "very not sweet," or "no sugar added," the evaluation label is very weak; if the user's evaluation description includes "slightly sweet" or "less sugar," the evaluation label is weak; if the user's evaluation description includes "just right," "moderate," or "not too much or too little," the evaluation label is moderate; if the user's evaluation description includes "too much sugar," "a little sweet," or "quite sweet," the evaluation label is strong; and if the user's evaluation description includes "too sweet," "too much sugar," or "very sweet," the evaluation label is very strong.

[0043] In this embodiment, based on the finished product evaluation data of the dough to be tested, the batch production data of each historical production, the finished product sales data, and the quantitative evaluation values ​​of all consumers based on all evaluation parameters in the finished product feedback data of each historical production of the dough to be tested, the parameter optimization factor of all consumers based on each evaluation parameter in the finished product feedback data of each historical production of the dough to be tested is calculated, and the historical reference value of each historical production of the dough to be tested is calculated. The calculation formula is expressed as follows: ; in, This represents the evaluation label of the j-th consumer based on the k-th evaluation parameter in the feedback data of the i-th historical production. This represents the quantitative evaluation value of the j-th consumer based on the k-th evaluation parameter in the feedback data of the finished product from the i-th historical production.

[0044] In this embodiment, Pearson correlation analysis is performed on the two types of data: all consumer evaluation quantification values ​​in the finished product feedback data of all historical productions and the historical parameter values ​​of all component parameters in the batch component data. By analyzing and quantifying the linear correlation between each evaluation parameter and each component parameter, the Pearson correlation value corresponding to each evaluation parameter and each component parameter is finally obtained. The value ranges from [-1, 1], with positive values ​​indicating positive correlation and negative values ​​indicating negative correlation. The larger the absolute value, the stronger the correlation.

[0045] In this embodiment, the preset correlation threshold ranges from 0 to 1, and can be 0.6. The absolute value of the Pearson correlation value of each evaluation parameter with all component parameters is compared with the threshold to filter out component parameters whose correlation value reaches or exceeds the threshold. These component parameters together constitute the correlation component set of the evaluation parameter. Each correlation component set contains at least one component parameter, which clarifies the core raw material indicators that affect each quality evaluation dimension.

[0046] In this embodiment, based on the finished product evaluation data of the dough to be tested, batch production data from multiple historical productions, finished product sales data, and finished product feedback data, the parameter optimization factor for each evaluation parameter of all consumers in the finished product feedback data of each historical production of the dough to be tested is calculated, and the historical reference value of each historical production of the dough to be tested is calculated. The calculation formula is expressed as follows: ; in, This represents the parameter optimization factor for all consumers based on the k-th evaluation parameter in the feedback data of the finished product from the i-th historical production. This represents the historical reference value for the i-th historical production. This represents the feedback evaluation value of the i-th historical production. This represents the consumption time of the j-th consumer in the feedback data of the i-th historical production. Let represent the production time, production cost, production quantity, and rework quantity in the batch production data of the i-th historical production, and let iN1 represent the number of consumers in the finished product feedback data of the i-th historical production. N represents the first-time decay factor, and N2 represents the number of evaluation parameters in the finished product evaluation data. This represents the evaluation weight of the k-th evaluation parameter in the finished product evaluation data. This represents the rework rate of the i-th historical production. This represents the unit cost of the i-th historical production. This represents the actual selling price and sales volume on day a within the selling time interval of the i-th historical production and sales data. N represents the early sales factor of the i-th historical production, and N3 represents the number of days within the sales time interval in the finished product sales data. This represents the second time decay factor. This represents the selling price in the historical sales data of the i-th production. Let represent the time discount factor for the i-th historical production, and 'a' represent the a-th day within the sales time interval.

[0047] In this embodiment, It represents the comprehensive parameter evaluation value of all consumers based on the k-th evaluation parameter in the finished product feedback data of the i-th historical production; In this embodiment, This represents the geometric mean of the early sales factor and the time discount factor of the i-th historical production. It comprehensively reflects the sales speed and sales quality of the i-th historical production. The faster the sales and the closer to the selling price, the larger this value, representing the sales performance of the i-th historical production.

[0048] In this embodiment, This represents the proportion of the unit cost of the i-th historical production to the selling price. The larger the value, the higher the cost relative to the price and the smaller the profit margin.

