Gelatin Automatic Feeding System and Feeding Method

Abstract

The present invention relates to the field of industrial automation technology, and particularly to a gelatin automatic feeding system and a feeding method optimized based on genetic algorithm and clustering algorithm. The present invention is specifically used for the gelatin and auxiliary materials feeding link in the production process of medicinal capsules, so as to solve the problems of insufficient batching accuracy, complex multi-formula switching, low efficiency of auxiliary materials management and lack of intelligent optimization ability in the prior art.

Description

Technical Field

[0001] The present invention relates to the field of industrial automation technology, and particularly relates to a gelatin automatic feeding system and a feeding method optimized based on genetic algorithm and clustering algorithm. Background Art

[0002] As a common drug carrier, pharmaceutical capsules are widely used because they can effectively protect drug ingredients and improve the patient's taking experience. In the production process of pharmaceutical capsules, the ingredient preparation link of gelatin and excipients is an important step in the whole production process, and its accuracy and efficiency are directly related to the quality and production cost of capsule products.

[0003] Currently, the gelatin ingredient preparation link of pharmaceutical capsules faces the following problems:

[0004] Insufficient formula accuracy: In the production of pharmaceutical capsules, the formula of gelatin must be strictly in accordance with specific ratios. However, in traditional feeding methods, there is more manual participation or the accuracy of semi-automatic equipment is insufficient, which easily leads to proportioning errors, thus affecting key properties such as the solubility and stability of capsules.

[0005] Complex multi-formula switching: The production of pharmaceutical capsules requires frequent switching of formulas according to different specifications and uses (such as adjusting the gelatin concentration, adding pigments or other excipients). Existing equipment takes a long time and wastes a lot when switching formulas, reducing production efficiency.

[0006] Low efficiency in excipient management: The excipients involved in the production of pharmaceutical capsules (such as plasticizers, pigments, flavoring agents, etc.) are numerous and have complex characteristics. Different excipients have different requirements for storage conditions, feeding order, and equipment parameters. Existing systems are difficult to adjust flexibly, increasing the management difficulty.

[0007] Lack of intelligent optimization ability: Traditional feeding systems generally run with fixed programs, lacking the real-time response ability to the production environment (such as raw material characteristics, equipment status), and unable to achieve dynamic optimization through data-driven, resulting in low equipment utilization and resource waste.

[0008] Pharmaceutical capsule production enterprises urgently need a more intelligent and precise automatic feeding system to improve ingredient accuracy, increase production efficiency, and meet the complex multi-formula production requirements. Based on this background, the intelligent feeding optimization technology combining genetic algorithm and clustering algorithm provides a new solution for the gelatin and excipient feeding system of pharmaceutical capsules. By optimizing the formula switching process and excipient grouping management, it can significantly improve the production efficiency and adaptability of the system, thus meeting the high-precision and high-efficiency production requirements in the field of pharmaceutical capsules. Summary of the Invention

[0009] The present invention provides a gelatin automatic feeding system and a feeding method optimized based on genetic algorithm and clustering algorithm, which are specifically used for the gelatin and auxiliary materials feeding link in the production process of medicinal capsules, so as to solve the problems of insufficient batching accuracy, complex multi-formula switching, low efficiency of auxiliary materials management, and lack of intelligent optimization ability in the prior art.

[0010] The technical solution of the gelatin automatic feeding system of the present invention is as follows. The system includes:

[0011] A raw material storage and transportation module, which is used to store gelatin and auxiliary materials and transport them to the feeding device;

[0012] An intelligent metering module, which is used to monitor and control the feeding amount of each raw material in real time;

[0013] An automatic mixing module, which is used to mix different raw materials;

[0014] A control and data management module, which optimizes the feeding management scheme by embedding genetic algorithm and clustering algorithm. Specifically, it includes:

[0015] The genetic algorithm is used to generate and optimize the feeding management scheme for multi-formula switching, so as to minimize the switching time, reduce raw material waste and improve equipment utilization rate;

[0016] The clustering algorithm is used to group the raw material attribute data to optimize the raw material feeding sequence and equipment operation parameters;

[0017] A human-computer interaction module, which is used to set formulas, monitor the feeding process and display the system operation status.

