An intelligent optimization method for the production and processing technology of sleep patches

Through intelligent optimization methods, including adaptive formula and continuous microfluidic mixing technologies, the problems of low quality consistency, efficiency and resource utilization in traditional sleep patch production are solved, and an efficient and flexible production process is achieved to adapt to market demand. change.

CN118707912BActive Publication Date: 2024-12-03LOOBI GUANGZHOU HEALTH IND CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411194153.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-28
Publication Date
2024-12-03
Estimated Expiration
2044-08-28

AI Technical Summary

Technical Problem

The traditional sleep patch production process has problems such as difficulty in ensuring product quality consistency, low production efficiency, energy consumption and waste of raw materials, and lacks flexibility, making it difficult to respond quickly to changes in market demand.

Method used

Using intelligent optimization methods for sleep patch production and processing technology, including obtaining raw material databases and target product parameters, establishing a digital twin model, determining the optimal ratio through adaptive formula optimization algorithm, and using continuous microfluidic mixing, precision coating and intelligent drying systems , combined with online cutting and intelligent quality control systems, realize full-process monitoring and dynamic task allocation.

Benefits of technology

It significantly improves product quality consistency, improves production efficiency, reduces energy consumption and raw material waste, realizes real-time quality control and flexibility in production processes, and can quickly adapt to formula and raw material changes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118707912B_ABST
    Figure CN118707912B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of process intelligent optimization technology, and more specifically, to an intelligent optimization method for the production and processing technology of a sleep patch, comprising the following steps: obtaining a sleep patch raw material database, a target product parameter set, and quality constraints; establishing a digital twin model of a sleep patch product; determining the optimal ratio through an adaptive formula optimization algorithm based on the raw material database, the target product parameter set, and the quality constraints, to obtain the sleep patch raw material formula; using a continuous microfluidic mixing system to mix the raw materials to form a uniform sleep patch drug mixture; obtaining a sleep patch base film with uniform thickness; obtaining a sleep patch semi-finished product with a water content that meets the requirements; forming a sleep patch with uniform specifications; ensuring the quality stability of the sleep patch; improving the production efficiency of the sleep patch; and continuously improving the sleep patch production process, and through intelligent control of the entire process, reducing the coefficient of variation of the product's active ingredient content to 2.1%, which is far superior to the existing technology level.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of process intelligent optimization, and more specifically, to a method for intelligent optimization of the production and processing technology of sleep patches. Background Art

[0002] As people pay more and more attention to sleep health, sleep patches, as a new type of sleep aid product, have been widely welcomed in the market in recent years. However, the production process of sleep patches faces many challenges. The traditional production and processing technology of sleep patches mainly relies on batch production methods, which has become inadequate in the current rapidly changing market environment.

[0003] At present, there are several main problems in the production of sleep patches:

[0004] First, it is difficult to ensure the consistency of product quality. Due to the characteristics of batch production, the product quality between different batches fluctuates greatly. For example, studies have shown that the coefficient of variation of the active ingredient content of sleep patches produced by traditional methods can be as high as 8% or more, which seriously affects the stability of the product's efficacy.

[0005] Secondly, the production efficiency is low. The traditional production method involves multiple independent processes, such as mixing, coating, drying, cutting, etc. There is often a waiting time between each process, resulting in a lengthy production cycle. According to statistics, it usually takes 7 to 8 hours to produce 1,000 sleep patches. This efficiency is difficult to meet the rapid changes in market demand.

[0006] Third, energy consumption and raw material waste are prominent issues. In traditional production methods, processes such as drying often use fixed parameters, which are difficult to dynamically adjust according to actual conditions, resulting in low energy efficiency. At the same time, due to the lack of precise control, the utilization rate of raw materials is generally not high, and usually only about 90% of the raw materials end up in the product.

[0007] Fourth, quality control lags behind. Existing quality control mainly relies on random inspections after the product is completed, which makes it difficult to find and solve problems in the production process in a timely manner. This not only increases the defective product rate, but also prolongs the production cycle.

[0008] In addition, the existing production methods lack flexibility. When the formula needs to be adjusted or the raw materials need to be replaced, it often takes a long process of process exploration, making it difficult to quickly respond to changes in market demand.

[0009] In response to the above problems, the industry has also made some attempts to improve. For example, some researchers have proposed a method of introducing intelligent control in a single link, such as using an intelligent mixing system. However, this local optimization method has limited effect. Taking the practice of a well-known pharmaceutical company as an example, they introduced an intelligent control system in the mixing link. Although it reduced the coefficient of variation of product quality from 8.7% to 7.2%, it is still far from the ideal level. At the same time, since other links still use traditional methods, the improvement of overall production efficiency is also very limited.

[0010] Therefore, there is an urgent need for a method that can comprehensively improve the production and processing technology of sleep patches to cope with the various challenges currently faced. Summary of the invention

[0011] The present invention is proposed in this context, and aims to address the limitations of the prior art and provide an intelligent optimization method for the production and processing technology of sleep patches.

[0012] The present invention provides a method for intelligently optimizing the production and processing technology of a sleep patch, comprising the following steps:

[0013] Obtaining a database of raw materials for a sleep patch, a target product parameter set, and quality constraints; establishing a digital twin model of a sleep patch product; the method also includes:

[0014] According to the raw material database, target product parameter set and quality constraints, the optimal ratio is determined by an adaptive formula optimization algorithm to obtain the raw material formula of the sleep patch; based on the optimal ratio, a continuous microfluidic mixing system is used to mix the raw materials to form a uniform sleep patch drug mixture; the mixed raw materials are coated into a film using a precision coating system to obtain a sleep patch base film with uniform thickness; the coated film product is subjected to intelligent drying treatment to obtain a sleep patch semi-finished product with a moisture content that meets the requirements; an online cutting system is used to accurately cut the dried product to form a sleep patch with uniform specifications; an intelligent quality control system is used to monitor the entire product process and detect anomalies to ensure the quality stability of the sleep patch; dynamic task allocation and production line optimization are achieved based on a production scheduling optimization algorithm to improve the production efficiency of the sleep patch; the optimized production parameters are fed back to the digital twin model to continuously improve the sleep patch production process.

