A corrugated board adhesive amount control method based on industrial artificial system optimization

Abstract

This application provides a method for controlling the adhesive amount of corrugated cardboard based on industrial artificial system optimization, comprising: acquiring adhesive amount data and coating thickness data of coating equipment; processing the adhesive amount data and coating thickness data using a machine learning algorithm to obtain adhesive amount control parameters and thickness optimization parameters; acquiring real-time monitoring signals from the optimized equipment configuration; transmitting the real-time monitoring signals for environmental factors during the substrate coating process to obtain environmental data; analyzing the environmental data using a machine learning algorithm to determine the relationship between humidity adjustment requirements and temperature control requirements; if the humidity adjustment requirement is higher than the temperature control requirement, prioritizing the determination of a humidity adjustment command; acquiring the current humidity value according to the humidity adjustment command; optimizing the adjustment path using an evolutionary algorithm to obtain a supplementary temperature control command; acquiring micro-control updates from the final adjustment scheme and transmitting them to an intelligent optimization network to obtain an overall operation optimization result.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence-based paperboard production technology, and in particular to a method for controlling the amount of adhesive used in corrugated paperboard based on industrial artificial system optimization. Background Technology

[0002] In the modern cardboard and corrugated board manufacturing industry, the adhesive coating process is one of the core elements determining product quality and production efficiency. The uniformity and precision of the adhesive coating directly affect the bonding strength, flatness, and subsequent printing and forming effects of the cardboard, thus becoming a key technology area for packaging manufacturers to enhance their competitiveness.

[0003] Currently, most production lines still rely on manual experience or mechanical equipment with fixed parameters for glue application. While this method is simple, it reveals significant shortcomings in actual production: because the moisture content, temperature, and thickness of the cardboard itself, as well as the ambient temperature and humidity, are constantly changing, it is difficult to maintain a stable and consistent amount and thickness of glue application. Especially in high-speed continuous production, even slight deviations in parameters can lead to excessive glue application in certain areas, causing glue overflow and cardboard warping, or insufficient glue application, resulting in quality problems such as delamination and peeling. These defects not only increase the scrap rate but also force companies to frequently stop production for adjustments, severely restricting production efficiency and consistency.

[0004] A deeper contradiction lies in the fact that the coating process involves multiple interdependent variables: the real-time humidity and temperature of the cardboard directly affect the adhesive's penetration and drying speed, while the pressure, speed, and glue tank level of the coating rollers collectively determine the actual coating thickness. These factors have complex dynamic relationships, and traditional equipment can only set static parameters, unable to respond and fine-tune in real time according to the actual condition of each meter or even centimeter of the cardboard. For example, when a batch of cardboard with a high moisture content suddenly appears during production, the adhesive is easily over-absorbed, leading to a decrease in bonding strength; and if the operator cannot adjust the amount of adhesive in time, another section of cardboard may develop bubbles or warp due to excessive adhesive. This contradiction of real-time fluctuations in dependent variables and the inability to accurately match them has become the most prominent technical bottleneck in current coating processes.

[0005] How to achieve real-time, microscopic, and dynamic control of glue quantity, coating thickness, and coating equipment operation under high-speed production conditions, and to accurately match and automatically optimize based on the actual conditions of the paperboard at each moment, such as humidity and temperature, has become a key issue in improving the quality and production efficiency of corrugated paperboard products. Summary of the Invention

[0006] This invention provides a method for controlling the amount of adhesive used in corrugated cardboard bonding based on industrial manual system optimization, mainly including: The coating equipment's adhesive quantity and thickness data are acquired. Machine learning algorithms are used to process these data to obtain adhesive quantity control parameters and thickness optimization parameters. The coating equipment is then fine-tuned based on these parameters, and the adjusted equipment status data is obtained. It is determined whether the equipment status data exceeds a preset threshold; if so, an evolutionary algorithm is used to recalculate the control parameters to determine the optimized equipment configuration. Real-time monitoring signals are obtained from the optimized equipment configuration, and these signals are transmitted to address environmental factors during the substrate coating process, resulting in environmental data. Machine learning algorithms are used to analyze the environmental data, determining the relationship between humidity adjustment requirements and temperature control requirements. If the humidity adjustment requirement is higher than the temperature control requirement, a humidity adjustment command is prioritized. The current humidity value is obtained based on the humidity adjustment command, and an evolutionary algorithm is used to optimize the adjustment path, resulting in a supplementary temperature control command. The coating optimization process is then processed based on the supplementary temperature control command and the environmental adjustment mechanism, determining whether the coating optimization indicators meet preset thresholds. If they do, operational optimization feedback is generated to determine the final adjustment scheme. Micro-control updates are obtained from the final adjustment scheme and transmitted to the intelligent optimization network to obtain the overall operational optimization result.

[0007] Furthermore, the step of acquiring adhesive quantity data and coating thickness data of the coating equipment, and processing the adhesive quantity data and coating thickness data using a machine learning algorithm to obtain adhesive quantity control parameters and thickness optimization parameters includes: acquiring the adhesive quantity data in real time through the adhesive quantity sensor of the coating equipment, acquiring the coating thickness data in real time through the thickness sensor, inputting the adhesive quantity data and coating thickness data into a pre-trained machine learning model, wherein the machine learning model performs feature extraction and mapping processing on the adhesive quantity data and coating thickness data, and outputs the adhesive quantity control parameters for equipment control and the thickness optimization parameters for thickness adjustment.

[0008] Furthermore, the step of fine-tuning the coating equipment according to the adhesive dosage control parameters and the thickness optimization parameters, obtaining the fine-tuned equipment status data, determining whether the equipment status data exceeds a preset threshold, and if so, recalculating the control parameters using an evolutionary algorithm to determine the optimized equipment configuration includes: transmitting the adhesive dosage control parameters and the thickness optimization parameters to the control unit of the coating equipment to perform parameter fine-tuning operations, collecting the operating status of the coating equipment after fine-tuning to form the equipment status data, comparing the equipment status data with a preset threshold, and if the equipment status data exceeds the preset threshold, starting an evolutionary algorithm to iteratively optimize the control parameters, and obtaining the optimized equipment configuration through multiple generations of evolutionary calculations.

[0009] Furthermore, the step of obtaining real-time monitoring signals from the optimized equipment configuration and transmitting the real-time monitoring signals for environmental factors during the substrate coating process to obtain environmental data includes: extracting real-time monitoring signals associated with the substrate coating process from the optimized equipment configuration, transmitting the real-time monitoring signals to the control system, and having the control system collect humidity and temperature values ​​during the substrate coating process based on the real-time monitoring signals and integrate them to form the environmental data.

[0010] Furthermore, the step of using machine learning algorithms to analyze the environmental data and determine the relationship between humidity adjustment needs and temperature control needs, and prioritizing the determination of humidity adjustment instructions if the humidity adjustment needs are higher than the temperature control needs, includes: inputting the environmental data into a machine learning analysis model, wherein the machine learning analysis model performs importance assessment and demand intensity calculation on the humidity and temperature components in the environmental data to obtain humidity adjustment demand indicators and temperature control demand indicators, and when the humidity adjustment demand indicator is higher than the temperature control demand indicator, the humidity adjustment instruction for humidity is generated first.

[0011] Furthermore, the step of obtaining the current humidity value according to the humidity adjustment instruction, optimizing the adjustment path using an evolutionary algorithm, and obtaining a supplementary temperature control instruction includes: extracting the current humidity value from the substrate monitoring system according to the humidity adjustment instruction, inputting the current humidity value as an initial condition into the evolutionary algorithm, and iteratively optimizing the adjustment path through population initialization, fitness evaluation, crossover mutation, and selection operations, while simultaneously generating the supplementary temperature control instruction for supplementary adjustment during the optimization process.

[0012] Furthermore, the step of processing the coating optimization process according to the temperature control supplementary command and the environmental adjustment mechanism, determining whether the coating optimization index meets the preset threshold, and generating operation optimization feedback if it does, and determining the final adjustment scheme, includes: combining the temperature control supplementary command with the environmental adjustment mechanism to execute the coating optimization process, monitoring the optimization index in real time during the coating process, comparing the optimization index with the preset threshold, generating the operation optimization feedback if the optimization index meets the preset threshold, and integrating various control results according to the operation optimization feedback to form the final adjustment scheme.

[0013] Furthermore, the step of obtaining micro-control updates from the final adjustment scheme and transmitting them to the intelligent optimization network to obtain the overall operation optimization result includes: extracting micro-control updates for the equipment and environment from the final adjustment scheme, transmitting the micro-control updates to the intelligent optimization network, and the intelligent optimization network performing global integration and verification processing on the micro-control updates to output the overall operation optimization result characterizing the overall coating process optimization.

