A heat treatment optimization method of Al-Si-Mg-Ti-Sr-Zr alloy based on machine learning regulation

Abstract

The application discloses an Al-Si-Mg-Ti-Sr-Zr alloy heat treatment optimization method based on machine learning regulation, and belongs to the technical field of aluminum alloy materials. The application is based on machine learning assisted component design, adopts vacuum medium frequency induction melting to prepare as-cast alloy, and obtains a compact-structure and uniform-component ingot through vacuum pumping, argon gas washing, grading current heating, melt shaking and mixing, and pouring cooling; then, T5 and T6 double-system heat treatment is carried out according to GB / T 1173-2013 specification, solid solution time, aging temperature and aging time are systematically optimized, and optimal process parameters are determined. After optimal process treatment, the alloy realizes complete spheroidization of eutectic Si, dispersion precipitation of (Al, Si)3(Zr, Ti) phase, and coherent strengthening of nanoscale Mg2Si phase, the maximum tensile strength can reach 341 MPa, the elongation is 5.5%, the comprehensive mechanical properties are significantly better than those of a commercial ZL101A alloy, the expensive Sc element is not used, the cost is low, the process is stable, the industrial production can be realized, and the application can be used in the field of lightweight structural parts such as automobile engine supports, gearbox housings and chassis structural parts.

Description

Technical Field

[0001] This invention relates to the field of aluminum alloy materials technology, specifically to an Al-Si-Mg-Ti-Sr-Zr alloy based on machine learning control, as well as a vacuum medium-frequency induction melting preparation method for this alloy, an optimized T5 / T6 heat treatment process, and its engineering application in the field of automotive structural parts. It belongs to the field of metal structural material design and preparation technology. Background Technology

[0002] Driven by the lightweight upgrade of high-end equipment, aluminum alloys, as lightweight metal structural materials with high specific strength, good formability, and high recyclability, have become core materials for achieving weight reduction, energy saving, and improved service performance in fields such as new energy vehicles, rail transit, and aerospace. Among them, Al-Si-Mg series cast aluminum alloys, with their excellent casting fluidity, weldability, corrosion resistance, and adjustable mechanical properties, are currently the most widely used aluminum alloy system in lightweight structural components, especially the ZL101A alloy, which has achieved large-scale application in key load-bearing components such as automotive engine brackets, gearbox housings, and chassis structural components.

[0003] With the continuous improvement of the driving range and safety performance requirements of new energy vehicles, and the stringent requirements of high-end equipment for long service life and adaptability to complex working conditions of structural components, the performance shortcomings of traditional Al-Si-Mg cast aluminum alloys are becoming increasingly prominent, and they can no longer meet the needs of high-end lightweight applications. First, there are inherent defects in the control of the as-cast microstructure. In the as-cast state of traditional alloys, eutectic Si mostly exists in the form of needles, plates, and short rods. Their sharp interfaces will severely cleave the α-Al matrix, which is prone to stress concentration during the stress process, becoming the core source of crack initiation and propagation, and significantly reducing the plasticity, toughness and fatigue performance of the alloy. Conventional modification treatment can only achieve partial refinement of eutectic Si, and cannot achieve complete globalization, so the modification effect is limited.

[0004] Secondly, existing toughening modification schemes suffer from a core contradiction between cost and performance. Current technologies often employ microalloying modifications using rare earth elements such as Sc, Er, and Yb to improve alloy strength and toughness. While this can improve mechanical properties to some extent, rare earth elements are scarce and expensive; the market price of metallic Sc ​​is hundreds of times higher than that of the Al matrix, directly leading to a significant increase in the cost of alloy raw materials. This makes it impossible to meet the cost control requirements of large-scale industrial production in civilian sectors such as automobiles, hindering its engineering application. Furthermore, existing microalloying systems are mostly single-element modifications, failing to achieve synergistic control of grain refinement, eutectic Si modification, and dispersion strengthening. This generally results in the industry pain point of "strength-plasticity imbalance," where strength is increased while plasticity is significantly lost, or high strength requirements are not achieved while ensuring plasticity.

[0005] Furthermore, the inadequacy of traditional preparation and heat treatment processes severely restricts the release of the alloy's performance potential. Traditional smelting often employs resistance furnaces or gas furnaces in an atmospheric environment, making the alloy melt prone to oxidation and gas absorption, resulting in casting defects such as oxide inclusions, porosity, and compositional segregation. This leads to large performance fluctuations between product batches and insufficient service stability. On the other hand, existing heat treatment processes often directly adopt the general parameters of commercial ZL101A, which have extremely poor compatibility with the modified alloy composition system. This can easily lead to insufficient solid solution, resulting in insufficient precipitation of strengthening phases, or excessive aging, causing coarsening of precipitates and performance degradation, thus failing to fully realize the alloy's performance potential.

[0006] In summary, the industry urgently needs to develop a new type of Al-Si-Mg cast aluminum alloy that does not rely on expensive rare earth elements, has controllable costs, combines high strength and high toughness, has a uniform and stable microstructure, and has a simple preparation process suitable for industrial mass production. At the same time, it needs to be equipped with a heat treatment optimization process that is precisely matched with the composition system to solve the core technical bottlenecks of traditional alloys, such as poor strength-ductility matching, high cost, insufficient microstructure stability, and poor mass production consistency, so as to meet the stringent requirements of lightweight structural components for high-end equipment. Summary of the Invention

[0007] The purpose of this invention is to provide a heat treatment optimization method for Al-Si-Mg-Ti-Sr-Zr alloys based on machine learning control. By synergistic microalloying of Ti, Sr, and Zr elements, combined with vacuum induction melting and T5 / T6 precise heat treatment, the method achieves eutectic Si spheroidization, grain refinement, and uniform precipitation of dispersed strengthening phases, significantly improving the alloy's tensile strength, yield strength, and elongation. The overall performance is superior to that of commercial ZL101A alloy, while the cost is controllable, the process is simple, and it is suitable for large-scale industrial production.

[0008] To achieve the above technical objectives, the present invention adopts the following technical solution:

[0009] A heat treatment optimization method for Al-Si-Mg-Ti-Sr-Zr alloys based on machine learning control includes the following steps:

[0010] S1. Database Construction: Collect Al-Si-Mg alloy data from literature and independent experiments, and after cleaning and standardization, construct a composition-property database containing 111 samples;

[0011] S2. Material Descriptor Generation: Using WenAlloys, Magpie, Deml, and Matminer feature extractors, the components are converted into 290 material descriptors containing physicochemical information;

[0012] S3. Key Descriptor Screening: Through a four-step method of feature importance ranking, Pearson correlation analysis, recursive elimination, and exhaustive screening, key descriptors affecting UTS and EL are obtained;

[0013] S4. Prediction Model Construction: Using key descriptors as input, the CatBoost algorithm is adopted, and after Bayesian optimization and ten-fold cross-validation, a high-precision prediction model for UTS and EL is established.