[0015] Specifically, the adaptive recipe optimization algorithm includes:

[0016] Establishing a comprehensive evaluation index that takes into account the raw material cost, the target product parameters and the quality constraints;

[0017] Assigning importance weights to each raw material in the sleep patch raw material database, and dynamically adjusting the weights according to their influence on the performance of the sleep patch;

[0018] By means of an iterative optimization method, the ratio of each raw material is gradually adjusted until the optimal value of the evaluation index is reached;

[0019] During the optimization process, the interaction between the raw materials is considered in real time to ensure that the synergistic effect of the raw material formula of the sleep patch is maximized;

[0020] The optimized formula is compared with the target product parameter set, and if the deviation exceeds a preset threshold, it is optimized again.

[0021] Specifically, the optimization of the continuous microfluidic mixing system includes:

[0022] Design a multi-stage microchannel structure in which the channel width decays exponentially along the flow direction and is superimposed with sinusoidal fluctuations;

[0023] The pressure drop and temperature change during the raw material mixing process are monitored in real time by a pressure sensor and a thermocouple;

[0024] Based on monitoring data, adaptive control algorithms are used to dynamically adjust flow rate and temperature to maximize mixing efficiency;

[0025] During the mixing process, the uniformity of the drug mixture of the sleep patch is evaluated in real time by a fluorescent tracer method;

[0026] When the mixing uniformity reaches a preset threshold, the next production stage is automatically entered.

[0027] Specifically, the optimization of the precision coating system includes:

[0028] A coating kinetics model is established to express the thickness of the base film of the sleep patch as a function of flow rate, substrate speed, viscosity, surface tension and contact angle;

[0029] A multi-objective optimization algorithm is used to simultaneously consider coating uniformity, surface tension gradient, and energy consumption;

[0030] Using a precision slot die driven by piezoelectric ceramics to achieve sub-micron thickness adjustment;

[0031] The thickness of the base film of the sleeping patch is measured in real time by a laser triangulation rangefinder to form a closed-loop control system;

[0032] Based on the measurement results, a fuzzy PID controller is used to dynamically adjust the coating parameters to ensure the uniformity of the base film of the sleeping patch.

[0033] Specifically, the intelligent drying process includes:

[0034] Establish a multi-physics coupled drying model, taking into account moisture diffusion, heat conduction and convection heat transfer;

[0035] Using distributed infrared moisture meters and temperature sensors, a real-time monitoring network for the moisture content and temperature distribution of the semi-finished sleeping patch is constructed;

[0036] Based on monitoring data, an adaptive multi-stage drying optimization algorithm is used to dynamically adjust the drying temperature and air flow speed;

[0037] Control the heating element power through pulse width modulation technology to achieve precise temperature control;

[0038] During the drying process, the quality indicators of the semi-finished sleeping patch, such as uniformity of moisture content and thermal stress, are continuously evaluated to ensure its quality stability.

[0039] Specifically, the optimization of the online cutting system includes:

[0040] Use high-speed CMOS cameras to collect images in real time, and combine deep learning algorithms for image processing and feature extraction;

[0041] Establish a visual positioning model to map image features to actual cropping coordinates;

[0042] Develop an adaptive cutting trajectory planning algorithm that comprehensively considers cutting accuracy, speed, and energy consumption;

[0043] The precision platform driven by piezoelectric ceramics is used to control the position of the cutting knife, achieving micron-level positioning accuracy.

[0044] Dynamically adjusting cutting energy through an ultrasonic generator to optimize edge smoothness and precision of the sleeping patch with uniform specifications;

[0045] During the cutting process, continuous visual inspection and feedback ensure consistent cutting quality.

[0046] Specifically, the intelligent quality control system includes:

[0047] Establishing a digital twin model of the sleep patch product, which can update and predict system status in real time;

[0048] Deploy a variety of online testing equipment, including near-infrared spectrometers, high-speed cameras, and mechanical properties testers;

[0049] Develop a multi-sensor data fusion algorithm and adopt an improved Kalman filter method to comprehensively evaluate the quality of the sleep patch;

[0050] Establish anomaly detection and prediction algorithms and use ensemble learning methods to improve the accuracy and robustness of anomaly identification;

[0051] The quality control results are fed back to each link of the production line in real time to form a closed-loop optimization mechanism.

[0052] Specifically, the intelligent quality control system also includes:

[0053] Based on historical production data and real-time monitoring results, a quality prediction model of the sleep patch is constructed;

[0054] The long short-term memory network (LSTM) is used as the core of the prediction model to capture the long-term dependencies in the production process.

[0055] Set multi-level warning thresholds, including normal, attention, warning and danger levels;

[0056] Depending on the severity of the warning information, different levels of response mechanisms are automatically triggered, including parameter fine-tuning, production suspension, and manual intervention;

[0057] The accuracy of the early warning system is continuously evaluated and the prediction model is continuously optimized through online learning methods.

[0058] Specifically, the production scheduling optimization algorithm includes:

[0059] Build a multi-objective optimization model that takes into account equipment utilization, raw material inventory levels, order delivery time, and energy consumption;

[0060] An improved ant colony optimization algorithm is used, and an adaptive pheromone update strategy and a local search mechanism are introduced;

[0061] Implement dynamic task allocation and use the improved Hungarian algorithm to quickly respond to production abnormalities;

[0062] Constructing a digital twin system of the sleep patch production line to simulate and predict production status in real time;

[0063] Based on the prediction results of the digital twin system, potential bottlenecks are identified in advance and preventive scheduling optimization is performed.

