Special-shaped surface treatment method and device for GPU main body structural component

Abstract

The invention discloses a special-shaped surface processing method for a GPU main body structural member. The method comprises the steps of generating a three-dimensional special-shaped surface model with an electromagnetic shielding function and a thermal radiation enhancement characteristic based on environment simulation data of a target application scene; a metal additive manufacturing process is adopted, a metal matrix is formed layer by layer according to a three-dimensional special-shaped surface model, a multi-stage fractal microstructure is formed on the surface of the metal matrix through an electrochemical micromachining process, and a gradient refractive index photonic crystal layer and a periodic nano-particle layer are sequentially deposited on the surface of the multi-stage fractal microstructure. Forming a composite functional layer with a super-hydrophobic surface and high infrared emissivity; and based on the hydrophobic angle and thermal radiation performance data detected in real time, processing parameters are adjusted through a feedback control system, iterative optimization is performed on the morphology of the composite functional layer, a high-performance surface with electromagnetic shielding, efficient heat dissipation and super-hydrophobic characteristics is formed, the reliability of the GPU structural part is integrally improved, and the service life of the GPU structural part is integrally prolonged.

Description

Technical Field

[0001] This invention relates to the field of GPU technology, and in particular to a method and apparatus for irregular surface treatment of GPU main structural components. Background Technology

[0002] With the increasing popularity of high-performance mobile computing devices, consumer electronics products such as thin and light gaming laptops and graphics workstations are placing higher demands on the heat dissipation and reliability of GPUs. When running high-load tasks (such as 4K rendering and ray tracing), the GPU chip temperature of these devices can exceed 100°C within seconds. At the same time, due to frequent user movement, environmental dust and fiber debris are easily drawn into the device, leading to dust accumulation and increased thermal resistance on the surface of the heat dissipation module.

[0003] Current GPU surface treatment technologies primarily employ a combination of metal heat sinks and thermal grease, with some high-end models incorporating vapor chambers or graphene coatings to enhance heat transfer efficiency. However, existing technologies generally rely on passive cooling designs with fixed structures, whose surface morphologies are mostly regular geometric shapes (such as parallel fins and grid arrays), making it impossible to dynamically adjust the heat dissipation path based on real-time heat load. More importantly, traditional surface treatment processes lack anti-contamination designs, allowing dust to easily accumulate in the gaps of the heat dissipation structure, forming an insulating layer and causing a significant decrease in heat dissipation efficiency over time. Tests show that GPUs equipped with conventional cooling solutions experience a reduction in heat flux density after six months of continuous operation in a dusty environment, becoming a key bottleneck restricting the development of highly reliable mobile computing devices.

[0004] Therefore, it is necessary to improve the existing GPU technology to solve the problems of performance degradation and device failure caused by low surface heat dissipation efficiency and contaminant adhesion. Summary of the Invention

[0005] The purpose of this invention is to provide a method and apparatus for irregular surface treatment of GPU main structural components, thereby solving the above-mentioned technical problems.

[0006] To achieve this objective, the present invention adopts the following technical solution: A method for irregularly shaped surface treatment of a GPU main structural component includes: Based on environmental simulation data of the target application scenario, a three-dimensional irregular surface model with electromagnetic shielding function and thermal radiation enhancement feature is generated through multi-physics field coupling simulation. Using metal additive manufacturing process, a metal substrate is formed layer by layer based on the three-dimensional irregular surface model. The metal substrate includes a grid-like support structure and embedded heat pipe channels. A multi-level fractal microstructure is formed on the surface of the metal substrate by an electrochemical micromachining process. The multi-level fractal microstructure includes a primary array, a secondary protrusion structure, and a composite coating. A gradient refractive index photonic crystal layer and a periodic nanoparticle layer are sequentially deposited on the surface of the multi-level fractal microstructure to form a composite functional layer with a superhydrophobic surface and high infrared emissivity. Based on real-time detected hydrophobic angle and thermal radiation performance data, the morphology of the composite functional layer is iteratively optimized by adjusting the processing parameters through a feedback control system.

[0007] Optionally, the environmental simulation data based on the target application scenario is used to generate a three-dimensional irregular surface model with electromagnetic shielding function and thermal radiation enhancement characteristics through multiphysics coupling simulation, specifically including: Collect environmental parameters of the target application scenario, and acquire dynamic heat load data, environmental dust concentration distribution data and electromagnetic interference spectrum data of electronic devices during operation through sensor networks to construct a multi-dimensional environmental feature database; The environmental characteristic data is input into the electromagnetic-thermal-fluid coupling simulation system, and the synergistic relationship between electromagnetic shielding requirements and thermal radiation efficiency is calculated through a dynamic weight allocation algorithm to generate an initial three-dimensional irregular surface model with surface energy distribution. Perform reverse topology optimization: Based on the initial three-dimensional irregular surface model, the reverse Monte Carlo algorithm is used to iteratively generate candidate irregular surface structures, and the electromagnetic shielding coefficient and infrared emissivity performance of each candidate irregular surface structure are predicted by a convolutional neural network in order to optimize the three-dimensional irregular surface model. Finite element electromagnetic compatibility simulation and discrete element dust adsorption simulation were performed on the optimized initial three-dimensional irregular surface model. The curvature and porosity of the microstructure were adjusted by an adaptive correction algorithm until the preset performance threshold was met. The validated structural parameters are converted into a manufacturable 3D model format, and electromagnetic shielding layer thickness gradient data and thermal radiation enhancement area marking information are embedded simultaneously.

