Method and system for optimizing blade coating with pressure control

Abstract

The application provides a doctor blade coating optimization method and system combined with pressure control, relates to the field of automatic control technology, generates a dynamic pressure distribution map through real-time pressure sensing based on a coating doctor blade; performs pressure deviation calculation according to the dynamic pressure distribution map, generates a multi-dimensional control instruction; performs coating through a multi-dimensional control instruction driving coating doctor blade execution mechanism, obtains a doctor blade coating parameter set; performs data feedback based on the doctor blade coating parameter set, formulates a pressure control strategy, performs coating optimization on the multi-dimensional control instruction through the pressure control strategy, and generates a coating optimization report. The application solves the technical problems that, due to the lack of real-time sensing and closed-loop control mechanism in the prior art, the coating doctor blade pressure regulation lags, the coating layer thickness is uneven, and the consistency is poor, achieves the technical effects of enhancing the adaptability of the coating process to material changes and process disturbances and improving the coating layer thickness uniformity and overall coating stability.

Description

Technical Field

[0001] This application relates to the field of automatic control technology, specifically to a method and system for optimizing scraper coating by incorporating pressure control. Background Technology

[0002] In modern industrial production, coating (or application) is a crucial step in product surface treatment. It involves using equipment such as doctor blades and rollers to evenly apply coating materials to the substrate surface, forming a coating structure that meets requirements for thickness, uniformity, and adhesion. Doctor blade coating, due to its advantages of simple structure, flexible operation, and strong adaptability, is widely used in high-end manufacturing scenarios such as thin films, lithium battery separators, and optical films. In existing technologies, the pressure adjustment of the coating doctor blade mainly relies on static preset parameters, using mechanical spring loading or single-point air pressure for pressure control. This lacks the dynamic sensing capability of the doctor blade's real-time pressure status and lacks an effective feedback mechanism. It cannot flexibly respond to changes in process disturbances, substrate unevenness, or coating rheological properties, leading to lag in pressure adjustment, insufficient control precision, and consequently, uneven coating thickness, resulting in defects such as edge buildup or thinning in the center, affecting coating consistency and yield. Summary of the Invention

[0003] This application provides a method and system for optimizing blade coating by combining pressure control. It solves the technical problem in the prior art that the lack of real-time sensing and closed-loop control mechanism leads to lag in blade pressure adjustment, resulting in uneven coating thickness and poor coating consistency. It achieves the technical effect of enhancing the coating process's adaptability to material changes and process disturbances, and improving coating thickness uniformity and overall coating stability.

[0004] In view of the above problems, this application provides a method for optimizing blade coating by combining pressure control. The method includes: real-time pressure sensing based on the coating blade to generate a dynamic pressure distribution map; calculating pressure deviation according to the dynamic pressure distribution map to generate multi-dimensional control commands; driving the coating blade execution mechanism to perform coating through the multi-dimensional control commands to obtain a blade coating parameter set; performing data feedback based on the blade coating parameter set to formulate a pressure control strategy; optimizing the coating through the multi-dimensional control commands using the pressure control strategy; and generating a coating optimization report.

[0005] On the other hand, this application also provides a doctor blade coating optimization system combined with pressure control. The system includes: a pressure sensing module for real-time pressure sensing based on the doctor blade and generating a dynamic pressure distribution map; a control command generation module for calculating pressure deviation according to the dynamic pressure distribution map and generating multi-dimensional control commands; a coating control module for driving the doctor blade execution mechanism to perform coating through the multi-dimensional control commands and obtaining a doctor blade coating parameter set; and a coating optimization module for performing data feedback based on the doctor blade coating parameter set, formulating a pressure control strategy, optimizing the coating through the multi-dimensional control commands using the pressure control strategy, and generating a coating optimization report.

[0006] One or more technical solutions provided in this application have at least the following beneficial effects: By real-time pressure sensing of the coating blade, a dynamic pressure distribution map is generated to perceive the uniformity of force under the current coating state, providing a data foundation for subsequent control. Pressure deviation calculation is performed, comparing the target pressure with the actual distribution to identify local deviations and generate corresponding multi-dimensional control commands for fine-grained adjustment. These multi-dimensional control commands drive the coating blade execution mechanism for coating, while simultaneously collecting key parameters during execution to form a blade coating parameter set for subsequent feedback learning. Based on data feedback from the blade coating parameter set, pressure control strategies are formulated, coating optimization is performed on the multi-dimensional control commands, and a coating optimization report is generated. This achieves experience accumulation and a control closed loop, continuously improving the coating effect and ensuring the stability and consistency of the coating process.

[0007] In summary, this application constructs a coating blade control method centered on real-time perception, multi-dimensional decision-making, and closed-loop optimization. This method enables real-time monitoring and precise control of blade pressure. Furthermore, by introducing execution feedback and strategy correction mechanisms, it enhances the coating process's adaptability to material changes and process disturbances, significantly improving the uniformity of coating thickness and the overall stability of coating. This provides a solution for high-quality coating in modern industry.

[0008] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0009] Figure 1 This is a schematic flowchart of a pressure-controlled scraper coating optimization method provided in an embodiment of this application.

[0010] Figure 2 This is a flowchart illustrating the construction of multi-dimensional control commands in the pressure-controlled scraper coating optimization method provided in this application embodiment.

[0011] Figure 3 A schematic diagram of the structure of the blade coating optimization system combined with pressure control provided in the embodiments of this application.

[0012] Explanation of reference numerals in the attached drawings: Pressure sensing module 10, control command generation module 20, coating control module 30, coating optimization module 40. Detailed Implementation

[0013] This application provides a method and system for optimizing doctor blade coating by combining pressure control. This solves the technical problem in the prior art where the lack of real-time sensing and closed-loop control mechanisms leads to lag in doctor blade pressure adjustment, resulting in uneven coating thickness and poor coating consistency. The method and system enhance the coating process's adaptability to material changes and process disturbances, and improve coating thickness uniformity and overall coating stability.

[0014] Example 1, as Figure 1 As shown in the embodiments of this application, a method for optimizing scraper coating combined with pressure control is provided, the method comprising: Step S1: Real-time pressure sensing based on the coating blade to generate a dynamic pressure distribution map.

