Search

CN-122021334-A - Intelligent optimization method for hot working process parameters of high heat conduction aluminum alloy component

CN122021334ACN 122021334 ACN122021334 ACN 122021334ACN-122021334-A

Abstract

The intelligent optimization method for the thermal processing technological parameters of the high heat conduction aluminum alloy component is characterized in that multidimensional input characteristic data of a target component are obtained, based on a chemical component spectrum analysis report of the target aluminum alloy, mass percent data of key alloy elements are extracted to form vectors, the mass percent of the key alloy elements comprise, by mass percent, 0.05-0.50% of Si, 0.20-1.00% of Mg, 0.50-2.00% of Fe, 0.05-0.25% of B and 0.05-0.40% of rare earth elements RE, and the technical scheme ensures that the thermal conductivity of the material is stable to be more than or equal to 200W/m.K, the heat dissipation requirement of a high-end precision device is remarkably higher than that of a conventional die casting aluminum alloy (generally < 150W/m.K), the extremely low silicon content of the aluminum alloy fundamentally eliminates defects of black spots, white ashes or uneven coloring caused by enrichment of silicon phase during anodic oxidation, and the color, uniform and attractive appearance, high added value of a film layer can be obtained, and the compact surface of the film layer is improved by strictly controlling the total amount of silicon (Si less than 0.5%).

Inventors

  • ZENG ZHUOHUA

Assignees

  • 东莞市佳铠精密金属制品有限公司

Dates

Publication Date
20260512
Application Date
20260209

Claims (10)