[0018] Preferably, the genetic algorithm evaluates the fitness of the feeding management scheme for multi-formula switching, with the switching time, raw material waste amount and equipment utilization rate as the optimization objectives. The fitness function formula is:

[0019] ;

[0020] Among them, is the switching time, W is the raw material waste amount, U is the equipment utilization rate, , , are weight coefficients.

[0021] Preferably, the clustering algorithm generates a feeding grouping management scheme through clustering analysis of raw material attribute data (including density, humidity, viscosity). The distance calculation formula is:

[0022] ;

[0023] Among them, is the raw material attribute feature vector, is the clustering center, m is the attribute dimension, represents the eigenvalue of the i-th sample point in the k-th dimension; represents the clustering center in the k-th dimension.

[0024] Preferably, when the genetic algorithm generates the feeding management plan, the formula switching plan is optimized through the crossover operation, and the crossover formula is:

[0025] ;

[0026] ;

[0027] where , is the parent feeding management plan, , is the offspring feeding management plan, is the crossover point; represents the slice Slice from index 0 to k−1, that is, the first k elements in the parent plan; represents the slice from index k to the end, that is, all elements from the k-th element to the end in the parent plan.

[0028] Preferably, the mutation operation of the genetic algorithm is optimized by randomly adjusting some parameters in the feeding management plan, and its mutation formula is:

[0029] ;

[0030] where is the j-th parameter in the feeding management plan, and V is the value after mutation.

[0031] Preferably, the feeding grouping management plan generated by the clustering algorithm is optimized by updating the clustering center, and its update formula is:

[0032] ;

[0033] where is the new clustering center of the cluster , is the set of sample points of the cluster , is the set of sample points of the cluster and

[0034] Preferably, the control and data management module automatically adjusts the feeding operation of the equipment according to the optimal feeding management plan generated by the genetic algorithm, and reduces the frequency of equipment parameter adjustment in combination with the grouping results of the clustering algorithm.

[0035] Preferably, the human-machine interaction module supports remote monitoring and control, and realizes real-time data transmission and remote operation through the industrial Internet interface.

[0036] The feeding method of the gelatin automatic feeding system of the present technical solution includes the following steps:

[0037] A. Raw material attribute collection: Collect attribute data of gelatin and auxiliary materials used for the production of pharmaceutical capsules, including but not limited to density, humidity, and viscosity;

[0038] B. Raw material grouping: Analyze the raw material attribute data through a clustering algorithm to generate a raw material grouping scheme, optimize the feeding order of raw materials and equipment parameters, where the clustering calculation formula is:

[0039] ;

[0040] Among them, is the raw material attribute feature vector, is the clustering center, m is the attribute dimension, represents the eigenvalue of the i-th sample point in the k-th dimension; represents the clustering center eigenvalue in the k-th dimension;

[0041] C. Generation of feeding management plan: Use a genetic algorithm combined with the raw material grouping scheme to generate an optimal feeding management plan for multi-formula switching, specifically including:

[0042] Set the fitness function, with the switching time, raw material waste, and equipment utilization rate as the optimization objectives, and the formula is:

[0043] ;

[0044] Among them, is the switching time, W is the raw material waste, U is the equipment utilization rate, , , are the weight coefficients;

[0045] Generate a feeding management plan through the crossover and mutation operations of the genetic algorithm, and the crossover formula is:

[0046] ;

[0047] ;

[0048] Among them, , is the parental feeding management plan, , It is a feeding management solution for offspring. It is an intersection point. It represents a slice from index 0 to k−1, that is, the first k elements in the parent solution. It represents a slice from index k to the end, that is, all elements from the k-th element to the end in the parent solution.

[0049] D. Precise feeding: According to the optimal feeding management solution, control the operating parameters of the intelligent metering module and the automatic mixing module, and precisely feed gelatin and auxiliary materials according to the formula requirements.

[0050] E. Multi-formula switching: After completing the feeding of the current formula, adjust the equipment parameters according to the optimal feeding management solution to complete the efficient switching of multiple formulas.

[0051] F. Data recording and feedback: Record the real-time data during the feeding process, including feeding accuracy, switching efficiency, and equipment utilization rate, and dynamically optimize the feeding management solution through the genetic algorithm.

