Milk powder based on probiotic composition and production process control method thereof

Abstract

The invention relates to the technical field of milk powder production control, and discloses milk powder based on a probiotic composition and a production process control method thereof.The production process control method of the milk powder comprises the steps that a mathematical model is constructed, first milk powder to be produced is extracted from a production database based on a regulation factor list, and the first milk powder to be produced is extracted; obtaining a milk powder production plan of first milk powder to be produced; the prediction model outputs an ingredient prediction result and an equipment performance prediction result within prediction time based on the first change sequence and the second change sequence; adjusting control parameters of the production equipment based on a comparison result, updating the regulation and control factor list in real time, and generating an updated regulation and control list; and the mathematical model extracts a milk powder production plan of the next milk powder to be produced based on the updated regulation and control list, and the step is repeated until all the milk powder to be produced is extracted in sequence. The production efficiency of the milk powder of the probiotic composition and the precision of production control are improved.

Description

Technical Field

[0001] The present application relates to the technical field of milk powder production control, and in particular to a milk powder based on a probiotic composition and a production process control method thereof. Background Art

[0002] In the field of food processing, there are abundant types of milk powder of probiotic compositions. As a common probiotic carrier product, milk powder of probiotic compositions faces many technical challenges in the production process. The formula of milk powder of probiotic compositions is complicated. In addition to basic raw materials such as milk powder, it is also necessary to add various functional ingredients such as probiotics, vitamins, minerals, and prebiotics. There are interactions between different ingredients, and their ratios directly affect the nutritional value of the product. Therefore, it is necessary to solve the problems of probiotic activity guarantee, ingredient ratio control, and production process control faced by milk powder of probiotic compositions in the production process.

[0003] Similar prior art includes a Chinese patent application with publication number CN118963303A, which discloses an intelligent monitoring method and system for milk powder production lines based on artificial intelligence, which obtains the target production line equipment status data from the milk powder production line, identifies the target abnormal node reflecting the target abnormal monitoring project, and generates a corresponding first state path vector. Then, based on the production line equipment status data of the current and next equipment, the abnormal monitoring project status data reflecting the abnormal monitoring project is obtained, and the corresponding second state path vector is generated. After being fused with the first state path vector and output as the target state path vector, the abnormal indicator of the abnormal monitoring project is determined, thereby obtaining the linkage characteristic parameters of the abnormal indicator of the abnormal monitoring project in the next equipment status data, thereby determining the abnormal monitoring project node reflecting the abnormal state partition of the target abnormal monitoring project in the next equipment status data, and then predicting the possible abnormalities.

[0004] The shortcomings of the existing technology are mainly manifested in that it only monitors the abnormalities of the equipment status data, and does not regulate the various production equipment in the milk powder production process. In actual situations, milk powders with various types of probiotic compositions need to be finely regulated according to the production status. Summary of the invention

[0005] The present application provides a milk powder based on a probiotic composition and a production process control method thereof, which are used to improve the production efficiency of the milk powder of the probiotic composition and the accuracy of production control.

[0006] In a first aspect, the present application provides a production process control method, which is applied to the production process of milk powder based on a probiotic composition, and the production process control method comprises: Obtaining the type names, ingredient ratios, process requirements and production volume requirements of all milk powders to be produced, establishing a production database based on the type names, ingredient ratios, process requirements and production volume requirements, summarizing the remaining amount of ingredients, equipment performance data and ingredient equipment distribution to generate a list of control factors, and constructing a mathematical model. The mathematical model extracts the first milk powder to be produced in the production database based on the list of control factors, and obtains a milk powder production plan for the first milk powder to be produced based on the list of control factors; Based on the milk powder production plan, the remaining amount of ingredients and the equipment performance data of the production equipment are respectively collected during the milk powder production process to generate a first change sequence and a second change sequence, and a prediction model is established based on a neural network, wherein the prediction model outputs an ingredient prediction result and an equipment performance prediction result within a prediction time based on the first change sequence and the second change sequence; Compare the equipment performance prediction result with the process requirement, output the comparison result, adjust the control parameters of the production equipment in the milk powder production process based on the comparison result, update the control factor list in real time based on the ingredient prediction result and the equipment performance prediction result, and generate an updated control factor list; Based on the production volume demand, the production end time of the first milk powder to be produced is obtained, the mathematical model extracts the milk powder production plan of the next milk powder to be produced in the production database based on the updated control list, and the production end time is set as the production start time of the next milk powder to be produced, and this step is repeated until all the milk powder to be produced are extracted in turn and the milk powder production plans of all the milk powder to be produced are completed.

[0007] In combination with the first aspect, constructing the mathematical model includes: Extracting the milk powder to be produced from the production database and setting it as the planned production milk powder; The first constraint condition for the planned production of milk powder is set based on the ingredient ratio and the remaining amount of the ingredient, the first constraint condition is divided into an integer constraint condition and a nonlinear constraint condition, the elements in the nonlinear constraint condition are replaced with different first variables, the first variable is vectorized to generate a vector variable, the nonlinear constraint condition is linearized based on the vector variable to generate a first linear condition, and the integer constraint is converted into a continuous variable constraint; An objective function is established based on the production volume demand, decision variables are set based on the ingredient ratio, and the objective function, the decision variables, the first linear condition and the continuous variable constraint are integrated to generate the mathematical model.

