基于深度学习的亚稳态高熵合金高应变率冲击响应预测

刘传志; 安稳; 熊启林

doi:10.11883/bzycj-2025-0259

基于深度学习的亚稳态高熵合金高应变率冲击响应预测

doi: 10.11883/bzycj-2025-0259

刘传志^{1, 2,},
安稳^{1, 2},
熊启林^{1, 2, ,}

1.
华中科技大学航空航天学院，武汉 430074
2.
华中科技大学工程结构分析与安全评定湖北省重点实验室，武汉 430074

基金项目: 国家自然科学基金（12522216）；冲击波物理与爆轰物理全国重点实验室基金（2023JCJQLB05403）

详细信息

作者简介:
刘传志（1997－　），男，博士研究生，137668501@qq.com

通讯作者:
熊启林（1987－　），男，教授，博士生导师，xiongql@hust.edu.cn

中图分类号: O347.3
计量
- 文章访问数: 228
- HTML全文浏览量: 26
- PDF下载量: 37
- 被引次数: 0
出版历程
- 收稿日期: 2025-08-11
- 修回日期: 2025-10-24
- 网络出版日期: 2025-11-04

Deep learning-based prediction of high-strain-rate shock response in metastable high-entropy alloys

LIU Chuanzhi^{1, 2
,},
AN Wen^{1, 2},
XIONG Qilin^{1, 2
, ,}

1.
Department of Mechanics, School of Aerospace Engineering, Huazhong University of Science & Technology, 1037 Luoyu Road, Wuhan 430074, China
2.
Hubei Key Laboratory of Engineering Structural Analysis and Safety Assessment, Huazhong University of Science & Technology Luoyu Road 1037, Wuhan 430074, China

摘要

摘要: 亚稳态高熵合金因其在高应变率下优异的力学性能而受到广泛关注；然而，由于对微观结构与冲击响应关系的认识不足，限制了其在高应变率下的工程应用。本研究采用一种结合晶体塑性有限元方法和卷积神经网络的深度学习框架，阐明了微观结构与冲击响应之间的关系。基于晶体塑性模拟收集数据集，该数据集包含高应变率下亚稳态高熵合金在拉伸、压缩及剪切载荷条件下不同织构的完整应力应变响应和相变体积分数的演变。构建了一个双分支卷积神经网络模型，输入为织构和载荷条件。该模型的两个分支用于预测不同的输出：即应力应变曲线与马氏体体积分数的演变。基于收集的数据集对卷积神经网络模型进行训练。结果表明，该模型能够准确预测高应变率条件下亚稳态高熵合金的冲击响应。该研究进一步证明了深度学习框架在保证预测精度的同时，相比晶体塑性有限元模拟具有显著的计算效率优势，为高效评估高应变率下亚稳态高熵合金的力学行为提供了一种新思路。
- 深度学习 /
- 冲击响应 /
- 晶体塑性 /
- 亚稳态高熵合金
Abstract: Metastable high-entropy alloys (HEA) have attracted considerable attention due to their exceptional mechanical properties at high strain rates. However, their engineering applications under high strain rates are limited, which stems from an inadequate understanding of the relationship between microstructure and impact response. An end-to-end deep learning framework has been implemented, combining the crystal plasticity finite element (CPFE) method with a convolutional neural network (CNN) to elucidate the mapping between microstructure and shock response. A crystal plasticity constitutive model, which couples dislocation slip and martensitic transformation mechanisms, has been developed and validated against experimental results, confirming the model's effectiveness. Based on this constitutive model, a dataset for training the deep learning model is generated, including the complete stress-strain response and martensite volume fraction evolution of metastable HEA with corresponding textures and loading conditions at high strain rates. The two-branch CNN model is used to extract microstructural features. Its input is microstructural information in image format and loading conditions, and its output consists of two branches corresponding to stress-strain curves and the evolution of martensite volume fraction. The collected dataset was used to train the CNN model. The results show that the model can accurately predict the shock response of metastable HEA under high strain rate conditions. This study demonstrates that the deep learning framework, while maintaining predictive accuracy, offers a significant computational efficiency advantage over CPFE simulations. It provides a novel approach for efficiently assessing the mechanical behavior of metastable high-entropy alloys under high strain rates.
- deep learning /
- high strain rate /
- crystal plasticity /
- metastable high entropy alloy

HTML全文

图 1 基于卷积神经网络的冲击响应预测工作流程

Figure 1. Schematic of shock response prediction based on convolutional neural networks