[0015] Specifically, multiple sets of thin-film pressure sensors or piezoelectric strain sensors are arranged along the contact edge of the scraper installation area. These sensors can acquire the contact pressure value per unit area at each location in real time. The sensor data is transmitted to the embedded control system or host computer software (such as LabVIEW or MATLAB) via a data acquisition card (DAQ) to form a two-dimensional pressure distribution map, enabling real-time, visualized pressure monitoring during the scraper coating process. This dynamic pressure distribution map is a visual representation of the spatial distribution of multiple pressure points on the scraper's working surface, providing accurate data for subsequent pressure compensation and control optimization, helping to identify pressure anomalies in advance and improve the foresight of control response.

[0016] Step S2: Calculate the pressure deviation according to the dynamic pressure distribution map and generate multi-dimensional control commands.

[0017] Specifically, the pressure map is converted into a pressure matrix, the deviation between the actual pressure and the ideal applied pressure at each monitoring point is calculated, and based on the pressure deviation calculation results, combined with the requirements of the coating process and empirical rules, a fuzzy control algorithm or a multi-parameter PID adjustment model is used to generate multi-dimensional control commands to dynamically adjust the scraper angle, pressure parameters and coating speed.

[0018] Step S3: Drive the coating blade execution mechanism to perform coating by the multi-dimensional control command to obtain the blade coating parameter set.

[0019] Specifically, multi-dimensional control commands are sent to the coating blade execution mechanism, driving the coating blade execution mechanism (such as servo motors, angle adjusters, and speed controllers) to perform coating. During the coating process, sensors (such as thickness sensors and speed sensors) are continuously synchronized with the controller, recording each adjustment behavior and real-time coating status, and generating a structured set of coating parameters, such as coating thickness, pressure distribution, and coating speed.

[0020] Step S4: Based on the scraper coating parameter set, data feedback is performed to formulate a pressure control strategy. The multi-dimensional control commands are then optimized using the pressure control strategy to generate a coating optimization report.

[0021] Specifically, the data from the doctor blade coating parameter set is fed back to the control system for analysis via a communication interface (such as an RS-485 interface or an Ethernet interface). The control system, based on a pre-set optimization algorithm, learns from a large amount of coating data to formulate a pressure control strategy that optimizes performance under current conditions. Then, this pressure control strategy is used to adjust the parameters in the multi-dimensional control commands and send them to the doctor blade execution mechanism for coating optimization. The optimized coating parameter set is collected, a coating optimization report is generated, and the coating performance is recorded.

[0022] The above steps achieve continuous optimization of the coating process through data feedback, adjust the pressure control strategy according to the actual coating situation, and then optimize multi-dimensional control commands to improve coating uniformity and coating stability. The generated coating optimization report provides a reference for subsequent similar coating projects and helps to continuously improve the coating process.

[0023] Furthermore, step S1 includes: Step S11: Arrange a pressure sensor array at equal intervals along the axial direction on the coating blade body.

[0024] Step S12: The pressure sensor array is used to sense the contact pressure between the coating blade and the coating substrate to obtain multiple local pressure values.

[0025] Step S13: Spatial identification of the pressure sensor array based on the multiple local pressure values, and determination of multiple spatial coordinates.

[0026] Step S14: Map the multiple spatial coordinates onto the surface of the coated substrate for dynamic identification, and generate the dynamic pressure distribution map.

[0027] Specifically, multiple equidistant positions are determined along the axial direction of the coating blade body. A pressure sensor, such as a miniature piezoelectric sensor or strain gauge sensor, is installed at each position, ensuring a tight fit between the sensor and the blade body to form a pressure sensor array. For example, for a 120mm long coating blade, eight pressure sensors are arranged axially at 15mm intervals and connected to a data acquisition module via wires or a flexible PCB. This enables real-time sensing and coverage of the pressure application interface on the blade, providing hardware support for subsequent assessment of pressure distribution and the construction of spatial images.

[0028] During the coating process, the pressure sensor array is activated to monitor the contact pressure between the coating blade and the substrate in real time. Each pressure sensor collects the local pressure value at its location and transmits the data to the data acquisition module. The data acquisition module records the local pressure value of each sensor in real time, achieving high-precision quantification of local pressure differences and providing a key raw data foundation for building a refined coating control model.

[0029] Based on the known sensor spacing and arrangement order, a spatial location index table is constructed to assign physical location coordinates to the data points corresponding to each sensor, facilitating subsequent visualization and control mapping. For example, the first sensor is assigned x=0mm, the second x=15mm, and so on. A coordinate matching algorithm (such as based on array indexing or hash tables) is used to bind each pressure value to its corresponding physical location. For example, eight sensors simultaneously collect pressure values, with the following results: (0.9, 1.0, 1.2, 1.4, 1.3, 1.1, 1.0, 0.8), in Newtons (N). These pressure values ​​are bound to locations x=(0, 15, 30, ..., 105) in millimeters (mm), forming (xi, Pi) pairs: (0, 0.9), (15, 1.0), ..., (105, 0.8).

[0030] The sensor coordinates are converted to their positions in the substrate coordinate system, that is, the sensing points on the scraper are projected onto the moving coating material. As the substrate moves and the pressure changes, the mapping relationship and pressure state are updated in real time. A dynamic pressure distribution map is generated by using Python, Matplotlib, or LabVIEW graphics controls based on the real-time mapping relationship and pressure state. This dynamic pressure distribution map can be in the form of a two-dimensional heat map or contour map, with the horizontal axis representing the scraper width and the vertical axis representing the pressure value (or color mapping).

[0031] The above steps involve deploying a pressure sensor array along the doctor blade axis to collect multi-point contact pressure data, thus constructing a complete perception chain from raw signals to spatial coordinates and then to image distribution. The resulting dynamic pressure distribution map not only has real-time performance and spatial resolution, but can also be used to drive subsequent control mechanisms to achieve feedforward control and local compensation in the coating process.

[0032] Furthermore, such as Figure 2 As shown, step S2 includes: Step S21: Traverse the dynamic pressure distribution map to extract the pressure distribution matrix, and divide the pressure distribution matrix into multiple pressure sub-regions.