[0008] Optionally, the metal additive manufacturing process is used to form a metal substrate layer by layer based on the three-dimensional irregular surface model. The metal substrate includes a mesh-like support structure and embedded heat pipe channels, specifically including the following steps: Copper powder, graphene nanosheets and binder are mixed and spherical composite powder with a preset particle size is prepared by gas atomization process. The graphene nanosheets are oriented along the surface of copper particles. The geometric features of the three-dimensional irregular surface model are analyzed, and the metal substrate is divided into a high thermal conductivity zone and a heat pipe functional channel zone according to the heat flux density distribution, generating partitioned laser power control parameters; Using a dual-beam laser system, under inert gas protection, spherical composite powder is melted layer by layer according to the zonal control parameters, and a mesh-like support structure and pre-embedded grooves for heat pipe channels are formed simultaneously. The pre-embedded groove is filled with phase change material, and the sidewalls of the groove are metallurgically bonded by local laser remelting to form an embedded heat pipe channel, thereby forming a metal matrix. The metal matrix is ​​subjected to hot isostatic pressing to eliminate internal residual stress and improve thermal conductivity.

[0009] Optionally, the phase change material is a bismuth-based alloy with a melting point of 58°C, and the specific process of the metallurgical bonding is as follows: After filling the pre-embedded groove with bismuth-based alloy powder, a short-pulse laser of a preset wavelength is used to locally scan the sidewall of the groove, so that the surface metal of the substrate and the bismuth-based alloy are fused together to form a diffusion bonding layer. The molten pool was then solidified at a gradient cooling rate of 10-15℃ / s, resulting in a Bi-Cu intermetallic compound layer of a preset thickness at the interface.

[0010] Optionally, the primary array is composed of orthogonally staggered rhomboid microgrooves, with each rhomboid unit having a side length of 400-600 μm, a groove depth of 80-120 μm, and a spacing of 50-100 μm between adjacent units; the long axis of the rhomboid microgrooves is consistent with the spatial distribution direction of the grid-like support structure of the metal substrate, forming a continuous heat flow guiding channel.

[0011] Optionally, the secondary protrusion structure is an array of conical micropillars distributed on the surface of the primary array of rhomboid microgrooves; the axis of the conical micropillars is at an angle of 15° to 30° to the normal direction of the bottom surface of the primary microgrooves, and its surface is covered with a composite coating of titanium dioxide / silicon carbide.

[0012] Optionally, the step of forming a multi-level fractal microstructure on the surface of the metal substrate through an electrochemical micromachining process specifically includes the following steps: The surface of the metal substrate is pretreated by electrolytic polishing using a phosphoric acid-sulfuric acid mixed solution under pulsed current to reduce the surface roughness to a preset qualified threshold. Based on the designed rhombic microgroove parameters, S32 uses a mask-assisted micro-electrolysis process to etch a rhombic microgroove array of preset depths onto the metal substrate surface using a sodium nitrate-glycerol electrolyte and pulsed voltage, thus fabricating a primary array. Switching to the ultrasonic vibration-assisted electrolytic processing mode, a conical microelectrode is used to perform three-dimensional scanning along the surface of the rhomboid microgroove. By adjusting the electrolyte flow rate and current density in real time, a conical micropillar of a preset height is generated to form a two-dimensional protrusion structure. After electrolytic processing, an atomized precursor containing tetraethyl titanate and tetraethyl orthosilicate is introduced into the surface of the primary array and secondary protrusion structure. The local plasma generated during electrolysis is used to excite vapor deposition to form a composite coating of titanium dioxide / silicon carbide with a predetermined thickness.

[0013] Optionally, a gradient refractive index photonic crystal layer and a periodic nanoparticle layer are sequentially deposited on the surface of the multi-level fractal microstructure to form a composite functional layer with a superhydrophobic surface and high infrared emissivity, specifically including the following steps: Plasma pretreatment was performed on the surface of a multi-level fractal microstructure under an inert gas atmosphere to remove surface oxides and activate chemical bonding sites. A magnetron sputtering-atomic layer deposition composite process was used to alternately deposit titanium dioxide and aluminum oxide thin films. The film thickness gradient was controlled by dynamic mask movement to form a gradient refractive index photonic crystal layer with a continuous refractive index ranging from 2.7 to 1.8. A single layer of polystyrene nanospheres was spin-coated onto the surface of the gradient refractive index photonic crystal layer as a mask, and silver nanoparticles were deposited by electron beam evaporation. The nanosphere mask was then removed to form a hexagonal close-packed periodic array to assemble a periodic nanoparticle layer. Annealing enhancement of composite functional layer: Annealing at 400℃ for 30 min in a reducing atmosphere promotes the formation of Ti-Ag intermetallic compound at the interface between silver nanoparticles and gradient layer, simultaneously improving superhydrophobic properties and infrared emissivity; wherein, the reducing atmosphere is H2 / N2=1:9.

[0014] Optionally, the morphology of the composite functional layer is iteratively optimized by adjusting processing parameters through a feedback control system based on real-time detected hydrophobic angle and thermal radiation performance data, specifically including the following steps: The hydrophobic angle data and preset band emissivity data of the composite functional layer are collected simultaneously by a contact angle meter and an infrared spectrometer. The surface morphology and topological features are obtained by scanning equipment, and a real-time performance database is constructed. Real-time performance data is input into a pre-trained multi-objective convolutional neural network model to generate optimized parameters for laser power correction, atomic layer deposition cycle number adjustment, and nanoparticle mask displacement compensation. The deposition process of the composite functional layer is adjusted according to the optimized parameters. Selective reprocessing of the gradient refractive index photonic crystal layer and the periodic nanoparticle layer is performed in local areas until the measured hydrophobic angle and infrared emissivity are within the qualified threshold range.

[0015] The present invention also provides a surface treatment apparatus for a GPU main body structure, applied to implement the irregular surface treatment method for the GPU main body structure as described above, wherein the surface treatment apparatus specifically includes: The additive manufacturing mechanism integrates a dual-beam laser melting system, a metal powder supply device, and a hot isostatic pressing chamber. The electrochemical processing mechanism is equipped with a mask positioning mechanism, an ultrasonic vibration electrolytic cell, and an in-situ plasma deposition module. Functional layer deposition equipment, including magnetron sputtering-atomic layer deposition composite coating machine, nanosphere self-assembly platform and electron beam evaporation device; The dynamic monitoring unit is equipped with a contact angle measuring instrument, an infrared spectral detection head, and a multi-axis precision motion platform; The control module integrates a modeling control unit and multiple decision-making models to coordinate process parameters in real time and generate iterative instructions.