[0064] Specifically, it also includes: realizing the digital twin system of the sleep patch production line, updating the status of each workstation at a frequency of 1 Hz; using the long short-term memory network LSTM to build a prediction model, with an input window size of 24 and a hidden layer containing 128 units; realizing 3D visualization through WebGL technology, supporting a rendering frame rate of more than 60fps, and displaying the real-time status of the sleep patch production line; integrating a natural language processing module, supporting voice interaction and intelligent command parsing, and realizing remote control of the sleep patch production and processing technology; comparing the simulation results of the digital twin system with the actual production data, and continuously optimizing the accuracy of the digital twin model of the sleep patch product; based on the prediction results of the digital twin system, automatically generating production reports and optimization suggestions to further improve the intelligence level of the sleep patch production and processing technology.

[0065] The beneficial effects of the present invention are mainly reflected in the following aspects:

[0066] 1. Significantly improve product quality consistency. Through intelligent control of the entire process, the method of the present invention reduces the coefficient of variation of the content of the active ingredient of the product to 2.1%, which is far superior to the existing technical level.

[0067] 2. Greatly improved production efficiency. Using the method of the present invention, it only takes 3.5 hours to produce 1,000 pieces of sleeping patches, which is more than double the efficiency of the traditional method.

[0068] 3. Effectively reduce energy consumption and raw material waste. The energy consumption of the method of the present invention is reduced by 46% compared with the traditional method, and the raw material utilization rate is increased to 98.5%.

[0069] 4. Real-time quality control is achieved. Through full-process intelligent monitoring, the defective product rate is reduced to 0.7%, greatly reducing the losses caused by quality problems.

[0070] 5. Improved flexibility of the production process. The method of the present invention can quickly adapt to changes in formula and raw materials, shortening the development cycle of new products.

[0071] In general, the method provided by the present invention not only solves many problems existing in the production of existing sleep patches, but also provides a reference example for the intelligent transformation of the entire pharmaceutical industry. This comprehensive intelligent production method is expected to play an important role in improving product quality, reducing production costs, and enhancing corporate competitiveness, and promote the entire industry to develop in a more efficient and intelligent direction. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 The figure is a flow chart of the method of the present invention.

[0073] Figure 2 It is the logic block diagram of the adaptive recipe optimization of the present invention.

[0074] Figure 3 This is a logic block diagram of the continuous microfluidic mixing of the present invention.

[0075] Figure 4 This is a logic block diagram of the intelligent drying of the present invention.

[0076] Figure 5 This is a logic block diagram of the intelligent quality control of the present invention. DETAILED DESCRIPTION

[0077] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects thereof are described in detail below in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0078] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0079] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0080] like Figure 1 As shown, the present invention provides a method for intelligently optimizing the production and processing technology of a sleep patch, comprising the following steps:

[0081] Step S1: Obtain the sleep patch raw material database, target product parameter set and quality constraints; establish a digital twin model of the sleep patch product.

[0082] Preferably, in one embodiment of the present invention, the sleep patch raw material database contains a set of physical and chemical parameters of each raw material, such as density, viscosity, solubility, etc. These parameters are crucial for subsequent formulation optimization and production process control. For example, the viscosity of the raw material will directly affect the difficulty of the mixing and coating process, while the solubility is related to the release effect of the active ingredient.

[0083] The target product parameter set usually includes key indicators such as the size, thickness, drug content, and release curve of the sleep patch. These indicators are important bases for evaluating the quality of the final product. For example, a typical sleep patch may require a size of 5cm×5cm, a thickness of 0.5mm±0.05mm, a drug content of 5mg±0.25mg per piece, and a drug release of no less than 90% within 24 hours.

[0084] The quality constraints are a series of restrictions based on product quality standards and production process requirements. This may include raw material purity requirements, production environment temperature and humidity ranges, equipment parameter restrictions, etc. For example, the production environment temperature may be limited to 20℃-25℃, and the relative humidity may be controlled between 40%-60%.

[0085] Next, a digital twin model of the sleep patch product is established. This model is the core of the entire intelligent optimization method. It digitizes each link in the actual production process and realizes the combination of virtual and real. Specifically, the model includes the following key components:

[0086] 1) Raw material model: describes the physicochemical properties of various raw materials and their interactions.

[0087] 2) Process model: simulates the processes of each production link such as mixing, coating, drying, and cutting.

[0088] 3) Equipment model: reflects the operating status and performance parameters of production equipment.

[0089] 4) Environmental model: simulates the temperature, humidity, cleanliness and other factors of the production environment.

[0090] 5) Quality model: predict and evaluate product quality indicators.

[0091] This digital twin model is continuously optimized through real-time data updates and machine learning algorithms, and can accurately reflect the actual production conditions, providing a reliable basis for subsequent optimization decisions.

[0092] Step S2: According to the raw material database, the target product parameter set and the quality constraint conditions, the optimal ratio is determined by an adaptive formula optimization algorithm to obtain the raw material formula of the sleep patch.

[0093] Please refer to Figure 2 In a preferred embodiment of the present invention, the adaptive recipe optimization algorithm adopts a multi-objective optimization method to comprehensively consider product performance, production cost and process feasibility. The core of the algorithm is to establish a comprehensive evaluation index , which is expressed as follows:

[0094]

[0095] in, Indicates The weight of the parameters, Indicates actual product parameters, represents the target product parameters, is the cost balancing factor, is the total production cost. The goal of the algorithm is to minimize value.

[0096] The specific implementation steps are as follows:

[0097] 1) Initialization: Set the initial recipe based on experience.