[0033] Step S22: Determine a preset pressure threshold, calculate the deviation of the multiple pressure sub-regions according to the preset pressure threshold, calculate the deviation of the multiple local pressure values ​​from the preset pressure threshold, and generate a local pressure deviation set.

[0034] Step S23: Perform gradient analysis based on the local pressure deviation set to identify and determine the pressure anomaly region, and mark the pressure anomaly region in combination with the multiple spatial coordinates to determine the pressure anomaly spatial coordinates.

[0035] Step S24: Determine the magnitude of the pressure anomaly region based on the set of local pressure deviations, determine the morphological characteristics of the pressure anomaly region based on the spatial coordinates of the pressure anomaly, and construct the multi-dimensional control command.

[0036] Specifically, the values ​​in the dynamic pressure distribution map are represented in matrix form to obtain a pressure distribution matrix. Each element of this matrix represents the pressure value at a certain spatial location, forming a two-dimensional array such as P(i,j) to represent the pressure distribution between the coating blade and the coating substrate. The pressure distribution matrix is ​​divided into multiple pressure sub-regions according to pre-defined division rules (such as equal area or equal spacing), with each sub-region containing multiple local pressure values. Simultaneously, a unique identifier is assigned to each pressure sub-region for subsequent processing.

[0037] A target pressure value is set as a preset pressure threshold according to the coating process requirements to assess whether the actual pressure is within the allowable range. For each pressure sub-region, the deviation between its local pressure value and the preset pressure threshold is calculated. All deviation values ​​are collected into a set to form a local pressure deviation set. Each element in this set corresponds to the local pressure value of a pressure sub-region, which is the average local pressure value at multiple coordinate points in that pressure sub-region.

[0038] The differential method is used to calculate the pressure changes between adjacent coordinates, obtaining the gradient of the local pressure deviation set. This gradient is then compared with a pre-set gradient threshold. Based on the comparison results, regions where the pressure deviation exceeds the allowable range are identified and output as pressure anomaly regions. The spatial coordinates corresponding to the pressure anomaly regions are then marked to form a pressure anomaly spatial coordinate set.

[0039] Deviation magnitude refers to the range of pressure deviation, used to assess the severity of the deviation. Morphological characteristics refer to the shape and distribution features of the pressure anomaly area, such as linear distribution, point distribution, etc.

[0040] The system analyzes the set of local pressure deviations, calculates the absolute value of the deviation or its ratio to a preset pressure threshold, and thus determines the magnitude of the deviation in the pressure anomaly region. This magnitude refers to the range of pressure deviation magnitude and is used to assess the severity of the deviation. Based on the spatial coordinates of the pressure anomaly, the shape and distribution characteristics of the pressure anomaly region are analyzed, such as linear distribution or point distribution. The magnitude and morphological characteristics of the deviation are input into a PID control algorithm or a fuzzy control algorithm to generate multi-dimensional control commands, including parameters for adjusting the scraper angle, vertical pressure, and coating speed.

[0041] The above steps involve matrix partitioning, deviation calculation, gradient identification, and anomaly analysis of the dynamic pressure distribution map. Ultimately, a multi-dimensional control command is constructed to address coating anomalies, enabling refined control of the coating process, reducing uneven coating thickness, and improving coating quality.

[0042] Furthermore, step S23 includes: Step S231: Perform gradient calculation on the set of local pressure deviations to generate lateral pressure gradient and longitudinal pressure gradient.

[0043] Step S232: When the lateral pressure gradient exceeds the first threshold and the longitudinal pressure gradient exceeds the second threshold, it is determined to be a pressure change region.

[0044] Step S233: Perform continuous pressure analysis on the pressure mutation region based on the local pressure deviation set to generate a pressure anomaly region.

[0045] Step S234: Extract the set of boundary coordinates of the pressure anomaly region, map and identify the multiple spatial coordinates, and determine the spatial coordinates of the pressure anomaly.

[0046] Specifically, for each element in the local pressure deviation set, the difference between it and its lateral and longitudinal adjacent elements is calculated to obtain the lateral pressure gradient and longitudinal pressure gradient corresponding to each pressure sub-region. Taking the lateral pressure gradient as an example, for an element in a certain row of the pressure deviation set, the difference between it and its left or right adjacent element is calculated, and these differences are organized in order to form the lateral pressure gradient.

[0047] The first and second thresholds are pre-set criteria for determining transverse and longitudinal pressure abrupt changes, used to judge whether abnormal changes in transverse or longitudinal pressure occur. According to process requirements, thresholds for transverse and longitudinal pressure gradients are set to obtain the first and second thresholds. The calculated transverse pressure gradient is compared with the first threshold, and the longitudinal pressure gradient is compared with the second threshold. When the transverse pressure gradient of a sub-region exceeds the first threshold and the longitudinal pressure gradient exceeds the second threshold, that region is determined to be a pressure abrupt change region.

[0048] By combining sensor time-series data, a sliding window is used to perform continuous pressure analysis on regions of sudden pressure changes to determine whether the pressure deviation consistently exceeds a set value. A time window is set, for example, 3 seconds. Within the time window, the pressure deviation value in the region of sudden pressure change is continuously monitored. If the pressure deviation value consistently exceeds a preset pressure threshold by a certain range (e.g., 20%) within the time window, the region of sudden pressure change is identified as a pressure anomaly region. Through continuous pressure analysis, effective differentiation between transient and persistent anomalies is achieved, avoiding occasional misjudgments and improving the robustness of the control response.

[0049] Extract the set of boundary coordinates from the pressure anomaly region, map the set of boundary coordinates to spatial coordinates, that is, convert the indices (such as i, j) in the two-dimensional matrix into real physical coordinates (such as mm-level positions), determine the specific location of the pressure anomaly region, and mark the mapped coordinates as the pressure anomaly spatial coordinates.

[0050] The above steps, through multi-dimensional gradient analysis and time persistence analysis, comprehensively construct a pressure anomaly identification mechanism with spatial positioning and dynamic persistence discrimination capabilities, effectively enhancing the accuracy, continuity, and real-time performance of pressure anomaly identification, and providing a clear and reliable response basis for the subsequent execution layer of the coating control system.

