Rice noodle processing raw material proportioning method based on rice quality

Abstract

The application discloses a rice noodle processing raw material proportioning method based on rice quality, and belongs to the technical field of food processing; the method comprises the following steps: obtaining historical rice noodle quality data; wherein the historical rice noodle quality data comprises rice raw material proportioning, proportioned rice quality data and rice noodle quality data; the rice noodle quality data comprises taste score, toughness score and cooking loss rate; a rice noodle quality prediction model is constructed according to the historical rice noodle quality data; rice quality data of different varieties of rice in a current purchase batch is obtained; wherein the rice quality data comprises amylose content, gel consistency, alkali-soluble value, chalkiness, peak viscosity, retrogradation value, decay value and reduction value; and based on a preset dynamic scheduling algorithm, an optimal rice proportioning scheme of the current purchase batch is obtained according to the rice quality data and the rice noodle quality prediction model.

Description

Technical Field

[0001] This invention belongs to the field of food processing technology, specifically relating to a method for proportioning raw materials for rice noodle processing based on rice quality. Background Technology

[0002] Rice noodles are a traditional staple food in southern my country, made from rice, and are known for their smooth texture and convenience. With the rapid development of the rice noodle industry, higher demands have been placed on the stability and consistency of rice noodle quality. However, the quality of rice noodles largely depends on the quality characteristics of the raw rice. Different varieties, origins, years, and even batches of rice exhibit significant differences in key quality indicators such as amylose content, gel consistency, alkali spreading value, chalkiness, and gelatinization characteristics. These differences directly affect the rice noodles' shaping, texture, elasticity, cooking loss rate, and breakage rate.

[0003] Currently, most companies classify rice based on rough information such as origin, variety name, or appearance when purchasing it, lacking precise testing of key quality indicators such as amylose content, gel consistency, alkali spreading value, chalky grain rate, chalkiness, retrogradation value, and peak viscosity. The amylose content directly affects the rice noodle's shaping, cooking loss rate, and palatability, while gel consistency and alkali spreading value also significantly impact taste. Furthermore, the quality of the same rice variety can fluctuate due to differences in planting conditions and storage time, and current technology lacks the ability to detect such fluctuations. Secondly, raw material ratios mainly rely on repeated trials and adjustments based on the personal experience of technicians, which is not only inefficient and time-consuming but also makes it difficult to guarantee the quality stability and consistency of different batches. Thirdly, due to the lack of a mechanism for dynamically fine-tuning the ratio based on the actual quality test results of each batch of purchased rice, the quality of rice noodles produced using the same ratio under different quality raw materials fluctuates significantly, seriously affecting the stability of product quality.

[0004] Therefore, there is an urgent need for a method for proportioning raw materials for rice noodle processing based on rice quality, in order to solve the problems existing in the current technology. Summary of the Invention

[0005] In view of this, the present invention provides a method for the proportioning of raw materials for rice noodle processing based on rice quality, which solves the problem in the prior art that the proportioning relies on manual experience and cannot be dynamically adjusted according to batch quality fluctuations, resulting in unstable taste and toughness of rice noodles and poor quality consistency.

[0006] To achieve the above objectives, the present invention provides a method for proportioning raw materials for rice noodle processing based on rice quality, comprising the following steps: Obtain historical rice noodle quality data; the historical rice noodle quality data includes: rice raw material ratio, rice quality data of the ratio, and rice noodle quality data; the rice noodle quality data includes: taste score, toughness score, and cooking loss rate; A rice noodle quality prediction model was constructed based on historical rice noodle quality data. Obtain rice quality data for different varieties of rice in the current batch of goods; the rice quality data includes: amylose content, gel consistency, alkali spreading value, chalkiness, peak viscosity, retrogradation value, decay value, and reduction value; Based on a preset dynamic scheduling algorithm, the optimal rice blending scheme for the current batch of goods is obtained according to rice quality data and rice noodle quality prediction model.

[0007] As an embodiment of the present invention, a rice noodle quality prediction model is constructed based on historical rice noodle quality data, including: The quality data of the mixed rice were determined based on the rice raw material ratio and the quality data of the mixed rice. Based on mixed rice quality data and rice noodle quality data, we trained a taste score prediction model, a toughness score prediction model, and a cooking loss rate prediction model, respectively. A rice noodle quality prediction model was obtained based on the taste score prediction model, the toughness score prediction model, and the cooking loss rate prediction model.

[0008] As an embodiment of the present invention, a taste score prediction model and a toughness score prediction model are trained based on mixed rice quality data and rice noodle quality data, respectively, including: Using indicators from mixed rice quality data as independent variables and taste score as the dependent variable, a taste score prediction model was established based on multi-source regression analysis, as shown below: In the formula, Indicates taste rating. Represents the intercept term. , , , , , , and All represent the regression coefficients of the taste rating prediction model. Indicates amylose content, Indicates the consistency of the glue, Indicates alkali digestion value, Indicates chalkiness, Indicates peak viscosity, Indicates regeneration value, Indicates the attenuation value and Indicates the decrease value; Using indicators from mixed rice quality data as independent variables and resilience score as the dependent variable, a resilience score prediction model was established based on multi-source regression analysis, as shown below: In the formula, Indicates resilience score, Represents the intercept term. , , , and All of these represent the regression coefficients of the resilience score prediction model.