[0014] S5.SHAP Interpretability Analysis: The SHAP method is used to quantify the contribution of descriptors and reveal the regulatory mechanism of electronic structure and crystal structure descriptors on strong plasticity;

[0015] S6. Multi-objective composition optimization: Traverse the composition space and select the optimal alloy composition with the goal of maximizing UTS and EL simultaneously;

[0016] S7. Crush the high-purity Al, Al-25Si master alloy, Mg, Al-5Ti, Al-10Sr, and Al-10Zr master alloy into small pieces, wipe them clean with anhydrous ethanol, weigh them accurately according to their composition, and add an additional 0.05wt.% of Mg to compensate for volatile burn-off.

[0017] S8. Clean the furnace cavity of the vacuum medium-frequency induction melting furnace with anhydrous ethanol, uniformly spray the inner wall of the graphite mold with boron nitride coating, and load the weighed raw materials into the graphite crucible in sequence.

[0018] S9. Evacuate the furnace for 10 minutes, purge with argon gas, evacuate again for 10 minutes, and perform staged heating melting under argon protective atmosphere: 225A heating for 4 minutes → 245A heating for 3 minutes → 255A heating for 2 minutes while shaking the crucible → 270A heating for 1 minute → 275A heating for 1 minute → 245A heating for 1 minute while stopping shaking.

[0019] S10. Quickly pour the well-mixed molten metal into a graphite mold, evacuate again and cool to obtain the cast alloy;

[0020] S11. The as-cast alloy obtained in step S10 is heat-treated using either T5 or T6 processes, with a solution temperature of 535℃. After solution treatment, it is quenched in cold water at room temperature and then air-cooled after aging.

[0021] (1) When using T5 heat treatment: aging temperature 155℃, solution treatment time 4-12h, aging time 4-8h;

[0022] (2) When using T6 heat treatment: aging temperature 180℃, solution treatment time 8-10h, aging time 4-8h.

[0023] Furthermore, the purity of the raw materials used in step S7 is Al ≥ 99.999%, Mg ≥ 99.999%, and the purity of the intermediate alloy used is ≥ 99.99%; the dimensions of the as-cast alloy sample are 50π mm. 2 ×80mm.

[0024] Further, the as-cast alloy described in step S10, based on an alloy mass fraction of 100%, includes 6.5-7.5% Si, 0.45-0.60% Mg, 0.10-0.20% Ti, 0.01-0.03% Sr, 0.15-0.25% Zr, with the balance being Al and unavoidable impurities.

[0025] Furthermore, the cast alloy, by mass fraction of 100%, comprises 7.0% Si, 0.55% Mg, 0.15% Ti, 0.025% Sr, and 0.2% Zr, with the balance being Al and unavoidable impurities.

[0026] Furthermore, in step S11, the optimal process for T5 is solution treatment at 535℃ for 10 hours followed by aging at 155℃ for 6 hours, denoted as T5.10.6. The tensile strength of the as-cast alloy after heat treatment and air cooling is 341 MPa, the yield strength is 288 MPa, and the elongation is 5.5%.

[0027] Furthermore, in step S11, the optimal process for T6 is solution treatment at 535℃ for 8 hours followed by aging at 180℃ for 6 hours, denoted as T6.8.6. The tensile strength of the as-cast alloy after heat treatment and air cooling is 318 MPa, the yield strength is 284 MPa, and the elongation is 4.6%.

[0028] The present invention has the following beneficial effects:

[0029] 1. Compared to existing Al-Si-Mg alloy modification technologies that commonly employ expensive rare earth elements such as Sc to achieve strength and toughness, this invention completely abandons high-cost rare earth elements. Based on machine learning-assisted composition design, it innovatively uses three low-cost elements—Ti, Sr, and Zr—to construct a microalloying system. This achieves mechanical properties far exceeding those of commercial ZL101A alloys while significantly reducing raw material costs. In existing Sc-modified aluminum alloys, the Sc element addition is typically no less than 0.15wt%, which alone increases raw material costs by more than 80%, resulting in extremely poor economics for large-scale production. In contrast, the Ti, Sr, and Zr elements used in this invention are commonly used auxiliary materials in the industrial production of aluminum alloys, with low procurement costs. The overall raw material cost of the alloy is basically the same as that of commercial ZL101A, and the cost is reduced by more than 60% compared to rare earth-modified alloys. At the same time, it has strong process adaptability, requiring no large-scale modification of existing casting production lines, perfectly adapting to the large-scale industrial production needs in fields such as automotive structural parts, and completely solving the industry problem of the traditional high-strength and high-toughness aluminum alloys where "high performance and low cost are mutually exclusive."

[0030] 2. Compared with traditional Al-Si-Mg alloy modification technologies, which can only achieve single-performance improvement and generally suffer from an imbalance between strength and plasticity, this invention achieves multi-dimensional microstructure control through the synergistic effect of multiple elements such as Ti, Sr, and Zr. This control enables eutectic Si spheroidization, α-Al grain refinement, and uniform precipitation of nano-dispersed phases, simultaneously improving the alloy's strength and plasticity, thus solving the inherent problem of "increasing strength inevitably leading to decreased plasticity" in traditional alloys. The existing commercial ZL101A alloy in the T6 state has a tensile strength of only about 275 MPa and an elongation of about 3%, while the optimal T5 state alloy of this invention achieves a tensile strength of 341 MPa, a yield strength of 288 MPa, and an elongation of 5.5%. The tensile strength is increased by 24% and the elongation by 83% compared to ZL101A. Among them, Sr element realizes the complete globalization of eutectic Si, eliminates the cutting effect of brittle relative to the matrix, and greatly improves plasticity; Ti and Zr elements work together to refine grains and pin grain boundaries, forming a dispersed strengthening phase to hinder dislocation movement and significantly improve strength; Mg and Si form a nano-precipitated phase to achieve precipitation strengthening. The three elements work together to greatly improve strength while further optimizing the alloy's plasticity and toughness, achieving the optimal match between strength and plasticity.

