Structured surface based on bionics and optimization method thereof

Abstract

The invention discloses a structured surface based on bionics and an optimization method thereof, and belongs to the technical field of thermal management, and according to the structured surface based on bionics and the optimization method thereof, through the multistage microstructure design of bionic ruellia grass blades, hexagonal columnar infinitesimal elements and conical mastoids are creatively fused, and collaborative optimization of wettability and boiling heat transfer performance is achieved. The I-level infinitesimal structure forms uniform fluid infusion microgrooves through a hexagonal columnar array, the surface roughness is remarkably improved, on the basis, the II-level infinitesimal structure adds a conical mastoid of a plane structure to the top of a columnar infinitesimal, the pinning effect is weakened through 5-micron round corner treatment, the apparent contact angle is further reduced, and the spreading diameter is further increased. According to the multi-stage structural design, vertical transportation of liquid is promoted through the capillary action and the Laplace pressure difference, the motion resistance of liquid drops is reduced through the smooth edge, the spreading limitation of a traditional structured surface is broken through, and a better wetting condition is provided for boiling heat transfer.

Description

Technical Field

[0001] The present invention belongs to the technical field of thermal management, and in particular relates to a bionics-based structured surface and an optimization method thereof. Background Art

[0002] With the rapid development of electronic devices towards miniaturization and integration, the heat flux density per unit area has increased significantly, which makes it easy for local hot spots to form inside the device. If the heat cannot be discharged in time, it will seriously affect the performance and reliability of the equipment. Traditional heat dissipation technologies (such as air cooling and liquid cooling) have been unable to meet the needs of high heat flux scenarios, and methods based on phase change heat transfer (such as boiling heat dissipation) have gradually become a research hotspot due to their high heat transfer coefficient and fast thermal response. In the existing technology, structured surfaces enhance wettability through micro-nano structures to improve boiling heat transfer performance, but the following problems still exist:

[0003] Difficulty in optimizing microstructure parameters: Parameters such as element shape, spacing, and height have complex effects on wettability and heat transfer performance, requiring extensive experimentation or simulation to determine, resulting in high costs and low efficiency.

[0004] Conflict between wettability and heat transfer performance: Although some structures can reduce the apparent contact angle (improve wettability), they may inhibit bubble detachment, resulting in a decrease in heat transfer coefficient;

[0005] Insufficient processing accuracy and stability: It is difficult to process special-shaped or multi-level microstructures, and they are prone to wear after long-term use, affecting performance stability.

[0006] To address the above problems, the present invention proposes a structured surface design that imitates the leaves of sedge, combining machine learning with a multi-objective genetic algorithm to optimize microstructure parameters and achieve a synergistic improvement in wettability and heat transfer performance. Summary of the Invention

[0007] The purpose of the present invention is to provide a bionics-based structured surface and its optimization method. By designing the microelement morphology through bionic principles and combining machine learning with genetic algorithms to optimize structural parameters, the surface wettability, boiling heat transfer coefficient and bubble detachment efficiency can be significantly improved, thereby solving the problem of small-scale heat dissipation of electronic devices.

[0008] The purpose of the present invention can be achieved through the following technical solutions:

[0009] A biomimetic structured surface, comprising:

[0010] Level I micro-element structure and level II micro-element structure;

[0011] The I-level micro-element structure is composed of hexagonal columnar micro-elements arranged in a mosaic form, with a micro-element spacing of 15-30 μm and a height of 10-25 μm.

[0012] The II-level micro-element structure is a conical papilla with a planar configuration in which the top of the micro-element is fused, and the height of the conical papilla is 30.5-50.5 μm.

[0013] Preferably, the microelement spacing is 24.5 μm, the height is 22.5 μm, the apparent contact angle of the structured surface is 48.3°, and the spreading diameter of the structured surface is 179.8 μm.

[0014] Preferably, the height of the conical papilla is 40.5 μm, the edge of the conical papilla is rounded to 5 μm, the apparent contact angle of the structured surface is 25.4°, and the spreading diameter of the structured surface is 198.0 μm.