[0014] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses an intelligent adaptive optimization method for the coating process. Addressing the unique business scenario where coating equipment is susceptible to environmental humidity and temperature fluctuations during adhesive amount and thickness control, leading to difficulties in maintaining coating uniformity and stability, the method achieves precise adaptive optimization through multi-level closed-loop intelligent control. First, adhesive amount and coating thickness data are collected in real time, input into a machine learning model to extract features, and output adhesive amount control parameters and thickness optimization parameters. Based on this, the equipment is fine-tuned, and the status data is verified to ensure it does not exceed thresholds. If the threshold is exceeded, an evolutionary algorithm is initiated to iteratively optimize the equipment configuration. Then, real-time monitoring signals are extracted from the optimized configuration to collect environmental data. Machine learning is used to analyze the priority of humidity and temperature requirements. When the humidity adjustment requirement is stronger, a humidity adjustment command is generated first, and the current humidity value is used as the initial condition to optimize the adjustment path using an evolutionary algorithm, simultaneously generating supplementary temperature control commands. Finally, the coating optimization process is executed in conjunction with the environmental adjustment mechanism. After monitoring and ensuring the optimization indicators meet the standards, operational optimization feedback is generated and integrated to obtain the final adjustment plan. The micro-control updates are then transmitted to the intelligent optimization network for global integration and verification, outputting the overall operational optimization result. This invention effectively solves the problem of coating quality fluctuation under environmental interference through a multi-level closed-loop feedback mechanism that deeply integrates machine learning and evolutionary algorithms, thereby achieving high-precision adaptive and stable control of the coating process and improving overall efficiency. Attached Figure Description

[0015] Figure 1 This is a flowchart of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization according to the present invention.

[0016] Figure 2 This is a schematic diagram of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization according to the present invention.

[0017] Figure 3 This is another schematic diagram of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization according to the present invention.

[0018] Figure 4 This is another schematic diagram of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization according to the present invention.

[0019] Figure 5 This is another schematic diagram of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization according to the present invention.

[0020] Figure 6 This is another schematic diagram of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization according to the present invention.

[0021] Figure 7This is another schematic diagram of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization according to the present invention.

[0022] Figure 8 This is another schematic diagram of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization according to the present invention.

[0023] Figure 9 This is another schematic diagram of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization according to the present invention.

[0024] Figure 10 This is another schematic diagram of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization according to the present invention.

[0025] Figure 11 This is another schematic diagram of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization according to the present invention.

[0026] Figure 12 This is another schematic diagram of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization according to the present invention.

[0027] Figure 13 This is another schematic diagram of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization according to the present invention.

[0028] Figure 14 This is another schematic diagram of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization according to the present invention. Detailed Implementation

[0029] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] like Figures 1-14 This embodiment of a method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization may specifically include: S101. Obtain the adhesive amount data and adhesive thickness data of the coating equipment, and process the adhesive amount data and adhesive thickness data using a machine learning algorithm to obtain adhesive amount control parameters and thickness optimization parameters.

[0031] S1. Obtain the adhesive amount data and adhesive thickness data of the coating equipment, and process the adhesive amount data and adhesive thickness data using a machine learning algorithm to obtain adhesive amount control parameters and thickness optimization parameters.

[0032] In one embodiment, obtaining adhesive quantity data and adhesive thickness data of the coating equipment in step S1 specifically includes step S11, monitoring the amount of adhesive consumed during the coating process in real time through an adhesive quantity sensor, for example, capturing the adhesive flow rate per second when the coating roller rotates, and measuring the adhesive layer height on the substrate surface through a thickness sensor, for example, obtaining micron-level precision data using laser ranging.

[0033] Step S12: The adhesive volume data and coating thickness data are processed using the support vector machine algorithm. The support vector machine is a supervised learning method that uses a hyperplane to separate data points to achieve regression prediction. First, the adhesive volume data and thickness data are used as input vectors to form a feature set. Then, a kernel function such as a radial basis function is calculated to map the high-dimensional space. The margin is optimized to minimize the error, and adhesive volume control parameters such as flow rate adjustment ratio and thickness optimization parameters such as layer height correction value are obtained.

[0034] For example, when processing glue volume data, the support vector machine learns the pattern of glue viscosity and flow rate relationship based on historical training samples. For instance, during the training phase, multiple sets of glue volume-thickness pairs are input, and the algorithm adjusts the weights to fit the curve, ensuring that the predicted parameters are accurate for subsequent equipment fine-tuning.

[0035] In one possible implementation, step S1 can be replaced by using a random forest algorithm to process the data. Random forest is an ensemble learning method consisting of multiple decision trees, each trained independently and voting to make decisions. First, a subset is constructed by randomly sampling from the glue amount data and thickness data. Then, at each tree node split, the best feature, such as glue amount deviation or thickness fluctuation, is selected. The outputs of all trees are aggregated to obtain glue amount control parameters and thickness optimization parameters. This method can reduce the risk of overfitting and improve robustness under noisy data. For example, in scenarios with large changes in coating speed, random forest can better capture nonlinear relationships, thereby making parameter optimization more stable and beneficial to improving coating uniformity.

[0036] It should be noted that in high-viscosity adhesive coating scenarios, support vector machines are suitable for precise boundary partitioning, while random forests are suitable for multivariate interactions. The choice between the two algorithms depends on the data scale. For example, when the number of data points exceeds 1000, random forests are preferred to accelerate the computation.

[0037] In one embodiment, step S1 is applied to the field of thin film coating. For example, when the substrate is a plastic film, the adhesive amount is set to 5 grams per square meter and the thickness target is 10 micrometers. After processing by a support vector machine, the adhesive amount control parameter is adjusted to reduce by 0.5 grams to avoid overflow, and the thickness optimization parameter is corrected to increase by 1 micrometer to ensure uniform coverage. This reduces waste and improves product adhesion.

[0038] For example, in the paper coating scenario, the initial adhesive amount is 8 grams per square meter and the thickness is 15 micrometers. After analysis by the random forest algorithm, the output adjustment parameter is to increase by 1 gram to compensate for the absorption rate, and the optimization parameter is to decrease by 2 micrometers to prevent blistering. The beneficial effect is to enhance the coating durability and reduce the defect rate.

[0039] In one possible implementation, after obtaining the adhesive volume control parameters and thickness optimization parameters in step S1, they are directly used for equipment feedback, such as transmitting the parameter values ​​to the controller to achieve closed-loop adjustment, avoid manual intervention, and improve the level of automation.

[0040] S102. Fine-tune the coating equipment according to the adhesive amount control parameters and the thickness optimization parameters, obtain the equipment status data after fine-tuning, determine whether the equipment status data exceeds the preset threshold, and if it does, use an evolutionary algorithm to recalculate the control parameters and determine the optimized equipment configuration.

[0041] S1, the coating equipment is fine-tuned according to the adhesive amount control parameters and the thickness optimization parameters.

[0042] In one embodiment, uniform distribution of adhesive and precise control of thickness are achieved by adjusting the roller speed and pressure parameters of the coating equipment.

[0043] Specifically, the glue quantity control parameters include the glue supply rate, and the thickness optimization parameters involve the adjustment of the scraper gap. These parameters are input into the equipment controller to complete the fine-tuning process.

[0044] S2, obtain the device status data after fine-tuning.

[0045] Specifically, data on roller speed, pressure, and adhesive flow are collected from equipment sensors to form an equipment status dataset.

[0046] S3, determine whether the device status data exceeds a preset threshold.

[0047] In one possible implementation, the collected data is compared with preset thresholds such as a maximum roller speed of 20 rpm and a maximum pressure of 5 MPa. If either data exceeds the threshold, the optimization process begins.

[0048] If S4 is exceeded, the control parameters are recalculated using an evolutionary algorithm.

[0049] S41, initialize the population, including multiple combinations of control parameters, such as the glue supply rate from 10 ml / s to 15 ml / s and the scraper gap from 0.1 mm to 0.3 mm.

[0050] S42, Calculate the fitness function, using mean squared error to evaluate the degree of improvement of each combination on the device state. The fitness function is defined as the sum of squared differences between the actual state data and the target state, where the target state is a stable value within a threshold.

[0051] S43 performs selection, crossover, and mutation operations. The selection operation uses a roulette wheel method, the crossover operation swaps parameter values, and the mutation operation randomly adjusts the parameters within a range of ±5%.

[0052] S44, iterate until convergence, setting the upper limit of the number of iterations to 50, or stop when the fitness change is less than 0.01.

[0053] In one embodiment, for high-viscosity adhesive scenarios, the initial population focuses on a lower supply rate, and the fitness function emphasizes thickness uniformity to avoid defects caused by uneven coating and improve optimization efficiency.

