Dynamic proportion control system and method in prebiotic production process

Abstract

The invention discloses a dynamic proportion control system and method in a prebiotic production process, relates to the technical field of production process control, and continuously reflects actual states of a raw material component proportion Red, a fermentation tank temperature Tem, a pH value Phv, a dissolved oxygen content Oxy and a flora metabolism rate Met in a current batch based on a real-time production state set Sta. Linkage adjustment is carried out on production parameters through a multivariable optimization input index set Inp and a current optimal dynamic matching strategy Opt, so that parameter changes in the production process are not mutually-separated local repair any more, but overall cooperative control is carried out around the same target; in this way, the problems that in the prior art, a fixed proportion is difficult to adapt to raw material batch fluctuation, parameter changes in the fermentation process cannot be responded in time, and finally the product quality and the output efficiency are unstable can be directly solved. And meanwhile, a closed-loop control state set Clo is formed in combination with a feedback signal Fbk in the execution process to continue correction.

Description

Technical Field

[0001] This invention relates to the field of production process control technology, specifically to a dynamic proportioning control system and method for prebiotic production process. Background Technology

[0002] As an important functional food component that promotes the proliferation of beneficial intestinal bacteria, the production of prebiotics involves not only the rational proportioning of raw materials but also complex technological processes such as fermentation, conversion, and post-processing. In actual industrial production, the types and components of prebiotics are diverse, and a single fixed proportion is difficult to meet the changing needs of different batches of raw materials and fermentation states.

[0003] In current production practices, prebiotic raw materials are mostly oligosaccharide combinations or fructooligosaccharide mixtures, and the production process typically employs fixed feed ratios and preset fermentation conditions. This static formulation model has significant limitations: on the one hand, different batches of raw materials vary in purity, moisture content, and active components, and a fixed ratio cannot adequately compensate for these differences; on the other hand, kinetic parameters, microbial metabolic rates, and environmental factors (temperature, pH) during fermentation are constantly changing, but traditional processes lack real-time response mechanisms, leading to significant fluctuations in product indicators and low utilization rates of functional components. Therefore, the dynamic interaction and optimization needs among multiple variables in the production process cannot be effectively met by existing static or single-objective control methods. A formulation control scheme capable of achieving multi-parameter synergistic optimization and dynamic closed-loop adjustment is needed to solve this problem. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a dynamic proportioning control system and method for prebiotic production processes, which solves the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a method for dynamic proportioning control in the prebiotic production process, comprising the following steps: S1. Collect production parameters during the prebiotic production process, including the proportion of raw material components Red, fermentation tank temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met, and form a real-time production status set Sta. S2. Based on the real-time production status set Sta, perform data preprocessing and normalization on each parameter to generate a multivariate optimization input index set Inp, which is used to represent the state relationship of each parameter with production status and product index, and is used to calculate the optimal dynamic ratio strategy Opt. S3. Based on the multivariate optimization input index set Inp, calculate the current optimal dynamic ratio strategy Opt, so that the real-time production state set Sta reaches the preset target, thereby controlling the product quality Qul, activity Aci and yield Yie within the expected range. S4. The optimal dynamic ratio strategy Opt is sent to the production execution unit to dynamically adjust the proportion of raw material components Red, fermentation tank temperature Tem, pH value Phv, and dissolved oxygen content Oxy. The ratio control is adjusted in real time according to the feedback signal Fbk generated during the execution process to form a closed-loop control state set Clo, which is used to monitor and dynamically adjust the ratio execution effect. S5. Based on the closed-loop control state set Clo, calculate the actual values ​​of product quality Qul, activity Aci, and yield Yie online, and generate an optimized dynamic ratio strategy Adj by adjusting the optimal dynamic ratio strategy Opt and the control strategy correction algorithm. This optimized dynamic ratio strategy Adj is used to adjust the production process parameters so that product quality Qul, activity Aci, and yield Yie are close to the preset targets.

[0006] Preferably, S1 includes S11; S11. During the prebiotic production process, online sensors and sampling units installed in the raw material feeding device and fermenter are used to sequentially collect data on the raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met. Among them, the raw material component ratio Red is obtained by measuring the input of each raw material using a raw material flow meter and weighing equipment and calculating the proportion; The fermenter temperature Tem is obtained by measuring the temperature values ​​at different heights of the tank using multi-point temperature sensors and calculating the average value. pH value Phv and dissolved oxygen content Oxy were continuously collected using an online pH electrode and dissolved oxygen sensor, and noise was eliminated using a smoothing filtering algorithm. The metabolic rate of the microbial community (Met) was obtained by measuring the changes in the concentration of key metabolites through online sampling and calculating it in combination with the rate of change in bacterial cell concentration. The raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met are combined according to a predefined data structure and order to generate a real-time production status set Sta.

[0007] Preferably, S2 includes S21; S21. Read the production parameter data from the real-time production status set Sta, including the raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met. Then preprocessing is performed, including unit conversion, dimension unification, noise filtering, and outlier identification; Among them, unit conversion is used to unify data from different measurement systems into standard engineering units, such as unifying temperature into degrees Celsius and concentration into grams per liter; Noise filtering uses a moving average to smooth continuously sampled data, reducing the impact of instantaneous sensor fluctuations on subsequent calculations; Outlier identification uses a dynamic thresholding method: the mean (Mean) is calculated using the valid historical data of each parameter, and a multiple threshold is set. If the current collected value is greater than three times the mean or less than three times the mean, it is determined to be an outlier. For data determined to be outlier, the valid data before and after the parameter is replaced by an interpolation algorithm. Obtain the preprocessed real-time production status set Sta, defined as the processed production parameter dataset.

[0008] Preferably, S2 further includes S22; S22. Based on the processed production parameter dataset, normalize each production parameter, then match it with a preset weight value for correction, and generate an indicator vector. The raw material component ratio Red is normalized and mapped to a standard range of 0 to 1. Then, the normalized raw material component ratio Red value is multiplied by a pre-set weight value to obtain the raw material component ratio Red index vector RedVec. The weight value is set according to the production standard and is used to reflect the contribution of the raw material component ratio Red to the production status and product indicators. Simultaneously, the fermenter temperature Tem is normalized and multiplied by the corresponding weight value to obtain the fermenter temperature Tem index vector TemVec; the pH value Phv is normalized and multiplied by the corresponding weight value to obtain the pH value Phv index vector PhvVec; the dissolved oxygen content Oxy is normalized and multiplied by the corresponding weight value to obtain the dissolved oxygen content Oxy index vector OxyVec; and the microbial metabolic rate Met is normalized and multiplied by the corresponding weight value to obtain the microbial metabolic rate Met index vector MetVec. Each index vector corresponds to its respective production parameter and is used to represent the degree of contribution of that production parameter to the production status and product indicators. Subsequently, the raw material component ratio Red index vector RedVec, fermenter temperature Tem index vector TemVec, pH value Phv index vector PhvVec, dissolved oxygen content Oxy index vector OxyVec, and microbial metabolic rate Met index vector MetVec are combined in a predefined order to generate a multivariate optimization input index set Inp.

