Control system and method for automatic mixing of materials for premix production

Abstract

The invention belongs to the field of premix production control, and in particular relates to a control system and method for automatic mixing of materials for premix production. The system realizes accurate prediction and adjustment of raw material supply, optimized control of a mixing process, and real-time and offline detection of finished product quality by integrating ratio prediction, feed metering and control, mixing control and quality detection modules. The ratio prediction unit predicts the raw material ratio and feeding sequence based on historical data, the feed metering unit and the feed control unit accurately calculate and control the raw material feeding amount and rate in real time, the mixing control module optimizes stirring parameters by using genetic algorithms and fuzzy PID control, the quality detection module ensures product quality by online and offline detection and provides feedback to optimize the ratio, and the integrated interactive module uniformly manages the operation of each module and the display of results, thereby improving production efficiency and the qualified ratio of finished products, and realizing automation and intelligence of premix production.

Description

Technical Field

[0001] The invention belongs to the field of premix production control, and in particular relates to a control system and method for automatically mixing materials used in premix production. Background Art

[0002] The veterinary drug industry, especially the production of veterinary premixes, has extremely strict requirements on the precise ratio of raw materials. Premixes are usually mixed with one or more active ingredients and carrier materials in a specific ratio to improve animal health, growth performance or disease prevention. These premixes need to be evenly distributed in animal feed to ensure that each feed contains an accurate dose of the drug, which is crucial to maintaining the health of the animal population.

[0003] For example, a Chinese patent with authorization announcement number CN102393634B discloses a single-objective multi-production line scheduling control method based on a PSO-GA hybrid optimization algorithm, including the following steps: inputting optional production lines for different workpieces, the corresponding preparation time for the first batch of processing, the mold change time for different workpieces on the optional production line, the processing time for different workpieces on the optional production line, a single objective function and a processing batch strategy; using random coding to initialize operating parameters; initializing the population; dividing P(t) into two sub-populations A and B with equal particle sizes; setting the number of iterations; performing a PSO operation based on SA local improvement on the particle group P1(t) of the A population; performing a GA operation based on SA local improvement on the particle group P2(t) of the B population; the A and B populations share information; judging whether the termination conditions are met, and if so, outputting the optimal individual.

[0004] For example, a Chinese patent with authorization announcement number CN117631632B discloses a PTA mixed feeding control method and system, the method comprising the following steps: obtaining PTA production parameters and configuring raw materials; setting a production factor control plan; obtaining production monitoring parameters, and setting production monitoring parameter safety thresholds and monitoring sensors; setting a safety interlock plan, and configuring a processing emergency protection plan based on the safety interlock plan; performing PTA mixed feeding processing according to the production factor control plan, and collecting real-time production monitoring parameters; performing interlock judgments based on the production monitoring parameter safety thresholds; implementing a processing emergency protection plan, adjusting and optimizing the safety interlock plan, and updating the processing emergency protection plan.

[0005] The above existing technologies have the following problems: 1) The raw material ratio often depends on manual experience or a fixed formula, and lacks adaptability to changes in the production environment; 2) Traditional feeding control may not be able to accurately control the amount and timing of each raw material feed. In order to solve the above problems, the present invention provides a control system and method for automatic mixing of materials for premix production. Summary of the invention

[0006] In view of the shortcomings of the prior art, the present invention proposes a control system and method for automatic mixing of materials for premix production. The system realizes accurate prediction and adjustment of raw material supply, optimized control of the mixing process and real-time and offline detection of finished product quality by integrating ratio prediction, feed metering and control, mixing control and quality detection modules. The ratio prediction unit predicts the raw material ratio and feeding order based on historical data, the feed metering unit and the feed control unit accurately calculate and control the raw material feeding amount and rate in real time, the mixing control module optimizes the stirring parameters by using genetic algorithm and fuzzy PID control, the quality detection module ensures product quality and feeds back the optimized ratio through online and offline detection, and the integrated interactive module uniformly manages the operation of each module and the display of results, thereby improving production efficiency and the qualified ratio of finished products, and realizing the automation and intelligence of premix production.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A control system for automatic mixing of materials for premix production, including: a raw material supply module and a mixing control module;

[0009] The raw material supply module includes a ratio prediction unit, a feed metering unit and a feed control unit;

[0010] The ratio prediction unit is used to obtain the historical premix production ratio process data, and use the obtained production ratio process data to construct a ratio prediction model to predict the ratio and placement order of raw materials under the corresponding production environment conditions; the feeding metering unit is used to calculate the number of raw material additions and the single amount of each raw material added according to the raw material formula, predicted ratio and production demand; the feeding control unit is used to generate a release control instruction based on the raw material properties, the results output by the feeding metering unit and the ratio prediction unit through the built-in collaborative rate control model, to control the rate and order of each raw material added in the production process in real time, and to collaboratively control the ratio of the added raw materials and the mixing and stirring process of the raw materials;

[0011] The mixing control module includes a stirring parameter optimization unit and a fuzzy control unit;

[0012] The mixing parameter optimization unit is used to obtain the initial optimized mixing and stirring control parameters through a genetic algorithm based on the historical mixing control parameters, effect data, and the rate and order of each raw material addition; the fuzzy control unit is used to generate mixing control instructions through a fuzzy PID control algorithm based on the initial optimized mixing and stirring control parameters, and control the mixing equipment to mix and stir the added raw materials.

[0013] Specifically, the historical premix production ratio process data includes raw material formula, raw material ratio, release order, production demand, raw material type and attribute, production environment and finished product qualified ratio data; the steps of constructing the ratio prediction model in the ratio prediction unit include:

[0014] A1. Obtain the production ratio process data corresponding to the optimal production process of the historical finished product qualified ratio of N types of premixes, and classify and save the obtained data after preprocessing according to the types of premixes;

[0015] A2. Construct input sequence set based on saved production ratio process data ,in Represents the input sequence corresponding to the nth premix;

[0016] A3. Based on the random forest algorithm, a ratio prediction model including N sub-models is constructed in an integrated manner, and the constructed input sequence set is input into the N sub-models for training in a one-to-one correspondence, and the predicted premix raw material ratio and finished product qualified ratio data are output accordingly;

[0017] A4. Obtain the single sub-model prediction error based on each predicted raw material ratio and finished product qualified ratio data and the actual raw material ratio and finished product qualified ratio data in the input sequence set, and use the single sub-model prediction error to construct a minimum error loss function through a cross entropy function;

[0018] A5. Set a loss threshold, use the constructed minimum error loss function to train the ratio prediction model, and obtain the trained ratio prediction model.

