AI方法在新钢种研发中的应用

Application of AI methods to the research and development of new steel grades

摘要

目前，我国钢铁企业主要采用传统的“试错法”开发新钢种. 这种传统的“研仿”过程，需花费大量的时间和研发经费. 为了提高研发效率，利用AI (Artificial intelligence) 方法实现新产品研发和现有产品的工艺参数优化，提高产品质量可靠性，是钢铁企业亟待解决的关键技术问题. 本文提出了基于AI方法的新产品研发模型，采用机器学习方法建立成分、工艺与性能间的预测模型，并通过核主成分分析算法将成分、工艺参数降维为二维主向量，采用高斯混合模型建立材料指纹图，然后利用AI的生成模型探索潜在的成分、工艺参数空间，替代传统的“试错法”，以提升新产品研发效率. 以深冲钢研发作为工业应用实例，采用AI方法实现了新钢种的快捷研发，通过“数字化试错法”可大幅提升研发效率、降低研发成本.

Abstract

Currently, machine learning and artificial intelligence (AI) methods are predominantly applied to the research and development of high-entropy alloys. Although high-entropy alloys exhibit superior properties and significant industrial potential, their manufacturing costs remain prohibitive, and established industrial applications are currently limited. Conversely, the vast majority of steel products are low-entropy alloys primarily composed of Fe, C, and trace amounts of Mn, Cr, Ni, Nb, and other elements. Presently, Chinese iron and steel enterprises rely largely on traditional “trial and error” methods to develop new steel grades. This traditional “research and imitation” process requires time and high research and development (R&D) expenditure. To enhance R&D efficiency, the utilization of AI methods to facilitate new product development, optimize existing process parameters of existing products, and improve product quality is a critical technological requirement for the industry. This paper proposes a novel product development model based on AI methods, utilizing IF steel R&D as an industrial case study. Given the high-dimensionality, strong coupling, and nonlinearity characteristic of industrial production data, eight machine learning models were evaluated to predict the mechanical properties of IF steel. To ensure model generalization, the data was regularized, the super-parameters were optimized, and the training set underwent five-fold cross-validation. The prediction accuracy and applicable scenarios for each model are subsequently discussed. Among the eight models for predicting the mechanical properties of interstitial-free (IF) steel, the random forest (RF) and deep neural network (DNN) models demonstrated superior accuracy, with R² > 0.97. Furthermore, the kernel principal component analysis (KPCA) model was employed to reduce the dimensionality of high-dimensional components and process parameter features into two-dimensional principal component vectors. A material fingerprint was then established using the Gaussian mixture module (GMM) model. This fingerprinting technique facilitated the visualization of high-dimensional feature data, enabling the analysis and observation of data distributions and the identification of potential mining intervals and the spatial location of the generated data. These potential spaces represent optimal search intervals for the discovery of novel materials. Finally, a generative model based on Wasserstein auto-encoders (WAE) was proposed to explore the potential composition and process parameter spaces. The WAE model can generate thousands of potentially valuable samples from which specific candidates can be selected for industrial testing. This digital “trial and error” approach serves as a robust alternative to traditional methods, accelerating product development while significantly reducing R&D costs and durations.

关键词

人工智能 / 机器学习 / 生成模型 / 深冲钢研发 / 工艺设计

Key words

artificial intelligence / machine learning / generative model / research and development of interstitial-free (IF) steel / process design

参考文献

[1] Tian S H, Jiang X, Wang W R, et al. Steel design based on a large language model[J]. Acta Mater, 2025, 285: 120663

[2] Xian Y H, Dang P F, Tian Y, et al. Compositional design of multicomponent alloys using reinforcement learning[J]. Acta Mater, 2024, 274: 120017

[3] Lin J, Zhao D F, Lu S P, et al. Conversational large—language—model artificial intelligence agent for accelerated synthesis of metal—organic frameworks catalysts in olefin hydrogenation[J]. ACS Nano, 2025, 19(26): 23840

[4] Wang H C, Fu T F, Du Y Q, et al. Scientific discovery in the age of artificial intelligence[J]. Nature, 2023, 620(7972): 47

[5] Jiang X, Xue D Z, Bai Y, et al. AI4Materials: Transforming the landscape of materials science and engineering[J]. Rev Mater Res, 2025, 1(1): 100010

[6] Rao Z Y, Tung P Y, Xie R W, et al. Machine learning—enabled high—entropy alloy discovery[J]. Science, 2022, 378(6615): 78