[0016] Compared with existing technologies, this invention has the following advantages: First, a three-dimensional irregular surface model with electromagnetic shielding and enhanced thermal radiation is generated through multiphysics coupling simulation, realizing the precise forming of complex electromagnetic shielding structures; based on this model, a metal substrate with mesh support and embedded heat pipes is formed layer by layer using metal additive manufacturing process; then, a multi-level fractal microstructure is constructed on the substrate surface using electrochemical micromachining, and a gradient refractive index photonic crystal layer and a periodic nanoparticle layer are deposited sequentially to form a composite functional layer. The multi-level fractal micromachining significantly improves the surface area and interfacial bonding force, enhancing the heat transfer efficiency of the heat dissipation channel; by real-time detection of hydrophobic angle and thermal radiation performance data, the processing parameters are adjusted to achieve dynamic iterative optimization of the composite functional layer, ensuring dynamic matching between process parameters and target performance, forming a high-performance surface with electromagnetic shielding, efficient heat dissipation and superhydrophobic properties, solving the defect of uncontrollable functional layer performance in traditional processes, and improving the overall reliability and service life of GPU structural components. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] The structures, proportions, sizes, etc., shown in the accompanying drawings of this specification are only for the purpose of assisting those skilled in the art in understanding and reading the content disclosed in the specification, and are not intended to limit the conditions under which the present invention can be implemented. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0019] Figure 1 This is one of the flowcharts illustrating the irregular surface treatment method for the GPU main structural component in this embodiment. Figure 2 This is the second flowchart illustrating the irregular surface treatment method for the GPU main structural component in this embodiment one; Figure 3This is a schematic diagram of the multi-level fractal microstructure of the GPU main structural component in this embodiment. Figure 4 This is a schematic diagram of the system structure of the surface treatment device for the GPU main structural component in this embodiment 2. Detailed Implementation

[0020] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0021] In the description of this invention, it should be understood that the terms "upper," "lower," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. It should be noted that when a component is considered to be "connected" to another component, it can be directly connected to the other component or there may be a component positioned centrally in the connection.

[0022] The technical solution of the present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0023] Example 1: Combination Figures 1 to 3 As shown, this embodiment of the invention provides a method for irregular surface treatment of a GPU main structural component, including: S1, based on environmental simulation data of the target application scenario, generates a three-dimensional irregular surface model with electromagnetic shielding function and thermal radiation enhancement characteristics through multi-physics field coupling simulation. By collecting environmental parameters of the target application scenario (such as heat load, electromagnetic interference spectrum, dust distribution, etc.), a three-dimensional irregular surface model is established using multiphysics coupling simulation technology (electromagnetic field, thermodynamics, fluid mechanics). This model balances electromagnetic shielding requirements with thermal radiation enhancement characteristics through a dynamic weight allocation algorithm, generating a surface morphology with a non-uniform topology.

[0024] Environmental data is directly converted into manufacturable geometric models, ensuring that the structural design for subsequent processing can proactively adapt to the actual working conditions.

[0025] S2 uses metal additive manufacturing process to form a metal substrate layer by layer according to a three-dimensional irregular surface model. The metal substrate includes a grid-like support structure 200 and an embedded heat pipe channel. Metal additive manufacturing technologies (such as selective laser melting) are used to form a metal substrate layer by layer. Through a zoned energy control strategy, a grid-like support structure 200 (to improve mechanical strength) and embedded heat pipe channels (to optimize heat conduction paths) are formed simultaneously. This integrated manufacturing of complex irregular structures solves the problem of coordinating the forming of internal cavities (such as heat pipe channels) and external irregular surfaces, which is difficult to achieve with traditional machining, providing a high-precision substrate for surface microstructure processing.

[0026] S3, a multi-level fractal microstructure 100 is formed on the surface of a metal substrate by electrochemical micromachining process. The multi-level fractal microstructure 100 includes a primary array 101, a secondary protrusion structure 102 and a composite coating. By increasing the effective surface area through multi-level microstructure design and combining the chemical properties of the composite coating, the hydrophobicity and wear resistance of the surface are simultaneously enhanced, providing an anchoring interface for the subsequent deposition of functional layers.

[0027] S4, a gradient refractive index photonic crystal layer and a periodic nanoparticle layer are sequentially deposited on the surface of the multi-level fractal microstructure 100 to form a composite functional layer with superhydrophobic surface and high infrared emissivity. The gradient refractive index layer improves thermal radiation efficiency by modulating the band structure in the infrared radiation band, while the nanoparticle array enhances hydrophobicity through surface plasmon effects. Together, they achieve an integrated function of heat dissipation and anti-pollution.

[0028] S5, based on real-time monitoring of hydrophobic angle and thermal radiation performance data, adjusts processing parameters through a feedback control system to iteratively optimize the morphology of the composite functional layer. A closed-loop feedback mechanism adapts to processing errors and environmental fluctuations, ensuring the stability and consistency of the final product's performance.

[0029] The working principle of this invention is as follows: First, a three-dimensional irregular surface model with electromagnetic shielding and enhanced thermal radiation is generated through multiphysics coupling simulation, realizing the precise forming of complex electromagnetic shielding structures. Based on this model, a metal substrate with grid support and embedded heat pipes is formed layer by layer using metal additive manufacturing technology. Then, a multi-level fractal microstructure 100 is constructed on the substrate surface using electrochemical micromachining, and a gradient refractive index photonic crystal layer and a periodic nanoparticle layer are deposited sequentially to form a composite functional layer. Multi-level fractal micromachining significantly improves the surface area and interfacial bonding force, and enhances the heat transfer efficiency of the heat dissipation channel. By real-time detection of hydrophobic angle and thermal radiation performance data, the processing parameters are adjusted to achieve dynamic iterative optimization of the composite functional layer, ensuring dynamic matching between process parameters and target performance, forming a high-performance surface with electromagnetic shielding, efficient heat dissipation and superhydrophobic properties. This solves the defect of uncontrollable performance of functional layers in traditional processes, and improves the overall reliability and service life of GPU structural components.