[0015] As a further solution of the present invention, a method for optimizing a structured surface based on bionics is provided, the method comprising the following steps:

[0016] Step 1: Construction of Level I structure: Analyze the surface microstructure of the Aleurites herb leaf and use hexagonal columnar microelements to construct the Level I microelement structure of the structured surface;

[0017] Step 2: Data acquisition: Use Latin hypercube sampling to obtain discrete combinations of element spacing and element heights, and obtain the corresponding apparent contact angle and spreading diameter data through numerical simulation;

[0018] Model training: Based on the BP neural network, a prediction model of micro-element spacing, micro-element height and wetting performance is established, and model training is performed to verify the effectiveness of the model;

[0019] Step 3: Algorithm optimization: Introduce CAWOA to improve the whale optimization algorithm to optimize the BP neural network;

[0020] Multi-objective optimization: The NSGA-II multi-objective genetic algorithm is used to perform multi-objective optimization with the dual objectives of minimizing the apparent contact angle and maximizing the spreading diameter to generate the Pareto frontier solution set;

[0021] Step 4: Weight calculation: Use the entropy weight method to determine the weight of the apparent contact angle and the weight of the spreading diameter;

[0022] Closeness calculation: Combined with the TOPSIS method, the relative closeness of the Pareto frontier solution to the positive and negative ideal solutions is calculated to sort the solutions and make decisions;

[0023] Step 5: Structural fusion: Based on the level I micro-element, the conical mastoids are fused to construct the level II structured surface by optimizing the mastoid height, shape and edge processing;

[0024] Heat transfer analysis: Construct a boiling heat transfer model to analyze the effects of different working fluids, heat flux densities, and hotspot sizes on heat transfer performance.

[0025] Furthermore, the microelement spacing in step 2 is the width of the fluid infusion microgroove, and fluid infusion microgrooves are formed between the microelements in the level I microelement structure.

[0026] Furthermore, the optimization process in step 3 includes chaos map initialization and adaptive weight adjustment; the chaos map initialization is to use quadratic polynomial chaos map to enhance population diversity:

[0027] X*(t)=ρ(1-X 2 (t));

[0028] Where ρ is the control parameter; X(t) is the vector, i.e. the initial population position; X*(t) is the updated population position;

[0029] The adaptive weight adjustment introduces an exponential decay strategy to balance global and local search capabilities:

[0030]

[0031] In the formula, w represents the weight, w max is the maximum weight, w min is the minimum weight, t is the current number of iterations, T is the maximum number of iterations, and m is the speed change factor.

[0032] Furthermore, in the NSGA-II multi-objective genetic algorithm model described in step 3:

[0033] Objective function: Minimize apparent contact angle: f1 = θ; maximize spreading diameter: f2 = D;

[0034] Constraints: Microelement spacing: 15μm≤X1≤30μm; Microelement height: 10μm≤X2≤25μm.

[0035] Furthermore, the prediction error MAPE of the optimized apparent contact angle in step 3 is 0.60%, and the prediction error MAPE of the spreading diameter is 1.85%.

[0036] Furthermore, the closeness calculation in step 4 is:

[0037] Where: They represent the negative ideal distance and positive ideal distance of the i-th solution individual respectively; the relative closeness c of the i-th solution individual to the ideal solution i .

[0038] Furthermore, the apparent contact angle weight in step 4 is 0.5311; and the spreading diameter weight is 0.4689.

[0039] Beneficial effects of the present invention:

[0040] This invention provides a biomimetic-based structured surface and its optimization method. By designing microelement morphology based on biomimetic principles and combining machine learning with genetic algorithms to optimize structural parameters, it significantly improves surface wettability, boiling heat transfer coefficient, and bubble detachment efficiency, solving the problem of microscale heat dissipation in electronic devices. Specific effects are as follows:

[0041] Multi-level bionic structure synergistically improves wetting and heat transfer performance: The present invention uses a multi-level microstructure design that mimics the leaves of reed grass, innovatively integrating hexagonal columnar microelements and conical papillae, to achieve synergistic optimization of wettability and boiling heat transfer performance. The I-level microelement structure forms uniform rehydration microgrooves through a hexagonal columnar array, significantly increasing the surface roughness, reducing the apparent contact angle of the droplet from 69° on a smooth surface to 48.3°, and increasing the spreading diameter to 179.8μm, thereby enhancing the droplet spreading and liquid replenishment capabilities. On this basis, the II-level microelement structure adds a planar-configured conical papillae on the top of the columnar microelement, and weakens the pinning effect through a 5μm fillet treatment, further reducing the apparent contact angle to 25.4° and increasing the spreading diameter to 198.0μm. This multi-level structural design not only promotes vertical liquid transport through capillary action and Laplace pressure difference, but also reduces the resistance to droplet movement through smooth edges, breaking through the spreading limitations of traditional structured surfaces and providing better wetting conditions for boiling heat transfer.

[0042] Intelligent optimization method efficiently searches for optimal microelement parameters: In response to the problems that traditional structured surface parameter optimization relies on a large number of experiments, is costly and inefficient, the present invention proposes an intelligent optimization method based on machine learning and multi-objective genetic algorithm. Discrete microelement parameter combinations are obtained through Latin hypercube sampling, and a prediction model of microelement spacing, height and wetting performance is established in combination with the BP neural network. The neural network is further optimized by introducing the improved whale optimization algorithm (CAWOA), which further reduces the prediction error and significantly improves the model accuracy. On this basis, the NSGA-II genetic algorithm is used for multi-objective optimization to generate the Pareto frontier solution set, and the optimal parameter combination is determined by the entropy weight-TOPSIS method. This method significantly improves the efficiency of parameter optimization, effectively reduces the cost of experimental trial and error, and provides efficient and accurate guidance for the design of structured surfaces.