[0009] Preferably, S3 includes S31; S31. Based on the multivariate optimization input index set Inp, read each index vector, including the raw material component ratio Red index vector RedVec, fermenter temperature Tem index vector TemVec, pH value Phv index vector PhvVec, dissolved oxygen content Oxy index vector OxyVec, and microbial metabolic rate Met index vector MetVec. Then, the data is input into the pre-trained dynamic optimization calculation model in a predefined order to generate the current optimal dynamic ratio strategy Opt. The dynamic optimization calculation model consists of three parts: Predictive model: Based on historical production data and experimental data, it is used to simulate the impact of various parameter combinations on production status and product indicators; the predictive model can adopt regression model, response surface model or data-driven machine learning model to predict the results of different parameter combinations; Objective function: used to quantify the production process objectives, transforming the deviations of production status and product indicators from preset target values ​​into optimization indicators. The objective function is defined as the weighted square of the deviations of each product indicator, used to guide the algorithm to find the optimal strategy. Constraints: These are used to limit the reasonable range of various production parameters and process restrictions, including the upper and lower limits of the raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met.

[0010] Preferably, S3 further includes S32; S32. Specifically, the generation of the current optimal dynamic ratio strategy Opt is as follows: the index vectors output by the dynamic optimization calculation model are compared with the input index vectors, and the percentage change corresponding to each index vector is calculated. Based on this percentage change, the corresponding production parameters in the real-time production status set Sta are adjusted to form the updated production status. The updated production status is recombined to generate the current optimal dynamic ratio strategy Opt. At the same time, the current optimal dynamic ratio strategy Opt is directly issued to the production execution unit to dynamically adjust the production process so that the product indicators are close to the preset target.

[0011] Preferably, S4 includes S41; S41. The generated current optimal dynamic ratio strategy Opt is sent to the production execution unit. The production execution unit dynamically adjusts the production process according to the combination of parameters in the strategy, including adjusting the raw material component ratio Red, fermenter temperature Tem, pH value Phv and dissolved oxygen content Oxy. During execution, the production execution unit continuously monitors and adjusts the parameter values ​​in the strategy according to the preset control frequency, adjusts each production parameter step by step according to the preset control frequency, and compares the deviation between the actual value and the target value in each control cycle to reduce the corresponding deviation.

[0012] Preferably, S4 further includes S42; S42. While performing adjustments, the production execution unit continuously collects production process feedback signals Fbk, which include the actual values ​​and dynamic changes of each parameter during the execution process. Based on the feedback signal Fbk, the production parameter adjustment amount for the current cycle is corrected in real time, the adjusted production parameters are updated in the real-time production state set Sta, and a closed-loop control state set Clo is generated to monitor the matching status of each parameter with the strategy and the adjustment effect.

[0013] Preferably, S5 includes S51; S51. Based on the closed-loop control state set Clo, read the pre-state product detection data of the corresponding batch before the execution of the current optimal dynamic ratio strategy Opt and the post-state product detection data of the corresponding batch after the execution of the current optimal dynamic ratio strategy Opt, and calculate the actual values ​​of product quality Qul, activity Aci and yield Yie before and after execution based on the pre-state product detection data and the post-state product detection data. The product quality Qul is calculated based on the content of the target prebiotic component and the content of the impurity component; The active Aci is calculated based on the functional activity test results corresponding to the target prebiotic; The yield (Yie) is calculated based on the actual output of the target prebiotic and the corresponding input of raw materials. Subsequently, the actual values ​​of product quality Qul, activity Aci, and yield Yie after execution were compared with the actual values ​​of product quality Qul, activity Aci, and yield Yie before execution to obtain the corresponding quality change rate, activity change rate, and yield change rate. The rate of change in quality, the rate of change in activity, and the rate of change in yield are compared with preset fluctuation ranges, respectively. When at least two of the quality change rate, activity change rate and yield change rate fall within the positive fluctuation range for a preset number of consecutive times, the current optimal dynamic ratio strategy Opt is determined to be effective, and the current optimal dynamic ratio strategy Opt continues to be executed. When at least two of the quality change rate, activity change rate and yield change rate fall within the negative fluctuation range for a preset number of consecutive times, the current optimal dynamic ratio strategy Opt is determined to be invalid. Based on the percentage of negative fluctuation, the corresponding parameter adjustment amount in the current optimal dynamic ratio strategy Opt is corrected, and the optimized dynamic ratio strategy Adj is output. When the rate of change in quality, the rate of change in activity, and the rate of change in yield are within a preset fluctuation range and have not reached a preset number of consecutive times, the current monitoring status is maintained and the detection data for the next cycle is collected.

[0014] A dynamic ratio control system for prebiotic production process includes a production parameter acquisition module, a parameter data processing module, a dynamic ratio generation module, a ratio distribution module, and a ratio verification and decision module. The production parameter acquisition module collects production parameters during the prebiotic production process, including the proportion of raw material components Red, fermentation tank temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met, and forms a real-time production status set Sta. The parameter data processing module preprocesses and normalizes each parameter based on the real-time production status set Sta, generating a multivariate optimization input index set Inp, which represents the state relationship between each parameter and the production status and product index, and is used to calculate the optimal dynamic ratio strategy Opt. The dynamic ratio generation module calculates the current optimal dynamic ratio strategy Opt based on the multivariate optimization input index set Inp, so that the real-time production state set Sta reaches the preset target, thereby controlling the product quality Qul, activity Aci and yield Yie within the expected range. The ratio distribution module distributes the optimal dynamic ratio strategy Opt to the production execution unit, dynamically adjusting the raw material component ratio Red, fermentation tank temperature Tem, pH value Phv, and dissolved oxygen content Oxy. Based on the feedback signal Fbk generated during the execution process, the ratio control is adjusted in real time to form a closed-loop control state set Clo, which is used to monitor and dynamically adjust the ratio execution effect. The ratio verification decision module calculates the actual values ​​of product quality Qul, activity Aci, and yield Yie online based on the closed-loop control state set Clo. It then generates an optimized dynamic ratio strategy Adj by adjusting the optimal dynamic ratio strategy Opt and the control strategy correction algorithm. This optimized dynamic ratio strategy Adj is used to adjust the production process parameters so that product quality Qul, activity Aci, and yield Yie are close to the preset targets.

[0015] This invention provides a dynamic proportioning control system and method for prebiotic production, which has the following beneficial effects: (1) Based on the real-time production status set Sta, the actual status of the raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met in the current batch is continuously reflected. Then, the production parameters are adjusted in a coordinated manner through the multivariate optimization input index set Inp and the current optimal dynamic ratio strategy Opt. This makes the parameter changes in the production process no longer isolated local repairs, but overall coordinated control around the same goal. This can directly improve the problems in the existing technology where the fixed ratio is difficult to adapt to the fluctuation of raw material batches, the parameter changes during fermentation cannot be responded to in a timely manner, and the product quality and output efficiency are unstable. When the actual utilization performance of the target component in a certain batch of raw materials decreases and the microbial metabolic rate Met drops during fermentation, it is not just a single parameter that is compensated. Instead, the raw material component ratio Red and fermentation conditions are adjusted synchronously according to the current optimal dynamic ratio strategy Opt. During the execution process, the feedback signal Fbk is combined to form a closed-loop control status set Clo for further correction. This can make the product quality Qul, activity Aci, and yield Yie after execution maintain a more stable upward trend compared with before execution.