[0052] Preferably, the mutation operation of the genetic algorithm generates a new solution by randomly adjusting the parameter values in the feeding management solution, and the mutation formula is:

[0053] ;

[0054] Among them, is the j-th parameter in the feeding management solution, and V is the mutated value.

[0055] The present invention has achieved the following remarkable technical effects in the production process of pharmaceutical capsules through a gelatin automatic feeding system optimized based on the genetic algorithm and the clustering algorithm:

[0056] 1. The accuracy of ingredient preparation has been significantly improved.

[0057] The intelligent metering module combines with the genetic algorithm to dynamically optimize the feeding management solution, so that the ingredient error of gelatin and auxiliary materials is controlled within ±0.5%, ensuring the formula accuracy and product consistency in the production process of pharmaceutical capsules. It reduces the quality problems caused by proportioning errors and improves the solubility, stability, and uniformity of the products.

[0058] 2. The efficiency of multi-formula switching has been greatly improved.

[0059] The genetic algorithm optimizes the multi-formula switching process. By dynamically adjusting the switching sequence and equipment parameters, the switching time is shortened by about 30%, and the raw material waste is reduced by about 20%. The system supports rapid response to the production requirements of multiple formulas and meets the efficient production scenarios of multiple varieties and small batches in the production of pharmaceutical capsules.

[0060] 3. The management of auxiliary materials is intelligent.

[0061] The clustering algorithm optimizes the grouping of raw material attributes (such as density, humidity, viscosity, etc.), making the raw material grouping more reasonable, and reducing the number of equipment adjustments by about 25%. It realizes the flexible allocation and management of different auxiliary materials, avoids equipment failures and operation mistakes caused by differences in raw material characteristics, and improves the stability of system operation.

[0062] 4. Advantages of the cooperation of the two algorithms

[0063] The global optimization ability of the genetic algorithm: It can quickly generate and optimize the feeding management plan for multi-formula switching, and find the switching path with the highest overall production efficiency.

[0064] The characteristic data processing ability of the clustering algorithm: Through the accurate analysis of raw material attributes, it realizes reasonable grouping and provides high-quality input data for the genetic algorithm.

[0065] The collaborative effect of the algorithms: The clustering algorithm provides a preliminary optimization basis for the genetic algorithm, and further adjusts the switching plan through the genetic algorithm, forming a closed-loop process from data grouping to global optimization.

[0066] This collaborative optimization method significantly improves the operation efficiency and adaptive ability of the system, and provides an efficient solution for complex feeding scenarios in the production of medicinal capsules.

[0067] 5. Significantly improved system operation efficiency

[0068] Through the combination of the genetic algorithm and the clustering algorithm, the system realizes real-time dynamic optimization of production conditions, and the equipment utilization rate is increased by about 15%-20%. The continuity and stability of system operation are greatly improved, the downtime is reduced, and the production efficiency is significantly improved.

[0069] 6. Improvement of the intelligent and automated level of the production process

[0070] The intelligent generation and real-time adjustment of the feeding management plan reduce manual intervention and improve the adaptive ability of the system. It is equipped with a man-machine interaction module, supports remote monitoring and operation, realizes data visualization management, and is convenient for the control and traceability of the production process.

[0071] 7. Reduction of production costs

[0072] By reducing raw material waste, shortening the switching time and improving equipment utilization rate, the raw material loss and energy consumption in the production process are significantly reduced, and the comprehensive cost is reduced by about 10%-15%.

[0073] 8. Enhancement of market competitiveness

[0074] The system optimizes the production process of pharmaceutical capsules through precise batching and efficient management, providing high-quality and high-efficiency production guarantee for enterprises in the drug carrier market. The integration and empowerment of the two algorithms enable the system to reach the leading level in the industry in terms of intelligence and economy, enhancing the market competitiveness and brand value of the enterprise.