[0043] Pinning effect regulation promotes bubble detachment and heat transfer enhancement: Traditional structured surfaces often inhibit bubble detachment due to the pinning effect at the edges of the microelement, resulting in a decrease in the heat transfer coefficient. This invention significantly weakens the pinning effect by rounding the edges of the conical nipples by 5μm, reducing the spreading resistance of droplets on the surface. This design not only improves wettability but also effectively enhances boiling stability by promoting bubble generation, fusion, and rapid detachment, thereby solving the problem of local hot spots under high heat flux density.

[0044] Improved Processing Feasibility and Stability: The structured surface design of this invention fully considers practical processing requirements. The regular arrangement of hexagonal columnar microelements and the design of planar papillae reduce the difficulty of micro-nanofabrication. The rounded edge treatment (5μm fillet) not only ensures structural strength, but also avoids the wear and tear of sharp edges during long-term use, improving the stability and durability of the surface.

[0045] In summary, the present invention achieves a breakthrough improvement in the wettability and heat transfer performance of structured surfaces through bionic principles, intelligent optimization methods and pinning effect regulation, solves the core problems in the existing technology, and has significant technological innovation and engineering application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The present invention will be further described below with reference to the accompanying drawings.

[0047] Figure 1 The following are scanning electron microscope images of the surface micromorphology of Aleurites salsa (a) and schematic diagram of the micromorphology and microstructural characteristics of the surface micromorphology of Aleurites salsa (b);

[0048] Figure 2 It is a schematic diagram of the level I microelement structure model;

[0049] Figure 3 It is a schematic diagram of the structural fusion process model;

[0050] Figure 4 This is a diagram of the research and development ideas for the optimization method of structured surfaces;

[0051] Figure 5 It is a flow chart for training a BP neural network prediction model for surface wetting characteristics of the present invention;

[0052] Figure 6 The data graphs are the comparison between the predicted value of the apparent contact angle Y1 and the actual value of the simulation under the CAWOA-BP optimization algorithm (left figure) and the error MAPE (right figure);

[0053] Figure 7 The data graphs are the comparison between the predicted value of spreading diameter Y2 and the actual value of simulation under CAWOA-BP optimization algorithm (left figure) and the error MAPE (right figure);

[0054] Figure 8 Schematic diagram of the core steps of the NSGA-II multi-objective optimization algorithm (a) Pareto front solution set distribution and (b) fast non-dominated sorting and congestion calculation under the elitist strategy;

[0055] Figure 9 Schematic diagram of the optimization steps of the wetting performance of a level I biomimetic surface based on machine learning and NSGA-II genetic algorithm;

[0056] Figure 10This is a schematic diagram of the apparent contact angle and spreading diameter test values of a droplet on a structured surface;

[0057] Figure 11 Schematic diagram of the time-varying behavior of structured surface boiling bubbles.

[0058] In the figure: 1. Columnar microelement of blade; 2. Conical papilla of blade; 3. Fluid-replenishing microgroove of blade; 4. Level I microelement structure; 41. Microelement; 42. Fluid-replenishing microgroove; 5. Conical papilla. DETAILED DESCRIPTION

[0059] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0060] Example 1

[0061] A structured surface based on biomimetics:

[0062] 1. Bionic prototype and design ideas:

[0063] Inspiration: Figure 1 From the existing literature "Koch K, Blecher IC, K Nig G, Kehraus S, Barthlott W. The superhydrophilic and superoleophilic leaf surface of Ruelliadevosiana (Acanthaceae): a biological model for spreading of water and oil on surfaces [J]. Functional Plant Biology. 2009; 36 (4): 339.", specifically Figure 1 As shown, the polygonal leaf columnar microelements (1), leaf fluid replenishment microgrooves (3) and leaf conical papillae (2) structures on the surface of the Aleurites herb leaf achieve superhydrophilicity through surface tension, capillary force and Laplace pressure difference.

[0064] Design goal: To construct a structured surface with high wettability to enhance droplet spreading and liquid replenishment capabilities, thereby improving boiling heat transfer efficiency.

[0065] 2. Level I structured (4) surface design:

[0066] Parameters of level I micro-element structure (4):

[0067] Composition: Figure 2 As shown, the level I microelement structure (4) is composed of hexagonal columnar microelements (41) arranged in a mosaic form, and fluid replenishment microgrooves (42) are formed between the microelements (41). The microelement spacing (n) is 15-30 μm, and the height (h) is 10-25 μm.