[0016] (2) By optimizing the input index set Inp through multivariate optimization, it is further transformed into the current optimal dynamic ratio strategy Opt, which can be directly applied to the production parameter adjustment level. This solves the practical problem in existing production where parameter changes can be observed but cannot be converted into an executable adjustment scheme. Using the pre-trained dynamic optimization calculation model, the raw material component ratio Red index vector RedVec, fermenter temperature Tem index vector TemVec, pH value Phv index vector PhvVec, dissolved oxygen content Oxy index vector OxyVec, and microbial metabolic rate Met index vector MetVec are combined with historical production patterns, objective functions, and constraints to directly output the optimization direction. The model output result is then compared with the index vectors before input, and finally recombined to generate the current optimal dynamic ratio strategy Opt. The strategy obtained in this way is not a mathematical result divorced from the actual process conditions, but an executable parameter combination that corresponds one-to-one with the current production state. This is further transformed into a basis for actual executable parameter adjustment.

[0017] (3) By truly implementing the current optimal dynamic ratio strategy Opt into production execution, instead of directly jumping to the target value all at once by adjusting the raw material component ratio Red, fermenter temperature Tem, pH value Phv and dissolved oxygen content Oxy, the parameters are adjusted step by step according to the preset control frequency. The deviation between the actual value and the target value is compared in each control cycle, making the parameter adjustment process more stable and reducing the process fluctuations caused by equipment response lag or sudden adjustment. Then, the feedback signal Fbk is used to correct the production parameter adjustment amount in the current cycle in real time, and the corrected result is updated to the real-time production state set Sta, forming a closed-loop control state set Clo. This makes the system not only execute the strategy, but also continuously judge how much the execution result is still different from the target. Attached Figure Description

[0018] Figure 1 This is a schematic diagram illustrating the steps of a dynamic ratio control method for prebiotic production process according to the present invention; Figure 2 This is a schematic diagram of a dynamic proportioning control system for the prebiotic production process according to the present invention. Figure 3 A schematic diagram is generated for the closed-loop execution, verification, and optimization of the dynamic matching strategy Adj. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0020] Example 1 This invention provides a method for dynamic ratio control in the production process of prebiotics. Please refer to [link / reference]. Figure 1 This includes the following steps: S1. Collect production parameters during the prebiotic production process, including the proportion of raw material components Red, fermentation tank temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met, and form a real-time production status set Sta. S2. Based on the real-time production status set Sta, perform data preprocessing and normalization on each parameter to generate a multivariate optimization input index set Inp, which is used to represent the state relationship of each parameter with production status and product index, and is used to calculate the optimal dynamic ratio strategy Opt. S3. Based on the multivariate optimization input index set Inp, calculate the current optimal dynamic ratio strategy Opt, so that the real-time production state set Sta reaches the preset target, thereby controlling the product quality Qul, activity Aci and yield Yie within the expected range. S4. The optimal dynamic ratio strategy Opt is sent to the production execution unit to dynamically adjust the proportion of raw material components Red, fermentation tank temperature Tem, pH value Phv, and dissolved oxygen content Oxy. The ratio control is adjusted in real time according to the feedback signal Fbk generated during the execution process to form a closed-loop control state set Clo, which is used to monitor and dynamically adjust the ratio execution effect. S5. Based on the closed-loop control state set Clo, calculate the actual values ​​of product quality Qul, activity Aci, and yield Yie online, and generate an optimized dynamic ratio strategy Adj by adjusting the optimal dynamic ratio strategy Opt and the control strategy correction algorithm. This optimized dynamic ratio strategy Adj is used to adjust the production process parameters so that product quality Qul, activity Aci, and yield Yie are close to the preset targets.

[0021] In this embodiment, through the above steps, the prebiotic production process no longer relies on a fixed feed ratio and single-set fermentation conditions. Instead, it first continuously reflects the actual state of the current batch's raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met based on the real-time production status set Sta. Then, it uses a multivariate optimization input index set Inp and the current optimal dynamic ratio strategy Opt to adjust the production parameters in a coordinated manner. This ensures that parameter changes during the production process are no longer isolated local repairs, but rather overall coordinated control around the same goal. This can directly improve the problems in the existing technology where fixed ratios are difficult to adapt to batch fluctuations in raw materials, parameter changes during fermentation cannot be responded to in a timely manner, and ultimately product quality and output efficiency are unstable. Specifically, when the actual utilization of the target component in a batch of raw materials declines, and the metabolic rate of the microbial community (Met) drops during fermentation, the solution does not only compensate for a single parameter. Instead, it simultaneously adjusts the proportion of raw material components (Red) and fermentation conditions based on the current optimal dynamic ratio strategy (Opt). During execution, it combines the feedback signal (Fbk) to form a closed-loop control state set (Clo) for further correction. This ensures that the product quality (Qul), activity (Aci), and yield (Yie) after execution maintain a more stable upward trend compared to before execution. For common situations in real-world production, such as product fluctuations in the same prebiotic production line due to changes in raw material condition, in-tank mass transfer, or metabolic activity in the morning and afternoon, this solution can compare the before-and-after state detection data to determine whether the current optimal dynamic ratio strategy (Opt) is effective. If ineffective, it outputs the optimized dynamic ratio strategy (Adj) for further correction, thereby reducing batch-to-batch purity fluctuations, activity declines, and yield drops. This makes production results more continuous, stable, and meets the consistency and traceability requirements of actual industrial production.

[0022] Example 2 Specifically: S1 includes S11; S11. During the prebiotic production process, online sensors and sampling units installed in the raw material feeding device and fermenter are used to sequentially collect data on the raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met. Among them, the raw material component ratio Red is obtained by measuring the input of each raw material using a raw material flow meter and weighing equipment and calculating the proportion; The fermenter temperature Tem is obtained by measuring the temperature values ​​at different heights of the tank using multi-point temperature sensors and calculating the average value. pH value Phv and dissolved oxygen content Oxy were continuously collected using an online pH electrode and dissolved oxygen sensor, and noise was eliminated using a smoothing filtering algorithm. The metabolic rate of the microbial community (Met) was obtained by measuring the changes in the concentration of key metabolites through online sampling and calculating it in combination with the rate of change in bacterial cell concentration. According to a predefined data structure and order, the raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met are combined to generate a real-time production status set Sta. It should be noted that: pH (Phv) and dissolved oxygen (Oxy) are continuously collected using online pH electrodes and dissolved oxygen sensors installed inside the fermenter. The data acquisition equipment needs to communicate with the production control system in real time to obtain the instantaneous values ​​at each measurement point. To ensure data stability and reduce occasional interference, the acquired signals should be smoothed and filtered before transmission, using methods such as moving averages or exponentially weighted averages to process the continuous sampling points, thereby eliminating instantaneous fluctuations and sensor noise. In this way, the obtained pH (Phv) and dissolved oxygen (Oxy) values ​​can reflect the acid-base state and dissolved oxygen level of the liquid environment in the fermenter in real time and continuously, providing stable and directly usable input data for dynamic proportioning control. The microbial metabolic rate (Met) is obtained by online sampling and analysis of changes in the concentration of key metabolites. Specifically, during fermentation, culture medium samples are periodically collected using an online sampler, and the concentrations of key metabolites are measured using an online analyzer (such as HPLC or optical density analyzer). Based on continuous sampling data, combined with the rate of change in cell concentration (obtainable through optical density OD values ​​or other online cell measurement methods) within the same time period, the rate of metabolite production per unit time is calculated, forming the microbial metabolic rate (Met) data.