[0030] In this embodiment, step S1 specifically includes: S11: Collect environmental parameters of the target application scenario, and acquire dynamic heat load data, environmental dust concentration distribution data and electromagnetic interference spectrum data of electronic devices during operation through sensor network to build a multi-dimensional environmental feature database; By deploying a multimodal sensor network (such as thermal sensors, dust concentration detectors, and electromagnetic probes) within electronic devices, dynamic heat load distribution, environmental dust particle concentration, and electromagnetic interference spectrum characteristics data are collected in real time during device operation. This data, after filtering and noise reduction, is used to construct a multidimensional environmental feature database along the time-space dimension. This provides realistic operating condition input for subsequent simulation modeling, ensuring that the design of the 3D model accurately matches the physical constraints of the actual application scenario (such as heat accumulation and electromagnetic interference intensity).

[0031] S12, input environmental characteristic data into the electromagnetic-thermal-fluid coupling simulation system, calculate the synergistic relationship between electromagnetic shielding requirements and thermal radiation efficiency through dynamic weight allocation algorithm, and generate an initial three-dimensional irregular surface model with surface energy distribution. The database was imported into the electromagnetic-thermal-fluid coupling simulation system. A dynamic weight allocation algorithm (based on the real-time ratio of thermal radiation efficiency to electromagnetic shielding performance requirements) was used to calculate the energy distribution priority, generating an initial three-dimensional irregular surface model. The model includes the electromagnetic shielding layer thickness gradient distribution and spatial coordinate markings for the thermal radiation enhancement region.

[0032] S13, Perform reverse topology optimization: Based on the initial three-dimensional irregular surface model, the reverse Monte Carlo algorithm is used to iteratively generate candidate irregular surface structures, and the electromagnetic shielding coefficient and infrared emissivity performance of each candidate irregular surface structure are predicted by a convolutional neural network in order to optimize the three-dimensional irregular surface model. Based on the initial model, candidate irregular surface structures (such as fractal grooves and asymmetric protrusions) are randomly generated using the inverse Monte Carlo algorithm. A pre-trained convolutional neural network is then used to quickly predict the electromagnetic shielding coefficient (SE) and infrared emissivity (ε) of each candidate structure, selecting the topological configuration with the best overall performance. By replacing traditional trial-and-error methods with data-driven intelligent optimization, the design cycle for complex irregular structures is significantly shortened, while avoiding local optimum traps.

[0033] S14. Finite element electromagnetic compatibility simulation and discrete element dust adsorption simulation are performed on the optimized initial three-dimensional irregular surface model. The microstructure curvature and porosity are adjusted by combining the adaptive correction algorithm until the preset performance threshold is met. Finite element electromagnetic compatibility simulation (to assess electromagnetic leakage risk) and discrete element dust adsorption simulation (to predict dust deposition distribution) were performed on the optimized model. The adaptive correction algorithm was used to dynamically adjust parameters such as the radius of curvature and porosity of the microstructure until the model met the preset electromagnetic shielding threshold and anti-dust accumulation performance index.

[0034] S15 converts the verified structural parameters into a manufacturable 3D model format, and simultaneously embeds electromagnetic shielding layer thickness gradient data and thermal radiation enhancement area marking information.

[0035] The validated structural parameters are converted into STL format 3D models compatible with additive manufacturing equipment, and process feature information (such as material deposition gradient instructions for electromagnetic shielding layers and laser energy density markings for thermal radiation enhancement regions) is embedded in the models.

[0036] In this embodiment, step S2 specifically includes the following steps: S21, copper powder, graphene nanosheets and binder are mixed and spherical composite powder with a preset particle size is prepared by gas atomization process, and the graphene nanosheets are oriented along the surface of copper particles. Copper powder and graphene nanosheets were mixed in a specific ratio and then atomized to prepare spherical composite powder. During this process, the graphene nanosheets adhered directionally to the surface of the copper particles through the flow field, forming a uniform coating structure. The addition of a binder ensured the stability of the powder during subsequent processing. The directional arrangement of graphene optimized the thermal conductivity path of the composite powder, providing a base material with both high thermal conductivity and structural strength for subsequent additive manufacturing, while avoiding the graphene agglomeration problem in traditional mixing processes.

[0037] S22, analyze the geometric features of the three-dimensional irregular surface model, divide the metal matrix into a high thermal conductivity zone (mesh support structure 200) and a heat pipe functional channel zone according to the heat flux density distribution, and generate partitioned laser power control parameters; Based on the geometric features and heat flux density distribution data of the 3D irregular surface model, the metal substrate is divided into a high thermal conductivity zone (grid-like support structure 200) and a heat pipe functional channel zone. By analyzing the laser power control parameters for generating the heat flux path partitions, for example, high energy density is used in the high thermal conductivity zone to enhance melt density, while lower energy is used in the heat pipe channel zone to prevent over-sintering. This enables differentiated processing of different functional regions of the metal substrate, ensuring the synergistic optimization of the mechanical properties of the grid support structure and the geometric accuracy of the heat pipe channels, providing a suitable physical carrier for subsequent thermal management functions.

[0038] S23 employs a dual-beam laser system to melt spherical composite powder layer by layer under inert gas protection, according to the zone control parameters, and simultaneously forms a grid-like support structure 200 and a pre-embedded groove for the heat pipe channel. A dual-wavelength laser system (such as near-infrared and ultraviolet lasers) is employed in an inert gas protected environment to melt composite powder layer by layer according to zonal control parameters. The long-wavelength laser is used for the rapid prototyping of large-area grid support structures, while the short-wavelength laser is focused on the precision machining of pre-embedded grooves in heat pipes, simultaneously achieving integrated manufacturing of macroscopic structures and microscopic features. This dual-beam energy distribution strategy resolves the efficiency and precision contradictions inherent in traditional single-laser systems for processing complex irregular structures, while also avoiding the impact of oxidation on material properties.