[0019] Specifically, the collaborative rate control model in the feeding control unit includes a delivery enhancement adjustment layer and a collaborative enhancement adjustment layer; the delivery enhancement adjustment layer includes m delivery enhancement adjustment sublayers; the specific steps of constructing the delivery enhancement adjustment sublayer include:

[0020] B1. Use the obtained production ratio process data of the nth premix, the single amount of each raw material, the number of times of addition and the order of addition to construct the input state ,in Represents the attribute data of the i-th raw material put into the production of the n-th premix at time t, It represents the single amount of raw material put into the production of n-th premix at time t. It indicates the number of times the i-th raw material is put into production of the n-th premix at time t. It indicates the order of placing the i-th raw material in the production of the n-th premix at time t. It represents the rate of release of the ith raw material for the production of the nth premix at time t, It represents the qualified ratio of the nth premix product produced at time t;

[0021] B2. Set the qualified ratio threshold of the nth premix finished product ,when When the information of adjusting the rate or order of the i-th raw material is triggered, When , the information of the current delivery rate and order of the i-th raw material is triggered.

[0022] Specifically, the specific steps of constructing the enhanced adjustment sub-layer include:

[0023] B3. Construct a joint execution action based on the trigger information obtained in B2 ,in Indicates the action of adjusting the raw material delivery rate when the adjustment of the i-th raw material delivery rate is triggered. Indicates the action of adjusting the order of raw materials when the adjustment of the order of raw materials is triggered. Indicates the action of maintaining the current delivery rate and order of the i-th raw material when the information of maintaining the current delivery rate or order of the i-th raw material is triggered;

[0024] B4. Construct a reward function based on the joint execution actions performed in B3 , specifically:

[0025] ,

[0026] in, Indicates a negative reward when the premix fails to meet the standards. Indicates that when the premix is ​​qualified, a positive reward is given. It represents the difference between the actual delivery rate of the ith raw material and the target delivery rate for the production of the nth premix at time t. It indicates the reward for producing the nth premix at time t and putting the ith raw material in the right order. , when the delivery order is wrong ; Represents the gain coefficient of the reward function;

[0027] B5. Construct a delivery control sub-model based on reinforcement learning, input the constructed input state, joint execution action and reward function into the delivery control sub-model, and generate the delivery control adjustment parameters corresponding to the execution action in B3.

[0028] Specifically, the collaborative rate control model in the feeding control unit also includes a delivery control layer and a collaborative control layer, and the delivery control layer corresponds to the delivery enhancement adjustment layer; the delivery control layer includes m delivery control sublayers, and the m delivery control sublayers correspond one-to-one to the m delivery enhancement adjustment sublayers; the specific steps of constructing the m delivery control sublayers include:

[0029] B6. According to the obtained and and Initialize the control parameters in the delivery control sublayer, use the delivery control adjustment parameters generated in B5 to generate a control signal for a single delivery of raw materials and the combined control signal of m kinds of raw materials , specifically:

[0030]

[0031]

[0032] in, represents the control signal generated to control the rate or order of adding the i-th raw material, , and Sequentially means that the i-th enhancement adjustment sublayer performs actions in sequence , and The corresponding evaluation value generated when It represents the difference between the qualified ratio of premix finished products and the corresponding threshold value, T represents the time length of a single batch of premix mixed raw materials, and Represents the execution action control factor, when and Execute simultaneously and ;when and When the corresponding action is executed ,when and When only execute ,when and When only execute ; Indicates the i-th The combined control factor of Corresponding When the rate or order of the corresponding i-th raw material is adjusted, of When the corresponding raw material delivery rate and order are not adjusted, , and They represent the proportional control parameters, integral control parameters and differential control parameters in the delivery control sublayer respectively.

[0033] Specifically, the collaborative enhancement adjustment layer corresponds to the collaborative control layer; the specific steps of constructing the collaborative enhancement adjustment layer include:

[0034] B7, set The qualified ratio of finished products is obtained by adjusting the raw material delivery rate and sequence through the delivery enhancement adjustment layer and the delivery control layer. When the control signal corresponding to the first collaborative control information is adjusted, the first collaborative control information is triggered, and when the control signal corresponding to the first collaborative control information is adjusted, the first collaborative control information is still satisfied. When the second coordinated control information is triggered;

[0035] B8. According to and Constructing collaborative input state , and construct a collaborative control action based on the first collaborative control information and the second collaborative control information obtained , Indicates the action of collaboratively adjusting the mixing rate of raw materials when the first collaborative control information is triggered. Indicates the action of collaboratively adjusting the ratio of raw materials to be put in when the second collaborative control information is triggered;

[0036] B9. Construct a collaborative control sub-model based on the reinforcement algorithm, and generate corresponding collaborative control adjustment parameters according to the constructed collaborative input state and collaborative control action.

[0037] Specifically, the specific steps of building the collaborative control layer include:

[0038] B10, adjust the parameters and control signals according to the coordinated control obtained in B9 and the synergistic signal generation function to generate the control synergistic control signal for producing the nth premix , the collaborative signal generation function is specifically:

[0039] ,

[0040] in, and Sequentially indicates execution of actions and The corresponding evaluation value is , and It represents the coordinated control factor of feeding adjustment, the coordinated control factor of stirring adjustment and the coordinated control factor of raw material ratio adjustment in turn;

[0041] B11. Under the premise of and Corresponding to the raw material input adjustment of the execution process from B1 to B5, when and No adjustment is performed for and The corresponding adjustment of the mixing process is performed; when and Correspondingly execute the coordinated adjustment of the dosing and mixing process;

[0042] B12. When When the ratio of raw materials is adjusted, the action of adjusting the ratio of raw materials is executed, and the process of B1-B11 is repeated, so that the nth premix is ​​always satisfied. condition.