[0009] As an embodiment of the present invention, based on a preset dynamic scheduling algorithm, the optimal rice blending scheme for the current batch of goods is obtained according to rice quality data and rice noodle quality prediction model, including: Obtain the standard quality data of rice noodles and the current rice raw material ratio input by the administrator; Initial mixed quality data were obtained based on the rice raw material ratio and the quality data of the current batch of rice. Input the initial mixed quality data into the trained rice noodle quality prediction model to obtain the initial predicted quality of rice noodles; Determine whether the quality difference between the initial predicted quality of rice noodles and the standard quality data of rice noodles is greater than the preset quality difference; if it is not greater, take the current rice raw material ratio as the optimal rice ratio scheme for the current batch of goods. If the ratio is greater than the target value, the optimal rice blending scheme for the current batch of rice is obtained based on the preset dynamic scheduling algorithm, according to the standard quality data of rice noodles and the quality data of rice in the current batch.

[0010] As an embodiment of the present invention, determining whether the quality difference between the initial predicted quality of rice noodles and the standard quality data of rice noodles is greater than a preset quality gap includes: The quality difference between the initial predicted quality of the rice noodles and the standard quality data is calculated as follows: In the formula, This indicates poor quality. This represents the taste score indicating the initial predicted quality of the rice noodles. This indicates the taste score, representing the standard quality data of rice noodles. This represents the initial predicted quality score of the rice noodles, indicating their resilience. The toughness score represents the standard quality data of rice noodles. This represents the cooking loss rate of the initial predicted quality of the rice noodles. This indicates the cooking loss rate, which represents the standard quality data for rice noodles. Determine whether the quality difference is greater than the preset quality gap.

[0011] As an embodiment of the present invention, based on a preset dynamic scheduling algorithm, the optimal rice blending scheme for the current batch of rice noodles is obtained according to the standard quality data of rice noodles and the quality data of the current batch of rice, including: Using the minimization of quality difference and the rice noodle quality prediction function as the objective function, the rice quality data of the current batch of rice is input into the preset dynamic scheduling algorithm, and the optimal rice blending scheme of the current batch of rice is output; wherein, the optimal rice blending scheme is the proportion of different types of rice in the current batch of rice.

[0012] As one embodiment of the present invention, it also includes: To obtain the test results of the quality inspection of the finished product produced according to the optimal rice ratio scheme; Determine whether the quality difference between the test results and the standard quality data for rice noodles is greater than the preset error; If the test results of three consecutive batches of goods are greater than the preset error, the test results and the rice quality data corresponding to the optimal rice ratio scheme will be used as training data to dynamically update the regression coefficients in the rice noodle quality prediction model.

[0013] As one embodiment of the present invention, it also includes: The gelatinization temperature and peak time of different types of ingredients in the obtained optimal rice formulation; The dispersion of the gelatinization characteristics of mixtures based on the gelatinization temperature and peak time of different types of ingredients is shown below: In the formula, This indicates the dispersion of gelatinization temperature. Indicates the dispersion of peak time. and These represent the optimal rice blending scheme. The ingredients and the first The gelatinization temperature of each ingredient. and These represent the optimal rice blending scheme. The ingredients and the first Peak time of each ingredient and These represent the optimal rice blending scheme. The ingredients and the first The proportions of each ingredient Indicates the preset ratio threshold; Determine whether the dispersion of the gelatinization characteristics meets the preset consistency condition; if it does, execute the optimal rice ratio scheme; if it does not, adjust the optimal rice ratio scheme to obtain the target rice ratio scheme.

[0014] As an embodiment of the present invention, the optimal rice blending scheme is adjusted to obtain the target rice blending scheme, including: Ingredients with a proportion greater than a preset threshold are selected as adjustment ingredients; Calculate and adjust the deviation of each ingredient in the batch, as shown below: In the formula, Indicates the first The deviation in the gelatinization temperature of each ingredient. Indicates the first The deviation of the peak time of each ingredient This indicates adjusting the average gelatinization temperature of the ingredients. This indicates the average peak time for adjusting the ingredients; The ingredient with the largest deviation in gelatinization temperature was selected as the gelatinization ingredient, and the ingredient with the largest deviation in peak time was selected as the peak ingredient. Determine if the gelatinized ingredient and the peak ingredient are the same; if they are the same, reduce the preset proportion of the gelatinized ingredient. If they are not the same, the ingredient with the larger deviation between the gelatinized ingredient and the peak ingredient is regarded as the main outlier, and the ingredient with the smaller deviation between the gelatinized ingredient and the peak ingredient is regarded as the secondary outlier. While keeping the proportion of the main outlier ingredient constant, the proportion of the secondary outlier ingredient is gradually shifted to the ingredient with the opposite gelatinization characteristics at a preset step size.