[0039] S24, fill the pre-embedded groove with phase change material, and achieve metallurgical bonding of the groove sidewall through local laser remelting, and encapsulate to form an embedded heat pipe channel to obtain a metal matrix; After filling the pre-embedded groove with bismuth-based phase change material, the sidewalls of the groove are remelted by local laser, causing a eutectic reaction between the base metal and the phase change material to form a metallurgically bonded sealed interface. This process ensures that the phase change material retains its preset melting point characteristics after encapsulation by precisely controlling the laser energy input.

[0040] S25 involves hot isostatic pressing of the metal substrate to eliminate internal residual stress and improve thermal conductivity.

[0041] The formed metal substrate is placed in a high-temperature and high-pressure environment for hot isostatic pressing. The internal porosity and residual stress generated during additive manufacturing are eliminated by isotropic pressure, while promoting further densification of the copper-graphene interface.

[0042] In this embodiment, it is further explained that the phase change material is a bismuth-based alloy with a melting point of 58°C, and the specific metallurgical bonding process is as follows: After filling the pre-embedded groove with bismuth-based alloy powder, a short-pulse laser of a preset wavelength (pulse width 100 ns, power density 2 × 10⁻⁶) is used. 4 W / cm 2Local scanning of the groove sidewalls allows the base metal and bismuth-based alloy to eutecticly fuse and form a diffusion bonding layer; short pulse widths limit the range of the heat-affected zone, preventing excessive vaporization of the phase change material (bismuth-based alloy); Subsequently, the solidification of the molten pool was controlled by a gradient cooling rate of 10-15℃ / s, and a Bi-Cu intermetallic compound layer with a preset thickness was generated at the interface; this suppressed the microcracks at the interface caused by rapid cooling and improved the thermal cycling stability of the metallurgical bonding layer.

[0043] In this embodiment, it is further explained that, in combination with Figure 3 As shown, the primary array 101 consists of orthogonally staggered rhomboid microgrooves. Each rhomboid unit has a side length of 400-600 μm, a groove depth of 80-120 μm, and a spacing of 50-100 μm between adjacent units. The long axis of the rhomboid microgrooves aligns with the spatial distribution direction of the mesh support structure 200 of the metal substrate, forming a continuous heat flow guiding channel. The orthogonally staggered rhomboid microgrooves form a continuous heat flow guiding channel through a regular geometric arrangement. The alignment of the rhomboid long axis with the mesh support structure 200 allows heat to be conducted along a predetermined path to the heat pipe region. This design, in conjunction with the heat flow path of the mesh support structure, reduces the tortuosity of the heat conduction path.

[0044] In this embodiment, it is further explained that the secondary protrusion structure 102 is a conical micropillar array distributed on the surface of the rhomboid microgroove of the primary array 101; the axis of the conical micropillar is at an angle of 15° to 30° to the normal direction of the bottom surface of the primary microgroove, and its surface is covered with a composite coating of titanium dioxide / silicon carbide.

[0045] Conical micropillars (20-30 μm in diameter at the base and 50-80 μm in height) are distributed at an angle of 15°-30° on the surface of rhomboid microgrooves, with their axial angle design guiding fluid flow in a specific direction. A dense protective layer is formed on the surface by chemical vapor deposition of a titanium dioxide / silicon carbide composite coating.

[0046] The conical micropillars are angled, and the angled design disrupts the static contact of surface droplets, enhancing the self-cleaning effect of hydrophobic properties; titanium dioxide provides photocatalytic self-cleaning properties, and silicon carbide improves the wear resistance of the coating. The combination of the two extends the life of the functional layer.

[0047] In this embodiment, step S3 specifically includes the following steps: S31, the surface of the metal substrate is pretreated by electrolytic polishing using a phosphoric acid-sulfuric acid mixed solution under pulsed current to reduce the surface roughness to a preset qualified threshold. Electropolishing of metal substrate surfaces was performed using a phosphoric acid-sulfuric acid mixed solution (volume ratio 3:1) under pulsed current. The strong oxidizing properties of phosphoric acid rapidly dissolved surface oxides, while the corrosive effect of sulfuric acid, combined with the intermittent electrolysis of the pulsed current, achieved uniform dissolution of microscopic protrusions on the surface, thereby reducing roughness to a preset threshold.

[0048] Based on the designed rhombic microgroove parameters, S32 uses a mask-assisted micro-electrolysis process to etch a rhombic microgroove array of a predetermined depth onto the metal substrate surface using a sodium nitrate-glycerol electrolyte and pulsed voltage, thus fabricating a primary array 101. Based on a mask-defined primary array pattern of 101, micro-electrochemical processing was performed using a sodium nitrate-glycerol electrolyte (15wt%) and pulsed voltage (10V, 50μs). Sodium nitrate provided a stable electrolytic environment, the viscosity of glycerol inhibited lateral corrosion, and the pulsed voltage precisely controlled the etching rate, forming a diamond-shaped microgroove array of a preset depth.

[0049] S33, switch to ultrasonic vibration assisted electrolytic processing mode, use conical microelectrode to perform three-dimensional scanning along the surface of rhomboid microgroove, and generate conical micropillars of preset height by adjusting electrolyte flow rate and current density in real time to form secondary protrusion structure 102; Switching to ultrasonic vibration-assisted electrolytic machining mode, a conical microelectrode with a tip diameter of 10 μm is used to scan along the surface of the rhomboid microgroove. The electrolyte flow rate (0.5-2 m / s) and current density (20-50 A / cm²) are adjusted in real time. 2 The ultrasonic cavitation effect is used to promptly remove electrolysis products, generating conical micropillars of a predetermined height under high aspect ratio conditions. This timely removal of electrolysis products through ultrasonic cavitation avoids the "blind hole" effect in deep hole machining. A dynamic parameter combination (high flow rate + low current density) suppresses over-etching at the top of the micropillars, ensuring a height tolerance of ±5μm. The conical electrode design matches the tapered morphology of the conical micropillars, achieving directional etching of asymmetric structures.

