Residual Stress Prediction of TA15 Alloy in Laser Powder Bed Fusion Based on Explainable Machine Learning

ZENG Yuan; HE Jiajia

doi:10.3969/j.issn.1674-6457.2026.08.021

PDF(14332 KB)

Journal of Netshape Forming Engineering ›› 2026, Vol. 18 ›› Issue (8) : 229-241. DOI: 10.3969/j.issn.1674-6457.2026.08.021

Additive Manufacturing

Residual Stress Prediction of TA15 Alloy in Laser Powder Bed Fusion Based on Explainable Machine Learning

ZENG Yuan^*, HE Jiajia

Author information +

History +

Abstract

In laser powder bed fusion (LPBF), the rapid heating and cooling of powder often generates significant residual stress, leading to part deformation and cracking. To efficiently predict the residual stress in TA15 alloy, the work aims to propose an explainable machine learning approach. Firstly, a dataset of residual stress was generated through finite element simulation and Latin hypercube experimental design. Prediction models were constructed through XGBoost, support vector regression (SVR), and back‑propagation neural network (BPNN), with laser power, scanning speed, and hatch spacing as inputs and residual stress of TA15 alloy as output. The models were optimized via five‑fold cross‑validation and particle swarm optimization (PSO), and their performance was comprehensively evaluated through statistical metrics. Finally, Shapley Additive Explanations (SHAP) were applied to interpret the model and investigate the interactive effects of process parameters on residual stress. PSO effectively improved the prediction accuracy of XGBoost, BPNN, and SVR, increasing their coefficients of determination by 7.9%, 10.6%, and 26%, respectively. Among them, PSO‑XGBoost achieved the highest accuracy, with a coefficient of determination, mean absolute error, and mean squared error of 0.849, 4.1, and 35.5, respectively. Through explainable machine learning, hatch spacing was identified as the key variable affecting residual stress. The proposed approach provides an effective tool for the accurate prediction and process optimization of residual stress in TA15 alloy during LPBF, and can serve as a reference for the forming processes of other materials.

Key words

laser powder bed fusion / TA15 alloy / residual stress simulation / explainable machine learning / extreme gradient boosting decision trees

Cite this article

EndNote

Ris (Procite)

Bibtex

Download Citations

ZENG Yuan, HE Jiajia. Residual Stress Prediction of TA15 Alloy in Laser Powder Bed Fusion Based on Explainable Machine Learning[J]. Journal of Netshape Forming Engineering. 2026, 18(8): 229-241 https://doi.org/10.3969/j.issn.1674-6457.2026.08.021