[0008] In combination with the first aspect, the mathematical model extracts the first milk powder to be produced in the production database based on the list of control factors, comprising the following steps: Combining the first linear condition and the continuous variable constraint to generate a constraint condition, constructing a first optimization problem based on the constraint condition, solving the first optimization problem based on the mathematical model based on a linear programming algorithm, and generating a first ingredient combination amount for the planned production of milk powder; Determine whether the first ingredient combination quantity meets the ingredient ratio corresponding to the planned production of milk powder; if so, update the vector variable based on the first ingredient combination quantity, generate new constraints based on the updated vector variables, and construct a second optimization problem based on the new constraints; if not, construct a second optimization problem based on the remaining constraints, and repeat this step until all constructed optimization problems are solved based on the objective function, wherein the optimization problem includes the first optimization problem and the second optimization problem; The number of conforming items corresponding to any of the planned production milk powders is counted based on the optimization problem, and the planned production milk powder corresponding to the most conforming items is set as the first milk powder to be produced.

[0009] In combination with the first aspect, obtaining a milk powder production plan for the first milk powder to be produced comprises: The production equipment is extracted from all the batching equipment based on the ingredient ratio of the first milk powder to be produced, a production route is generated based on the arrangement order of the production equipment, the production time is calculated based on the production route, and the production route, the production time and the ingredient ratio are combined and set as the milk powder production plan for the first milk powder to be produced.

[0010] Combined with the first aspect, the ingredient prediction results within the prediction time are output, including: sequentially obtaining first changes of the first change sequence between any two adjacent acquisition time points, generating a change sequence by counting all the first changes based on the acquisition time points, marking the change sequence as an outflow label, an inflow label, and a coincidence label based on the size of the first change, verifying the accuracy of the data sequence corresponding to the outflow label based on the milk powder production plan, performing data correction on the second change sequence based on the accuracy and generating a third change sequence; After setting data category labels for the first change sequence and the third change sequence, the first change sequence and the third change sequence are combined based on the acquisition time point to generate a combined sequence, and the correlation between any two groups of subsequences in the combined sequence is obtained. If the correlation is greater than or equal to a first threshold, the correlation is set as a supplementary feature of the two groups of subsequences, and the combined sequence is pre-processed and then input into the prediction model. The prediction model trains the combined sequence based on all the supplementary features, and outputs a prediction result within the prediction time based on the data category labels; Verifying the accuracy of the prediction model, and adjusting the prediction result based on the accuracy of the prediction model to generate a standard prediction result; The standard prediction result whose data category label belongs to the remaining amount of the ingredient is set as the ingredient prediction result.

[0011] In combination with the first aspect, a standard prediction result in which the data category label belongs to device performance is set as the device performance prediction result.

[0012] In combination with the first aspect, adjusting the control parameters of the production equipment in the milk powder production process based on the comparison result includes: Setting a controller, the controller collects the control value of the production equipment at the end of the last time, sets the result of dividing the control value by the first preset value as the second change amount, and the controller adjusts the current control parameter to the control value based on the second change amount within a predetermined time; Obtaining parameter types included in the control parameters, obtaining a difference between the equipment performance prediction result and the process requirement corresponding to the same parameter type, obtaining a prediction time point corresponding to the equipment performance prediction result, fitting the difference to generate a first curve based on the prediction time point, and splitting the first curve into a stable state and an unstable state based on the slope of the first curve; The controller adjusts the control parameters based on the stable state or the unstable state, wherein the control parameters include parameters of the control device corresponding to the temperature, pressure and ingredient flow rate in the production equipment.

[0013] In combination with the first aspect, the controller adjusts the control parameter based on the stable state, including: In the stable state, the controller monitors the control value based on the equipment performance prediction result. If the control value is greater than the corresponding value in the process requirement, the control parameter is reduced based on the second change amount before the prediction time point. If the control value is less than the corresponding value in the process requirement, the control parameter is increased based on the second change amount before the prediction time point.

[0014] In combination with the first aspect, the controller adjusts the control parameter based on the unstable state, including: In the unstable state, the third and fourth changes are set based on the second change. If the difference is greater than the third change, the control parameter is reduced based on the third change before the predicted time point. If the difference is less than the fourth change, the control parameter is increased based on the fourth change before the predicted time point.

[0015] In a second aspect, the present application provides a milk powder based on a probiotic composition, which is produced using the above-mentioned production process control method.

[0016] In the technical solution provided by the present application, firstly, by constructing a mathematical model based on linear programming, and combining methods such as multi-objective optimization, dynamic adjustment and hierarchical solution, the production database is screened, the milk powder to be produced is extracted in turn and the corresponding milk powder production plan is formulated, the optimal configuration of ingredients, equipment and time in the production process of probiotic composition milk powder is achieved, and the amount of each ingredient used can be accurately controlled to avoid ingredient waste or product quality problems caused by inaccurate ingredient ratios. Then, by establishing a prediction model, and combining methods such as data correction, correlation analysis and model verification, the state of the production equipment of the milk powder to be produced in the production process is predicted and analyzed in turn, and accurate ingredient prediction and performance prediction of production equipment can be achieved.

[0017] This application also applies the ingredient prediction results and equipment performance prediction results to the control of the production process. By comparing the results, the control parameters of the production equipment in the milk powder production process are adjusted to achieve refined control of the production process. Predictions are made based on data category labels and prediction model output results, and standard prediction results are generated, providing data support for production decisions. This makes the production process more scientific and intelligent, and helps improve product quality and production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0019] Figure 1 This is a schematic diagram of an embodiment of a production process control method based on a probiotic composition milk powder in the embodiment of the present application. DETAILED DESCRIPTION

[0020] The present application embodiment provides a milk powder based on a probiotic composition and a production process control method thereof. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0021] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 , an embodiment of a production process control method in the embodiments of the present application, the production process control method is applied to the production process of milk powder based on the probiotic composition, comprising: S101. Obtain the type names, ingredient ratios, process requirements and production volume requirements of all milk powders to be produced, establish a production database based on the type names, ingredient ratios, process requirements and production volume requirements, summarize the remaining ingredients, equipment performance data and ingredient equipment distribution to generate a list of control factors, build a mathematical model, extract the first milk powder to be produced in the production database based on the list of control factors, and obtain a milk powder production plan for the first milk powder to be produced based on the list of control factors.

