Sheep milk powder production process parameter self-optimization regulation method and system

Abstract

The present application relates to the optimization control technology of goat milk powder production process parameters, especially a goat milk powder production process parameter self-optimization control method and system, the method comprises the following steps: recording historical production data, the historical production data includes the raw milk characteristic data of historical batch, the process parameter data covering the key process of goat milk powder production adopted for the raw milk characteristic data of historical batch, and the retention level data of specific heat-sensitive functional ingredients in the final product produced by adopting the process parameter data, the key process of goat milk powder production at least includes sterilization process, concentration process and drying process; through recording historical production data, the current raw material characteristics are obtained, the best historical batch is determined based on similarity comparison and target screening, which effectively solves the problem that the prior art is difficult to optimize and adjust process parameters according to raw material fluctuation and process influence to improve the retention level of specific heat-sensitive functional ingredients.

Description

TECHNICAL FIELD

[0001] The present application relates to the optimization and control technology of goat milk powder production process parameters, and in particular to a goat milk powder production process parameter self-optimization control method and system. BACKGROUND

[0002] In the production and manufacturing process of goat milk powder, the traditional process is based on fixed parameters (such as pasteurization, concentration, spray drying) to ensure that the basic nutrition, microbial safety, solubility, etc. meet the standards. The operator mainly monitors the consistency of the parameters. With the improvement of market demand for goat milk powder quality, consumers have begun to pay attention to the content and activity retention level of specific heat-sensitive functional ingredients (such as certain active proteins, probiotics, etc.) in the product. These high-value functional ingredients are extremely sensitive to heat, shear force, osmotic pressure and other conditions during the processing process, and are easily inactivated or denatured in key processes such as sterilization, concentration and drying. Even if the conventional parameters are strictly controlled, the retention rate is still unstable and has large batch differences.

[0003] Production technicians try to improve the retention of functional ingredients by adjusting existing process parameters, such as fine-tuning the sterilization temperature or time. But this often falls into a dilemma: reducing the intensity of heat treatment may affect microbial safety, and the high-temperature environment of subsequent processes (such as concentration, drying) may still cause further loss. At the same time, in order to protect sensitive ingredients, a substantial adjustment of parameters (such as reducing the drying temperature) may result in excessive product moisture, poor flowability, decreased solubility, impaired storage stability, and even affect production efficiency.

[0004] More complex is that each process of goat milk powder production is not isolated, and parameter adjustment will have a chain reaction. For example, changes in sterilization parameters may affect the physical properties of the milk liquid, thereby affecting the efficiency and effectiveness of the concentration and drying processes, and ultimately affecting the particle morphology and physical properties of the milk powder. Operators often have difficulty comprehensively evaluating the overall impact of multi-point parameter adjustment on the overall process and various product attributes, and it is difficult to find an optimal parameter combination that can simultaneously consider the retention of specific functional ingredients, product basic quality, and production economy.

[0005] Enterprises also try to increase the detection of specific functional ingredients in intermediate products and final products, but such detection is usually time-consuming and costly, making it difficult to achieve high-frequency online monitoring. The delayed detection results are difficult to effectively guide the tracing and adjustment of specific process parameters. Although a large amount of process data and product detection data has been accumulated during the production process, how to systematically use these historical data, understand the complex relationship between process parameters and the retention rate of specific functional ingredients, and form a scheme that covers multiple processes and can actively optimize and adjust parameters according to fluctuations in raw materials to significantly improve the retention rate and support continuous improvement is a key technical problem that needs to be solved.

[0006] In view of the above problems, the prior art needs to be improved. SUMMARY

[0007] The present application aims to solve the problems in the prior art and proposes a goat milk powder production process parameter self-optimization control method and system.

[0008] In a first aspect, the present application provides a goat milk powder production process parameter self-optimization control method, which comprises the following steps:

[0009] Record historical production data, which includes raw milk characteristic data of historical batches, process parameter data covering goat milk powder production key processes adopted for the raw milk characteristic data of the historical batches, and retention level data of specific heat-sensitive functional ingredients in final products produced by using the process parameter data, the goat milk powder production key processes at least including sterilization process, concentration process and drying process;

[0010] Obtain the characteristic data of the current batch of raw milk to be produced;

[0011] Based on the obtained characteristic data of the current batch of raw milk to be produced, compare the characteristic data of the current batch of raw milk to be produced with the raw milk characteristic data of each historical batch in the recorded historical production data for similarity to determine one or more historical batches matching the characteristic of the current batch of raw milk to be produced;

[0012] Based on the determined one or more historical batches, and according to the preset retention level target of specific heat-sensitive functional ingredients and other production constraints, select at least one target historical batch from the one or more historical batches, the retention level data of specific heat-sensitive functional ingredients in the final product of the target historical batch meeting the preset retention level target;

[0013] Extract the process parameter data covering the goat milk powder production key processes corresponding to the selected at least one target historical batch as the recommended process parameter for the current batch of raw milk to be produced.

[0014] In a second aspect, a goat milk powder production process parameter self-optimization control system is provided, which comprises:

[0015] A data recording module is configured to record historical production data, which includes raw milk characteristic data of historical batches, process parameter data covering goat milk powder production key processes adopted for the raw milk characteristic data of the historical batches, and retention level data of specific heat-sensitive functional ingredients in final products produced by using the process parameter data, the goat milk powder production key processes at least including sterilization process, concentration process and drying process;

[0016] A current data acquisition module is configured to acquire characteristic data of the current batch of raw milk to be produced;

[0017] A similarity comparison module is configured to compare the characteristic data of the current batch of raw milk to be produced with the characteristic data of each historical batch of raw milk recorded in the historical production data based on the acquired characteristic data of the current batch of raw milk to be produced, so as to determine one or more historical batches that match the characteristic data of the current batch of raw milk to be produced;

[0018] A target batch screening module is configured to screen at least one target historical batch from the one or more historical batches based on the determined one or more historical batches and according to a preset target retention level of a specific heat-sensitive functional ingredient and other production constraints, wherein the retention level data of the specific heat-sensitive functional ingredient in the final product of the target historical batch meets the preset target retention level.

[0019] A parameter extraction module is configured to extract process parameter data of the screened at least one target historical batch covering key processes of the production of sheep milk powder as recommended process parameters for the current batch of raw milk to be produced.