[0075] The present invention not only solves the problems of inaccurate feeding, complex multi-formula switching, and inefficient auxiliary material management in the production of pharmaceutical capsules, but also realizes the intelligent optimization of the whole process through the synergistic effect of genetic algorithm and clustering algorithm, providing an efficient, accurate and intelligent solution for the industry. Brief Description of the Drawings

[0076] Figure 1 is the structural block diagram of the present invention;

[0077] Figure 2 is the step diagram of the operation method of the present invention. Detailed Description of the Invention

[0078] Embodiment 1:

[0079] The technical solution of the gelatin automatic feeding system in this embodiment is as follows. The system includes:

[0080] The raw material storage and conveying module is used to store gelatin and auxiliary materials and convey them to the feeding device;

[0081] The intelligent metering module is used to monitor and control the feeding amount of each raw material in real time;

[0082] The automatic mixing module is used to mix different raw materials;

[0083] The control and data management module optimizes the feeding management scheme by embedding genetic algorithm and clustering algorithm, specifically including:

[0084] The genetic algorithm is used to generate and optimize the feeding management scheme for multi-formula switching to minimize the switching time, reduce raw material waste and improve equipment utilization rate;

[0085] The clustering algorithm is used to group the raw material attribute data to optimize the raw material feeding order and equipment operation parameters;

[0086] The human-computer interaction module is used to set formulas, monitor the feeding process and display the system operation status.

[0087] Embodiment 2:

[0088] The difference between this embodiment and Embodiment 1 is that this embodiment further includes:

[0089] The genetic algorithm evaluates the fitness of the feeding management scheme for multi-formula switching, with the switching time, raw material waste, and equipment utilization rate as the optimization objectives. The fitness function formula is as follows:

[0090] ;

[0091] Among them, is the switching time, W is the raw material waste, U is the equipment utilization rate, , , are the weight coefficients.

[0092] Preferably, the clustering algorithm generates a feeding grouping management scheme by performing clustering analysis on the raw material attribute data (including density, humidity, and viscosity). The distance calculation formula is as follows:

[0093] ;

[0094] Among them, is the raw material attribute feature vector, is the clustering center, m is the attribute dimension, represents the eigenvalue of the i-th sample point in the k-th dimension; represents the clustering center 's eigenvalue in the k-th dimension.

[0095] Preferably, when the genetic algorithm generates a feeding management scheme, it optimizes the formula switching scheme through crossover operation. The crossover formula is as follows:

[0096] ;

[0097] ;

[0098] Among them, , are the parental feeding management schemes, , are the offspring feeding management schemes, is the crossover point; represents the slice Slice from index 0 to k−1, that is, the first k elements in the parental scheme; represents the slice from index k to the end, that is, all elements from the k-th element to the end in the parental scheme.

[0099] Preferably, the mutation operation of the genetic algorithm is optimized by randomly adjusting some parameters in the feeding management scheme. The mutation formula is as follows:

[0100] ;

[0101] Among them, is the j-th parameter in the feeding management plan, and V is the value after variation.

[0102] Preferably, the feeding grouping management plan generated by the clustering algorithm is optimized by updating the clustering center, and its update formula is:

[0103] ;

[0104] Among them, is the new clustering center of the cluster , is the set of sample points of the cluster , is the number of sample points of the cluster .

[0105] Preferably, the control and data management module automatically adjusts the feeding operation of the equipment according to the optimal feeding management plan generated by the genetic algorithm, and reduces the frequency of equipment parameter adjustment in combination with the grouping result of the clustering algorithm.

[0106] Preferably, the human-computer interaction module supports remote monitoring and control, and realizes real-time data transmission and remote operation through the industrial Internet interface.

[0107] Example 3:

[0108] The feeding method of a gelatin automatic feeding system in this example includes the following steps:

[0109] A. Raw material attribute collection: Collect the attribute data of gelatin and excipients used for the production of medicinal capsules, including but not limited to density, humidity, and viscosity;

[0110] B. Raw material grouping: Analyze the raw material attribute data through the clustering algorithm to generate a raw material grouping plan, and optimize the feeding order and equipment parameters of the raw materials. The clustering calculation formula is:

[0111] ;

[0112] Among them, is the raw material attribute feature vector, is the clustering center, m is the attribute dimension, represents the eigenvalue of the i-th sample point on the k-th dimension; represents the eigenvalue of the clustering center on the k-th dimension;