[0051] Furthermore, step S3 includes: Step S31: Receive the multi-dimensional control command, parse it, and drive the multi-axis linkage module of the coating blade execution mechanism according to the parsing result.

[0052] Step S32: Monitor the synchronous adjustment response of the coating process through the multi-axis linkage module to obtain multiple monitoring datasets.

[0053] Step S33: Perform time-series correlation between the multiple monitoring datasets and the parameters of the multi-dimensional control command to construct the scraper coating parameter set.

[0054] Specifically, the coating blade control system receives and analyzes multi-dimensional control commands via a PLC controller or embedded chip, extracting the blade angle adjustment amount, vertical pressure correction value, and coating speed compensation coefficient. Based on the analysis results, the control system sends corresponding control signals to the multi-axis linkage module of the coating blade execution mechanism, driving the multi-axis linkage module to adjust the blade's angle, pressure, and speed. The multi-axis linkage module can simultaneously control multiple axes (corresponding to different control dimensions of the blade, such as angle, pressure, and speed). This module receives control commands via a CAN bus and feeds back the actual execution status data of each axis. For example, the multi-axis linkage module can employ a combination of an electric screw drive system, a hydraulic servo unit, and a speed-regulating conveyor motor.

[0055] After receiving control commands via the CAN bus, the multi-axis linkage module begins synchronous adjustment, simultaneously changing the scraper tilt angle, hydraulic pressure, and substrate conveying speed. During this process, sensors (such as displacement sensors to measure scraper displacement, pressure sensors to measure actual pressure values, and speed sensors to measure coating line speed) are used to collect dynamic response data in real time, obtaining multiple monitoring datasets which are then fed back to the control system. These monitoring datasets include dynamic response data such as scraper displacement, actual pressure values, and coating line speed.

[0056] Based on the timestamp, each monitoring data point in multiple monitoring datasets is associated with the corresponding control command parameters to construct a comprehensive dataset, namely the scraper coating parameter set, which provides a complete data foundation for subsequent coating optimization operations.

[0057] The above steps achieve closed-loop control, encompassing control analysis, multi-axis linkage execution, and feedback data construction. By structurally matching intelligent commands with real-time response data, high-precision linkage capabilities, real-time response monitoring capabilities, and data-driven adaptive capabilities are ensured when executing control strategies, providing underlying data support for subsequent strategy optimization and coating quality improvement.

[0058] Furthermore, step S33 includes: Step S331: Iterate through the multi-dimensional control instructions to add data and obtain multi-dimensional control timestamps and multi-dimensional environmental parameter labels.

[0059] Step S332: Perform actual execution analysis based on the multiple monitoring datasets to determine multiple actual execution parameters.

[0060] Step S333: Encode the multi-dimensional control instructions and the multiple actual execution parameters into a matrix according to the multi-dimensional control timestamp to generate a matrix encoding result.

[0061] Step S334: Associate the multi-dimensional environmental parameter labels with the matrix encoding results to construct the scraper coating parameter set.

[0062] Specifically, the system iterates through multi-dimensional control commands, adding a multi-dimensional control timestamp to each command element. This timestamp can be provided by the clock module in the control system; for example, the current time can be obtained and marked on the control command using the built-in timer function in the microcontroller. Simultaneously, sensors (such as temperature, humidity, and tension sensors) are used to collect current environmental parameters, including temperature, humidity, and substrate tension, and corresponding multi-dimensional environmental parameter labels are added to the multi-dimensional control commands based on these parameters.

[0063] Statistical processing is performed on multiple monitoring datasets to calculate key indicators such as average value, maximum deviation, and response delay. These key indicators are then extracted as result parameters, such as actual angle change, actual applied pressure, and actual coating line speed. Next, based on multi-dimensional control timestamps, multi-dimensional control commands and multiple actual execution parameters are matrix-encoded to generate matrix encoding results. In this matrix, rows represent the time series, and columns represent control command parameters, including angle, pressure, and speed. An example matrix encoding result is shown in Table 1. Table 1—Matrix Encoding Results By associating multi-dimensional environmental parameter labels with the matrix encoding results, new columns can be added to the matrix encoding results to represent environmental parameters, or a data index can be established to associate environmental parameters with each row (corresponding to different time series) in the matrix encoding results. Simultaneously, coating thickness detection results are collected using downstream measurement equipment (such as a laser thickness gauge) and integrated into the matrix encoding results to construct a scraper coating parameter set.

[0064] The above steps unify the encoding of control commands, execution feedback, environmental status, and final effect data, and establish a time-series matrix and a multi-dimensional fusion dataset to form a structured scraper coating parameter set. This comprehensively records various parameters in the coating process, providing rich data support for subsequent optimization and feedback.

[0065] Furthermore, step S4 includes: Step S41: Extract the historical doctor blade coating parameter set from the process database, and perform feature extraction on the doctor blade coating parameter set to obtain doctor blade pressure features and coating thickness features.

[0066] Step S42: Construct a pressure-thickness correlation feature matrix by performing correlation mapping on the scraper pressure feature and the coating thickness feature.

[0067] Step S43: Perform regression analysis based on the pressure-thickness correlation feature matrix to generate a dynamic pressure adjustment rule set.

[0068] Step S44: Perform data backtracking on the scraper coating parameter set according to the pressure-thickness correlation feature matrix to formulate a pressure control strategy.

[0069] Step S45: Execute the pressure control strategy according to the dynamic pressure adjustment rule set to perform multi-objective collaborative optimization of the multi-dimensional control command, determine the collaborative optimization parameters, execute the coating process of the coating blade according to the collaborative optimization parameters and perform quality analysis, and generate the coating optimization report.

[0070] Specifically, the process database stores data on the coating blade coating process, including historical control commands, actual execution data, and coating thickness detection results. Then, data analysis software is used to extract features from these parameter sets to obtain blade pressure characteristics and coating thickness characteristics. For blade pressure characteristics, statistical analysis methods can be used to calculate statistics such as the mean and standard deviation of the pressure data to obtain blade pressure characteristic indicators such as average pressure and pressure change rate. Similarly, for coating thickness characteristics, the average and range of the thickness data can be calculated to obtain coating thickness characteristic indicators such as average thickness and thickness fluctuation rate.