[0020] Compared with the prior art, the present application has the following beneficial effects:

[0021] By recording the historical production data, acquiring the current raw material characteristics, determining the best historical batch based on similarity comparison and target screening, and extracting the process parameters thereof as recommendations, the problem that the prior art is difficult to optimize and adjust the process parameters according to the fluctuations of raw materials and the influence between processes to improve the retention level of specific heat-sensitive functional ingredients is effectively solved, and the present application has the advantages of being able to automatically recommend optimized process parameters based on historical data according to the current raw material characteristics, thereby improving the retention level of specific heat-sensitive functional ingredients, coping with fluctuations in raw materials, and realizing self-optimization and regulation of the production process. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 The method flowchart of the present application.

[0023] Figure 2 The system structure schematic diagram of the present application.

[0024] In the figure: 201, data recording module; 202, current data acquisition module; 203, similarity comparison module; 204, target batch screening module; 205, parameter extraction module. DETAILED DESCRIPTION

[0025] Embodiments of the present application are described in detail below with reference to several drawings. The embodiments described below are exemplary and are not intended to be limiting of the present application, unless otherwise explicitly indicated herein.

[0026] The terms "first", "second", etc. are used only for the purpose of description and are not to be construed as indicating or implying relative importance or an indicated number of features. Thus, features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "a plurality of" is two or more, unless otherwise explicitly specified.

[0027] To this end, the present application provides a method for self-optimizing control of process parameters for producing goat milk powder, as shown in the accompanying drawings, which is applied to a goat milk powder production process targeting maximum retention of specific heat-sensitive functional ingredients. The method comprises the following steps: Figure 1

[0028] S101, record historical production data, including raw milk characteristic data of historical batches, process parameter data covering key goat milk powder production processes adopted for the raw milk characteristic data of historical batches, and retention level data of specific heat-sensitive functional ingredients in final products produced by adopting the process parameter data, the key goat milk powder production processes at least including sterilization process, concentration process and drying process;

[0029] S102, obtain characteristic data of the current batch of raw milk to be produced;

[0030] S103, based on the obtained characteristic data of the current batch of raw milk to be produced, compare the characteristic data of the current batch of raw milk to be produced with the raw milk characteristic data of each historical batch in the recorded historical production data for similarity, to determine one or more historical batches matching the characteristic data of the current batch of raw milk to be produced;

[0031] S104, based on the determined one or more historical batches, and according to the pre-set retention level target of specific heat-sensitive functional ingredients and other production constraints, screen at least one target historical batch from the one or more historical batches, the retention level data of specific heat-sensitive functional ingredients in the final product of the target historical batch meeting the pre-set retention level target;

[0032] S105, extract the process parameter data covering the key goat milk powder production processes corresponding to the at least one target historical batch screened out, as the recommended process parameters for the current batch of raw milk to be produced.

[0033] ​The historical production data refers to a set of relevant records of past production batches, which specifically includes historical batch raw milk property data, process parameter data for covering key processes of goat milk powder production based on historical batch raw milk property data, and retention level data of specific heat-sensitive functional ingredients in the final product produced by using the process parameter data, and the purpose is to provide data basis and learning samples for subsequent process parameter recommendation; the raw milk property data refers to data used to describe the physical, chemical or biological properties of raw milk, such as ingredient content, microbial indicators, physicochemical indicators, etc., and the purpose is to characterize the inherent properties of different batches of raw milk; the process parameter data refers to the numerical values of various control parameters set or actually run in the key processes of goat milk powder production, such as temperature, pressure, flow, time, etc., and the purpose is to record the processing conditions affecting the product formation process; the retention level data of specific heat-sensitive functional ingredients refers to the content or activity of specific heat-sensitive functional ingredients in the final product relative to the initial content in the raw milk, and the purpose is to quantify the influence of the process on the target functional ingredients; the key process of goat milk powder production refers to the core processing link that significantly affects the retention of specific heat-sensitive functional ingredients, at least including sterilization process, concentration process and drying process, and the purpose is to define the production range concerned and regulated by the method; the similarity comparison refers to calculating or evaluating the similarity between the current batch raw milk property data and the historical batch raw milk property data, and the purpose is to identify the historical production situation similar to the current raw milk property; the target historical batch refers to the batch selected from the historical batches with similar raw milk properties, whose final product specific heat-sensitive functional ingredient retention level meets the preset target and meets other production constraint conditions, and the purpose is to determine the reference sample that can achieve the expected product effect; the recommended process parameter refers to the process parameter setting set obtained by extracting or synthesizing the target historical batch, which is suitable for the current batch of raw milk, and the purpose is to provide optimized process guidance for the current production.

[0034] The scheme of the present application builds an experience knowledge base by recording and accumulating historical production data containing raw material characteristics, process parameters and product results. When a new batch of raw milk to be produced is available, its characteristic data is first obtained, and then compared with the characteristics of the historical data to find one or more historical batches similar to the current raw material characteristics. This means that the system can identify similar raw material situations encountered in the past production. Then, among these similar historical batches, according to the preset specific heat-sensitive functional ingredient retention level target and other production constraints (such as microbial safety, basic physicochemical indicators, production efficiency, etc.), those batches that have successfully achieved or approached the expected retention level in the actual production in the past are screened out, and these screened batches are the target historical batches. Finally, the process parameter data used in the key processes such as sterilization, concentration and drying of these target historical batches are extracted and used as recommended process parameters for the current batch of raw milk. The whole process forms a closed-loop optimization and control mechanism based on historical experience, which can dynamically recommend verified process parameter combinations for different batches of raw material characteristics, thereby effectively dealing with raw material fluctuations and complex effects between processes, and improving the retention rate of specific heat-sensitive functional ingredients.

[0035] As an embodiment of the present application, when there is no single historical batch that meets both the preset specific heat-sensitive functional ingredient retention level target and all other production constraints in one or more historical batches, the step of screening at least one target historical batch from the one or more historical batches based on the determined one or more historical batches and according to the preset specific heat-sensitive functional ingredient retention level target and other production constraints comprises:

[0036] Setting the retention level target of the specific heat-sensitive functional ingredient as the priority target, and setting the acceptance range for the other production constraints;

[0037] Based on the determined one or more historical batches and according to the set specific heat-sensitive functional ingredient retention level target corresponding to the priority target and the acceptance range of the other production constraints, among the one or more historical batches, screening out the historical batches whose specific heat-sensitive functional ingredient retention level data meets the preset retention level target and each of the other production constraints is within the respective acceptance range as the target historical batches;

[0038] If multiple target historical batches that meet the conditions are obtained based on the screening, a final target historical batch is determined from the multiple target historical batches according to a preset emphasis strategy.