[0049] In this embodiment, This represents the geometric mean of the rework rate and cost percentage of the i-th historical production, reflecting the issues of production efficiency and cost efficiency. The higher the rework rate or the higher the cost percentage, the larger this average value will be.

[0050] In this embodiment, It represents a comprehensive index of production efficiency and cost efficiency. The higher the value, the less rework is required and the lower the cost percentage, which means better production quality, better cost control, and larger profit margins.

[0051] The beneficial effects of the above technologies are as follows: analyzing historical batch data to determine the set of associated components for each evaluation parameter, calculating the parameter optimization factor for each evaluation parameter in each historical production, and calculating the historical reference value for each historical production of the dough to be tested can identify the key components that affect the quality evaluation of the finished product, improve the scientificity and pertinence of component ratio adjustment, and reduce the waste of resources caused by blind optimization. Example 6:

[0052] Based on Example 5, a method for non-destructive testing and proportion optimization of dough components is proposed. This method determines a reference production set for the dough based on historical reference values ​​from all previous production runs. Then, based on the reference production set and the associated component set for each evaluation parameter, it determines the component optimization data for the dough, including: Sort the historical reference values ​​of all historical productions of the dough to be tested from largest to smallest, and select the historical productions corresponding to the first specified number of historical reference values ​​after sorting as the reference production set of the dough to be tested. Based on the batch composition data of each historical production in the reference production set, historical reference values, parameter optimization factors of all evaluation parameters, the associated composition set of each evaluation parameter, and the composition optimization model, the optimization parameter range and optimal parameter value of each composition parameter are determined. Based on the optimized parameter range and optimal parameter values ​​of all component parameters, the optimized component data of the dough to be tested are determined.

[0053] In this embodiment, historical reference values ​​corresponding to all historical production runs of the dough to be tested are collected and arranged in descending order of value. A specified number is set according to actual needs. The specified number is determined based on the number of historical batches of all historical production runs. The larger the number of historical batches, the larger the specified number; the smaller the number of historical batches, the smaller the specified number, but it is at least 5 and can be 10. The top-ranking reference values ​​of this specified number are selected from the sorted historical reference values. The historical production batches corresponding to these reference values ​​together constitute the reference production set of the dough to be tested. The historical productions in the set are all cases of better past performance and have high reference value.

[0054] In this embodiment, batch component data from each historical production run in the reference production set is extracted. Simultaneously, historical reference values ​​for each historical production run, parameter optimization factors for all evaluation parameters, and the associated component set for each evaluation parameter are summarized. This multi-dimensional data is input into a pre-built component optimization model. The model integrates and analyzes this information, comprehensively considering the correlation between component parameters and finished product quality, market performance, etc., to determine the adjustable range of each component parameter during the optimization process, and the specific value within that range that achieves the optimal effect.

[0055] In this embodiment, batch component data provides the actual component values ​​of each historical batch in the reference production set, serving as the basic samples for model learning. The distribution range of its historical parameter values ​​directly determines the initial boundary of the optimization parameter range, ensuring that the optimization results do not deviate from actual production feasibility. Historical reference values ​​reflect the comprehensive performance of the corresponding batches and serve as the optimization target guide for the model, leading the optimal parameter values ​​towards the component values ​​of batches with higher historical reference values, ensuring that the overall performance of the optimized product is better. The parameter optimization factors of all evaluation parameters clarify the adjustment direction and magnitude of each component parameter on the quality of the finished product, providing the model with a basis for targeted adjustments, avoiding deviation from core quality requirements during the optimization process, and refining the reasonable range of optimization parameters. The associated component set of each evaluation parameter filters out the core component parameters that have a significant impact on quality, allowing the model to focus on optimizing key variables, filtering out interference from weakly associated parameters, ensuring the targeting of the optimization parameter range and the optimal parameter value, and improving optimization efficiency and accuracy.

[0056] In this embodiment, the optimized parameter range and optimal parameter value of all component parameters are summarized. These data are then integrated to form the component optimization data of the dough to be tested. This data clarifies the adjustment direction, adjustment boundary and ideal value of each component parameter, providing a direct and accurate basis for the subsequent optimization of the component ratio of the dough to be tested.