[0113] C. Generation of feeding management plan: Use the genetic algorithm in combination with the raw material grouping plan to generate an optimal feeding management plan for multi-formula switching, specifically including:

[0114] Set the fitness function, with the switching time, raw material waste, and equipment utilization rate as the optimization objectives. The formula is:

[0115] ;

[0116] Among them, is the switching time, W is the raw material waste, U is the equipment utilization rate, , , are the weight coefficients;

[0117] Generate the feeding management plan through the crossover and mutation operations of the genetic algorithm. The crossover formula is:

[0118] ;

[0119] ;

[0120] Among them, , are the parent feeding management plans, , are the offspring feeding management plans, is the crossover point; represents the slice Slice from index 0 to k−1, that is, the first k elements in the parent plan; represents the slice from index k to the end, that is, all elements from the kth element to the end in the parent plan;

[0121] D. Precise feeding: According to the optimal feeding management plan, control the operating parameters of the intelligent metering module and the automatic mixing module, and accurately put in gelatin and auxiliary materials according to the formula requirements;

[0122] E. Multi-formula switching: After completing the feeding of the current formula, adjust the equipment parameters according to the optimal feeding management plan to complete the efficient switching of multiple formulas;

[0123] F. Data recording and feedback: Record the real-time data during the feeding process, including feeding accuracy, switching efficiency, and equipment utilization rate, and dynamically optimize the feeding management plan through the genetic algorithm.

[0124] The mutation operation of the genetic algorithm generates a new plan by randomly adjusting the parameter values in the feeding management plan. The mutation formula is:

[0125] ;

[0126] Among them, is the jth parameter in the feeding management plan, and V is the mutated value

[0127] Example 4:

[0128] This example is a specific application example of Example 3, as follows:

[0129] Example background

[0130] A pharmaceutical capsule manufacturing enterprise needs to produce two capsule specifications:

[0131] Formula A: High-concentration gelatin (90%) and low-concentration glycerol (10%).

[0132] Formula B: Medium-concentration gelatin (70%), glycerol (20%), and pigment (10%).

[0133] Production requirements:

[0134] 1. Produce 100 kg per batch, and the proportioning error of each raw material is required to be controlled within ±0.5%.

[0135] 2. Switch the formula once per hour, and the switching time should be as short as possible to reduce raw material waste.

[0136] 3. Raw material properties: Gelatin humidity is 12% - 15%, glycerol has high viscosity, and pigment has low density. Equipment parameters need to be adjusted dynamically.

[0137] 1. Collection of raw material properties

[0138] Collect the property data of raw materials through sensors, including:

[0139] Gelatin: Humidity 12.5%, density 1.4 g / cm³, viscosity 0.5 Pa·s;

[0140] Glycerol: Humidity 0%, density 1.26 g / cm³, viscosity 0.89 Pa·s;

[0141] Pigment: Humidity 1%, density 1.2 g / cm³, viscosity 0.1 Pa·s.

[0142] 2. Grouping of raw materials

[0143] Use the clustering algorithm to group the raw material property data, and the steps are as follows:

[0144] (1) Initialize the clustering centers

[0145] Select K = 3 (3 categories), and randomly initialize 3 clustering centers:

[0146] : Humidity 10%, density 1.3 g / cm³, viscosity 0.3 Pa·s;

[0147] : Humidity 0%, density 1.25 g / cm³, viscosity 0.8 Pa·s;

[0148] : Humidity 1%, density 1.2 g / cm³, viscosity 0.1 Pa·s.

[0149] (2) Calculate the distance from the sample to the cluster center

[0150] Use the formula:

[0151] ;

[0152] Calculate the distances from gelatin, glycerol, and pigment to , , respectively.

[0153] Calculation results:

[0154] · Gelatin: Classified into ;

[0155] · Glycerol: Classified into ;

[0156] · Pigment: Classified into .

[0157] (3) Update the cluster center

[0158] Calculate the new cluster center for each group and finally generate the grouping scheme:

[0159] · Group 1 (humidity-sensitive group): Gelatin;

[0160] · Group 2 (high-viscosity group): Glycerol;

[0161] · Group 3 (low-density group): Pigment.