[0054] For example, in precision electronic coating applications, the initial parameter combination is 10 ml / s of adhesive and 0.2 mm of thickness. If the status data exceeds the threshold, after iteration through an evolutionary algorithm, the new parameters are 12 ml / s and 0.15 mm, which reduces waste and improves the product qualification rate.

[0055] In another embodiment, for low-viscosity adhesives, the population initialization covers a wider range, and the fitness function incorporates a weight for adhesive quantity stability to ensure that the algorithm is robust in variable environments and improve the adaptability of the coating process.

[0056] For example, the initial combination was a supply rate of 15 ml / s and a gap of 0.3 mm, which was later adjusted to 13 ml / s and 0.25 mm, reducing equipment vibration and improving overall operational stability.

[0057] It should be noted that evolutionary algorithms, such as genetic algorithms, can effectively search the parameter space by simulating the natural selection process, avoid local optima, achieve global optimization, and are beneficial for parameter adjustment of complex coating systems.

[0058] S5, determine the optimized equipment configuration.

[0059] Specifically, the device controller settings are updated from the optimal parameter combination output by the evolutionary algorithm to form an optimized configuration, ensuring the stability of the coating process.

[0060] S103. Obtain real-time monitoring signals from the optimized equipment configuration, transmit the real-time monitoring signals for environmental factors during the substrate coating process, and obtain environmental data.

[0061] Step S3: Obtain real-time monitoring signals from the optimized equipment configuration, transmit the real-time monitoring signals for environmental factors during the substrate coating process, and obtain environmental data.

[0062] In one embodiment, step S3 specifically includes step S31, extracting real-time monitoring signals from the optimized device configuration, which are derived from coating process data collected by the device sensors.

[0063] Step S32: For environmental factors such as humidity and temperature during the substrate coating process, real-time monitoring signals are transmitted to the control system, wherein the transmission adopts a wireless communication protocol to ensure data real-time performance.

[0064] Step S33: The transmitted signals are processed by the control system to generate environmental data for subsequent analysis.

[0065] For example, this implementation is applied to a coating equipment, where the optimized equipment configuration has been adjusted for parameters using an evolutionary algorithm. Real-time monitoring signals include sensor readings, and data integrity is ensured during transmission, taking into account factors such as humidity and temperature.

[0066] It should be noted that environmental factors are prioritized during transmission; for example, humidity signals precede temperature signals to match coating requirements. This improves coating uniformity and reduces defects.

[0067] In one embodiment, in step S31, a real-time monitoring signal is obtained from the optimized device configuration, specifically through an embedded sensor network, such as a signal acquisition device installed on the coating roller, which collects data once per second to ensure that the signal is updated in real time.

[0068] Step S311: Identify the signal sources in the device configuration, such as the monitoring points associated with the glue volume sensor.

[0069] Step S312: Extract signal values ​​to form data packets.

[0070] For example, in a glass substrate coating scenario, signal acquisition targets humidity-sensitive areas at a frequency of 10Hz. This allows for timely detection of environmental fluctuations, helping to maintain stable coating thickness and avoid quality problems caused by uneven adhesive layers.

[0071] It should be noted that step S32 transmits real-time monitoring signals for environmental factors during the substrate coating process. These environmental factors include humidity and temperature. The transmission process uses an Internet of Things (IoT) protocol to send signals from the device to the central control system. Humidity and temperature are critical factors because excessive humidity during coating can cause adhesive dilution, and temperature changes affect the drying speed.

[0072] In one embodiment, the transmission path is optimized to a low-latency link, such as using a 5G module for transmission, to ensure signal latency of less than 50ms.

[0073] For example, in thin film substrate coating, humidity signal transmission is prioritized, and an alarm signal is immediately transmitted when the humidity exceeds 80%. This approach enables rapid response to environmental changes, improves overall coating efficiency, and reduces scrap rates.

[0074] Step S33 obtains environmental data and aggregates and transmits signals through the control system, such as combining humidity and temperature values ​​into a data vector to determine adjustment needs.

[0075] For example, the environmental data output is a structured table containing timestamps, humidity readings such as 75%, and temperature readings such as 25°C. This data facilitates analysis by machine learning algorithms and is beneficial for optimizing the stability of the coating process.

[0076] S104. Analyze the environmental data using machine learning algorithms to determine the relationship between humidity adjustment needs and temperature control needs. If the humidity adjustment needs are higher than the temperature control needs, then the humidity adjustment command will be determined first.

[0077] In one embodiment, a machine learning algorithm is used to analyze the environmental data to determine the relationship between humidity adjustment requirements and temperature control requirements. If the humidity adjustment requirements are higher than the temperature control requirements, a humidity adjustment command is determined first. Specifically, this includes step S1, collecting environmental data, including humidity and temperature values ​​during the substrate coating process. These data are transmitted from real-time monitoring signals to the control system.

[0078] Step S2: Use the Support Vector Machine (SVM) algorithm to process the environmental data. The SVM algorithm is a supervised learning method that separates different categories of data by constructing a hyperplane. Here, it is used to classify the relationship between humidity adjustment needs and temperature control needs. First, the environmental data is converted into feature vectors. The feature vectors include humidity deviation rate and temperature deviation rate. The humidity deviation rate is the difference between the current humidity and the target humidity divided by the target humidity. The temperature deviation rate is calculated similarly.

[0079] Step S3: Based on the classification results of the support vector machine algorithm, calculate the quantization scores of humidity adjustment demand and temperature control demand. The quantization scores are obtained through the decision function of the algorithm. The decision function is a linear or nonlinear mapping obtained based on the training data. If the quantization score of humidity adjustment demand is higher than the quantization score of temperature control demand, then it is determined that the humidity adjustment demand is higher than the temperature control demand.

[0080] Step S4: When it is determined that the humidity adjustment requirement is higher, a humidity adjustment command is generated first. The command includes specific parameters for adjusting the humidity to the target range, such as increasing the power of the dehumidifier.

[0081] For example, when coating equipment processes paper substrates, environmental data shows that the humidity deviation rate is 0.15 and the temperature deviation rate is 0.08. The support vector machine algorithm calculates the humidity adjustment requirement score as 0.75 and the temperature control requirement score as 0.45. Since 0.75 is higher than 0.45, the humidity adjustment command is given priority, adjusting the humidity from the current 65% to the target 55%. This can prevent the glue from being diluted excessively and improve the uniformity of the coating thickness.

[0082] In one possible implementation, the training process of the support vector machine algorithm in step S2 uses historical environmental data as the training set. The historical data includes humidity and temperature records and corresponding coating quality indicators in past coating operations. The hyperplane is optimized by minimizing the classification error to ensure that the algorithm accurately judges the demand relationship.

[0083] Specifically, the benefits of this method are that it optimizes environmental factors in real time and reduces coating defects such as uneven thickness, because prioritizing humidity can stabilize adhesive volume data more quickly and avoid the chain reaction caused by temperature fluctuations.

[0084] For example, in another scenario where plastic substrates are processed, the environmental data shows a humidity value of 70% and a temperature value of 28 degrees Celsius, with deviation rates of 0.20 and 0.10, respectively. The algorithm scores are 0.82 for humidity and 0.50 for temperature. The priority humidity adjustment instruction reduces the humidity to 50%, thereby improving the accuracy of the thickness optimization parameters.

[0085] In one embodiment, the quantization score calculation in step S3 can also be combined with a radial basis function kernel, which maps the feature vector to a higher-dimensional space through a Gaussian distribution, thereby improving the accuracy of nonlinear relationship judgment. For example, in a coating environment with strong humidity and temperature coupling, the score calculation is more accurate and avoids misjudgment.

[0086] For example, in coating processes where adhesive volume control parameters are sensitive, environmental data analysis shows that when humidity demand is dominant, priority commands can improve coating optimization indicators by 15%, because humidity directly affects adhesive viscosity, while temperature is a secondary factor, thus ensuring the effectiveness of the final adjustment plan.

[0087] S105. Obtain the current humidity value according to the humidity adjustment instruction, optimize the adjustment path using an evolutionary algorithm, and obtain a temperature control supplementary instruction; process the coating optimization process according to the temperature control supplementary instruction and the environmental adjustment mechanism, determine whether the coating optimization index meets the preset threshold, and if it does, generate operation optimization feedback and determine the final adjustment scheme.

[0088] Step S1: Obtain the current humidity value according to the humidity adjustment command, optimize the adjustment path using an evolutionary algorithm, and obtain a supplementary temperature control command.

[0089] Step S11: Read the current humidity value from the substrate monitoring system, for example, data collected in real time by a humidity sensor.

[0090] Step S12: Input the current humidity value as the initial population into the genetic algorithm. The genetic algorithm is an optimization method based on natural selection and genetics, which searches for the optimal solution by simulating crossover, mutation and selection operations.