[0039] The setting of the retention level of the specific heat-sensitive functional ingredient as a priority target refers to giving priority to whether the retention level of the ingredient meets the preset requirement when screening historical batches, and the purpose is to ensure the high added value characteristics of the final product; the setting of an acceptable range for other production constraints refers to setting an acceptable fluctuation interval instead of a strict fixed value for other production-related limiting conditions except the retention level of the specific heat-sensitive functional ingredient, and the purpose is to increase the number of historical batches that meet the conditions and improve data utilization; the preset focus strategy refers to a decision rule or priority order used to determine the final recommended process parameters when there are multiple historical batches that meet the screening conditions, which can be set according to actual production needs, and the purpose is to select a scheme that best meets the current production target from multiple alternative schemes.

[0040] The scheme of the present application adds a flexible screening mechanism for complex production constraints under the historical data similarity matching and target screening framework provided by the basic scheme. When one or more historical batches matching the current raw material characteristics are determined through similarity comparison, if it is found that none of these batches can simultaneously meet the preset specific heat-sensitive functional ingredient retention level target and all other production constraint conditions, the scheme no longer simply discards these data, but ensures that the selected batch meets the core indicator by taking the retention level of the specific heat-sensitive functional ingredient as the primary consideration. At the same time, it is recognized that other production constraint conditions are difficult to completely fix in practice, so a reasonable acceptable range is set for these conditions. Then, a secondary screening is performed among the matched historical batches to find batches whose specific heat-sensitive functional ingredient retention level meets the preset target and other production constraint conditions are within their respective acceptable ranges. This way expands the set of potential target batches and avoids excluding valuable historical data due to minor deviations in individual constraint conditions. If multiple batches that meet the conditions are still found after flexible screening, further decision-making is needed. At this time, according to the preset focus strategy, such as giving priority to cost, efficiency or other quality indicators, a final target historical batch is determined from the multiple alternative batches. The process parameters corresponding to this final target historical batch will be extracted as the recommended process parameters for the current batch. This method provides a practical and effective data utilization strategy when perfect matching historical data cannot be found, and can balance other production requirements while ensuring the retention of core functional ingredients, thereby solving the problem of difficult effective utilization of historical data under complex production constraints. This processing method increases the ability to deal with actual production complexity under the framework of historical data similarity matching and target screening provided by the basic scheme, so that the entire self-optimizing control method can still provide valuable process parameter reference when facing an undesirable historical data distribution, improving the robustness and practicality of the method.

[0041] As an embodiment of the present application, the step of determining a final target historical batch from a plurality of target historical batches according to a preset emphasis strategy comprises:

[0042] assigning priorities to a plurality of emphasis dimensions included in the preset emphasis strategy to obtain a plurality of assigned-priority emphasis dimensions;

[0043] for the plurality of target historical batches, and according to the plurality of assigned-priority emphasis dimensions, starting from the highest-priority emphasis dimension, performing the following operations step by step until a final target historical batch is determined:

[0044] obtaining performance data of each target historical batch under the currently processed priority dimension;

[0045] based on the obtained performance data, evaluating the performance of each target historical batch under the currently processed priority dimension to identify one or more candidate historical batches that meet the predetermined criteria of the dimension;

[0046] if the identified candidate historical batches are one, then the one candidate historical batch is the final target historical batch;

[0047] if the identified candidate historical batches are multiple, then the multiple candidate historical batches are taken as the objects of processing in the next priority dimension;

[0048] if after all priority dimensions are processed, the objects obtained in the last sub-step are still multiple candidate historical batches, then from these candidate historical batches, the historical batch that performs best in the preset dimension related to the specific heat-sensitive functional ingredient retention level is selected as the final target historical batch.

[0049] The preset emphasis strategy refers to a set of pre-set rules for guiding how to select an optimal batch from multiple historical batches that meet the basic conditions when multiple historical batches meet the basic conditions. Specifically, it can include a series of aspects to be considered and their importance ranking. The purpose is to enable the finally selected batch to better meet the specific production preferences or optimization targets. The multiple emphasis dimensions refer to the specific aspects of consideration that constitute the preset emphasis strategy. For example, they can be certain specific production constraint conditions (such as energy consumption, material loss), certain non-constrained performance indicators (such as production efficiency, equipment wear degree), or more refined indicators related to the retention level of specific heat-sensitive functional ingredients (such as the denaturation rate of a certain specific protein, the survival rate of a certain probiotic). The purpose is to provide a multi-angle evaluation perspective. The priority assignment refers to assigning importance levels to these emphasis dimensions and determining their consideration order in the screening process. The purpose is to ensure that the most important factors are given priority consideration in the selection process. The multiple emphasis dimensions assigned with priorities refer to the set of emphasis dimensions with clear importance order after priority assignment. The purpose is to provide a basis for subsequent step-by-step screening. The performance data refers to the specific numerical values or qualitative descriptions of each target historical batch on the specific emphasis dimension. The purpose is to provide basic information for evaluation. The predetermined standard refers to the threshold or condition for judging whether the batch meets the requirements on a certain emphasis dimension. For example, it can be a certain upper or lower limit, or a certain range. The purpose is to define the qualified range on that dimension. The candidate historical batch refers to the batch set that remains after the screening of the current priority dimension and may become the final target. The purpose is to gradually narrow down the selection range.

