A Method for Optimizing Aluminum Plate Texture Based on Surface Tension Adjustment

Abstract

This invention provides a method for optimizing aluminum plate texture based on surface tension adjustment, belonging to the field of aluminum plate processing technology. The invention reduces the roughness of the aluminum plate surface through plasma treatment, configures a functional liquid containing a surfactant mixture and trace polymers, constructs an array-type local heating unit and a closed-loop temperature control system to achieve programmable temperature field distribution, inputs liquid film surface temperature distribution data and initial thickness data into a flow field prediction model to obtain Marangoni convection velocity field distribution and liquid film morphology evolution prediction results, dynamically adjusts the heating power distribution mode of each region based on the prediction results, and achieves directional correction by increasing the concentration of eddy current suppressing polymers and reducing the temperature gradient when the texture direction deflection angle exceeds a threshold. This solves the technical problem of accurately controlling the texture randomness caused by Marangoni convection during liquid film evaporation.

Description

Technical Field

[0001] This invention belongs to the field of aluminum plate processing technology, and more specifically, relates to a method for optimizing aluminum plate texture based on surface tension adjustment. Background Technology

[0002] In the field of aluminum plate surface texture preparation, traditional techniques mainly achieve surface texture pattern formation by controlling the liquid film evaporation process, and are widely used in the manufacture of decorative materials, optical devices, and functional coatings. These methods rely on temperature gradient-driven Marangoni convection to regulate liquid film flow, thereby affecting solute deposition distribution and texture morphology. However, in existing technologies, the velocity field distribution of Marangoni convection is difficult to predict and dynamically adjust in real time, leading to the formation of disordered vortex structures within the liquid film, causing significant deflection of the texture direction. Simultaneously, the evolution path of the liquid film boundary cannot be accurately tracked, resulting in randomness and non-repeatability of the texture morphology. Furthermore, the excessive accumulation of solute at the edges caused by the coffee ring effect further exacerbates texture defects. In current aluminum plate surface treatment processes, due to the lack of precise prediction methods and real-time control mechanisms for liquid film dynamics, traditional methods struggle to achieve directional control of texture patterns. In other words, existing technologies suffer from the technical problem of difficulty in precisely controlling the texture randomness caused by Marangoni convection during liquid film evaporation. Summary of the Invention

[0003] In view of this, the present invention provides an aluminum plate texture optimization method based on surface tension adjustment, which can solve the technical problem in the prior art that the texture randomness caused by Marangoni convection during liquid film evaporation is difficult to control precisely.

[0004] This invention is implemented as follows: It provides a method for optimizing aluminum plate texture based on surface tension regulation, including plasma treatment of the aluminum plate surface, preparation of a functional liquid and spraying it onto the aluminum plate surface to form a liquid film, activation of a closed-loop temperature control system, construction of an array-type local heating unit, input of the temperature distribution data of the liquid film surface into a flow field prediction model, output of the Marangoni convection velocity field distribution and liquid film morphology evolution prediction results from the flow field prediction model, adjustment of the power distribution mode of the array-type local heating unit based on the Marangoni convection velocity field distribution and liquid film morphology evolution prediction results, increasing the concentration of eddy current suppressor polymer and decreasing the temperature gradient of the corresponding region to achieve texture orientation correction when the texture direction deflection angle output by the flow field prediction model exceeds a threshold, controlling the ambient humidity and establishing selective evaporation conditions for the mixed solvent, utilizing the reverse Marangoni flow generated by the preferential evaporation of low-boiling-point components to counteract the coffee ring effect, and achieving accurate tracking of the liquid film boundary position and real-time prediction and control of Marangoni convection through the integration of the flow field prediction model with a phase-field model interface tracking algorithm.

[0005] The plasma treatment employs Qi and The mixture of gases has a gas flow ratio of 4:1, a processing power of 200W to 300W, and a processing time of 180s to 240s.

[0006] The plasma treatment reduces the surface roughness of the aluminum plate to below 50 nm and constructs an array-type local heating unit on the treated aluminum plate surface. The array-type local heating unit consists of multiple independently controlled resistance heating elements with a spacing of 5 mm between the heating elements.

[0007] The functional liquid is composed of a base solvent, a surfactant mixture, and a trace polymer, wherein the surfactant mixture has an HLB value of 8 to 12 and the trace polymer has a mass fraction of 0.3% to 0.8%.

[0008] The trace polymer is a polyvinylpyrrolidone or polyacrylamide polymer with a molecular weight of [missing information]. to .

[0009] The closed-loop temperature control system integrates multi-point platinum resistance sensors and uses a PID control algorithm to control the substrate temperature fluctuation within ±0.2℃.

[0010] In the PID control algorithm of the closed-loop temperature control system, the proportional coefficient is set to 0.5 to 0.8, the integral time constant is set to 10s to 20s, and the derivative time constant is set to 2s to 5s.

[0011] The temperature distribution data of the liquid film surface is obtained by a multi-point platinum resistance sensor at 25 measurement points on the aluminum plate surface, with a sampling frequency of 10Hz. The temperature distribution data includes the temperature value and spatial coordinate value of each measurement point.

[0012] Before inputting the temperature distribution data of the liquid film surface into the flow field prediction model, the process includes measuring the initial thickness data of the liquid film using a laser triangulation distance sensor. The measurement range covers six regions on the aluminum plate surface, each region having an area of ​​50 mm. .

[0013] The input layer of the flow field prediction model receives temperature distribution data and initial thickness data of the liquid film surface. The input layer contains 256 input nodes. The first 250 input nodes receive temperature values ​​and spatial coordinate values ​​of 25 measurement points, and the last 6 input nodes receive the average value of the initial thickness data of the liquid film in 6 regions.

[0014] The phase field model interface tracking algorithm module introduces a phase field variable to characterize the interface position between the liquid film and air. The phase field variable takes a value of 1 in the liquid phase region and a value of 0 in the gas phase region, and smoothly transitions from 0 to 1 at the interface.

[0015] The phase-field model interface tracking algorithm module establishes the spatiotemporal evolution equation of the phase-field variables. This equation includes a diffusion term, a convection term, and an interface energy term. The diffusion coefficient of the diffusion term is set as follows: to .

[0016] The velocity field of the convection term is obtained by calculating the surface tension gradient from the temperature distribution data. The calculation coefficient of the surface tension gradient is -0.1 mN / m·K. The surface tension coefficient of the interface energy term is set to 0.02 N / m to 0.05 N / m according to the physical properties of the functional liquid.

[0017] The adjustment of the power distribution mode is achieved by changing the heating power of each resistance heating element in the array-type local heating unit. When the Marangoni convection velocity field distribution shows that the flow velocity in a certain area is too high, the heating power of the resistance heating element corresponding to that area is reduced by 10% to 30%.

[0018] The texture direction deflection angle is the angle between the main flow direction of the Marangoni convection velocity field distribution and the preset texture direction. It is obtained by extracting the main flow direction through principal component analysis of the Marangoni convection velocity field distribution. The threshold for the texture direction deflection angle is 15 degrees.

[0019] The mixed solvent is composed of a low-boiling-point component and a high-boiling-point component mixed in a volume ratio of 3:7, wherein the low-boiling-point component is selected from... or High-boiling-point components are selected or The ambient humidity should be controlled between 60% and 75%.