[0091] Step S121: Initialize the population, including multiple possible adjustment paths, each path consisting of transition parameters from humidity to temperature.

[0092] Step S122: Perform fitness assessment and calculate the impact score of each path on coating uniformity. The score is based on the weighted sum of humidity deviation and temperature compensation.

[0093] Step S123: Exchange path parameters through crossover operation to generate a new path, and apply mutation to introduce random changes.

[0094] Step S124: Iteratively select a high-fitness path until convergence, and output the optimized path as a supplementary instruction for temperature control.

[0095] In one embodiment, the genetic algorithm used in step S1 can handle multi-objective optimization in the coating process. For example, in the scenario where the substrate is a thin film, the initial population size is set to 50 and the number of iterations is 100. This can quickly converge to the humidity adjustment path, avoid uneven coating thickness, and improve equipment stability.

[0096] For example, in a scenario where the amount of adhesive applied is 10 grams per square meter and the current humidity is 60%, the temperature control supplementary instruction obtained after optimization by the genetic algorithm is to lower the temperature by 2 degrees Celsius. This helps to compensate for the changes in adhesive flow caused by excessive humidity and improve the coating quality.

[0097] Step S2: Based on the temperature control supplementary instruction and the environmental adjustment mechanism, process the coating optimization process, determine whether the coating optimization index meets the preset threshold, and if it does, generate operation optimization feedback to determine the final adjustment plan.

[0098] Step S21: Input the temperature control supplementary command into the environmental adjustment mechanism, which adjusts the humidity and temperature parameters through a feedback loop.

[0099] Step S22: Perform the coating optimization process and monitor optimization indicators such as thickness uniformity and adhesive distribution.

[0100] Step S23: Compare the optimization index with the preset threshold, for example, a thickness deviation of less than 0.05 mm is considered to meet the condition.

[0101] Step S24: If satisfied, generate operation optimization feedback, including a summary of adjusted parameters, and determine the final adjustment scheme as output.

[0102] In one embodiment, the judgment process in step S2 can be applied to a high-speed coating production line. When the temperature control supplementary instruction indicates that the temperature should be reduced, the environmental adjustment mechanism activates the fan adjustment. The feedback generated after the optimization index meets the threshold helps to reduce energy consumption and improve production efficiency.

[0103] For example, in a coating process with an ambient temperature of 25 degrees Celsius, the optimization index showed that the thickness uniformity reached 98%, exceeding the preset threshold of 95%, thus determining the final adjustment plan, including maintaining the current humidity control, which is beneficial for long-term equipment optimization.

[0104] S106. Obtain micro-control updates from the final adjustment scheme, transmit them to the intelligent optimization network, and obtain the overall operation optimization result.

[0105] In one embodiment, step S7 obtains micro-control updates from the final adjustment scheme and transmits them to the intelligent optimization network to obtain the overall operation optimization result. Specifically, step S71 extracts micro-control updates from the final adjustment scheme, which include fine-grained parameter adjustment values ​​for the amount and thickness of adhesive on the coating equipment.

[0106] Step S72: The extracted micro-control update is transmitted to the intelligent optimization network. The intelligent optimization network is a neural network-based structure used to integrate multi-source data for global optimization. Specifically, the update value is input into the input layer of the network through a data interface, and the weights of each node are calculated through forward propagation within the network.

[0107] Step S73: Based on the transmission results, the overall operation optimization results are obtained by processing in the intelligent optimization network. The processing uses a convolutional neural network algorithm to analyze the correlation between the updated values ​​and historical device data. The convolutional neural network algorithm is a commonly used image and sequence data processing method. It extracts features through convolutional layers, reduces dimensions through pooling layers, and outputs optimization results through fully connected layers.

[0108] For example, in the operating scenario of a coating equipment, suppose the final adjustment plan updates the micro-control to a glue dosage control parameter of 0.5 g / m. 2 With a thickness optimization parameter of 10μm, the intelligent optimization network calculated the overall operational optimization result after transmission, resulting in a 15% improvement in equipment efficiency. This was achieved through the network's global integration of parameters.

[0109] In one embodiment, the specific process of obtaining the overall operation optimization result in step S73 further includes step S731, combining the transmitted micro-control update with the coating history data stored in the network to form an input dataset.

[0110] Step S732: Apply the backpropagation algorithm to optimize network parameters. The backpropagation algorithm is a gradient descent method used to minimize the loss function. By calculating the error gradient, the weights are adjusted to ensure accurate output.

[0111] Step S733: Generate overall operation optimization results based on the optimized network, such as the adjusted device configuration scheme.

[0112] For example, regarding humidity adjustment during substrate coating, if the micro-control update indicates that the humidity value needs to be reduced by 5%, the results obtained after network processing include the integration of temperature supplementation instructions, which brings beneficial effects such as a 20% improvement in coating uniformity and a reduction in defect rate.

[0113] In one embodiment, considering different parameter applications, step S7 is applied to a high-precision coating scenario. After micro-adjustment update transmission, the intelligent optimization network processes time-series data through a recurrent neural network algorithm. The recurrent neural network algorithm is a model for processing sequences, retaining historical state information, and outputting a predicted optimization curve, which is beneficial to long-term equipment stability.

[0114] For example, if the update value is a thickness parameter of 15μm, the network will produce an overall adhesive volume control path, which will optimize energy consumption and reduce it by 10%.

[0115] In one embodiment, the transmission process in step S72 can be extended to encrypted transmission to ensure data security. Combined with the optimized path of the evolutionary algorithm in the technical disclosure, the reasoning extends to verifying the consistency of the update immediately after the network receives the data, which is beneficial for real-time response to environmental factors such as changes in humidity and temperature.

[0116] For example, in the scenario of supplementary temperature control instructions, the network quickly generates results after transmission and updates, preventing coating optimization indicators from exceeding the threshold.

[0117] S107. The step of acquiring adhesive quantity data and adhesive thickness data of the coating equipment, and processing the adhesive quantity data and adhesive thickness data using a machine learning algorithm to obtain adhesive quantity control parameters and thickness optimization parameters includes: acquiring the adhesive quantity data in real time through the adhesive quantity sensor of the coating equipment, acquiring the adhesive thickness data in real time through the thickness sensor, inputting the adhesive quantity data and adhesive thickness data into a pre-trained machine learning model, wherein the machine learning model performs feature extraction and mapping processing on the adhesive quantity data and adhesive thickness data, and outputs the adhesive quantity control parameters for equipment control and the thickness optimization parameters for thickness adjustment.

[0118] The adhesive quantity data is collected in real time by the adhesive quantity sensor of the coating equipment, and the adhesive thickness data is collected in real time by the thickness sensor. The adhesive quantity data and the adhesive thickness data are input into a pre-trained machine learning model. The machine learning model performs feature extraction and mapping processing on the adhesive quantity data and the adhesive thickness data, and outputs the adhesive quantity control parameters for equipment control and the thickness optimization parameters for thickness adjustment.

[0119] In one embodiment, the adhesive volume data is collected in real time by an adhesive volume sensor of the coating equipment, including: the adhesive volume sensor is installed in the adhesive supply system of the coating equipment, the sensor adopts the principle of flow meter, and collects the adhesive flow rate once per second to obtain adhesive volume data such as the adhesive volume value per minute.

[0120] The coating thickness data is collected in real time by a thickness sensor, including: the thickness sensor is placed behind the coating roller, and a laser ranging method is used to measure the distance difference between the coating surface and the substrate every second to obtain coating thickness data such as micron-level thickness values.

[0121] The glue volume data and the glue coating thickness data are input into a pre-trained machine learning model. The machine learning model performs feature extraction and mapping processing on the glue volume data and the glue coating thickness data, including: the machine learning model adopts the support vector machine algorithm, which maps the input data to a high-dimensional space through a kernel function. First, the mean and variance of the glue volume data are extracted as features, and then the fluctuation range of the thickness data is extracted as another feature. Then, the mapping processing is performed to associate these features with historical control parameters.

[0122] The output includes the glue quantity control parameters for equipment regulation and the thickness optimization parameters for thickness adjustment, including: based on the mapping results, the model calculates the glue quantity control parameters such as the glue supply pump speed adjustment value and the thickness optimization parameters such as the roller pressure correction value.

[0123] For example, in the scenario of coating paper substrate, the glue volume sensor collects glue volume data of 500 ml per minute, and the thickness sensor collects thickness data of 20 micrometers. After inputting into the support vector machine model, the model extracts features such as the average glue volume of 480 ml and the thickness fluctuation of 2 micrometers. Through mapping processing, the glue volume control parameter is output as increasing the rotation speed by 10%, and the thickness optimization parameter is reducing the pressure by 5%. This can improve the coating uniformity by 15% and reduce waste.