[0068] Spacing and height: Through orthogonal experimental optimization, the optimal parameters were determined to be a spacing of 24.5μm and a height of 22.5μm. At this time, the apparent contact angle dropped to 48.3° and the spreading diameter reached 179.8μm.

[0069] Mechanism of improving wetting performance: The I-level microstructure (4) increases the surface roughness and promotes the spreading of droplets; the liquid replenishment microgrooves (42) produce capillary action and enhance the liquid replenishment capacity.

[0070] 3. Level II micro-element structure surface design:

[0071] Structural fusion: e.g. Figure 3 As shown, a conical papilla (5) is added to the top of the level I microelement to form a multi-level structure, and the height of the conical papilla (5) is 30.5-50.5 μm.

[0072] Key parameter optimization:

[0073] When the height of the conical papilla (5) is 40.5 μm, the apparent contact angle decreases to 33.8°.

[0074] Morphology and edge treatment: The planar configuration of the conical papilla (5) combined with the 5 μm rounded corner treatment weakens the pinning effect, further reducing the apparent contact angle to 25.4° and increasing the spreading diameter to 198.0 μm.

[0075] Wetting enhancement mechanism: The conical papilla (5) promotes vertical transport of liquid through Laplace pressure difference, and the smooth edge reduces pinning resistance and improves the spreading ability of droplets.

[0076] 4. Numerical simulation and experimental verification:

[0077] Model construction: The VOF model is used to simulate droplet spreading, and the continuous surface tension model (CSF) and Lee phase change model are combined to analyze boiling heat transfer.

[0078] Verification method:

[0079] Mesh independence verification: Ensure calculation accuracy through polyhedral mesh division.

[0080] Experimental comparison: Compared with smooth surface, the wetting performance of structured surface is significantly improved, and the boiling heat transfer coefficient is as high as 21.15kW / (m 2 ·K).

[0081] 5. Machine Learning and Optimization Algorithms:

[0082] Prediction model: Based on the BP neural network, a nonlinear relationship between microelement parameters and wetting performance was established, and the prediction errors (MAPE) were 4.15% (contact angle) and 3.73% (spreading diameter).

[0083] Optimization algorithm: The improved whale optimization algorithm (CAWOA) was introduced to optimize the BP neural network, and the prediction error was reduced to 0.60% and 1.85%; the NSGA-II genetic algorithm was combined for multi-objective optimization to obtain the Pareto frontier solution set, and the optimal parameters were determined by the entropy weight-TOPSIS method.

[0084] 6. Analysis of boiling heat transfer performance:

[0085] Influence of working fluid: High surface tension fluids (such as water) have better heat transfer performance due to their advantages in thermal conductivity and specific heat; low surface tension fluids (such as HFE-7100) have better wettability but lower heat transfer coefficient.

[0086] Heat flux density and hot spot size: The heat transfer coefficient increases with the increase of heat flux density, and reaches a peak of 21.15kW / (m 2 ·K), the excessive size (218μm) leads to deterioration of heat transfer.

[0087] Bubble behavior: The II-level structure promotes bubble generation, fusion and detachment, improving boiling stability.

[0088] Example 2

[0089] A biomimetic structured surface and its optimization method, see Figure 4 As shown, the following steps are included:

[0090] 1. Bionic prototype analysis and level I structure design:

[0091] Mechanism research: An in-depth analysis of the surface microstructure of Aleurites sinensis leaves clarified that its wetting properties are formed by the combined action of polygonal columnar microelements, fluid-replenishing microgrooves, and conical papillae, involving key mechanisms such as surface tension, capillary force, and Laplace pressure difference.

[0092] Structural Design: A Level I structured surface was constructed using pentagonal and hexagonal columnar microelements. Experiments and simulations compared the effects of different microelement types, spacing, and heights on the apparent contact angle and spreading diameter. The results showed that hexagonal microelements were chosen due to their superior microgroove uniformity. When the spacing was 15 μm and the height was 10 μm, the apparent contact angle decreased to 59.2°, and the spreading diameter increased by 12.8%.

[0093] 2. Machine learning prediction model construction:

[0094] Data acquisition: Latin hypercube sampling is used to obtain discrete combinations of element spacing and element height, and the corresponding apparent contact angle and spreading diameter data are obtained through numerical simulation.