[0020] This invention solves the technical problem of accurately controlling the texture randomness caused by Marangoni convection during liquid film evaporation by constructing a flow field prediction model integrating a phase-field model interface tracking algorithm and combining it with dynamic power adjustment of an array-type local heating unit. The flow field prediction model transforms temperature distribution data into quantitative predictions of Marangoni convection velocity field distribution and liquid film morphology evolution, enabling early identification of the development trend of disordered vortex structures and providing a reliable basis for texture orientation correction. The phase-field model interface tracking algorithm automatically handles the topological changes of liquid film rupture and fusion by solving spatiotemporal evolution equations, accurately quantifying the continuously changing liquid film boundary positions into the spatial distribution of phase-field variables, eliminating the numerical instability of traditional interface tracking methods. The array-type local heating unit adjusts the heating power and temperature gradient distribution of each region in real time based on the flow field prediction results, suppressing the development of small-scale vortex structures and guiding the main flow direction towards the preset texture direction. The reverse Marangoni flow generated by the selective evaporation mechanism of the mixed solvent effectively counteracts the edge accumulation phenomenon caused by the coffee ring effect. In summary, the present invention solves the technical problem mentioned in the background art of the difficulty in accurately controlling the texture randomness caused by Marangoni convection during liquid film evaporation. Attached Figure Description

[0021] Figure 1 This is a flowchart of the method of the present invention.

[0022] Figure 2 This is a vector diagram showing the predicted velocity field distribution of the Marangoni convection.

[0023] Figure 3 This is a graph showing the spatiotemporal evolution of the velocity field of the reverse Marangoni flow. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below.

[0025] like Figure 1 The diagram shows a flowchart of an aluminum plate texture optimization method based on surface tension adjustment provided by the present invention. This method includes the following steps:

[0026] S01. Plasma treatment is performed on the surface of the aluminum plate to reduce the surface roughness to below 50nm, and an array-type local heating unit is constructed on the treated aluminum plate surface. The array-type local heating unit consists of multiple independently controlled resistance heating elements with a spacing of 5mm between the heating elements.

[0027] S02. Prepare a functional liquid, wherein the functional liquid is composed of a base solvent, a surfactant mixture, and a trace polymer, wherein the surfactant mixture has an HLB value of 8 to 12, and the trace polymer has a mass fraction of 0.3% to 0.8%;

[0028] S03. The prepared functional liquid is sprayed onto the surface of the aluminum plate to form a liquid film. At the same time, the closed-loop temperature control system is started. The closed-loop temperature control system integrates multi-point platinum resistance sensors and uses a PID control algorithm to control the temperature fluctuation of the substrate within ±0.2℃.

[0029] S04. Input the temperature distribution data of the liquid film surface into the flow field prediction model. The flow field prediction model outputs the Marangoni convection velocity field distribution and liquid film morphology evolution prediction results. Adjust the power distribution mode of the array-type local heating unit according to the Marangoni convection velocity field distribution and liquid film morphology evolution prediction results.

[0030] S05. When the texture direction deflection angle output by the flow field prediction model exceeds 15 degrees, increase the concentration of the eddy suppression polymer in the corresponding region to 1.2%, and at the same time reduce the temperature gradient in the corresponding region to below 3K / mm to achieve texture orientation correction.

[0031] S06. Control the ambient humidity to 60% to 75% and establish selective evaporation conditions for the mixed solvent, wherein the mixed solvent is a mixture of low-boiling-point components and high-boiling-point components in a volume ratio of 3:7, and the reverse Marangoni flow generated by the preferential evaporation of the low-boiling-point components is used to counteract the coffee ring effect.

[0032] The plasma treatment employs Qi and The mixture of gases has a gas flow ratio of 4:1, a processing power of 200W to 300W, and a processing time of 180s to 240s. The surface oxide layer is removed and the surface microstructure is reconstructed by bombardment with high-energy particles.

[0033] The HLB value is the hydrophilic-lipophilic balance value, which characterizes the hydrophilic-lipophilic properties of the surfactant. Surfactant mixtures with HLB values ​​in the range of 8 to 12 can preferentially adsorb onto the rough peaks on the surface of the aluminum plate, smoothing the effective contact surface between the liquid and the solid, and reducing the discontinuous liquid expansion phenomenon caused by the contact line pinning effect.

[0034] The trace polymer is a polyvinylpyrrolidone or polyacrylamide polymer with a molecular weight of [missing information]. to By increasing the viscoelasticity of the liquid, small-scale vortex structures in Marangoni convection are suppressed, thus reducing texture randomness.

[0035] The array-type local heating unit achieves programmable temperature field distribution by independently controlling the current of each resistance heating element. The temperature response time of the resistance heating element is less than 2s, and the spatial temperature resolution reaches 5mm. It can construct complex two-dimensional temperature gradient patterns to guide the flow direction of the liquid film.

[0036] In the PID control algorithm of the closed-loop temperature control system, the proportional coefficient is set to 0.5 to 0.8, the integral time constant is set to 10s to 20s, and the derivative time constant is set to 2s to 5s. The closed-loop temperature control system collects temperature signals from multi-point platinum resistance sensors in real time and calculates the control deviation. It achieves precise temperature control by adjusting the heating power.

[0037] The temperature distribution data of the liquid film surface is obtained by a multi-point platinum resistance sensor at 25 measurement points on the aluminum plate surface, with a sampling frequency of 10Hz. The temperature distribution data includes the temperature value and spatial coordinate value of each measurement point.

[0038] The flow field prediction model is structured such that an input layer receives temperature distribution data and initial liquid film thickness data from the liquid film surface. This input layer contains 256 input nodes, with the first 250 nodes receiving temperature and spatial coordinate values ​​from 25 measurement points, and the last 6 nodes receiving the average value of the initial liquid film thickness data across 6 regions. A phase-field model interface tracking algorithm module is positioned between the input layer and the first hidden layer. This module performs interface morphology evolution calculations on the temperature distribution data and initial liquid film thickness data from the input layer, and incorporates phase-field variables. The phase field variable characterizes the interface position between the liquid film and air. The phase field model interface tracking algorithm module establishes phase field variables by taking a value of 1 in the liquid phase region and a value of 0 in the gas phase region, with a smooth transition from 0 to 1 at the interface. The spatiotemporal evolution equation describes the change of the interface position over time. This equation includes a diffusion term, a convection term, and an interface energy term. The diffusion coefficient of the diffusion term is set to... to The control interface thickness ranges from 0.1 mm to 0.3 mm. The velocity field of the convection term is calculated from the temperature distribution data using the surface tension gradient, with a calculation coefficient of -0.1 mN / m·K. The surface tension coefficient of the interface energy term is set to 0.02 N / m to 0.05 N / m based on the physical properties of the functional liquid. The phase field model interface tracking algorithm module obtains the phase field variables by numerically solving the spatiotemporal evolution equation. The spatial distribution values ​​at 32 grid points, and based on the phase field variables. The spatial distribution values ​​are used to extract the liquid film boundary position coordinates and liquid film thickness distribution data, wherein the liquid film boundary position coordinates are defined as phase field variables. The set of spatial points with a value of 0.5, the liquid film thickness distribution data is obtained by adjusting the phase field variables. The phase-field model interface tracking algorithm module outputs the liquid film boundary position coordinates and liquid film thickness distribution data as enhanced features to the first hidden layer after vertical integration. The first hidden layer contains 128 neurons and uses a modified linear unit activation function. It simultaneously receives the original features output from the 256 input nodes of the input layer and the enhanced features composed of the liquid film boundary position coordinates and liquid film thickness distribution data output from the phase-field model interface tracking algorithm module. The first hidden layer fuses the original features and enhanced features through a fully connected method, outputting a flow field feature vector containing interface morphology information. The second hidden layer contains 64 neurons and uses a modified linear unit activation function. The second hidden layer receives the flow field feature vector output from the first hidden layer. The flow field feature vector is processed by the second hidden layer, which performs a nonlinear transformation to extract the velocity components of the Marangoni convection. These velocity components include x-direction and y-direction velocity components. The third hidden layer contains 32 neurons and receives the velocity components output by the second hidden layer. The third hidden layer performs temporal correlation analysis on these velocity components to predict future changes in liquid film morphology. The output layer contains 130 output nodes, which output the Marangoni convection velocity field distribution and liquid film morphology evolution prediction results. The first 64 output nodes output the x-direction and y-direction velocity components of the Marangoni convection velocity field distribution at 32 grid points, while the last 66 output nodes output the liquid film morphology evolution prediction results, including the liquid film thickness field distribution and texture direction deflection angle for the next 10 time steps.