[0031] 3. Compared with existing technologies that generally use commercial alloy general heat treatment processes, have poor matching with modified alloy compositions, and cannot fully realize the potential of material performance, this invention addresses the microstructure evolution law of the Ti-Sr-Zr microalloying system. Based on the national standard GB / T1173-2013, it systematically optimizes the solution treatment-aging process parameters and develops two sets of standardized optimal heat treatment processes, T5 and T6, adapted to different application scenarios, achieving directional and controllable regulation of alloy properties. Existing general heat treatment processes are prone to problems such as excessively high solution temperatures leading to overheating, insufficient solution time resulting in incomplete dissolution of strengthening phases, or improper aging parameters leading to coarsening of precipitates and performance degradation. Through extensive testing, this invention has determined that 535℃ is the optimal solution temperature, ensuring sufficient solid solution of the Mg2Si strengthening phase while avoiding over-burning and coarsening of the eutectic Si. Simultaneously, two optimal processes, T5.10.6 and T6.8.6, have been optimized. The T5 process is suitable for structural components requiring high toughness and fatigue resistance, while the T6 process is suitable for load-bearing components requiring high yield strength and high load-bearing capacity. The appropriate process can be flexibly selected according to different application scenarios, fully leveraging the alloy's performance potential and ensuring the stability and consistency of product performance.

[0032] 4. Compared to traditional atmospheric melting processes, which are prone to defects such as oxidation inclusions, porosity, and compositional segregation, leading to large performance fluctuations and insufficient service stability between product batches, this invention develops an integrated preparation process combining vacuum medium-frequency induction melting, argon protection, staged heating, and melt homogenization. This ensures high purity and compositional uniformity of the alloy melt from the source, significantly improving batch consistency in industrial mass production. In traditional atmospheric melting, reactive elements such as Al and Mg readily react with oxygen and water vapor to generate oxidation inclusions. Simultaneously, melt gas absorption causes porosity defects in castings, and compositional segregation results in performance differences of over 15% between different parts. This invention completely isolates the oxidizing environment during the melting process through dual-stage vacuuming and argon gas scrubbing protection, avoiding oxidation inclusions and porosity defects. The staged heating melting process achieves gradient melting of raw materials with different melting points, and the melt sloshing process ensures complete dissolution and uniform dispersion of high-melting-point Ti and Zr master alloys, completely solving the compositional segregation problem. Furthermore, the Mg element burn-off compensation design ensures precise control of the alloy composition. The resulting as-cast alloy has a dense structure and uniform composition, with no obvious casting defects. The performance fluctuation between product batches is controlled within 3%, demonstrating excellent mass production stability and service reliability. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of the heat treatment optimization method for Al-Si-Mg-Ti-Sr-Zr alloy based on machine learning control according to the present invention. Figure 1 (a) is a design framework diagram of Al-Si-Mg-Ti-Sr-Zr alloy based on machine learning, showing the complete R&D process of database construction, material descriptor generation, key descriptor screening, prediction model construction, composition optimization and experimental verification. Figure 1 (b) is a roadmap for alloy preparation and heat treatment technologies;

[0034] Figure 2 This is a flowchart of the material descriptor screening process for this invention; wherein, Figure 2 (a) Figure 2 (b) is a ranking chart of the importance of the top 15 features. Figure 2 (c) Figure 2 (d) is a correlation analysis plot of material descriptors. Figure 2 (e) Figure 2 (f) is a graph showing the material descriptor results after Pearson screening;

[0035] Figure 3 Regression plots showing the prediction performance of different screening algorithms and models; among them, Figure 3 (a) Figure 3 (b) is a graph showing the results of recursive feature elimination screening. Figure 3 (c) Figure 3 (d) is a graph showing the results of the exhaustive search method. Figure 3 (e) Figure 3 (f) are the R² and RMSE regression analysis plots of the CatBoost algorithm model predicting the performance of UTS and EL, respectively;

[0036] Figure 4 This is a SHAP overview diagram of the alloy property prediction model of the present invention; wherein, Figure 4 (a) is a graph showing the ranking of feature importance and contribution percentage of the UTS prediction model. Figure 4 (b) A chart showing the ranking of feature importance and contribution percentage of the EL prediction model;

[0037] Figure 5 This is a distribution diagram of the experimental performance of the initial data sample and the optimized alloy of this invention;

[0038] Figure 6 This is an engineering stress-strain curve diagram of the alloy of the present invention under warm water quenching and cold water quenching; wherein, Figure 6 (a) shows the curve of the sample quenched in warm water. Figure 6 (b) shows the curve of the cold water quenched sample;

[0039] Figure 7 The figures show the engineering stress-strain curves of the alloy after T5 heat treatment with different process parameters according to the present invention; wherein, Figure 7 (a) shows the sample curve for process T5.6.6. Figure 7 (b) shows the sample curve for process T5.8.6. Figure 7 (c) shows the sample curve for process T5.10.6. Figure 7 (d) is the sample curve for process T5.12.6;

[0040] Figure 8 These are backscattered imagery of Al-7Si-0.55Mg-0.15Ti-0.025Sr-0.2Zr alloys heat-treated with different process parameters (T5) according to this invention; wherein, Figure 8 (a) is a sample from process T5.6.6. Figure 8 (b) is the sample from process T5.8.6. Figure 8 (c) is a sample from process T5.10.6. Figure 8 (d) is the sample from process T5.12.6;

[0041] Figure 9 The figures show the engineering stress-strain curves of the alloy after T6 heat treatment with different process parameters according to the present invention; wherein, Figure 9 (a) is the sample curve for process T6.8.4. Figure 9 (b) shows the sample curve for process T6.8.6. Figure 9 (c) shows the sample curve for process T6.8.8. Figure 9 (d) is the sample curve for process T6.10.6;

[0042] Figure 10 These are backscattered imagery of Al-7Si-0.55Mg-0.15Ti-0.025Sr-0.2Zr alloys subjected to T6 heat treatment with different process parameters according to this invention; wherein, Figure 10 (a) is a sample from process T6.8.4. Figure 10 (b) is a sample from process T6.8.6. Figure 10 (c) is a sample from process T6.8.8. Figure 10 (d) is a sample from process T6.10.6;

[0043] Figure 11 The images show the XRD patterns of the alloy of this invention after T5 heat treatment and T6 heat treatment, respectively; wherein, Figure 11 (a) is the XRD pattern of the alloy in the T5 state. Figure 11 (b) is the XRD pattern of the T6 state alloy;

[0044] Figure 12 The images shown are TEM images and EDS energy dispersive spectroscopy (EDS) spectra of the T5.10.6 and T6.8.6 heat-treated alloys of this invention; wherein, Figure 12 (a) Figure 12 (c) Bright-field transmission electron microscopy images of the T5.10.6 process sample and the T6.8.6 process sample, respectively. Figure 12 (b) Figure 12 (d) are the selected area morphology and EDS surface scan analysis images of the T5.10.6 process sample and the T6.8.6 process sample, respectively;

[0045] Figure 13 These are HRTEM images and corresponding FFT images of the precipitated phase along the