[0095] Model training: Based on the BP neural network, a prediction model of micro-element spacing, micro-element height and wetting performance is established, and the model training is performed as follows Figure 5 As shown, the BP neural network is a multi-layer feedforward neural network trained using the back-propagation error algorithm. X1 and X2 are the input layer, and Y1 and Y2 are the output layer. The input layer is the starting point of the BP neural network, and the output layer is the end point. It is responsible for predicting the apparent contact angle Y1 and spreading diameter Y2 of the droplet on the biomimetic surface. The hidden layer is the core of the BP neural network, extracting features from the training data and performing nonlinear transformations to extract the prediction model. Back-propagation is the core of the BP algorithm, updating the model weights and thresholds in the direction opposite to the gradient of the loss function. Its goal is to minimize the loss function, that is, minimize the prediction error. Through continuous iterative optimization of the trained model, a neural network prediction model is obtained that can represent the complex relationship between X1, X2 and Y1, Y2. After training, the prediction error of the apparent contact angle is 4.15% and the prediction error of the spreading diameter is 3.73%, verifying the effectiveness of the model.

[0096] 3. Optimization algorithm improvement and multi-objective optimization:

[0097] Algorithm optimization: However, the prediction errors of the apparent contact angle Y1 and the spreading diameter Y2 under the BP neural network meet the requirements but still have room for improvement. Therefore, the present invention introduces the improved whale optimization algorithm (CAWOA) to optimize the BP neural network and gradually improves the network training effect by adjusting the algorithm parameters. Figure 6-Figure 7 As shown in the figure, the prediction error MAPE of the apparent contact angle after optimization is 0.60%, and the prediction error MAPE of the spreading diameter is 1.85%.

[0098] Multi-objective optimization: Finally, the NSGA-II (Non-dominated Sorting Genetic Algorithm II) genetic algorithm is used to perform multi-objective optimization with the dual objectives of minimizing the apparent contact angle and maximizing the spreading diameter to generate a Pareto frontier solution set and provide candidate solutions for subsequent decision-making. The details are as follows Figure 8As shown, the present invention obtains two combinations of spacing and height, that is, two solutions a and b, through the NSGA-H genetic algorithm. If a is not inferior to b in both optimization objectives Y1 and Y2, and is better than b in at least one objective, then a is said to dominate b. According to this rule, several groups of Pareto excellent solutions are obtained, and the set of these solutions is called the Pareto frontier. The solutions in this solution set may be better than other solutions in one objective, but cannot achieve all the optimal solutions in both objectives. In the dual-objective optimization of surface wetting, the apparent contact angle is f1, and the smaller one is optimized; the spreading diameter is f2, and the larger one is optimized. Therefore, in the subsequent multi-objective optimization, it is expected to obtain the same Figure 8 (a) Distribution of Pareto front solutions with similar trends in II. NSGA-II uses elitist strategy, fast non-dominated sorting and congestion calculation to achieve efficient multi-objective optimization, such as Figure 8 (b). The combination of the distance between the microelement X1 and the height X2 is the individual solution. The elite strategy merges the parent and child solution populations Pt and Qt, and selects the next parent population based on the quality of the individual to retain the excellent solutions in the evolution process, ensuring that the algorithm converges to the global optimal Pareto front. Fast non-dominated sorting divides the solution population into different non-dominated levels. Figure 8 In (b), F(1), H(2), and I(3) provide a basis for subsequent selection operations. The crowding calculation evaluates the distribution of individual solutions within each non-dominated level and calculates its crowding degree based on the two objective functions of apparent contact angle Y1 and spreading diameter Y2, taking into account the uniformity of the individual solution distribution and maintaining population diversity. Through these three core steps, the NSGA-II genetic algorithm efficiently and accurately finds the microelement spacing and height combination that optimizes the overall wetting performance of the biomimetic surface.

[0099] 4. Optimal solution decision based on entropy weight-TOPSIS method:

[0100] Weight calculation: The entropy weight method was used to determine the weight of each evaluation index, where the weight of the apparent contact angle was 0.5311 and the weight of the spreading diameter was 0.4689.

[0101] Closeness calculation: The TOPSIS method is used to calculate the relative closeness of the Pareto frontier solution to the positive and negative ideal solutions, and the solutions are sorted and decided. The optimal element spacing is finally determined to be 24.5 μm and the height is 22.5 μm.

[0102] 5. Level II structured surface design and performance verification:

[0103] Structural fusion: Conical papillae are fused on the basis of level I microelements to construct a level II structured surface. By optimizing the papilla height (40.5μm), morphology (planar configuration) and edge treatment (5μm smoothing radius), the wetting performance is further improved, the apparent contact angle is reduced to 25.4°, and the spreading diameter is increased to 198.0μm.

[0104] Heat transfer analysis: A boiling heat transfer model was constructed to analyze the effects of different working fluids (such as water and HFE-7100), heat flux density, and hot spot size on the heat transfer performance. The results showed that when the droplet working fluid was liquid water and the heat flux density was 700W / cm 2 When the hot spot size is 169 μm, the heat transfer coefficient of the surface of the II-level micro-element structure can reach 21.15kW / (m 2 ·K), and can effectively promote the detachment of bubbles and improve the heat dissipation capacity of local hot spots.