[0023] In this embodiment, through S11, the solution does not merely complete general parameter acquisition, but simultaneously acquires the raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met in the same acquisition link, and further combines them to generate a real-time production status set Sta. This allows subsequent control to no longer rely on a single temperature value or a single feed value for local judgment, but can identify the current fermentation status based on the correspondence of multiple parameters at the same time. This has a direct effect on solving the problem in existing production where parameters are collected but are fragmented and difficult to form a unified judgment basis. In particular, by continuously collecting pH value (Phv) and dissolved oxygen content (Oxy) and performing smoothing filtering before transmission, and combining this with the microbial metabolic rate (Met) obtained through changes in the concentration of key metabolites and the rate of change in cell concentration, the interference of instantaneous fluctuations, sensor noise, or single sampling deviations on state judgment can be reduced. This makes the resulting real-time production state set (Sta) closer to the continuous changes in the actual fermentation process. For example, in actual production, the same fermenter may experience aeration disturbances or localized uneven stirring in a short period of time. If only a single dissolved oxygen content (Oxy) reading is used, it is easy to misjudge the oxygen supply status. However, this scheme also introduces the processed pH value (Phv) and the microbial metabolic rate (Met) for joint reflection, which can more accurately distinguish between short-term disturbances and real metabolic changes. This provides a more stable and reliable data foundation for the subsequent generation of the multivariate optimization input index set (Inp). This benefit is difficult to achieve with ordinary single-parameter acquisition or simple batch recording methods.

[0024] Example 3 Specifically: S2 includes S21; S21. Read the production parameter data from the real-time production status set Sta, including the raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met. Then preprocessing is performed, including unit conversion, dimension unification, noise filtering, and outlier identification; Among them, unit conversion is used to unify data from different measurement systems into standard engineering units, such as unifying temperature into degrees Celsius and concentration into grams per liter; Noise filtering uses a moving average to smooth continuously sampled data, reducing the impact of instantaneous sensor fluctuations on subsequent calculations; Outlier identification uses a dynamic thresholding method: the mean (Mean) is calculated using the valid historical data of each parameter, and a multiple threshold (e.g., three times) is set. If the current collected value is greater than three times the mean or less than three times the mean, it is determined to be an outlier. For data determined to be outlier, the valid data before and after is used to replace it through an interpolation algorithm, or the valid data value at the previous moment is used to cover it, so as to ensure data continuity and availability. Obtain the preprocessed real-time production status set Sta, defined as the processed production parameter dataset.

[0025] S2 further includes S22; S22. Based on the processed production parameter dataset, normalize each production parameter, then match it with a preset weight value for correction, and generate an indicator vector. The raw material component ratio Red is normalized and mapped to a standard range of 0 to 1. Then, the normalized raw material component ratio Red value is multiplied by a pre-set weight value to obtain the raw material component ratio Red index vector RedVec. The weight values ​​are set according to production standards and are used to reflect the contribution of the proportion of raw material components Red to the production status and product indicators (including product quality Qul, activity Aci and yield Yie). Simultaneously, the fermenter temperature Tem is normalized and multiplied by the corresponding weight value to obtain the fermenter temperature Tem index vector TemVec; the pH value Phv is normalized and multiplied by the corresponding weight value to obtain the pH value Phv index vector PhvVec; the dissolved oxygen content Oxy is normalized and multiplied by the corresponding weight value to obtain the dissolved oxygen content Oxy index vector OxyVec; and the microbial metabolic rate Met is normalized and multiplied by the corresponding weight value to obtain the microbial metabolic rate Met index vector MetVec. Each index vector corresponds to its respective production parameter and is used to represent the degree of contribution of that production parameter to the production status and product indicators. Subsequently, the raw material component ratio Red index vector RedVec, fermenter temperature Tem index vector TemVec, pH value Phv index vector PhvVec, dissolved oxygen content Oxy index vector OxyVec, and microbial metabolic rate Met index vector MetVec are combined in a predefined order to generate a multivariate optimization input index set Inp. The multivariate optimization input index set Inp can fully represent the quantitative relationship between each production parameter and the current production state and product index, and serve as the structured input for calculating the optimal dynamic ratio strategy Opt. It should be noted that: Each production parameter (such as the proportion of raw material components Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met) needs to be matched with a corresponding weight value before generating the index vector. This weight value reflects the contribution of the parameter to the production status and product indicators (including product quality Qul, activity Aci, and yield Yie). The acquisition and application of this weight value can be implemented according to the following methods: Weight values ​​are determined by analyzing the actual contribution of each production parameter to product indicators based on historical production data and experimental verification results.

[0026] For parameters with a significant impact, a higher weight value can be set; for parameters with a smaller impact, a lower weight value can be set.

[0027] The weight values ​​can be standardized values ​​of 0 to 1 to ensure that the normalization interval is not changed when the index vector is multiplied.

[0028] Matching method: After normalization, the normalized value of each parameter is matched with its preset weight value.

[0029] Implementers can automatically match parameters by looking up tables or controlling the system to read the weight values ​​corresponding to the parameter names.

[0030] For example, if the weight of the raw material component ratio Red is set to 0.4 in the control system, then the normalized raw material component ratio Red value multiplied by 0.4 will yield the raw material component ratio Red index vector RedVec.

[0031] The effect of multiplying by the weight value: The index vector obtained after multiplying by the weight value can quantify the relative importance of each parameter to the production status and product index, so that the optimization algorithm can fully consider the contribution of each parameter when calculating the optimal dynamic ratio strategy Opt.

[0032] In this embodiment, the special advantage of this solution achieved through S21 and S22 is not simply in collecting more parameters, but in processing production data from different sources, with different dimensions, and different stability into a unified input that is comparable, usable, and continuously calculable. Then, it is further compressed into a multivariate optimization input index set Inp that can be directly entered into the optimization model. This solves the practical problems in the existing production process where, although parameters have been collected, they cannot be directly compared, outliers interfere with the judgment, and the importance of different parameters in the results cannot be distinguished. Specifically, step S21 first performs unit conversion, dimension unification, noise filtering, and outlier identification on the raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met. This ensures that data from different sources are organized into a processed production parameter dataset under the same processing standard, preventing amplification and misjudgment of certain parameters due to unit differences, short-term fluctuations, or occasional outliers in subsequent optimization. Subsequently, step S22 converts each production parameter into a raw material component ratio Red index vector RedVec, a fermenter temperature Tem index vector TemVec, a pH value Phv index vector PhvVec, a dissolved oxygen content Oxy index vector OxyVec, and a microbial metabolic rate Met index vector MetVec. By using preset weight values, the contribution differences of each parameter to product quality Qul, activity Aci, and yield Yie are clearly expressed. In this way, the subsequently generated multivariate optimization input index set Inp is no longer just a stack of organized raw data, but a structured input with parameter priority and contribution differentiation. In practical terms, for example, in actual production, a certain batch might experience significant fluctuations in the fermenter temperature (Tem) but have limited impact on the current batch's product, while the microbial metabolic rate (Met) might only change slightly but directly lead to a subsequent decrease in activity. Without the processing described in S21 and S22, the control system could easily misjudge the former as the main disturbance. However, this solution, through preprocessing and weighted correction, allows the multivariate optimization input index set (Inp) to more accurately reflect which parameter is more worthy of priority adjustment. This enables the subsequent generation of the current optimal dynamic ratio strategy (Opt) to be based on more stable and discriminative data. This benefit differs from the benefits of simultaneous acquisition of multiple parameters and accurate reflection of the status mentioned earlier; it focuses more on transforming messy production data that cannot be directly used for decision-making into effective inputs that can be directly used for optimization decisions.