[7] He J Y, Li Z Z, Lin J C, et al. Machine learning—assisted design of refractory high—entropy alloys with targeted yield strength and fracture strain[J]. Mater Des, 2024, 246: 113326

[8] Sohail Y, Zhang C L, Xue D Z, et al. Machine—learning design of ductile FeNiCoAlTa alloys with high strength[J]. Nature, 2025, 643(8070): 119

[9] Wen C, Zhang Y, Wang C X, et al. Machine—learning—assisted compositional design of refractory high—entropy alloys with optimal strength and ductility[J]. Engineering, 2025, 46: 214

[10] Xu G, Niu G, Wang Y W, et al. Machine learning—based research of new refractory high—entropy alloys using guided multiobjectives search strategy[J]. Mater Genome Eng Adv, 2025, 3(4): e70030

[11] Xie J X. Prospects of materials genome engineering frontiers[J]. Mater Genome Eng Adv, 2023, 1(2): e17

[12] Xie J X, Su Y J, Zhang D W, et al. A vision of materials genome engineering in China[J]. Engineering, 2022, 10: 10

[13] Xu P C, Ji X B, Li M J, et al. Small data machine learning in materials science[J]. npj Comput Mater, 2023, 9: 42

[14] Jain A. Machine learning in materials research: Developments over the last decade and challenges for the future[J]. Curr Opin Solid State Mater Sci, 2024, 33: 101189

[15] Xu G, Xue Y, Geng X X, et al. A machine learning—based crystal graph network and its application in development of functional materials[J]. Mater Genome Eng Adv, 2024, 2(3): e38

[16] Qian C, Kaminer I, Chen H S. A guidance to intelligent metamaterials and metamaterials intelligence[J]. Nat Commun, 2025, 16: 1154

[17] Senkov O N, Wilks G B, Scott J M, et al. Mechanical properties of Nb₂₅Mo₂₅Ta₂₅W₂₅ and V₂₀Nb₂₀Mo₂₀Ta₂₀W₂₀ refractory high entropy alloys[J]. Intermetallics, 2011, 19(5): 698

[18] Liu C X, Wang Y C, Zhang Y T, et al. Deformation mechanisms of additively manufactured TiNbTaZrMo refractory high—entropy alloy: The role of cellular structure[J]. Int J Plast, 2024, 173: 103884

[19] Huang E W, Lee W J, Singh S S, et al. Machine—learning and high—throughput studies for high—entropy materials[J]. Mater Sci Eng R Rep, 2022, 147: 100645

[20] Abramson J, Adler J, Dunger J, et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3[J]. Nature, 2024, 630(8016): 493

[21] He J Y, Li Z Z, Zhao P L, et al. Machine learning—assisted design of high—entropy alloys with superior mechanical properties[J]. J Mater Res Technol , 2024, 33: 260

[22] Cai J M, Han M X, Yan X R, et al. A process—synergistic active learning framework for high—strength Al—Si alloys design[J]. npj Comput Mater, 2025, 11: 228

[23] Serefoglu Cabuk K, Cengiz S K, Guler M G, et al. Chasing the objective upper eyelid symmetry formula; R², RMSE, POC, MAE, and MSE[J]. Int Ophthalmol, 2024, 44: 303

[24] Pei X L, Pei J Q, Hou H, et al. Optimizing casting process using a combination of small data machine learning and phase—field simulations[J]. npj Comput Mater, 2025, 11: 27

[25] Liu S S, Bocklund B, Diffenderfer J, et al. A comparative study of predicting high entropy alloy phase fractions with traditional machine learning and deep neural networks[J]. npj Comput Mater, 2024, 10: 172

[26] Murakami Y, Furushima R, Shiga K, et al. Mechanical property prediction of aluminium alloys with varied silicon content using deep learning[J]. Acta Mater, 2025, 286: 120683

[27] van Duijvenbode J R, Cloete L M, Shishvan M S, et al. Material fingerprinting as a tool to investigate between and within material type variability with a focus on material hardness[J]. Miner Eng, 2022, 189: 107885

[28] Tolstikhin I, Bousquet O, Gelly S, et al. Wasserstein auto—encoders[PP/OL]. arXiv (2017—11—05) [2025—11—19]. https://arxiv.org/abs/1711.01558

基金资助

国家自然科学基金资助项目(52371002)