References

[1] 李博洋, 刘强, 陈宜, 等. SLM成形拱形悬垂结构的过渡圆角优化仿真与验证[J]. 精密成形工程, 2025, 17(9): 126-135.
LI B Y, LIU Q, CHEN Y, et al.Simulation and Validation of Transition Fillet Optimization for SLM-Formed Arch Overhang Structures[J]. Journal of Netshape Forming Engineering, 2025, 17(9): 126-135.
[2] 周运龙, 马毅, 管迎春. 面向航空发动机高性能制造的激光选区熔化技术研究进展[J]. 航空学报, 2024, 45(13): 43-80.
ZHOU Y L, MA Y, GUAN Y C.Research Progress on Laser Selective Melting Technology for High-Performance Manufacturing of Aero-Engines[J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(13): 43-80.
[3] 魏取龙, 姜丽红, 刘征, 等. 选区激光熔化制备TPMS晶格结构及力学性能[J]. 航空学报, 2025, 46(3): 296-311.
WEI Q L, JIANG L H, LIU Z, et al.Lattice Structure and Mechanical Properties of TPMS Prepared by Selective Laser Melting[J]. Acta Aeronautica et Astronautica Sinica, 2025, 46(3): 296-311.
[4] VOISIN T, CALTA N P, KHAIRALLAH S A, et al.Defects-Dictated Tensile Properties of Selective Laser Melted Ti-6Al-4V[J]. Materials & Design, 2018, 158: 113-126.
[5] KUMAR P, RAMAMURTY U.High Cycle Fatigue in Selective Laser Melted Ti-6Al-4V[J]. Acta Materialia, 2020, 194: 305-320.
[6] LIVERANI E, TOSCHI S, CESCHINI L, et al.Effect of Selective Laser Melting (SLM) Process Parameters on Microstructure and Mechanical Properties of 316L Austenitic Stainless Steel[J]. Journal of Materials Processing Technology, 2017, 249: 255-263.
[7] LI Z H, YANG S, LIU B, et al.Simulation of Temperature Field and Stress Field of Selective Laser Melting of Multi-Layer Metal Powder[J]. Optics & Laser Technology, 2021, 140: 106782.
[8] ZHANG C L, WANG Y, LIANG H Y, et al.Effect of Process Parameters on Residual Stresses in SLM-Formed Bionic Porous Titanium Alloy Structures[J]. Materials Today Communications, 2024, 39: 108539.
[9] 吴昌鼎, 高月华, 刘其鹏, 等. 同轴送粉激光增材工艺对温度场及应力场的影响[J]. 科学技术与工程, 2023, 23(14): 5970-5977.
WU C D, GAO Y H, LIU Q P, et al.Effect of Process Parameters on Temperature and Stress Fields in Laser Additive Manufacturing with Coaxial Powder-Feeding[J]. Science Technology and Engineering, 2023, 23(14): 5970-5977.
[10] 孙金娥, 贺亚飞. SLM成形Al-Mg-Li-Cu合金残余应力及组织性能[J]. 特种铸造及有色合金, 2023, 43(3): 313-317.
SUN J E, HE Y F. Residual Stress, Microstructure and Properties of SLM-Formed Al-Mg-Li-Cu Alloy[J]. Special Casting & Nonferrous Alloys, 2023, 43(3): 313-317.
[11] 张喜龙, 周永军, 郭鹏程, 等. 基于机器学习和数值模拟的选区激光熔化Al-Mg-Sc-Zr合金成形工艺优化[J]. 中国有色金属学报, 2025, 35(2): 487-502.
ZHANG X L, ZHOU Y J, GUO P C, et al.Forming Process Optimization of Al-Mg-Sc-Zr Alloy by Selective Laser Melting Based on Machine Learning and Numerical Simulation[J]. The Chinese Journal of Nonferrous Metals, 2025, 35(2): 487-502.
[12] 洪学银, 高尚. 基于机器学习方法预测3D打印零件的性能[J]. 中国塑料, 2025, 39(7): 72-79.
HONG X Y, GAO S.Performance Prediction of 3D Printed Parts Based on Machine Learning Methods[J]. China Plastics, 2025, 39(7): 72-79.
[13] 刘少杰, 彭逸琦, 赵宇凡, 等. 基于机器学习与多目标优化算法的电弧增材制造过程优化研究[J]. 铸造技术, 2025, 46(1): 29-38.
LIU S J, PENG Y Q, ZHAO Y F, et al.Parameter Optimization of Wire Arc Additive Manufacturing via Machine Learning and a Multiobjective Optimization Algorithm[J]. Foundry Technology, 2025, 46(1): 29-38.
[14] LU C Y, SHI J.Relative Density Prediction of Additively Manufactured Inconel 718: A Study on Genetic Algorithm Optimized Neural Network Models[J]. Rapid Prototyping Journal, 2022, 28(8): 1425-1436.
[15] GUO W, GAWADE V, ZHANG B, et al.Explainable AI for Layer-Wise Emission Prediction in Laser Fusion[J]. CIRP Annals, 2023, 72(1): 437-440.
[16] LEE J A, SAGONG M J, JUNG J, et al.Explainable Machine Learning for Understanding and Predicting Geometry and Defect Types in Fe-Ni Alloys Fabricated by Laser Metal Deposition Additive Manufacturing[J]. Journal of Materials Research and Technology, 2023, 22: 413-423.
