Dynamic physics-guided neural network for predicting hot deformation behavior of TiAl-based intermetallic alloys

Abstract

Ti-Al-based intermetallic compounds are promising candidates for high-temperature structural applications owing to their outstanding mechanical properties. Ti₂AlNb alloys, characterized by complex multiphase microstructures, present significant challenges for hot deformation modeling because of their atypical flow behavior and sensitivity to processing conditions. In this study, we systematically investigated the hot deformation behavior of Ti₂AlNb through experiments and compared conventional constitutive models with advanced machine learning approaches. The conventional strain-compensated Sellars (SCS) model showed limited accuracy for Ti₂AlNb, especially across complex microstructural transitions, while performing well for simpler alloy systems like Ti4822. To address these limitations, we developed a dynamic physics-guided neural network (DPGNN) that integrates physical constraints with data-driven learning via an adaptive gating mechanism. The DPGNN model significantly outperformed the SCS model and three purely data-driven baselines, achieving high accuracy (test R² > 0.98) and robust generalization across both Ti₂AlNb and Ti4822 alloys. These findings highlight the value of embedding physical principles within machine learning frameworks, providing a robust and generalizable tool for predicting hot deformation behavior in advanced alloys.

Keywords

deformation behavior / physics-guided neural network / Ti₂AlNb / TiAl-based intermetallic compound

References

[1] Nó ML, Klein T, Clemens H, Juan JMS. High-temperature microstructure evolution of an advanced intermetallic nano-lamellar γ-TiAl-based alloy and associated diffusion processes. Acta Mater. 2023;261:119380.

[2] Burtscher M, Weibensteiner I, Wartbichler R, et al. Precipitation behavior of hexagonal carbides in a C containing intermetallic γ-TiAl based alloy. J Alloys Compd. 2023;969:172400.

[3] Kagerera S, Hudaka OE, Wojcik T, et al. Oxidation protection of TNM alloys with Al-rich γ-TiAl-based coatings. J Alloys Compd. 2023;969:172343.

[4] Duan BH, Yang YC, He SY, et al. History and development of γ-TiAl alloys and the effect of alloying elements on their phase transformations. J Alloys Compd. 2022;909:164811.

[5] Genc O, Unal R. Development of gamma titanium aluminide (γ-TiAl) alloys: a review. J Alloys Compd. 2022;929:167262.

[6] Chi YD, Dong ZX, Cui MC, et al. Comparative study on machinability and surface integrity of γ-TiAl alloy in laser assisted milling. J Mater Res Technol. 2024;33:3743-3755.

[7] Anwar S, Ahmed N, Pervaiz S, Ahmad S, Mohammad A, Saleh M. On the turning of electron beam melted gamma-TiAl with coated and uncoated tools: a machinability analysis. J Mater Process Technol. 2020;282:116664.

[8] Lu Z-G, Wu J, Guo R-P, Xu L, Yang R. Hot deformation mechanism and ring rolling behavior of powder metallurgy Ti2AlNb intermetallics. Acta Metall Sin. 2017;30(7):621-629.

[9] Jia ZY, Zhang Y, Meng FC, et al. Improving the uniform deformation ability and ductility of powder metallurgical Ti2AlNb intermetallic with single-step solution heat treatment. J Alloys Compd. 2025;1011:178377.

[10] Banerjee D. The intermetallic Ti2AlNb. Prog Mater Sci. 1997;42(1-4):135-158.

[11] Yang J, Wang G, Jiao X, Li Y, Liu Q. High-temperature deformation behavior of the extruded Ti-22Al-25Nb alloy fabricated by powder metallurgy. Mater Charact. 2018;137:170-179.

[12] Zhang HY, Yan N, Liang HY, Liu YC. Phase transformation and microstructure control of TiAlNb-based alloys: a review. J Mater Sci Technol. 2021;80:203-216.

[13] Zhang H, Zhang Y, Liang H, Liu Y. Influence of cooling rates on microstructure and tensile properties of a heat treated Ti2AlNb-based alloy. Mater Sci Eng A. 2021;817:141345.

[14] Zhang H, Yan N, Liang H, Liu Y. Phase transformation and microstructure control of Ti2AlNb-based alloys: a review. J Mater Sci Technol. 2021;80:203-216.

[15] Zhang SW, Xi MZ, Sun XX, et al. The effect of post heat treatment on microstructural evolution and mechanical properties of Ti2AlNb intermetallic alloy prepared by point-forging and laser-deposition. Mater Sci Eng A. 2024;916:147331.

[16] Ayadh W, Denand B, Halkoum A, et al. Effect of prior α2 phase on precipitation kinetics of O-phase in advanced Ti2AlNb alloy. Acta Mater. 2023;252:118930.