  1. 1. An intelligent optimization method for thermal processing technological parameters of a high heat conduction aluminum alloy component is characterized by comprising the following steps: S1, acquiring multi-dimensional input characteristic data of a target component, wherein the multi-dimensional input characteristic data comprises: A. The material composition characteristic vector is based on a chemical composition spectrum analysis report of a target aluminum alloy, and the mass percent data of key alloy elements are extracted to form a vector, wherein the key alloy elements comprise, by mass percent, 0.05-0.50% of Si, 0.20-1.00% of Mg, 0.50-2.00% of Fe, 0.05-0.25% of B, 0.05-0.40% of rare earth elements RE, and the balance of Al and unavoidable impurities with the total amount not exceeding 0.50%; B. Calculating a group of geometric characteristic parameters representing the thermal processing deformation sensitivity of the component by a geometric characteristic extraction algorithm based on a three-dimensional computer aided design model of the target component, wherein the geometric characteristic parameters at least comprise a maximum section thickness (H-max), a minimum section thickness (H-min) and a thickness variation gradient (G-t); S2, after normalizing the material component feature vector and the component geometric feature vector, splicing the material component feature vector and the component geometric feature vector into a comprehensive feature vector, and inputting the comprehensive feature vector into a pre-trained process parameter prediction model; S3, calculating the process parameter prediction model based on the comprehensive feature vector, and outputting an optimal hot working process parameter set aiming at the target component; The process parameter prediction model is a machine learning model trained based on a data set comprising a plurality of groups of historical production data, and each group of historical production data comprises multidimensional input characteristic data corresponding to the step S1, process parameter vectors which are actually adopted and verified successfully and measured performance data of a machined component.
  2. 2. The method according to claim 1, wherein the target aluminum alloy member optimized and treated by the method has properties such that the surface hardness is equal to or more than HV60, the thermal conductivity is equal to or more than 200W/mK, the yield strength is equal to or more than 120MPa, the tensile strength is 180-240MPa, the elongation is equal to or more than 15%, the average size of precipitated phases in a microstructure is equal to or less than 5 μm, and the average grain size of a matrix is equal to or less than 80 μm; the aluminum alloy component can be uniformly adsorbed with dye for coloring after anodic oxidation treatment to form a uniform color oxide film, and black spots formed by enrichment of silicon elements do not exist on the surface of the oxide film.
  3. 3. The method of claim 1, wherein the component geometry feature vector further comprises a ratio (S/V) of equivalent cooling surface area (S) to volume (V), and a structural asymmetry coefficient (A-S); The thickness change gradient (G-t) is calculated in a mode of G-t= (H-max-H-min)/L-path, wherein L-path is the shortest surface path length from the H-max position to the H-min position in the three-dimensional model; The structural asymmetry coefficient (A-s) is calculated in a mode of A-s= |V-left-V-right|/V-total, wherein V-left and V-right are left and right partial volumes divided in a specific main plane by the theoretical gravity center of the component, and V-total is the total volume.
  4. 4. A method according to claim 1 or 3, wherein the process parameter prediction model is an ensemble learning model trained based on a gradient boosting decision tree algorithm; the training process comprises the steps of carrying out Principal Component Analysis (PCA) dimension reduction on input features in the training data set, and carrying out super-parameter optimization by adopting grid search and five-fold cross validation, wherein Euclidean distance between a process parameter vector predicted by a minimized model and a historical success parameter vector is used as an optimization target.
  5. 5. The method according to claim 1, wherein the content of the key alloy elements in the material composition characteristic vector is, in mass percent, 0.05% -0.30% of Si, 0.2% -0.8% of Mg, 0.5% -1.5% of Fe, 0.05% -0.20% of B and 0.05% -0.25% of rare earth element RE; The rare earth element RE is at least one of lanthanum (La), cerium (Ce), scandium (Sc) and yttrium (Y), and the characteristic vector also comprises a derivative characteristic parameter consisting of the mass ratio of Si to Mg (Si/Mg) and the total content of Fe and Mg (Fe+Mg).
  6. 6. The method according to any one of claims 1 to 5, wherein the thermal process is solution heat treatment, and the model output optimal thermal process parameter set is expressed in terms of vectors, including at least a step heating rate combination [ H1, H2], a target soaking temperature T-target±Δt, a holding time T-hold calculated dynamically based on a member thickness, and a zone cooling rate combination [ C-surface, C-core ]; The step heating rate combination [ H1, H2] is specifically that the first heating stage rate H1 is not more than 3 ℃ per minute and is suitable for a heating process from room temperature to 300 ℃, and the second heating stage rate H2 is 5-10 ℃ per minute and is suitable for a heating process from 300 ℃ to solid solution temperature; the target soaking temperature T-target is a dynamic solid solution temperature determined based on the dissolution dynamics of a main strengthening phase of the target alloy, and the range is 470-550 ℃; The heat preservation time t-hold is in direct proportion to the square value of the maximum section thickness (H-max) of the component, and a specific calculation formula is obtained by model learning based on historical data; The parameter set also includes an upper quench delay time limit t-delay-max, which is strictly defined to be within 25 seconds, and which correlates to recommended types of water quenched or polymer quench media with temperature ranges.