[0015] The beneficial effects of this invention are as follows: It solves the problem in the prior art that the proportions rely on manual experience and cannot be dynamically adjusted according to batch quality fluctuations, resulting in unstable taste and toughness of rice noodles and poor quality consistency. By replacing traditional experience-based decision-making with a prediction model driven by historical data, the determination of the proportions is transformed from repeated trial and adjustment to a one-time accurate calculation based on a quantitative model, which greatly shortens the formula adjustment cycle and eliminates the uncertainty caused by human factors. At the same time, the dynamic scheduling algorithm can calculate the optimal proportions in real time based on the actual test data of each batch of rice, so that the proportion scheme adapts to the fluctuations in raw material quality. This ensures that rice noodle products with stable taste and toughness and controllable cooking loss rate can still be produced even when there are significant differences in quality indicators among different batches of rice, effectively improving the batch-to-batch consistency of rice noodle processing quality.

[0016] Other advantages, objectives, and features of the invention will be set forth in the following description and will be apparent to those skilled in the art in some respects, or may be learned by practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0017] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the following figures are provided for illustration: Figure 1 This is a schematic diagram of the process of the present invention; Figure 2 This is a schematic diagram of the process for constructing a rice noodle quality prediction model according to the present invention. Detailed Implementation

[0018] like Figures 1-2 As shown, this invention provides a method for proportioning raw materials for rice noodle processing based on rice quality, comprising the following steps: Obtain historical rice noodle quality data; the historical rice noodle quality data includes: rice raw material ratio, rice quality data of the ratio, and rice noodle quality data; the rice noodle quality data includes: taste score, toughness score, and cooking loss rate; A rice noodle quality prediction model was constructed based on historical rice noodle quality data. Obtain rice quality data for different varieties of rice in the current batch of goods; the rice quality data includes: amylose content, gel consistency, alkali spreading value, chalkiness, peak viscosity, retrogradation value, decay value, and reduction value; Based on a preset dynamic scheduling algorithm, the optimal rice blending scheme for the current batch of goods is obtained according to rice quality data and rice noodle quality prediction model.

[0019] The working principle of the above technical solution is as follows: This technical solution constructs a training dataset by collecting historical production data. The historical data includes rice raw material ratio schemes, mixed raw materials under each ratio, and the taste, toughness, and cooking loss rate of the rice noodles actually processed under that ratio, forming a complete corresponding sample of "ratio → raw material quality → rice noodle quality". Based on this, using quality indicators such as amylose content, gel consistency, alkali spreading value, chalkiness, peak viscosity, retrogradation value, decay value, and reduction value of the mixed raw materials as input features, and rice noodle taste, toughness, and cooking loss rate as output targets, a rice noodle quality prediction model is established using regression analysis, realizing the prediction from raw material quality to rice noodle quality. The quality is quantitatively mapped; for different varieties of rice in the current batch, after obtaining the above quality indicators of each variety by detection, the preset dynamic scheduling algorithm uses the prediction model as the core evaluation function. Under the constraints of the total ratio being 1, the ratio of each variety being non-negative and not exceeding the upper limit, and the key quality indicators of the mixed raw materials being in the appropriate range (amylose content 20%~28%, gel consistency not higher than 45mm, alkali spreading value not lower than level 6, and retrogradation value not lower than 1600cP), the optimal ratio scheme that maximizes the predicted taste score and toughness score while minimizing the cooking loss rate is searched, so as to realize the dynamic adaptive adjustment of the raw material ratio according to the actual quality of each batch of rice. The beneficial effects of the above technical solution are as follows: This solution solves the problem in existing technologies where the proportions rely on manual experience and cannot be dynamically adjusted according to batch quality fluctuations, leading to unstable texture and chewiness of rice noodles and poor quality consistency. By replacing traditional experience-based decision-making with a historical data-driven predictive model, the proportion determination is transformed from repeated trial adjustments to a one-time precise calculation based on a quantitative model. This significantly shortens the formula adjustment cycle and eliminates the uncertainty caused by human factors. Simultaneously, the dynamic scheduling algorithm can calculate the optimal proportions in real time based on the actual test data of each batch of incoming rice, allowing the proportions to adapt to fluctuations in raw material quality. This ensures that even when different batches of rice have significant differences in quality indicators, rice noodle products with stable and acceptable texture and chewiness, and controllable cooking loss rates can still be produced, effectively improving the batch-to-batch consistency of rice noodle processing quality.