[0033] Example 4 Specifically: S3 includes S31; S31. Based on the multivariate optimization input index set Inp, read each index vector, including the raw material component ratio Red index vector RedVec, fermenter temperature Tem index vector TemVec, pH value Phv index vector PhvVec, dissolved oxygen content Oxy index vector OxyVec, and microbial metabolic rate Met index vector MetVec. Then, the data is input into the pre-trained dynamic optimization calculation model in a predefined order to generate the current optimal dynamic ratio strategy Opt. The dynamic optimization calculation model consists of three parts: Predictive model: Based on historical production data and experimental data, it is used to simulate the impact of various parameter combinations on production status and product indicators (including product quality Qul, activity Aci, and yield Yie); the predictive model can use regression model, response surface model, or data-driven machine learning model to predict the results of different parameter combinations; Objective function: used to quantify the production process objectives, transforming the deviations of production status and product indicators from preset target values ​​into optimization indicators. The objective function is defined as the weighted square of the deviations of each product indicator, used to guide the algorithm to find the optimal strategy. Constraints: These are used to limit the reasonable range and process restrictions of each production parameter, including the upper and lower limits of the raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met, to ensure that the strategy obtained from the optimization calculation meets the production feasibility. It should be noted that: In the dynamic proportioning control method for prebiotic production, the dynamic optimization calculation model includes a prediction model, an objective function, and constraints. Its training and finalization can be achieved as follows: Predictive model training and finalization: Data Preparation: Collect historical production and experimental data, including actual values ​​of each production parameter (raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met) for each batch or experimental run, and corresponding product indicators (product quality Qul, activity Aci, yield Yie). Ensure data covers different parameter combinations and production conditions, including high, low, and medium parameter levels. Training process: Using regression models, response surface models, or data-driven machine learning models (such as random forests, support vector machines, or neural networks), historical production parameters are used as inputs, and product indicators are used as outputs to fit the model. Cross-validation or batch training methods can be employed during training to ensure the model has good predictive ability for different parameter combinations. Modeling and Validation: After training, the prediction accuracy is validated using an independent validation set or a new batch of data. The model hyperparameters are adjusted or new feature inputs are selected based on the prediction error until the prediction error meets production requirements. Once the model is finalized, it can serve as the basis for the relationship between input indicators and product results in the optimization algorithm, and can be directly used to predict the impact of different parameter combinations on product indicators. Definition and shaping of the objective function: Objective function construction: The deviations of production status and product indicators from preset target values ​​are quantified to form calculable optimization indicators. The objective function can be defined as the weighted sum of squares of the deviations of each product indicator, where the weight of each product indicator is set according to production requirements and indicator importance to ensure that the optimization algorithm can focus on key indicators.

[0034] Implementation: Based on the target product specifications, the implementers determine the target values ​​and weights of each product indicator. After constructing the formula, it can be directly called in the optimization algorithm to evaluate the merits of each parameter combination.

[0035] Constraint setting and verification: Constraint construction: Based on process limitations and equipment capabilities, set allowable upper and lower limits for each production parameter, including the proportion of raw material components Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met.

[0036] Verification and Adjustment: By simulating different parameter combinations, the feasibility of the constraints is verified, ensuring that all parameter combinations output by the optimization calculation can be directly implemented in production without exceeding the equipment capacity or process safety range.

[0037] S3 further includes S32; S32. Specifically, the generation of the current optimal dynamic ratio strategy Opt is as follows: the index vectors output by the dynamic optimization calculation model are compared with the input index vectors, and the percentage change corresponding to each index vector is calculated. Based on this percentage change, the corresponding production parameters in the real-time production status set Sta are adjusted to form the updated production status. The updated production status is recombined to generate the current optimal dynamic ratio strategy Opt. At the same time, the current optimal dynamic ratio strategy Opt is directly issued to the production execution unit to dynamically adjust the production process so that the product indicators (including product quality Qul, activity Aci and yield Yie) are close to the preset target.

[0038] In this embodiment, the special advantage of this solution achieved through S31 and S32 is not simply that the data is fed into the model, but that the multivariate optimization input index set Inp formed in the previous step is further transformed into the current optimal dynamic ratio strategy Opt that can be directly implemented at the level of production parameter adjustment. This solves the practical problem in existing production where parameter changes can be seen, but the parameter changes cannot be transformed into an executable adjustment plan. Specifically, step S31 utilizes a pre-trained dynamic optimization calculation model to combine the raw material component ratio Red index vector RedVec, fermenter temperature Tem index vector TemVec, pH value Phv index vector PhvVec, dissolved oxygen content Oxy index vector OxyVec, and microbial metabolic rate Met index vector MetVec with historical production patterns, objective functions, and constraints. Instead of relying on manual experience to determine which parameter should be adjusted first and by how much, it directly outputs the optimization direction. Step S32 then compares the model output with the index vectors before input, calculates the percentage change for each index vector, and applies this percentage change back to the corresponding production parameters in the real-time production state set Sta. Finally, it recombines these parameters to generate the current optimal dynamic ratio strategy Opt. The resulting strategy is not a mathematical result divorced from actual process conditions, but rather an executable parameter combination that corresponds one-to-one with the current production state. In practical terms, for example, in actual production, a batch might simultaneously exhibit slightly low pH (Phv), insufficient dissolved oxygen (Oxy) at certain stages, and a decrease in the microbial metabolic rate (Met). Relying solely on manual experience often leads to prioritizing adjustments to a single oxygen supply or temperature parameter, easily overlooking the interconnected effects between these parameters. This solution, through steps S31 and S32, transforms the combined effects of these parameters into a comprehensive optimization adjustment scheme for the raw material component ratio (Red), fermenter temperature (Tem), pH (Phv), dissolved oxygen (Oxy), and microbial metabolic rate (Met). This ensures that the subsequent optimal dynamic ratio strategy (Opt) is constrained by historical patterns and limited by current process boundaries, thereby more specifically improving the efficiency of product quality (Qul), activity (Aci), and yield (Yie) converging towards the target state. This advantage differs from the previous steps, which focused on collecting real-world data and generating effective input; it more significantly transforms structured input into a basis for practically executable parameter adjustments.

[0039] Example 5 Please see Figure 3 Specifically: S4 includes S41; S41. The generated current optimal dynamic ratio strategy Opt is sent to the production execution unit. The production execution unit dynamically adjusts the production process according to the combination of parameters in the strategy, including adjusting the raw material component ratio Red, fermenter temperature Tem, pH value Phv and dissolved oxygen content Oxy. During execution, the production execution unit continuously monitors and adjusts the parameter values ​​in the strategy according to the preset control frequency, adjusts each production parameter step by step according to the preset control frequency, and compares the deviation between the actual value and the target value in each control cycle to reduce the corresponding deviation. It should be noted that: The production execution unit dynamically adjusts production parameters according to the issued current optimal dynamic proportioning strategy (Opt). Implementation personnel can implement this strategy in the following ways: Control frequency settings: In the production execution unit control system, preset the parameter update frequency, such as every 1 minute, 5 minutes, or set a fixed sampling period according to process requirements.