[0124] In one embodiment, for high-speed coating scenarios, the machine learning model can be switched to a neural network algorithm, which includes an input layer, a hidden layer, and an output layer. First, the adhesive volume data and thickness data are normalized. Then, feature extraction is performed in the hidden layer, such as calculating the Fourier transform of the adhesive volume data to capture periodic changes. Then, it is mapped to the output layer to generate parameters. In this way, when the substrate speed is 100 meters per minute, more accurate control parameters can be output, improving the response speed.

[0125] For example, assuming the adhesive volume is 600 ml per minute and the thickness is 25 micrometers, the neural network model extracts features and maps the output adhesive volume control parameters to reduce the rotation speed by 8% and the thickness optimization parameters to increase the pressure by 3%, thereby controlling the thickness deviation within 1 micrometer, which is beneficial for high-precision coating applications.

[0126] In one embodiment, for coating adhesives of different viscosities, a viscosity variable is added during model training. Feature extraction includes calculating the correlation coefficient between adhesive volume data and viscosity, and the mapping process uses a regression method to predict parameter adjustments, thus covering low-viscosity to high-viscosity scenarios.

[0127] For example, in low-viscosity adhesive coating, the collected data were 400 ml per minute and 15 micrometers thickness. After adjusting the model output parameters, the coating efficiency was improved by 20%, and the adhesive splatter problem was reduced.

[0128] S108. The step of fine-tuning the coating equipment according to the adhesive quantity control parameters and the thickness optimization parameters, obtaining the fine-tuned equipment status data, determining whether the equipment status data exceeds a preset threshold, and if it does, recalculating the control parameters using an evolutionary algorithm to determine the optimized equipment configuration includes: transmitting the adhesive quantity control parameters and the thickness optimization parameters to the control unit of the coating equipment to perform parameter fine-tuning operations, collecting the operating status of the coating equipment after fine-tuning to form the equipment status data, comparing the equipment status data with a preset threshold, and if the equipment status data exceeds the preset threshold, starting an evolutionary algorithm to iteratively optimize the control parameters, and obtaining the optimized equipment configuration through multiple generations of evolutionary calculations.

[0129] Step S1: The adhesive amount control parameters and the thickness optimization parameters are transmitted to the control unit of the coating equipment to perform parameter fine-tuning.

[0130] In one embodiment, the adhesive quantity control parameters include the adhesive supply rate and distribution uniformity values, while the thickness optimization parameters include the target thickness range and deviation correction coefficient. These parameters are transmitted to the control unit via a data interface, and the control unit adjusts the coating roller speed and adhesive pump pressure based on the parameters to achieve fine-tuning.

[0131] Step S2: Collect the operating status of the coating equipment after fine-tuning to form the equipment status data.

[0132] The data acquisition process uses built-in sensors to monitor the equipment's vibration frequency and adhesive flow stability in real time, generating equipment status data such as vibration amplitude and adhesive flow fluctuation rate.

[0133] Step S3: Compare the device status data with a preset threshold.

[0134] For example, the vibration amplitude value is compared with a preset threshold of 10 Hz, and the glue flow fluctuation rate is compared with a preset threshold of 5%.

[0135] Step S4: If the device status data exceeds the preset threshold, an evolutionary algorithm is started to iteratively optimize the control parameters.

[0136] In one embodiment, the evolutionary algorithm refers to a genetic algorithm, which optimizes parameters by simulating the natural selection process. Step S41: Initialize the population, including randomly generating multiple sets of control parameters, such as the glue supply rate from an initial value of 20 ml / min to a mutated value of 25 ml / min. Step S42: Calculate the fitness function, defined as minimizing the deviation between the equipment state data and a threshold, for example, fitness value = 1 / (vibration amplitude - threshold)^2. Step S43: Perform selection, crossover, and mutation operations, selecting individuals with high fitness, exchanging parameter components during crossover, such as the exchange rate and deviation coefficient, and randomly adjusting values ​​during mutation. Step S44: Iterate through multiple generations until the fitness converges. This evolutionary algorithm effectively handles nonlinear changes in the coating process and improves optimization efficiency.

[0137] Step S5: Obtain the optimized device configuration through multi-generational evolutionary calculations.

[0138] For example, after 10 generations of evolution, a configuration with a glue supply rate of 22 ml / min and a thickness deviation correction factor of 0.8 is obtained, ensuring that the equipment status data does not exceed the threshold.

[0139] In one embodiment, for high-viscosity adhesive coating scenarios, the population size of the evolutionary algorithm in step S4 is set to 50, and the number of iterations is 15. This can handle the problem of unstable adhesive flow and bring beneficial effects such as reducing waste by 5%. In another embodiment, in low-speed coating scenarios, the fitness function is weighted by thickness uniformity. The optimized configuration improves production efficiency by 10%. These aspects support the robustness of the algorithm under different parameters, forming a consistent optimization argument.

[0140] S109. Obtaining real-time monitoring signals from the optimized equipment configuration, transmitting the real-time monitoring signals for environmental factors during the substrate coating process, and obtaining environmental data includes: extracting real-time monitoring signals associated with the substrate coating process from the optimized equipment configuration, transmitting the real-time monitoring signals to the control system, and the control system collecting humidity and temperature values ​​during the substrate coating process based on the real-time monitoring signals and integrating them to form the environmental data.

[0141] In one embodiment, step S3 obtains real-time monitoring signals from the optimized equipment configuration, transmits the real-time monitoring signals for environmental factors during the substrate coating process, and obtains environmental data. Specifically, step S31 extracts real-time monitoring signals associated with the substrate coating process from the optimized equipment configuration.

[0142] Step S311: Identify the signal sources in the optimized equipment configuration and select monitoring points related to substrate coating, such as locating the output interfaces of the adhesive quantity sensor and thickness sensor in the coating equipment.

[0143] Step S312: Extract real-time signal data based on the selected monitoring points to ensure that the signal is correlated with environmental factors such as humidity and temperature.

[0144] Specifically, this implementation can automatically extract signals during the operation of the coating equipment, process the previously obtained adhesive amount control parameters and thickness optimization parameters through machine learning algorithms, form an optimized configuration, and directly extract signals from the configuration, avoiding manual intervention and improving real-time performance.

[0145] In one embodiment, step S32 involves transmitting the real-time monitoring signal to the control system.

[0146] Specifically, the signal is transmitted wirelessly to ensure that the transmission delay is lower than the preset value.

[0147] For example, in a glass substrate coating scenario, the control system responds immediately after signal transmission to maintain a stable coating process.

[0148] In one embodiment, in step S33, the control system collects the humidity and temperature values ​​during the substrate coating process based on the real-time monitoring signal.

[0149] Step S331: Analyze the real-time monitoring signal and activate the humidity sensor and temperature sensor to collect data.

[0150] Step S332: Record the collected humidity and temperature values, for example, the humidity range is 40% to 60% and the temperature range is 20℃ to 30℃.

[0151] Step S333: Verify the accuracy of the collected data. If the deviation exceeds the threshold, recollect the data.

[0152] Specifically, this implementation uses an evolutionary algorithm to optimize the device configuration, ensuring that the data acquisition process is linked with the previously fine-tuned device status data. For example, when the device status data exceeds a threshold, the acquisition prioritizes areas with high humidity adjustment requirements.

[0153] For example, in thin film substrate coating applications, the control system collects humidity values ​​of 45% and temperature values ​​of 25°C based on signals. This helps determine whether the environment is suitable for optimizing the coating thickness, bringing beneficial effects such as reducing coating defects and improving product uniformity.

[0154] In one embodiment, the data acquisition process in step S33 is illustrated from several perspectives. First, in principle, the acquisition is based on real-time signal-triggered sensors. The sensors convert environmental factors into numerical data through electrical signals, avoiding human error. Second, in terms of specific values, assuming the substrate coating rate is 5 meters per minute, the acquired humidity value is 50% and the temperature value is 22°C. The analysis process shows that when the humidity is higher than the temperature, the humidity is adjusted first, resulting in technical effects such as improving coating efficiency by 10%. Third, from a business perspective, the equipment has been fine-tuned before acquisition, and the acquired data is used for subsequent analysis. If the humidity value is too high, it may lead to uneven glue application, resulting in an increased product scrap rate. Therefore, timely acquisition can prevent such problems, forming a tight logical chain.

[0155] In one embodiment, step S34 involves integrating the environmental data to form the environmental data.

[0156] Specifically, the collected humidity and temperature values ​​are combined into a data package and output as environmental data.

[0157] For example, in paper substrate coating, integrated environmental data is used to determine adjustment needs and optimize the overall process.