[0043] Specifically, the mixing parameter optimization unit obtains the initial optimized mixing control parameters in the following specific steps:

[0044] C1. Obtain the mixing and stirring control parameters corresponding to the optimal production process of the qualified finished product ratio of N types of premixes and the data on the rate and sequence of each raw material in the feeding process, and use the obtained data to construct the input chromosome corresponding to the genetic algorithm ,in, represents the stirring rate when producing the nth premix at time t, It represents the stirring time of producing the nth premix at time t. Indicates the number of stirring times per unit time. Indicates the stirring direction. It represents the rate of release of the ith raw material for the production of the nth premix at time t, It indicates the order of placing the i-th raw material in producing the n-th premix at time t;

[0045] C2: Randomly initialize the constructed input chromosome and use the data obtained in C1 to construct the fitness function And constraints, specifically:

[0046] , , ,

[0047] in, Represents the mapping function between raw material feeding rate and stirring rate, represents the penalty term for incorrect stirring direction. When the stirring direction is correct, , when the stirring direction is wrong, , G represents the total number of times the nth premix is ​​stirred at time t; It indicates the uniformity of mixing and stirring of the nth premix produced at time t. Indicates the threshold value of mixing uniformity corresponding to the production of the nth premix; and They represent the minimum stirring rate and the maximum stirring rate when producing the nth premix at time t respectively;

[0048] C3, set the fitness threshold and maximum iteration cycle, and construct the input chromosome and fitness function The constraints are input into the parameter optimization model constructed by the genetic algorithm for iterative training. When the fitness function value is obtained When it is greater than the fitness threshold or reaches the maximum iteration cycle, the iteration is stopped and the parameter optimization model after training is obtained;

[0049] C4. The obtained parameter optimization model is built into the mixing parameter optimization unit, and the mixing and stirring control parameters are initially optimized according to the obtained input raw material properties, input rate and order.

[0050] A control method for automatically mixing materials for premix production, comprising the steps of:

[0051] S1. Obtain historical production data of premixes, obtain the ratio and order of raw materials under corresponding production conditions through a ratio prediction model, and calculate the number of raw material additions and the single amount according to the obtained ratio and order of addition and production demand;

[0052] S2, according to the obtained number of times of raw material delivery, the single delivery amount of each raw material and the delivery order, a delivery control instruction is generated through a collaborative rate control model to control the delivery rate and order of each raw material;

[0053] S3. According to the obtained raw material feeding rate and order and historical mixing control parameters, the initial optimized mixing and stirring control parameters are obtained through genetic algorithm, and the initial optimized mixing and stirring control parameters are input into the fuzzy PID control algorithm to generate mixing control instructions, so as to control the mixing equipment to mix and stir the fed raw materials.

[0054] A computer-readable storage medium stores computer instructions, which, when executed, execute a control method for automatically mixing materials for premix production.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] The present invention aims at the deficiencies of the prior art, and by integrating historical data to predict the optimal raw material ratio and delivery order, effectively responds to changes in the production environment, and makes production more flexible and efficient. The precise calculation capabilities of the feed metering unit and the feed control unit ensure the accuracy of the raw material delivery amount and rate, avoid errors caused by human factors, and greatly reduce the waste and cost of raw materials. Secondly, the introduction of the collaborative rate control model realizes the dynamic optimization and adjustment of the raw material delivery process and the mixing process through delivery reinforcement adjustment and collaborative reinforcement adjustment. Thirdly, the genetic algorithm optimizes the mixing parameters, and combined with the fuzzy PID control algorithm, it realizes the fine regulation of the mixing process, improves the mixing efficiency and mixing effect, and ensures the quality stability of the premix. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 This is a module diagram of a control system for automatic mixing of materials for premix production according to Example 1 of the present invention;

[0058] Figure 2 This is a working structure diagram of corresponding units of a control system for automatic mixing of materials for premix production according to Embodiment 2 of the present invention;

[0059] Figure 3 This is a flow chart of a control method for automatically mixing materials for premix production according to Example 3 of the present invention. DETAILED DESCRIPTION

[0060] Example 1

[0061] See also Figure 1 , an embodiment provided by the present invention: a control system for automatic mixing of materials for premix production, comprising: a raw material supply module, a mixing control module, a quality detection module and an integrated interaction module;

[0062] The raw material supply module controls the supply amount and supply rate of raw materials in the production process by accurately measuring and adjusting the flow rate of raw materials; the raw material supply module includes a ratio prediction unit, a feed metering unit and a feed control unit;

[0063] A ratio prediction unit is used to obtain historical premix production ratio process data, and use the obtained production ratio process data to predict the ratio and placement order of raw materials under corresponding production environment conditions through a constructed ratio prediction model; further, in this embodiment, the historical premix production ratio process data includes raw material formula, raw material ratio, placement order, production demand, raw material type and attribute, production environment and finished product qualified ratio data; further, the production environment data includes temperature, pressure and humidity data; the raw material attribute data includes the particle size, shape, density, fluidity, moisture content and reactivity of the raw material;

[0064] The steps of constructing the ratio prediction model in the ratio prediction unit include:

[0065] A1. Obtain the production ratio process data corresponding to the optimal production process of the historical finished product qualified ratio of N types of premixes, and classify and save the obtained data after preprocessing according to the types of premixes;

[0066] A2. Construct input sequence set based on saved production ratio process data ,in Represents the input sequence corresponding to the nth premix;

[0067] A3. Based on the random forest algorithm, a ratio prediction model including N sub-models is constructed in an integrated manner, and the constructed input sequence set is input into the N sub-models for training in a one-to-one correspondence, and the predicted premix raw material ratio and finished product qualified ratio data are output accordingly;