[0071] A correlation mapping is performed between the characteristics of the scraper pressure and the coating thickness to establish a matrix structure representing their relationship, namely the pressure-thickness correlation feature matrix. For example, historical data entries are used as row elements of the matrix, and the scraper pressure and coating thickness characteristics are used as column elements.

[0072] Regression analysis was performed on the data in the pressure-thickness correlation feature matrix using data analysis tools such as Python (scikit-learn) and the MATLAB regression toolbox to determine the regression equation between the scraper pressure characteristics and the coating thickness characteristics. Then, a dynamic pressure adjustment rule set was developed based on this regression model. For example, if the regression model shows a linear relationship between pressure and thickness, then pressure adjustment rules for different thickness requirements are developed based on this relationship. For instance, when the pressure increases by 1 kPa, the thickness increases by 0.5 μm.

[0073] The pressure-thickness correlation feature matrix is ​​used to perform retrospective analysis on historical coating parameter sets. This involves reverse analysis and querying of the blade coating parameter set based on the pressure-thickness correlation feature matrix to formulate pressure control strategies. For example, if the pressure value in the pressure-thickness correlation feature matrix is ​​found to correspond to a thickness that does not meet requirements, the adjustment of other parameters (such as speed and angle) under similar pressures is searched in the blade coating parameter set to formulate a pressure control strategy.

[0074] The pressure control strategy is executed according to a dynamic pressure adjustment rule set, and optimization algorithms such as genetic algorithms, particle swarm optimization, or rule-based strategies are used. Simultaneously, multi-dimensional control commands are adjusted from three dimensions: angle, pressure, and speed, to determine co-optimization parameters. Then, the coating process of the coating blade is executed according to the co-optimization parameters, and thickness uniformity, defect rate, and energy efficiency are measured and statistically analyzed during the process. Finally, a coating optimization report is generated.

[0075] The above steps construct a feature regression model driven by historical data to extract pressure-thickness control patterns from historical experience and then map these patterns back to current control commands, achieving accurate prediction and optimization of the doctor blade pressure adjustment behavior. Simultaneously, through a collaborative multi-objective optimization mechanism, the doctor blade angle, pressure, and speed form a closed-loop linkage, effectively reducing thickness unevenness and defect rates, and outputting a coating optimization report including process evaluation indicators. Overall, this significantly improves the stability of the coating process and the uniformity of coating thickness.

[0076] Furthermore, in step S45, the pressure control strategy is executed according to the dynamic pressure adjustment rule set to perform multi-objective collaborative optimization of the multi-dimensional control command, determining the collaborative optimization parameters, including: Step S451: Retrieve the historical best coating control parameters as the initial solution, and execute the pressure control strategy on the initial solution according to the dynamic pressure adjustment rule set to generate multiple pressure control solutions.

[0077] Step S452: Perform fitness scoring on the multiple pressure control solutions, and set control constraint terms based on the multiple fitness scores.

[0078] Step S453: Perform cross-mutation verification on the multi-dimensional control instructions according to the control constraints, perform iterative optimization based on the verification results, until the preset number of iterations is reached, and generate the collaborative optimization parameters.

[0079] Specifically, the process begins by identifying the best-performing control parameters in historical data to meet various coating performance indicators such as thickness uniformity, defect rate, and energy efficiency. These historically optimal coating control parameters are then used as the initial solution. Next, based on a dynamic pressure adjustment rule set, a pressure control strategy is executed on this initial solution, generating multiple pressure control solutions. For example, the historically optimal solution is: angle = 46°, pressure = 0.20 MPa, speed = 4.0 m / min. Multiple solutions are generated based on the dynamic pressure adjustment rule set, such as solution 1: angle +1°, pressure +0.01 MPa, speed -0.2 m / min; solution 2: angle -1°, pressure -0.005 MPa, speed +0.1 m / min, etc.

[0080] For each pressure control solution, its fitness is scored according to pre-defined evaluation criteria. These criteria include indicators such as thickness uniformity, defect rate, and energy efficiency. Then, based on the fitness scores of multiple pressure control solutions, control constraints are set to limit the scope and conditions of the optimization process. For example, if some pressure control solutions are found to perform poorly in terms of thickness uniformity, the optimization process can be limited by the parameter ranges of these solutions, such as: pressure variation not exceeding ±0.03 MPa, coating speed fluctuation within ±5%, etc., to avoid meaningless optimization within these poorly performing parameter ranges.

[0081] Crossover and mutation operations are performed on multi-dimensional control instructions according to control constraints. This involves swapping or randomly changing some parameters of two or more control instructions, and then verifying the resulting new instructions to check if they satisfy the control constraints. If they do, the next round of iterative optimization continues; otherwise, the crossover and mutation operation is repeated. This process is repeated until a preset number of iterations is reached, ultimately generating the co-optimized parameters. In practice, genetic algorithm libraries (such as the DEAP library in Python) can be used to implement crossover and mutation verification and iterative optimization operations.

[0082] Starting with the historically optimal combination of coating parameters, the above steps generate multiple candidate pressure control solutions under the guidance of a dynamic pressure adjustment rule set. The comprehensive performance of each control solution is evaluated using a fitness scoring mechanism, and constraints are established accordingly. Under these constraints, a cross-mutation mechanism is further used to iteratively optimize multi-dimensional control commands. Through multiple rounds of evolution, a collaboratively optimized parameter combination that satisfies the balance of multi-objective performance is finally determined, significantly improving the autonomous control capability and process stability of the coating system.

[0083] Furthermore, step S45, which involves executing the coating process of the coating blade according to the collaborative optimization parameters and performing quality analysis to generate the coating optimization report, includes: Step S454: Drive the coating blade to perform the coating process based on the collaborative optimization parameters, and collect and acquire coating process execution data in real time.

[0084] Step S455: Based on the coating process execution data, perform a thickness scan on the coating substrate to obtain a coating thickness distribution curve.