[001] Al zone axis of the present invention; wherein, Figure 13 (a) Figure 13 (c) Selected area HRTEM images and corresponding precipitate phase HRTEM images of the alloys in heat-treated states T5.10.6 and T6.8.6, respectively; Figure 13 (b) Figure 13 (d) are the FFT spectra corresponding to the alloy precipitation in the T5.10.6 heat-treated state and the T6.8.6 heat-treated state, respectively;

[0046] Figure 14 The images shown are TEM images and EDS energy dispersive spectroscopy (EDS) analyses of large-size precipitate phases within the T5.10.6 and T6.8.6 heat-treated alloys of this invention; wherein, Figure 14 (a) Figure 14 (c) Bright-field transmission electron microscopy images of the T5.10.6 process sample and the T6.8.6 process sample, respectively. Figure 14 (b) Figure 14(d) are the selected area morphology, EDS point composition analysis diagrams and corresponding FFT images of the T5.10.6 process sample and the T6.8.6 process sample, respectively. Detailed Implementation

[0047] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0048] Unless otherwise defined, the technical terms used in this invention have their conventional meaning in the art; the experimental materials and equipment are all commercially available. The raw materials used are Al (99.999%), Al-25Si master alloy (99.99%), Mg (99.999%), Al-5Ti (99.99%), Al-10Sr (99.99%), and Al-10Zr master alloy (99.99%). The melting equipment is a vacuum medium-frequency induction melting furnace, and the tensile testing equipment is an LD26.105 universal testing machine. Microscopic characterization was performed using XRD, EPMA, and TEM equipment.

[0049] In an embodiment of the present invention, an Al-Si-Mg-Ti-Sr-Zr alloy, based on an alloy mass fraction of 100%, comprises 6.5-7.5% Si, 0.45-0.60% Mg, 0.10-0.20% Ti, 0.01-0.03% Sr, and 0.15-0.25% Zr, with the balance being Al and unavoidable impurities.

[0050] The preparation method of the Al-Si-Mg-Ti-Sr-Zr alloy includes the following steps:

[0051] S1. Crush high-purity Al, Al-25Si master alloy, Mg, Al-5Ti, Al-10Sr, and Al-10Zr master alloy into small pieces, wipe them clean with anhydrous ethanol, weigh them accurately according to their composition, and add an extra 0.05wt.% of Mg to compensate for volatile burn-off.

[0052] S2. Clean the furnace cavity of the vacuum medium-frequency induction melting furnace with anhydrous ethanol, uniformly spray the inner wall of the graphite mold with boron nitride coating, and load the weighed raw materials into the graphite crucible in sequence.

[0053] S3. Evacuate the furnace for 10 minutes, purge with argon gas, evacuate again for 10 minutes, and perform staged heating melting under argon protective atmosphere: 225A heating for 4 minutes → 245A heating for 3 minutes → 255A heating for 2 minutes while shaking the crucible → 270A heating for 1 minute → 275A heating for 1 minute → 245A heating for 1 minute while stopping shaking.

[0054] S4. Quickly pour the thoroughly mixed molten metal into a graphite mold, evacuate and cool it again to obtain the as-cast Al-7Si-0.55Mg-0.15Ti-0.025Sr-0.2Zr alloy.

[0055] The purity of the raw materials used was Al ≥ 99.999% and Mg ≥ 99.999%, and the purity of the master alloy used was ≥ 99.99%; the dimensions of the cast alloy sample were 50π mm. 2 ×80mm.

[0056] The heat treatment optimization method for the Al-Si-Mg-Ti-Sr-Zr alloy includes two heat treatment processes, T5 or T6, with a solution temperature of 535℃. After solution treatment, the alloy is quenched in cold water at room temperature and then air-cooled after aging.

[0057] (1) When using T5 heat treatment: aging temperature 155℃, solution treatment time 4-12h, aging time 4-8h;

[0058] (2) When using T6 heat treatment: aging temperature 180℃, solution treatment time 8-10h, aging time 4-8h.

[0059] Technical principle of the invention:

[0060] (I) The core role of each raw material in the alloy

[0061] This invention uses an Al-Si-Mg alloy as the basic system, and through micro-alloying modification with Ti, Sr, and Zr, each component is precisely matched and performs its specific function. The core functions are as follows:

[0062] 1. Aluminum (Al): As the alloy matrix, high-purity aluminum with a purity of ≥99.999% is used to provide the alloy with excellent plasticity, corrosion resistance and castability. At the same time, it provides a carrier for solid solution and precipitation of other alloying elements and is the basis for the comprehensive performance of the alloy.

[0063] 2. Silicon (Si): The mass fraction is controlled at 6.5-7.5%, which is close to the Al-Si eutectic composition. On the one hand, it can greatly improve the casting fluidity of the alloy melt and reduce the forming defects of complex structural parts; on the other hand, it can form Mg2Si nano-reinforcing phase with Mg element, which is the core source of alloy precipitation strengthening; the modified eutectic Si phase can also serve as a second phase to improve the wear resistance and high temperature stability of the alloy.

[0064] 3. Magnesium (Mg): The mass fraction is controlled at 0.45-0.60%. Its core role is to combine with Si to form Mg2Si nano-coherent precipitates, which are uniformly dispersed during the aging heat treatment process. By hindering dislocation movement, it achieves significant precipitation strengthening and is the core element for improving the strength and hardness of the alloy. At the same time, the trace amount of dissolved Mg can optimize the corrosion resistance of the matrix.

[0065] 4. Titanium (Ti): The mass fraction is controlled at 0.10-0.20%. As a highly efficient grain refiner, it preferentially forms Al3Ti intermetallic compounds during melt solidification. Its lattice constant is highly matched with the α-Al matrix, and it can serve as the core for non-spontaneous nucleation of α-Al grains, greatly improving the nucleation rate and significantly refining the α-Al matrix grains. Through grain refinement, it simultaneously improves the strength and plasticity of the alloy. At the same time, it can synergistically form a composite dispersed phase with Zr, further strengthening the grain boundaries.

[0066] 5. Strontium (Sr): The mass fraction is controlled at 0.01-0.03%. As a long-lasting eutectic Si modifier, it can inhibit the preferential growth of eutectic Si by adsorbing on the growth interface of eutectic Si, changing its growth morphology, and completely transforming the needle-like and plate-like eutectic Si in the cast state into uniform and fine spherical particles. This completely eliminates the cutting effect of brittle eutectic Si relative to the α-Al matrix, greatly reduces the risk of stress concentration, and significantly improves the plasticity, toughness and fatigue performance of the alloy. Moreover, Sr modification does not show a decay phenomenon, making it suitable for long-term industrial smelting production.