[0068] A4. Obtain the single sub-model prediction error based on each predicted raw material ratio and finished product qualified ratio data and the actual raw material ratio and finished product qualified ratio data in the input sequence set, and use the single sub-model prediction error to construct a minimum error loss function through a cross entropy function;

[0069] A5. Setting a loss threshold, using the constructed minimum error loss function to train the ratio prediction model, and obtaining a trained ratio prediction model; further, in this embodiment, the loss threshold is 0.1;

[0070] A6. The trained ratio prediction model is built into the ratio prediction unit, and the optimal ratio and order of raw materials are calculated in real time according to the type of premix, production demand, type and attribute data of corresponding raw materials and production environment data;

[0071] This process builds an input sequence set based on historical optimal production data, and integrates N sub-models in combination with the random forest algorithm, effectively capturing the complex relationship between raw material ratio and finished product qualified ratio. The accuracy of model prediction is ensured through training and adjustment of the minimum error loss function. The built-in trained ratio prediction model can automatically calculate the optimal raw material ratio and delivery order in real time based on production demand, raw material type and production environment data, greatly simplifying the production process and reducing human errors.

[0072] The feeding metering unit is used to calculate the number of times the raw materials are put in and the single amount of each raw material according to the raw material formula, the predicted ratio and the production demand; further, in this embodiment, the feeding metering of the raw materials adopts a one-hundred-thousandth balance to carry out the specific metering process;

[0073] The feeding control unit is used to generate a feeding control instruction according to the raw material properties, the feeding metering unit and the output of the ratio prediction unit through the built-in collaborative rate control model, to control the rate and order of each raw material feeding in the production process in real time, and to collaboratively control the ratio of the raw materials fed and the mixing and stirring process of the raw materials; further, the collaborative rate control model in the feeding control unit in this embodiment includes a feeding enhancement adjustment layer and a collaborative enhancement adjustment layer; the feeding enhancement adjustment layer includes m feeding enhancement adjustment sublayers; the specific steps of constructing the feeding enhancement adjustment sublayer include:

[0074] B1. Use the obtained production ratio process data of the nth premix, the single amount of each raw material, the number of times of addition and the order of addition to construct the input state ,in Represents the attribute data of the i-th raw material put into the production of the n-th premix at time t, It represents the single amount of raw material put into the production of n-th premix at time t. It indicates the number of times the i-th raw material is put into production of the n-th premix at time t. It indicates the order of placing the i-th raw material in the production of the n-th premix at time t. It represents the rate of release of the ith raw material for the production of the nth premix at time t, It represents the qualified ratio of the nth premix finished product produced at time t; further, in this embodiment, the qualified ratio of the nth premix finished product is obtained by regular detection by an offline detection unit;

[0075] B2. Set the qualified ratio threshold of the nth premix finished product ,when When the information of adjusting the rate or order of the i-th raw material is triggered, When , the information of the current delivery rate and order of the i-th raw material is maintained;

[0076] B3. Construct a joint execution action based on the trigger information obtained in B2 ,in Indicates the action of adjusting the raw material delivery rate when the adjustment of the i-th raw material delivery rate is triggered. Indicates the action of adjusting the order of raw materials when the adjustment of the order of raw materials is triggered. Indicates the action of maintaining the current delivery rate and order of the i-th raw material when the information of maintaining the current delivery rate or order of the i-th raw material is triggered;

[0077] B4. Construct a reward function based on the joint execution actions performed in B3 , specifically:

[0078] ,

[0079] in, Indicates a negative reward when the premix fails to meet the standards. Indicates that when the premix is ​​qualified, a positive reward is given. It represents the difference between the actual delivery rate of the ith raw material and the target delivery rate for the production of the nth premix at time t. It indicates the reward for producing the nth premix at time t and putting the ith raw material in the right order. , when the delivery order is wrong ; Represents the gain coefficient of the reward function;

[0080] B5. Construct a delivery control sub-model based on reinforcement learning, input the constructed input state, joint execution action and reward function into the delivery control sub-model, and generate delivery control adjustment parameters corresponding to the execution action in B3; further, the reinforcement learning in this embodiment includes SAC algorithm, Q-Learning, REINFORCE and DDPG;

[0081] Furthermore, the collaborative rate control model in the feeding control unit also includes a delivery control layer and a collaborative control layer, and the delivery control layer corresponds to the delivery enhancement adjustment layer; the delivery control layer includes m delivery control sublayers, and the m delivery control sublayers correspond one-to-one to the m delivery enhancement adjustment sublayers; the specific steps of constructing the m delivery control sublayers include:

[0082] B6. According to the obtained and and Initialize the control parameters in the delivery control sublayer, use the delivery control adjustment parameters generated in B5 to generate a control signal for a single delivery of raw materials and the combined control signal of m kinds of raw materials , specifically:

[0083]

[0084]

[0085] in, represents the control signal generated to control the rate or order of adding the i-th raw material, , and Sequentially means that the i-th enhancement adjustment sublayer performs actions in sequence , and The corresponding evaluation value generated when It represents the difference between the qualified ratio of premix finished products and the corresponding threshold value, T represents the time length of a single batch of premix mixed raw materials, and Represents the execution action control factor, when and Execute simultaneously and ;when and When the corresponding action is executed ,when and When only execute ,when and When only execute ; Indicates the i-th The combined control factor of Corresponding When the rate or order of the corresponding i-th raw material is adjusted, of When the corresponding raw material delivery rate and order are not adjusted; , and represents the proportional control parameter, integral control parameter and differential control parameter in the delivery control sublayer in turn; further, in this embodiment, correspond , correspond ; Further, in this embodiment , and Execute the corresponding , and The estimated value of production at the time of

[0086] This process acquires and analyzes production ratio process data in real time, accurately adjusts the rate and order of raw material delivery, thereby significantly improving the qualified ratio of premix finished products, effectively reducing the production of unqualified products, and improving production efficiency and product quality. Secondly, by setting a threshold for the qualified ratio of finished products and triggering corresponding adjustments or maintenance actions, intelligent control of the production process is achieved, human intervention is reduced, and production stability and reliability are improved. In addition, combined with reinforcement learning algorithms such as SAC and Q-Learning, a delivery control sub-model is constructed to further optimize the delivery control strategy, enabling the system to autonomously learn and adapt to changes in the production process, further improving production efficiency and flexibility.