[0162] According to the grouping results, the device dynamically adjusts the parameters:

[0163] · Gelatin: Feeding speed 6 kg / min;

[0164] · Glycerol: Feeding speed 8 kg / min;

[0165] · Pigment: Feeding speed 10 kg / min.

[0166] 3. Generation of the feeding management plan

[0167] Use the genetic algorithm to generate the optimal feeding management plan for multi-formula switching. The steps are as follows:

[0168] (1) Set the fitness function

[0169] Optimization objective:

[0170] ;

[0171] Among them:

[0172] T = 15 minutes (switching time);

[0173] W = 3 kg (raw material waste);

[0174] U = 85% (equipment utilization rate).

[0175] (2) Crossover operation

[0176] Randomly select two parent solutions:

[0177] Parent 1: Recipe switching order A → B;

[0178] Parent 2: Recipe switching order B → A.

[0179] Generate child solutions:

[0180] Child 1: A → B;

[0181] Child 2: B → A.

[0182] 4. Precise feeding

[0183] Execute precise feeding according to the optimal solution:

[0184] Recipe A: Feed 90 kg of gelatin and 10 kg of glycerol;

[0185] Recipe B: Feed 70 kg of gelatin, 20 kg of glycerol and 10 kg of pigment.

[0186] Real-time monitor the feeding accuracy through the intelligent metering module, and control the error within ±0.4%.

[0187] 5. Multi-recipe switching

[0188] Complete recipe switching according to the optimal solution:

[0189] Dynamically adjust the equipment parameters;

[0190] Reduce the cleaning time from 12 minutes to 8 minutes;

[0191] Reduce the switching time from 15 minutes to 10 minutes;

[0192] Reduce the raw material waste from 3 kg to 1 kg.

[0193] 6. Data recording and feedback

[0194] Real-time recording:

[0195] Feeding accuracy: error ±0.4%;

[0196] Switching time: 10 minutes;

[0197] Equipment utilization rate: 90%.

[0198] The system dynamically optimizes the fitness function by recording data to support the optimization of the next generation.

[0199] If there is mutation:

[0200] Mutation operation

[0201] (1) Trigger mutation

[0202] Perform mutation operation on the offspring solution with a probability of 10%:

[0203] Original solution: A → B;

[0204] Mutated solution: B → A.

[0205] (2) Adjust parameters

[0206] Perform mutation operation on the feeding speed:

[0207] Original value: Glycerol feeding speed 8 kg / min;

[0208] After mutation: Glycerol feeding speed 9 kg / min.

[0209] (3) Evaluate and retain

[0210] Evaluate the mutated solution through the fitness function:

[0211] If the fitness of the mutated solution is better than the current solution, retain it.

[0212] Example effect:

[0213] Feeding accuracy: Feeding error controlled within ±0.5%;

[0214] Improved switching efficiency: Switching time reduced by 33%, raw material waste reduced by 66%;

[0215] Improved equipment utilization rate: Through mutation optimization, the equipment utilization rate reaches 92%.

Claims

1. A gelatin automatic feeding system, characterized in that: The system includes: A raw material storage and conveying module is used to store gelatin and auxiliary materials and convey them to the feeding device; Intelligent metering module, used to monitor and control the amount of each raw material in real time; Automatic mixing module, used to mix different raw materials; The control and data management module optimizes the feeding management plan by embedding genetic algorithms and clustering algorithms, including: The genetic algorithm is used to generate and optimize the feeding management plan for multi-recipe switching to minimize the switching time, reduce the waste of raw materials and improve the utilization rate of equipment. The genetic algorithm evaluates the fitness of the feeding management plan for multi-recipe switching, taking the switching time, the waste of raw materials and the utilization rate of equipment as the optimization targets. The fitness function formula is: ； Among them, T is the switching time, W is the amount of raw material waste, U is the equipment utilization rate, w1, w2, w3 are weight coefficients; The clustering algorithm is used to group the raw material attribute data to optimize the raw material feeding sequence and equipment operating parameters; the clustering algorithm generates a feeding grouping management plan by clustering the raw material attribute data. The clustering algorithm calculation formula is: ； Among them, x i is the raw material attribute feature vector, c j is the cluster center, m is the attribute dimension, x ik represents the eigenvalue of the i-th sample point in the k-th dimension; c jk Represents the cluster center c j The eigenvalue in the kth dimension; The human-computer interaction module is used to set the recipe, monitor the feeding process and display the system operation status.