[0067] 6. Zirconium (Zr): The mass fraction is controlled at 0.15-0.25%. On the one hand, it can form the Al3Zr phase during solidification, which works synergistically with Al3Ti to further improve the grain refinement effect. At the same time, the Al3Zr phase can pin grain boundaries, inhibit grain coarsening during heat treatment and service, and improve the thermal stability of the alloy. On the other hand, Zr can combine with Ti and Si to form (Al,Si)3(Zr,Ti) nano-dispersed phase, which is uniformly distributed in the grain and at the grain boundaries, hindering dislocation slip and grain boundary migration, and significantly improving the room temperature and high temperature strength of the alloy without losing plasticity.

[0068] (II) Multi-element synergistic toughening mechanism

[0069] The core innovation of this invention lies in the synergistic effect of Ti, Sr, and Zr, combined with Mg-Si precipitation strengthening, achieving multi-dimensional synergistic regulation of "modification-refinement-dispersion strengthening," completely solving the pain point of the imbalance between strength and plasticity in traditional alloys. The synergistic mechanism is as follows:

[0070] 1. Pre-modification basis of Sr element: Sr element first realizes the complete globalization of eutectic Si, eliminating the most important brittle phase and crack source in the alloy, fundamentally solving the problem of plasticity loss caused by matrix splitting, reserving sufficient plasticity margin for improving the strength of the alloy, and avoiding the inherent defect of "strength improvement inevitably leads to plasticity reduction" in traditional modification.

[0071] 2. Synergistic Grain and Grain Boundary Regulation by Ti and Zr: Ti provides a large number of heterogeneous nucleation sites, achieving significant refinement of α-Al grains; the Al3Zr phase formed by Zr pins grain boundaries, inhibiting grain growth. Together, they achieve ultra-fine grains, enhancing both the strength and plasticity of the alloy through grain refinement strengthening. Simultaneously, the (Al,Si)3(Zr,Ti) nano-dispersed phase formed by the two elements is uniformly distributed within the grains, creating a multi-scale strengthening effect with the Mg2Si precipitate phase, further enhancing the alloy's strength. The dispersed phase at grain boundaries pins the grain boundaries, inhibiting grain boundary slip and improving the alloy's thermal stability and fatigue performance.

[0072] 3. Synergistic precipitation strengthening effect of multiple phases: The Mg2Si nano-coherent phase formed by Mg and Si is the core of the alloy strength improvement during aging; while the dispersed phase formed by Ti-Zr can serve as non-uniform nucleation sites for the Mg2Si phase, promoting the uniform and dispersed precipitation of the Mg2Si phase, avoiding the agglomeration and coarsening of the precipitated phase, and further enhancing the precipitation strengthening effect. At the same time, the dispersed phase can inhibit grain coarsening during aging, ensuring the microstructure stability of the alloy after heat treatment, and ultimately achieving the optimal match between strength and plasticity.

[0073] (III) Necessity and Importance of Selecting Preparation Process and Heat Treatment Parameters

[0074] 1. Core Design Logic of Melting and Preparation Process Parameters

[0075] The vacuum intermediate frequency induction melting process of this invention has each parameter designed specifically for the characteristics of the alloy system, which is the core to ensure accurate alloy composition, uniform microstructure, and absence of casting defects.

[0076] Raw material pretreatment and composition compensation design: High-purity raw materials are used to reduce the negative impact of impurity elements on alloy performance from the source; In view of the characteristics of Mg element, which has high vapor pressure and is easy to volatilize and burn off during the melting process, an additional 0.05wt.% of Mg is added to ensure that the final composition of the alloy falls accurately within the design range and to avoid insufficient strengthening effect due to loss of Mg element.

[0077] Vacuum and protective atmosphere process: The process of two-stage vacuuming (10 minutes each time) + argon gas washing can completely remove active gases such as oxygen and water vapor from the furnace cavity. Melting is carried out in an inert argon atmosphere, which completely avoids the oxidation of active elements such as Al and Mg. This eliminates casting defects such as oxide inclusions and porosity from the source and ensures the high purity of the alloy melt.

[0078] The graded heating process employs a gradient current for graded heating, the core of which is to adapt to the melting point differences of different raw materials, achieving uniform melting and composition homogenization. Heating at 225A for 4 minutes at a low temperature melts the Al matrix, forming a stable molten pool and avoiding localized overheating; heating at 245A for 3 minutes melts various intermediate alloys, achieving initial dispersion of alloying elements; heating at 255A for 2 minutes while shaking the crucible promotes complete dissolution of high-melting-point Zr and Ti intermediate alloys, achieving uniform dispersion of alloying elements through forced convection, completely resolving compositional segregation issues; short-term high-temperature heating at 270A and 275A for 1 minute each ensures complete melting of all refractory intermediate alloys; finally, heating at 245A for 1 minute stabilizes the melt temperature, eliminates bubbles, and avoids excessive Mg loss due to prolonged high temperatures.

[0079] Casting and cooling process: The inner wall of the graphite mold is sprayed with boron nitride coating to prevent the melt from sticking to the mold and to ensure the surface quality of the casting; after casting, vacuum cooling is performed again to further suppress gas absorption and oxidation during the solidification process of the casting and to ensure that the as-cast structure is dense and uniform.

[0080] 2. Core Design Logic of Heat Treatment Process Parameters

[0081] The T5 / T6 heat treatment process of this invention is precisely designed based on the microstructure evolution law of the Ti-Sr-Zr microalloying system, and is the core of fully leveraging the alloy's strength and toughness potential.

[0082] The necessity of a solution temperature of 535℃: The eutectic temperature of Al-Si-Mg alloys is 577℃. Excessively high solution temperatures can lead to over-melting of the eutectic Si and grain coarsening, resulting in a sharp decline in performance. Conversely, excessively low solution temperatures prevent the Mg2Si strengthening phase from fully dissolving into the α-Al matrix, leading to insufficient subsequent aging precipitation and limited strengthening effect. This invention selects 535℃ as the solution temperature, ensuring complete dissolution of the Mg2Si phase, providing sufficient solute atoms for aging precipitation, while completely avoiding the risk of over-melting. Simultaneously, it promotes the uniform distribution of Ti and Zr elements in the matrix, ensuring the stable precipitation of subsequent dispersed phases.

[0083] The necessity of quenching process: After solution treatment, room temperature cold water quenching can achieve ultra-fast cooling, completely retain the supersaturated solid solution in the solid solution state to room temperature, avoid the premature precipitation of Mg2Si phase during the cooling process, provide a basis for the uniform dispersion of nano-precipitates during the subsequent aging process, and is the key to ensuring precipitation strengthening effect.