[0087] Furthermore, the collaborative enhancement adjustment layer corresponds to the collaborative control layer; the specific steps of constructing the collaborative enhancement adjustment layer include:

[0088] B7, set The qualified ratio of finished products is obtained by adjusting the raw material delivery rate and sequence through the delivery enhancement adjustment layer and the delivery control layer. When the control signal corresponding to the first collaborative control information is adjusted, the first collaborative control information is triggered, and when the control signal corresponding to the first collaborative control information is adjusted, the first collaborative control information is still satisfied. When the second coordinated control information is triggered;

[0089] B8. According to and Constructing collaborative input state , and construct a collaborative control action based on the first collaborative control information and the second collaborative control information obtained , Indicates the action of collaboratively adjusting the mixing rate of raw materials when the first collaborative control information is triggered. Indicates the action of collaboratively adjusting the ratio of raw materials to be put in when the second collaborative control information is triggered;

[0090] B9. Construct a collaborative control sub-model based on the reinforcement algorithm, and generate corresponding collaborative control adjustment parameters according to the constructed collaborative input state and collaborative control action.

[0091] B10, adjust the parameters and control signals according to the coordinated control obtained in B9 and the synergistic signal generation function to generate the control synergistic control signal for producing the nth premix , the collaborative signal generation function is specifically:

[0092] ,

[0093] in, and Sequentially indicates execution of actions and The corresponding evaluation value is , and represents the coordinated control factor of the feeding adjustment, the coordinated control factor of the stirring adjustment and the coordinated control factor of the raw material ratio adjustment in turn; further, in this embodiment, and To coordinate the execution of sub-models and The generated evaluation value;

[0094] B11. Under the premise of and Corresponding to the raw material input adjustment of the execution process from B1 to B5, when and No adjustment is performed for and The corresponding adjustment of the mixing process is performed; when and Correspondingly execute the coordinated adjustment of the dosing and mixing process;

[0095] B12. When When the ratio of raw materials is adjusted, the action of adjusting the ratio of raw materials is executed, and the process of B1-B11 is repeated, so that the nth premix is ​​always satisfied. condition;

[0096] A mixing control module is used to optimize the mixing control parameters and control the mixing equipment through the optimized control parameters and control strategy to evenly mix the put-in raw materials; the mixing control module includes a mixing parameter optimization unit and a fuzzy control unit;

[0097] The stirring parameter optimization unit is used to obtain the initially optimized mixing and stirring control parameters through a genetic algorithm according to the acquired historical mixing control parameters, effect data, and the rate and order of each raw material addition; further, in this embodiment, the stirring parameter optimization unit obtains the initially optimized mixing and stirring control parameters in the specific steps of:

[0098] C1. Obtain the mixing and stirring control parameters corresponding to the optimal production process of the qualified finished product ratio of N types of premixes and the data on the rate and sequence of each raw material in the feeding process, and use the obtained data to construct the input chromosome corresponding to the genetic algorithm ,in, represents the stirring rate when producing the nth premix at time t, It represents the stirring time of producing the nth premix at time t. Indicates the number of stirring times per unit time. Indicates the stirring direction. It represents the rate of release of the ith raw material for the production of the nth premix at time t, It indicates the order of placing the i-th raw material in producing the n-th premix at time t;

[0099] C2: Randomly initialize the constructed input chromosome and use the data obtained in C1 to construct the fitness function And constraints, specifically:

[0100] , , ,

[0101] in, Represents the mapping function between raw material feeding rate and stirring rate, represents the penalty term for incorrect stirring direction. When the stirring direction is correct, , when the stirring direction is wrong, , G represents the total number of times the nth premix is ​​stirred at time t; It indicates the uniformity of mixing and stirring of the nth premix produced at time t. represents the threshold value of mixing uniformity corresponding to the production of the nth premix; further, in this embodiment, Acquired through online detection unit detection; and represents the minimum stirring rate and the maximum stirring rate when producing the nth premix at time t; further, in this embodiment, The historical feeding rate and stirring rate are obtained by support vector machine fitting, and updated by the real-time feeding rate and stirring rate in each production process;

[0102] C3, set the fitness threshold and maximum iteration cycle, and construct the input chromosome and fitness function The constraints are input into the parameter optimization model constructed by the genetic algorithm for iterative training. When the fitness function value is obtained When it is greater than the fitness threshold or reaches the maximum iteration cycle, the iteration is stopped and the parameter optimization model after training is obtained;

[0103] C4. The obtained parameter optimization model is built into the mixing parameter optimization unit, and the mixing and stirring control parameters are initially optimized according to the obtained input raw material properties, input rate and order.

[0104] This process optimizes the mixing parameters by integrating genetic algorithms, achieving precise control of the mixing process. It constructs fitness functions and constraints, and based on historical optimal production process data, it can automatically adjust parameters such as mixing rate, duration, number of times and direction, while optimizing the rate and order of raw material addition.

[0105] The fuzzy control unit is used to generate mixing control instructions through the fuzzy PID control algorithm according to the initially optimized mixing control parameters. , control the stirring device to mix and stir the put raw materials; further, the specific formula for the fuzzy control unit of this embodiment to generate the fuzzy control instruction includes:

[0106] ,

[0107] express represents the stirring rate when producing the nth premix The difference between the raw material feeding and the raw material mixing and stirring can be further controlled by the difference, so that the raw material feeding and stirring can be kept in the best synchronization state in real time; It represents the directional stirring factor, that is, the ratio of the number of times the stirring direction is correct to the total number of stirring times G when producing the nth premix. , and In turn, they represent the proportional control parameter, integral control parameter and differential control parameter in the mixed control instruction;

[0108] The quality inspection module is used to inspect the produced premixed product online and offline and provide feedback on the inspection results; the quality inspection module includes an online inspection unit and an offline inspection unit;

[0109] The online detection unit monitors the changing trend of the physicochemical parameters of the raw material mixture in the production process in real time through sensors, and obtains the evaluation score of the mixture through an evaluation algorithm based on the obtained changing trend of the physicochemical parameters, and feeds back the obtained changing trend and evaluation score to the feeding control unit and the stirring parameter optimization unit; further, the evaluation algorithm in this embodiment includes a comprehensive fuzzy algorithm and a grey theory algorithm; further, in this embodiment, the mixing uniformity score of the premix raw material mixture is obtained through the evaluation algorithm, and the higher the score, the more uniform the mixture.