[0050] S34, an atomized precursor containing tetraethyl titanate and tetraethyl orthosilicate is introduced into the surface of the primary array 101 and the secondary protrusion structure 102 after electrolytic processing. The local plasma generated during the electrolysis process is used to excite vapor deposition to form a composite coating of titanium dioxide / silicon carbide with a preset thickness.

[0051] An atomized precursor (tetraethyl titanate + tetraethyl orthosilicate) is introduced onto the surface of the microstructure after electrolytic processing. The localized plasma residue from the electrolysis activates the precursor decomposition, depositing a TiO2 / SiC composite coating (300±50 nm thick). Tetraethyl titanate generates photocatalytic TiO2, while tetraethyl orthosilicate provides the wear-resistant SiC phase; the combination of these two enhances the durability of the functional layer. In-situ plasma excitation requires no additional energy source, and the coating thickness uniformity is controlled by linking the atomization rate with the electrolysis termination time.

[0052] In this embodiment, step S4 specifically includes the following steps: S41, Plasma pretreatment is performed on the surface of the multi-level fractal microstructure 100 in an inert gas atmosphere to remove surface oxides and activate chemical bonding sites. In an inert gas atmosphere (such as argon), the surface of the multi-level fractal microstructure 100 was bombarded for 5 minutes using argon plasma with a power of 300W. The high-energy ions in the plasma underwent physical sputtering and chemical reactions (such as reduction reactions) with the surface oxides, removing contaminants and exposing fresh metal surfaces, while activating chemical bonding sites such as hydroxyl groups (-OH).

[0053] Argon is chosen as the inert gas for its chemical inertness, which avoids secondary oxidation of the surface and ensures the interface purity of the subsequent deposited layer. The 300W power balances the surface activation efficiency and the risk of thermal damage within 5 minutes (avoiding microstructure deformation). The activation of chemical bonding sites provides active sites for the chemical adsorption of the gradient refractive index layer, enhancing the interfacial bonding strength.

[0054] S42 employs a magnetron sputtering-atomic layer deposition composite process to alternately deposit titanium dioxide and alumina thin films. The thickness gradient of the thin films is controlled by dynamic mask movement (the thickness of titanium dioxide gradually changes from 50nm to 200nm, and the thickness of alumina gradually changes from 200nm to 50nm), forming a gradient refractive index photonic crystal layer with a continuously varying refractive index from 2.7 to 1.8. A composite process employing alternating magnetron sputtering and atomic layer deposition (ALD) is used, with the thickness gradient of the two materials controlled by continuous displacement of a dynamic mask. This thickness gradient variation results in a continuous transition of the equivalent refractive index after material mixing from 2.7 (high titanium dioxide content) to 1.8 (high alumina content). The magnetron sputtering-ALD composite process enables rapid titanium dioxide deposition, while ALD ensures the density and thickness accuracy of the alumina film.

[0055] Refractive index range design (2.7→1.8): to match the photonic bandgap requirements of the infrared band (8-14μm) and enhance thermal radiation efficiency.

[0056] S43, a single layer of polystyrene nanospheres is spin-coated on the surface of a gradient refractive index photonic crystal layer as a mask, silver nanoparticles are deposited by electron beam evaporation, and then the nanosphere mask is removed to form a hexagonal close-packed periodic array to assemble a periodic nanoparticle layer. A single layer of polystyrene nanospheres was spin-coated onto the surface of a gradient refractive index layer as a mask, and silver nanoparticles were deposited by electron beam evaporation. After the nanospheres were removed, the silver particles formed a hexagonal close-packed periodic array on the surface, and the gaps between them generated a localized surface plasmon resonance effect.

[0057] Nanosphere particle size control matches the target silver particle size, ensuring precise correspondence between mask aperture and particle deposition position; hexagonal close-packed array periodic structure enhances surface plasmon coupling effect, improving hydrophobicity and infrared absorption characteristics.

[0058] S44, Annealing Enhancement of Composite Functional Layer: Annealing at 400℃ for 30 min in a reducing atmosphere promotes the formation of Ti-Ag intermetallic compounds at the interface between silver nanoparticles and gradient layer, simultaneously improving superhydrophobic properties and infrared emissivity; wherein, the reducing atmosphere is H2 / N2=1:9.

[0059] In a reducing atmosphere of H2 / N2 = 1:9, annealing at 400℃ for 30 minutes promotes the formation of Ti-Ag intermetallic compounds at the interface between silver particles and the TiO2 layer. A low concentration of hydrogen (10%) reduces the risk of explosion while reducing the oxide; the 400℃ temperature matches the Ti-Ag eutectic point, ensuring sufficient compound formation and preventing the silver particles from melting and agglomerating. The Ti-Ag compound enhances interfacial bonding through chemical bonding and simultaneously adjusts the surface electronic state distribution, synergistically improving superhydrophobicity and infrared emissivity.

[0060] In this embodiment, step S5 specifically includes the following steps: S51 uses a contact angle meter and an infrared spectrometer to simultaneously collect hydrophobic angle data and preset band emissivity data of the composite functional layer, and combines scanning equipment to obtain surface morphology topology features to build a real-time performance database. Hydrophobic angle data and emissivity data in the 8-14 μm band (a preset key infrared radiation band) of the composite functional layer were simultaneously acquired using a contact angle meter and an infrared spectrometer. Surface morphology and topological features (such as microstructure height and spacing) were then obtained using a high-resolution scanning electron microscope. The synchronous acquisition mode ensures the temporal and spatial consistency of performance data, and the preset band selection targets the peak range of infrared radiation spectrum specific to heat dissipation scenarios, providing accurate input for subsequent optimization. The correlation between surface morphology data and performance parameters forms a multi-dimensional performance database, supporting the reliability of dynamic optimization decisions.

