Data driven multi objective optimization of valve seat lapping using machine learning surrogates and NSGA-II
DOI:
https://doi.org/10.15282/jmes.20.3.2026.6.0883Keywords:
Valve seat lapping, Multi-objective optimisation, Response surface methodology, Artificial neural network, Random forest, NSGA-IIAbstract
This paper presents a data-driven multi-objective optimization of a valve-seat lapping process by varying spindle speed, axial force, and lapping time using a three-level full-factorial design (3³ = 27 operating conditions). Each condition was replicated three times (n = 81), where the mean response at each design point was used for modelling and optimization, while replicate data supported pure-error diagnostics. Four responses are optimised simultaneously: leakage time (t_leak), surface roughness (Ra), electrical energy (E), and wear volume (V_wear). Quadratic response surface methodology (RSM), a regularized artificial neural network (ANN), and a random forest (RF) regressor are developed as surrogate models and evaluated using repeated 5-fold cross-validation (20 repeats), reported as out-of-fold mean ± SD for R² and RMSE. The quadratic RSM achieves the most consistent accuracy across all responses (R² ≈ 0.95–0.99), outperforming ANN and RF in this low-dimensional, smoothly varying design space. The surrogates are then coupled with NSGA-II to generate Pareto fronts, and optimization quality is quantified using a normalized hypervolume indicator. RSM-based optimization yields the highest and most stable hypervolume (HV = 0.74 ± 0.01), compared with ANN (0.71 ± 0.02) and RF (0.66 ± 0.03). Mechanically, increasing speed/force/time generally improves sealing performance (higher t_leak) and reduces Ra, but at the expense of higher E and V_wear; representative conservative, balanced, and aggressive regimes provide practical operating guidance. Overall, for the present dataset and within the tested bounds, classical quadratic response surfaces offer a reliable and interpretable basis for multi-objective optimisation, while more flexible machine-learning surrogates provide no clear advantage.
References
[1] G.E.P. Box and K.B. Wilson, “On the experimental attainment of optimum conditions,” Journal of the Royal Statistical Society, Series B, vol. 13, no. 1, pp. 1–45, 1951. https://doi.org/10.1111/j.2517-6161.1951.tb00067.x
[2] K. Deb, A. Pratap, S. Agarwal, and T. Meyarivan, “A fast and elitist multiobjective genetic algorithm: NSGA-II,” IEEE Transactions on Evolutionary Computation, vol. 6, no. 2, pp. 182–197, 2002. https://doi.org/10.1109/4235.996017
[3] L. Breiman, “Random forests,” Machine Learning, vol. 45, no. 1, pp. 5–32, 2001. https://doi.org/10.1023/A:1010933404324
[4] V.L. Trinh, B.T. Vu, and N.V. Thai, “A review of the surface roughness prediction methods in machining,” Engineering, Technology & Applied Science Research, vol. 14, no. 4, pp. 10941–10953, 2024. https://doi.org/10.48084/etasr.7710
[5] M.A. Pontes, A.J.C. da Silva, P.D. Sampaio, and R.M. de Souza, “Artificial neural networks for machining processes surface roughness modeling,” The International Journal of Advanced Manufacturing Technology, vol. 49, pp. 879–902, 2010. https://doi.org/10.1007/s00170-009-2456-2
[6] Z.W. Zhong, L.P. Khoo, and S.T. Han, “Prediction of surface roughness of turned surfaces using neural networks,” The International Journal of Advanced Manufacturing Technology, vol. 28, no. 7–8, pp. 688–693, 2006. https://doi.org/10.1007/s00170-004-2429-4
[7] B. Beatrice, B. Anandakrishnan, and M.R. Shunmugam, “Surface roughness prediction using artificial neural network in hard turning of AISI H13 steel with minimal cutting fluid application,” Procedia Engineering, vol. 97, pp. 205–211, 2014. https://doi.org/10.1016/j.proeng.2014.12.245
[8] S. Kalidass, M.P. Samraj, and R. Pugazhenthi, “Prediction of surface roughness for AISI 304 steel with regression and artificial neural network models,” Arabian Journal for Science and Engineering, vol. 39, pp. 299–311, 2014. https://doi.org/10.1007/s13369-014-1346-6