[0110] The offline detection unit is used to regularly obtain samples from the premix finished products produced, and obtain the parameter properties and qualified ratio of the finished products through high-precision detection equipment, and feed back the obtained parameter properties and qualified ratio of the finished products to the ratio prediction unit to optimize the ratio of the raw materials of the premix in the same production batch; further, in this embodiment, the qualified ratio of the finished product represents the ratio of the amount of qualified premix produced in a production process of one demand to the amount of the total premix produced; further, in this embodiment, the detection of the offline samples adopts a high performance liquid chromatograph or an infrared spectrophotometer for specific offline detection;

[0111] The integrated interactive module is used to control the deployment and operation of all module units in the system and interactively display the operation results.

[0112] Example 2

[0113] See also Figure 2 Another embodiment provided by the present invention: The workflow of the corresponding unit of the control system for automatic mixing of materials for premix production specifically includes:

[0114] First, through the ratio prediction unit in the integrated interactive module, the ratio and order of raw materials under the corresponding production conditions are obtained according to the historical premix production data, and the obtained ratio and production demand are input into the feeding metering unit to calculate the number of raw material inputs and the single amount of each raw material input; second, the calculated number of raw material inputs, the single amount of each raw material input, and the order of input are input into the feeding control unit to generate a feeding control instruction to control the rate and order of each raw material input in the production process in real time, and coordinately control the ratio of raw materials input by the feeding metering unit and the mixing and stirring process of raw materials in the fuzzy control unit;

[0115] Second, the obtained rate and order of each raw material addition are input into the stirring parameter optimization unit to initially optimize the stirring control parameters of the current mixing production process, and the initially optimized stirring control parameters are input into the fuzzy control unit to generate a mixing control instruction corresponding to the premix raw material mixing and stirring process, and the stirring equipment is controlled to mix and stir the added raw materials. At the same time, during the stirring process, the physical and chemical parameter change trend of the raw material mixture is detected in real time through the online detection unit to obtain the evaluation score of the mixture, and the obtained evaluation score is fed back to the feeding control unit and the stirring parameter optimization unit to optimize and adjust the rate and order of each raw material addition and the mixing and stirring control parameters;

[0116] Third, during the premix production process, the offline detection unit is used to regularly detect the finished product parameter properties and finished product qualified ratio of the same production batch, and the acquired finished product parameter properties are fed back to the ratio prediction unit to optimize and adjust the raw material ratio of the premix of the same production batch;

[0117] Fourth, the changing trends of the physical and chemical parameters of the raw material mixture obtained in real time by the online detection unit and the results of regular detection by the offline detection unit are visualized through the integrated interactive module, and adjustments are made through the integrated interactive module based on the displayed results.

[0118] This process significantly improves the accuracy and efficiency of premix production through a highly integrated intelligent control process. Ratio prediction and real-time feeding control ensure accurate raw material delivery, optimizing production costs and product quality. Dynamic optimization of mixing parameters combined with fuzzy control enhances the mixing effect and achieves uniform product performance. At the same time, the dual-layer monitoring mechanism of online and offline detection not only provides instant feedback on production status, but also optimizes the ratio and mixing parameters through data circulation, ensuring continuous and stable product quality. Visual display provides operators with an intuitive production monitoring interface, which facilitates rapid response and adjustment, further improving the level of intelligent production management.

[0119] Example 3

[0120] See also Figure 3Another embodiment provided by the present invention is a control method for automatically mixing materials for premix production, the steps comprising:

[0121] S1. Obtaining historical production data of premixes, obtaining the ratio and order of raw materials under corresponding production conditions through a ratio prediction model, and calculating the number of raw material additions and the single amount of raw materials according to the obtained ratio and order of addition and production requirements; further, in this embodiment, the corresponding production conditions include temperature and humidity conditions during production and attribute parameters corresponding to each mixed raw material;

[0122] S2, according to the obtained number of times of raw material delivery, the single delivery amount of each raw material and the delivery order, a delivery control instruction is generated through a collaborative rate control model to control the delivery rate and order of each raw material;

[0123] S3, according to the obtained raw material feeding rate and order and historical mixing control parameters, the initial optimized mixing and stirring control parameters are obtained through genetic algorithm, and the initial optimized mixing and stirring control parameters are input into the fuzzy PID control algorithm to generate a mixing control instruction, and the mixing equipment is controlled to mix and stir the fed raw materials;

[0124] S4, monitoring the physical and chemical parameter change trend of the raw material mixture in real time through sensors during S3, and feeding back the acquired physical and chemical parameter change trend to the feeding control unit and the stirring parameter optimization unit, so as to optimize and adjust the feeding rate of each raw material and the mixing and stirring control parameters;

[0125] S5. Regularly sample finished products of the same batch at different time points in the S3 production process, and obtain the corresponding finished product property parameters and finished product qualified ratio, and feed back the obtained finished product property parameters to the ratio prediction unit through the integrated interactive module to optimize and adjust the ratio of premix raw materials in the same production batch.

[0126] Example 4

[0127] A computer-readable storage medium stores computer instructions, which, when executed, execute a control method for automatically mixing materials for premix production.

[0128] An electronic device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements a control method for automatically mixing materials for premix production when executing the computer program.

[0129] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in the field may also change, modify, replace and modify the above-mentioned embodiments without departing from the scope of protection of the purpose of the present invention and the claims, and all of these are within the protection of the present invention.