[0084] T5 heat treatment process parameter design: The T5 process employs low-temperature aging at 155℃, primarily to adapt to applications requiring high toughness. Low-temperature aging allows the Mg2Si phase to precipitate as finer, more uniform nano-sized particles, achieving high strength while maximizing the preservation of the alloy's plasticity and toughness. Through system optimization, the optimal process was determined to be 535℃ solution treatment for 10 hours followed by aging at 155℃ for 6 hours. The 10-hour solution treatment ensures complete compositional homogeneity, while the 6-hour aging reaches the peak aging state, avoiding over-aging that could lead to coarsening of the precipitated phases, ultimately achieving an excellent balance between strength and plasticity.

[0085] T6 heat treatment process parameter design: The T6 process adopts 180℃ medium-temperature aging, the core of which is to adapt to the efficiency requirements of industrial mass production and the performance requirements of high load-bearing structural components. Medium-temperature aging can accelerate precipitation kinetics, significantly shorten the aging cycle, and improve production efficiency; at the same time, it can obtain higher yield strength, adapting to the use requirements of high load-bearing components. The optimal process is 535℃ solution treatment for 8 hours + 180℃ aging for 6 hours. The 8 hours of solution treatment can achieve sufficient solution of the Mg2Si phase, and the 6 hours of aging reaches the peak aging. The resulting alloy has both high yield strength and good plasticity, perfectly adapting to the needs of industrial mass production.

[0086] (iv) Unexpected technical effects achieved by the present invention

[0087] 1. Low cost achieves comprehensive performance far exceeding commercial alloys: This invention completely abandons expensive rare earth elements and adopts a low-cost Ti-Sr-Zr microalloying system. The raw material cost is basically the same as that of commercial ZL101A, but the optimal T5 state alloy has a tensile strength of 341MPa, which is 24% higher than ZL101A, and an elongation of 5.5%, which is 83% higher than ZL101A. It achieves a simultaneous and significant improvement in strength and plasticity, breaking through the industry's inherent perception that "low cost cannot achieve high strength and toughness".

[0088] 2. Achieved complete spheroidization of eutectic Si and multi-dimensional synergistic regulation of its structure: Traditional Sr modification can only refine eutectic Si, but cannot achieve complete spheroidization. However, this invention achieves 100% spheroidization of acicular eutectic Si through precise control of the amount of Sr added, combined with the synergistic effect of Ti-Zr. At the same time, it achieves the synergistic effect of grain refinement, dispersion strengthening, and precipitation strengthening, which completely solves the core pain points of matrix fragmentation and strength-plasticity imbalance in traditional Al-Si-Mg alloys.

[0089] 3. Excellent microstructure stability and mass production consistency: The alloy system of this invention, through the grain boundary pinning effect of the Ti-Zr composite phase, exhibits no significant grain coarsening during heat treatment and long-term service, resulting in excellent microstructure thermal stability. At the same time, the accompanying smelting process can completely eliminate defects such as oxide inclusions, porosity, and compositional segregation. The performance fluctuation between product batches is controlled within 3%, which is far superior to the fluctuation level of more than 15% in traditional processes. It has excellent industrial mass production stability and solves the industry problem of being unable to balance high performance in the laboratory with mass production stability.

[0090] 4. Full-process process adaptable to existing industrial production lines: The smelting and heat treatment process of this invention does not require large-scale modification of existing aluminum alloy casting production lines. It can be directly adapted to existing general equipment such as vacuum medium-frequency melting furnaces and heat treatment furnaces. The process parameters are clear and controllable, and the replicability is extremely high. It can directly replace the traditional ZL101A alloy for high-end automotive structural parts and has extremely high engineering application value.

[0091] To make the present invention more fully disclosed, more specific embodiments are described below.

[0092] Example 1

[0093] In this invention, a schematic diagram of the heat treatment optimization method for Al-Si-Mg-Ti-Sr-Zr alloy based on machine learning regulation is shown below. Figure 1 As shown, the design framework for Al-Si-Mg-Ti-Sr-Zr alloys based on machine learning control is as follows: Figure 1 As shown in (a).

[0094] (1) Database construction

[0095] The system collects composition-property data of as-cast Al-Si-Mg alloys from authoritative domestic and international academic literature and independent experiments. Data sources include CNKI, Web of Science, ScienceDirect, and independent experimental tests. The raw data was cleaned by removing duplicates, samples with unclear experimental conditions, and those exhibiting abnormal performance. Standardization was performed: the unit of composition was standardized to mass percentage (wt.%), and the standard for mechanical property testing was standardized to room temperature tensile testing. Finally, a database containing 111 valid samples was constructed, covering Al, Si, Mg, Ti, Sr, Zr, Zn, Cu, Mn, and Fe elements, along with their corresponding UTS and EL data.

[0096] (2) Material descriptor generation

[0097] Based on the Matminer library, four types of feature extractors—WenAlloys, Magpie, Deml, and Matminer—were used to convert alloy composition data into 290 material descriptors, including physicochemical features such as valence electron concentration, atomic size differences, electronegativity differences, mixing enthalpy, and mixing entropy, thus constructing a high-dimensional material descriptor database.

[0098] (3) Key material descriptor screening

[0099] A four-step screening strategy was adopted: Feature Importance Ranking: The CatBoost algorithm was used to calculate the contribution of descriptors and select the top 15 key descriptors for UTS and EL; Pearson Correlation Analysis: Strongly correlated redundant descriptors with |ρ|>0.95 were eliminated; Recursive Feature Elimination: Inefficient features were iteratively eliminated, retaining the optimal feature subset; Exhaustive Screening: The candidate subset was traversed to lock the globally optimal combination of key descriptors. Finally, 6 key material descriptors affecting UTS and 8 affecting EL were identified, eliminating redundant interference and improving model accuracy. Figure 2 and Figure 3 As shown.

[0100] (4) Construction of prediction model

[0101] Using key material descriptors as input and UTS and EL as outputs, a prediction model is constructed using the CatBoost algorithm, and hyperparameters are tuned through Bayesian optimization combined with ten-fold cross-validation. Figure 3 As shown in (e)(f), the optimized UTS model R 2 =0.96, RMSE=3.19, EL model R 2 =0.95, RMSE=0.71, the model has high accuracy and no overfitting, providing support for component optimization.

[0102] (5) SHAP interpretability analysis

[0103] The SHAP method is used to quantify the performance contribution of key descriptors, and the results are as follows: Figure 4 As shown, electronic structure descriptors primarily improve UTS, while crystal structure compatibility and valence electron mismatch-related descriptors primarily improve EL. This clarifies the intrinsic relationship between descriptors and performance, providing a physical basis for alloy design.