[17] FUTAGAMI K, FUKAZAWA Y, KAPOOR N, et al.Pairwise Acquisition Prediction with SHAP Value Interpretation[J]. The Journal of Finance and Data Science, 2021, 7: 22-44.
[18] XIAO Z X, CHEN C P, ZHU H H, et al.Study of Residual Stress in Selective Laser Melting of Ti₆Al₄V[J]. Materials & Design, 2020, 193: 108846.
[19] PARRY L, ASHCROFT I A, WILDMAN R D.Understanding the Effect of Laser Scan Strategy on Residual Stress in Selective Laser Melting through Thermo-Mechanical Simulation[J]. Additive Manufacturing, 2016, 12: 1-15.
[20] LI Z Q, QU H T, CHEN F L, et al.Deformation Behavior and Microstructural Evolution during Hot Stamping of TA15 Sheets: Experimentation and Modelling[J]. Materials, 2019, 12(2): 223.
[21] EDWARDS P, RAMULU M.Fatigue Performance Evaluation of Selective Laser Melted Ti-6Al-4V[J]. Materials Science and Engineering: A, 2014, 598: 327-337.
[22] LI C, LIU J F, FANG X Y, et al.Efficient Predictive Model of Part Distortion and Residual Stress in Selective Laser Melting[J]. Additive Manufacturing, 2017, 17: 157-168.
[23] ZOU S, XIAO H B, YE F P, et al.Numerical Analysis of the Effect of the Scan Strategy on the Residual Stress in the Multi-Laser Selective Laser Melting[J]. Results in Physics, 2020, 16: 103005.
[24] WAQAR S, GUO K, SUN J.FEM Analysis of Thermal and Residual Stress Profile in Selective Laser Melting of 316L Stainless Steel[J]. Journal of Manufacturing Processes, 2021, 66: 81-100.
[25] CAI C, GAO X Y, TENG Q, et al.A Novel Hybrid Selective Laser Melting/Hot Isostatic Pressing of Near-Net Shaped Ti-6Al-4V Alloy Using an In-Situ Tooling: Interfacial Microstructure Evolution and Enhanced Mechanical Properties[J]. Materials Science and Engineering: A, 2018, 717: 95-104.
[26] WANG C, TAN X P, TOR S B, et al.Machine Learning in Additive Manufacturing: State-of-the-Art and Perspectives[J]. Additive Manufacturing, 2020, 36: 101538.
[27] QIU Y G, ZHOU J, KHANDELWAL M, et al.Performance Evaluation of Hybrid WOA-XGBoost, GWO-XGBoost and BO-XGBoost Models to Predict Blast-Induced Ground Vibration[J]. Engineering with Computers, 2022, 38(5): 4145-4162.
[28] DAI H C, MACBETH C.Effects of Learning Parameters on Learning Procedure and Performance of a BPNN[J]. Neural Networks, 1997, 10(8): 1505-1521.
[29] CHAUHAN V K, DAHIYA K, SHARMA A.Problem Formulations and Solvers in Linear SVM: A Review[J]. Artificial Intelligence Review, 2019, 52(2): 803-855.
[30] MARJANOVIĆ M, KOVAČEVIĆ M, BAJAT B, et al. Landslide Susceptibility Assessment Using SVM Machine Learning Algorithm[J]. Engineering Geology, 2011, 123(3): 225-234.
[31] JORDEHI A R. Particle Swarm Optimisation (PSO) for Allocation of FACTS Devices in Electric Transmission Systems: A Review[J]. Renewable and Sustainable Energy Reviews, 2015, 52: 1260-1267.
[32] ZHU H H, WANG Y, WANG K S, et al.Particle Swarm Optimization (PSO) for the Constrained Portfolio Optimization Problem[J]. Expert Systems with Applications, 2011, 38(8): 10161-10169.
[33] KUMAR S, GOPI T, HARIKEERTHANA N, et al.Machine Learning Techniques in Additive Manufacturing: A State of the Art Review on Design, Processes and Production Control[J]. Journal of Intelligent Manufacturing, 2023, 34(1): 21-55.
[34] MEGAHED M, MINDT H W, N'DRI N, et al. Metal Additive-Manufacturing Process and Residual Stress Modeling[J]. Integrating Materials and Manufacturing Innovation, 2016, 5(1): 61-93.

Funding

Research on a Novel Teaching Model Integrating Learning Pass with the BOPPPS Model in Experimental Instruction for Mechanical Innovation Design (JXJG-24-29-10)