[0057] The beneficial effects of the above technology are as follows: Based on the historical reference values ​​of all historical productions of the dough to be tested, the reference production set of the dough to be tested is determined. Based on the reference production set of the dough to be tested and the associated component set of each evaluation parameter, the component optimization data of the dough to be tested is determined. This can achieve targeted optimization of the component ratio, improve the scientificity and adaptability of the component optimization data, and provide a precise and feasible basis for adjusting the ratio for subsequent production. Example 7:

[0058] Based on Example 6, a method for non-destructive testing and proportion optimization of dough components is provided. This method determines the testing label, optimization label, and optimized component data of the dough based on the detected component data, finished product proportion data, and component optimization data. Batch production of the dough is then carried out based on the optimized component data, including: The component values ​​of each component parameter in the component data of the dough to be tested and the target parameter range of each component parameter in the component optimization data are judged. If the component values ​​of all component parameters are within the corresponding target parameter range, the test label of the dough to be tested is determined to be qualified. If the component value of any component parameter is not within the corresponding target parameter range, the test label of the dough to be tested is determined to be unqualified. If the test label of the dough to be tested is unqualified, the dough to be tested will be optimized to meet the qualified ratio based on all the component parameters in the test component data that are not within the target parameter range; If the test label of the dough to be tested is qualified, based on the test component data and component optimization data, the sales ratio of the dough to be tested is optimized to determine the optimized component data of the dough to be tested; Based on the optimized composition data of the dough to be tested, batch production of the dough to be tested was carried out.

[0059] In this embodiment, the actual component value of each component parameter in the component analysis data of the dough to be tested is extracted, and the target parameter range corresponding to each component parameter in the component optimization data is defined. The actual component value of each component parameter is compared with its corresponding target parameter range one by one. If the actual component values ​​of all component parameters are within their respective target parameter ranges, the test label of the dough to be tested is determined to be qualified; if the actual component value of any component parameter exceeds its corresponding target parameter range, the test label of the dough to be tested is determined to be unqualified.

[0060] In this embodiment, if the test label of the dough to be tested is unqualified, it indicates that some of its component parameters do not meet the basic quality requirements. At this time, it is necessary to focus on all component parameters whose actual component values ​​in the test component data are not within the corresponding target parameter range, and adjust and optimize these non-compliant component parameters. The core goal is to make the values ​​of all component parameters fall within the corresponding target parameter range, thereby achieving the qualified ratio optimization of the dough to be tested.

[0061] In this embodiment, once the optimized ingredient data of the dough to be tested is determined, this data will be used as the basis for production. The dough to be tested will be mass-produced according to this optimized ingredient data to ensure that the produced dough not only meets the quality requirements but also adapts to the actual application scenario.

[0062] The beneficial effects of the above technologies are as follows: Based on the test component data, finished product ratio data, and component optimization data of the dough to be tested, the test label, optimization label, and optimized component data of the dough to be tested can be determined. Based on the optimized component data, batch production of the dough to be tested can be carried out, which can achieve precise matching of dough ratios, meet sales needs, take into account quality compliance and market adaptability, improve the targeting and efficiency of ratio optimization, and enhance the connection between production and the market. Example 8:

[0063] Based on Example 7, a method for non-destructive testing and proportion optimization of dough components is provided. If the test label of the dough to be tested is qualified, the optimized component data of the dough to be tested is determined by optimizing the sales proportion based on the test component data and component optimization data, including: The component values ​​of each component parameter in the component data of the dough to be tested and the optimized parameter range of each component parameter in the finished product ratio data are judged. If the component values ​​of all component parameters are within the corresponding target parameter range, the optimization label of the dough to be tested is determined to be no optimization required. If the component value of any component parameter is not within the corresponding optimized parameter range, the optimization label of the dough to be tested is determined to be optimization required. If the optimization label of the dough to be tested is that it needs to be optimized, the selling ratio of the dough to be tested is optimized based on the component parameters corresponding to all component values ​​that are not within the optimization parameter range and the optimal parameter value of each component parameter. Based on all optimized component values ​​that are outside the optimized parameter range, and all component values ​​that are within the optimized parameter range, the optimized component data of the dough to be tested is determined.

[0064] In this embodiment, the component value of each component parameter in the component data of the dough to be tested is extracted, and the optimized parameter range corresponding to each component parameter in the finished product formulation data is obtained. The component value of each component parameter is compared with the corresponding optimized parameter range one by one. If the component values ​​of all component parameters are within their respective optimized parameter ranges, the optimized label of the dough to be tested is determined to be no optimization required; if the component value of any component parameter exceeds its corresponding optimized parameter range, the optimized label of the dough to be tested is determined to be optimization required.