[0020] In one embodiment, a rice noodle quality prediction model is constructed based on historical rice noodle quality data, including: The quality data of the mixed rice were determined based on the rice raw material ratio and the quality data of the mixed rice. Based on mixed rice quality data and rice noodle quality data, we trained a taste score prediction model, a toughness score prediction model, and a cooking loss rate prediction model, respectively. A rice noodle quality prediction model was obtained based on the taste score prediction model, the toughness score prediction model, and the cooking loss rate prediction model. Among them, taste score prediction models and toughness score prediction models were trained based on mixed rice quality data and rice noodle quality data, respectively, including: Using indicators from mixed rice quality data as independent variables and taste score as the dependent variable, a taste score prediction model was established based on multi-source regression analysis, as shown below: In the formula, Indicates taste rating. Represents the intercept term. , , , , , , and All represent the regression coefficients of the taste rating prediction model. Indicates amylose content, Indicates the consistency of the glue, Indicates alkali digestion value, Indicates chalkiness, Indicates peak viscosity, Indicates regeneration value, Indicates the attenuation value and Indicates the decrease value; Using indicators from mixed rice quality data as independent variables and resilience score as the dependent variable, a resilience score prediction model was established based on multi-source regression analysis, as shown below: In the formula, Indicates resilience score, Represents the intercept term. , , , and All of these represent the regression coefficients of the resilience score prediction model.

[0021] The working principle and beneficial effects of the above technical solution are as follows: First, a training dataset is constructed based on historical rice noodle quality data, including the rice raw material ratio, quality indicators of each rice variety, and the taste score, toughness score, and cooking loss rate of the rice noodles actually processed from this ratio. Based on the proportion of each variety in the rice raw material ratio, the quality indicators of each rice variety are weighted and averaged to calculate the mixed rice quality data. This data covers key indicators such as amylose content, gel consistency, alkali spreading value, chalkiness, peak viscosity, retrogradation value, decay value, and reduction value. Subsequently, using the above comprehensive quality indicators of the mixed rice as independent variables, and taste score, toughness score, and cooking loss rate as dependent variables, three independent sub-prediction models are trained using multiple regression analysis. During the model training process... Furthermore, by combining stepwise regression, with a significance level less than 0.05 as the inclusion criterion and greater than 0.1 as the exclusion criterion, rice quality indicators that have a significant impact on various rice noodle qualities are automatically screened, and insignificant indicator variables are eliminated. Simplified taste score prediction models, toughness score prediction models, and cooking loss rate prediction models are then established. These three sub-models are then combined to form a rice noodle quality prediction model. The taste score prediction model is used to evaluate the sensory attributes of rice noodles, such as smoothness, hardness, and elasticity; the toughness score prediction model is used to evaluate the mechanical attributes of rice noodles, such as tensile strength and breakage rate; and the cooking loss rate prediction model is used to evaluate the yield and processing loss characteristics of rice noodles. The three sub-models comprehensively reflect the overall quality of rice noodles from different dimensions, thereby achieving a comprehensive quantitative prediction of rice noodle quality. By establishing three independent sub-prediction models for taste score, toughness score, and cooking loss rate, the sensory quality, mechanical quality, and economic processing quality of rice noodles can be accurately predicted in multiple dimensions. This overcomes the shortcomings of a single comprehensive model that cannot take into account multiple quality attributes. Furthermore, by combining stepwise regression method to automatically screen significant indicators, the model structure is simplified while ensuring the accuracy of model prediction.