[0050] The scheme of the present application assigns priority to multiple focus dimensions contained in the preset focus strategy, thereby clarifying the factors that need to be prioritized when selecting the optimal batch. Subsequently, for multiple target historical batches, the assigned priority of the multiple focus dimensions is used to perform a screening operation step by step, starting from the highest priority focus dimension. In each level of screening, the performance data of each target historical batch in that dimension is first obtained, and then the performance of each batch in that dimension is evaluated based on these data, and one or more candidate historical batches that meet the predetermined criteria of that dimension are identified. If there is only one candidate historical batch, then that batch is determined as the final target historical batch, and the screening process ends. If there are multiple candidate historical batches, then these batches are taken as the object of processing in the next level of priority dimension, and the screening continues. This step-by-step screening method ensures that the performance of low-priority dimensions is considered after meeting the requirements of high-priority dimensions, thereby systematically selecting a batch with better overall performance from multiple qualified batches. Even if there are still multiple candidate historical batches after all priority dimensions have been processed, the scheme provides a fallback mechanism, i.e., selecting the historical batch with the best performance in the preset dimension related to the specific heat-sensitive functional ingredient retention level as the final target, further ensuring that the finally selected batch has an advantage in the core target. Based on the above scheme that screens multiple batches that meet the specific heat-sensitive functional ingredient retention level target and part of other production constraint conditions, the present application further provides a refined method for optimal selection based on preset preferences, so that the entire process parameter recommendation process can better balance multiple demands, improving the relevance and reliability of the recommended parameters.

[0051] As an embodiment of the present application, the step of similarity comparison between the characteristic data of the current batch of raw milk to be produced and the characteristic data of each historical batch in the recorded historical production data includes:

[0052] According to the current value of at least one preset reference characteristic index in the characteristic data of the current batch of raw milk to be produced, the current batch of raw milk is classified into a predefined raw material state interval;

[0053] For the classified raw material state interval, from the pre-set multiple sets of weight distribution strategies corresponding to different raw material state intervals, a set of weight distribution strategies matching the current classified raw material state interval is selected, and the weight distribution strategy specifies the weight values of each characteristic index in the raw milk characteristic data in the similarity comparison, except for the reference characteristic index;

[0054] The selected weight distribution strategy is applied to the characteristic data of the current batch of raw milk to be produced and the characteristic data of each historical batch of raw milk in the recorded historical production data to perform weighted difference calculation on the corresponding characteristic indicators, and the weighted differences are integrated to complete the similarity comparison between the characteristic data of the current batch of raw milk to be produced and the characteristic data of each historical batch of raw milk in the recorded historical production data, and the similarity between the current batch of raw milk to be produced and each historical batch of raw milk is obtained.

[0055] The preset reference characteristic indicator refers to one or more indicators that are previously determined to have a significant or basic impact on the overall state of raw milk or subsequent processing characteristics among the many characteristic indicators of raw milk, which can be determined based on expert experience, historical data analysis or pre-set rules; the pre-defined raw material state interval refers to dividing the possible state of raw milk into several discrete classification intervals with specific meanings according to the numerical range of the preset reference characteristic indicator, which can be pre-defined by threshold setting, clustering analysis or manual experience based on the reference characteristic indicator; the pre-configured multiple sets of weight distribution strategies corresponding to different raw material state intervals refer to that for each pre-defined raw material state interval, a specific set of rules for the importance weight of each characteristic indicator in similarity comparison is previously configured, which can be pre-configured and corresponded based on experimental data, expert knowledge or machine learning model training results of the influence degree of each characteristic indicator on the retention level of specific heat-sensitive functional ingredients under different raw material states; the weight distribution strategy specifies the weight value of each characteristic indicator other than the reference characteristic indicator in the raw milk characteristic data in similarity comparison refers to that in similarity comparison, each characteristic indicator participating in comparison is assigned a numerical value other than the reference characteristic indicator used for classifying the state of raw milk, which reflects the relative importance of the characteristic indicator in evaluating the similarity of raw milk under the specific raw material state interval; the weighted difference calculation refers to calculating the numerical difference of each characteristic indicator other than the reference characteristic indicator when comparing the characteristic data of the current batch of raw milk to be produced and the historical batch of raw milk, and multiplying the difference by the weight value of the characteristic indicator corresponding to the currently selected weight distribution strategy, which can be calculated by multiplying the absolute difference by the weight, multiplying the squared difference by the weight, etc.; the integrated weighted difference refers to aggregating the weighted difference values of all characteristic indicators participating in the weighted difference calculation to obtain a single numerical value, which reflects the overall similarity or difference between the current batch of raw milk to be produced and a certain historical batch of raw milk after considering the importance difference of each characteristic indicator, which can be integrated by summing, weighted averaging or distance measurement based on weighted difference.

[0056] The scheme of the present application can classify the current raw milk to be produced into a predefined raw material state interval according to the current value of at least one preset reference characteristic index of the current raw milk to be produced, so as to divide the raw milk with similar characteristics into the same interval, and lay the foundation for subsequent adoption of different weight allocation strategies, because the influence degree of the characteristic indexes of raw milk of different states on the final product may be different. On this basis, for the classified raw material state interval, one set of weight allocation strategies corresponding to different raw material state intervals is selected from the preset multiple sets of weight allocation strategies, and the weight allocation strategy specifies the weight value of each characteristic index in the raw milk characteristic data in the similarity comparison except the reference characteristic index. By configuring different weight allocation strategies for different raw material state intervals, the importance of each characteristic index in the similarity comparison can be more flexibly adjusted, so as to more accurately reflect the influence degree of different characteristic indexes on the final product, and avoid the deviation caused by adopting a unified weight. Subsequently, the selected weight allocation strategy is applied to perform weighted difference calculation on the corresponding characteristic indexes in the characteristic data of the current raw milk to be produced and the characteristic data of each historical batch of raw milk in the recorded historical production data, and the weighted differences are integrated to complete the similarity comparison between the characteristic data of the current raw milk to be produced and the characteristic data of each historical batch of raw milk in the recorded historical production data, so as to obtain the similarity between the current raw milk to be produced and each historical batch of raw milk. Through the weighted difference calculation, the influence of important characteristic indexes can be highlighted, and the influence of unimportant characteristic indexes can be weakened, so as to more accurately evaluate the similarity between the current raw milk to be produced and the historical batch of raw milk. This way of dynamically adjusting the comparison emphasis according to the characteristics of the raw material itself makes the similarity comparison result more reflect the potential behavior of the raw milk in actual production, so as to provide more accurate input for subsequent process parameter recommendation, and significantly improves the achievement effect of the recommended parameters on the retention target of specific heat-sensitive functional ingredients.