[0104] (6) Multi-objective component optimization

[0105] With Si = 7.0 wt.%, Mg = 0.55 wt.%, Zn, Cu, Mn, and Fe = 0, and Ti = 0~0.2 wt.%, Sr = 0~0.05 wt.%, and Zr = 0~0.3 wt.%, 1617 virtual composition points were generated by meshing. The composition was input into the prediction model, and with the goal of maximizing UTS and EL, the optimal composition was selected as follows: based on 100% alloy mass fraction, including Si 7.0%, Mg 0.55%, Ti 0.15%, Sr 0.025%, Zr 0.2%, with the balance being Al and unavoidable impurities. The optimal composition was calculated as Al-7Si-0.55Mg-0.15Ti-0.025Sr-0.2Zr. Figure 5 The experimental performance distribution of the alloy was designed to provide initial data samples and optimize its design.

[0106] Alloy preparation methods include:

[0107] S1. Crush high-purity Al, Al-25Si, Mg, Al-5Ti, Al-10Sr, and Al-10Zr into uniform small pieces. Wipe the surfaces with anhydrous ethanol to remove oxide scale, oil, and impurities. Weigh each raw material precisely according to the optimal composition. Add an additional 0.05 wt.% of Mg to compensate for volatile loss during the smelting process. The weighing accuracy is to three decimal places.

[0108] S2. Thoroughly clean the inner wall of the vacuum medium-frequency induction melting furnace with cotton moistened with anhydrous ethanol to ensure the furnace cavity is clean; uniformly spray a boron nitride coating onto the inner wall of the graphite mold to facilitate subsequent demolding; load the weighed raw materials into the graphite crucible in order of increasing melting point.

[0109] S3. Close the furnace door, first evacuate for 10 minutes to remove air and water vapor from the furnace, then introduce high-purity argon gas for purging, evacuate again for 10 minutes, and heat under argon protection throughout the process.

[0110] S4. Staged heating and melting: Heat with 225A current for 4 minutes, 245A current for 3 minutes, 255A current for 2 minutes and start shaking the crucible in the middle, heat with 270A current for 1 minute, 275A current for 1 minute, 245A current for 1 minute and stop shaking the crucible.

[0111] S5. Quickly pour the molten metal into the graphite mold. After pouring, evacuate again and cool. After cooling to room temperature, open the furnace door, remove the mold, and obtain the cast alloy sample.

[0112] Alloy preparation and heat treatment process as follows Figure 1 As shown in (b).

[0113] Example 2: Comparison Test of Quenching Methods

[0114] The as-cast alloy from Example 1 was solution-treated at 535℃ for 4 hours, followed by quenching in warm water at 60℃ and cold water at room temperature, respectively. Then, it underwent a uniform aging treatment at 155℃ for 6 hours. The alloy's engineering stress-strain curve is shown below. Figure 6 As shown in the figure. The results indicate that the elongation of the warm water-quenched sample is slightly higher, but the tensile strength is lower; the cold water-quenched sample has higher strength and better overall performance. Therefore, this invention determines that cold water quenching is used as the standard quenching method.

[0115] Example 3: Comparison of T5 series heat treatment processes

[0116] The as-cast alloy samples prepared in Example 1 were subjected to T5 heat treatment. The solution treatment temperature was fixed at 535℃, the aging temperature at 155℃, and the aging time at 6h. The solution treatment times were set at 4h, 6h, 8h, 10h, and 12h respectively. After solution treatment, the samples were quenched in cold water at room temperature and then air-cooled after aging to obtain samples T5.4.6, T5.6.6, T5.8.6, T5.10.6, and T5.12.6.

[0117] Room temperature tensile test results are as follows Figure 7 As shown: T5.4.6 alloy UTS=253MPa, YS=218MPa, EL=5.1%; T5.6.6 alloy UTS=256MPa, YS=221MPa, EL=6.2%; T5.8.6 alloy UTS=274MPa, YS=236MPa, EL=7.0%; T5.10.6 alloy UTS=341MPa, YS=288MPa, EL=5.5%; T5.12.6 alloy UTS=285MPa, YS=239MPa, EL=4.5%.

[0118] EPMA characterization results are as follows Figure 8 As shown: during the solution treatment period of 4-6 hours, the eutectic Si only underwent preliminary spheroidization, and needle-like and short rod-like structures remained; during the solution treatment period of 8 hours, the spheroidization degree of the eutectic Si was significantly improved; during the solution treatment period of 10 hours, the eutectic Si was completely spheroidized and distributed as dispersed particles, and the (Al,Si)3(Zr,Ti) phase was uniformly precipitated without agglomeration; during the solution treatment period of 12 hours, the eutectic Si and the strengthening phase were significantly coarsened, micropores appeared in the matrix, and the structure deteriorated.

[0119] Example 4: Comparison of T6 series heat treatment processes

[0120] The as-cast alloy samples prepared in Example 1 were subjected to T6 heat treatment. The solution treatment temperature was fixed at 535℃, the aging temperature at 180℃, and the aging time was fixed at 4h, 6h, and 8h. The solution treatment time was set at 8h and 10h respectively. After solution treatment, the samples were quenched in cold water and then air-cooled after aging to obtain samples T6.8.4, T6.8.6, T6.8.8, and T6.10.6.

[0121] Room temperature tensile test results are as follows Figure 9 As shown: T6.8.4 alloy UTS=314MPa, YS=280MPa, EL=3.5%; T6.8.6 alloy UTS=318MPa, YS=284MPa, EL=4.6%; T6.8.8 alloy UTS=297MPa, YS=262MPa, EL=5.9%; T6.10.6 alloy UTS=252MPa, YS=216MPa, EL=4.6%.

[0122] EPMA characterization results are as follows Figure 10 As shown: In alloy T6.8.6, the eutectic Si is fully spheroidized, and the (Al,Si)3(Zr,Ti) phase is fine and dispersed, resulting in the best microstructure uniformity; In alloy T6.10.6, the strengthening phase and eutectic Si are significantly coarsened, a large number of pores appear in the matrix, the risk of stress concentration is increased, and the mechanical properties are reduced.

[0123] Example 5 Microstructure Characterization

[0124] The T5.10.6 sample from Example 2 and the T6.8.6 sample from Example 3 were characterized by XRD, EPMA, and TEM. The XRD results are as follows: Figure 11 As shown, the alloys under both processes consist of an α-Al matrix phase and a single-element Si phase, with no harmful coarse phases. The nano-reinforcing phase was not detected due to its extremely fine size.

[0125] EPMA results as follows Figure 8 and Figure 10 As shown, under the optimal process, the eutectic Si is completely transformed from needle-like to uniform spherical particles, and the (Al,Si)3(Zr,Ti) phase is uniformly distributed in short rod shape without continuous network or agglomeration, effectively eliminating the cutting effect of brittle relative to the matrix.