[0022] In one embodiment, based on a preset dynamic scheduling algorithm, the optimal rice blending scheme for the current batch of goods is obtained according to rice quality data and a rice noodle quality prediction model, including: Obtain the standard quality data of rice noodles and the current rice raw material ratio input by the administrator; Initial mixed quality data were obtained based on the rice raw material ratio and the quality data of the current batch of rice. Input the initial mixed quality data into the trained rice noodle quality prediction model to obtain the initial predicted quality of rice noodles; Determine whether the quality difference between the initial predicted quality of rice noodles and the standard quality data of rice noodles is greater than the preset quality difference; if it is not greater, take the current rice raw material ratio as the optimal rice ratio scheme for the current batch of goods. If the ratio is greater than the standard quality data of rice noodles and the quality data of rice in the current batch, the optimal rice ratio scheme for the current batch is obtained based on the preset dynamic scheduling algorithm and the standard quality data of rice noodles and the current batch of rice. Determine whether the quality difference between the initial predicted quality of rice noodles and the standard quality data is greater than the preset quality gap, including: The quality difference between the initial predicted quality of the rice noodles and the standard quality data is calculated as follows: In the formula, This indicates poor quality. This represents the taste score indicating the initial predicted quality of the rice noodles. This indicates the taste score, representing the standard quality data of rice noodles. This represents the initial predicted quality score of the rice noodles, indicating their resilience. The toughness score represents the standard quality data of rice noodles. This represents the cooking loss rate of the initial predicted quality of the rice noodles. This indicates the cooking loss rate, which represents the standard quality data for rice noodles. Determine whether the quality difference is greater than the preset quality gap; Based on a preset dynamic scheduling algorithm, the optimal rice blending scheme for the current batch of rice noodles is obtained according to the standard quality data of rice noodles and the quality data of the rice in the current batch, including: Using the minimization of quality difference and the rice noodle quality prediction function as the objective function, the rice quality data of the current batch of rice is input into the preset dynamic scheduling algorithm, and the optimal rice blending scheme for the current batch of rice is output; where the optimal rice blending scheme is the proportion of different types of rice in the current batch of rice. The working principle and beneficial effects of the above technical solution are as follows: First, the standard quality data of rice noodles pre-set by the administrator is obtained as a target reference, and the rice raw material ratio scheme currently used by the enterprise is also obtained. Then, based on the proportion of each variety of rice in the currently used ratio scheme, the quality indicators of each variety of rice obtained from the current batch of incoming goods are used to calculate the initial mixed rice quality data. The initial mixed quality data is input into the trained rice noodle quality prediction model, and the taste score prediction sub-model, toughness score prediction sub-model, and cooking loss rate prediction sub-model are called respectively to output the initial rice noodle predicted quality. Subsequently, the solution calculates the quality difference between the initial rice noodle predicted quality and the standard quality data of rice noodles. This quality difference is comprehensively quantitatively evaluated by considering the deviation of taste score, toughness score, and cooking loss rate. It is determined whether the quality difference is greater than the preset quality gap: if it is not greater, it means that the currently used ratio scheme can still meet the production requirements under the quality of this batch of rice, and the current ratio is directly used as the optimal scheme; if it is greater, it means that the current ratio can no longer adapt to the quality fluctuation of this batch of rice, and the preset dynamic scheduling algorithm is activated. Among them, the preset dynamic... The dynamic scheduling algorithm aims to minimize the aforementioned quality differences. Using the rice noodle quality prediction model as the core evaluation of the objective function, it takes the quality data of each rice variety in the current batch as input. Under constraints such as the total proportion being equal to the whole, the proportion of each variety being non-negative and not exceeding the upper limit, and the key quality indicators of the mixed raw materials being within a suitable range, it iteratively searches and optimizes the proportion scheme until it finds the proportion combination that minimizes the difference between the predicted quality and the standard quality. Finally, it outputs the optimal rice proportion scheme for the current batch. For example, suppose the current batch contains three types of rice… If the initial formulation causes the predicted toughness score to exceed the standard upper limit due to the higher amylose content of the first type of rice than the benchmark value, the dynamic scheduling algorithm will first slightly reduce the proportion of the first type of rice within the constraints and correspondingly increase the proportion of the second or third type of rice. It will calculate the predicted quality difference under the new formulation. If the difference is reduced, it will continue to adjust in this direction. If the difference is increased, it will adjust in the opposite direction. Through repeated iterative search, it will eventually converge to the optimal formulation combination that minimizes the comprehensive deviation of the taste score, toughness score and cooking loss rate, thus obtaining the optimal rice formulation scheme for the current batch of goods. Through the above technical solution, by establishing a quality pre-verification mechanism for the initial formulation, it is possible to determine whether the current formulation is still suitable for the new batch of raw materials without conducting actual processing tests, thus avoiding the risk of quality fluctuations caused by blindly following the original formulation. When quality differences exceed the standard, the dynamic scheduling algorithm can automatically perform targeted optimization based on the actual quality data of each variety of rice in the current batch, with the standard quality as the target, and output the optimal formulation within one calculation cycle. This completely replaces the traditional manual adjustment mode of repeated trial and error, and significantly shortens the formulation response time. At the same time, this solution constructs a quality difference evaluation system through three dimensions: comprehensive taste score, toughness score, and cooking loss rate, ensuring that taste and toughness meet the standards simultaneously and that processing losses are controllable during the formulation optimization process, effectively guaranteeing a high degree of consistency in rice noodle quality between different batches of raw materials.

[0023] In one embodiment, it also includes: To obtain the test results of the quality inspection of the finished product produced according to the optimal rice ratio scheme; Determine whether the quality difference between the test results and the standard quality data for rice noodles is greater than the preset error; If the test results of three consecutive batches of goods are greater than the preset error, the test results and the rice quality data corresponding to the optimal rice ratio scheme will be used as training data to dynamically update the regression coefficients in the rice noodle quality prediction model.