[0098] 2) Evaluation: Calculate the J value of the current recipe.

[0099] 3) Optimization: Use the improved particle swarm algorithm to adjust the formula and find a better solution.

[0100] 4) Iteration: Repeat steps 2 and 3 until the preset number of iterations or optimization goal is reached.

[0101] 5) Verification: Compare the optimized formula with the quality constraints to ensure that all constraints are met.

[0102] For example, for a sleep patch containing melatonin, the initial formula may be: 5% melatonin, 15% polyvinyl pyrrolidone (PVP), 20% propylene glycol, 60% water. After optimization, a better formula may be obtained: 5.2% melatonin, 13.5% PVP, 18% propylene glycol, 63.3% water. This new formula may improve production efficiency and reduce costs while ensuring the efficacy of the drug.

[0103] Step S3: Based on the optimal ratio, a continuous microfluidic mixing system is used to mix the raw materials to form a uniform sleeping patch drug mixture.

[0104] Please refer to Figure 3 The present invention adopts an advanced continuous microfluidic mixing system, which has the advantages of high mixing efficiency and good product consistency compared with the traditional batch mixing method. The core of the system is a carefully designed microchannel structure with a channel width of Following flow direction The changes satisfy the following relationship:

[0105]

[0106] in, is the initial channel width, usually in the range of 100-500μm; α is the contraction coefficient, which controls the contraction rate along the channel, with a typical value of 0.01-0.1mm -1 is the fluctuation coefficient, which introduces periodic disturbance to enhance the mixing effect, and is generally taken as 0.1-0.3; is the total length of the channel, which may be in the range of 10-100 mm depending on the mixing requirements.

[0107] The optimization of the mixing process is mainly achieved by adjusting the flow rate and temperature. The system is equipped with high-precision pressure sensors and thermocouples, which collect data at a frequency of 100Hz and 10Hz respectively. Based on these real-time data, an adaptive control algorithm is used to dynamically adjust the parameters:

[0108] 1. Flow rate regulation: Use a PI controller to adjust the syringe pump speed according to pressure changes.

[0109]

[0110]

[0111] in, To control the output (pump speed), is the pressure error, and are the proportional and integral coefficients respectively.

[0112] 2. Temperature regulation: Fuzzy PID controller is used to adjust the heating power according to the temperature deviation and change rate. is the heating power adjustment, is the temperature deviation, is the dynamically adjusted PID parameter.

[0113] During the mixing process, the system uses fluorescence tracing to evaluate the mixing uniformity in real time. Specifically, a trace amount of fluorescent dye is added to one of the raw materials, and then the fluorescence intensity distribution is measured in the mixed fluid. Mixing uniformity Defined as:

[0114]

[0115] in, is the standard deviation of the actual fluorescence intensity, is the theoretical standard deviation when there is no mixing. When the value reaches the preset threshold (usually 0.95), the system determines that the mixing is sufficient and automatically enters the next production stage.

[0116] Step S4: Use a precision coating system to coat the mixed raw materials into a film to obtain a sleeping patch base film with uniform thickness.

[0117] The present invention adopts slit coating technology, combined with a precise closed-loop control system, to achieve high-precision and uniform coating effects. The core of the coating process is to establish a coating dynamics model to adjust the thickness of the base film of the sleep patch. Expressed as a function of multiple process parameters:

[0118]

[0119] in, For flow, is the substrate speed, is the viscosity, is the surface tension, is the contact angle. Preferably, in one embodiment of the present invention, a multi-objective optimization algorithm is used to simultaneously consider coating uniformity, surface tension gradient and energy consumption. The specific optimization objective function is:

[0120]

[0121] in, is the thickness uniformity index, is the surface tension gradient, For energy consumption, , , is the weight coefficient.

[0122] In order to achieve precise thickness control, the system uses a precision slot die driven by piezoelectric ceramics, with a response frequency of up to 1kHz and an adjustment range of ±50μm. The coating thickness is measured in real time by a laser triangulation rangefinder with a scanning frequency of 100Hz and a measurement accuracy of ±0.1μm.

[0123] Based on the measurement results, the system uses a fuzzy PID controller to dynamically adjust the coating parameters. The input to the controller is the thickness deviation and the rate of change of deviation , the output is the die gap adjustment amount The following are examples of fuzzy rules:

[0124] 1.

[0125] 2. If and then

[0126] Among them, PS, NS, ZO, NM, etc. are fuzzy sets, representing positive small, negative small, zero, and negative medium linguistic variables, respectively.

[0127] Through this precise control, the thickness of the sleeping patch base film can be controlled within ±2% of the target value, ensuring the consistency of product quality.

[0128] Step S5: performing intelligent drying treatment on the coated film-forming product to obtain a semi-finished sleeping patch with a moisture content that meets the requirements.

[0129] Please refer to Figure 4 The present invention adopts a multi-stage temperature and humidity controllable tunnel drying system, combined with a multi-physics field coupled drying model, to achieve precise drying control. The core of the drying process is to solve the following set of partial differential equations:

[0130]

[0131]

[0132] in, is the water content, is the temperature, is the diffusion coefficient, is the air flow velocity, is the thermal conductivity, is the volume heat source. In order to achieve precise control, the system deploys a distributed infrared moisture meter and temperature sensor network with a sampling frequency of 1Hz. Based on these real-time data, the adaptive multi-stage drying optimization algorithm (AMDO) is used to dynamically adjust the drying parameters. The objective function of AMDO is:

[0133]

[0134] in, For energy consumption, For drying time, is the time weight factor. The optimization process is subject to the following constraints:

[0135]

[0136]

[0137]

[0138] in, is the maximum allowable moisture content, is the maximum allowable temperature, The maximum permissible temperature gradient. During the drying process, the system precisely controls the heating element power through pulse width modulation (PWM) technology:

[0139]

[0140] in, is the maximum power, The modulation frequency can be dynamically adjusted within the range of 0.1-10Hz. Through this intelligent drying process, the final moisture content of the semi-finished sleeping patch can be controlled within the range of 2%-3%, while maximizing the protection of the active ingredients and improving product quality.