[0040] The control frequency should ensure the timeliness of parameter acquisition and adjustment, while not exceeding the equipment's response capability, in order to avoid oscillations or safety risks caused by excessively rapid adjustments.

[0041] Adjust parameters gradually: For each production parameter (including the proportion of raw material components Red, fermenter temperature Tem, pH value Phv, and dissolved oxygen content Oxy), the execution unit makes incremental adjustments according to the target value specified in the strategy.

[0042] Each adjustment can be set based on the difference between the current value and the target value, such as adjusting proportionally or using a fixed step size, to ensure that the parameter gradually approaches the strategy value rather than jumping all at once.

[0043] S4 also includes S42; S42. While performing adjustments, the production execution unit continuously collects production process feedback signals Fbk, which include the actual values ​​and dynamic changes of each parameter during the execution process. Based on the feedback signal Fbk, the production parameter adjustment amount for the current cycle is corrected in real time, the adjusted production parameters are updated to the real-time production status set Sta, and a closed-loop control status set Clo is generated to monitor the matching status and adjustment effect of each parameter and strategy. It should be noted that: During the dynamic adjustment process of S4.1, real-time values ​​of production parameters are continuously collected, and a feedback signal Fbk is generated. Implementers can execute the process as follows: Real-time parameter comparison: For each production parameter (including the proportion of raw material components Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met), the actual measured value is compared with the target value in the current optimal dynamic ratio strategy Opt, and the deviation is calculated.

[0044] Deviation can be expressed as a percentage or an absolute value, making it easier to determine the adjustment range.

[0045] Parameter fine-tuning: Based on the calculated deviation, each production parameter is fine-tuned step by step. For example, linear interpolation, fixed step size, or proportional correction can be used to gradually adjust the actual value closer to the strategy value.

[0046] During fine-tuning, the minimum step size and maximum adjustment range can be set to ensure smooth adjustment without exceeding the equipment's capabilities.

[0047] Update the real-time production status set Sta: The fine-tuned production parameters are updated to the real-time production status set Sta to keep the real-time status synchronized with the strategy objectives. Each update can record a timestamp and adjustment range for monitoring and traceability.

[0048] Generate the closed-loop control state set Clo: The adjusted real-time production status set Sta, along with the corresponding deviation information and feedback signal Fbk, are integrated to form the closed-loop control status set Clo.

[0049] The closed-loop control state set Clo can be used to monitor the matching of each parameter with the strategy objective and serve as input for subsequent dynamic matching strategy correction, thereby realizing closed-loop control and dynamic adjustment.

[0050] S5 includes S51; S51. Based on the closed-loop control state set Clo, read the pre-state product detection data of the corresponding batch before the execution of the current optimal dynamic ratio strategy Opt and the post-state product detection data of the corresponding batch after the execution of the current optimal dynamic ratio strategy Opt, and calculate the actual values ​​of product quality Qul, activity Aci and yield Yie before and after execution based on the pre-state product detection data and the post-state product detection data. The product quality Qul is calculated based on the content of the target prebiotic component and the content of the impurity component; The active Aci is calculated based on the functional activity test results corresponding to the target prebiotic; The yield (Yie) is calculated based on the actual output of the target prebiotic and the corresponding input of raw materials. Subsequently, the actual values ​​of product quality Qul, activity Aci, and yield Yie after execution were compared with the actual values ​​of product quality Qul, activity Aci, and yield Yie before execution to obtain the corresponding quality change rate, activity change rate, and yield change rate. The rate of change in quality, the rate of change in activity, and the rate of change in yield are compared with preset fluctuation ranges, respectively. When at least two of the quality change rate, activity change rate and yield change rate fall within the positive fluctuation range for a preset number of consecutive times, the current optimal dynamic ratio strategy Opt is determined to be effective, and the current optimal dynamic ratio strategy Opt continues to be executed. When at least two of the quality change rate, activity change rate and yield change rate fall within the negative fluctuation range for a preset number of consecutive times, the current optimal dynamic ratio strategy Opt is determined to be invalid. Based on the percentage of negative fluctuation, the corresponding parameter adjustment amount in the current optimal dynamic ratio strategy Opt is corrected, and the optimized dynamic ratio strategy Adj is output. When the rate of change in quality, the rate of change in activity, and the rate of change in yield are within a preset fluctuation range and have not reached a preset number of consecutive times, the current monitoring status is maintained and the next cycle of detection data is collected. It should be noted that: The product quality Qul is obtained based on the component detection results of the product samples within the current testing cycle; During implementation, before and after executing the current optimal dynamic proportioning strategy Opt, corresponding batches of samples are collected, and the content of the target prebiotic component Tar and the content of the impurity component Imp in the samples are detected by liquid chromatography. Subsequently, the content of the target prebiotic component Tar is divided by the sum of the content of the target prebiotic component Tar and the content of the impurity component Imp to obtain the product quality Qul. In this way, implementers can directly obtain the product quality Qul based on the actual detection results of the target component and impurity component in the sample, and use it for direct comparison of the quality status before and after execution.

[0051] The active Aci is obtained based on the promoting effect of the product sample on the target bacterial community within the current detection period; During implementation, before and after executing the current optimal dynamic ratio strategy (Opt), corresponding batches of samples were added to the same target bacterial culture system, while a blank control group without samples was set up. After completing the culture under the same inoculum size, culture temperature, and culture time, the increase in bacterial optical density and acid production in the sample group and the blank control group were measured respectively. The ratio of the increase in bacterial optical density and the increase in acid production in the sample group relative to the blank control group were averaged to obtain the activity Aci. In this way, the implementer can directly obtain the activity Aci based on the degree to which the sample promotes the proliferation and metabolism of the target bacterial population, and use it for direct comparison of the activity state before and after the execution.

[0052] The yield (Yie) is obtained based on the actual output of the target prebiotic and the amount of raw materials input during the current testing period. During implementation, the actual output (Out) of the target prebiotic in the current cycle is read through the production metering unit, and the raw material input (Raw) in the same cycle is read through the raw material metering unit. Then, the actual output (Out) of the target prebiotic is divided by the raw material input (Raw) to obtain the yield (Yie). In this way, implementers can directly obtain the yield (Yie) based on the production metering results and use it for direct comparison of output efficiency before and after execution.