[0158] S1010. The step of using a machine learning algorithm to analyze the environmental data and determine the relationship between humidity adjustment needs and temperature control needs, and prioritizing the determination of humidity adjustment instructions if the humidity adjustment needs are higher than the temperature control needs, includes: inputting the environmental data into a machine learning analysis model, wherein the machine learning analysis model performs importance assessment and demand intensity calculation on the humidity and temperature components in the environmental data to obtain humidity adjustment demand indicators and temperature control demand indicators, and when the humidity adjustment demand indicator is higher than the temperature control demand indicator, prioritizing the generation of humidity adjustment instructions.

[0159] In one embodiment, the step of using a machine learning algorithm to analyze the environmental data and determine the relationship between humidity adjustment needs and temperature control needs, and if the humidity adjustment needs are higher than the temperature control needs, then the humidity adjustment command is determined first, including the following steps.

[0160] Step S1: Input the environmental data into the machine learning analysis model.

[0161] Specifically, environmental data obtained from real-time monitoring signals acquired from the optimized equipment configuration and transmitted to the control system is directly input into a pre-trained machine learning analysis model. This model is built based on the Support Vector Machine (SVM) algorithm, a supervised learning method that classifies or regresses data by finding the optimal separating hyperplane between data points; here, it is used to handle the multidimensional features of the environmental data.

[0162] Step S2: The machine learning analysis model assesses the importance and calculates the demand intensity of the humidity and temperature components in the environmental data.

[0163] In one possible implementation, step S2 specifically includes: step S21, extracting the humidity component and temperature component from the environmental data, where the humidity component refers to the numerical sequence related to humidity in the environmental data, such as the relative humidity percentage, and the temperature component refers to the numerical sequence related to temperature, such as the Celsius value; step S22, using a support vector machine algorithm to perform importance assessment, which is achieved by calculating the feature weight of each component. For example, the weight of the humidity component is obtained by mapping it to a high-dimensional space through the kernel function in the support vector machine and then solving an optimization problem. Similarly, the weight of the temperature component is calculated; step S23, performing demand intensity calculation, which is based on the importance weight multiplied by the deviation value of the component, where the deviation value is the difference between the current component value and the preset optimal value. This calculation quantifies the adjustment urgency of each component.

[0164] For example, in a scenario where a coating machine processes paper-based substrates, environmental data shows a humidity component of 80% and a temperature component of 25°C. The preset optimal humidity is 60% and the optimal temperature is 22°C. The deviations are 20% and 3°C, respectively. If a support vector machine (SVM) evaluates humidity with a weight of 0.7 and temperature with a weight of 0.3, the humidity demand intensity is 20% * 0.7 = 14, and the temperature intensity is 3 * 0.3 = 0.9, thus highlighting the priority of humidity adjustment. The beneficial effect of this method is that it can accurately quantify the impact of environmental factors on coating thickness, avoiding wasted adhesive due to blind adjustments.

[0165] In another embodiment, for glass substrate coating applications, the support vector machine model can adjust the kernel function to a radial basis function to better capture nonlinear relationships. For example, when the humidity component in the environmental data fluctuates greatly, the importance assessment will increase its weight to 0.8, and the demand intensity calculation will consider historical data sequences and enhance prediction accuracy by calculating the moving average deviation. The beneficial effect is to improve the stability of the equipment in high humidity environments and reduce the problem of uneven thickness.

[0166] Step S3 yields the humidity adjustment requirement index and the temperature control requirement index.

[0167] Specifically, through the demand intensity calculation in step S2, the humidity adjustment demand index is directly output as 14 and the temperature control demand index is 0.9. These indicators are used as quantitative results for subsequent comparisons.

[0168] Step S4: When the humidity adjustment requirement index is higher than the temperature control requirement index, the humidity adjustment command for humidity is generated first.

[0169] For example, during the coating process of paper substrates, if the humidity index of 14 is higher than the temperature index of 0.9, an instruction such as "reduce humidity to 60%" will be generated, and the humidity adjustment will be executed first. The beneficial effect is that it can quickly respond to environmental changes, ensure that the coating thickness data is stable within the preset threshold, and avoid deviation of the adhesive volume control parameters due to excessive humidity.

[0170] In one embodiment, step S4 further includes: step S41, comparing the magnitudes of two indicators; if the humidity indicator is higher than the temperature indicator, then activating the humidity adjustment module; step S42, generating specific instructions based on the intensity value output by the support vector machine, such as calculating the adjustment range as a percentage of the deviation value, for example, 50% of 20% is a 10% reduction, thereby forming an instruction. The beneficial effect of this approach is to achieve automated priority management and improve the efficiency of the overall coating optimization process.

[0171] For example, in actual operation, for a case where the humidity requirement index is 15 and the temperature is 5, after the humidity command is generated first, the system can observe that the adhesive thickness data has returned to normal from exceeding the threshold, which proves the effectiveness of the judgment relationship, avoids the interference of small temperature fluctuations with the main adjustment, and enhances the stability of equipment status data.

[0172] S1011. Obtaining the current humidity value according to the humidity adjustment instruction, optimizing the adjustment path using an evolutionary algorithm, and obtaining a supplementary temperature control instruction includes: extracting the current humidity value from the substrate monitoring system according to the humidity adjustment instruction, inputting the current humidity value as an initial condition into the evolutionary algorithm, wherein the evolutionary algorithm iteratively optimizes the adjustment path through population initialization, fitness evaluation, crossover mutation, and selection operations, and synchronously generates the supplementary temperature control instruction for supplementary adjustment during the optimization process.

[0173] In one embodiment, the current humidity value is extracted from the substrate monitoring system according to the humidity adjustment command, and the current humidity value is used as an initial condition input to an evolutionary algorithm. The evolutionary algorithm iteratively optimizes the adjustment path through population initialization, fitness evaluation, crossover mutation, and selection operations, and simultaneously generates the temperature control supplementary command for supplementary adjustment during the optimization process. Specifically, this includes step S51, extracting the current humidity value from the substrate monitoring system according to the humidity adjustment command.

[0174] In step S511, after receiving the humidity adjustment command, the substrate monitoring system queries real-time sensor data and extracts the current humidity value. For example, during the operation of the coating equipment, the humidity value is directly read through the embedded sensor to ensure the real-time nature of the data.

[0175] Step S512: Record the extracted current humidity value as a numerical value for subsequent input.

[0176] In one embodiment, the extraction process can respond quickly to instructions, avoid uneven adhesive layer caused by environmental fluctuations during the coating process, and improve coating accuracy.

[0177] Step S52: Input the current humidity value as an initial condition into the evolutionary algorithm.

[0178] Step S521: Set the extracted humidity value as the starting parameter of the algorithm and initialize the input vector.

[0179] In one embodiment, this input ensures that the algorithm is based on the actual environment and optimizes for coating requirements that better fit the substrate.

[0180] Step S53: The evolutionary algorithm iteratively optimizes and adjusts the path through population initialization, fitness evaluation, crossover mutation, and selection operations.

[0181] Step S531: Initialize the population and generate an initial population, where each individual represents a possible adjustment path. The path parameters include the humidity change step size and time interval. The population size is set to 50 individuals. Based on the principle of genetic algorithm, which is an optimization method that simulates natural evolution, the optimal solution is searched by encoding the solution into chromosomes.

[0182] Step S532: Perform fitness evaluation. Calculate the fitness function for each individual. The fitness function is defined as minimizing the squared difference between the humidity deviation and the target value. For example, if the target humidity is 60% and the current humidity is 65%, then evaluate the deviation and quantify the path efficiency. Fitness evaluation in the genetic algorithm is a process of measuring the quality of individuals through a predefined function. In specific calculations, first simulate the humidity change curve for the path, and then calculate the integral difference between the curve and the target curve as a score.

[0183] In step S533, crossover and mutation operations are performed. High-fit individuals are crossovered to exchange path parameter fragments, and random mutation is introduced to increase diversity. The crossover probability is set to 0.8 and the mutation probability is 0.1. The crossover and mutation of the genetic algorithm is to simulate the biological genetic mechanism. Crossover is to exchange some genes of two parent individuals to generate offspring, while mutation is to randomly change gene values ​​to avoid local optima.

[0184] Step S534: Perform the selection operation, retaining individuals with high fitness to enter the next generation. Iterate for 10 generations until convergence. The selection operation of the genetic algorithm adopts the roulette wheel method, selecting individuals according to the fitness ratio to ensure that excellent paths are retained.

[0185] In one embodiment, the evolutionary algorithm is applied in a coating device. When the substrate is paper, the initial population takes into account the moisture absorption characteristics of paper, and the fitness evaluation incorporates a moisture absorption rate factor. For example, the humidity is optimized from 70% to 50%. After iteration, the path adjustment time is shortened by 20%, which helps to reduce glue waste.