[0024] The working principle and beneficial effects of the above technical solution are as follows: First, the finished rice noodles produced according to the optimal rice ratio are subjected to quality testing to obtain actual taste scores, toughness scores, and cooking loss rates. Then, the actual test results are compared with the rice noodle standard quality data preset by the administrator. The quality difference between the two is calculated by comprehensively considering the deviations in taste scores, toughness scores, and cooking loss rates, and it is determined whether this quality difference exceeds a preset error threshold. If the quality difference between the current batch's test results and the standard quality does not exceed the preset error, it indicates that the current prediction model still has good prediction accuracy under the conditions of this batch of raw materials, and no model adjustment is required. If it exceeds the preset error, the batch is marked as abnormal. When the test results of three consecutive batches of incoming goods are all greater than the preset error, it indicates that the prediction model has a systematic deviation under the current trend of raw material quality changes, and the model prediction accuracy can no longer meet the production requirements. At this time, the model update mechanism is triggered: the actual test results of the finished products of these three batches and the mixed rice quality data calculated by the optimal rice ratio scheme of the corresponding batches are used as new training samples and added to the historical training dataset. Regression analysis is performed again, and the regression coefficients of the taste score prediction sub-model, toughness score prediction sub-model and cooking loss rate prediction sub-model in the rice noodle quality prediction model are dynamically updated so that the prediction law of the model is adapted to the latest raw material quality change trend, thereby restoring and maintaining the prediction accuracy of the model. Through the above technical solution, by establishing an error monitoring mechanism between the finished product quality test results and the standard quality, the prediction accuracy of the rice noodle quality prediction model is continuously tracked and evaluated. When the prediction deviation of multiple consecutive batches exceeds the allowable range, the model is automatically updated, and the actual production data is fed back to the model for dynamic correction of the regression coefficient. This enables the model to adaptively track long-term fluctuations such as seasonal changes in raw material quality, changes in production location, and evolution of storage conditions.

[0025] In one embodiment, it also includes: The gelatinization temperature and peak time of different types of ingredients in the obtained optimal rice formulation; The dispersion of the gelatinization characteristics of mixtures based on the gelatinization temperature and peak time of different types of ingredients is shown below: In the formula, This indicates the dispersion of gelatinization temperature. Indicates the dispersion of peak time. and These represent the optimal rice blending scheme. The ingredients and the first The gelatinization temperature of each ingredient. and These represent the optimal rice blending scheme. The ingredients and the first Peak time of each ingredient and These represent the optimal rice blending scheme. The ingredients and the first The proportions of each ingredient Indicates the preset ratio threshold; Determine whether the dispersion of the gelatinization characteristics meets the preset consistency condition; if it does, execute the optimal rice ratio scheme; if it does not, adjust the optimal rice ratio scheme to obtain the target rice ratio scheme. The optimal rice blending scheme is adjusted to obtain the target rice blending scheme, which includes: Ingredients with a proportion greater than a preset threshold are selected as adjustment ingredients; Calculate and adjust the deviation of each ingredient in the batch, as shown below: In the formula, Indicates the first The deviation in the gelatinization temperature of each ingredient. Indicates the first The deviation of the peak time of each ingredient This indicates adjusting the average gelatinization temperature of the ingredients. This indicates the average peak time for adjusting the ingredients; The ingredient with the largest deviation in gelatinization temperature was selected as the gelatinization ingredient, and the ingredient with the largest deviation in peak time was selected as the peak ingredient. Determine if the gelatinized ingredient and the peak ingredient are the same; if they are the same, reduce the preset proportion of the gelatinized ingredient. If they are not the same, the ingredient with the larger deviation between the gelatinized ingredient and the peak ingredient is regarded as the main outlier, and the ingredient with the smaller deviation between the gelatinized ingredient and the peak ingredient is regarded as the secondary outlier. While keeping the proportion of the main outlier ingredient constant, the proportion of the secondary outlier ingredient is gradually shifted to the ingredient with the opposite gelatinization characteristics at a preset step size.