[0057] As an embodiment of the present application, when the selected target historical batch is multiple, the step of extracting the process parameter data covering the key process of sheep milk powder production corresponding to the selected at least one target historical batch as the recommended process parameter for the current raw milk to be produced includes:

[0058] Collecting parameter values of various process parameters for each key process of sheep milk powder production from multiple target historical batches;

[0059] Obtaining the retention level data of specific heat-sensitive functional ingredients in the final products of multiple target historical batches respectively;

[0060] identify process parameters that have a preset influence degree on the retention level of the specific heat-sensitive functional ingredient among the process parameters based on the parameter values of the collected process parameters;

[0061] determine a single parameter value of the process parameter according to the parameter values of the process parameter corresponding to the multiple target historical batches;

[0062] determine a single parameter value of the process parameter according to the parameter values of the process parameter corresponding to the multiple target historical batches;

[0063] combine the single parameter values determined for each process parameter to form a set of single recommended process parameters covering the key processes of goat milk powder production as the recommended process parameters for the raw milk of the current batch to be produced.

[0064] Among them, each process parameter of each key process of goat milk powder production refers to the specific setting value or measured value used to control the running state of the equipment and the processing conditions of the material in each operation link (such as sterilization, concentration, drying) that has an important influence on product quality, especially the retention level of specific heat-sensitive functional ingredients, such as temperature, time, pressure, flow, concentration, etc. The retention level data of specific heat-sensitive functional ingredients refers to the quantitative data of the content, activity or relative initial content of one or more heat-sensitive ingredients with specific biological activity (such as active protein, probiotics) in the final product of goat milk powder. Identifying process parameters that have a preset influence degree on the retention level of specific heat-sensitive functional ingredients among the process parameters refers to determining which process parameters have a significant, preset threshold correlation or causal relationship with the retention level of specific heat-sensitive functional ingredients in the final product by analyzing historical data or based on prior knowledge, which can be achieved by statistical analysis methods, machine learning models or expert system rules. The preset influence degree refers to a quantitative standard or threshold value preset when identifying the influence of process parameters on the retention level of specific heat-sensitive functional ingredients, which is used to distinguish between significant (with a preset influence degree) and insignificant parameters. This standard can be set based on statistical significance, influence size or business importance. Determining a single parameter value of the process parameter refers to deriving a unique parameter setting value for the current production batch from the parameter values of the multiple target historical batches for a specific process parameter through a certain calculation or decision method. For parameters with a preset influence degree, the determination method considers their relationship with the retention level; for parameters without a preset influence degree, the determination method may focus more on the central tendency of the parameter values.

[0065] The scheme of the present application provides a comprehensive data basis for subsequent analysis by collecting process parameter values and corresponding specific heat-sensitive functional ingredient retention level data from multiple target historical batches. Based on these data, the scheme further identifies key process parameters that have a preset degree of influence on the specific heat-sensitive functional ingredient retention level. It is precisely because of the distinction between key parameters and non-key parameters that the scheme can adopt different processing strategies for different types of parameters: for key parameters with significant influence, comprehensive analysis is performed based on the retention level data and parameter values of historical batches to determine a single parameter value that is more conducive to high retention levels; while for non-key parameters with insignificant influence, a relatively simple method can be used to determine a single parameter value. This divide-and-conquer strategy focuses on the most important parameters, which can more effectively utilize the complex relationship between parameters and effects contained in historical data, especially for those parameters that have the greatest impact on the target effect (specific heat-sensitive functional ingredient retention), thereby generating a set of recommended parameters that are more likely to achieve high retention levels. This method, combined with the selection of multiple batches that meet the preset retention level target in the prior scheme, can make full use of the information of these excellent batches, rather than wasting most of the data, overcoming the randomness and suboptimality that may be caused by simple selection, and improving the reliability and optimization of the recommended parameters. Finally, the single parameter values determined for each process parameter are combined to form a set of single recommended process parameters covering the key processes of goat milk powder production, which can be directly applied to the raw milk treatment of the current batch to be produced, thereby more targetedly optimizing the production process and improving the retention level of specific heat-sensitive functional ingredients.

[0066] As an embodiment of the present application, for the identified process parameters with a preset degree of influence, the step of determining a single parameter value of the process parameter based on the obtained retention level data of multiple target historical batches and the parameter values of the process parameter corresponding to these historical batches includes:

[0067] From the multiple target historical batches that meet the standard, one or more target historical batches that meet the standard are selected according to the retention level data of the specific heat-sensitive functional ingredient in the final product of each target historical batch, wherein the specific heat-sensitive functional ingredient retention level data of the one or more target historical batches that meet the standard meets the preset standard.

[0068] If the selected target historical batch that meets the standard is one, the parameter value of the process parameter with a preset degree of influence corresponding to the one target historical batch that meets the standard is used as a single parameter value of the process parameter with a preset degree of influence.

[0069] If the screened target historical batches meeting the standard are multiple, a single parameter value of the process parameter with the preset influence degree is determined based on a set of parameter values of the process parameter with the preset influence degree corresponding to the multiple target historical batches meeting the standard, through a preset parameter processing rule.

[0070] The preset standard refers to a threshold or range for measuring whether the retention level of the specific heat-sensitive functional ingredient meets the expected requirement, which can be represented by a specific percentage value, a numerical interval, or a rating, and aims to identify those historical batches that perform well in terms of specific functional ingredient retention. The preset parameter processing rule refers to a calculation method or logic for synthesizing multiple parameter values to determine a single representative value, which can be achieved by using methods such as average calculation, weighted average calculation, median selection, or prediction based on statistical models, and aims to extract a representative parameter value that is helpful for optimizing the retention level from the data of multiple excellent historical batches.

[0071] The scheme of the present application determines a single parameter value of the process parameter with the preset influence degree according to the retention level data and corresponding parameter values of multiple target historical batches. First, from the multiple target historical batches meeting the standard, one or more target historical batches meeting the standard are selected according to the retention level data of the specific heat-sensitive functional ingredient in the final product of each batch, which meets the preset standard. This selection process focuses on those batches that perform well in terms of target ingredient retention, which is equivalent to a quality filtration of the original data, excluding the interference of batch data with low retention level on parameter determination. Then, according to the selection result, branch processing is performed: if the screened target historical batches meeting the standard are only one, the parameter value of the process parameter with the preset influence degree corresponding to this batch is directly used as the single parameter value, which ensures that when there is a clear optimal historical batch, its successful experience can be directly adopted; if the screened target historical batches meeting the standard are multiple, a single parameter value is determined based on the set of parameter values of the process parameter with the preset influence degree corresponding to these excellent batches, through a preset parameter processing rule. This processing method can integrate the parameter information of multiple excellent batches, balance the data of different batches through certain algorithms or logic, and obtain a more robust and representative recommended parameter value, avoiding the accidental bias that may be caused by relying on single batch data. This process refines and optimizes the step of determining a single parameter value after simply collecting parameter values and identifying the influencing parameters in the basic scheme, so that the value of the key influencing parameter is more accurately anchored on the historical data that have proven to achieve high retention level in actual production, thereby improving the effectiveness of the final recommended process parameter and helping to better achieve the maximum retention goal of the specific heat-sensitive functional ingredient in actual production.