[17] Shao B, Zong YY, Wen DS, Tian YT, Shan DB. Investigation of the phase transformations in Ti-22Al-25Nb alloy. Mater Charact. 2016;114:75-78.

[18] Wei JX, Jia Y, Zhang YY, et al. Effect of hot rolling process on the evolution of microstructure and mechanical properties of Ti2AlNb-based alloy foil during cold rolling. Mater Charact. 2024;210:113784.

[19] Sui XC, Lin J, Cheng S, Liu YD, Wang GF, Li ZW. Controlling the tensile properties of a high-strength-ductility Ti2AlNb alloy by hot rolling. J Mater Res Technol. 2024;33:1846-1859.

[20] Wang J, Yang P, Wang D. Constitutive equation and microstructure analysis of Al0.6CoCrFeNi high entropy alloy during hot deformation. Philos Mag. 2022;102(17):1684-1707.

[21] Jeong HT, Park HK, Kim WJ. Dynamic recrystallization and hot deformation mechanisms of a eutectic Al0.7CoCrFeMnNi high-entropy alloy. J Alloys Compd. 2021;871:159488.

[22] Wu K, Liu GQ, Hu BF, et al. Effect of processing parameters on hot compressive deformation behavior of a new Ni–Cr–Co based P/M superalloy. Mater Sci Eng A. 2011;528(13-14):4620-4629.

[23] Xiong J, He J-C, Leng X-S, Zhang T-Y. Gaussian process regressions on hot deformation behaviors of FGH98 nickel-based powder superalloy. J Mater Sci Technol. 2023;146:177-185.

[24] Guan Y, Liu Y, Ma Z, Li H, Yu H. Hot deformation behavior and recrystallization mechanism in an as-cast CoNi-based superalloy. Met Mater Int. 2021;28(6):1488-1498.

[25] Guo NN, Wang L, Luo LS, et al. Hot deformation characteristics and dynamic recrystallization of the MoNbHfZrTi refractory high-entropy alloy. Mater Sci Eng A. 2016;651:698-707.

[26] Tian SW, Jiang HT, Guo WQ, Zhang GH, Zeng SW. Hot deformation and dynamic recrystallization behavior of TiAl-based alloy. Intermetallics. 2019;112:106521.

[27] Kong FT, Chen YY, Zhang DL, Zhang SZ. High temperature deformation behavior of Ti-46Al-2Cr-4Nb-0.2Y alloy. Mater Sci Eng A. 2012;539:107-114.

[28] Abdelghany AW, Jaskari M, Hamada AS, et al. Hot deformation behavior and constitutive modeling of a cost-effective Al8Cr12Mn25Ni20Fe35 high-entropy alloy. J Alloys Compd. 2022;928:167028.

[29] Wang H, Liu Y, Zhang W, Wang L, Liu B. Development of constitutive relationship for the hot deformation of Ti-47Al-2Cr-2Nb-0.2W alloy. Adv Mater Res. 2011;156-157:625-632.

[30] Murty SVSN, Sarkar A, Narayanan PR, Venkitakrishnan PV, Mukhopadhyay J. Development of processing maps and constitutive relationship for thermomechanical processing of aluminum alloy AA2219. J Mater Eng Perform. 2017;26(5):2190-2203.

[31] He ZB, Wang ZB, Li P. A comparative study on Arrhenius and Johnson-Cook constitutive models for high-temperature deformation of Ti2AlNb-based alloys. Metals. 2019;9(2):123.

[32] Zhang B, Shang XD, Yao S, et al. A comparative study on Johnson-Cook, modified Johnson-Cook, modified Zerilli-Armstrong and Arrhenius-type constitutive models to predict hot deformation behavior of TA2. High Temp Mater Process. 2019;38:699-714.

[33] Safari A, Imran M, Weiss S. A comparative study on modified Johnson-Cook and arrhenius-type constitutive models to predict the hot deformation behaviour of molybdenum-hafnium-carbide alloy. J Mater Eng Perform. 2021;30(3):1945-1956.

[34] Zhan HY, Wang G, Kent D, Dargusch M. Constitutive modelling of the flow behaviour of a β titanium alloy at high strain rates and elevated temperatures using the Johnson-Cook and modified Zerilli-Armstrong models. Mater Sci Eng A. 2014;612:71-79.

[35] Cheng L, Xue XY, Tang B, Kou HC, Li JS. Flow characteristics and constitutive modeling for elevated temperature deformation of a high Nb containing TiAl alloy. Intermetallics. 2014;49:23-28.