[0065] In this embodiment, if the optimized label of the dough to be tested indicates that optimization is needed, all component parameters whose actual component values ​​in the tested component data are not within the corresponding optimized parameter range are first screened out. For these component parameters, the optimal parameter value corresponding to each component parameter is used as the adjustment target to optimize the sales ratio, adjusting the values ​​of these component parameters to approach the optimal parameter value.

[0066] The beneficial effects of the above technology are as follows: If the test label of the dough to be tested is qualified, based on the test component data and component optimization data, the optimized component data of the dough to be tested can be determined by optimizing the sales ratio. This can further improve the accuracy and efficiency of component ratio optimization, so that the final optimized component data not only meets the market sales demand, but also ensures the rationality of component values, thereby enhancing the market competitiveness and quality stability of the finished product.

[0067] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for non-destructive detection and proportion optimization of dough components, characterized in that, include: S1: Obtain the target finished product of the dough to be tested, determine the finished product category, finished product ratio data, finished product evaluation data and dough test data of the dough to be tested, and determine the test component data of the dough based on the dough test data and finished product ratio data of the dough to be tested; S2: Obtain historical batch data of the dough to be tested from multiple historical productions, analyze the historical batch data to determine the set of associated components for each evaluation parameter, calculate the parameter optimization factor for each evaluation parameter in each historical production, and calculate the historical reference value for each historical production of the dough to be tested. S3: Based on the historical reference values ​​of all historical productions of the dough to be tested, determine the reference production set of the dough to be tested. Based on the reference production set of the dough to be tested and the associated component set of each evaluation parameter, determine the component optimization data of the dough to be tested. S4: Based on the test component data, finished product ratio data, and component optimization data of the dough to be tested, determine the test label, optimization label, and optimized component data of the dough to be tested, and carry out batch production of the dough to be tested based on the optimized component data.

2. The method for non-destructive detection and proportion optimization of dough components according to claim 1, characterized in that, Obtain the target finished product of the dough to be tested, and determine the finished product category, finished product ratio data, finished product evaluation data, and dough testing data, including: Obtain the target finished product of the dough to be tested, and determine the finished product category, finished product ratio data, and finished product evaluation data based on the finished product target. The finished product ratio data includes multiple component parameters and the target parameter range for each component parameter. The finished product evaluation data includes multiple evaluation parameters and the evaluation weight for each evaluation parameter. The dough testing data is determined based on the finished product category of the dough to be tested. The dough testing data includes multiple testing parameters and the testing equipment label for each testing parameter.

3. The method for non-destructive detection and proportion optimization of dough components according to claim 2, characterized in that, Based on the dough testing data and finished product formulation data of the dough to be tested, the component data of the dough to be tested are determined, including: Based on all the testing device tags in the dough testing data, non-destructive testing is performed on the dough to be tested to determine the testing parameter data of the dough to be tested. The testing parameter data includes multiple testing parameters and the testing parameter value of each testing parameter. Map all detection parameters in the detection parameter data and all component parameters in the finished product ratio data to determine the mapping detection parameter or mapping detection parameter set for each component parameter in the finished product ratio data. The mapping detection parameter set includes at least two or more detection parameters. The detection parameter value of the mapping detection parameter of each component parameter in the finished product ratio data or the detection parameter value of all detection parameters in the mapping detection parameter set is input into the parameter mapping model, and the component value of each component parameter in the finished product ratio data is determined based on the output of the parameter mapping model. Based on the component values ​​of all component parameters in the finished product formulation data, the component data for the dough to be tested are determined.

4. The method for non-destructive detection and proportion optimization of dough components according to claim 1, characterized in that, Historical batch data includes: batch production data, batch ingredient data, finished product sales data, and finished product feedback data. Among them, batch production data includes production time, production cost, production volume, and rework volume; batch ingredient data includes multiple ingredient parameters and historical parameter values ​​for each ingredient parameter; finished product sales data includes sales time range, sales price, actual daily sales price and sales volume within the sales time range; and finished product feedback data includes the consumption time of multiple consumers and evaluation descriptions based on each evaluation parameter in the finished product evaluation data.