[0072] As an embodiment of the present application, the step of determining a single parameter value of the process parameter with preset influence degree through preset parameter processing rules comprises:

[0073] According to the data of the retention level of the specific heat-sensitive functional ingredient in the final product of each of the target historical batches meeting the standard, a weighted average of the parameter values of the process parameter with preset influence degree is performed to obtain a single parameter value of the process parameter with preset influence degree.

[0074] Wherein, the plurality of target historical batches meeting the standard refers to a set of historical production batches whose data of the retention level of the specific heat-sensitive functional ingredient in the final product meet the preset standard after preliminary screening and further screening. The data of the retention level of the specific heat-sensitive functional ingredient in the final product refers to a quantitative value reflecting the retention degree of the specific heat-sensitive functional ingredient obtained by detecting and analyzing the final goat milk powder product completed in each batch of production, which can be expressed in the form of percentage, content per unit mass, or relative activity, etc. The process parameter with preset influence degree refers to a process control variable that is identified after analysis and is determined to have significant or important influence on the retention level of the specific heat-sensitive functional ingredient in the final product in the key process of goat milk powder production. The parameter value refers to the specific numerical value setting or measurement value actually adopted or recorded for the process parameter with preset influence degree in the production process of a specific historical batch. The weighted average refers to a statistical average method which calculates the average value according to the contribution of each data point to the final result, which can be realized by different weight calculation and application methods. The single parameter value refers to a unique recommended setting value for representing the process parameter with preset influence degree obtained by weighted average calculation.

[0075] The scheme of the present application obtains a single parameter value of the process parameter with preset influence degree by weighted averaging the parameter values of the process parameter with preset influence degree according to the retention level data of the specific heat-sensitive functional ingredient in the final product of each target historical batch meeting the standard. Specifically, after identifying the process parameters with preset influence degree on the retention level of the specific heat-sensitive functional ingredient, and after screening a plurality of target historical batches with retention level meeting the preset standard, the scheme uses the corresponding retention level data of these batches as weight information. The higher the retention level, the more beneficial the process parameter value corresponding to the batch is to achieving high retention level, so it is given a higher weight in weighted averaging. In this way, the single parameter value finally calculated will be more inclined to those parameter settings that have proven to bring higher retention level in historical practice. This method, combined with the aforementioned data recording, preliminary matching screening, key parameter identification, and high retention level batch screening, forms a complete self-optimizing control process. It provides a method of fine and comprehensive multiple high retention level batch parameter values on the basis of screening a plurality of high retention level batches. Instead of simply selecting or averaging, it is weighted according to the actual retention effect, so that the better experiences in the historical data are more fully utilized. This makes the final recommended process parameters more accurately reflect the correlation between high retention level, so as to better guide parameter setting in actual production and improve the retention level of the specific heat-sensitive functional ingredient. This weighted averaging based on effect feedback enables the system to learn from historical experience and optimize parameter setting, cope with raw material fluctuations and process effects, and achieve continuous improvement for specific targets.

[0076] As an embodiment of the present application, the step of weighted averaging the parameter values of the process parameter with preset influence degree includes:

[0077] According to the retention level data of the specific heat-sensitive functional ingredient in the final product of each target historical batch meeting the standard, the retention level data of the specific heat-sensitive functional ingredient is normalized;

[0078] The normalized retention level data is used as weight to weighted average the parameter values of the process parameter with preset influence degree, to obtain a single parameter value of the process parameter with preset influence degree.

[0079] The normalization processing refers to converting data of different dimensions or value ranges to a unified scale, which can be realized by using methods such as minimum-maximum normalization and Z-score normalization, and the purpose is to eliminate the dimensional influence between data and make them comparable; the weighted average refers to giving different weights to different data points when calculating the average value, and the weight reflects the importance of the data point in the calculation, and the purpose is to make important data points have greater influence on the result.

[0080] The scheme of the present application firstly normalizes the retention level data of the specific heat-sensitive functional ingredient in the final product of each target historical batch meeting the standard, so as to unify the retention level data of different historical batches to a comparable range, eliminate the influence caused by the difference in the original data dimension or value range, and make the subsequent weighted average more reasonable and effective. Then, the normalized retention level data is used as the weight to perform weighted average on the parameter values of the process parameters with a preset influence degree, and a single parameter value of the process parameter with the preset influence degree is obtained. This means that the historical batch with a higher retention level has a greater proportion in the final single parameter value. This processing method is combined with the overall framework of determining the recommended parameter from multiple target historical batches, and especially after multiple historical batches meeting the specific retention level standard are screened out, the scheme can more finely utilize the data of these batches. By normalizing the retention level data and using it as the weight, the experience of the historical batches that perform better in the retention of the specific heat-sensitive functional ingredient can be more fully utilized, so that it is more likely to obtain a process parameter value that can better retain the specific functional ingredient, and the deficiency that the simple average may ignore the difference in the retention level is overcome.

[0081] As an embodiment of the present application, the step of using the normalized retention level data as the weight to perform weighted average on the parameter values of the process parameters with a preset influence degree to obtain a single parameter value of the process parameter with the preset influence degree comprises:

[0082] calculating the product of the weight value and the parameter value of each batch to obtain a weighted parameter value;

[0083] summing all the weighted parameter values to obtain a weighted sum;

[0084] summing all the weight values to obtain a weight sum;

[0085] dividing the weighted sum by the weight sum to obtain a single parameter value.