[0105] Example 3

[0106] The specific process of the optimization method (see Figure 9 shown):

[0107] 1. Overall framework of optimization method:

[0108] The optimization method of this invention is based on the principles of biomimetics, combined with machine learning and multi-objective genetic algorithms, to perform multi-level optimization of the microelement parameters of the structured surface, aiming to achieve a synergistic improvement in wettability and boiling heat transfer performance. The optimization process includes the following core steps:

[0109] (1) Construction of level I microelement structure and parameter optimization.

[0110] (2) Data collection and machine learning model training.

[0111] (3) Improved whale optimization algorithm (CAWOA) to optimize neural networks.

[0112] (4) NSGA-II multi-objective genetic algorithm optimization.

[0113] (5) Entropy weight-TOPSIS method is used to determine the optimal parameter combination.

[0114] (6) Level II micro-element structure fusion and heat transfer analysis.

[0115] 2. Construction of Level I micro-element structure and parameter optimization:

[0116] 2.1 Analysis of bionic prototype: Design goal: to construct a multi-level microstructure, enhance surface roughness, reduce apparent contact angle, and improve liquid spreading ability.

[0117] 2.2I-level structural parameter design:

[0118] (1) Microelement morphology: Hexagonal columnar microelements are arranged in mosaics to form uniform fluid replenishment microgrooves.

[0119] (2) Key parameters: cell spacing (n): 15-30 μm; cell height (h): 10-25 μm.

[0120] (3) Optimization objectives: Minimize the apparent contact angle (θ); maximize the spreading diameter (D).

[0121] 2.3 Parameter orthogonal experiment: The effects of different combinations of microelement spacing and height on wettability were analyzed through numerical simulation (VOF model). The test results are as follows:

[0122] Microelement pitch (μm) Microelement height (μm) Apparent contact angle (°) Spread diameter (μm) 15 10 59.2 169.3 24.5 22.5 48.3 179.8 30 25 63.3 172.0

[0123] 3. Data collection and machine learning model training:

[0124] 3.1 Latin Hypercube Sampling:

[0125] (1) Method: Latin hypercube sampling (LHS) is used to generate discrete parameter combinations covering the design space of element spacing (15-30μm) and height (10-25μm).

[0126] (2) Sample size: 31 sets of parameter combinations, including training sets (28 sets) and validation sets (3 sets).

[0127] 3.2BP neural network modeling:

[0128] (1) Model structure:

[0129] Input layer: microelement spacing (X1), microelement height (X2).

[0130] Output layer: apparent contact angle (Y1), spreading diameter (Y2).

[0131] Hidden layer: 15 neurons, activation function is Sigmoid.

[0132] Training algorithm: Levenberg-Marquardt algorithm, loss function is mean square error (MSE).

[0133] (2) Performance evaluation:

[0134] Training error: θ prediction MAPE = 4.15%, D prediction MAPE = 3.73%.

[0135] Validation error: θ prediction MAPE = 4.74%, D prediction MAPE = 3.73%.

[0136] 4. Improved Whale Optimization Algorithm (CAWOA) to optimize neural networks:

[0137] 4.1 Algorithm improvement strategy:

[0138] (1) Chaotic map initialization: A quadratic polynomial chaotic map (the following formula) is used to improve population diversity.

[0139] X*(t)=ρ(1-X2 (t));

[0140] Where ρ is the control parameter; X(t) is the vector, i.e., the initial population position; X*(t) is the updated population position.

[0141] (2) Adaptive weight adjustment: An exponential decay strategy (the following formula) is introduced to balance global and local search capabilities.

[0142]

[0143] In the formula, w represents the weight, w max is the maximum weight, w min is the minimum weight, t is the current iteration number, T is the maximum iteration number, and m is the speed change factor. Based on the current and maximum iteration numbers, the exponential term gradually shifts from near 1 (at the beginning of an iteration) to 0 (at the end of an iteration). This is the key to achieving adaptive adjustment. This process continuously adjusts the weights to achieve an adaptive strategy that changes with the number of iterations, balancing and improving global and local search capabilities.

[0144] 4.2 Optimization effect comparison:

[0145] (1) Improved prediction accuracy:

[0146]

[0147]

[0148] 5. NSGA-II multi-standard genetic algorithm optimization:

[0149] 5.1 Multi-objective optimization model:

[0150] (1) Objective function: Minimize the apparent contact angle: f1 = θ Maximize the spreading diameter: f2 = D;

[0151] (2) Constraints: microelement spacing: 15μm≤X1≤30μm; microelement height: 10μm≤X2≤25μm.