[0126] TEM results as follows Figure 12 and Figure 13 As shown, by Figure 12 It can be seen that the precipitated phases in both structures are mainly nano-sized particles, with the main precipitate phases being uniformly dispersed and primarily formed by the segregation of Mg and Si elements. Figure 13 Crystal plane determination of the FFT diffraction pattern spots confirmed that the precipitate phase was a monoclinic β″-Mg5Si6 phase. Figure 13 It can be seen that both optimal process samples precipitate 2-5 nm nanoscale phases, maintaining a completely coherent relationship with the α-Al matrix, without interfacial mismatch or dislocation defects. Figure 14 (a) Figure 14 (c) As can be seen, both alloys contain precipitates with sizes up to the micrometer scale. Combining the EDS point scan analysis of this phase at points 1 and 2, and... Figure 14 (b) Figure 14(d) Crystal plane calibration of FFT diffraction spots determined that the phase is (Al,Si)3(Zr,Ti) phase. The synergistic effect of the nano phase and the micro phase achieves significant precipitation strengthening and fine grain strengthening.

[0127] Example 6: Performance Comparison with ZL101A Alloy

[0128] The performance of the optimal process alloy of this invention was compared with that of the commercial ZL101A alloy: ZL101A alloy UTS=265-295MPa, EL=3-4%; the T5.10.6 alloy of this invention has UTS increased to 341MPa and EL increased to 5.5%; the T6.8.6 alloy has UTS increased to 318MPa and EL increased to 4.6%. The strength and plasticity of the two optimal process alloys are superior to ZL101A alloy in all aspects, and they can directly replace it for key structural components such as automotive engine brackets and gearbox housings.

[0129] Summarize:

[0130] This invention provides a heat treatment optimization method for Al-Si-Mg-Ti-Sr-Zr alloys based on machine learning. Through machine learning-aided composition design, Ti / Sr / Zr multi-element microalloying, and precise T5 / T6 heat treatment, it achieves eutectic Si spheroidization, grain refinement, and nano-phase dispersion strengthening, effectively solving the problems of poor strength-ductility matching, high cost, and insufficient microstructural stability in traditional Al-Si-Mg alloys. The resulting alloy exhibits significantly superior mechanical properties compared to commercially available ZL101A, and the process is simple, low-cost, and suitable for industrial production, making it applicable to lightweight structural components in the automotive and machinery industries, and possessing significant engineering application value.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A heat treatment optimization method for Al-Si-Mg-Ti-Sr-Zr alloys based on machine learning regulation, characterized in that: Includes the following steps: S1. Database Construction: Collect Al-Si-Mg alloy data from literature and independent experiments, and after cleaning and standardization, construct a composition-property database containing 111 samples; S2. Material Descriptor Generation: Using WenAlloys, Magpie, Deml, and Matminer feature extractors, the components are converted into 290 material descriptors containing physicochemical information; S3. Key Descriptor Screening: Through a four-step method of feature importance ranking, Pearson correlation analysis, recursive elimination, and exhaustive screening, key descriptors affecting UTS and EL are obtained; S4. Prediction Model Construction: Using key descriptors as input, the CatBoost algorithm is adopted, and after Bayesian optimization and ten-fold cross-validation, a high-precision prediction model for UTS and EL is established. S5.SHAP Interpretability Analysis: The SHAP method is used to quantify the contribution of descriptors and reveal the regulatory mechanism of electronic structure and crystal structure descriptors on strong plasticity; S6. Multi-objective composition optimization: Traverse the composition space and select the optimal alloy composition with the goal of maximizing UTS and EL simultaneously; S7. Crush the high-purity Al, Al-25Si master alloy, Mg, Al-5Ti, Al-10Sr and Al-10Zr master alloy into small pieces, wipe them clean with anhydrous ethanol, weigh them precisely according to the optimal alloy composition, and add an additional 0.05wt.% of Mg to compensate for volatile burn-off. S8. Clean the furnace cavity of the vacuum medium-frequency induction melting furnace with anhydrous ethanol, uniformly spray the inner wall of the graphite mold with boron nitride coating, and load the weighed raw materials into the graphite crucible in sequence. S9. Evacuate the furnace for 10 minutes, purge with argon gas, evacuate again for 10 minutes, and perform staged heating melting under argon protective atmosphere: 225A heating for 4 minutes → 245A heating for 3 minutes → 255A heating for 2 minutes while shaking the crucible → 270A heating for 1 minute → 275A heating for 1 minute → 245A heating for 1 minute while stopping shaking. S10. The thoroughly mixed molten metal is quickly poured into a graphite mold, vacuumed again and cooled to obtain a cast alloy. The cast alloy, based on an alloy mass fraction of 100%, includes Si 6.5-7.5%, Mg 0.45-0.60%, Ti 0.10-0.20%, Sr 0.01-0.03%, Zr 0.15-0.25%, with the balance being Al and unavoidable impurities. S11. The as-cast alloy obtained in step S10 is heat-treated using either T5 or T6 processes, with a solution temperature of 535℃. After solution treatment, it is quenched in cold water at room temperature and then air-cooled after aging. (1) When using T5 heat treatment: aging temperature 155℃, solution treatment time 4-12h, aging time 4-8h; (2) When using T6 heat treatment: aging temperature 180℃, solution treatment time 8-10h, aging time 4-8h.

2. The heat treatment optimization method for Al-Si-Mg-Ti-Sr-Zr alloy based on machine learning control according to claim 1, characterized in that, The purity of the raw materials used in step S7 is Al≥99.999%, Mg≥99.999%, and the purity of the intermediate alloy used is≥99.99%; the dimensions of the as-cast alloy sample are 50π mm. 2 ×80mm.

3. The heat treatment optimization method for Al-Si-Mg-Ti-Sr-Zr alloy based on machine learning control according to claim 1, characterized in that, The cast alloy, by mass fraction of 100%, comprises 7.0% Si, 0.55% Mg, 0.15% Ti, 0.025% Sr, and 0.2% Zr, with the balance being Al and unavoidable impurities.

4. The heat treatment optimization method for Al-Si-Mg-Ti-Sr-Zr alloy based on machine learning control according to claim 1, characterized in that, In step S11, the optimal process for T5 is solution treatment at 535℃ for 10 hours followed by aging at 155℃ for 6 hours, denoted as T5.10.

6. The tensile strength of the cast alloy after heat treatment and air cooling is 341 MPa, the yield strength is 288 MPa, and the elongation is 5.5%.

5. The heat treatment optimization method for Al-Si-Mg-Ti-Sr-Zr alloy based on machine learning control according to claim 1, characterized in that, In step S11, the optimal process for T6 is solution treatment at 535℃ for 8 hours followed by aging at 180℃ for 6 hours, denoted as T6.8.

6. The tensile strength, yield strength, and elongation of the cast alloy after heat treatment and air cooling are 318 MPa, 284 MPa, and 4.6%, respectively.

Images