[0141] Step S6: using an online cutting system to accurately cut the dried product to form sleeping patches of uniform specifications.

[0142] The online cutting system of the present invention adopts machine vision technology combined with ultrasonic cutting to achieve a high-precision and high-efficiency cutting process. The core of the system includes the following parts:

[0143] 1) Image acquisition: Use a high-speed CMOS camera with a frame rate ≥ 1000fps and a resolution ≥ 2048x2048 pixels.

[0144] 2) Image processing: Use deep learning algorithms for real-time image segmentation and feature extraction. Specifically, a convolutional neural network with a U-Net architecture is used, and its loss function is:

[0145]

[0146] in, is the true label, is the predicted probability.

[0147] 3. Visual positioning: Establish the mapping relationship between image coordinates and actual cutting coordinates:

[0148]

[0149] in, is the actual coordinate, is the image feature vector, and is the matrix of unknown coefficients,

[0150] 4. Trajectory planning: Develop an adaptive cutting trajectory planning algorithm (ACTP), whose objective function is:

[0151]

[0152] in, is the actual cutting point, ,target is the target cutting point, Cut time for the president, is the time weight factor.

[0153] 5. Precision drive: The precision platform driven by piezoelectric ceramics controls the position of the cutting knife and the response frequency kHz, positioning accuracy ±1μm.

[0154] 6. Ultrasonic cutting: Dynamically adjust the cutting energy through the ultrasonic generator to ensure the smoothness and accuracy of the cutting edge. Cutting energy The control function is:

[0155]

[0156] in, is the base energy, is the modulation factor (usually in the range of 0.1-0.3), is the modulation frequency (adjustable in the range of 20-60kHz).

[0157] It is worth mentioning that during the cutting process, the system will continuously perform visual inspection and feedback to ensure the consistency of cutting quality. If any abnormality is detected, such as uneven edges or dimensional deviations outside the allowable range, the system will automatically adjust the cutting parameters or issue an alarm to intervene in the production process in a timely manner.

[0158] Through this precise online cutting system, we can control the size accuracy of the sleep patch within ±0.1mm, which greatly improves the consistency and appearance quality of the product.

[0159] Step S7: Use the intelligent quality control system to monitor the entire product process and detect abnormalities to ensure the quality stability of the sleep patch.

[0160] Please refer to Figure 5 In this step, we built a comprehensive intelligent quality control system, which is like the nerve center of the entire production line. The core of this system is a constantly updated digital twin model that reflects every link of the production line in real time.

[0161] Specifically, we deploy a variety of online testing equipment at key points of the production line. For example, we use near-infrared spectrometers to monitor the composition of raw materials and products, high-speed cameras to check the appearance quality, and mechanical properties testers to evaluate the physical properties of products. The massive amount of data generated by these devices is the basis of our quality control.

[0162] In order to make full use of this data, we developed a multi-sensor data fusion algorithm. The core of this algorithm is the improved Kalman filter method, which can effectively integrate information from different sensors and give a comprehensive quality assessment. The core formula of the algorithm is as follows:

[0163]

[0164]

[0165]

[0166] in, is the state estimate, is the Kalman gain, is the observation vector, is the observation matrix, is the covariance matrix, is the observation noise covariance. The beauty of this algorithm is that it not only gives the current quality assessment, but also predicts the quality problems that may occur in the future. This allows us to take action in advance and prevent problems before they occur.

[0167] In addition, we have built an anomaly detection system based on machine learning. This system uses an algorithm called Isolation Forest. Its basic idea is that abnormal data points are usually easier to be "isolated". The algorithm's anomaly score is calculated as follows:

[0168]

[0169] in, It is a data point The average path length, The sample size is The average path length of a binary search tree.

[0170] The system can quickly identify anomalies in the production process, such as when a parameter suddenly deviates from the normal range or an abnormal pattern begins to emerge. Once an anomaly is detected, the system will immediately issue an alarm so that operators can deal with it in time.

[0171] Finally, all quality control results will be fed back to each link of the production line in real time, forming a closed-loop optimization mechanism. In this way, the entire production process can continuously improve itself and continuously improve product quality.

[0172] Step S8: Implement dynamic task allocation and production line optimization based on the production scheduling optimization algorithm to improve the production efficiency of the sleep patch.

[0173] In this step, we introduced an intelligent production scheduling system, which is like the commander of the entire production line. The core of this system is a multi-objective optimization model, which takes into account multiple factors such as equipment utilization, raw material inventory level, order delivery time and energy consumption.

[0174] We used an improved ant colony optimization algorithm to solve this complex optimization problem. This algorithm was inspired by the foraging behavior of ants, but we made some innovative improvements to it. For example, we introduced an adaptive pheromone update strategy:

[0175]

[0176]

[0177] in, Is the path The pheromone concentration on is the evaporation coefficient, is a constant, It's an ant We also added a local search mechanism, which allows the algorithm to find a good solution faster. The probability of an ant choosing the next node is calculated as follows:

[0178]

[0179] here, and are weight factors, which are dynamically adjusted according to the search process to balance global and local search. The advantage of this algorithm is that it can quickly adapt to changes in the production environment. For example, if an urgent order is suddenly received or a certain equipment needs temporary maintenance, the system can quickly reschedule the production plan to ensure that the overall efficiency is maximized.

[0180] In addition, we also implemented a dynamic task allocation function. We used an improved Hungarian algorithm, whose time complexity is , but it performs quite well in practical applications. The core of the algorithm is to construct a cost matrix:

[0181]

[0182] in, is the time matrix, is the energy consumption matrix, is the mass matrix, , , is the weight coefficient.