[0026] The working principle and beneficial effects of the above technical solution are as follows: In the actual production of rice noodles, there are significant differences in the gelatinization temperature and peak time of different varieties of rice. Gelatinization temperature refers to the temperature required for rice starch granules to begin absorbing water and swelling, disintegrating the crystal structure and forming a paste during the heating process, reflecting the ease or difficulty of starch gelatinization; peak time refers to the time elapsed from the start of heating until the starch gelatinization viscosity reaches its highest value, reflecting the speed of starch gelatinization. When the differences in gelatinization temperature or peak time among multiple varieties in the mixed raw materials are too large, during the same heating process in the gelatinization chamber, some varieties will be over-gelatinized while others will not be fully gelatinized, resulting in an uneven starch gel network structure inside the rice noodles during extrusion forming, ultimately manifesting as inconsistent texture and large fluctuations in toughness. Therefore, this technical solution first obtains the gelatinization temperature and peak time of each variety of rice in the current optimal ratio scheme, and calculates the dispersion of the mixed raw materials on these two indicators, i.e., the difference between the maximum and minimum values, to quantitatively evaluate the consistency of each variety in gelatinization characteristics. Based on this, a preset ratio threshold is set to exclude varieties with excessively low proportions. That is, when the proportion of a certain variety in the ratio is extremely low, its quality characteristics have a negligible impact on the overall gelatinization characteristics of the mixed raw materials. Including it in the dispersion calculation would cause interference and lead to inaccurate evaluation results. Then, it is determined whether the dispersion meets the preset consistency condition. If it does, the proportioning scheme is directly implemented; if not, the proportioning scheme is adjusted. During the adjustment process, only varieties with proportions greater than the preset threshold are selected as adjustment targets. For each adjusted variety, the deviation of its gelatinization temperature from the average gelatinization temperature of all adjusted varieties and the deviation of its peak time from the average peak time of all adjusted varieties are calculated. The variety with the largest deviation is selected as the gelatinization ingredient and the peak ingredient, respectively, and it is determined whether the two are the same variety. If the ingredients are of the same variety, it indicates that the variety deviates significantly from the mixing mean in both the gelatinization temperature and peak time dimensions, which has the greatest impact on the asynchronous gelatinization. In this case, the proportion of the variety is reduced by a preset amount, and the reduced portion is distributed to other varieties according to their original proportions to keep the adjusted proportion structure relatively balanced. If the gelatinization ingredients and the peak ingredients are different varieties, the one with the larger deviation is taken as the main outlier and its proportion is kept unchanged during the adjustment. The proportion of the secondary outlier ingredients with smaller deviations is gradually transferred to varieties with gelatinization characteristics opposite to those of the main outlier ingredients at preset step sizes. For example, if the gelatinization temperature of the main outlier ingredient is too high, a variety with a lower gelatinization temperature is found as the receiving variety, and the proportion of the secondary outlier ingredients is gradually transferred to the receiving variety. The dispersion is recalculated at each transfer step until the dispersion meets the consistency condition or reaches the maximum number of iterations to obtain the target rice proportion scheme. By using the above technical solution as an independent evaluation dimension for the quality of the formulation scheme, the consistency of gelatinization can be pre-tested before the formulation scheme is put into production, thus avoiding the problem of uneven gelatinization in the optimal formulation scheme. At the same time, by setting a preset ratio threshold, the interference of low-proportion varieties on the dispersion calculation is eliminated, so that the evaluation results focus on the main varieties that truly play a decisive role in the gelatinization characteristics of the mixed raw materials, thereby improving the accuracy of the judgment.

[0027] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made to it in form and detail without departing from the scope defined by the claims of the present invention.

Claims

1. A method for proportioning raw materials for rice noodle processing based on rice quality, characterized in that, Includes the following steps: Obtain historical rice noodle quality data; the historical rice noodle quality data includes: rice raw material ratio, rice quality data of the ratio, and rice noodle quality data; the rice noodle quality data includes: taste score, toughness score, and cooking loss rate; A rice noodle quality prediction model was constructed based on historical rice noodle quality data. Obtain rice quality data for different varieties of rice in the current batch of goods; the rice quality data includes: amylose content, gel consistency, alkali spreading value, chalkiness, peak viscosity, retrogradation value, decay value, and reduction value; Based on a preset dynamic scheduling algorithm, the optimal rice blending scheme for the current batch of goods is obtained according to rice quality data and rice noodle quality prediction model.

2. The method for proportioning raw materials for rice noodle processing based on rice quality according to claim 1, characterized in that, A rice noodle quality prediction model was constructed based on historical rice noodle quality data, including: The quality data of the mixed rice were determined based on the rice raw material ratio and the quality data of the mixed rice. Based on mixed rice quality data and rice noodle quality data, we trained a taste score prediction model, a toughness score prediction model, and a cooking loss rate prediction model, respectively. A rice noodle quality prediction model was obtained based on the taste score prediction model, the toughness score prediction model, and the cooking loss rate prediction model.

3. The method for proportioning raw materials for rice noodle processing based on rice quality according to claim 1, characterized in that, Based on mixed rice quality data and rice noodle quality data, taste score prediction models and chewiness score prediction models were trained respectively, including: Using indicators from mixed rice quality data as independent variables and taste score as the dependent variable, a taste score prediction model was established based on multi-source regression analysis, as shown below: In the formula, Indicates taste rating. Represents the intercept term. , , , , , , and All represent the regression coefficients of the taste rating prediction model. Indicates amylose content, Indicates the consistency of the glue, Indicates alkali digestion value, Indicates chalkiness, Indicates peak viscosity, Indicates regeneration value, Indicates the attenuation value and Indicates the decrease value; Using indicators from mixed rice quality data as independent variables and resilience score as the dependent variable, a resilience score prediction model was established based on multi-source regression analysis, as shown below: In the formula, Indicates resilience score, Represents the intercept term. , , , and All of these represent the regression coefficients of the resilience score prediction model.