[0061] S52, Dynamic Parameter Optimization Decision: Input real-time performance data into a pre-trained multi-objective convolutional neural network model to generate optimized parameters for laser power correction, atomic layer deposition cycle number adjustment, and nanoparticle mask displacement compensation. Real-time performance data is input into a pre-trained multi-objective convolutional neural network (CNN) model to generate laser power correction values ​​(±15W), atomic layer deposition (ALD) cycle number adjustment (±5 times), and nanoparticle mask displacement compensation parameters (±5μm). The CNN model, trained on a historical process-performance dataset, can identify the mapping relationship between morphological defects (such as coating inhomogeneity) and performance indicators (hydrophobicity angle, emissivity), balancing heat dissipation and hydrophobicity requirements through a multi-objective optimization algorithm. The parameter adjustment range (e.g., ±15W laser power) is set based on the equipment's process window to avoid overshooting that could damage the functional layer.

[0062] S53, closed-loop feedback iterative correction: adjust the deposition process of the composite functional layer according to the optimized parameters, and perform selective reprocessing of the gradient refractive index photonic crystal layer and the periodic nanoparticle layer in the local area until the measured hydrophobic angle and infrared emissivity are within the qualified threshold range.

[0063] The deposition equipment (e.g., laser power, ALD precursor flux) is adjusted based on optimized parameters to selectively reprocess the gradient refractive index layer and nanoparticle layer in localized areas. Through iterative detection-correction cycles, the process gradually approaches acceptable thresholds (e.g., hydrophobic angle ≥160°, infrared emissivity ≥0.95). This localized reprocessing strategy only corrects areas that do not meet performance standards, reducing material waste, while gradient cooling control prevents interface delamination caused by thermal stress.

[0064] Example 2: Combination Figure 4 As shown, the present invention also provides a surface treatment apparatus for a GPU main body structure, applied to implement the irregular surface treatment method for the GPU main body structure as described in Embodiment 1. The surface treatment apparatus specifically includes: The additive manufacturing mechanism 10 integrates a dual-beam laser melting system 11, a metal powder supply device, and a hot isostatic pressing chamber; The electrochemical processing mechanism 20 is equipped with a mask positioning mechanism 21, an ultrasonic vibration electrolytic cell 22, and an in-situ plasma deposition module 23. Functional layer deposition equipment 30 includes a magnetron sputtering-atomic layer deposition composite coating machine 31, a nanosphere self-assembly platform 32, and an electron beam evaporation device 33; The dynamic monitoring unit 40 is equipped with a contact angle measuring instrument, an infrared spectral detection head, and a multi-axis precision motion platform; The control module 50 integrates a modeling control unit and multiple decision models to coordinate process parameters in real time and generate iterative instructions.

[0065] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for surface treatment of irregular shapes in a GPU main structural component, characterized in that, include: Based on environmental simulation data of the target application scenario, a three-dimensional irregular surface model with electromagnetic shielding function and thermal radiation enhancement feature is generated through multi-physics field coupling simulation. Using metal additive manufacturing process, a metal substrate is formed layer by layer based on the three-dimensional irregular surface model. The metal substrate includes a grid-like support structure and embedded heat pipe channels. A multi-level fractal microstructure is formed on the surface of the metal substrate by an electrochemical micromachining process. The multi-level fractal microstructure includes a primary array, a secondary protrusion structure, and a composite coating. A gradient refractive index photonic crystal layer and a periodic nanoparticle layer are sequentially deposited on the surface of the multi-level fractal microstructure to form a composite functional layer with a superhydrophobic surface and high infrared emissivity. Based on real-time detected hydrophobic angle and thermal radiation performance data, the morphology of the composite functional layer is iteratively optimized by adjusting the processing parameters through a feedback control system.

2. The method for irregular surface treatment of GPU main structural components according to claim 1, characterized in that, The environmental simulation data based on the target application scenario is used to generate a three-dimensional irregular surface model with electromagnetic shielding function and enhanced thermal radiation characteristics through multi-physics field coupling simulation, specifically including: Collect environmental parameters of the target application scenario, and acquire dynamic heat load data, environmental dust concentration distribution data and electromagnetic interference spectrum data of electronic devices during operation through sensor networks to construct a multi-dimensional environmental feature database; The environmental characteristic data is input into the electromagnetic-thermal-fluid coupling simulation system, and the synergistic relationship between electromagnetic shielding requirements and thermal radiation efficiency is calculated through a dynamic weight allocation algorithm to generate an initial three-dimensional irregular surface model with surface energy distribution. Perform reverse topology optimization: Based on the initial three-dimensional irregular surface model, the reverse Monte Carlo algorithm is used to iteratively generate candidate irregular surface structures, and the electromagnetic shielding coefficient and infrared emissivity performance of each candidate irregular surface structure are predicted by a convolutional neural network in order to optimize the three-dimensional irregular surface model. Finite element electromagnetic compatibility simulation and discrete element dust adsorption simulation were performed on the optimized initial three-dimensional irregular surface model. The curvature and porosity of the microstructure were adjusted by an adaptive correction algorithm until the preset performance threshold was met. The validated structural parameters are converted into a manufacturable 3D model format, and electromagnetic shielding layer thickness gradient data and thermal radiation enhancement area marking information are embedded simultaneously.

3. The method for irregular surface treatment of GPU main structural components according to claim 1, characterized in that, The process employs metal additive manufacturing, forming a metal substrate layer by layer based on the three-dimensional irregular surface model. The metal substrate includes a mesh-like support structure and embedded heat pipe channels. Specifically, the process includes the following steps: Copper powder, graphene nanosheets and binder are mixed and spherical composite powder with a preset particle size is prepared by gas atomization process. The graphene nanosheets are oriented along the surface of copper particles. The geometric features of the three-dimensional irregular surface model are analyzed, and the metal substrate is divided into a high thermal conductivity zone and a heat pipe functional channel zone according to the heat flux density distribution, generating partitioned laser power control parameters; Using a dual-beam laser system, under inert gas protection, spherical composite powder is melted layer by layer according to the zonal control parameters, and a mesh-like support structure and pre-embedded grooves for heat pipe channels are formed simultaneously. The pre-embedded groove is filled with phase change material, and the sidewalls of the groove are metallurgically bonded by local laser remelting to form an embedded heat pipe channel, thereby forming a metal matrix. The metal matrix is ​​subjected to hot isostatic pressing to eliminate internal residual stress and improve thermal conductivity.