[0183] The advantage of this algorithm is that it can quickly adapt to changes in the production environment. For example, if an urgent order is suddenly received or a certain equipment needs temporary maintenance, the system can quickly reschedule the production plan to ensure that the overall efficiency is maximized.

[0184] Another highlight of this system is that it is closely integrated with the digital twin system we mentioned earlier. The digital twin system can simulate and predict production status in real time, and this information is directly fed back to the scheduling system for optimized decision-making.

[0185] Through this intelligent scheduling system, we have not only improved production efficiency, but also greatly enhanced production flexibility and adaptability. In actual application, we have seen that the production cycle has been shortened by about 20% and the equipment utilization rate has increased by about 15%, which is very helpful for improving the market competitiveness of sleep patches.

[0186] Step S9: Feedback the optimized production parameters to the digital twin model to continuously improve the sleep patch production process.

[0187] This step is the finishing touch of the entire method. It establishes a feedback loop to achieve continuous optimization and self-improvement of the entire production process.

[0188] Specifically, our digital twin system updates the status of each workstation at a frequency of 1 Hz. This frequency is the optimal value obtained through repeated experiments. It can ensure the real-time nature of the data without placing too much computing burden on the system.

[0189] The core of the digital twin model is a prediction model based on the long short-term memory network (LSTM). The structure of this model is as follows:

[0190] 1. Input layer: receives historical data of 24 time steps;

[0191] 2.LSTM layer: contains 128 hidden units;

[0192] 3. Fully connected layer: maps the output of LSTM to the prediction result;

[0193] 4. Output layer: predict future production status.

[0194] The model is trained using the Back Propagation Through Time (BPTT) algorithm, and the loss function uses the Mean Square Error (MSE). This model can accurately predict the production status in the next few hours, providing an important basis for production decisions.

[0195] It is worth mentioning that we have also developed a 3D visualization system based on WebGL technology. This system can render the status of the entire production line in real time at a frame rate of more than 60fps. It can not only intuitively display the current production situation, but also simulate the effects of different production strategies, which is a very useful tool for production managers.

[0196] In addition, we have integrated a natural language processing module that supports voice interaction and intelligent command parsing. This module uses a BERT-based transfer learning model to accurately understand the operator's verbal instructions, greatly improving the system's usability.

[0197] Finally, the system will regularly generate detailed production reports and optimization suggestions. These reports not only contain regular production data, but also provide some in-depth analysis and suggestions. For example, it may suggest adjusting the parameters of a certain process or replacing a certain raw material to further improve product quality or reduce production costs.

[0198] Through this continuous feedback and optimization mechanism, our sleep patch production process can continue to evolve and always stay at the forefront of technology. In actual application, we have observed that the consistency of product quality has been significantly improved, and the defective rate has been reduced by about 40%, which has undoubtedly greatly improved the market competitiveness of the product.

[0199] In general, the intelligent optimization method for the production and processing of sleep patches provided by the present invention realizes the comprehensive intelligence and optimization of the production process by combining advanced technologies such as artificial intelligence, the Internet of Things, and big data analysis with traditional pharmaceutical processes. It not only greatly improves production efficiency and product quality, but also significantly reduces production costs and resource consumption.

[0200] More importantly, this method is highly adaptable and scalable. Although we developed this system for sleep patches, its core ideas and technologies can be fully extended to the production of other types of drugs or medical devices. I believe that with the continuous advancement and improvement of technology, this intelligent production method will play an increasingly important role in the pharmaceutical industry, providing patients with better quality and more reliable drugs.

[0201] Of course, we have also encountered some challenges in practical applications. For example, how to ensure the stability and reliability of the system, how to handle massive amounts of real-time data, how to balance automation and manual intervention, etc. We are actively researching and solving these problems. We believe that through continuous efforts and innovation, we will be able to develop a more complete and intelligent production system and make greater contributions to the development of the pharmaceutical industry.

[0202] In order to verify the superiority of the present invention, we conducted a series of comparative experiments. The following is a detailed description of the embodiments and comparative examples, as well as the corresponding test results and analysis.

[0203] Example 1: Intelligent optimization method for the production and processing technology of sleep patches using the present invention

[0204] We selected a transdermal sleep patch containing melatonin as the test object. According to the method described in the present invention, intelligent management is adopted throughout the process from raw material formula optimization to production line scheduling. Specifically:

[0205] 1. Use adaptive recipe optimization algorithm to determine the best recipe;

[0206] 2. Use a continuous microfluidic mixing system to mix raw materials;

[0207] 3. Use precision coating system for coating;

[0208] 4. Apply intelligent drying system to control the drying process;

[0209] 5. Use online cutting system for precise cutting;

[0210] 6. Use intelligent quality control system to monitor the production process throughout;

[0211] 7. Apply production scheduling optimization algorithm for dynamic task allocation.

[0212] Comparative Example 1: Traditional batch production method.

[0213] Using traditional batch production methods, including:

[0214] 1. Fixed formula;

[0215] 2. Batch mixing;

[0216] 3. Coating by ordinary coating machine;

[0217] 4. Fixed parameter oven drying;

[0218] 5. Offline cutting;

[0219] 6. Manual quality inspection;

[0220] 7.Fixed production plan.

[0221] Comparative Example 2: Partially intelligent method.

[0222] Based on the traditional methods, some intelligent measures are introduced, including:

[0223] 1. Formula optimization software-assisted design;

[0224] 2. Improved batch mixing equipment;

[0225] 3. Automatic coating system;

[0226] 4. Program-controlled drying system;

[0227] 5. Automatic cutting equipment;

[0228] 6. Automatic testing equipment assists quality control;

[0229] 7. Computer-aided production scheduling.