下载: 全尺寸图片幻灯片

图 2 卷积与池化示意图

Figure 2. Schematic of Convolution and pooling

下载: 全尺寸图片幻灯片

图 3 随机取向多晶模型的RVE

Figure 3. RVE of the randomly oriented polycrystal model

下载: 全尺寸图片幻灯片

图 4 晶体塑性模拟与实验的比较

Figure 4. Comparison of crystal plasticity simulation and experiment

下载: 全尺寸图片幻灯片

图 5 数据集中代表性样本

Figure 5. Representative samples in the dataset

下载: 全尺寸图片幻灯片

图 6 训练集、验证集和测试集中的数据分布

Figure 6. Data distribution in the training, validation, and test sets

下载: 全尺寸图片幻灯片

图 7 具有双分支回归器的卷积神经网络模型的详细架构

Figure 7. Convolutional neural network model architecture with a two-branch regressor

下载: 全尺寸图片幻灯片

图 8 训练集与测试集上损失函数与误差随训练过程的演化

Figure 8. Evolution of loss function and error on training and test sets during training

下载: 全尺寸图片幻灯片

图 9 测试集上的预测结果

Figure 9. Prediction results on the test set

下载: 全尺寸图片幻灯片

图 10 测试集上预测值与真实值的对比

Figure 10. Comparison of predicted and true values on the test set

下载: 全尺寸图片幻灯片

图 11 不同学习率的训练过程

Figure 11. Training process with different learning rates

下载: 全尺寸图片幻灯片

图 12 不同优化器的训练过程

Figure 12. Training process of different optimizers

下载: 全尺寸图片幻灯片

图 13 晶体塑性模拟与卷积神经网络计算效率比较

Figure 13. Comparison of computational efficiency between crystal plasticity simulation and convolutional neural network

下载: 全尺寸图片幻灯片

图 14 变加载条件下晶体塑性模拟与卷积神经网络结果对比（实线：晶体塑性有限元模拟；虚线：卷积神经网络）

Figure 14. Comparison of crystal plasticity simulation and convolutional neural network results under variable loading conditions (solid line: crystal plasticity finite element simulation; dashed line: convolutional neural network)

下载: 全尺寸图片幻灯片

表 1 晶体塑性模型本构模型参数

Table 1. Constitutive model parameters of crystal plasticity model

参数	名称	取值	来源
ρ/(Kg·m⁻³)	密度	7 655	[34]
C_p/(J·K⁻¹·Kg⁻¹)	比热容	430	[34]
C₁₁/GPa	弹性参数	174.2	[35]
C₁₂/GPa	弹性参数	97.9	[35]
C₄₄/GPa	弹性参数	139.7	[35]
α_SR	短程阻力系数	0.29	本研究
α_LR	长程阻力系数	0.7	本研究
ξ_αβ	滑移与相变的相互作用系数	0（共面）	本研究
ξ_αβ	滑移与相变的相互作用系数	0.6（非共面）	本研究
α_mult	位错增值系数	0.001 7	本研究
α_anni	位错湮灭系数	0.24	本研究
f_cr/MPa	临界相变阻力	180	本研究

下载: 导出CSV

参考文献(35)