[0130] If the disclosed technical solution involves personal information, the product using the disclosed technical solution has clearly informed the personal information processing rules and obtained the individual's voluntary consent before processing the personal information. If the disclosed technical solution involves sensitive personal information, the product using the disclosed technical solution has obtained the individual's separate consent before processing the sensitive personal information, and at the same time meets the "explicit consent" requirement. For example, on personal information collection devices such as cameras, clear and prominent signs are set to inform that the personal information collection scope has been entered and personal information will be collected. If the individual voluntarily enters the collection scope, it is deemed that he or she agrees to the collection of his or her personal information; or on the device that processes personal information, the personal information processing rules are notified by obvious signs / information, and the individual's authorization is obtained through pop-up information or by asking the individual to upload his or her personal information; among them, the personal information processing rules may include the personal information processor, the purpose of personal information processing, the processing method, and the type of personal information processed.

Claims

1. A control system for automatic mixing of materials for premix production, characterized in that: include: Raw material supply module, mixing control module; The raw material supply module includes a ratio prediction unit, a feed metering unit and a feed control unit; The ratio prediction unit is used to obtain historical premix production ratio process data, and use the obtained production ratio process data to predict the ratio and placement order of raw materials under corresponding production environment conditions through the constructed ratio prediction model; the feeding metering unit is used to calculate the number of raw material placements and the single placement amount of each raw material according to the raw material formula, predicted ratio and production demand; the feeding control unit is used to generate a placement control instruction according to the raw material properties, the feeding metering unit and the results output by the ratio prediction unit through the built-in collaborative rate control model, so as to control the rate and order of each raw material placement in the production process in real time, and collaboratively control the ratio of the raw materials and the mixing and stirring process of the raw materials; The mixing control module includes a stirring parameter optimization unit and a fuzzy control unit; The mixing parameter optimization unit is used to obtain the initially optimized mixing and stirring control parameters through a genetic algorithm according to the acquired historical mixing control parameters, effect data and the rate and order of each raw material addition; the fuzzy control unit is used to generate a mixing control instruction through a fuzzy PID control algorithm according to the initially optimized mixing and stirring control parameters, and control the mixing equipment to mix and stir the added raw materials; The historical premix production proportioning process data includes raw material formula, raw material proportion, release sequence, production demand, raw material type and property, production environment and finished product qualified ratio data; The steps of constructing the ratio prediction model in the ratio prediction unit include: A1. Obtain the production ratio process data corresponding to the optimal production process of the historical finished product qualified ratio of N types of premixes, and classify and save the obtained data after preprocessing according to the types of premixes; The collaborative rate control model in the feeding control unit includes a delivery enhancement adjustment layer and a collaborative enhancement adjustment layer; the delivery enhancement adjustment layer includes m delivery enhancement adjustment sublayers; the specific steps of constructing the delivery enhancement adjustment sublayer include: B1. Use the obtained production ratio process data of the nth premix, the single amount of each raw material, the number of times of addition and the order of addition to construct the input state ,in Represents the attribute data of the i-th raw material put into the production of the n-th premix at time t, It represents the single amount of raw material put into the production of n-th premix at time t. It indicates the number of times the i-th raw material is put into production of the n-th premix at time t. It indicates the order of placing the i-th raw material in the production of the n-th premix at time t. It represents the rate of release of the ith raw material for the production of the nth premix at time t, It represents the qualified ratio of the nth premix product produced at time t; B2. Set the qualified ratio threshold of the nth premix finished product ,when When the information of adjusting the rate or order of the i-th raw material is triggered, When , the information of the current delivery rate and order of the i-th raw material is triggered.

2. The control system for automatic mixing of materials for premix production according to claim 1, characterized in that: The step of constructing the ratio prediction model in the ratio prediction unit also includes: A2. Construct input sequence set based on saved production ratio process data ,in Represents the input sequence corresponding to the nth premix; A3. Based on the random forest algorithm, a ratio prediction model including N sub-models is constructed in an integrated manner, and the constructed input sequence set is input into the N sub-models for training in a one-to-one correspondence, and the predicted premix raw material ratio and finished product qualified ratio data are output accordingly; A4. Obtain the single sub-model prediction error based on each predicted raw material ratio and finished product qualified ratio data and the actual raw material ratio and finished product qualified ratio data in the input sequence set, and use the single sub-model prediction error to construct a minimum error loss function through a cross entropy function; A5. Set a loss threshold, use the constructed minimum error loss function to train the ratio prediction model, and obtain the trained ratio prediction model.

3. The control system for automatic mixing of materials for premix production according to claim 2, characterized in that: The specific steps of constructing the delivery enhancement adjustment sublayer also include: B3. Construct a joint execution action based on the trigger information obtained in B2 ,in Indicates the action of adjusting the raw material delivery rate when the adjustment of the i-th raw material delivery rate is triggered. Indicates the action of adjusting the order of raw materials when the adjustment of the order of raw materials is triggered. Indicates the action of maintaining the existing delivery rate and order of the i-th raw material when the information of maintaining the existing delivery rate and order of the i-th raw material is triggered; B4. Construct a reward function based on the joint execution actions performed in B3 , specifically: ， in, Indicates a negative reward when the premix fails to meet the standards. Indicates that when the premix is ​​qualified, a positive reward is given. It represents the difference between the actual delivery rate of the ith raw material and the target delivery rate for the production of the nth premix at time t. It indicates the reward for producing the nth premix at time t and putting the ith raw material in the right order. , when the delivery order is wrong ; Represents the gain coefficient of the reward function; B5. Construct a delivery control sub-model based on reinforcement learning, input the constructed input state, joint execution action and reward function into the delivery control sub-model, and generate the delivery control adjustment parameters corresponding to the execution action in B3.