[0086] Among them, the retention level data of specific heat-sensitive functional ingredients in the final products of multiple target historical batches that meet the standards are normalized, which can convert these data to a unified scale, thereby ensuring the comparability of data between different batches. Using the normalized retention level data as a weight means that the higher the retention level of the batch, the greater the contribution of its corresponding process parameters to the final recommended value. Calculating the product of the weight value and the parameter value of each batch to obtain a weighted parameter value is to combine the information of each batch on the retention level and process parameters. Summing all weighted parameter values ​​to obtain a weighted sum is the weighted contribution of all target historical batches. Summing all weight values ​​to obtain the weighted sum is the weighted weight of all batches. Dividing the weighted sum by the weighted sum to obtain a single parameter value is a standard method for performing a weighted average calculation, thereby determining a recommended parameter value that comprehensively considers the experience of multiple high retention level batches.

[0087] Specifically, after identifying multiple target historical batches that meet the retention target for a specific heat-sensitive functional ingredient and other production constraints, a single recommended parameter value is determined from the process parameters corresponding to these batches. To ensure that the determined parameter value better reflects the experience of batches with high retention levels, a weighted average approach is employed. First, the retention level data for the specific heat-sensitive functional ingredient in the final product of each target historical batch is normalized. This converts the retention level data from different batches to a unified scale, ensuring data comparability and providing a reliable basis for subsequent weighted calculations. The normalized retention level data is used as the weight. Higher retention levels correspond to larger normalized weights, indicating that the process parameters for that batch have greater reference value in achieving high retention levels. The weight value for each batch is multiplied by the parameter value of its corresponding process parameter with a preset impact level to obtain the weighted parameter value for that batch. The weighted parameter values ​​for all batches are then accumulated to obtain the weighted sum. The weighted sum is then calculated by summing the weighted values ​​for all batches. A single parameter value of the process parameter with a preset degree of influence is calculated by dividing the weighted sum by the weight sum.

[0088] like Figure 2 A self-optimizing and controlling system for process parameters of goat milk powder production is shown, and the system comprises:

[0089] Data recording module 201 is used to record historical production data, including raw milk property data of historical batches, process parameter data covering key production processes of goat milk powder based on the raw milk property data of historical batches, and retention level data of specific heat-sensitive functional ingredients in the final product produced using the process parameter data. The key production processes of goat milk powder include at least sterilization, concentration, and drying.

[0090] The current data acquisition module 202 is configured to acquire characteristic data of the current batch of raw milk to be produced;

[0091] The similarity comparison module 203 is configured to perform similarity comparison between the characteristic data of the current batch of raw milk to be produced and the characteristic data of each historical batch of raw milk recorded in the historical production data based on the acquired characteristic data of the current batch of raw milk to be produced, so as to determine one or more historical batches matching the characteristic data of the current batch of raw milk to be produced;

[0092] The target batch screening module 204 is configured to screen at least one target historical batch from the one or more historical batches based on the determined one or more historical batches and according to a preset target retention level of a specific heat-sensitive functional ingredient and other production constraints, wherein the retention level data of the specific heat-sensitive functional ingredient in the final product of the target historical batch satisfies the preset target retention level.

[0093] The parameter extraction module 205 is configured to extract process parameter data covering key processes of the production of sheep milk powder corresponding to the at least one screened target historical batch as recommended process parameters for the current batch of raw milk to be produced.

[0094] The basic principles, main features and advantages of the present application are shown and described above. It should be understood by those skilled in the art that the present application is not limited to the above-mentioned embodiments, and the above-mentioned embodiments and descriptions in the specification are only the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application.

Claims

1. A method for self-optimization and control of process parameters for goat milk powder production, applied to a goat milk powder production process with the goal of maximizing the retention of specific heat-sensitive functional ingredients, characterized in that: The method comprises the following steps: Recording historical production data, including raw milk property data of historical batches, process parameter data covering key steps in goat milk powder production based on the raw milk property data of the historical batches, and retention level data of specific heat-sensitive functional ingredients in the final product produced using the process parameter data; Obtain characteristic data of the current batch of raw milk to be produced; Based on the acquired characteristic data of the current batch of raw milk to be produced, performing a similarity comparison between the characteristic data of the current batch of raw milk to be produced and the characteristic data of raw milk of each historical batch in the recorded historical production data, to determine one or more historical batches whose characteristics match the characteristics of the current batch of raw milk to be produced; Based on the determined one or more historical batches and according to a preset retention level target for the specific heat-sensitive functional ingredient and other production constraints, at least one target historical batch is selected from the one or more historical batches, wherein the retention level data of the specific heat-sensitive functional ingredient in the final product of the target historical batch meets the preset retention level target; Extracting process parameter data covering key processes of goat milk powder production corresponding to at least one selected target historical batch as recommended process parameters for the current batch of raw milk to be produced; When there is no single historical batch that satisfies both the preset target retention level of the specific heat-sensitive functional ingredient and all other production constraints among the one or more historical batches, the step of selecting at least one target historical batch from the one or more historical batches based on the determined one or more historical batches and in accordance with the preset target retention level of the specific heat-sensitive functional ingredient and other production constraints includes: setting a target retention level of the specific heat-sensitive functional ingredient as a priority target and setting an acceptable range for the other production constraints; Based on the determined one or more historical batches, and according to the set retention level target of the specific heat-sensitive functional ingredient corresponding to the priority target and the set acceptable range of the other production constraints, from the one or more historical batches, historical batches whose retention level data of the specific heat-sensitive functional ingredient meets the preset retention level target and whose other production constraints are all within the respective acceptable ranges are screened as target historical batches; If multiple target historical batches meeting the conditions are obtained based on the screening, a final target historical batch is determined from the multiple target historical batches according to a preset emphasis strategy; The step of determining a final target historical batch from multiple target historical batches according to a preset emphasis strategy includes: Prioritizing the multiple emphasis dimensions included in the preset emphasis strategy to obtain multiple emphasis dimensions with assigned priorities; For multiple target historical batches, and based on the multiple priority-based dimensions, starting with the highest priority dimension, the following operations are performed step by step until a final target historical batch is determined: Obtain the performance data of each target historical batch under the current processing priority dimension; Based on the obtained performance data, evaluating the performance of each target historical batch under the priority dimension currently being processed to identify one or more candidate historical batches that meet predetermined criteria for the dimension; If only one candidate historical batch is identified, then the candidate historical batch becomes the final target historical batch; If multiple candidate historical batches are identified, these multiple candidate historical batches will be processed as objects of the next level of priority dimension; If after all priority dimensions are processed, the processing object obtained in the previous sub-step is still multiple candidate historical batches, then from these candidate historical batches, the historical batch with the best performance in the preset dimension related to the retention level of the specific thermosensitive functional ingredient is selected as the final target historical batch.