[0039] The initial thickness data of the liquid film was obtained immediately after the functional liquid spraying was completed using a laser triangulation distance sensor. The measurement range covered six areas on the aluminum plate surface, each area being 50 mm. .

[0040] The steps for establishing the training dataset for the flow field prediction model include: conducting liquid film flow experiments under different temperature gradient conditions and recording the Marangoni convection velocity field distribution data using particle image velocimetry; simultaneously measuring the liquid film thickness evolution process using laser confocal microscopy; using the temperature distribution data and initial thickness data of the liquid film surface as input features; and using the Marangoni convection velocity field distribution and liquid film thickness evolution process as label data to establish a training dataset containing 5000 samples; and normalizing the training dataset so that the values ​​of the input features and label data are both between 0 and 1.

[0041] The training steps of the flow field prediction model include: dividing the training dataset into a training set and a validation set in an 8:2 ratio; updating the weight parameters of the flow field prediction model using an adaptive moment estimation optimization algorithm; setting the initial learning rate to 0.001 and decreasing exponentially with each training round; setting the batch size to 32; stopping training after 300 rounds when the validation set loss function no longer decreases for 20 consecutive rounds; the loss function is composed of a weighted sum of the mean square error terms of the Marangoni convection velocity field distribution and the mean square error terms of the liquid film thickness field distribution, with weighting coefficients of 0.6 and 0.4, respectively.

[0042] The phase-field model interface tracking algorithm module provides accurate liquid film morphology evolution information for the flow field prediction model, enabling the model to predict the actual morphology and texture formation location of the liquid film after temperature distribution data adjustment, thus avoiding texture defects caused by liquid film rupture or uneven contraction. The phase-field model interface tracking algorithm module uses smooth phase-field variables to represent continuous physical interfaces. This indicates that the numerical instability of traditional interface tracking methods in handling interface merging and separation is eliminated. By solving the spatiotemporal evolution equation, the topological changes of liquid film rupture and fusion are automatically handled without explicit assumptions about the interface shape. The velocity field of the convection term in the spatiotemporal evolution equation is obtained from temperature distribution data through surface tension gradient calculation, realizing the physical coupling between the temperature field and interface evolution. When the temperature distribution data changes, the surface tension gradient changes accordingly, driving the phase field variables. The distribution adjustment reflects the real-time displacement of the liquid film boundary. The diffusion term ensures the phase field variables. A smooth transition at the interface avoids numerical oscillations during interface tracking, and the set range of the diffusion coefficient ensures that the interface thickness remains at a physically reasonable scale. The interface energy term maintains the physical constraint of the liquid film surface tension, ensuring that the predicted liquid film morphology conforms to the principle of energy minimization. The phase-field model interface tracking algorithm module transmits the interface morphology information of the liquid film to the first hidden layer through the liquid film boundary position coordinates and liquid film thickness distribution data, enabling the flow field prediction model to incorporate interface physical properties at the feature extraction stage, significantly improving the accuracy of the flow field prediction model in predicting the dynamic behavior of liquid films driven by Marangoni convection. The phase-field model interface tracking algorithm module can predict the liquid film boundary movement and thickness changes caused by temperature field adjustments, providing a reliable theoretical basis for the real-time optimization and adjustment of the power distribution mode of the array-type local heating unit, ensuring the orientation and repeatability of the texture pattern.

[0043] The adjustment of the power distribution mode is achieved by changing the heating power of each resistance heating element in the array-type local heating unit. When the Marangoni convection velocity field distribution output by the flow field prediction model shows that the flow velocity in a certain area is too high, the heating power of the resistance heating element corresponding to that area is reduced by 10% to 30%. When the liquid film morphology evolution prediction result output by the flow field prediction model shows that the liquid film thickness in a certain area is less than 0.05 mm, the heating power of the resistance heating element corresponding to that area is increased by 15% to 40%.

[0044] The texture direction deflection angle is the angle between the main flow direction of the Marangoni convection velocity field distribution and the preset texture direction. It is obtained by extracting the main flow direction through principal component analysis of the Marangoni convection velocity field distribution. When the texture direction deflection angle exceeds 15 degrees, it indicates that the Marangoni convection exhibits a significant disordered vortex structure.

[0045] The increase in the concentration of the vortex-suppressing polymer is achieved by supplementing the corresponding area with a high-concentration polymer solution. The polymer mass fraction of the high-concentration polymer solution is 5% to 8%. The amount of supplementation is precisely controlled by a micro-injection device, so that the micro-polymer concentration in the corresponding area is increased to 1.2% to suppress the development of small-scale vortices.

[0046] The temperature gradient reduction in the corresponding region is achieved by reducing the heating power of the resistance heating element in the array-type local heating unit of the corresponding region. The heating power is reduced by 30% to 50%, so that the temperature gradient in the corresponding region is reduced from the initial above 5K / mm to below 3K / mm, thereby weakening the Marangoni convection intensity in the corresponding region.

[0047] The low-boiling-point component of the mixed solvent is selected from... or The boiling point is 78°C to 82°C, and the high-boiling-point component of the mixed solvent is selected from... or The boiling point is 180°C to 195°C. The low-boiling-point component preferentially volatilizes in the early stage of evaporation, which leads to a decrease in the surface temperature of the liquid film. A temperature difference is formed between the edge and the center of the liquid film, generating a surface tension gradient from the edge to the center, which drives the liquid to flow back.

[0048] The reverse Marangoni flow is a flow from the edge of the liquid film toward the center. The reverse Marangoni flow carries solute from the edge to the center, compensating for the outward flow driven by evaporation, so that the deposition distribution of solute after the liquid film dries tends to be uniform, and the ratio of solute concentration at the edge to solute concentration at the center is reduced to below 1.2.

[0049] The coffee ring effect refers to the phenomenon where solute migrates to the edge and accumulates excessively in the edge region during the liquid film evaporation process, resulting in a solute concentration in the edge region that is 3 to 5 times higher than that in the center region, causing excessive thickening of the texture boundary.

[0050] The specific implementation methods of the above steps are described in detail below.

[0051] The specific implementation of step S01 is as follows: First, the aluminum plate is fixed in the vacuum chamber of the plasma processing equipment, and a vacuum is drawn until the chamber pressure is below 10 Pa to establish a low-pressure environment. Then, air is introduced into the chamber. Qi and A mixture of gases is precisely controlled by a mass flow controller to maintain a 4:1 flow ratio between the two gases, stabilizing the chamber pressure between 50 Pa and 80 Pa. Then, a radio frequency (RF) power supply is activated to ionize the gas mixture and generate plasma. The RF power is set to 200 W to 300 W, producing high-energy plasma. Ions and Atoms, accelerated by an electric field, bombard the surface of an aluminum plate, removing the oxide layer and organic contaminants. Atoms react with the aluminum substrate to form a uniform aluminum oxide thin layer. The processing time is controlled between 180s and 240s. During the processing, the luminescence intensity of the plasma is monitored by an optical emission spectrometer to determine the surface cleanliness. When the intensity of the characteristic spectral lines reaches a stable value, the surface treatment is completed. After the treatment, the surface roughness of the aluminum plate is measured by an atomic force microscope, confirming that the roughness has been reduced to below 50nm. The plasma treatment utilizes the dual effects of physical bombardment and chemical reaction of high-energy particles to reconstruct the microstructure of the aluminum plate surface, eliminating the inherent peak and valley structure of the surface and providing a flat substrate for the uniform expansion of the subsequent liquid film. After processing, multiple resistance heating elements are attached to the surface of the aluminum plate in a square array with a 5mm spacing. The resistance heating elements are thin-film resistors with a thickness of less than 0.5mm. Each resistance heating element is connected to a programmable power supply through an independent current control channel. The programmable power supply outputs different current values ​​according to the preset heating mode. The resistance of the resistance heating elements is 10Ω to 20Ω. When current is applied, heat is generated according to the Joule heating effect. Thermal grease is applied between the resistance heating elements and the surface of the aluminum plate to improve heat conduction efficiency. The construction of the array-type local heating unit enables the formation of a programmable two-dimensional temperature distribution on the surface of the aluminum plate, providing a hardware basis for actively guiding the Marangoni convection direction.