[0022] Specifically, the milk powder to be produced refers to the milk powder of a variety of probiotic compositions to be produced, and the ingredients of the milk powder of the probiotic composition include but are not limited to the basic milk powder component, bifidobacterium, lactobacillus, thermophilic streptococcus, etc. The species name refers to the name of the milk powder to be produced after the different probiotic compositions are formulated. The ingredient ratio refers to the planned detailed ratio of various ingredients in the milk powder to be produced, for example, milk powder, skim milk powder, whey protein powder, etc. of the basic milk powder component, the type and amount of probiotics added, for example, bifidobacterium, lactobacillus acidophilus, etc., and vitamins and minerals of other auxiliary materials. The process requirements refer to the production process parameters of each ingredient in the milk powder to be produced, for example, strain activity, mixing time, temperature, sterilization temperature, etc. The production volume demand refers to the planned production quantity of the milk powder to be produced. The above-mentioned acquired data are summarized to establish a structured production database. The remaining amount of ingredients refers to the remaining amount information of various ingredients, and the equipment performance data refers to the data of factors associated with the process requirements in the equipment used to produce the ingredients, for example, performance data such as pressure, temperature, and flow. The distribution of batching equipment refers to the spatial distribution of various production equipment when mixing the ingredients for the milk powder to be produced. Usually, the batching equipment is connected by pipelines or conveyor belts, where the production equipment refers to the equipment used to produce or store the ingredients associated with the milk powder to be produced. The above-mentioned remaining ingredients, equipment performance data and the information contained in the distribution of batching equipment are summarized to generate a list of control factors for the formulation of subsequent production plans.

[0023] Mathematical model refers to a set of mathematical equations and constraints used to describe and solve specific problems. Mathematical model usually includes objective function and a set of constraints, which together define the feasible solution space and optimal solution of the problem. Therefore, the mathematical model selects the appropriate milk powder to be produced from the production database according to the list of control factors and formulates the corresponding production plan. The factors that the mathematical model may need to consider include but are not limited to the sufficiency of ingredients corresponding to the remaining amount of ingredients, the environmental state corresponding to the equipment performance data, and the utilization rate of the batching equipment corresponding to the distribution of the batching equipment. The first milk powder to be produced refers to the first milk powder to be produced extracted by the mathematical model from all the milk powders to be produced through the production database. Among them, various milk powders to be produced are batch produced according to type. Therefore, it is necessary to extract each milk powder to be produced in turn before production, which can improve the refined control of the milk powder production process, thereby improving production efficiency, ensuring product quality and optimizing resource utilization.

[0024] S102. Based on the milk powder production plan, the remaining amount of ingredients and equipment performance data of the production equipment are collected during the milk powder production process to generate a first change sequence and a second change sequence, and a prediction model is established based on a neural network. The prediction model outputs ingredient prediction results and equipment performance prediction results within a prediction time based on the first change sequence and the second change sequence.

[0025] Specifically, the remaining amount of ingredients and equipment performance data are key factors that affect the milk powder to be produced in the milk powder production process. Therefore, it is necessary to collect the remaining amount of ingredients and equipment performance data in real time, and generate the first change sequence and the second change sequence based on the collection time point. Since there are multiple ingredients corresponding to the remaining amount of ingredients and multiple equipment performances corresponding to the equipment performance data, there are multiple first change sequences and second change sequences. Select a suitable neural network model, such as a long short-term memory network (LSTM), as a prediction model for processing time series data. The prediction model predicts the remaining amount of ingredients and changes in the production environment within the future prediction time based on the current production data, that is, the ingredient prediction results and the equipment performance prediction results.

[0026] S103, comparing the equipment performance prediction results with the process requirements, outputting the comparison results, adjusting the control parameters of the production equipment in the milk powder production process based on the comparison results, updating the control factor list in real time based on the ingredient prediction results and the equipment performance prediction results, and generating an updated control list.

[0027] Specifically, the predicted equipment performance prediction results are compared with the pre-set process requirements, such as whether the parameters such as temperature, humidity, and pressure are within the range of process requirements. According to the comparison results, the control parameters of the production equipment are adjusted, for example, the mixing time, temperature, speed and other parameters are adjusted to ensure the activity of probiotics, the drying temperature, time and other parameters are adjusted to ensure that the moisture content of milk powder meets the standard, and the sterilization temperature, time and other parameters are adjusted to ensure product safety. Among them, the control parameters refer to the parameters of the control device corresponding to the performance of each equipment.

[0028] According to the batching prediction results and environmental prediction results, the list of control factors is updated in real time, for example, the remaining amount of ingredients, equipment performance data, batching equipment distribution and other information are updated to generate an updated control list for subsequent production plan adjustments.

[0029] S104. Based on the production volume demand, the production end time of the first milk powder to be produced is obtained. The mathematical model extracts the milk powder production plan of the next milk powder to be produced from the production database based on the updated control list, and the production end time is set as the production start time of the next milk powder to be produced. This step is repeated until all the milk powder to be produced are extracted in sequence and the milk powder production plans of all the milk powder to be produced are completed.

[0030] Specifically, the mathematical model extracts the next type of milk powder to be produced from the production database based on the updated control list. According to the ingredient ratio, process requirements and production volume requirements of the milk powder type, combined with the current updated control list, the production plan of the milk powder type is adjusted. Repeat the above steps until all types of milk powder to be produced are extracted in turn, and the production plans of all types of milk powder are completed. Among them, the production end time of the first milk powder to be produced can be calculated according to the mathematical model based on the production demand, and the production start time of the next milk powder to be produced can be set according to the production end time of the first milk powder to be produced, which can effectively improve the production efficiency of all milk powders to be produced.