4. The method for irregular surface treatment of GPU main structural components according to claim 3, characterized in that, The phase change material is a bismuth-based alloy with a melting point of 58°C, and the specific process of the metallurgical bonding is as follows: After filling the pre-embedded groove with bismuth-based alloy powder, a short-pulse laser of a preset wavelength is used to locally scan the sidewall of the groove, so that the surface metal of the substrate and the bismuth-based alloy are fused together to form a diffusion bonding layer. The molten pool was then solidified at a gradient cooling rate of 10-15℃ / s, resulting in a Bi-Cu intermetallic compound layer of a preset thickness at the interface.

5. The method for irregular surface treatment of GPU main structural components according to claim 1, characterized in that, The primary array is composed of orthogonally staggered rhomboid microgrooves, with each rhomboid unit having a side length of 400-600 μm, a groove depth of 80-120 μm, and a spacing of 50-100 μm between adjacent units. The long axis of the rhomboid microgrooves is aligned with the spatial distribution direction of the grid-like support structure of the metal substrate, forming a continuous heat flow guiding channel.

6. The method for irregular surface treatment of GPU main structural components according to claim 5, characterized in that, The secondary protrusion structure is an array of conical micropillars distributed on the surface of the primary array of rhomboid microgrooves; the axis of the conical micropillars is at an angle of 15° to 30° to the normal direction of the bottom surface of the primary microgrooves, and its surface is covered with a composite coating of titanium dioxide / silicon carbide.

7. The method for irregular surface treatment of GPU main structural components according to claim 6, characterized in that, The process of forming a multi-level fractal microstructure on the surface of the metal substrate using an electrochemical micromachining technique specifically includes the following steps: The surface of the metal substrate is pretreated by electrolytic polishing using a phosphoric acid-sulfuric acid mixed solution under pulsed current to reduce the surface roughness to a preset qualified threshold. Based on the designed rhombic microgroove parameters, S32 uses a mask-assisted micro-electrolysis process to etch a rhombic microgroove array of preset depths onto the metal substrate surface using a sodium nitrate-glycerol electrolyte and pulsed voltage, thus fabricating a primary array. Switching to the ultrasonic vibration-assisted electrolytic processing mode, a conical microelectrode is used to perform three-dimensional scanning along the surface of the rhomboid microgroove. By adjusting the electrolyte flow rate and current density in real time, a conical micropillar of a preset height is generated to form a two-dimensional protrusion structure. After electrolytic processing, an atomized precursor containing tetraethyl titanate and tetraethyl orthosilicate is introduced into the surface of the primary array and secondary protrusion structure. The local plasma generated during electrolysis is used to excite vapor deposition to form a composite coating of titanium dioxide / silicon carbide with a predetermined thickness.

8. The method for irregular surface treatment of GPU main structural components according to claim 1, characterized in that, A gradient refractive index photonic crystal layer and a periodic nanoparticle layer are sequentially deposited on the surface of the multi-level fractal microstructure to form a composite functional layer with a superhydrophobic surface and high infrared emissivity. The specific steps include: Plasma pretreatment was performed on the surface of a multi-level fractal microstructure under an inert gas atmosphere to remove surface oxides and activate chemical bonding sites. A magnetron sputtering-atomic layer deposition composite process was used to alternately deposit titanium dioxide and aluminum oxide thin films. The film thickness gradient was controlled by dynamic mask movement to form a gradient refractive index photonic crystal layer with a continuous refractive index ranging from 2.7 to 1.

8. A single layer of polystyrene nanospheres was spin-coated onto the surface of the gradient refractive index photonic crystal layer as a mask, and silver nanoparticles were deposited by electron beam evaporation. The nanosphere mask was then removed to form a hexagonal close-packed periodic array to assemble a periodic nanoparticle layer. Annealing enhancement of composite functional layer: Annealing at 400℃ for 30 min in a reducing atmosphere promotes the formation of Ti-Ag intermetallic compound at the interface between silver nanoparticles and gradient layer, simultaneously improving superhydrophobic properties and infrared emissivity; wherein, the reducing atmosphere is H2 / N2=1:

9.

9. The method for irregular surface treatment of GPU main structural components according to claim 1, characterized in that, The morphology of the composite functional layer is iteratively optimized by adjusting processing parameters through a feedback control system based on real-time detected hydrophobic angle and thermal radiation performance data. Specifically, this includes the following steps: The hydrophobic angle data and preset band emissivity data of the composite functional layer are collected simultaneously by a contact angle meter and an infrared spectrometer. The surface morphology and topological features are obtained by scanning equipment, and a real-time performance database is constructed. Real-time performance data is input into a pre-trained multi-objective convolutional neural network model to generate optimized parameters for laser power correction, atomic layer deposition cycle number adjustment, and nanoparticle mask displacement compensation. The deposition process of the composite functional layer is adjusted according to the optimized parameters. Selective reprocessing of the gradient refractive index photonic crystal layer and the periodic nanoparticle layer is performed in local areas until the measured hydrophobic angle and infrared emissivity are within the qualified threshold range.

10. A surface treatment device for a GPU main structural component, characterized in that, The surface treatment apparatus for implementing the irregular surface treatment method of the GPU main body structure as described in any one of claims 1 to 9 specifically includes: The additive manufacturing mechanism integrates a dual-beam laser melting system, a metal powder supply device, and a hot isostatic pressing chamber. The electrochemical processing mechanism is equipped with a mask positioning mechanism, an ultrasonic vibration electrolytic cell, and an in-situ plasma deposition module. Functional layer deposition equipment, including magnetron sputtering-atomic layer deposition composite coating machine, nanosphere self-assembly platform and electron beam evaporation device; The dynamic monitoring unit is equipped with a contact angle measuring instrument, an infrared spectral detection head, and a multi-axis precision motion platform; The control module integrates a modeling control unit and multiple decision-making models to coordinate process parameters in real time and generate iterative instructions.

Images