[0052] The specific implementation of step S02 involves preparing the functional liquid on a thermostatic magnetic stirrer. First, a base solvent is added to a beaker as a dispersion medium. The base solvent is selected from deionized water or low molecular weight alcohols, with a volume of 200 mL to 300 mL. Then, a surfactant mixture is added to the base solvent. This surfactant mixture consists of a nonionic surfactant and anionic surfactant mixed in a mass ratio of 6:4. The HLB value of the mixture is adjusted to 8 to 12 by adjusting the ratio of the two types of surfactants. The amount added is 2% to 5% of the volume of the base solvent. The surfactant mixture can reduce the surface tension of the functional liquid and change the wetting properties of the solid-liquid interface. Next, a trace polymer is added to the solution. The trace polymer is polyvinylpyrrolidone or polyacrylamide powder with a molecular weight of [missing information]. to The amount added is such that the mass fraction is 0.3% to 0.8%. Start the magnetic stirrer and stir at 600 rpm for 60 to 90 minutes to ensure that the trace polymer is completely dissolved and uniformly dispersed in the solution. During the stirring process, the viscosity change of the solution is monitored in real time using a rheometer. When the viscosity reaches a stable value, it indicates that the trace polymer has been fully dissolved. The long chain structure of the trace polymer increases the viscoelasticity of the functional liquid, which can dissipate the kinetic energy of small-scale eddies during liquid film flow, thereby suppressing turbulent pulsation in Marangoni convection. After preparation, the functional liquid is transferred to a sealed container and allowed to stand for 12 hours to remove air bubbles in the solution. During the standing process, the temperature is kept constant at 25°C to avoid solvent evaporation.

[0053] The specific implementation of step S03 is as follows: the prepared functional liquid is filled into the reservoir of the pneumatic spray gun, the nozzle diameter of the pneumatic spray gun is adjusted to 0.5mm to 0.8mm, the spraying pressure is set to 0.2MPa to 0.3MPa, the spraying distance is 150mm to 200mm, and the functional liquid is sprayed on the surface of the aluminum plate in a uniform scanning manner with a scanning speed of 50mm / s to 80mm / s. The initial thickness of the liquid film is made to reach 0.2mm to 0.5mm by controlling the spraying time. During the spraying process, the nozzle is kept perpendicular to the surface of the aluminum plate to ensure the uniformity of the liquid film thickness. Simultaneously with the spraying process, a closed-loop temperature control system is activated. This system consists of 25 platinum resistance sensors, a data acquisition card, an industrial control computer, and a programmable power supply. The platinum resistance sensors are arranged in a square array with a 5mm spacing on the surface of the aluminum plate. Each platinum resistance sensor has a measurement accuracy of ±0.1℃. The data acquisition card reads the resistance value of the platinum resistance sensor at a sampling frequency of 10Hz and converts it into a temperature value. The industrial control computer runs a PID control algorithm to calculate the heating power required for each heating element in real time. The proportional term of the PID control algorithm calculates the instantaneous response based on the deviation between the current temperature and the target temperature, with the proportional coefficient set to 0.5 to 0.8. The integral term accumulates and compensates for historical deviations to eliminate steady-state errors, with the integral time constant set to 10 to 20 seconds. The derivative term predicts future deviations based on the rate of temperature change and adjusts the output in advance, with the derivative time constant set to 2 to 5 seconds. The control signal output by the PID control algorithm drives the programmable power supply to adjust the power supply current of each resistive heating element, achieving precise control of the aluminum plate surface temperature. This keeps the substrate temperature fluctuation within ±0.2℃. The closed-loop temperature control system eliminates the interference of ambient temperature fluctuations and uneven heat conduction on the liquid film flow, ensuring the controllability of the surface tension gradient and the repeatability of texture formation.

[0054] The specific implementation of step S04 involves immediately collecting temperature distribution data on the surface of the liquid film after spraying. This temperature distribution data is obtained in real-time by 25 platinum resistance sensors in a closed-loop temperature control system. The temperature value and spatial coordinates of each measurement point form a data tuple, and the 25 data tuples constitute a complete temperature distribution data matrix. Simultaneously, a laser triangulation sensor is used to measure the initial thickness data of the liquid film. The measurement range covers six regions defined on the aluminum plate surface, each region having an area of ​​50 mm. The sensor emits a laser beam towards the liquid film surface and calculates the liquid film thickness by receiving the reflected light. The arithmetic mean of the thickness values ​​at multiple measurement points within each region is then used to obtain the region's average thickness. Temperature distribution data and initial liquid film thickness data are input into a flow field prediction model. This model first organizes the temperature, spatial coordinates, and thickness values ​​into a 256-dimensional input vector at the input layer. This input vector is then passed to the phase-field model interface tracking algorithm module for interface morphology evolution calculation. This module establishes phase-field variables describing the position of the liquid film-air interface. The spatiotemporal evolution equation is constructed based on the Cahn-Hilliard theory or the Allen-Cahn theory. The diffusion term in the equation describes the natural diffusion behavior of the interface, and the diffusion coefficient is set to... to To maintain the interface thickness within a physically reasonable range of 0.1 mm to 0.3 mm, the convection term in the equation describes the interface displacement driven by fluid motion. The convection velocity is calculated from the temperature distribution data using the surface tension gradient. The dependence of surface tension on temperature is approximated linearly, and the gradient calculation coefficient is set to -0.1 mN / m·K. The interface energy term in the equation describes the effect of surface tension at the interface, and the surface tension coefficient is set to 0.02 N / m to 0.05 N / m based on the physical properties of the functional liquid. The phase-field model interface tracking algorithm module uses the finite difference method to numerically solve the spatiotemporal evolution equation. The aluminum plate surface is discretized into 32 grid points, and the time step is set to 0.01 s to 0.05 s. The phase-field variables of each grid point in the future multiple time steps are obtained through iterative calculation. The value, based on the phase field variables The position equal to 0.5 determines the coordinates of the liquid film boundary, by adjusting the phase field variables. The liquid film thickness distribution data is obtained by numerical integration in the vertical direction. The phase field model interface tracking algorithm module outputs the liquid film boundary position coordinates and liquid film thickness distribution data as enhanced features. The first hidden layer of the flow field prediction model receives the original features from the input layer and the enhanced features from the phase field model interface tracking algorithm module. Feature fusion is performed through a fully connected layer of 128 neurons. Each neuron calculates the weighted sum of the input features and processes them through a modified linear unit activation function to output a flow field feature vector containing interface morphology information. The 64 neurons of the second hidden layer perform nonlinear transformation on the flow field feature vector to extract the velocity components of Marangoni convection. The 32 neurons of the third hidden layer perform temporal correlation analysis on the velocity components to predict future liquid film morphology changes. The 130 output nodes of the output layer output the predicted results of Marangoni convection velocity field distribution and liquid film morphology evolution. The flow intensity of each region is determined based on the Marangoni convection velocity field distribution output by the flow field prediction model. When the flow velocity in a certain region is too high and exceeds the threshold of 20 mm / s, the heating power of the corresponding resistance heating element in that region is reduced by 10% to 30% to weaken the flow intensity. When the liquid film morphology evolution prediction result shows that the liquid film thickness in a certain region is lower than the threshold of 0.05 mm, the heating power of the corresponding resistance heating element in that region is increased by 15% to 40% to supplement the liquid film thickness. The adjustment of the power distribution mode is achieved by changing the supply current of each resistance heating element in real time through a programmable power supply. The adjusted temperature field reguides the direction and intensity of Marangoni convection, causing the liquid film flow to tend towards the preset texture direction.