[0031] In a specific embodiment, constructing a mathematical model includes the following steps: (1) Extract a milk powder to be produced from the production database and set it as the planned production milk powder.

[0032] (2) The first constraint condition for planning the production of milk powder is set based on the ingredient ratio and the remaining amount of the ingredients. The first constraint condition is divided into an integer constraint condition and a nonlinear constraint condition. The elements in the nonlinear constraint condition are replaced by different first variables. The first variable is vectorized to generate a vector variable. The nonlinear constraint condition is linearized based on the vector variable to generate a first linear condition. The integer constraint is converted into a continuous variable constraint.

[0033] (3) Establish an objective function based on production volume requirements, set decision variables based on ingredient ratios, integrate the objective function, decision variables, first linear conditions, and continuous variable constraints, and generate a mathematical model.

[0034] Specifically, a milk powder to be produced is randomly selected as the planned production milk powder, and a mathematical model is constructed based on the planned production milk powder. The first constraint condition is set based on the ingredient ratio and the remaining amount of ingredients. For example, the ingredient ratio information of the planned production milk powder is obtained from the production database, for example, 50% milk powder, 20% skim milk powder, 15% whey protein powder, 0.1% probiotics, and 14.9% other auxiliary materials. The current remaining amount of ingredients is obtained from the list of control factors, for example, 1000kg of milk powder remaining, 500kg of skim milk powder remaining, 300kg of whey protein powder remaining, 1kg of probiotics remaining, and 200kg of other auxiliary materials remaining. The first constraint condition refers to setting the ingredient usage restriction conditions for the planned production of milk powder based on the ingredient ratio and the remaining ingredient quantity. Since there are multiple types of ingredient ratios and ingredient remaining quantities, there are also multiple first constraints. For example, the usage of milk powder ≤1000kg, the usage of skimmed milk powder ≤500kg, the usage of whey protein powder ≤300kg, the usage of probiotics ≤1kg, the usage of other auxiliary materials ≤200kg, etc.

[0035] Integer constraints refer to constraints involving integer amounts of ingredients, for example, the amount of probiotics used (grams). Nonlinear constraints refer to constraints involving nonlinear relationships between ingredients. For example, the amount of some auxiliary ingredients used may be proportional to the amount of other ingredients used, which can be described by a mathematical formula: Other auxiliary ingredients used = k × (milk powder used + skim milk powder used + whey protein powder used), where k is the proportional coefficient. Replacing the elements in a nonlinear constraint with different first variables means using the factors in the nonlinear constraint as the first variable. For example, the amount of milk powder used, the amount of skim milk powder used, and the amount of whey protein powder used are replaced by the first variables x1, x2, and x3, respectively, and the amount of other auxiliary ingredients used is replaced by the first variable x4. The nonlinear constraint corresponding to the above mathematical formula can be expressed as: x4 = k × (x1 + x2 + x3). Similarly, the remaining amount of ingredients can be inferred from the amount of ingredients used.

[0036] Vectorization refers to replacing the first variable in the nonlinear constraint with a set of new variables (vector variables). These new variables make the original nonlinear relationship linearized, for example, vector variable X=[x1, x2, x3, x4]. Based on the vector variables, the nonlinear constraint condition is linearized to generate the first linear condition. For example, the above nonlinear constraint condition is linearized to: x4-kx1-kx2-kx3=0. Convert integer constraints to continuous variable constraints, for example, the amount of probiotics used (g) ∈ [0, 1000]. In the subsequent optimization process, the continuous variable results can be converted to integers by rounding or other methods.

[0037] The objective function usually aims to minimize production costs or maximize production efficiency. For example, assuming that the objective function is to minimize production costs: objective function = c1*x1+c2*x2+c3*x3+c4*x4, where c1, c2, c3, and c4 are the unit costs of milk powder, skim milk powder, whey protein powder, and other auxiliary materials, respectively. The decision variable is the element factor corresponding to the first variable contained in the objective function. For example, the decision variables are x1, x2, x3, and x4. Integrate the objective function, decision variables, first linear conditions, and continuous variable constraints established in the above steps to generate the final mathematical model.

[0038] In a specific embodiment, the mathematical model extracts the first milk powder to be produced from the production database based on the list of control factors, including the following steps: (1) The first linear condition and the continuous variable constraint are combined to generate constraint conditions, a first optimization problem is constructed based on the constraint conditions, and the mathematical model is used to solve the first optimization problem based on a linear programming algorithm to generate a first ingredient combination quantity for planned production of milk powder.

[0039] (2) Determine whether the first ingredient combination quantity meets the ingredient ratio corresponding to the planned production of milk powder. If so, update the vector variable based on the first ingredient combination quantity, generate new constraints based on the updated vector variables, and construct a second optimization problem based on the new constraints. If not, construct a second optimization problem based on the remaining constraints. Repeat this step until all constructed optimization problems are solved based on the objective function, wherein the optimization problem includes the first optimization problem and the second optimization problem.

[0040] (3) Based on the optimization problem, the number of eligible items corresponding to any planned production of milk powder is counted, and the planned production of milk powder corresponding to the most eligible items is set as the first milk powder to be produced.

[0041] In a specific embodiment, obtaining a milk powder production plan for a first milk powder to be produced includes: Based on the ingredient ratio of the first milk powder to be produced, production equipment is extracted from all the ingredient equipment, a production route is generated based on the arrangement order of the production equipment, the production time is calculated based on the production route, and the production route, production time and ingredient ratio are combined and set as the milk powder production plan for the first milk powder to be produced.