5. The method for non-destructive detection and proportion optimization of dough components according to claim 4, characterized in that, Analyze historical batch data to determine the set of associated components for each evaluation parameter, calculate the parameter optimization factor for each evaluation parameter in each historical production, and calculate the historical reference value for each historical production of the dough to be tested, including: In the feedback data of finished products from each historical production of the dough to be tested, word sense recognition is performed on the evaluation sub-description of each evaluation parameter in the finished product evaluation data for each consumer in the finished product evaluation data. The evaluation labels are determined for each consumer in the feedback data of finished products from each historical production of the dough to be tested, based on each evaluation parameter in the finished product evaluation data. The evaluation labels include very weak, weak, moderate, strong, and very strong. The evaluation labels of each consumer based on each evaluation parameter in the finished product evaluation data in each historical production batch are quantified to determine the evaluation quantification value of each consumer based on each evaluation parameter in the finished product feedback data in each historical production batch. Pearson correlation analysis was performed on the quantitative evaluation values ​​of all consumers based on all evaluation parameters in the finished product feedback data of all historical productions and the historical parameter values ​​of all component parameters in the batch component data to determine the Pearson correlation value of each evaluation parameter and each component parameter. Based on the Pearson correlation values ​​of each evaluation parameter and all component parameters, as well as the preset correlation threshold, the set of associated components for each evaluation parameter is determined, wherein the set of associated components includes at least one or more component parameters. Based on the finished product evaluation data of the dough to be tested, the batch production data of each historical production, the finished product sales data, and the finished product feedback data, the quantitative evaluation values ​​of all consumers based on all evaluation parameters are calculated in the finished product feedback data of each historical production of the dough to be tested, and the parameter optimization factor of all consumers based on each evaluation parameter is calculated. The historical reference value of each historical production of the dough to be tested is also calculated.

6. The method for non-destructive detection and proportion optimization of dough components according to claim 5, characterized in that, Based on historical reference values ​​from all previous production runs of the dough to be tested, a reference production set for the dough to be tested is determined. Based on this reference production set and the associated component set for each evaluation parameter, the component optimization data for the dough to be tested is determined, including: Sort the historical reference values ​​of all historical productions of the dough to be tested from largest to smallest, and select the historical productions corresponding to the first specified number of historical reference values ​​after sorting as the reference production set of the dough to be tested. Based on the batch composition data of each historical production in the reference production set, historical reference values, parameter optimization factors of all evaluation parameters, the associated composition set of each evaluation parameter, and the composition optimization model, the optimization parameter range and optimal parameter value of each composition parameter are determined. Based on the optimized parameter range and optimal parameter values ​​of all component parameters, the optimized component data of the dough to be tested are determined.

7. The method for non-destructive detection and proportion optimization of dough components according to claim 6, characterized in that, Based on the tested component data, finished product ratio data, and component optimization data of the dough to be tested, the testing label, optimization label, and optimized component data of the dough to be tested are determined. Based on the optimized component data, batch production of the dough to be tested is carried out, including: The component values ​​of each component parameter in the component data of the dough to be tested and the target parameter range of each component parameter in the component optimization data are judged. If the component values ​​of all component parameters are within the corresponding target parameter range, the test label of the dough to be tested is determined to be qualified. If the component value of any component parameter is not within the corresponding target parameter range, the test label of the dough to be tested is determined to be unqualified. If the test label of the dough to be tested is unqualified, the dough to be tested will be optimized to meet the qualified ratio based on all the component parameters in the test component data that are not within the target parameter range; If the test label of the dough to be tested is qualified, based on the test component data and component optimization data, the sales ratio of the dough to be tested is optimized to determine the optimized component data of the dough to be tested; Based on the optimized composition data of the dough to be tested, batch production of the dough to be tested was carried out.

8. The method for non-destructive detection and proportion optimization of dough components according to claim 7, characterized in that, If the test label of the dough to be tested is qualified, based on the test component data and component optimization data, the optimized component data of the dough to be tested is determined by optimizing the sales ratio, including: The component values ​​of each component parameter in the component data of the dough to be tested and the optimized parameter range of each component parameter in the finished product ratio data are judged. If the component values ​​of all component parameters are within the corresponding target parameter range, the optimization label of the dough to be tested is determined to be no optimization required. If the component value of any component parameter is not within the corresponding optimized parameter range, the optimization label of the dough to be tested is determined to be optimization required. If the optimization label of the dough to be tested is that it needs to be optimized, the selling ratio of the dough to be tested is optimized based on the component parameters corresponding to all component values ​​that are not within the optimization parameter range and the optimal parameter value of each component parameter. Based on all optimized component values ​​that are outside the optimized parameter range, and all component values ​​that are within the optimized parameter range, the optimized component data of the dough to be tested is determined.

Images