[0055] The specific implementation of step S05 is to perform principal component analysis on the Marangoni convection velocity field distribution output by the flow field prediction model to extract the main flow direction. The principal component analysis method calculates the covariance matrix of the velocity field and solves its eigenvalues ​​and eigenvectors. The eigenvector corresponding to the largest eigenvalue represents the main direction of the velocity field. The vector angle between the main direction and the preset texture direction is calculated to obtain the texture direction deflection angle. When the texture direction deflection angle exceeds the threshold of 15 degrees, it is determined that the Marangoni convection has a significant disordered vortex structure and texture orientation correction is required. Texture orientation correction is achieved through two synergistic measures. First, the concentration of vortex-suppressing polymer in the deflection region is increased. A high-concentration polymer solution with a polymer mass fraction of 5% to 8% is added to the deflection region using a micro-injection device. The injection volume is precisely controlled by a micro-pump, and the required replenishment amount is calculated based on the area of ​​the deflection region and the thickness of the liquid film, thereby increasing the micro-polymer concentration in the deflection region from the initial 0.3% to 0.8% to 1.2%. The increase in micro-polymer concentration enhances the viscoelasticity of the liquid film. Viscoelasticity dampens small-scale vortices, consuming their rotational kinetic energy and suppressing their development and propagation. Simultaneously, the temperature gradient in the deflection region is reduced. The heating power of the resistance heating element corresponding to the deflection region is reduced by 30% to 50% through the programmable power supply of the closed-loop temperature control system. The reduction in heating power reduces the temperature difference between the deflection region and the surrounding region, and the temperature gradient is reduced from the initial value of more than 5 K / mm to less than 3 K / mm. The reduction in temperature gradient directly reduces the surface tension gradient. According to the Marangoni effect theory, the surface tension gradient is the driving force for liquid flow. The reduction in surface tension gradient reduces the intensity of Marangoni convection and weakens the generation of disordered vortices. The synergistic effect of the two measures simultaneously suppresses the vortex structure from both dynamic and rheological perspectives, allowing the liquid film flow to return to a laminar state and flow orderly along the preset texture direction.

[0056] The specific implementation of step S06 involves adjusting the humidity of the liquid film evaporation environment through an environmental control system. This system consists of a humidifier, a dehumidifier, and a humidity sensor forming a closed-loop control circuit. The humidity sensor monitors the ambient humidity in real time and feeds the signal back to the controller. The controller adjusts the operating status of the humidifier or dehumidifier according to a set value of 60% to 75%, stabilizing the ambient humidity within the target range. Controlling the ambient humidity affects the partial pressure of water vapor on the liquid film surface, thereby altering the evaporation rate. Higher ambient humidity reduces the water vapor concentration gradient between the liquid film surface and the environment, establishing restricted evaporation conditions and making the evaporation process slower and more uniform. Simultaneously, the selective evaporation characteristics of the mixed solvent are used to counteract the coffee ring effect. The mixed solvent consists of low-boiling-point components. or With high-boiling-point components or The components are mixed in a volume ratio of 3:7. The boiling point of the low-boiling-point component is 78℃ to 82℃, and the boiling point of the high-boiling-point component is 180℃ to 195℃. This difference in boiling points leads to a significant difference in evaporation rates. In the initial stage of liquid film evaporation, the low-boiling-point component preferentially evaporates from the liquid film surface. The latent heat absorbed during evaporation causes the surface temperature of the liquid film to decrease. Due to its larger surface area, the evaporation rate is faster at the edge of the liquid film, resulting in a more significant temperature decrease. A temperature difference is formed between the edge and the center of the liquid film, creating a surface tension gradient. Since surface tension decreases with increasing temperature, the lower temperature at the edge of the liquid film... In the high-temperature zone where the surface tension is greater than that at the center, the surface tension gradient drives the liquid to flow from the edge to the center, forming a reverse Marangoni flow. The direction of the reverse Marangoni flow is opposite to the outward flow direction driven by conventional evaporation. The reverse Marangoni flow carries solute from the edge to the center, compensating for the solute migration to the edge caused by evaporation. This makes the solute deposition distribution after the liquid film dries more uniform, and the ratio of solute concentration at the edge to solute concentration at the center decreases from 3 to 5 times that of the coffee ring effect to less than 1.2, eliminating the excessive thickening of the texture boundary.

[0057] It should be noted that one of the key technical ideas of this invention is to embed the phase-field model interface tracking algorithm module between the input layer and the first hidden layer of the flow field prediction model. By introducing phase-field variables to describe the continuous evolution process of the liquid film-air interface, the phase-field method treats the interface as a transition region with finite thickness, avoiding the difficulty of explicitly tracking the interface position required by traditional interface tracking methods. The spatiotemporal evolution equation established by the phase-field model interface tracking algorithm module naturally includes the physical mechanisms of topological changes of the interface, such as liquid film rupture and fusion. By numerically solving the spatiotemporal evolution equation, the spatiotemporal evolution information of the interface position and liquid film thickness is obtained. This information is used as an enhanced feature input to the neural network, enabling the flow field prediction model to incorporate the physical characteristics of the interface in the feature extraction stage. This significantly improves the model's prediction accuracy for the liquid film dynamics driven by Marangoni convection. Compared with the traditional pure data-driven neural network model, the flow field prediction model with enhanced physical information can follow the basic conservation laws of fluid mechanics and still give prediction results that conform to physical laws even with limited training data, providing a reliable basis for the real-time optimization and adjustment of array-type local heating units.

[0058] The second key technical concept of this invention is to construct an array-type local heating unit to achieve programmable temperature field distribution and combine it with a closed-loop temperature control system to achieve high-precision temperature regulation. Traditional uniform heating methods can only generate a single temperature gradient direction and cannot actively guide liquid film flow. However, the array-type local heating unit can construct arbitrarily complex two-dimensional temperature distribution patterns on the aluminum plate surface by independently controlling the power of each resistance heating element. The temperature distribution directly determines the surface tension distribution. According to the Marangoni effect theory, the surface tension gradient is the driving force for liquid film flow. By designing a specific temperature distribution pattern, Marangoni convection can be actively guided to flow in a preset direction, achieving directional control of the texture. The closed-loop temperature control system uses a PID control algorithm to control temperature fluctuations within ±0.2℃, eliminating the interference of ambient temperature fluctuations on the surface tension gradient and ensuring the repeatability of liquid film flow behavior. Compared with the traditional open-loop heating method, the closed-loop temperature control system can compensate for the effects of uneven heat conduction and environmental disturbances in real time, maintaining the spatiotemporal stability of the temperature field.

[0059] The third key technical idea of ​​this invention is to counteract the coffee ring effect by generating a reverse Marangoni flow through the selective evaporation of a mixed solvent. In traditional single-solvent systems, capillary flow from the center to the edge inevitably occurs during evaporation, carrying solutes to the edge and causing excessive accumulation at the edge. However, the mixed solvent system utilizes the difference in evaporation rates between low-boiling-point and high-boiling-point components to establish a reverse temperature gradient on the liquid film surface. The low-boiling-point component preferentially evaporates and absorbs latent heat, resulting in a lower temperature in the rapidly evaporating edge region than in the slowly evaporating center region. The surface tension gradient generated by the temperature difference drives the liquid to flow back from the edge to the center, forming a reverse Marangoni flow. The reverse Marangoni flow is opposite to the direction of capillary flow and can carry solutes back to the center region, achieving a dynamic balance of solute transport and making the solute deposition distribution more uniform. Compared with the traditional method of adding macromolecular substances to suppress the coffee ring, the mixed solvent selective evaporation method fundamentally eliminates the driving force for solute migration to the edge and avoids complex effects on the rheological properties of the liquid film without introducing additional additives.