[0042] Specifically, the above linear conditions and continuous variable constraints are combined to form a complete set of constraints, that is, they are all constraints. The first optimization problem refers to the linearized problem obtained according to the constraints, which can simplify the original problem and make it easier to solve. For example, the amount of probiotics used (g) ∈ [0, 1000] in the integer constraint is converted into a continuous variable constraint 0≤m1≤100, where m1 is the amount of probiotics used. Linear Programming (LP) is a mathematical optimization method used to find the optimal value (maximum or minimum value) of the objective function under a given set of linear constraints. The first optimization problem is solved based on the linear programming algorithm to generate the first ingredient combination amount, wherein the simplex method, interior point method and other linear programming solving algorithms can be used. The first ingredient combination amount refers to the ingredient usage used for the production plan to produce milk powder, for example, x1=800kg, x2=400kg, x3=200kg, x4=140kg.

[0043] Determine whether the first ingredient combination quantity meets the ingredient ratio corresponding to the planned production of milk powder. For example, the ingredient ratio of the planned production of milk powder is 50% milk powder, 20% skim milk powder, 15% whey protein powder, 0.1% probiotics, and 14.9% other auxiliary materials. The determination method is: calculate the proportion of each ingredient in the first ingredient combination quantity, which is milk powder ratio = 800 / (800+400+200+140)≈52.6%, skim milk powder ratio = 400 / (800+400+200+140)≈26.3%, The proportion of whey protein powder = 200 / (800+400+200+140)≈13.2%, the proportion of other auxiliary materials = 140 / (800+400+200+140)≈9.2%, among which, the proportion of milk powder (52.6%)>50%, meets the requirements; the proportion of skimmed milk powder (26.3%)>20%, meets the requirements; the proportion of whey protein powder (13.2%)<15%, does not meet the requirements; the proportion of other auxiliary materials (9.2%)<14.9%, does not meet the requirements, and there are 2 corresponding non-compliant items.

[0044] If the first ingredient combination quantity meets the ingredient ratio, the vector variable is updated based on the first ingredient combination quantity. The vector variable is updated by substituting the first ingredient combination quantity into the vector variable, for example, g=[x1, x2, x3, x4]=[800, 400, 200, 140]. Based on the updated vector variable, a new constraint condition can be generated.

[0045] Since different constraints can generate different optimization problems, the number of optimization problems can be the same as the number of constraints. Then, the objective function corresponding to each of the planned production milk powders extracted is solved in turn according to the optimization problem, so that all optimization problems are tested, wherein the optimization problem includes but is not limited to the first optimization problem and the second optimization problem.

[0046] When all the milk powder to be produced is extracted as the planned production milk powder, the number of conforming items of each planned production milk powder can be counted in turn, and the planned production milk powder with the most conforming items can be used as the first milk powder to be produced, which can effectively improve the efficiency control of the milk powder production process.

[0047] Production equipment refers to the batching equipment associated with the first milk powder to be produced. For example, milk powder batching equipment, skimmed milk powder batching equipment, etc., determine the batching equipment to be used based on the batching ratio. For example, if the batching ratio is 50% milk powder, milk powder batching equipment is required. The arrangement order refers to the physical arrangement order of the production equipment in the production workshop. For example, the production route is milk powder batching equipment → skimmed milk powder batching equipment → whey protein powder batching equipment → other auxiliary ingredients batching equipment. According to the processing capacity of the production equipment, the number of ingredients, the production speed and other factors, calculate the production time required for each batching equipment, accumulate the required time of all production equipment, and get the production time, which can be summarized to generate the milk powder production plan for the first milk powder to be produced.

[0048] In a specific embodiment, outputting the batching prediction result within the prediction time may specifically include the following steps: (1) The first change amount of the first change sequence between any two adjacent collection time points is obtained in sequence, and a change amount sequence is generated by counting all the first change amounts based on the collection time points. The change amount sequence is marked as an outflow label, an inflow label, and an overlap label based on the size of the first change amount. The accuracy of the data sequence corresponding to the outflow label is verified based on the milk powder production plan. The second change sequence is corrected based on the accuracy and a third change sequence is generated.

[0049] (2) After setting data category labels for the first change sequence and the third change sequence, they are combined based on the acquisition time point to generate a combined sequence, and the correlation between any two groups of subsequences in the combined sequence is obtained. If the correlation is greater than or equal to a first threshold, the correlation is set as a supplementary feature of the two groups of subsequences. The combined sequence is preprocessed and then input into the prediction model. The prediction model trains the combined sequence based on all the supplementary features and outputs the prediction results within the prediction time based on the data category labels.

[0050] (3) Verify the accuracy of the prediction model and adjust the prediction results based on the accuracy of the prediction model to generate standard prediction results.

[0051] (4) The standard prediction result whose data category label belongs to the remaining amount of ingredients is set as the ingredient prediction result.

[0052] In a specific embodiment, a standard prediction result whose data category label belongs to device performance is set as the device performance prediction result.

[0053] Specifically, when producing the first milk powder to be produced, it is necessary to monitor each production equipment in real time. The first change refers to the change in the remaining amount of the ingredient between two adjacent collection time points. For example, the remaining amount of milk powder at time t1 is 100, and at time t2 it is 95, then the first change is -5, wherein, if the first change is a negative number, it indicates that the ingredient is flowing out, which can correspond to the outflow label, if the first change is a positive number and is greater than the ingredient ratio, it means that the ingredient is being added, which can correspond to the inflow label, otherwise, it means that the ingredient is both flowing out and being added at the same time, which can correspond to the overlap label. The change sequence is a collection of first changes, which are arranged by collection time points, thereby, any change in the change sequence can be marked with the corresponding label (outflow label, inflow label or overlap label) in turn.