[1]	GEORGE E P, CURTIN W A, TASAN C C. High entropy alloys: a focused review of mechanical properties and deformation mechanisms [J]. Acta Materialia, 2020, 188: 435–474. DOI: 10.1016/j.actamat.2019.12.015.
[2]	HE Z F, JIA N, WANG H W, et al. The effect of strain rate on mechanical properties and microstructure of a metastable FeMnCoCr high entropy alloy [J]. Materials Science and Engineering: A, 2020, 776: 138982. DOI: 10.1016/j.msea.2020.138982.
[3]	ZHANG N B, ZHANG C X, LI B, et al. Impact response of metastable dual-phase high-entropy alloy Cr₁₀Mn₃₀Fe₅₀Co₁₀ [J]. Journal of Alloys and Compounds, 2023, 965: 171341. DOI: 10.1016/j.jallcom.2023.171341.
[4]	XU J, LIANG L, TONG W, et al. Role of strain rate in phase stability and deformation mechanism of non-equiatomic Fe_38-xMn₃₀Co₁₅Cr₁₅Ni₂Gd_x high-entropy alloy [J]. Materials Characterization, 2022, 194: 112356. DOI: 10.1016/j.matchar.2022.112356.
[5]	WANG P, BU Y Q, LIU J B, et al. Atomic deformation mechanism and interface toughening in metastable high entropy alloy [J]. Materials Today, 2020, 37: 64–73. DOI: 10.1016/j.mattod.2020.02.017.
[6]	WANG K Y, CHENG Z J, LIU C Y, et al. Deformation behavior and strengthening mechanisms of high-entropy alloys under high strain rate across wide temperature ranges [J]. International Journal of Plasticity, 2025, 189: 104321. DOI: 10.1016/j.ijplas.2025.104321.
[7]	WANG X, DE VECCHIS R R, LI C Y, et al. Design metastability in high-entropy alloys by tailoring unstable fault energies [J]. Science Advances, 2022, 8(36): eabo7333. DOI: 10.1126/sciadv.abo7333.
[8]	WERNER K V, NAEEM M, NIESSEN F, et al. Experimental and computational assessment of the temperature dependency of the stacking fault energy in face-centered cubic high-entropy alloys [J]. Acta Materialia, 2024, 278: 120271. DOI: 10.1016/j.actamat.2024.120271.
[9]	SONG H, KIM D G, KIM D W, et al. Effects of strain rate on room- and cryogenic-temperature compressive properties in metastable V10Cr10Fe45Co35 high-entropy alloy [J]. Scientific Reports, 2019, 9(1): 6163. DOI: 10.1038/s41598-019-42704-x.
[10]	YANG J P, AN W, LIU C Z, et al. Strength-plasticity synergy of metastable Fe₄₀Mn₄₀Co₁₀Cr₁₀ high entropy alloy at high strain rate and cryogenic temperature [J]. Materials Science and Engineering: A, 2025, 941: 148635. DOI: 10.1016/j.msea.2025.148635.
[11]	XIE H, MA Z, ZHANG W, et al. Phase transition in shock compressed high-entropy alloy FeNiCrCoCu [J]. International Journal of Mechanical Sciences, 2023, 238: 107855. DOI: 10.1016/j.ijmecsci.2022.107855.
[12]	LIU S S, FENG G Z, XIAO L J, et al. Shock-induced dynamic response in single and nanocrystalline high-entropy alloy FeNiCrCoCu [J]. International Journal of Mechanical Sciences, 2023, 239: 107859. DOI: 10.1016/j.ijmecsci.2022.107859.
[13]	XU W Z, GENG Y X, ZHENG H Z, et al. Microstructural evolution of FeCoNiCrMn high-entropy alloy subjected to laser shock peening: molecular dynamics simulation study [J]. Next Materials, 2025, 7: 100523. DOI: 10.1016/j.nxmate.2025.100523.
[14]	EUSER V K, JONES D R, MARTINEZ D T, et al. The effect of microstructure on the dynamic shock response of 1045 steel [J]. Acta Materialia, 2023, 250: 118874. DOI: 10.1016/j.actamat.2023.118874.
[15]	YANG Y, YANG S J, WANG H M. Effects of microstructure on the evolution of dynamic damage of Fe₅₀Mn₃₀Co₁₀Cr₁₀ high entropy alloy [J]. Materials Science and Engineering: A, 2021, 802: 140440. DOI: 10.1016/j.msea.2020.140440.
[16]	GAO T J, ZHAO D, ZHANG T W, et al. Strain-rate-sensitive mechanical response, twinning, and texture features of NiCoCrFe high-entropy alloy: experiments, multi-level crystal plasticity and artificial neural networks modeling [J]. Journal of Alloys and Compounds, 2020, 845: 155911. DOI: 10.1016/j.jallcom.2020.155911.
[17]	ZHOU Z X, HUO Y M, WANG Z J, et al. Experimental investigation and crystal plasticity modelling of dynamic recrystallisation in dual-phase high entropy alloy during hot deformation [J]. Materials Science and Engineering: A, 2025, 922: 147634. DOI: 10.1016/j.msea.2024.147634.