  7. 7. The method of claim 6, wherein the thermal process further comprises an aging treatment performed after the solution heat treatment, the model further comprising for a parameter set of an aging treatment output: if the two-stage aging is performed, the specific output of the first-stage aging temperature T-aging1 is 110-130 ℃, the heat preservation time T-aging1 is 2-8 hours, the second-stage aging temperature T-aging2 is 150-180 ℃, and the heat preservation time T-aging2 is 12-24 hours; Controlling the ageing temperature rise rate V-aging of the appearance of the precipitated phase, wherein the rate is limited to be 1.5-2.5 ℃ per minute; an aging process window, taking the target performance as the core, outputs an allowable temperature-time floating range, and ensures that the fluctuation of the yield strength of the component after being processed within the range does not exceed 5% of a standard value.
  8. 8. The method of claim 6, wherein the thermal processing process is an annealing process, and the set of parameters output by the model further comprises: the annealing subclass judges and targets, and definitely aims at relieving stress annealing of internal stress or complete annealing of softened materials, and correspondingly outputs a target residual stress reduction rate or a target Brinell hardness value; An annealing heating rate V-ane inversely related to cold working deflection, the V-ane recommended not to exceed 5 ℃ per minute for components having deflection greater than 15%; the annealing heat preservation temperature T-ane calculated based on the alloy recrystallization temperature ranges from 330 ℃ to 410 ℃, and the heat preservation time T-ane is dynamically adjusted according to the stacking density of the components; The recommended subsequent cooling modes comprise furnace cooling, air cooling or forced air cooling, and the specific cooling rate needs to be controlled between 30 ℃ per hour and 150 ℃ per hour.
  9. 9. An intelligent optimization system for the hot working process parameters of the aluminum alloy component for realizing the method of any one of claims 1-8, which is characterized by comprising a data acquisition and processing module, a feature fusion module, a parameter prediction module and a process file generation module: The data acquisition and processing module is used for acquiring and processing the material component characteristic vector and the component geometric characteristic vector, and the module further comprises: The material component analysis unit is configured to receive an aluminum alloy chemical component report from a spectrum analyzer or an input interface, extract Si, mg, fe, B and mass percent data of RE key elements, verify the data validity according to a preset alloy brand rule base and finally generate a structured material component feature vector; A geometric feature calculation unit configured to import a three-dimensional CAD model file of a target member, automatically recognize and calculate the maximum section thickness (H-max), the minimum section thickness (H-min), the thickness variation gradient (G-t), the equivalent cooling surface area to volume ratio (S/V), and the structural asymmetry coefficient (A-S) by a built-in geometric engine, and package as the member geometric feature vector; The feature fusion module is in communication connection with the data acquisition and processing module and is used for receiving and preprocessing the two feature vectors, and the module further comprises: The normalization processing unit adopts a maximum-minimum normalization method to map each dimension value in the material component feature vector and the component geometric feature vector to a [0,1] interval respectively so as to eliminate dimension influence; The vector splicing unit splices the normalized two feature vectors into a comprehensive feature vector according to a preset sequence, and the comprehensive feature vector is used as unified input of a machine learning model; the parameter prediction module is at the heart of the pre-trained process parameter prediction model and is configured to: loading the process parameter prediction model, wherein the model is trained by the historical data set and is stored in a system memory in a model file form; Receiving the comprehensive feature vector output by the feature fusion module, and calling the model to calculate and infer; Outputting an optimal thermal process parameter set for the current target component, the parameter set being a structured data vector, the contents of which dynamically contain specific parameters as defined in claim 6, 7 or 8 according to the selected process type; The process file generation module is in communication connection with the parameter prediction module and is used for converting an abstract process parameter set into an executable instruction, and the module further comprises: An instruction compilation unit that compiles a set of process parameters (including temperature, time, rate, etc.) output by the model into a specific format (e.g., G-code, XML file, or device-specific instruction set) according to a predefined device communication protocol; and the man-machine interface unit is used for providing a visual interface for displaying the optimal process parameter set and the process curve preview, and comprises a control interface for an operator to confirm, finely tune and issue to the heat treatment equipment.
  10. 10. A method for preparing a high heat conduction aluminum alloy casting, comprising the steps of: s1, smelting, namely weighing raw materials according to the component range of claim 1 or 5, and smelting the raw materials into an aluminum alloy melt with uniform components at 720-760 ℃; S2, refining, namely adding a refining agent into the aluminum alloy melt for refining, wherein the refining temperature is 730-750 ℃, the refining agent is a mixture of hexachloroethane and cryolite, and inert gas is introduced for rotary blowing at the temperature of 730-750 ℃ for at least 15 minutes during refining; S3, casting, namely casting the refined melt into a metal mold preheated to 200-300 ℃ to obtain an ingot; S4, homogenizing, namely preserving the temperature of the cast ingot at 540-580 ℃ for 4-10 hours, and then performing air cooling or water quenching; s5, optimizing and treating the technological parameters, namely taking the cast ingot treated in the step S4 or a component machined by the cast ingot as a target component, adopting the method of any one of claims 1-8 to determine the optimal technological parameters of solution heat treatment, aging treatment or annealing treatment, and carrying out heat treatment according to the parameters.