[0085] Step S456: Surface capture of the coating substrate is performed based on the coating process execution data to obtain a coating surface image.

[0086] Step S457: Perform energy consumption analysis based on the coating process execution data to obtain the coating energy consumption value.

[0087] Step S458: Align the coating thickness distribution curve, the coating surface image, and the coating energy consumption value according to the coating timestamp to construct a multimodal coating time series dataset, and add the multimodal coating time series dataset to the coating optimization report.

[0088] Specifically, a multi-axis linkage module is used to drive the coating blade execution mechanism through collaborative optimization parameters. During this process, various sensors collect coating process execution data in real time, including thickness distribution, surface defects, and process energy consumption. For example, a thickness sensor is used to obtain coating thickness distribution data, image acquisition devices (such as cameras) capture surface images of the coating substrate, and energy consumption monitoring instruments record process energy consumption.

[0089] Data processing software is used to plot the thickness distribution data from the coating process execution data on a coordinate system according to the measurement locations, thus obtaining the coating thickness distribution curve. Image processing software is used to process and fuse the surface images from the coating process execution data to obtain a coating surface image, which reflects the condition of the coating surface, such as the presence of defects. Process energy consumption data is extracted from the coating process execution data, and the total energy consumed by the motor drive, solenoid valves, hydraulic system, etc., throughout the entire coating cycle is calculated to obtain the coating energy consumption value.

[0090] Based on the coating timestamp, the coating thickness distribution curve, coating surface image, and coating energy consumption value are aligned to construct a multimodal coating time-series dataset. For example, if a set of thickness values, a surface image, and the corresponding energy consumption value are collected at a certain time point t1, these three sets of data are arranged in the multimodal coating time-series dataset according to the time stamp t1. Then, the constructed multimodal coating time-series dataset is added to the coating optimization report, and based on this multimodal coating time-series dataset, key indicators such as coating quality, energy saving level, and stability are evaluated according to preset rules, and also added to the coating optimization report.

[0091] The surface of the coated substrate is captured using image capture equipment (such as a high-resolution camera). During capture, it is crucial to ensure image clarity and accuracy to accurately reflect the surface condition. Image quality can be optimized by adjusting parameters such as camera focal length and lighting. This report not only visually reflects coating quality, stability, and energy efficiency, but can also be used for subsequent strategy correction and model training, thereby achieving closed-loop optimization and continuous quality improvement and energy saving.

[0092] In summary, the pressure-controlled scraper coating optimization method provided in this application has the following beneficial effects: By arranging an array of pressure sensors equidistantly along the axial direction of the coating blade body and performing real-time pressure sensing, a dynamic pressure distribution map is generated, providing accurate data support for subsequent pressure deviation calculation and control optimization. The pressure distribution matrix is ​​extracted by traversing the dynamic pressure distribution map and divided into multiple pressure sub-regions. A preset pressure threshold is determined for deviation calculation, generating a local pressure deviation set. Gradient analysis is then performed based on this set to identify abnormal pressure areas, which are then marked using spatial coordinates. Finally, multi-dimensional control commands are constructed, achieving dynamic and refined control of the coating blade. The coating blade execution mechanism is driven by these multi-dimensional control commands for coating. During execution, a multi-axis linkage module monitors the synchronous adjustment response of the coating process, obtaining multiple monitoring datasets. These datasets are then correlated with the parameters of the multi-dimensional control commands in a time sequence to construct a blade coating parameter set, providing structured data support for data feedback and optimization. Furthermore, by extracting historical blade coating parameter sets from the process database and performing feature extraction, a pressure-thickness correlation feature matrix is ​​constructed. Regression analysis is then performed to generate a dynamic pressure adjustment rule set, which is used to formulate a pressure control strategy. Multi-dimensional control commands are then subjected to multi-objective collaborative optimization to determine the collaborative optimization parameters, thereby achieving intelligent control and continuous improvement of the coating process. Finally, by collecting coating process execution data in real time, performing thickness scanning, surface capture, and energy consumption analysis, a coating optimization report is generated to comprehensively evaluate coating quality and provide a scientific basis for production decisions.

[0093] Overall, the embodiments of this application construct a coating blade control method with real-time perception, multi-dimensional decision-making, and closed-loop optimization as its core, realizing real-time monitoring and fine control of blade pressure. Furthermore, by introducing execution feedback and strategy correction mechanisms, the coating process is enhanced to adapt to material changes and process disturbances, significantly improving the uniformity of coating thickness and the overall stability of coating, thus providing a solution for high-quality coating in modern industry.

[0094] Example 2, as Figure 3 As shown, based on the same inventive concept as in Embodiment 1 above, this application provides a doctor blade coating optimization system combined with pressure control, the system comprising: The pressure sensing module 10 is used for real-time pressure sensing based on the coating blade to generate a dynamic pressure distribution map.

[0095] The control command generation module 20 is used to calculate the pressure deviation according to the dynamic pressure distribution map and generate multi-dimensional control commands.

[0096] The coating control module 30 is used to drive the coating blade execution mechanism to perform coating through the multi-dimensional control commands and obtain the blade coating parameter set.

[0097] The coating optimization module 40 is used to perform data feedback based on the scraper coating parameter set, formulate a pressure control strategy, optimize the coating of the multi-dimensional control commands through the pressure control strategy, and generate a coating optimization report.

[0098] Furthermore, in this embodiment of the application, the pressure sensing module 10 is also used to perform the following steps: A pressure sensor array is equidistantly arranged along the axial direction on the coating blade body; the pressure sensor array senses the contact pressure between the coating blade and the coating substrate to obtain multiple local pressure values; the pressure sensor array is spatially labeled based on the multiple local pressure values ​​to determine multiple spatial coordinates; the multiple spatial coordinates are mapped onto the surface of the coating substrate for dynamic labeling to generate the dynamic pressure distribution map.