[0054] The ingredient ratio shows the amount of ingredients used for the production equipment. The first variation belonging to the outflow tag is extracted from the variation sequence, that is, the data sequence. The difference between any first variation in the data sequence and the amount of ingredients used is calculated, and all differences are averaged to generate the accuracy of the data sequence. Since the second variation sequence will change due to the first variation sequence, the data is corrected according to the product of the accuracy and the second variation sequence, and the second variation sequence is corrected to the third variation sequence. Since the equipment performance data may be multiple, the second variation sequence will also be multiple, and the corresponding third variation sequence is also multiple.

[0055] The data category label is the table label information used to label the first change sequence data and the third change sequence data, including but not limited to the ingredient remaining quantity label and the equipment performance label. Preprocessing refers to data preprocessing, including but not limited to data cleaning, data standardization, data segmentation (training set, validation set and test set), etc.

[0056] The combined sequence is input into the prediction model to obtain the correlation between any two groups of subsequences in the combined sequence. A subsequence refers to any group of change sequences (the first change sequence or the third change sequence) in the combined sequence. The correlation calculation can use methods such as the Pearson correlation coefficient and mutual information to calculate the correlation between the two groups of subsequences. If the correlation is greater than or equal to the first threshold, the correlation is set as a supplementary feature of the two groups of subsequences, and the correlation is added to the original data as a new supplementary feature to enhance the prediction ability of the prediction model.

[0057] In the prediction model, the combined sequence and supplementary features are used as input, and the prediction target corresponding to the data category label is used as output. The prediction model is trained using the training set data, the model parameters are optimized, and the prediction results within the prediction time are output. The product of the accuracy rate and the prediction result is used as the standard prediction result, and then the ingredient prediction result and the equipment performance prediction result are output according to the data category label. For example, the ingredient prediction result corresponds to the predicted value of the remaining amount of each ingredient in the future period, and the equipment performance prediction result corresponds to the predicted value of each equipment performance indicator in the future period.

[0058] In a specific embodiment, adjusting the control parameters of the production equipment in the milk powder production process based on the comparison result may specifically include the following steps: (1) A controller is set, the controller collects the control value of the production equipment at the last end, divides the control value by a first preset value and sets the result as a second change amount, and the controller adjusts the current control parameter to the control value based on the second change amount within a predetermined time.

[0059] (2) Obtaining the parameter type included in the control parameter, obtaining the difference between the equipment performance prediction result and the process requirement corresponding to the same parameter type, obtaining the prediction time point corresponding to the equipment performance prediction result, fitting the difference based on the prediction time point to generate a first curve, and splitting the first curve into a stable state and an unstable state based on the slope of the first curve.

[0060] (3) The controller adjusts control parameters based on a stable state or an unstable state, wherein the control parameters include parameters of control devices corresponding to temperature, pressure, and ingredient flow in the production equipment.

[0061] Specifically, the controller can be a programmable logic controller (PLC) or a similar control system. The last end refers to the time before the control device of the production equipment is adjusted. The control value refers to the set value of each control device (for example, temperature, pressure, batching flow controller, etc.) in the production equipment at the end of the last production. The control parameter refers to the parameter of the control device used to adjust the control value in the production equipment. The first preset value is a pre-set value used to calculate the second change amount. For example, if the first preset value is 5, the second change amount is 1 / 5 of the control value. The second change amount is used to gradually adjust the current control parameter to the control value within a predetermined time. The predetermined time refers to the time required from the start of production to the stable operation of the system, which can be adjusted according to actual conditions. The controller gradually increases or decreases the control parameter so that it reaches the control value within a predetermined time. For example, if the control value is 100 and the first change amount is 20, the controller gradually increases the control parameter from 0 to 100 within 15 seconds, increasing by 20 each time.

[0062] Parameter type refers to the category information of the control parameter, such as temperature, pressure, batching flow, etc. Equipment performance prediction results refer to the predicted values ​​of the above-mentioned prediction model for the performance of the production equipment (such as temperature, pressure, flow, etc.). Process requirements refer to the required values ​​of equipment performance data in the production process, such as target temperature, target pressure, target flow, etc. The same parameter type is calculated for difference, where difference = equipment performance prediction result - process requirements. With time as the horizontal axis and difference as the vertical axis, the trend of the difference over time is plotted into a curve, namely the first curve. The stable state refers to the stage where the difference changes slowly or tends to be stable, and the slope is close to 0. The unsteady state refers to the stage where the difference changes violently or fluctuates greatly, and the absolute value of the slope is large. Generally, when the production equipment flows out or flows in the batching, it will affect the internal equipment performance data. Therefore, it is necessary to perform state analysis on the first curve.

[0063] Under different conditions, the control and adjustment of different control devices can make the production equipment meet the needs and realize the refined control of the production equipment parameters, thereby ensuring the stable and efficient operation of the production process.

[0064] In a specific embodiment, the controller adjusts the control parameters based on the stable state, including the following steps: In the stable state, the controller monitors the control value based on the equipment performance prediction result. If the control value is greater than the corresponding value in the process requirement, the control parameter is reduced based on the second change amount before the prediction time point. If the control value is less than the corresponding value in the process requirement, the control parameter is increased based on the second change amount before the prediction time point.

[0065] In a specific embodiment, the controller adjusts the control parameters based on the non-stable state, including: In an unstable state, the third and fourth changes are set based on the second change. If the difference is greater than the third change, the control parameter is reduced based on the third change before the predicted time point. If the difference is less than the fourth change, the control parameter is increased based on the fourth change before the predicted time point.