[0152] 5.2 Pareto frontier solution set: Generate the Pareto optimal solution set through NSGA-II and select 5 groups of excellent solutions:

[0153] Solution number Microelement pitch (μm) Microelement height (μm) θ(°) D(μm) 1 24.5 22.5 47.14 184.9 2 24.5 22.5 47.13 184.8

[0154] 6. Entropy weight-TOPSIS method to determine the optimal parameters:

[0155] 6.1 Weight calculation:

[0156] (1) Entropy weight method was used to determine the index weights: apparent contact angle weight: 0.5311; spreading diameter weight: 0.4689;

[0157] 6.2 Closeness Sorting: Calculate the relative closeness between the Pareto solution and the ideal solution (following formula), and select the optimal solution:

[0158]

[0159] Where: They represent the negative ideal distance and positive ideal distance of the i-th solution individual respectively; the relative closeness c of the i-th solution individual to the ideal solution i .

[0160] Final parameters: microelement spacing 24.5μm, height 22.5μm, θ=48.3°, D=179.8μm.

[0161] 7. Level II micro-element structure fusion and heat transfer analysis:

[0162] 7.1 Conical mastoid design:

[0163] (1) Morphological optimization: planar configuration of conical papillae, 40.5 μm in height, and 5 μm rounded corners at the edges (to reduce the pinning effect).

[0164] (2) Improved wettability: The apparent contact angle decreased from 48.3° to 25.4°; the spreading diameter increased from 179.8 μm to 198.0 μm, see attached. Figure 10 shown.

[0165] 7.2 Boiling heat transfer performance:

[0166] (1) Influence of working fluid: Liquid water (high surface tension) at a heat flux density of 700W / cm 2 When the heat transfer coefficient reaches 21.15kW / (m 2 ·K).

[0167] (2) Heat flux density adaptability: The heat transfer performance is optimal when the local hot spot size is 169 μm, and the bubble detachment efficiency is increased by 30%.

[0168] 8. Technical effects of optimization methods:

[0169] (1) Wettability is significantly improved: After the optimization of the level I structure, the apparent contact angle is reduced to 48.3°, and the spreading diameter is increased by 12.8%; after the fusion of the level II structure, the contact angle is further reduced to 25.4°, and the spreading diameter is increased to 198.0 μm.

[0170] (2) Enhanced heat transfer performance: The boiling heat transfer coefficient is as high as 21.15kW / (m 2K), which is 3 times higher than that of smooth surface, and the bubble detachment speed is accelerated, which effectively inhibits the formation of local hot spots. Specifically, take the gas phase cloud of the structured surface when the local hot spot size D1 is 40μm as an example, see the attached Figure 11 As shown in the figure, the bubble motion behavior is observed and analyzed, and the influence of the wettability of the structured surface of the fused cone nipple on the boiling heat transfer and bubble behavior is analyzed. Figure 11 is the temporal variation of the gas phase volume fraction. First, during the initial boiling phase, bubbles form from the structured surface and the bottom of the refill microgrooves at t = 0.3 ms. The interface between the microelement's sidewalls and bottom serves as the primary nucleation site. As boiling continues, bubbles at the interface grow and merge into larger bubbles at t = 2.5 ms. At t = 4.9 ms, the bubbles envelop the central microelement. As mentioned in Section 5.3.2, the VOF model treats interfaces based on volume fraction, and boiling actually occurs in regions with a gas phase volume fraction of approximately 0.5. Therefore, during this time period, boiling occurs to varying degrees throughout the microelement and in the smooth outer regions. As boiling continues, from t = 6.2 to 6.8 ms, the surrounding gas phase gradually approaches and merges with the larger bubbles. The bubble volume increases, and the liquid phase in the refill microgrooves formed by the microelement is replenished. When the bubbles reach a certain volume, they begin to move upward along the central papilla microelement, gradually detaching from the surface and overflowing from the upper outlet. During the time period t = 11.0-11.3ms, the first bubble almost completely detached from the surface and overflowed from the outlet. At t = 11.0ms, the structured surface began to generate a second bubble at the nucleation site. As boiling continued, the bubble gradually grew and merged with the surrounding gas. At t = 11.5ms, a bubble was observed rising along the conical papilla surface and merging with another bubble, ultimately moving upward. Overall, compared to the Level I structured surface in Section 4.4, the structured surface after the fusion of the conical papilla accelerated the boiling generation of the gas phase and promoted the fusion and detachment of bubbles on the surface.

[0171] Based on the analysis previously presented in this section, the designed II-level microstructure improves the wetting properties of the structured surface, promoting liquid replenishment in the bottom refill microgrooves and vertically. Furthermore, the wedge-shaped angles formed by the microstructure provide nucleation sites required for boiling, helping to enhance the boiling heat transfer performance of the structured surface. Furthermore, after integrating the smoothed conical nipples, the Laplace pressure difference and capillary action formed between the microelements facilitate the growth, fusion, and upward detachment of bubbles, which indirectly helps maintain boiling and enhances the structured surface's heat dissipation capacity in localized high heat flux areas.