[0060] The synergistic effect of the three key technical approaches forms a complete liquid film flow control system. The phase-field model interface tracking algorithm module provides a predictive basis for the temperature field optimization of the array-type local heating unit, enabling temperature field adjustment to specifically avoid liquid film rupture and uneven thickness. The programmable temperature field generated by the array-type local heating unit not only guides the Marangoni convection direction but also works synergistically with the selective evaporation of the mixed solvent. By adjusting the local temperature, it affects the evaporation rate of low-boiling-point components, thereby controlling the intensity of the reverse Marangoni flow. The reverse Marangoni flow generated by the selective evaporation of the mixed solvent compensates for the temperature-driven forward flow, establishing a stable circulation structure inside the liquid film. The circulation structure can both transport solutes to achieve uniform deposition and maintain the uniformity of liquid film thickness. The synergistic effect of the three technical approaches ensures that the liquid film is in a precisely controllable state throughout the entire process from spraying to drying, achieving a comprehensive improvement in the orientation, uniformity, and repeatability of texture patterns. Compared with the traditional method of passively relying on the natural flow of the liquid film to form textures, the active control system established in this invention can dynamically adjust process parameters based on real-time feedback information, adapting to different substrate conditions and environmental conditions, significantly improving the robustness and engineering practicality of the texture optimization method.

[0061] It should be noted that this invention also solves the following technical problem: the inability to track the evolution path of the liquid film boundary in real time, leading to difficulties in preventing texture defects. This invention embeds a phase-field model interface tracking algorithm module into the flow field prediction model, introducing phase-field variables to characterize the interface position between the liquid film and air, and establishing a spatiotemporal evolution equation containing diffusion, convection, and interface energy terms to describe the dynamic changes of the interface over time. The phase-field model obtains the distribution values ​​of phase-field variables at spatial grid points by numerically solving the spatiotemporal evolution equation, extracting the liquid film boundary position coordinates and liquid film thickness distribution data, and passing this interface morphology information as enhanced features to the hidden layer of the neural network for fusion processing. When the flow field prediction model output shows that the liquid film thickness in a certain area is below a critical threshold, the system immediately increases the power of the corresponding resistance heating element in that area, preventing liquid film rupture by enhancing the control of the local evaporation rate, thereby implementing preventative intervention before texture defects form and ensuring the integrity and uniformity of the texture pattern.

[0062] Specifically, the principle of this invention is as follows: By embedding a physically driven phase-field model into a neural network architecture, this invention achieves accurate prediction and real-time control of liquid film dynamics, thereby solving the technical problem of uncontrollable texture randomness. The phase-field model interface tracking algorithm transforms temperature distribution data into a convection velocity field driving interface evolution through surface tension gradient calculation, establishing a deterministic physical coupling relationship between the temperature field and the liquid film morphology, enabling quantitative prediction of the impact of heating power adjustment on texture formation. The diffusion term ensures a smooth transition of phase-field variables at the interface, while the interface energy term maintains the physical constraint of surface tension, ensuring that the predicted liquid film morphology conforms to the principle of energy minimization, avoiding prediction bias caused by the lack of physical mechanisms in purely data-driven models. The flow field prediction model incorporates the liquid film boundary position and thickness distribution output by the phase-field model as enhancement features into the feature extraction process of the neural network, enabling the model to identify vortex structure features that cause texture deflection. The array-type local heating unit dynamically adjusts the temperature gradient of each region based on the flow field prediction results, achieving directional correction of texture direction by reducing excessive Marangoni convection intensity and supplementing vortex-suppressing polymer concentration. The preferential evaporation of low-boiling-point components in the mixed solvent results in a temperature drop at the edge of the liquid film, forming a surface tension gradient from the edge to the center. This drives liquid reflux to counteract the outward evaporation flow, making the solute deposition distribution more uniform.

[0063] The following provides a specific embodiment 1 of the present invention. The specific implementation of steps S01 and S02 in this embodiment 1 is the same as that described above, and will not be repeated in detail here. The specific implementation of other steps is described in detail below.

[0064] The specific implementation of step S03 is as follows: The prepared functional liquid is sprayed onto the surface of the aluminum plate to form a liquid film. Simultaneously, a closed-loop temperature control system is activated. This system integrates multi-point platinum resistance sensors and employs a proportional-integral-derivative (PID) control algorithm to control the substrate temperature fluctuation within ±0.2℃. The control output formula of the PID control algorithm is expressed as follows:

[0065] ;

[0066] In the formula, To control output power, the unit is W; Temperature control deviation, in °C; The reference temperature deviation is set at 1℃; For reference time, the value is 1 second; The reference power is set to 1W. This is a proportionality coefficient, with an empirical value of 0.5 to 0.8; The integral coefficients are obtained through the integration time constant. The calculation yields the following formula: , The value ranges from 10s to 20s; The differential coefficients are obtained by differentiating the time constant. The calculation yields the following formula: , The value ranges from 2s to 5s; The variable is the time integral variable, with units in seconds (s). The current time is expressed in seconds. The closed-loop temperature control system acquires temperature signals from multiple platinum resistance sensors in real time and calculates the control deviation. Control deviation The calculation formula is In the formula To set the temperature value, the unit is ℃. This refers to the actual temperature value measured by a multi-point platinum resistance sensor, in °C. The heating power is adjusted accordingly. Achieve precise temperature control.

[0067] The specific implementation of step S04 is as follows: Temperature distribution data of the liquid film surface is input into the flow field prediction model. The flow field prediction model outputs the Marangoni convection velocity field distribution and liquid film morphology evolution prediction results. Based on the Marangoni convection velocity field distribution and liquid film morphology evolution prediction results, the power distribution mode of the array-type local heating unit is adjusted. Phase field variables in the phase field model interface tracking algorithm module... The spatiotemporal evolution equation is expressed as follows:

[0068] ;

[0069] In the formula, This is a phase field variable, taking a value of 1 in the liquid phase region and a value of 0 in the gas phase region; The variable is time, and the unit is seconds (s). The time feature scale is set to 1 second. This is the convective velocity field vector, with units of m / s; The velocity characteristic scale is set to a value of [value]. m / s; Phase field variables The gradient operator; The diffusion coefficient has a range of values ​​of 1000. to ; The length characteristic scale is set to 0.01m. Phase field variables The Laplace operator; The interfacial energy coefficient is set to 0.02 N / m to 0.05 N / m based on the physical properties of the functional liquid. This is the interface thickness parameter, with a value ranging from 0.1mm to 0.3mm; The density feature scale is set to a value of 1. Convection velocity field vector The calculation formula is expressed as follows:

[0070] ;

[0071] In the formula, The surface tension temperature coefficient is -0.1 mN / m·K. This is the temperature gradient vector, with units of K / m, obtained from the temperature distribution data on the liquid film surface through differential calculation; The functional fluid dynamic viscosity is, empirically, [value]. Pa·s. Coordinates of the liquid film boundary position. The extraction formula is expressed as follows:

[0072] ;

[0073] In the formula, This is a spatial position vector, with units of mm; This is the set of coordinates for the liquid film boundary. Liquid film thickness distribution data. The calculation formula is expressed as follows:

[0074] ;

[0075] In the formula, The liquid film thickness is expressed in mm. This is a position vector parallel to the surface of the aluminum plate, in mm. The coordinates are perpendicular to the surface of the aluminum plate, and the unit is mm; and These are the lower and upper limits of the integration, typically set to 0mm and 2mm respectively.