[0099] Furthermore, in this embodiment of the application, the control instruction generation module 20 is also used to perform the following steps: The dynamic pressure distribution map is traversed to extract the pressure distribution matrix, which is then divided into multiple pressure sub-regions. A preset pressure threshold is determined, and the deviations of the multiple pressure sub-regions are calculated according to the preset pressure threshold. The deviations of the multiple local pressure values ​​from the preset pressure threshold are calculated to generate a local pressure deviation set. Gradient analysis is performed based on the local pressure deviation set to identify and determine pressure anomaly regions. These pressure anomaly regions are then identified by combining the multiple spatial coordinates to determine the pressure anomaly spatial coordinates. The deviation magnitude of the pressure anomaly regions is determined based on the local pressure deviation set, and the morphological characteristics of the pressure anomaly regions are determined based on the pressure anomaly spatial coordinates. Finally, the multi-dimensional control command is constructed.

[0100] Furthermore, in this embodiment of the application, the control instruction generation module 20 is also used to perform the following steps: Gradient calculation is performed on the local pressure deviation set to generate a lateral pressure gradient and a longitudinal pressure gradient. When the lateral pressure gradient exceeds a first threshold and the longitudinal pressure gradient exceeds a second threshold, it is determined to be a pressure abrupt change region. Based on the local pressure deviation set, continuous pressure analysis is performed on the pressure abrupt change region to generate a pressure anomaly region. The boundary coordinate set of the pressure anomaly region is extracted and mapped to the multiple spatial coordinates to determine the pressure anomaly spatial coordinates.

[0101] Furthermore, in this embodiment of the application, the coating control module 30 is also used to perform the following steps: The system receives and parses the multi-dimensional control commands, and drives the multi-axis linkage module of the coating blade execution mechanism based on the parsing results. The system then monitors the coating synchronization adjustment response through the multi-axis linkage module to obtain multiple monitoring datasets. Finally, the system correlates the multiple monitoring datasets with the parameters of the multi-dimensional control commands in a time sequence to construct the blade coating parameter set.

[0102] Furthermore, in this embodiment of the application, the coating control module 30 is also used to perform the following steps: The multi-dimensional control instructions are iterated through and data is added to obtain multi-dimensional control timestamps and multi-dimensional environmental parameter labels. Based on the multiple monitoring datasets, actual execution analysis is performed to determine multiple actual execution parameters. The multi-dimensional control instructions and the multiple actual execution parameters are matrix-encoded according to the multi-dimensional control timestamps to generate matrix encoding results. The multi-dimensional environmental parameter labels are associated with the matrix encoding results to construct the scraper coating parameter set.

[0103] Furthermore, in this embodiment of the application, the coating optimization module 40 is also used to perform the following steps: Historical doctor blade coating parameter sets are extracted from the process database. Feature extraction is performed on these parameter sets to obtain doctor blade pressure and coating thickness features. A pressure-thickness correlation feature matrix is ​​constructed by mapping the pressure and thickness features together. Regression analysis is performed based on this matrix to generate a dynamic pressure adjustment rule set. Data backtracking is conducted on the doctor blade coating parameter sets according to the pressure-thickness correlation feature matrix to formulate a pressure control strategy. The pressure control strategy is executed according to the dynamic pressure adjustment rule set to perform multi-objective collaborative optimization of the multi-dimensional control commands, determining the collaborative optimization parameters. The doctor blade coating process is then executed based on these parameters, and quality analysis is performed to generate a coating optimization report.

[0104] Furthermore, in this embodiment of the application, the coating optimization module 40 is also used to perform the following steps: The historically optimal coating control parameters are retrieved as the initial solution. A pressure control strategy is executed on the initial solution according to the dynamic pressure adjustment rule set to generate multiple pressure control solutions. The fitness scores of the multiple pressure control solutions are calculated, and control constraints are set based on the fitness scores. The multi-dimensional control instructions are cross-mutated and verified according to the control constraints. Iterative optimization is performed based on the verification results until a preset number of iterations is reached to generate the collaborative optimization parameters.

[0105] Furthermore, in this embodiment of the application, the coating optimization module 40 is also used to perform the following steps: The coating process is driven by the co-optimized parameters, and the coating process execution data is collected in real time. The coating thickness distribution curve is obtained by scanning the coating substrate based on the coating process execution data. The surface of the coating substrate is captured based on the coating process execution data to obtain the coating surface image. Energy consumption analysis is performed based on the coating process execution data to obtain the coating energy consumption value. The coating thickness distribution curve, the coating surface image, and the coating energy consumption value are aligned according to the coating timestamp to construct a multimodal coating time series dataset, and the multimodal coating time series dataset is added to the coating optimization report.

[0106] Through the foregoing detailed description of the pressure-controlled scraper coating optimization method, those skilled in the art can clearly understand that the pressure-controlled scraper coating optimization system in this embodiment, as disclosed in Embodiment 2, corresponds to the method disclosed in Embodiment 1, and has corresponding functional modules and beneficial effects. For relevant details, please refer to the method section.

[0107] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for optimizing blade coating by combining pressure control, characterized in that, The method includes: Real-time pressure sensing based on coating blade generates dynamic pressure distribution map; Pressure deviation is calculated based on the dynamic pressure distribution diagram to generate multi-dimensional control commands. The coating blade execution mechanism is driven by the multi-dimensional control commands to perform coating, thereby obtaining a set of blade coating parameters. Based on the scraper coating parameter set, data feedback is performed to formulate a pressure control strategy. The multi-dimensional control commands are then optimized using the pressure control strategy to generate a coating optimization report. Among them, the method for generating a dynamic pressure distribution map based on real-time pressure sensing using a coating blade includes: An array of pressure sensors is arranged at equal intervals along the axial direction on the coating blade body; The pressure sensor array is used to sense the contact pressure between the coating blade and the coating substrate to obtain multiple local pressure values. The pressure sensor array is spatially identified based on the multiple local pressure values, and multiple spatial coordinates are determined. The multiple spatial coordinates are mapped onto the surface of the coated substrate for dynamic identification, thereby generating the dynamic pressure distribution map; The method for calculating pressure deviation and generating multi-dimensional control commands based on the dynamic pressure distribution diagram includes: The pressure distribution matrix is ​​extracted by traversing the dynamic pressure distribution map, and the pressure distribution matrix is ​​divided into multiple pressure sub-regions. A preset pressure threshold is determined, and the deviation of the multiple pressure sub-regions is calculated according to the preset pressure threshold. The deviation of the multiple local pressure values ​​from the preset pressure threshold is calculated, and a local pressure deviation set is generated. Gradient analysis is performed based on the set of local pressure deviations to identify and determine pressure anomaly regions. These pressure anomaly regions are then identified in combination with the multiple spatial coordinates to determine the spatial coordinates of the pressure anomalies. The deviation magnitude of the pressure anomaly region is determined based on the set of local pressure deviations, the morphological characteristics of the pressure anomaly region are determined based on the spatial coordinates of the pressure anomaly, and the multi-dimensional control command is constructed.