[0053] In this embodiment, through S41, S42 and S51, the special benefits achieved by this solution are no longer limited to the ability to generate strategies, but further put the current optimal dynamic ratio strategy Opt into a continuous closed loop of production execution, feedback correction and effect verification. This solves the practical problem in existing production where even if an adjustment plan is given, it is difficult to determine whether it is really effective after execution, and it is difficult to make timely corrections after it becomes ineffective. Specifically, S41 does not directly jump to the target values ​​for the raw material component ratio Red, fermenter temperature Tem, pH value Phv, and dissolved oxygen content Oxy all at once. Instead, it adjusts them step by step according to a preset control frequency, comparing the deviation between the actual values ​​and the target values ​​in each control cycle. This makes the parameter adjustment process smoother and reduces process fluctuations caused by equipment response lag or sudden adjustments. S42 further uses the feedback signal Fbk to correct the production parameter adjustment amount of the current cycle in real time and updates the corrected result to the real-time production state set Sta, forming a closed-loop control state set Clo. This allows the system to not only execute the strategy but also continuously judge how much the execution result is different from the target. On this basis, S51 does not simply judge whether the indicators have increased after execution. Instead, it directly compares the actual values ​​of product quality Qul, activity Aci, and yield Yie before and after execution to obtain the quality change rate, activity change rate, and yield change rate. It then combines the preset fluctuation range and the preset number of consecutive times to determine whether the current optimal dynamic ratio strategy Opt is effective or ineffective. If ineffective, it outputs the optimized dynamic ratio strategy Adj for further correction. In practical terms, for example, in actual production, after a batch of products has executed the current optimal dynamic proportioning strategy Opt, the proportions of raw material components Red and dissolved oxygen content Oxy have been adjusted according to the strategy. However, due to the lag in the actual metabolic response within the fermenter, the product quality Qul shows a slight improvement, the activity Aci shows no significant change, and the yield Yie fluctuates slightly during the first test. Without the closed-loop verification mechanism described in S42 and S51, this execution could easily be misjudged as a failure or a success. This solution records the execution status through a closed-loop control state set Clo, and then judges the effectiveness of the strategy by measuring the rate of change in quality, activity, and yield over consecutive periods. This allows for the differentiation between short-term fluctuations and the true trend, thus avoiding premature abandonment of effective strategies and long-term maintenance of ineffective strategies. Ultimately, the output of the optimized dynamic proportioning strategy Adj is based on the actual execution results. This advantage differs from the previous steps, which focused on collecting actual states, constructing effective inputs, and generating executable strategies. It emphasizes the true closure of strategy execution, effect verification, and failure callback.

[0054] Example 6 A dynamic proportioning control system for the prebiotic production process, please refer to... Figure 2Specifically, it includes a production parameter acquisition module, a parameter data processing module, a parameter dynamic ratio generation module, a ratio distribution module, and a ratio verification and decision-making module. The production parameter acquisition module collects production parameters during the prebiotic production process, including the proportion of raw material components Red, fermentation tank temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met, and forms a real-time production status set Sta. The parameter data processing module preprocesses and normalizes each parameter based on the real-time production status set Sta, generating a multivariate optimization input index set Inp, which represents the state relationship between each parameter and the production status and product index, and is used to calculate the optimal dynamic ratio strategy Opt. The dynamic ratio generation module calculates the current optimal dynamic ratio strategy Opt based on the multivariate optimization input index set Inp, so that the real-time production state set Sta reaches the preset target, thereby controlling the product quality Qul, activity Aci and yield Yie within the expected range. The ratio distribution module distributes the optimal dynamic ratio strategy Opt to the production execution unit, dynamically adjusting the raw material component ratio Red, fermentation tank temperature Tem, pH value Phv, and dissolved oxygen content Oxy. Based on the feedback signal Fbk generated during the execution process, the ratio control is adjusted in real time to form a closed-loop control state set Clo, which is used to monitor and dynamically adjust the ratio execution effect. The ratio verification decision module calculates the actual values ​​of product quality Qul, activity Aci, and yield Yie online based on the closed-loop control state set Clo. It then generates an optimized dynamic ratio strategy Adj by adjusting the optimal dynamic ratio strategy Opt and the control strategy correction algorithm. This optimized dynamic ratio strategy Adj is used to adjust the production process parameters so that product quality Qul, activity Aci, and yield Yie are close to the preset targets.

[0055] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended technical solutions and their equivalents.

Claims

1. A method for dynamic proportioning control in the production process of prebiotics, characterized in that: Includes the following steps: S1. Collect production parameters during the prebiotic production process, including the proportion of raw material components Red, fermentation tank temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met, and form a real-time production status set Sta. S2. Based on the real-time production status set Sta, perform data preprocessing and normalization on each parameter to generate a multivariate optimization input index set Inp, which is used to represent the state relationship of each parameter with production status and product index, and is used to calculate the optimal dynamic ratio strategy Opt. S3. Based on the multivariate optimization input index set Inp, calculate the current optimal dynamic ratio strategy Opt, so that the real-time production state set Sta reaches the preset target, thereby controlling the product quality Qul, activity Aci and yield Yie within the expected range. S4. The optimal dynamic ratio strategy Opt is sent to the production execution unit to dynamically adjust the proportion of raw material components Red, fermentation tank temperature Tem, pH value Phv, and dissolved oxygen content Oxy. The ratio control is adjusted in real time according to the feedback signal Fbk generated during the execution process to form a closed-loop control state set Clo, which is used to monitor and dynamically adjust the ratio execution effect. S5. Based on the closed-loop control state set Clo, calculate the actual values ​​of product quality Qul, activity Aci, and yield Yie online, and generate an optimized dynamic ratio strategy Adj by adjusting the optimal dynamic ratio strategy Opt and the control strategy correction algorithm. This optimized dynamic ratio strategy Adj is used to adjust the production process parameters so that product quality Qul, activity Aci, and yield Yie are close to the preset targets.

2. The method for dynamic proportioning control in the prebiotic production process according to claim 1, characterized in that: S1 includes S11; S11. During the prebiotic production process, online sensors and sampling units installed in the raw material feeding device and fermenter are used to sequentially collect data on the raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met. Among them, the raw material component ratio Red is obtained by measuring the input of each raw material using a raw material flow meter and weighing equipment and calculating the proportion; The fermenter temperature Tem is obtained by measuring the temperature values ​​at different heights of the tank using multi-point temperature sensors and calculating the average value. pH value Phv and dissolved oxygen content Oxy were continuously collected using an online pH electrode and dissolved oxygen sensor, and noise was eliminated using a smoothing filtering algorithm. The metabolic rate of the microbial community (Met) was obtained by measuring the changes in the concentration of key metabolites through online sampling and calculating it in combination with the rate of change in bacterial cell concentration. The raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met are combined according to a predefined data structure and order to generate a real-time production status set Sta.

3. The method for dynamic proportioning control in the prebiotic production process according to claim 2, characterized in that: S2 includes S21; S21. Read the production parameter data from the real-time production status set Sta, including the raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met. Then preprocessing is performed, including unit conversion, dimension unification, noise filtering, and outlier identification; Among them, unit conversion is used to unify data from different measurement systems into standard engineering units; Noise filtering uses a moving average to smooth continuously sampled data; Outlier identification uses a dynamic thresholding method: the mean (Mean) is calculated using the valid historical data of each parameter, and a multiple threshold is set. If the current collected value is greater than three times the mean or less than three times the mean, it is determined to be an outlier. For data determined to be outlier, the valid data before and after the parameter is replaced by an interpolation algorithm. The preprocessed real-time production status set Sta is defined as the processed production parameter dataset.

4. The method for dynamic proportioning control in the prebiotic production process according to claim 3, characterized in that: S2 further includes S22; S22. Based on the processed production parameter dataset, normalize each production parameter, then match it with a preset weight value for correction, and generate an indicator vector. The raw material component ratio Red is normalized and mapped to a standard range of 0 to 1. Then, the normalized raw material component ratio Red value is multiplied by a pre-set weight value to obtain the raw material component ratio Red index vector RedVec. The weight value is set according to the production standard and is used to reflect the contribution of the raw material component ratio Red to the production status and product indicators. Simultaneously, the fermenter temperature Tem is normalized and multiplied by the corresponding weight value to obtain the fermenter temperature Tem index vector TemVec; the pH value Phv is normalized and multiplied by the corresponding weight value to obtain the pH value Phv index vector PhvVec; the dissolved oxygen content Oxy is normalized and multiplied by the corresponding weight value to obtain the dissolved oxygen content Oxy index vector OxyVec; and the microbial metabolic rate Met is normalized and multiplied by the corresponding weight value to obtain the microbial metabolic rate Met index vector MetVec. Each index vector corresponds to its respective production parameter and is used to represent the degree of contribution of that production parameter to the production status and product indicators. Subsequently, the raw material component ratio Red index vector RedVec, fermenter temperature Tem index vector TemVec, pH value Phv index vector PhvVec, dissolved oxygen content Oxy index vector OxyVec, and microbial metabolic rate Met index vector MetVec are combined in a predefined order to generate a multivariate optimization input index set Inp.