[0172] It should be noted that, in this document, terms such as "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or apparatus that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements that are inherent to such process, method, article or apparatus.

[0173] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A structured surface based on bionics, characterized in that: The structured surface comprises: Level I micro-element structure and level II micro-element structure; The I-level micro-element structure is composed of hexagonal columnar micro-elements arranged in a mosaic form, with a micro-element spacing of 15-30 μm and a height of 10-25 μm; The II-level micro-element structure is a conical papilla with a planar configuration in which the top of the micro-element is fused, and the height of the conical papilla is 30.5-50.5 μm.

2. The bionics-based structured surface according to claim 1, characterized in that: The microelement pitch is 24.5 μm, the height is 22.5 μm, the apparent contact angle of the structured surface is 48.3°, and the spreading diameter of the structured surface is 179.8 μm.

3. The bionics-based structured surface according to claim 1, characterized in that: The height of the conical papilla is 40.5 μm; the edge of the conical papilla is rounded to 5 μm, the apparent contact angle of the structured surface is 25.4°, and the spreading diameter of the structured surface is 198.0 μm.

4. A bionics-based structured surface optimization method according to any one of claims 1 to 3, characterized in that: The following steps are involved: Step 1: Construction of Level I structure: Analyze the surface microstructure of the Aleurites herb leaf and use hexagonal columnar microelements to construct the Level I microelement structure of the structured surface; Step 2: Data acquisition: Use Latin hypercube sampling to obtain discrete combinations of element spacing and element heights, and obtain the corresponding apparent contact angle and spreading diameter data through numerical simulation; Model training: Based on the BP neural network, a prediction model of micro-element spacing, micro-element height and wetting performance is established, and model training is performed to verify the effectiveness of the model; Step 3: Algorithm optimization: Introduce CAWOA to improve the whale optimization algorithm to optimize the BP neural network; Multi-objective optimization: The NSGA-II multi-objective genetic algorithm is used to perform multi-objective optimization with the dual objectives of minimizing the apparent contact angle and maximizing the spreading diameter to generate the Pareto frontier solution set; Step 4: Weight calculation: Use the entropy weight method to determine the weight of the apparent contact angle and the weight of the spreading diameter; Closeness calculation: Combined with the TOPSIS method, the relative closeness of the Pareto frontier solution to the positive and negative ideal solutions is calculated to sort the solutions and make decisions; Step 5: Structural fusion: Based on the level I micro-element, the conical mastoids are fused to construct the level II structured surface by optimizing the mastoid height, shape and edge processing; Heat transfer analysis: Construct a boiling heat transfer model to analyze the effects of different working fluids, heat flux densities, and hotspot sizes on heat transfer performance.

5. The bionics-based structured surface according to claim 4, characterized in that: The microelement spacing in step 2 is the width of the fluid infusion microgroove, and fluid infusion microgrooves are formed between the microelements in the level I microelement structure.

6. The bionics-based structured surface according to claim 4, characterized in that: The optimization process in step 3 includes chaos map initialization and adaptive weight adjustment; the chaos map initialization is to use quadratic polynomial chaos map to improve population diversity: X*(t)=ρ(1-X 2 (t)); Where ρ is the control parameter; X(t) is the vector, i.e. the initial population position; X*(t) is the updated population position; The adaptive weight adjustment introduces an exponential decay strategy to balance global and local search capabilities: In the formula, w represents the weight, w ma× is the maximum weight, W min is the minimum weight, t is the current number of iterations, T is the maximum number of iterations, and m is the speed change factor.

7. The bionics-based structured surface according to claim 4, characterized in that: In the NSGA-II multi-objective genetic algorithm model described in step 3: Objective function: Minimize apparent contact angle: f1 = θ; maximize spreading diameter: f2 = D; Constraints: Microelement spacing: 15μm≤X1≤30μm; Microelement height: 10μm≤X2≤25μm.

8. The bionics-based structured surface according to claim 4, characterized in that: After the optimization in step 3, the prediction error MAPE of the apparent contact angle is 0.60%, and the prediction error MAPE of the spreading diameter is 1.85%.

9. The bionics-based structured surface according to claim 4, characterized in that: The closeness calculation described in step 4: Where: They represent the negative ideal distance and positive ideal distance of the i-th solution individual respectively; the relative closeness c of the i-th solution individual to the ideal solution i .

10. The bionics-based structured surface according to claim 4, characterized in that: The apparent contact angle weight in step 4 is 0.5311; the spreading diameter weight is 0.4689.

Images