[9] T. Zhou, L. He, J. Wu, F. Du, and Z. Zou, “Prediction of surface roughness of 304 stainless steel and multi-objective optimization of cutting parameters based on GA-GBRT,” Applied Sciences, vol. 9, no. 18, p. 3684, 2019. https://doi.org/10.3390/app9183684
[10] C. Avram, C. Cismaru, G. Oancea, and M. Arghir, “Surface roughness determination with the help of artificial neural networks as enabler of metal machining process controlling system,” International Journal of Computers Communications & Control, vol. 20, no. 2, Art. no. 7028, 2025. https://doi.org/10.15837/ijccc.2025.2.7028
[11] A. Osan and T. Drenta, “Application of artificial neural networks in predicting surface quality and machining time,” Machines, vol. 13, no. 7, p. 561, 2025. https://doi.org/10.3390/machines13070561
[12] A.K. Ghosh, S. Debnath, and S. Ghosh, “Artificial neural networks modelling for surface finish during MQL turning,” in AIP Conference Proceedings, vol. 3360, 2023, p. 020019. https://doi.org/10.1063/5.0195537
[13] A. Balasuadhakar, S.T. Kumaran, and M. Uthayakumar, “Machine learning prediction of surface roughness in sustainable machining of AISI H11 tool steel,” Smart Materials in Manufacturing, vol. 1, no. 1, p. 100075, 2025. https://doi.org/10.1016/j.smmf.2025.100075
[14] K. Cevik, B.O. Yildiz, and U. Orak, “Machine learning-driven optimization of machining parameters: A comparative study of surrogate models and evolutionary algorithms,” Applied Sciences, vol. 15, no. 19, p. 6553, 2025. https://doi.org/10.3390/app15196553
[15] J. Guo, Q. Li, and H. Zhang, “A random forest regression with Bayesian optimization-based method for fatigue strength prediction,” Engineering Fracture Mechanics, vol. 283, p. 109714, 2023. https://doi.org/10.1016/j.engfracmech.2023.109714
[16] R. Li, Z. Wang, and J. Yan, “Multi-objective optimization of the process parameters of a grinding robot using LSTM-MLP-NSGAII,” Machines, vol. 11, no. 9, p. 882, 2023. https://doi.org/10.3390/machines11090882
[17] V.R. Khullar, R. Kumar, and I. Singh, “RSM- and NSGA-II-based multiple performance characteristics optimization in turning,” Arabian Journal for Science and Engineering, vol. 42, pp. 4181–4193, 2017. https://doi.org/10.1007/s13369-016-2399-5
[18] P.K. Singh and M. Shukla, “Multi-objective optimization of selected non-traditional machining processes using NSGA-II,” Decision Science Letters, vol. 9, no. 2, pp. 267–278, 2020. https://doi.org/10.5267/j.dsl.2019.9.003
[19] P. Senthilkumar, S. Raghuraman, and K.S. Shah, “Optimization of electrochemical machining process parameters using NSGA-II,” Journal of Minerals and Materials Characterization and Engineering, vol. 11, no. 11, pp. 1079–1088, 2012. https://doi.org/10.4236/jmmce.2012.1110091
[20] S. Cong, Y. Jiang, and F. Wei, “Multi-objective optimization based on response surface methodology and MOPSO for offshore pipeline network design,” Journal of Marine Science and Engineering, vol. 13, no. 6, p. 1037, 2025. https://doi.org/10.3390/jmse13061037
[21] H. Kahhal, H. Sebaaly, M.F. El-Kassas, and H.A. El-Hofy, “Multi-objective optimization of friction stir welding parameters using response surface methodology and MOPSO,” Scientific Reports, vol. 12, p. 2529, 2022. https://doi.org/10.1038/s41598-022-06652-3
[22] Q. Zhang, Y. Yu, Q. Zhao, and Y. Liu, “An experimental study of the leakage mechanism in static seals with iso-height asperity surfaces,” Applied Sciences, vol. 8, no. 8, p. 1404, 2018. https://doi.org/10.3390/app8081404
[23] W. Zhang, Y. Yu, and J. Wu, “Leakage model of metal static seal with two rough surfaces based on the theory of porous medium,” Journal of Mechanical Engineering, vol. 58, no. 21, pp. 215–224, 2022. https://doi.org/10.3901/JME.2022.21.215
[24] Y. Chun, C. Cho, J. Cho, and C. Lee, “A study of exhaust valve and seat insert wear depending on cycle numbers,” Wear, vol. 263, no. 7–12, pp. 1147–1157, 2007. https://doi.org/10.1016/j.wear.2007.02.006