2. The automatic gelatin feeding system according to claim 1, characterized in that: When the genetic algorithm generates a feeding management plan, the recipe switching plan is optimized through a crossover operation, and the crossover formula is: ； ； Among them, P1 and P2 are the parent generation feeding management plans, C1 and C2 are the child generation feeding management plans, and k is the intersection point; [:k] represents the slice from index 0 to k-1, that is, the first k elements in the parent generation plan; [k:] represents the slice from index k to the end, that is, all elements from the kth element to the end in the parent generation plan.

3. The automatic gelatin feeding system according to claim 1, characterized in that: The mutation operation of the genetic algorithm is optimized by randomly adjusting some parameters in the feeding management scheme, and its mutation formula is: C i '[j]=V; Among them, C i '[j] is the jth parameter in the feeding management plan, and V is the value after mutation.

4. The automatic gelatin feeding system according to claim 3, characterized in that: The feeding grouping management scheme generated by the clustering algorithm is optimized by updating the clustering center, and the updating formula is: ； Among them, c j is the new cluster center of cluster j, S j is the set of sample points of cluster j, n j is the number of sample points in cluster j.

5. The automatic gelatin feeding system according to claim 1, characterized in that: The control and data management module automatically adjusts the feeding operation of the equipment according to the optimal feeding management plan generated by the genetic algorithm, and reduces the frequency of equipment parameter adjustment in combination with the grouping result of the clustering algorithm.

6. The automatic gelatin feeding system according to claim 1, characterized in that: The human-computer interaction module supports remote monitoring and control, and realizes real-time data transmission and remote operation through the industrial Internet interface.

7. A feeding method for the gelatin automatic feeding system according to any one of claims 1 to 6, characterized in that: The following steps are involved: A. Raw material attribute collection: Collect attribute data of gelatin and excipients used in the production of pharmaceutical capsules, including but not limited to density, humidity and viscosity; B. Raw material grouping: Analyze the raw material attribute data through clustering algorithm, generate raw material grouping plan, optimize the raw material feeding sequence and equipment parameters, and the clustering calculation formula is: ； Among them, x i is the raw material attribute feature vector, c j is the cluster center, m is the attribute dimension, x ik represents the eigenvalue of the i-th sample point in the k-th dimension; c jk Represents the cluster center c j The eigenvalue in the kth dimension; C. Feeding management plan generation: Generate the optimal feeding management plan for multi-recipe switching by combining the raw material grouping plan with the genetic algorithm, which specifically includes: The fitness function is set to optimize the switching time, raw material waste and equipment utilization rate. The formula is: ； Among them, T is the switching time, W is the amount of raw material waste, U is the equipment utilization rate, w1, w2, w3 are weight coefficients; The feeding management plan is generated through the crossover and mutation operations of the genetic algorithm. The crossover formula is: ； ； Among them, P1 and P2 are the parent generation feeding management plans, C1 and C2 are the child generation feeding management plans, and k is the intersection point; [:k] represents the slice from index 0 to k-1, that is, the first k elements in the parent generation plan; [k:] represents the slice from index k to the end, that is, all elements from the kth element to the end in the parent generation plan; D. Accurate feeding: According to the optimal feeding management plan, control the operating parameters of the intelligent metering module and the automatic mixing module, and accurately feed gelatin and auxiliary materials according to the formula requirements; E. Multi-recipe switching: After completing the feeding of the current recipe, adjust the equipment parameters according to the optimal feeding management plan to complete the efficient switching of multiple recipes; F. Data recording and feedback: Record real-time data during the feeding process, including feeding accuracy, switching efficiency and equipment utilization, and dynamically optimize the feeding management plan through genetic algorithms.

8. The feeding method of the gelatin automatic feeding system according to claim 7, characterized in that: The mutation operation of the genetic algorithm generates a new scheme by randomly adjusting the parameter values ​​in the feeding management scheme. The mutation formula is: C i '[j]=V; Among them, C i '[j] is the jth parameter in the feeding management plan, and V is the value after mutation.

Images