[18]	CONNOLLY D S, KOHAR C P, INAL K. A novel crystal plasticity model incorporating transformation induced plasticity for a wide range of strain rates and temperatures [J]. International Journal of Plasticity, 2022, 152: 103188. DOI: 10.1016/j.ijplas.2021.103188.
[19]	DAI W, WANG H M, GUAN Q, et al. Studying the micromechanical behaviors of a polycrystalline metal by artificial neural networks [J]. Acta Materialia, 2021, 214: 117006. DOI: 10.1016/j.actamat.2021.117006.
[20]	LIU C Z, ZHANG X Y, LIU X, et al. Mechanical field guiding structure design strategy for meta‐fiber reinforced hydrogel composites by deep learning [J]. Advanced Science, 2024, 11(22): 2310141. DOI: 10.1002/advs.202310141.
[21]	WANG K, SUN W. A multiscale multi-permeability poroplasticity model linked by recursive homogenizations and deep learning [J]. Computer Methods in Applied Mechanics and Engineering, 2018, 334: 337–380. DOI: 10.1016/j.cma.2018.01.036.
[22]	YUAN M F, PARADISO S, MEREDIG B, et al. Machine learning–based reduce order crystal plasticity modeling for ICME applications [J]. Integrating Materials and Manufacturing Innovation, 2018, 7(4): 214–230. DOI: 10.1007/s40192-018-0123-x.
[23]	MELLBIN Y, HALLBERG H, RISTINMAA M. Accelerating crystal plasticity simulations using GPU multiprocessors [J]. International Journal for Numerical Methods in Engineering, 2014, 100(2): 111–135. DOI: 10.1002/nme.4724.
[24]	DE OCA ZAPIAIN D M, LIM H, PARK T, et al. Predicting plastic anisotropy using crystal plasticity and Bayesian neural network surrogate models [J]. Materials Science and Engineering: A, 2022, 833: 142472. DOI: 10.1016/j.msea.2021.142472.
[25]	IBRAGIMOVA O, BRAHME A, MUHAMMAD W, et al. A new ANN based crystal plasticity model for FCC materials and its application to non-monotonic strain paths [J]. International Journal of Plasticity, 2021, 144: 103059. DOI: 10.1016/j.ijplas.2021.103059.
[26]	TRAN A, WILDEY T. Solving stochastic inverse problems for property–structure linkages using data-consistent inversion and machine learning [J]. JOM, 2021, 73(1): 72–89. DOI: 10.1007/s11837-020-04432-w.
[27]	LIU J Q, HUANG M S, LI Z H, et al. A deep learning method for predicting microvoid growth in heterogeneous polycrystals [J]. Engineering Fracture Mechanics, 2022, 264: 108332. DOI: 10.1016/j.engfracmech.2022.108332.
[28]	HAN S Y, WANG C C, LAI Q Q, et al. Fitting-free mechanical response prediction in dual-phase steels by crystal plasticity theory guided deep learning [J]. Acta Materialia, 2025, 289: 120936. DOI: 10.1016/j.actamat.2025.120936.
[29]	IBRAGIMOVA O, BRAHME A, MUHAMMAD W, et al. A convolutional neural network based crystal plasticity finite element framework to predict localised deformation in metals [J]. International Journal of Plasticity, 2022, 157: 103374. DOI: 10.1016/j.ijplas.2022.103374.
[30]	HERRIOTT C, SPEAR A D. Predicting microstructure-dependent mechanical properties in additively manufactured metals with machine- and deep-learning methods [J]. Computational Materials Science, 2020, 175: 109599. DOI: 10.1016/j.commatsci.2020.109599.
[31]	LIU C Z, XIONG Q L, AN W. Shear localization in gradient high-entropy alloy at high strain rates: crystal plasticity modeling [J]. Extreme Mechanics Letters, 2024, 70: 102194. DOI: 10.1016/j.eml.2024.102194.
[32]	GUI Y, AN D Y, HAN F B, et al. Multiple-mechanism and microstructure-based crystal plasticity modeling for cyclic shear deformation of TRIP steel [J]. International Journal of Mechanical Sciences, 2022, 222: 107269. DOI: 10.1016/j.ijmecsci.2022.107269.
[33]	ALLEY E S, NEU R W. A hybrid crystal plasticity and phase transformation model for high carbon steel [J]. Computational Mechanics, 2013, 52(2): 237–255. DOI: 10.1007/s00466-012-0810-y.
[34]	HARIDAS R S, AGRAWAL P, THAPLIYAL S, et al. Strain rate sensitive microstructural evolution in a TRIP assisted high entropy alloy: experiments, microstructure and modeling [J]. Mechanics of Materials, 2021, 156: 103798. DOI: 10.1016/j.mechmat.2021.103798.
[35]	SOLTANI M, FERDOUSI S, HARIDAS R S, et al. A crystal plasticity finite element—machine learning combined approach for phase transformation prediction in high entropy alloy [J]. International Journal of Applied Mechanics, 2024, 16(2): 2450024. DOI: 10.1142/S1758825124500248.