[0186] In one embodiment, for the metal substrate scenario, the path parameters are adjusted to a smaller step size during population initialization, crossover mutation emphasizes temperature linkage, iterates for 15 generations, optimizes the path to stabilize the humidity at 55%, and improves the coating uniformity by 15%.

[0187] Step S54: During the optimization process, the temperature control supplementary command for supplementary adjustment is generated synchronously.

[0188] Step S541: In each iteration, based on the currently optimized humidity path, calculate the temperature compensation value and generate supplementary instructions. For example, if the humidity path shows a rapid decrease, the supplementary instruction is to increase the temperature by 2 degrees to balance the evaporation rate.

[0189] Step S542 integrates the generated instructions with the humidity path to ensure synchronized output.

[0190] In one embodiment, this synchronous generation prioritizes humidity followed by temperature supplementation in coating optimization, which helps maintain environmental stability and avoid thickness fluctuations.

[0191] For example, in the paper coating scenario, the extracted humidity value is 68%. After inputting it into the evolutionary algorithm, the population is initialized to generate path individuals. The fitness evaluation selects the path with the smallest deviation. After crossover and mutation, the best path is selected. The temperature supplement instruction is generated simultaneously to reduce the temperature by 1 degree. After optimization, the coating thickness deviation is reduced by 10%, which is beneficial to improving production efficiency.

[0192] For example, in another parameter setting, the humidity value is 62%. When iteratively optimizing the path, the fitness function takes into account the time cost and generates supplementary instructions to maintain a constant temperature, which is beneficial for energy saving.

[0193] In one embodiment, for a high initial humidity value such as 75%, the evolutionary algorithm generates a fine path through more iterations, with supplementary instructions including gradual temperature adjustment, which helps prevent substrate deformation.

[0194] S1012. The step of processing the coating optimization process according to the temperature control supplementary command and the environmental adjustment mechanism, determining whether the coating optimization index meets the preset threshold, generating operation optimization feedback if it meets the threshold, and determining the final adjustment scheme includes: combining the temperature control supplementary command with the environmental adjustment mechanism to execute the coating optimization process, monitoring the optimization index in real time during the coating process, comparing the optimization index with the preset threshold, generating the operation optimization feedback if the optimization index meets the preset threshold, and integrating various control results according to the operation optimization feedback to form the final adjustment scheme.

[0195] The coating optimization process is executed by combining the temperature control supplementary command with the environmental adjustment mechanism. The optimization indicators during the coating process are monitored in real time, and the optimization indicators are compared with preset thresholds. If the optimization indicators meet the preset thresholds, the operation optimization feedback is generated. The operation optimization feedback is used to integrate the various control results to form the final adjustment scheme.

[0196] In one embodiment, the coating optimization process is executed by combining the temperature control supplementary command with the environmental adjustment mechanism, including: step S61, inputting the temperature control supplementary command into the environmental adjustment mechanism, and adjusting the temperature parameters through the feedback loop of the environmental adjustment mechanism so that it acts synchronously with the humidity adjustment command on the coating equipment.

[0197] Specifically, the environmental adjustment mechanism refers to a closed-loop control system based on sensor data, used to correct the impact of environmental factors on the coating process in real time. For example, it uses a proportional-integral-derivative (PID) controller to handle temperature deviations and ensure the uniformity of the coating layer. Step S611: Read the target temperature value from the supplementary temperature control instruction and compare it with the current ambient temperature to calculate the deviation value. Step S612: Activate the heating or cooling module of the environmental adjustment mechanism based on the deviation value to perform preliminary temperature correction. Step S613: Monitor the impact of the corrected temperature on coating flowability and generate an execution signal to transmit to the coating optimization process.

[0198] Real-time monitoring of optimization metrics during the coating process includes continuously collecting data on coating thickness uniformity and adhesive distribution using optical sensors and flow meters installed on the coating equipment, which serve as optimization metrics.

[0199] The optimization index is compared with a preset threshold, including comparing the collected thickness uniformity data with a preset threshold such as 0.05 mm.

[0200] If the optimization index meets the preset threshold, the operation optimization feedback is generated, including generating a feedback signal indicating optimization if the thickness uniformity is less than 0.05 mm.

[0201] The final adjustment plan is formed by integrating various control results based on the operation optimization feedback, including: step S65, summarizing the glue volume control parameters, thickness optimization parameters and environmental data based on the operation optimization feedback, and integrating them into a unified plan.

[0202] Specifically, the fusion process uses a weighted average method to calculate the contribution of each parameter. For example, the weight of the adhesive volume control parameter is 0.4, the weight of the thickness optimization parameter is 0.3, and the weight of environmental data is 0.3, ensuring that the solution comprehensively covers the coating process. Step S651: Extract the success indicators from the operation optimization feedback and map them to various control results. Step S652: Apply the weighted average to calculate the integrated value and form a draft adjustment plan. Step S653: Verify whether the draft covers all environmental factors; if so, confirm the final adjustment plan.

[0203] In one possible implementation, the coating optimization process is executed by combining temperature control supplementary instructions with an environmental adjustment mechanism. In the lithium battery electrode coating scenario, the temperature control supplementary instructions are set to increase by 2 degrees Celsius, and the environmental adjustment mechanism regulates airflow through a fan. After the process is executed, coating optimization indicators such as thickness deviation are reduced by 20%, which is beneficial to improving battery performance stability.

[0204] Specifically, the supplementary instructions first correct the temperature, and then the mechanism adjusts the humidity to ensure that the viscosity of the adhesive is appropriate and to avoid the formation of air bubbles.

[0205] For example, in the coating of solar cell substrates, the temperature control supplementary instruction is to reduce the temperature by 1 degree Celsius. Combined with the dehumidification function of the environmental adjustment mechanism, the optimized process is executed. Monitoring indicators show that the uniformity is improved by 15%, thereby reducing the defect rate and benefiting the photoelectric conversion efficiency.

[0206] In one embodiment, the final adjustment scheme is formed by integrating various control results based on the operation optimization feedback. In the application of electronic circuit board coating, after the feedback shows that the optimization index meets the threshold, the adhesive amount parameter is integrated as 50 grams per square meter, the thickness parameter is 0.1 mm, and the environmental data is 40% humidity, forming a scheme that is beneficial to the circuit insulation performance.

[0207] It should be noted that the integration process ensures that the results are complementary, such as adjusting the amount of adhesive to compensate for thickness changes, thus avoiding fragmentation of the solution.

[0208] For example, if the feedback indicates a slight deviation, environmental data is prioritized during integration. The humidity in the solution is adjusted to 35%, and combined with temperature control, this ensures that the coating adhesion is enhanced by 10%, which is beneficial for long-term durability.

[0209] S1013. Obtaining micro-control updates from the final adjustment scheme, transmitting them to the intelligent optimization network, and obtaining overall operation optimization results includes: extracting micro-control updates for equipment and environment from the final adjustment scheme, transmitting the micro-control updates to the intelligent optimization network, the intelligent optimization network performing global integration and verification processing on the micro-control updates, and outputting the overall operation optimization results characterizing the overall coating process optimization.

[0210] Micro-control updates are obtained from the final adjustment scheme and transmitted to the intelligent optimization network to obtain the overall operation optimization result.

[0211] In one embodiment, micro-control updates for the device and environment are extracted from the final adjustment scheme, and the micro-control updates are transmitted to the intelligent optimization network. The intelligent optimization network performs global integration and verification processing on the micro-control updates, and outputs the overall operation optimization result characterizing the overall coating process optimization. Specifically, this includes step S1, extracting micro-control updates for the device and environment from the final adjustment scheme.

[0212] Step S11: By parsing the data fields in the final adjustment plan, identify equipment control parameters such as adhesive quantity control parameters and thickness optimization parameters, as well as environmental control parameters such as humidity adjustment instructions and temperature control supplementary instructions. These parameters constitute micro-control updates.

[0213] Step S12: Package the extracted micro-control updates into a standardized data packet for easy subsequent transmission.

[0214] For example, when the coating equipment processes glass substrates, the final adjustment scheme sets the adhesive amount control parameter to 2.5 grams per square meter, the thickness optimization parameter to 15 micrometers, the humidity adjustment instruction to reduce to 45%, and the temperature control supplementary instruction to maintain at 22 degrees Celsius. These updates are extracted to form micro-control updates, ensuring precise matching of equipment and environmental parameters.

[0215] In one embodiment, the extraction process in step S1 can bring beneficial effects, such as improving the accuracy of the coating process and avoiding uneven coating caused by parameter omissions.

[0216] Step S2: The micro-control update is transmitted to the intelligent optimization network.

[0217] Specifically, data packets are sent to network nodes via a wireless communication module to achieve real-time transmission.

[0218] Step S3: The intelligent optimization network performs global integration and verification processing on the micro-control update.