2. A method for self-optimizing and controlling process parameters for producing goat milk powder according to claim 1, characterized in that: The step of performing similarity comparison between the characteristic data of the current batch of raw milk to be produced and the characteristic data of each historical batch of raw milk in the recorded historical production data comprises: Classifying the current batch of raw milk to be produced into a predefined raw material state interval according to a current value of at least one preset benchmark characteristic indicator in the characteristic data of the current batch of raw milk to be produced; For the classified raw material state interval, a set of weight allocation strategies that matches the currently classified raw material state interval is selected from multiple sets of preset weight allocation strategies corresponding to different raw material state intervals, wherein the weight allocation strategy specifies the weight values ​​of each characteristic index other than the baseline characteristic index in the raw milk characteristic data in the similarity comparison; Applying the selected weight allocation strategy, a weighted difference calculation is performed on the characteristic data of the current batch of raw milk to be produced and the corresponding characteristic indicators in the characteristic data of each historical batch of raw milk in the recorded historical production data, and the weighted differences are integrated to complete the similarity comparison between the characteristic data of the current batch of raw milk to be produced and the characteristic data of each historical batch of raw milk in the recorded historical production data, so as to obtain the similarity between the current batch of raw milk to be produced and the raw milk of each historical batch.

3. A method for self-optimizing and controlling process parameters for producing goat milk powder according to claim 1, characterized in that: When there are multiple target historical batches screened out, the step of extracting process parameter data covering key processes of goat milk powder production corresponding to at least one of the screened out target historical batches as recommended process parameters for the current batch of raw milk to be produced includes: Gather the parameter values ​​of various process parameters for each key process of goat milk powder production from multiple target historical batches; Obtaining data on the retention levels of specific heat-sensitive functional ingredients in the final product for each of multiple target historical batches; Based on the collected parameter values ​​of the various process parameters, identifying a process parameter that has a predetermined degree of influence on the retention level of the specific heat-sensitive functional ingredient among the various process parameters; For the identified process parameter with a preset impact, determining a single parameter value of the process parameter based on the obtained retention level data of multiple target historical batches and the parameter values ​​of the process parameter corresponding to these historical batches; For a process parameter that is not identified as having a preset impact level, determining a single parameter value of the process parameter based on parameter values ​​of the process parameter corresponding to multiple target historical batches; The single parameter values ​​determined for the various process parameters are combined to form a set of single recommended process parameters covering the key processes of goat milk powder production, which are used as recommended process parameters for the current batch of raw milk to be produced.

4. A method for self-optimizing and controlling process parameters for producing goat milk powder according to claim 3, characterized in that: The step of determining a single parameter value of the process parameter identified as having a preset impact based on the obtained retention level data of multiple target historical batches and the parameter values ​​of the process parameter corresponding to these historical batches includes: From a plurality of target historical batches that meet the criteria, screening out one or more target historical batches that meet the criteria, based on the retention level data of the specific heat-sensitive functional ingredient in the final product of each target historical batch; If only one target historical batch that meets the criteria is screened out, the parameter value of the process parameter with the preset degree of influence corresponding to the target historical batch that meets the criteria is adopted as a single parameter value of the process parameter with the preset degree of influence; If there are multiple target historical batches that meet the criteria, a single parameter value of the process parameter with a preset degree of influence is determined through a preset parameter processing rule based on the set of parameter values ​​of the process parameters with a preset degree of influence corresponding to the multiple target historical batches that meet the criteria.

5. A method for self-optimizing and controlling process parameters for producing goat milk powder according to claim 4, characterized in that: The step of determining a single parameter value of a process parameter with a preset influence degree by using a preset parameter processing rule includes: Based on the retention level data of specific heat-sensitive functional ingredients in the final products of multiple target historical batches that meet the standards, the parameter values ​​of the process parameters with a preset impact degree are weighted averaged to obtain a single parameter value of the process parameter with the preset impact degree.

6. A method for self-optimizing and controlling process parameters for producing goat milk powder according to claim 5, characterized in that: The step of performing weighted averaging on the parameter values ​​of the process parameters with a preset influence degree comprises: normalizing the retention level data of the specific heat-sensitive functional ingredient according to the retention level data of the specific heat-sensitive functional ingredient in the final products of each of the plurality of target historical batches that meet the criteria; The normalized retention level data is used as a weight to perform weighted averaging on the parameter values ​​of the process parameters with a preset degree of influence, thereby obtaining a single parameter value of the process parameter with a preset degree of influence.

7. A method for self-optimizing and controlling process parameters for goat milk powder production according to claim 6, characterized in that: The steps of taking the normalized retention level data as weights and performing weighted averaging on the parameter values ​​of the process parameters with a preset influence degree to obtain a single parameter value of the process parameter with a preset influence degree include: Calculate the product of the weight value of each batch and the parameter value to obtain the weighted parameter value; Sum all weighted parameter values ​​to obtain the weighted sum; Sum all weight values ​​to get the total weight; The weighted sum is divided by the weight sum to obtain the single parameter value.

8. A goat milk powder production process parameter self-optimization control system, characterized in that: The system includes: A data recording module for recording historical production data, the historical production data including raw milk property data of historical batches, process parameter data covering key steps in goat milk powder production based on the raw milk property data of the historical batches, and retention level data of specific heat-sensitive functional ingredients in the final product produced using the process parameter data; The current data acquisition module is used to obtain the characteristic data of the current batch of raw milk to be produced; a similarity comparison module for performing a similarity comparison between the acquired characteristic data of the current batch of raw milk to be produced and the characteristic data of each historical batch of raw milk in the recorded historical production data, so as to determine one or more historical batches whose characteristics match those of the current batch of raw milk to be produced; a target batch screening module, configured to screen at least one target historical batch from the one or more historical batches based on the determined one or more historical batches and according to a preset retention level target for the specific heat-sensitive functional ingredient and other production constraints, wherein the retention level data of the specific heat-sensitive functional ingredient in the final product of the target historical batch meets the preset retention level target; The parameter extraction module is used to extract the process parameter data covering the key processes of goat milk powder production corresponding to at least one selected target historical batch as the recommended process parameters for the current batch of raw milk to be produced.

Images