Description

Intelligent optimization method for hot working process parameters of high heat conduction aluminum alloy component Technical Field The invention relates to the technical field of aluminum alloy, in particular to an intelligent optimization method for hot working process parameters of a high heat conduction aluminum alloy component. Background With the rapid development of science and technology and industrial economy in recent years, the requirements for aluminum alloy are increased, aluminum alloy has been widely applied to heat dissipation parts in the fields of electronic appliances, communication equipment, automobile heat dissipation, aerospace and the like due to the advantages of light weight, proper strength, good processing performance and the like, and with the continuous improvement of power density of electronic components, equipment is rapidly developed in the directions of miniaturization and integration, higher requirements are provided for the heat conduction performance of heat dissipation materials, and the high heat conduction aluminum alloy can effectively transfer and dissipate heat, so that the stable operation and the service life of the equipment are ensured, and the development of the aluminum alloy has important market value and technical significance. In the prior art, in order to improve the heat conduction performance of aluminum alloy, a common method is to adjust the types and the contents of alloy elements. The higher the purity of aluminum, the better the heat conducting property, but the lower the strength of pure aluminum, and the difficulty in meeting the requirements of structural members. Therefore, elements such as silicon (Si), copper (Cu), magnesium (Mg) and the like are often added to aluminum in industry to improve strength, but the addition of such elements tends to introduce more lattice defects and phase interfaces, and the scattering of thermally conductive carriers is significantly increased, thereby causing the thermal conductivity of the material to be reduced. The addition of silicon, while improving casting fluidity and strength to some extent, has a particularly significant negative effect on thermal conductivity. On the other hand, many heat dissipating components not only require good heat conducting capability, but often require surface treatments to enhance corrosion resistance, aesthetics, or to achieve specific functional characteristics. Anodic oxidation is a widely used surface treatment process for aluminum alloys, which can form hard, dense oxide films and can obtain diverse appearances by coloring. However, this process is very sensitive to the composition of the matrix material. When the silicon content in the aluminum alloy is high (for example, exceeds a certain threshold value), silicon phase (or silicide) is easy to be enriched in an oxide film or at the interface of a film layer and a matrix in the anodic oxidation process, so that uneven black spots or stripes are formed, and the appearance uniformity is seriously affected. Meanwhile, silica fume can be generated on the surface of the material, so that an oxide film is not firmly attached and has abnormal porosity, further the adsorption and deposition of dye are hindered, coloring is difficult, the color is dull or uneven, and even the aluminum alloy falls off soon after coloring, and the application of the aluminum alloy with high silicon content in products requiring attractive color appearance or specific functional surfaces is greatly limited. However, even though an aluminum alloy with high heat conduction potential and a well-colored substrate is obtained by a sophisticated composition design, the realization of its final properties is extremely dependent on a precise and adapted hot working process. The heat dissipation component usually has complex geometric shapes (such as thin walls, ribs and asymmetric structures), and in the key heat treatment processes of solid solution, aging and the like, the distribution of a temperature field and a stress field is extremely uneven, and the precise regulation and control of microstructure (such as precipitated phase size and grain size) are difficult to realize in the traditional processing mode based on a trial-and-error method or a fixed technological rule. This results in alloys of the same composition, which may exhibit substantial fluctuations in properties (e.g., strength, thermal conductivity) due to minor differences in component geometry or process parameters, even induce deformation, cracking, or affect the uniformity of subsequent anodization. Therefore, in the prior art, there is a remarkable contradiction that in order to obtain high heat conduction performance, it is desirable to control the content of alloy elements, especially silicon, at a low level, but in order to maintain or improve the mechanical properties and other process performances of the material, a certain amount of alloy elements is often required to be add