[0230] Comparative Example 3: Single-link intelligent method, which only uses intelligent methods in one link of the production line (such as the mixing process), while other links maintain traditional methods.

[0231] Test indicators and methods:

[0232] 1. Product quality consistency: Calculate the coefficient of variation (CV) by measuring the active ingredient content of 100 samples;

[0233] Test method: High performance liquid chromatography (HPLC).

[0234] 2. Production efficiency: record the time required to produce 1,000 pieces of product;

[0235] Test method: Timer recording.

[0236] 3. Energy consumption: record the electricity consumed to produce 1,000 pieces of product.

[0237] Test method: Smart meter measurement.

[0238] 4. Raw material utilization rate: calculate the proportion of input raw materials that eventually enter the product;

[0239] Test method: precision weighing.

[0240] 5. Defective product rate: Count the number of defective products among 1,000 products;

[0241] Test method: full inspection method.

[0242] The test results are shown in the following table:

[0243] Test indicators Example 1 Comparative Example 1 Comparative Example 2 Comparative Example 3 Quality consistency (CV) 2.1 8.7 5.3 7.2 Production efficiency (h) 3.5 8.2 5.7 7.1 Energy consumption (kWh) 42 60 70 70 Raw material utilization rate (%) 98.5 90.2 94.7 92.1 Defective rate (%) 0.7 4.5 2.8 3.9

[0244] It can be clearly seen from the test results that the method of the present invention (Example 1) shows obvious advantages in all indicators.

[0245] 1. Quality consistency: The coefficient of variation of Example 1 is only 2.1%, which is much lower than that of other comparative examples. This shows that the method of the present invention can effectively control the production process and ensure the stability of product quality. This is mainly due to the application of continuous microfluidic mixing system and precision coating system, as well as full-process intelligent quality control.

[0246] 2. Production efficiency: Example 1 only takes 3.5 hours to complete 1,000 pieces of products, which is more than twice as fast as the traditional method. This efficiency improvement mainly comes from the continuous production process and the optimization of the intelligent scheduling system.

[0247] 3. Energy consumption: The energy consumption of Example 1 is significantly reduced to only 42 kWh. This is because the method of the present invention can accurately control the energy input of each production link, avoiding unnecessary energy waste.

[0248] 4. Raw material utilization rate: Example 1 achieved a high utilization rate of 98.5%, which was due to precise formulation optimization and efficient production process control, greatly reducing the waste of raw materials.

[0249] 5. Defective product rate: The defective product rate of Example 1 is only 0.7%, which is much lower than that of other methods. This once again proves the superiority of the method of the present invention in quality control.

[0250] It is worth noting that although comparative examples 2 and 3 also adopt some intelligent measures, their effects are far inferior to those of the fully intelligent embodiment 1. This shows that only by systematically optimizing the entire production process intelligently can the advantages of intelligent manufacturing be fully utilized.

[0251] In general, this set of comparative experiments fully demonstrates the superiority of the method of the present invention. It can not only significantly improve product quality and production efficiency, but also effectively reduce energy consumption and raw material waste. This is of great significance for improving the competitiveness and sustainable development capabilities of enterprises.

[0252] In practical applications, we have found that this intelligent method is also highly adaptable. For example, when it is necessary to change raw material suppliers or adjust product specifications, the system can quickly re-optimize production parameters, greatly shortening the product development and process adjustment cycle. This flexibility is particularly important in today's rapidly changing market environment.

[0253] Of course, we also realize that implementing this comprehensive intelligent production method requires a large initial investment. However, in the long run, this investment is very worthwhile considering the improved production efficiency, reduced operating costs and improved product quality. We believe that with the further development of technology and the reduction of costs, this intelligent production method will be more and more widely used in the pharmaceutical industry.

[0254] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modification, replacement, and improvement made within the principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent optimization method for the production and processing technology of sleep patches, The following steps are involved: Obtain the raw material database of sleep patches, target product parameter sets and quality constraints; Establish a digital twin model of a sleep patch product; characterized in that the method also includes: According to the raw material database, target product parameter set and quality constraints, the optimal ratio is determined by an adaptive formula optimization algorithm to obtain the raw material formula of the sleep patch; based on the optimal ratio, a continuous microfluidic mixing system is used to mix the raw materials to form a uniform sleep patch drug mixture; the mixed raw materials are coated into a film using a precision coating system to obtain a sleep patch base film with uniform thickness; the coated film product is subjected to intelligent drying treatment to obtain a sleep patch semi-finished product with a moisture content that meets the requirements; an online cutting system is used to accurately cut the dried product to form a sleep patch with uniform specifications; an intelligent quality control system is used to monitor the product throughout the entire process and detect anomalies to ensure the quality stability of the sleep patch; dynamic task allocation and production line optimization are achieved based on a production scheduling optimization algorithm to improve the production efficiency of the sleep patch; the optimized production parameters are fed back to the digital twin model to continuously improve the sleep patch production process; The adaptive formula optimization algorithm adopts a multi-objective optimization method, comprehensively considering product performance, production cost and process feasibility. The core of the algorithm is to establish a comprehensive evaluation index , which is expressed as follows: ; in, Indicates The weight of the parameters, Indicates actual product parameters, represents the target product parameters, is the cost balancing factor, is the total production cost, and the goal of the algorithm is to minimize value; Microchannel structure of continuous microfluidic mixing system, and its channel width Following flow direction The changes satisfy the following relationship: ; middle, is the initial channel width, α is the contraction coefficient, is the volatility coefficient, is the total length of the channel; The precision coating system includes the establishment of a coating dynamics model to adjust the thickness of the base film of the sleeping patch. Expressed as a function of multiple process parameters: ; in, For flow, is the substrate speed, is the viscosity, is the surface tension, is the contact angle A multi-objective optimization algorithm is used to simultaneously consider coating uniformity, surface tension gradient, and energy consumption; The optimization objective function is: ; in, is the thickness uniformity index, is the surface tension gradient, For energy consumption, , , is the weight coefficient.