[0066] Specifically, in a stable state, the difference between the equipment performance prediction result and the process requirement is small, and the system runs relatively smoothly. The controller needs to be fine-tuned on this basis to ensure that the system runs stably and is as close to the process requirement as possible. If the control value is greater than the process requirement, the control parameter is reduced based on the second change before the prediction time point. Among them, the second change refers to the amplitude of each adjustment of the control parameter in a stable state, which is usually less than the first change to avoid causing drastic fluctuations in the system. The adjustment method is that the controller gradually reduces the control parameter so that it is close to the process requirement before the prediction time point. For example, if the current temperature control value is 100°C, the process requirement is 95°C, and the second change is 1°C / minute, the controller adjusts the temperature control parameter within 5 minutes, so that the control value is gradually reduced from 100°C to 95°C. If the control value is less than the process requirement, the control parameter is increased based on the second change before the prediction time point. The adjustment method is that the controller gradually increases the control parameter so that it is close to the process requirement before the prediction time point. For example, if the current pressure control value is 3MPa, the process requirement is 3.5MPa, and the second change is 0.1MPa / minute, the controller increases the pressure control parameter within 5 minutes, so that the pressure control value gradually increases from 3MPa to 3.5MPa. If the predicted result changes, the controller adjusts the increase and decrease amplitude and speed of the control parameter accordingly. For example, if the predicted temperature rises suddenly, the controller can speed up the reduction speed of the temperature control parameter.

[0067] In an unstable state, the difference between the equipment performance prediction result and the process requirement is large, and the system operation fluctuates greatly. The controller needs to adopt a more active adjustment strategy to quickly restore the system to a stable state. The third variation refers to the larger amplitude of each adjustment of the control parameter in an unstable state, which is used to quickly respond to the change of the difference. The fourth variation refers to the smaller amplitude of each adjustment of the control parameter in an unstable state, which is used for fine adjustment. If the difference is greater than the third variation, the control parameter is reduced based on the third variation before the prediction time point. The adjustment method is that the controller quickly reduces the control parameter to quickly reduce the equipment performance. For example, if the current temperature prediction value is 110°C, the process requirement is 95°C, and the third variation is 2°C / minute, the controller reduces the control value corresponding to the temperature control parameter from 110°C to 95°C within 7.5 minutes. If the difference is not greater than the third variation, it is determined whether the difference is less than the fourth variation. If the difference is less than the fourth variation, the control parameter is increased based on the fourth variation before the prediction time point. The adjustment method is that the controller slowly increases the control parameter to finely adjust the equipment performance. For example, if the current pressure prediction value is 3.2MPa, the process requirement is 3.5MPa, and the fourth change amount is 0.1MPa / minute, the controller increases the control value corresponding to the pressure control parameter from 3.2MPa to 3.5MPa within 3 minutes. If the difference is not less than the fourth change amount, the controller keeps the current control parameter unchanged or makes fine adjustments according to the actual situation.

[0068] The controller continuously monitors the difference and dynamically adjusts the third and fourth changes according to the change in the difference. The dynamic adjustment is: if the difference changes greatly, the controller can adjust the third and fourth changes to adapt to the system changes. For example, if the difference suddenly increases, the controller can increase the third change to speed up the adjustment. By setting different changes and dynamically adjusting them in combination with the predicted time point, the controller can effectively respond to various situations in the production process and ensure product quality and production efficiency.

[0069] The present application also provides a milk powder based on a probiotic composition, which is produced using the above-mentioned production process control method.

[0070] Through the synergy of the above components, firstly, by constructing a mathematical model based on a linear programming algorithm, and combining methods such as multi-objective optimization, dynamic adjustment and hierarchical solution, the production database is screened, the milk powder to be produced is extracted in turn and the corresponding milk powder production plan is formulated, the optimal configuration of ingredients, equipment and time in the production process of probiotic composition milk powder is achieved, and the amount of each ingredient used can be accurately controlled to avoid ingredient waste or product quality problems caused by inaccurate ingredient ratios. Then, by establishing a prediction model, and combining methods such as data correction, correlation analysis and model verification, the status of the production equipment in the production process of the milk powder to be produced is predicted and analyzed in turn, and accurate ingredient prediction and performance prediction of production equipment can be achieved.

[0071] This application also applies the ingredient prediction results and equipment performance prediction results to the control of the production process. By comparing the results, the control parameters of the production equipment in the milk powder production process are adjusted to achieve refined control of the production process. Predictions are made based on data category labels and prediction model output results, and standard prediction results are generated, providing data support for production decisions. This makes the production process more scientific and intelligent, and helps improve product quality and production efficiency.

[0072] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0073] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0074] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A production process control method for milk powder based on a probiotic composition, characterized in that: The production process control method comprises: Obtaining the type names, ingredient ratios, process requirements and production volume requirements of all milk powders to be produced, establishing a production database based on the type names, ingredient ratios, process requirements and production volume requirements, summarizing the remaining amount of ingredients, equipment performance data and ingredient equipment distribution to generate a list of control factors, and constructing a mathematical model. The mathematical model extracts the first milk powder to be produced in the production database based on the list of control factors, and obtains a milk powder production plan for the first milk powder to be produced based on the list of control factors; Based on the milk powder production plan, the remaining amount of ingredients and the equipment performance data of the production equipment are respectively collected during the milk powder production process to generate a first change sequence and a second change sequence, and a prediction model is established based on a neural network, wherein the prediction model outputs an ingredient prediction result and an equipment performance prediction result within a prediction time based on the first change sequence and the second change sequence; Compare the equipment performance prediction result with the process requirement, output the comparison result, adjust the control parameters of the production equipment in the milk powder production process based on the comparison result, update the control factor list in real time based on the ingredient prediction result and the equipment performance prediction result, and generate an updated control factor list; Based on the production volume demand, the production end time of the first milk powder to be produced is obtained, the mathematical model extracts the milk powder production plan of the next milk powder to be produced in the production database based on the updated control list, and the production end time is set as the production start time of the next milk powder to be produced, and this step is repeated until all the milk powder to be produced are extracted in turn and the milk powder production plans of all the milk powder to be produced are completed.