5. The method for dynamic proportioning control in the prebiotic production process according to claim 4, characterized in that: S3 includes S31; S31. Based on the multivariate optimization input index set Inp, read each index vector, including the raw material component ratio Red index vector RedVec, fermenter temperature Tem index vector TemVec, pH value Phv index vector PhvVec, dissolved oxygen content Oxy index vector OxyVec, and microbial metabolic rate Met index vector MetVec. Then, the data is input into the pre-trained dynamic optimization calculation model in a predefined order to generate the current optimal dynamic ratio strategy Opt. The dynamic optimization calculation model consists of three parts: Predictive model: Based on historical production data, it is used to simulate the impact of various parameter combinations on production status and product indicators; Objective function: used to quantify the production process objectives, transforming the deviations of production status and product indicators from preset target values ​​into optimization indicators. The objective function is defined as the weighted square of the deviations of each product indicator. Constraints: These are used to limit the reasonable range of various production parameters and process restrictions, including the upper and lower limits of the raw material component ratio Red, fermenter temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met.

6. The method for dynamic proportioning control in the prebiotic production process according to claim 5, characterized in that: S3 further includes S32; S32. Specifically, the generation of the current optimal dynamic ratio strategy Opt is as follows: the index vectors output by the dynamic optimization calculation model are compared with the input index vectors, and the percentage change corresponding to each index vector is calculated. Based on this percentage change, the corresponding production parameters in the real-time production status set Sta are adjusted to form the updated production status. The updated production status is recombined to generate the current optimal dynamic ratio strategy Opt. At the same time, the current optimal dynamic ratio strategy Opt is directly issued to the production execution unit to dynamically adjust the production process so that the product indicators are close to the preset target.

7. The method for dynamic proportioning control in the prebiotic production process according to claim 6, characterized in that: S4 includes S41; S41. The generated current optimal dynamic ratio strategy Opt is sent to the production execution unit. The production execution unit dynamically adjusts the production process according to the combination of parameters in the strategy, including adjusting the raw material component ratio Red, fermenter temperature Tem, pH value Phv and dissolved oxygen content Oxy. During execution, the production execution unit continuously monitors and adjusts the parameter values ​​in the strategy according to the preset control frequency, adjusts each production parameter step by step according to the preset control frequency, and compares the deviation between the actual value and the target value in each control cycle to reduce the corresponding deviation.

8. The method for dynamic proportioning control in the prebiotic production process according to claim 7, characterized in that: S4 also includes S42; S42. While performing adjustments, the production execution unit continuously collects production process feedback signals Fbk, which include the actual values ​​and dynamic changes of each parameter during the execution process. Based on the feedback signal Fbk, the production parameter adjustment amount for the current cycle is corrected in real time, the adjusted production parameters are updated in the real-time production state set Sta, and a closed-loop control state set Clo is generated to monitor the matching status of each parameter with the strategy and the adjustment effect.

9. The method for dynamic proportioning control in the prebiotic production process according to claim 8, characterized in that: S5 includes S51; S51. Based on the closed-loop control state set Clo, read the pre-state product detection data of the corresponding batch before the execution of the current optimal dynamic ratio strategy Opt and the post-state product detection data of the corresponding batch after the execution of the current optimal dynamic ratio strategy Opt, and calculate the actual values ​​of product quality Qul, activity Aci and yield Yie before and after execution based on the pre-state product detection data and the post-state product detection data. The product quality Qul is calculated based on the content of the target prebiotic component and the content of the impurity component; The active Aci is calculated based on the functional activity test results corresponding to the target prebiotic; The yield (Yie) is calculated based on the actual output of the target prebiotic and the corresponding input of raw materials. Subsequently, the actual values ​​of product quality Qul, activity Aci, and yield Yie after execution were compared with the actual values ​​of product quality Qul, activity Aci, and yield Yie before execution to obtain the corresponding quality change rate, activity change rate, and yield change rate. The rate of change in quality, the rate of change in activity, and the rate of change in yield are compared with preset fluctuation ranges, respectively. When at least two of the quality change rate, activity change rate and yield change rate fall within the positive fluctuation range for a preset number of consecutive times, the current optimal dynamic ratio strategy Opt is determined to be effective, and the current optimal dynamic ratio strategy Opt continues to be executed. When at least two of the quality change rate, activity change rate and yield change rate fall within the negative fluctuation range for a preset number of consecutive times, the current optimal dynamic ratio strategy Opt is determined to be invalid. Based on the percentage of negative fluctuation, the corresponding parameter adjustment amount in the current optimal dynamic ratio strategy Opt is corrected, and the optimized dynamic ratio strategy Adj is output. When the rate of change in quality, the rate of change in activity, and the rate of change in yield are within a preset fluctuation range and have not reached a preset number of consecutive times, the current monitoring status is maintained and the detection data for the next cycle is collected.

10. A dynamic proportioning control system for prebiotic production process, applied to the dynamic proportioning control method for prebiotic production process described in any one of claims 1 to 9, characterized in that: It includes a production parameter acquisition module, a parameter data processing module, a parameter dynamic ratio generation module, a ratio distribution module, and a ratio verification and decision-making module; The production parameter acquisition module collects production parameters during the prebiotic production process, including the proportion of raw material components Red, fermentation tank temperature Tem, pH value Phv, dissolved oxygen content Oxy, and microbial metabolic rate Met, and forms a real-time production status set Sta. The parameter data processing module preprocesses and normalizes each parameter based on the real-time production status set Sta, generating a multivariate optimization input index set Inp, which represents the state relationship between each parameter and the production status and product index, and is used to calculate the optimal dynamic ratio strategy Opt. The dynamic ratio generation module calculates the current optimal dynamic ratio strategy Opt based on the multivariate optimization input index set Inp, so that the real-time production state set Sta reaches the preset target, thereby controlling the product quality Qul, activity Aci and yield Yie within the expected range. The ratio distribution module distributes the optimal dynamic ratio strategy Opt to the production execution unit, dynamically adjusting the raw material component ratio Red, fermentation tank temperature Tem, pH value Phv, and dissolved oxygen content Oxy. Based on the feedback signal Fbk generated during the execution process, the ratio control is adjusted in real time to form a closed-loop control state set Clo, which is used to monitor and dynamically adjust the ratio execution effect. The ratio verification decision module calculates the actual values ​​of product quality Qul, activity Aci, and yield Yie online based on the closed-loop control state set Clo. It then generates an optimized dynamic ratio strategy Adj by adjusting the optimal dynamic ratio strategy Opt and the control strategy correction algorithm. This optimized dynamic ratio strategy Adj is used to adjust the production process parameters so that product quality Qul, activity Aci, and yield Yie are close to the preset targets.

Images