[0076] The specific implementation of step S05 is as follows: When the texture direction deflection angle output by the flow field prediction model exceeds 15 degrees, the concentration of the eddy current suppressor polymer in the corresponding region is increased to 1.2%, while the temperature gradient in the corresponding region is reduced to below 3K / mm, thereby achieving texture orientation correction. Texture direction deflection angle The calculations were obtained by extracting the main flow direction and deflection angle through principal component analysis of the Marangoni convection velocity field distribution. The calculation formula is expressed as follows:

[0077] ;

[0078] In the formula, This represents the texture direction deflection angle, in degrees. The velocity component in the x-direction of the main flow direction extracted by principal component analysis is expressed in m / s. The velocity component in the y-direction of the main flow direction extracted by principal component analysis, in m / s; This is the preset texture direction angle, in degrees; Pi, with a value of 3.14159.

[0079] The specific implementation of step S06 is as follows: The ambient humidity is controlled to be 60% to 75%, and selective evaporation conditions for the mixed solvent are established. The mixed solvent consists of a low-boiling-point component and a high-boiling-point component mixed in a volume ratio of 3:7. The reverse Marangoni flow generated by the preferential evaporation of the low-boiling-point component counteracts the coffee ring effect. Surface tension gradient. The calculation formula is expressed as follows:

[0080] ;

[0081] In the formula, This represents the surface tension gradient, in N / m². ; The surface tension temperature coefficient has an empirical value of -0.1 mN / m·K. The liquid film surface temperature gradient, in units of K / m, is obtained by the temperature difference between the liquid film edge and center caused by the preferential evaporation of low-boiling-point components, resulting in a decrease in the liquid film surface temperature. The surface tension characteristic scale is set to 0.072 N / m. The ratio of edge solute concentration to central solute concentration. The calculation formula is expressed as follows:

[0082] ;

[0083] In the formula, This is the ratio of the solute concentration at the periphery to the solute concentration at the center. This refers to the solute concentration in the edge region, expressed in g / ; The concentration of solute in the central region is expressed in g / L. The effect of the reverse Marangoni flow makes Reduced to below 1.2.

[0084] The specific implementation method for adjusting the power distribution mode is as follows: When the Marangoni convection velocity field distribution output by the flow field prediction model shows that the flow velocity in a certain region is too high, the heating power of the corresponding resistance heating element in that region is reduced by 10% to 30%. When the liquid film morphology evolution prediction result output by the flow field prediction model shows that the liquid film thickness in a certain region is less than 0.05 mm, the heating power of the corresponding resistance heating element in that region is increased by 15% to 40%. (Heating power adjustment amount) The calculation formula is expressed as follows:

[0085] ;

[0086] In the formula, This refers to the adjustment amount of heating power, expressed in watts (W). This is the power adjustment coefficient, which takes a value of -0.1 to -0.3 when the flow velocity is too high, and a value of 0.15 to 0.4 when the liquid film thickness is too low; This is the initial heating power, expressed in W. The Marangoni convection velocity is a local region, expressed in m / s, and is obtained by averaging the Marangoni convection velocity field distribution in that region as output by the flow field prediction model. The standard flow velocity is, empirically, [value]. m / s.

[0087] It should be noted that the variables involved in this embodiment are explained in detail in Table 1.

[0088] Table 1. Variable Explanation Table

[0089]

[0090] To better understand and implement this invention, a specific application scenario, Example 2, is provided below: The technical team first performs plasma treatment on the surface of an aluminum plate with dimensions of 200mm × 150mm. The treatment equipment adopts... Qi and The gas mixture is precisely controlled at a gas flow ratio of 4:1. The air flow rate is set to 80 sccm. The gas flow rate was set to 20 sccm. The plasma treatment power was set to 250 W, the treatment time to 210 s, and the treatment chamber pressure was maintained at 15 Pa. After treatment, the surface roughness of the aluminum plate was measured using an atomic force microscope. The measurement results showed that the surface roughness decreased from 180 nm before treatment to 42 nm, meeting the requirements of subsequent processes. An array-type local heating unit was constructed on the surface of the treated aluminum plate. This unit consists of 25 independently controlled resistance heating elements arranged in a 5×5 pattern, with a spacing of 5 mm between the heating elements. Each resistance heating element uses a platinum thin-film resistor with a resistance of 50 Ω and a temperature response time of 1.8 s.

[0091] The technical team formulated a functional liquid for texture pattern formation. Deionized water was used as the base solvent, and the surfactant mixture consisted of nonionic surfactants Tween 80 and Span 80 mixed in a 3:2 mass ratio, with a calculated HLB value of 9.8. The trace polymer was polyvinylpyrrolidone (PVP), with a molecular weight of [missing information]. The mass fraction was configured as 0.6%. The total volume of the functional liquid was 50 mL, of which the volume of the base solvent was 49 mL, the mass of the surfactant mixture was 0.8 g, and the mass of polyvinylpyrrolidone was 0.3 g. The mixed solvent was prepared from... and Mix at a volume ratio of 3:7, where The volume is 15 mL. The volume is 35 mL, and the boiling point of the mixed solvent is between 78 °C and 188 °C.

[0092] The technical team used an ultrasonic spraying device to spray the prepared functional liquid onto the surface of the aluminum plate. The spraying pressure was set to 0.3 MPa, the spraying distance to 150 mm, and the spraying speed to 20 mm / s, forming a liquid film with an initial thickness of approximately 0.3 mm. Immediately after spraying, a closed-loop temperature control system was activated. This system integrates 25 platinum resistance sensors, each installed near a heating element in an array-type local heating unit, with a sensor accuracy of ±0.05℃. The proportional gain of the PID control algorithm was set to 0.65, the integral time constant to 15 s, and the derivative time constant to 3.5 s. The system collected temperature data from 25 measurement points in real time at a sampling frequency of 10 Hz, and simultaneously recorded the spatial coordinates of each measurement point. In the initial stage of liquid film formation, the temperature of the aluminum substrate was controlled at 45℃, and the actual measured temperature fluctuation range was 44.85℃ to 45.15℃, meeting the control accuracy requirement of ±0.2℃.

[0093] The technical team used a laser triangulation sensor to measure the initial thickness of the liquid film. The sensor has a resolution of 1 μm and the measurement range covers six regions on the aluminum plate surface, each region having an area of ​​50 mm. The measurement results are shown in Table 2.

[0094] Table 2 Initial thickness measurement data of liquid film

[0095]

[0096] The technical team input the temperature distribution data and initial thickness data of the liquid film surface into the flow field prediction model. The input layer of the flow field prediction model contains 256 input nodes. The first 250 input nodes receive the temperature and spatial coordinate values ​​of 25 measurement points, and the last 6 input nodes receive the average initial thickness of the liquid film from 6 regions in Table 1. The phase field model interface tracking algorithm module introduces phase field variables to characterize the interface position between the liquid film and air, and the diffusion coefficient of the diffusion term in the spatiotemporal evolution equation is set to... The interface thickness was controlled at 0.2 mm. The velocity field of the convection term was calculated from the temperature distribution data using the surface tension gradient, with a surface tension gradient calculation coefficient of -0.1 mN / m·K. The surface tension coefficient of the interface energy term was set to 0.035 N / m based on the functional liquid properties. The phase field model was numerically solved at 32 grid points with a grid spacing of 6.25 mm and a time step of 0.05 s. Figure 2As shown, the flow field prediction model outputs the Marangoni convection velocity field distribution and liquid film morphology evolution prediction results. The velocity of Marangoni convection in the central region of the liquid film reaches 0.8 mm / s, while the velocity in the edge region is 0.3 mm / s.