[36] Xiong J, Shi SQ, Zhang TY. Machine learning of phases and mechanical properties in complex concentrated alloys. J Mater Sci Technol. 2021;87:133-142.

[37] Frie C, Durmaz AR, Eberl C. Exploration of materials fatigue influence factors using interpretable machine learning. Fatig Fract Eng Mater Struct. 2024;47(8):2752-2773.

[38] Pei CH, Ma QS, Zhang JW, et al. A novel model to predict oxidation behavior of superalloys based on machine learning. J Mater Sci Technol. 2025;235:232-243.

[39] Su YT, Liang CX, Wang D. Composition- and temperature-dependence of β to ω phase transformation in Ti-Nb alloys. J Mater Inf. 2023;3(3).

[40] Wang P, Zhao S, Zhou CL, et al. Simple formula learned via machine learning for creep rupture life prediction of high-temperature titanium alloys. J Mater Inf. 2024;4(4).

[41] Zhao S, Li J, Liao W, Yuan R. Machine learning-guided plasticity model in refractory high-entropy alloys. Mater Genome Eng Adv. 2025;3(2):e70022.

[42] Lookman T, Xue D. Multi-objective optimization and its application in materials science. Mater Genome Eng Adv. 2023;1(2):e14.

[43] Zhao R, He JC, Tian H, Jing YJ, Xiong J. Application of constitutive models and machine learning models to predict the elevated temperature flow behavior of TiAl alloy. Materials. 2023;16(14):4987.

[44] Ge GW, Wang ZM, Zhang LQ, Lin JP. Hot deformation behavior and artificial neural network modeling of β-γ TiAl alloy containing high content of Nb. Mater Today Commun. 2021;27:102405.

[45] Wang QH, Qin X, Xia SX, et al. Interpretable machine learning excavates a low-alloyed magnesium alloy with strength-ductility synergy based on data augmentation and reconstruction. J Magnesium Alloys. 2025;13(6):2866-2883.

[46] Qin X, Wang QH, Zhao XQ, et al. PCS: property-composition-structure chain in Mg-Nd alloys through integrating sigmoid fitting and conditional generative adversarial network modeling. Scr Mater. 2025;265:116762.

[47] Xiong J, Zhang T-Y. Data-driven glass-forming ability criterion for bulk amorphous metals with data augmentation. J Mater Sci Technol. 2022;121:99-104.

[48] Borate P, Rivière J, Marone C, Mali A, Kifer D, Shokouhi P. Using a physics-informed neural network and fault zone acoustic monitoring to predict lab earthquakes. Nat Commun. 2023;14(1):3693.

[49] Do MH, Ammar K, Castaing NG, Madiot F. Physics Informed Neural Networks for the mixed dual form of the neutron diffusion equation with heterogeneous coefficients. Ann Nucl Energy. 2025;223:111607.

[50] Goyal K, Sardana N. Phase stability and microstructural evolution of Ti2AlNb alloys-a review. Mater Today: Proc. 2021;41:951-968.

[51] Sellars CM, Tegart WJM. Hot workability. Int Mater Rev. 1972;17(1):1-24.

[52] Sellars CM. The role of computer modelling in deformation processing. Trans Indian Inst Met. 1997;50(6):563-571.

[53] Sellars CM, Tegart WJM. On the mechanism of hot deformation. Acta Metall. 1966;14(9):1136-1138.

[54] Zhang PH, Zeng WD, Jia RC, et al. Tensile behavior and deformation mechanism for Ti-22Al-25Nb alloy with lamellar O microstructures. Mater Sci Eng A. 2021;803:140492.

[55] Zhao Q, Lv MQ, Cui ZS. Investigation on transformation-related recrystallization behavior of Ti2AlNb-based alloy. Intermetallics. 2021;138:107302.

[56] Ma X, Zeng WD, Xu B, Sun Y, Xue C, Han YF. Characterization of the hot deformation behavior of a Ti-22Al-25Nb alloy using processing maps based on the Murty criterion. Intermetallics. 2012;20(1):1-7.

[57] Zhang HY, Li C, Ma ZQ, Yu LM, Li HJ, Liu YC. Morphology and quantitative analysis of O phase during heat treatment of hot-deformed Ti2AlNb-based alloy. Int J Miner Metall Mater. 2018;25(10):1191-1200.

[58] Rao KP, Prasad YVRK, Suresh K. Hot working behavior and processing map of a γ-TiAl alloy synthesized by powder metallurgy. Mater Des. 2011;32(10):4874-4881.

RIGHTS & PERMISSIONS

2025 The Author(s). Materials Genome Engineering Advances published by Wiley-VCH GmbH on behalf of University of Science and Technology Beijing.