4. The control system for automatic mixing of materials for premix production according to claim 3, characterized in that: The collaborative rate control model in the feeding control unit further includes a delivery control layer and a collaborative control layer, wherein the delivery control layer corresponds to the delivery enhancement adjustment layer; the delivery control layer includes m delivery control sublayers, wherein the m delivery control sublayers correspond one-to-one to the m delivery enhancement adjustment sublayers; and the specific steps of constructing the m delivery control sublayers include: B6. According to the obtained and and Initialize the control parameters in the delivery control sublayer, use the delivery control adjustment parameters generated in B5 to generate a control signal for a single delivery of raw materials and the combined control signal of m kinds of raw materials , specifically: ； ； in, represents the control signal generated to control the rate or order of adding the i-th raw material, , and Sequentially means that the i-th enhancement adjustment sublayer performs actions in sequence , and The corresponding evaluation value generated when It represents the difference between the qualified ratio of premix finished products and the corresponding threshold value, T represents the time length of a single batch of premix mixed raw materials, and Represents the execution action control factor, when and Execute simultaneously and ;when and Execute the corresponding action ,when and When only execute ,when and When only execute ; Indicates the i-th The combined control factor of Corresponding When the rate or order of the corresponding i-th raw material is adjusted, of When the corresponding raw material delivery rate and order are not adjusted, , and They represent the proportional control parameters, integral control parameters and differential control parameters in the delivery control sublayer respectively.

5. The control system for automatic mixing of materials for premix production according to claim 4, characterized in that: The collaborative enhancement adjustment layer corresponds to the collaborative control layer; the specific steps of constructing the collaborative enhancement adjustment layer include: B7, set The qualified ratio of finished products is obtained by adjusting the raw material delivery rate and sequence through the delivery enhancement adjustment layer and the delivery control layer. When the control signal corresponding to the first collaborative control information is adjusted, the first collaborative control information is triggered. When the second coordinated control information is triggered; B8. According to and Constructing collaborative input state , and construct a collaborative control action based on the first collaborative control information and the second collaborative control information obtained , Indicates the action of collaboratively adjusting the mixing rate of raw materials when the first collaborative control information is triggered. Indicates the action of collaboratively adjusting the ratio of raw materials to be put in when the second collaborative control information is triggered; B9. Construct a collaborative control sub-model based on the reinforcement algorithm, and generate corresponding collaborative control adjustment parameters according to the constructed collaborative input state and collaborative control action.

6. The control system for automatic mixing of materials for premix production according to claim 5, characterized in that: The specific steps of constructing the collaborative control layer include: B10, adjust the parameters and control signals according to the coordinated control obtained in B9 and the synergistic signal generation function to generate the control synergistic control signal for producing the nth premix , the collaborative signal generation function is specifically: ， in, and Sequentially indicates execution of actions and The corresponding evaluation value is , and It represents the coordinated control factor of feeding adjustment, the coordinated control factor of stirring adjustment and the coordinated control factor of raw material ratio adjustment in turn; B11. Under the premise of and Corresponding to the raw material input adjustment of the execution process from B1 to B5, when and No adjustment is performed for and The corresponding adjustment of the mixing process is performed; when and Correspondingly execute the coordinated adjustment of the dosing and mixing process; B12. When When the ratio of raw materials is adjusted, the action of adjusting the ratio of raw materials is executed, and the process of B1-B11 is repeated, so that the nth premix is ​​always satisfied. condition.

7. The control system for automatic mixing of materials for premix production according to claim 6, characterized in that: The specific steps of the mixing parameter optimization unit for obtaining the initially optimized mixing and stirring control parameters include: C1. Obtain the mixing and stirring control parameters corresponding to the optimal production process of the qualified finished product ratio of N types of premixes and the data on the rate and sequence of each raw material in the feeding process, and use the obtained data to construct the input chromosome corresponding to the genetic algorithm ,in, represents the stirring rate when producing the nth premix at time t, It represents the stirring time of producing the nth premix at time t. Indicates the number of stirring times per unit time. Indicates the stirring direction. It represents the rate of release of the ith raw material for the production of the nth premix at time t, It indicates the order of placing the i-th raw material in producing the n-th premix at time t; C2: Randomly initialize the constructed input chromosome and use the data obtained in C1 to construct the fitness function And constraints, specifically: ， ， ， in, Represents the mapping function between raw material feeding rate and stirring rate, represents the penalty term for incorrect stirring direction. When the stirring direction is correct, , when the stirring direction is wrong, , G represents the total number of times the nth premix is ​​stirred at time t; It indicates the uniformity of mixing and stirring of the nth premix produced at time t. Indicates the threshold value of mixing uniformity corresponding to the production of the nth premix; and They represent the minimum stirring rate and the maximum stirring rate when producing the nth premix at time t respectively; C3, set the fitness threshold and maximum iteration cycle, and construct the input chromosome and fitness function The constraints are input into the parameter optimization model constructed by the genetic algorithm for iterative training. When the fitness function value is obtained When it is greater than the fitness threshold or reaches the maximum iteration cycle, the iteration is stopped and the parameter optimization model after training is obtained; C4. The obtained parameter optimization model is built into the mixing parameter optimization unit, and the mixing and stirring control parameters are initially optimized according to the obtained input raw material properties, input rate and order.

8. A control method for automatic mixing of materials for premix production, which is implemented based on the control system for automatic mixing of materials for premix production according to any one of claims 1 to 7, characterized in that the steps include: S1. Obtain historical production data of premixes, obtain the ratio and order of raw materials under corresponding production conditions through a ratio prediction model, and calculate the number of raw material additions and the single amount according to the obtained ratio and order of addition and production demand; S2, according to the obtained number of times of raw material delivery, the single delivery amount of each raw material and the delivery order, a delivery control instruction is generated through a collaborative rate control model to control the delivery rate and order of each raw material; S3. According to the obtained raw material feeding rate and order and historical mixing control parameters, the initial optimized mixing and stirring control parameters are obtained through genetic algorithm, and the initial optimized mixing and stirring control parameters are input into the fuzzy PID control algorithm to generate mixing control instructions, so as to control the mixing equipment to mix and stir the fed raw materials.

9. A computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and when the computer instructions are executed, the control method for automatic mixing of materials for premix production as described in claim 8 is executed.

Images