[0097] The technical team adjusted the power distribution mode of the array-type local heating unit based on the output of the flow field prediction model. Initially, the power of each heating element was set to 8W. When the model predicted that the flow velocity in the upper left corner of the liquid film was too high, reaching 1.2 mm / s, the team reduced the power of the three resistance heating elements in that area by 20% to 6.4W, decreasing the temperature gradient in that area from 5.8 K / mm to 4.2 K / mm. After 180 seconds of liquid film evaporation, the texture direction deflection angle output by the flow field prediction model reached 18 degrees in the right side of the liquid film, exceeding the set threshold of 15 degrees. The team used principal component analysis to extract the main flow direction of the Marangoni convection velocity field distribution, confirming the presence of significant disordered vortex structures in that area. The team immediately replenished the area with a high-concentration polymer solution using a micro-injection device. The replenishment solution contained 6.5% polyvinylpyrrolidone by mass, with a replenishment volume of 0.8 mL, increasing the micro-polymer concentration in that area from 0.6% to 1.2%. At the same time, the power of the four resistance heating elements in this area was reduced by 40% to 4.8W, which reduced the temperature gradient from 5.2K / mm to 2.8K / mm, successfully achieving texture orientation correction.

[0098] The technical team maintained the ambient humidity at 68%, using a precision humidity control system to keep humidity fluctuations within ±2%. (In the mixed solvent) As a low-boiling-point component, it preferentially volatilizes in the initial stage of evaporation, causing a decrease in the surface temperature of the liquid film by approximately 3°C, creating a temperature difference between the edge and center of the liquid film. This temperature difference generates a surface tension gradient from the edge to the center, driving the formation of a reverse Marangoni flow. Figure 3 As shown, the velocity of the reverse Marangoni flow reaches 0.5 mm / s in the liquid film edge region, carrying solute from the edge to the center, effectively counteracting the outward flow driven by evaporation. After the liquid film was completely dried, the technical team measured the ratio of edge solute concentration to center solute concentration to be 1.15, which is significantly lower than the 3.8 of the traditional method, successfully suppressing the coffee ring effect.

[0099] The technical team conducted quality inspections on the aluminum plate surface after texture optimization. A laser confocal microscope was used to scan the entire aluminum plate surface, and the distribution of texture direction deflection angles was statistically analyzed. The inspection results are shown in Table 3.

[0100] Table 3. Statistical distribution of texture direction deflection angle

[0101]

[0102] As shown in Table 2, the area with a texture direction deflection angle of less than 15 degrees accounted for 98.6%, a significant improvement compared to the 60% achieved by traditional methods. The technical team also measured the thickness difference between the edge and center regions after the liquid film dried. The measurement results showed that the average thickness at the edge was 15.2 μm, and the average thickness at the center was 13.8 μm, with a thickness difference of only 1.4 μm, while the thickness difference of traditional methods typically exceeds 8 μm.

[0103] The technological advancements of this invention compared to traditional methods are mainly reflected in the following aspects. Traditional methods rely on empirical temperature control parameter settings, which cannot predict the velocity field distribution and liquid film morphology evolution path of Marangoni convection in real time, resulting in unavoidable texture randomness. This invention integrates a phase-field model interface tracking algorithm with a flow field prediction model, transforming temperature distribution data into quantitative predictions of Marangoni convection velocity field distribution, enabling the early identification of disordered vortex structures. The array-type local heating unit dynamically adjusts the heating power and temperature gradient distribution of each region based on the flow field prediction results, realizing a shift from passive response to active control. The phase-field model interface tracking algorithm automatically handles the topological changes of liquid film rupture and fusion by solving the spatiotemporal evolution equation, eliminating the numerical instability of traditional interface tracking methods when dealing with complex interface morphologies, and enabling precise quantification of the liquid film boundary position. The reverse Marangoni flow generated by the selective evaporation mechanism of the mixed solvent establishes a material transport path from the edge to the center, fundamentally changing the evaporation-driven unidirectional flow mode in traditional methods and effectively offsetting the edge accumulation phenomenon caused by the coffee ring effect. The dynamic adjustment of the concentration of the suppressor polymer and the precise control of the temperature gradient effectively suppress the small-scale vortex structure of Marangoni convection, enabling the main flow direction of the texture to be guided to the preset direction, thus solving the technical problem of the difficulty in correcting the deflection of the texture direction in traditional methods.

[0104] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for optimizing aluminum plate texture based on surface tension adjustment, characterized in that, The process includes plasma treatment of the aluminum plate surface, preparation of functional liquid and spraying onto the aluminum plate surface to form a liquid film, activation of a closed-loop temperature control system, construction of an array-type local heating unit, input of the temperature distribution data of the liquid film surface into a flow field prediction model, output of the Marangoni convection velocity field distribution and liquid film morphology evolution prediction results from the flow field prediction model, adjustment of the power distribution mode of the array-type local heating unit based on the Marangoni convection velocity field distribution and liquid film morphology evolution prediction results, increasing the concentration of eddy current suppressing polymer and reducing the temperature gradient of the corresponding region when the texture direction deflection angle output by the flow field prediction model exceeds a threshold to achieve texture orientation correction, controlling the ambient humidity and establishing selective evaporation conditions for the mixed solvent to utilize the reverse Marangoni flow generated by the preferential evaporation of low-boiling-point components to counteract the coffee ring effect, and achieving accurate tracking of the liquid film boundary position and real-time prediction and control of Marangoni convection through the integration of the phase field model interface tracking algorithm into the flow field prediction model.

2. The aluminum plate texture optimization method based on surface tension adjustment according to claim 1, characterized in that, The plasma treatment employs Qi and The mixture of gases has a gas flow ratio of 4:1, a processing power of 200W to 300W, and a processing time of 180s to 240s.

3. The aluminum plate texture optimization method based on surface tension adjustment according to claim 2, characterized in that, The plasma treatment reduces the surface roughness of the aluminum plate to below 50 nm and constructs an array-type local heating unit on the treated aluminum plate surface. The array-type local heating unit consists of multiple independently controlled resistance heating elements with a spacing of 5 mm between the heating elements.

4. The aluminum plate texture optimization method based on surface tension adjustment according to claim 3, characterized in that, The functional liquid consists of a base solvent, a surfactant mixture, and a trace polymer, wherein the surfactant mixture has an HLB value of 8 to 12 and the trace polymer has a mass fraction of 0.3% to 0.8%.

5. The aluminum plate texture optimization method based on surface tension adjustment according to claim 4, characterized in that, The trace polymer is a polyvinylpyrrolidone or polyacrylamide-based polymer material with a molecular weight of [missing information]. to .

6. The method for optimizing aluminum plate texture based on surface tension adjustment according to claim 5, characterized in that, The closed-loop temperature control system integrates multi-point platinum resistance sensors and uses a PID control algorithm to control the substrate temperature fluctuation within ±0.2℃.

7. The aluminum plate texture optimization method based on surface tension adjustment according to claim 6, characterized in that, In the PID control algorithm of the closed-loop temperature control system, the proportional coefficient is set to 0.5 to 0.8, the integral time constant is set to 10s to 20s, and the derivative time constant is set to 2s to 5s.

8. The method for optimizing aluminum plate texture based on surface tension adjustment according to claim 7, characterized in that, The temperature distribution data of the liquid film surface was obtained by a multi-point platinum resistance sensor at 25 measurement points on the aluminum plate surface, with a sampling frequency of 10Hz. The temperature distribution data includes the temperature value and spatial coordinate value of each measurement point.

9. The method for optimizing aluminum plate texture based on surface tension adjustment according to claim 8, characterized in that, Before inputting the temperature distribution data of the liquid film surface into the flow field prediction model, the initial thickness data of the liquid film is measured using a laser triangulation sensor. The measurement range covers six regions on the aluminum plate surface, each region having an area of ​​50 mm. .

10. The method for optimizing aluminum plate texture based on surface tension adjustment according to claim 9, characterized in that, The input layer of the flow field prediction model receives temperature distribution data and initial thickness data of the liquid film surface. The input layer contains 256 input nodes. The first 250 input nodes receive temperature values ​​and spatial coordinate values ​​of 25 measurement points, and the last 6 input nodes receive the average value of the initial thickness data of the liquid film in 6 regions.

Images