2. The method according to claim 1, It is characterized in that The adaptive recipe optimization algorithm includes: Assigning importance weights to each raw material in the sleep patch raw material database, and dynamically adjusting the weights according to their influence on the performance of the sleep patch; By means of an iterative optimization method, the ratio of each raw material is gradually adjusted until the optimal value of the evaluation index is reached; During the optimization process, the interaction between the raw materials is considered in real time to ensure that the synergistic effect of the raw material formula of the sleep patch is maximized; The optimized formula is compared with the target product parameter set, and if the deviation exceeds a preset threshold, it is optimized again.

3. The method according to claim 1, It is characterized in that The optimization of the continuous microfluidic mixing system includes: The pressure drop and temperature change during the raw material mixing process are monitored in real time by a pressure sensor and a thermocouple; Based on monitoring data, adaptive control algorithms are used to dynamically adjust flow rate and temperature to maximize mixing efficiency; During the mixing process, the uniformity of the drug mixture of the sleep patch is evaluated in real time by a fluorescent tracer method; When the mixing uniformity reaches the preset threshold, it automatically enters the next production stage.

4. The method according to claim 1, It is characterized in that The optimization of the precision coating system includes: Using a precision slot die driven by piezoelectric ceramics to achieve sub-micron thickness adjustment; The thickness of the base film of the sleeping patch is measured in real time by a laser triangulation rangefinder to form a closed-loop control system; Based on the measurement results, a fuzzy PID controller is used to dynamically adjust the coating parameters to ensure the uniformity of the base film of the sleeping patch.

5. The method according to claim 1, It is characterized in that The intelligent drying process comprises: Establish a multi-physics coupled drying model, taking into account moisture diffusion, heat conduction and convection heat transfer; Using distributed infrared moisture meters and temperature sensors, a real-time monitoring network for the moisture content and temperature distribution of the semi-finished sleeping patch is constructed; Based on monitoring data, an adaptive multi-stage drying optimization algorithm is used to dynamically adjust the drying temperature and air flow speed; Control the heating element power through pulse width modulation technology to achieve precise temperature control; During the drying process, the quality indicators, moisture content uniformity and thermal stress of the semi-finished sleeping patch are continuously evaluated to ensure its quality stability.

6. The method according to claim 1, It is characterized in that The optimization of the online cutting system includes: Use high-speed CMOS cameras to collect images in real time, and combine deep learning algorithms for image processing and feature extraction; Establish a visual positioning model to map image features to actual cropping coordinates; Develop an adaptive cutting trajectory planning algorithm that comprehensively considers cutting accuracy, speed, and energy consumption; The precision platform driven by piezoelectric ceramics is used to control the position of the cutting knife, achieving micron-level positioning accuracy. Dynamically adjusting cutting energy through an ultrasonic generator to optimize edge smoothness and precision of the sleeping patch with uniform specifications; During the cutting process, continuous visual inspection and feedback ensure consistent cutting quality.

7. The method according to claim 1, It is characterized in that The intelligent quality control system comprises: Establishing a digital twin model of the sleep patch product, which can update and predict system status in real time; Deploy a variety of online testing equipment, including near-infrared spectrometers, high-speed cameras, and mechanical properties testers; Develop a multi-sensor data fusion algorithm and adopt an improved Kalman filter method to comprehensively evaluate the quality of the sleep patch; Establish anomaly detection and prediction algorithms and use ensemble learning methods to improve the accuracy and robustness of anomaly identification; The quality control results are fed back to each link of the production line in real time to form a closed-loop optimization mechanism.

8. The method according to claim 7, It is characterized in that The intelligent quality control system also includes: Based on historical production data and real-time monitoring results, a quality prediction model of the sleep patch is constructed; The long short-term memory network (LSTM) is used as the core of the quality prediction model to capture the long-term dependencies in the production process. Set multi-level warning thresholds, including normal, attention, warning and danger levels; Depending on the severity of the warning information, different levels of response mechanisms are automatically triggered, including parameter fine-tuning, production suspension, and manual intervention; The accuracy of the early warning system is continuously evaluated and the prediction model is continuously optimized through online learning methods.

9. The method according to claim 1, It is characterized in that The production scheduling optimization algorithm includes: Build a multi-objective optimization model that takes into account equipment utilization, raw material inventory levels, order delivery time, and energy consumption; An improved ant colony optimization algorithm is used, and an adaptive pheromone update strategy and a local search mechanism are introduced; Implement dynamic task allocation and use the improved Hungarian algorithm to quickly respond to production abnormalities; Constructing a digital twin system of the sleep patch production line to simulate and predict production status in real time; Based on the prediction results of the digital twin system, potential bottlenecks are identified in advance and preventive scheduling optimization is performed.

10. The method according to claim 1, It is characterized in that Also includes: Implement the digital twin system of the sleep patch production line, and update the status of each workstation at a frequency of 1 Hz; use the long short-term memory network LSTM to build a prediction model, with an input window size of 24 and a hidden layer containing 128 units; use WebGL technology to achieve 3D visualization, support a rendering frame rate of more than 60fps, and display the real-time status of the sleep patch production line; integrate a natural language processing module to support voice interaction and intelligent command parsing, and realize remote control of the sleep patch production and processing technology; compare the simulation results of the digital twin system of the production line with the actual production data, and continuously optimize the accuracy of the digital twin model of the sleep patch product; based on the prediction results of the digital twin system, automatically generate production reports and optimization suggestions.