[0219] Step S31: Use a convolutional neural network to process the device parameters and environmental parameters in the micro-control update. The convolutional neural network extracts features through multi-layer convolution and pooling operations, such as feature fusion of adhesive amount control parameters and humidity adjustment instructions to form an integrated global parameter vector.

[0220] Step S32: The integrated global parameter vector is validated using a support vector machine. The support vector machine uses a classification hyperplane to determine whether the parameters meet the optimization criteria, such as checking whether the glue volume control parameter is within the preset range. If it exceeds the range, it is marked as invalid.

[0221] Step S33: If the verification passes, a verification report is generated, recording the accuracy of the integration process, for example, an accuracy of 95% in a simulation test.

[0222] Step S34: In the actual coating scenario, adjust the learning rate of the network to 0.01 and iterate the training 10 times to optimize the integration effect.

[0223] For example, in the coating process of plastic substrates, the micro-control update includes the glue amount control parameter of 3.0 grams per square meter, the thickness optimization parameter of 20 micrometers, the humidity adjustment instruction of increasing to 55%, and the temperature control supplementary instruction of adjusting to 25 degrees Celsius. The spatiotemporal features of these parameters are extracted by convolutional neural network, and their stability is verified by support vector machine. The results show that the parameter combination can reduce the coating defect rate by 15%, which is beneficial to improving the overall efficiency.

[0224] In one embodiment, the global integration and verification process in step S3 ensures the robustness of the coating process and reduces errors in high humidity environments through feature extraction from neural networks and judgment by machine learning classifiers.

[0225] Specifically, in another scenario, for coating of metal substrates, the adhesive amount control parameter in the micro-control update is 2.8 grams per square meter, the thickness optimization parameter is 18 micrometers, the humidity adjustment instruction is to maintain 50%, and the temperature control supplementary instruction is to reduce to 20 degrees Celsius; during integration, the convolutional neural network processes the correlation between parameters, and the support vector machine verifies and outputs the optimized vector, which has the beneficial effect of improving coating uniformity by 20%.

[0226] Step S4: Output the overall operation optimization results that characterize the overall coating process optimization.

[0227] Specifically, the validated global parameter vector is converted into an executable optimization result file for use by the coating equipment.

[0228] For example, in glass substrate optimization, the overall operational optimization results include an adjusted adhesive amount of 2.5 grams and a humidity of 45%, achieving comprehensive optimization of the coating process.

[0229] If the technical solution of this application involves the processing of personal information, the relevant products have established a sound user authorization mechanism: before collecting, using, or sharing personal information, the obligation to inform is fulfilled in accordance with the law, and the individual's voluntary and explicit consent is obtained; if sensitive personal information is involved, the user's separate and explicit consent is further obtained. Specific measures include, but are not limited to: setting up prominent prompts in the information collection area, or clearly displaying the processing rules (including the processor, purpose, method, information type, etc.) through electronic interfaces such as pop-ups, checkboxes, and active submissions, to ensure that users voluntarily authorize based on their knowledge. All personal information processing activities strictly comply with national laws and regulations, especially the relevant provisions of the "Personal Information Protection Law of the People's Republic of China," to effectively safeguard the legitimate rights and interests of personal information subjects. Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for controlling the amount of adhesive used in corrugated cardboard bonding based on industrial manual system optimization, characterized in that, include: Acquire adhesive amount data and adhesive thickness data from the coating equipment, and process the adhesive amount data and adhesive thickness data using machine learning algorithms to obtain adhesive amount control parameters and thickness optimization parameters; The coating equipment is fine-tuned according to the adhesive volume control parameters and the thickness optimization parameters. The equipment status data after fine-tuning is obtained. It is determined whether the equipment status data exceeds the preset threshold. If it does, the control parameters are recalculated using an evolutionary algorithm to determine the optimized equipment configuration. Real-time monitoring signals are obtained from the optimized equipment configuration, and environmental data is obtained by transmitting the real-time monitoring signals for environmental factors during the substrate coating process. The environmental data is analyzed using machine learning algorithms to determine the relationship between humidity adjustment needs and temperature control needs. If the humidity adjustment needs are higher than the temperature control needs, the humidity adjustment command is given priority. The current humidity value is obtained according to the humidity adjustment command, and the adjustment path is optimized using an evolutionary algorithm to obtain the temperature control supplementary command; The coating optimization process is processed according to the temperature control supplementary instruction and the environmental adjustment mechanism. It is determined whether the coating optimization index meets the preset threshold. If it does, operation optimization feedback is generated to determine the final adjustment plan. Micro-control updates are obtained from the final adjustment scheme and transmitted to the intelligent optimization network to obtain the overall operation optimization result.

2. The method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization as described in claim 1, characterized in that, The process of acquiring adhesive volume data and coating thickness data from the coating equipment, and then processing the adhesive volume data and coating thickness data using machine learning algorithms to obtain adhesive volume control parameters and thickness optimization parameters includes: The adhesive quantity data is collected in real time by the adhesive quantity sensor of the coating equipment, and the adhesive thickness data is collected in real time by the thickness sensor. The adhesive quantity data and the adhesive thickness data are input into a pre-trained machine learning model. The machine learning model performs feature extraction and mapping processing on the adhesive quantity data and the adhesive thickness data, and outputs the adhesive quantity control parameters for equipment control and the thickness optimization parameters for thickness adjustment.

3. The method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization as described in claim 1, characterized in that, The process of fine-tuning the coating equipment based on the adhesive dosage control parameters and the thickness optimization parameters, obtaining the fine-tuned equipment status data, determining whether the equipment status data exceeds a preset threshold, and if so, recalculating the control parameters using an evolutionary algorithm to determine the optimized equipment configuration includes: The adhesive volume control parameters and the thickness optimization parameters are transmitted to the control unit of the coating equipment to perform parameter fine-tuning. The operating status of the coating equipment after fine-tuning is collected to form the equipment status data. The equipment status data is compared with a preset threshold. If the equipment status data exceeds the preset threshold, an evolutionary algorithm is started to iteratively optimize the control parameters. The optimized equipment configuration is obtained through multi-generation evolutionary calculations.

4. The method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization as described in claim 1, characterized in that, The step of acquiring real-time monitoring signals from the optimized equipment configuration, transmitting the real-time monitoring signals for environmental factors during the substrate coating process, and obtaining environmental data includes: The optimized equipment configuration extracts real-time monitoring signals associated with the substrate coating process, and transmits the real-time monitoring signals to the control system. The control system collects humidity and temperature values ​​during the substrate coating process based on the real-time monitoring signals and integrates them to form the environmental data.

5. The method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization as described in claim 1, characterized in that, The environmental data is analyzed using machine learning algorithms to determine the relationship between humidity adjustment needs and temperature control needs. If the humidity adjustment needs are higher than the temperature control needs, a humidity adjustment command is prioritized, including: The environmental data is input into a machine learning analysis model, which assesses the importance and calculates the demand intensity of the humidity and temperature components in the environmental data to obtain humidity adjustment demand index and temperature control demand index. When the humidity adjustment demand index is higher than the temperature control demand index, the humidity adjustment instruction is generated first.

6. The method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization as described in claim 1, characterized in that, The step involves obtaining the current humidity value based on the humidity adjustment command, optimizing the adjustment path using an evolutionary algorithm, and obtaining a supplementary temperature control command, including: According to the humidity adjustment command, the current humidity value is extracted from the substrate monitoring system and used as the initial condition to input the evolutionary algorithm. The evolutionary algorithm iteratively optimizes the adjustment path through population initialization, fitness evaluation, crossover mutation and selection operations. During the optimization process, the temperature control supplementary command for supplementary adjustment is generated simultaneously.

7. The method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization as described in claim 1, characterized in that, The coating optimization process, based on the temperature control supplementary command and environmental adjustment mechanism, determines whether the coating optimization indicators meet preset thresholds. If they do, operation optimization feedback is generated to determine the final adjustment plan, including: The coating optimization process is executed by combining the temperature control supplementary command with the environmental adjustment mechanism. The optimization indicators during the coating process are monitored in real time, and the optimization indicators are compared with preset thresholds. If the optimization indicators meet the preset thresholds, the operation optimization feedback is generated. The operation optimization feedback is used to integrate the various control results to form the final adjustment scheme.

8. The method for controlling the amount of adhesive used in corrugated cardboard based on industrial manual system optimization as described in claim 1, characterized in that, The step of obtaining micro-control updates from the final adjustment scheme, transmitting them to the intelligent optimization network, and obtaining the overall operation optimization result includes: Micro-control updates for equipment and environment are extracted from the final adjustment scheme, and the micro-control updates are transmitted to the intelligent optimization network. The intelligent optimization network performs global integration and verification of the micro-control updates and outputs the overall operation optimization result characterizing the overall coating process optimization.

Images