2. The optimized scraper coating method combining pressure control as described in claim 1, characterized in that, The method for performing gradient analysis based on the set of local pressure deviations to identify and determine pressure anomaly regions, and then identifying the pressure anomaly regions by combining the multiple spatial coordinates to determine the spatial coordinates of the pressure anomalies, includes: Gradient calculations are performed on the set of local pressure deviations to generate lateral pressure gradients and longitudinal pressure gradients; When the lateral pressure gradient exceeds a first threshold and the longitudinal pressure gradient exceeds a second threshold, it is determined to be a pressure abrupt change region; Based on the set of local pressure deviations, continuous pressure analysis is performed on the pressure mutation region to generate a pressure anomaly region. Extract the set of boundary coordinates of the pressure anomaly region, map and identify the multiple spatial coordinates, and determine the spatial coordinates of the pressure anomaly.

3. The optimized scraper coating method combining pressure control as described in claim 1, characterized in that, The coating process is performed by driving the coating blade execution mechanism through the multi-dimensional control commands to obtain a set of blade coating parameters. The method includes: The system receives and parses the multi-dimensional control commands, and drives the multi-axis linkage module of the coating blade execution mechanism based on the parsing results. The multi-axis linkage module is used to monitor the synchronous adjustment response of the coating process, and multiple monitoring datasets are obtained. The multiple monitoring datasets are correlated with the parameters of the multi-dimensional control command in a time sequence to construct the scraper coating parameter set.

4. The optimized scraper coating method combining pressure control as described in claim 3, characterized in that, The method for constructing the scraper coating parameter set by temporally associating the multiple monitoring datasets with the parameters of the multi-dimensional control command includes: The multi-dimensional control instructions are iterated through to add data, and multi-dimensional control timestamps and multi-dimensional environmental parameter labels are obtained. Based on the multiple monitoring datasets, actual execution analysis is performed to determine multiple actual execution parameters; The multi-dimensional control instructions and the multiple actual execution parameters are matrix-encoded according to the multi-dimensional control timestamps to generate matrix encoding results. The multi-dimensional environmental parameter labels are associated with the matrix encoding results to construct the scraper coating parameter set.

5. The optimized scraper coating method combining pressure control as described in claim 1, characterized in that, Based on the scraper coating parameter set, data feedback is performed to formulate a pressure control strategy. This pressure control strategy is then used to optimize the coating process for the multi-dimensional control commands, generating a coating optimization report. The method includes: Historical scraper coating parameter sets are extracted from the process database, and feature extraction is performed on the scraper coating parameter sets to obtain scraper pressure features and coating thickness features. The pressure-thickness correlation feature matrix is ​​constructed by mapping the scraper pressure feature and the coating thickness feature together. Regression analysis is performed based on the pressure-thickness correlation feature matrix to generate a dynamic pressure adjustment rule set. Based on the pressure-thickness correlation feature matrix, the data backtracking of the scraper coating parameter set is performed to formulate a pressure control strategy; According to the dynamic pressure adjustment rule set, the pressure control strategy is executed to perform multi-objective collaborative optimization of the multi-dimensional control command, determine the collaborative optimization parameters, execute the coating process of the coating blade according to the collaborative optimization parameters, perform quality analysis, and generate the coating optimization report.

6. The optimized scraper coating method combining pressure control as described in claim 5, characterized in that, The pressure control strategy is executed according to the dynamic pressure adjustment rule set to perform multi-objective collaborative optimization of the multi-dimensional control commands, and the collaborative optimization parameters are determined. The method includes: The historical best coating control parameters are retrieved as the initial solution, and a pressure control strategy is executed on the initial solution according to the dynamic pressure adjustment rule set to generate multiple pressure control solutions. The fitness scores of the multiple pressure control solutions are calculated, and control constraints are set based on the fitness scores. The multi-dimensional control instructions are cross-mutated and validated according to the control constraints. Iterative optimization is performed based on the validation results until a preset number of iterations is reached, and the collaborative optimization parameters are generated.

7. The optimized scraper coating method combining pressure control as described in claim 5, characterized in that, The coating process of the coating blade is executed according to the aforementioned collaborative optimization parameters, and quality analysis is performed to generate the coating optimization report. The method includes: The coating process is driven by the coating blade based on the aforementioned collaborative optimization parameters, and the coating process execution data is collected in real time. Based on the coating process execution data, the coating substrate is thickness scanned to obtain the coating thickness distribution curve; Based on the coating process execution data, the surface of the coated substrate is captured to obtain an image of the coated surface. Energy consumption analysis is performed based on the coating process execution data to obtain the coating energy consumption value; Align the coating thickness distribution curve, the coating surface image, and the coating energy consumption value according to the coating timestamp to construct a multimodal coating time series dataset, and add the multimodal coating time series dataset to the coating optimization report.

8. A pressure-controlled scraper coating optimization system, characterized in that, The system is used to perform the blade coating optimization method combined with pressure control as described in any one of claims 1-7, including: The pressure sensing module is used to perform real-time pressure sensing based on the coating blade and generate a dynamic pressure distribution map. The control command generation module is used to calculate the pressure deviation according to the dynamic pressure distribution map and generate multi-dimensional control commands. The coating control module is used to drive the coating blade execution mechanism to perform coating through the multi-dimensional control commands and obtain the blade coating parameter set. The coating optimization module is used to formulate a pressure control strategy based on the data feedback of the scraper coating parameter set, optimize the coating of the multi-dimensional control commands through the pressure control strategy, and generate a coating optimization report.

Images