4. The method for proportioning raw materials for rice noodle processing based on rice quality according to claim 1, characterized in that, Based on a preset dynamic scheduling algorithm, the optimal rice blending scheme for the current batch of rice is obtained according to rice quality data and rice noodle quality prediction model, including: Obtain the standard quality data of rice noodles and the current rice raw material ratio input by the administrator; Initial mixed quality data were obtained based on the rice raw material ratio and the quality data of the current batch of rice. Input the initial mixed quality data into the trained rice noodle quality prediction model to obtain the initial predicted quality of rice noodles; Determine whether the quality difference between the initial predicted quality of rice noodles and the standard quality data of rice noodles is greater than the preset quality difference; if it is not greater, take the current rice raw material ratio as the optimal rice ratio scheme for the current batch of goods. If the ratio is greater than the target value, the optimal rice blending scheme for the current batch of rice is obtained based on the preset dynamic scheduling algorithm, according to the standard quality data of rice noodles and the quality data of rice in the current batch.

5. The method for proportioning raw materials for rice noodle processing based on rice quality according to claim 1, characterized in that, Determine whether the quality difference between the initial predicted quality of rice noodles and the standard quality data is greater than the preset quality gap, including: The quality difference between the initial predicted quality of the rice noodles and the standard quality data is calculated as follows: In the formula, This indicates poor quality. This represents the taste score indicating the initial predicted quality of the rice noodles. This indicates the taste score, representing the standard quality data of rice noodles. This represents the initial predicted quality score of the rice noodles, indicating their resilience. The toughness score represents the standard quality data of rice noodles. This represents the cooking loss rate of the initial predicted quality of the rice noodles. This indicates the cooking loss rate, which represents the standard quality data for rice noodles. Determine whether the quality difference is greater than the preset quality gap.

6. The method for proportioning raw materials for rice noodle processing based on rice quality according to claim 1, characterized in that, Based on a preset dynamic scheduling algorithm, the optimal rice blending scheme for the current batch of rice noodles is obtained according to the standard quality data of rice noodles and the quality data of the rice in the current batch, including: Using the minimization of quality difference and the rice noodle quality prediction function as the objective function, the rice quality data of the current batch of rice is input into the preset dynamic scheduling algorithm, and the optimal rice blending scheme of the current batch of rice is output; wherein, the optimal rice blending scheme is the proportion of different types of rice in the current batch of rice.

7. The method for proportioning raw materials for rice noodle processing based on rice quality according to claim 1, characterized in that, Also includes: To obtain the test results of the quality inspection of the finished product produced according to the optimal rice ratio scheme; Determine whether the quality difference between the test results and the standard quality data for rice noodles is greater than the preset error; If the test results of three consecutive batches of goods are greater than the preset error, the test results and the rice quality data corresponding to the optimal rice ratio scheme will be used as training data to dynamically update the regression coefficients in the rice noodle quality prediction model.

8. The method for proportioning raw materials for rice noodle processing based on rice quality according to claim 1, characterized in that, Also includes: The gelatinization temperature and peak time of different types of ingredients in the obtained optimal rice formulation; The dispersion of the gelatinization characteristics of mixtures based on the gelatinization temperature and peak time of different types of ingredients is shown below: In the formula, Indicates the dispersion of gelatinization temperature. Indicates the dispersion of peak time. and These represent the optimal rice blending scheme. The ingredients and the first The gelatinization temperature of each ingredient. and These represent the optimal rice blending scheme. The ingredients and the first Peak time of each ingredient and These represent the optimal rice blending scheme. The ingredients and the first The proportions of each ingredient Indicates the preset ratio threshold; Determine whether the dispersion of the gelatinization characteristics meets the preset consistency condition; if it does, execute the optimal rice ratio scheme; if it does not, adjust the optimal rice ratio scheme to obtain the target rice ratio scheme.

9. The method for proportioning raw materials for rice noodle processing based on rice quality according to claim 8, characterized in that, The optimal rice blending scheme is adjusted to obtain the target rice blending scheme, which includes: Ingredients with a proportion greater than a preset threshold are selected as adjustment ingredients; Calculate and adjust the deviation of each ingredient in the batch, as shown below: In the formula, Indicates the first The deviation in the gelatinization temperature of each ingredient. Indicates the first The deviation of the peak time of each ingredient This indicates adjusting the average gelatinization temperature of the ingredients. This indicates the average peak time for adjusting the ingredients; The ingredient with the largest deviation in gelatinization temperature was selected as the gelatinization ingredient, and the ingredient with the largest deviation in peak time was selected as the peak ingredient. Determine if the gelatinized ingredient and the peak ingredient are the same; if they are the same, reduce the preset proportion of the gelatinized ingredient. If they are not the same, the ingredient with the larger deviation between the gelatinized ingredient and the peak ingredient is regarded as the main outlier, and the ingredient with the smaller deviation between the gelatinized ingredient and the peak ingredient is regarded as the secondary outlier. While keeping the proportion of the main outlier ingredient constant, the proportion of the secondary outlier ingredient is gradually shifted to the ingredient with the opposite gelatinization characteristics at a preset step size.

Images