[25] L. Wang, W. Yan, and Z. Guo, “The effect of operating conditions on heavy-duty engine valve seat wear,” Wear, vol. 201, no. 1–2, pp. 15–25, 1996. https://doi.org/10.1016/S0043-1648(96)06945-1
[26] Q. Zhao, Q. Jiao, Q. Wang, and J. Liu, “Wear mechanism analysis of engine exhaust valve seats,” Tribology Transactions, vol. 40, no. 4, pp. 580–586, 1997. https://doi.org/10.1080/10402009708983647
[27] A. Ootani, M. Ando, K. Tanaka, and N. Ujidono, “Impact wear characteristics of engine valve and valve seat insert materials at high temperature,” Wear, vol. 188, no. 1–2, pp. 174–184, 1995. https://doi.org/10.1016/0043-1648(95)06656-X
[28] R.T. Setyawan, I. Hajar, and G. Gunawan, “Full-model electric vehicle modeling and simulation for analysis of suspension and regenerative braking interaction,” Jurnal Teknik Mesin, vol. 22, no. 2, pp. 108–116, 2025. https://doi.org/10.9744/jtm.22.2.108-116 (in Indonesian)
[29] N. Crozet, M. Olivi-Tran, D. Bouvard, R. Van der Weide, and S. Vairac, “Valve-seat components in a diesel engine: A tribological approach to limit wear,” Mechanics & Industry, vol. 22, no. 4, p. 44, 2021. https://doi.org/10.1051/meca/2021043
[30] E.E. Vera-Cardenas, R. Lewis, and T. Slatter, “Sliding wear study on the valve-seat insert contact,” Open Journal of Applied Sciences, vol. 7, no. 2, pp. 42–49, 2017. https://doi.org/10.4236/ojapps.2017.72004
[31] C. Luo, Y. Zhang, X. Liu, and F. Li, “Impact wear behavior of the valve cone surface after plasma alloying treatment,” Applied Sciences, vol. 14, no. 11, p. 4811, 2024. https://doi.org/10.3390/app14114811
[32] Q. Yang, Z. Hu, X. Liu, Y. Cai, and Y. Song, “Investigation of contact characteristics and sealing performance of metal seal in ultra-high vacuum valve,” Processes, vol. 13, no. 4, p. 968, 2025. https://doi.org/10.3390/pr13040968
[33] S. Li, W. Xu, X. Tang, and J. Xie, “Theoretical analysis and numerical simulation of the wear of hydrogen ball valve seat,” ACS Omega, vol. 8, no. 40, pp. 37693–37704, 2023. https://doi.org/10.1021/acsomega.3c04506
[34] J.-H. Lee, G.-A. Kang, and J.-S. Kwak, “Valve seat design for full contact effect using grey relational analysis,” Journal of the Korean Society of Manufacturing Process Engineers, vol. 17, no. 4, pp. 17–22, 2018. https://doi.org/10.14775/ksmpe.2018.17.4.017
[35] M. Kumar, “Experimental and theoretical analysis of material removal in poppet valve magnetorheological finishing,” Proceedings of the Institution of Mechanical Engineers, Part E: Journal of Process Mechanical Engineering, early access, 2022. https://doi.org/10.1177/09544089221139102
[36] G. Zhang, M. Liu, J. Li, W. Ming, X. Shao, and Y. Huang, “Multi-objective optimization for surface grinding process using a hybrid particle swarm optimization algorithm,” The International Journal of Advanced Manufacturing Technology, vol. 71, pp. 1861–1872, 2014. https://doi.org/10.1007/s00170-013-5571-z
[37] Z. Gu, Q. Dai, W. Huang, and X. Wang, “Effect of surface topography and roughness on the leakage of static seals,” Surface Topography: Metrology and Properties, vol. 12, no. 2, p. 025014, 2024. https://doi.org/10.1088/2051-672X/ad4570
[38] C.M. Bishop, “Pattern recognition and machine learning,” Journal of Electronic Imaging, vol. 16, no. 4, p. 049901, 2007. https://doi.org/10.1117/1.2819119
[39] D.C. Montgomery, Design and Analysis of Experiments, 8th ed. Hoboken, NJ, USA: John Wiley & Sons, 2012.
[40] R.H. Myers, D.C. Montgomery, and C.M. Anderson-Cook, Response Surface Methodology: Process and Product Optimization Using Designed Experiments, 4th ed. Hoboken, NJ, USA: John Wiley & Sons, 2016.
[41] C.M. Bishop, Pattern Recognition and Machine Learning. New York, NY, USA: Springer, 2006.
[42] T. Hastie, R. Tibshirani, and J. Friedman, The Elements of Statistical Learning: Data Mining, Inference, and Prediction, 2nd ed. New York, NY, USA: Springer, 2009. https://doi.org/10.1007/978-0-387-84858-7
Downloads
Published
Issue
Section
License
Copyright (c) 2026 The Author(s)
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