2. The method according to claim 1, characterized in that The mathematical model construction comprises: Extracting the milk powder to be produced from the production database and setting it as the planned production milk powder; The first constraint condition for the planned production of milk powder is set based on the ingredient ratio and the remaining amount of the ingredient, the first constraint condition is divided into an integer constraint condition and a nonlinear constraint condition, the elements in the nonlinear constraint condition are replaced with different first variables, the first variable is vectorized to generate a vector variable, the nonlinear constraint condition is linearized based on the vector variable to generate a first linear condition, and the integer constraint is converted into a continuous variable constraint; An objective function is established based on the production volume demand, decision variables are set based on the ingredient ratio, and the objective function, the decision variables, the first linear condition and the continuous variable constraint are integrated to generate the mathematical model.

3. The method according to claim 2, characterized in that The mathematical model extracts the first milk powder to be produced in the production database based on the list of control factors, comprising the following steps: Combining the first linear condition and the continuous variable constraint to generate a constraint condition, constructing a first optimization problem based on the constraint condition, solving the first optimization problem based on the mathematical model based on a linear programming algorithm, and generating a first ingredient combination amount for the planned production of milk powder; Determine whether the first ingredient combination quantity meets the ingredient ratio corresponding to the planned production of milk powder; if so, update the vector variable based on the first ingredient combination quantity, generate new constraints based on the updated vector variables, and construct a second optimization problem based on the new constraints; if not, construct a second optimization problem based on the remaining constraints, and repeat this step until all constructed optimization problems are solved based on the objective function, wherein the optimization problem includes the first optimization problem and the second optimization problem; The number of conforming items corresponding to any of the planned production milk powders is counted based on the optimization problem, and the planned production milk powder corresponding to the most conforming items is set as the first milk powder to be produced.

4. The method according to claim 3, characterized in that Obtaining a milk powder production plan for the first milk powder to be produced, including: The production equipment is extracted from all the batching equipment based on the ingredient ratio of the first milk powder to be produced, a production route is generated based on the arrangement order of the production equipment, the production time is calculated based on the production route, and the production route, the production time and the ingredient ratio are combined and set as the milk powder production plan for the first milk powder to be produced.

5. The method according to claim 1, characterized in that Output the batching prediction results within the prediction time, including: sequentially obtaining first changes of the first change sequence between any two adjacent acquisition time points, generating a change sequence by counting all the first changes based on the acquisition time points, marking the change sequence as an outflow label, an inflow label, and a coincidence label based on the size of the first change, verifying the accuracy of the data sequence corresponding to the outflow label based on the milk powder production plan, performing data correction on the second change sequence based on the accuracy and generating a third change sequence; After setting data category labels for the first change sequence and the third change sequence, the first change sequence and the third change sequence are combined based on the acquisition time point to generate a combined sequence, and the correlation between any two groups of subsequences in the combined sequence is obtained. If the correlation is greater than or equal to a first threshold, the correlation is set as a supplementary feature of the two groups of subsequences, and the combined sequence is pre-processed and then input into the prediction model. The prediction model trains the combined sequence based on all the supplementary features, and outputs a prediction result within the prediction time based on the data category labels; Verifying the accuracy of the prediction model, and adjusting the prediction result based on the accuracy of the prediction model to generate a standard prediction result; The standard prediction result whose data category label belongs to the remaining amount of the ingredient is set as the ingredient prediction result.

6. The method according to claim 5, characterized in that The standard prediction result whose data category label belongs to device performance is set as the device performance prediction result.

7. The method according to claim 1, characterized in that Adjusting control parameters of production equipment in the milk powder production process based on the comparison result includes: Setting a controller, the controller collects the control value of the production equipment at the end of the last time, sets the result of dividing the control value by the first preset value as the second change amount, and the controller adjusts the current control parameter to the control value based on the second change amount within a predetermined time; Obtaining parameter types included in the control parameters, obtaining a difference between the equipment performance prediction result and the process requirement corresponding to the same parameter type, obtaining a prediction time point corresponding to the equipment performance prediction result, fitting the difference to generate a first curve based on the prediction time point, and splitting the first curve into a stable state and an unstable state based on the slope of the first curve; The controller adjusts the control parameters based on the stable state or the unstable state, wherein the control parameters include parameters of the control device corresponding to the temperature, pressure and ingredient flow rate in the production equipment.

8. The method according to claim 7, characterized in that The controller adjusts the control parameter based on the stable state, including: In the stable state, the controller monitors the control value based on the equipment performance prediction result. If the control value is greater than the corresponding value in the process requirement, the control parameter is reduced based on the second change amount before the prediction time point. If the control value is less than the corresponding value in the process requirement, the control parameter is increased based on the second change amount before the prediction time point.

9. The method according to claim 7, characterized in that: The controller adjusts the control parameter based on the unstable state, including: In the unstable state, the third and fourth changes are set based on the second change. If the difference is greater than the third change, the control parameter is reduced based on the third change before the predicted time point. If the difference is less than the fourth change, the control parameter is increased based on the fourth change before the predicted time point.

10. A milk powder based on a probiotic composition, characterized in that: The milk powder is produced using the production process control method as described in any one of claims 1 to 9.

Images