Fatigue Life Prediction of Gray Cast Iron for Cylinder Head Based on Microstructure and Machine Learning

doi:10.1007/s40195-023-01566-z

Abstract

Abstract:

Conventional fatigue tests on complex components are difficult to sample, time-consuming and expensive. To avoid such problems, several popular machine learning (ML) algorithms were used and compared to predict fatigue life of gray cast iron (GCI) with the complex microstructures. The feature analysis shows that the fatigue life of GCI is mainly influenced by the external environment such as the stress amplitude, and the internal microstructure parameters such as the percentage of graphite, graphite length, stress concentration factor at the graphite tip, matrix microhardness and Brinell hardness. For simplicity, collected datasets with some of the above features were used to train ML models including back-propagation neural network (BPNN), random forest (RF) and eXtreme gradient boosting (XGBoost). The comparison results suggest that the three models could predict the fatigue lives of GCI, while the implemented RF algorithm is the best performing model. Moreover, the S-N curves fitted by the Basquin relation in the predicted data have a mean relative error of 15% compared to the measured data. The results have demonstrated the advantages of ML, which provides a generic way to predict the fatigue life of GCI for reducing time and cost.

Key words: Gray cast iron, Microstructure feature, Machine learning, High-cycle fatigue life

Xiaoyuan Teng, Jianchao Pang, Feng Liu, Chenglu Zou, Xin Bai, Shouxin Li, Zhefeng Zhang. Fatigue Life Prediction of Gray Cast Iron for Cylinder Head Based on Microstructure and Machine Learning[J]. Acta Metallurgica Sinica (English Letters), 2023, 36(9): 1536-1548.

Figures/Tables 18

Table 1 Chemical composition of GCI (wt%)

CE	C	Si	Mn	P	S	Cu	Sn	Fe
3.05	2.49	1.80	0.84	0.018	0.059	0.67	0.02	Balance
CE = C + 0.31Si + 0.33P [22]

Fig. 1 Shapes and dimensions of tensile a and fatigue b specimens (units in mm), metallographic c and 3D graphite morphology d samples

Fig. 2 Fatigue life prediction process: search for fatigue life related features a, data feature selection based on decision tree algorithm b, fatigue life prediction of GCI by BPNN, RF and XGBoost algorithms c

Fig. 3 Schematic diagrams of the artificial neural network a and random forest b

Fig. 4 Schematic diagram of the XGBoost model structure

Table 2 Hyperparameters for XGBoost

Booster	Objective	Gamma	Lambda	Alpha	Subsample	Eta
gbtree	reg: gamma	0	3	1	0.5	0.6

Table 3 Tensile properties of the GCI

Material	UTS (MPa)	YS (MPa)	EF (%)
HT300	269 ± 11	234 ± 13	0.45 ± 0.15

Table 4 Fatigue testing results for GCI specimens run in load-control mode at R = − 1

Specimens	σ_a (MPa)	F_max (kN)	N_f (cycles)
S-23	120	2.16	13,700
S-24	120	2.17	15,100
S-21	110	2.02	16,900
S-19	100	1.84	26,800
S-7	70	1.31	57,400
S-2	85	1.59	62,600
S-11	70	1.28	82,300
S-22	110	2.03	91,300
S-20	90	1.63	116,300
S-4	75	1.42	145,400
S-18	100	1.89	168,400
S-6	75	1.33	195,200
S-1	90	1.68	220,300
S-12	65	1.23	253,500
S-16	80	1.48	309,900
S-3	80	1.49	497,200
S-10	75	1.35	1,573,900
S-13	60	1.10	> 10,000,000
S-8	65	1.17	> 10,000,000
S-14	65	1.20	> 10,000,000
S-5	70	1.29	> 10,000,000
S-9	70	1.26	> 10,000,000
S-15	80	1.45	> 10,000,000
S-17	85	1.59	> 10,000,000

Fig. 5 2D morphologies of graphite in GCI observed by OM a, SEM b and XRT c

Fig. 6 Spatial structure and morphology of graphite of GCI a graphite on the surface, b 3D morphology of graphite with the sizes less than 351 µm and c 3D morphology of graphite with the sizes about 25 µm to 125 µm

Fig. 7 GCI microstructure characteristics: a graphite in the matrix, OM, b graphite shape, SEM

Fig. 8 Comparison of feature importance under decision tree algorithm

Fig. 9 Heatmap of Pearson correlation coefficients: feature correlation coefficients (upper triangle) and scatter plot (bottom triangle)

Fig. 10 ML evaluation of GCI with dataset ratio (the numbers 1, 2 and 3 indicate the 1st, 2nd and 3rd splits in the training set)

Fig. 11 ML model predictive capability based on the BPNN a, RF b and XGBoost c algorithms

Fig. 12 Comparison of the predictive performances (R2, MAE and RMSE) of different ML (BPNN, RF and XGBoost) models for GCI

Fig. 13 Evaluating the importance of features on BPNN a, RF b and XGBoost c models

Fig. 14 Predicted results based on the Basquin relation: testing data a, BPNN b, RF c, XGBoost d, comparison of the three predicted results with the testing results e, fitted curves error relationship f

References 41

[1]	W.M. Zhao, W.Z. Zhang, Study on thermal fatigue life prediction of cylinder head. In: International Conference on Materials Science and Information Technology (MSIT 2011, Singapore, 2011), pp. 3-8
[2]	B.B. Guo, W.Z. Zhang, X.S. Wang, Adv. Mech. Eng. 6, 862853 (2014) DOI URL
[3]	Y. Chen, J.C. Pang, S.X. Li, Z.F. Zhang, Acta Metall. Sin. -Engl. Lett. 35, 1117 (2022)
[4]	Q.Y. Liu, G. Zhu, J.C. Pang, F. Liu, S.X. Li, C.L. Guo, A.L. Jiang, Z.F. Zhang, Mater. Sci. Eng. A 764, 138248 (2019)
[5]	C.L. Zou, J.C. Pang, M.X. Zhang, Y. Qiu, S.X. Li, L.J. Chen, J.P. Li, Z. Yang, Z.F. Zhang, Mater. Sci. Eng. A 724, 606 (2018)
[6]	L. He, Z.L. Wang, H. Akebono, A. Sugeta, J. Mater. Sci. Technol 90, 9 (2021)
[7]	J.C. Pang, S.X. Li, Z.G. Wang, Z.F. Zhang, Mater. Sci. Eng. A 564, 331 (2013)
[8]	J.C. Pang, S.X. Li, Z.G. Wang, Z.F. Zhang, Fatigue Fract. Eng. Mater. Struct. 37, 958 (2014) DOI URL
[9]	W.J. Yang, J.C. Pang, L. Wang, S.G. Wang, Y.Z. Liu, L. Hui, S.X. Li, Z.F. Zhang, Mater. Sci. Eng. A 814, 141244 (2021)
[10]	Y.Y. Zhang, J.C. Pang, R.L. Shen, Y. Qiu, S.X. Li, Z.F. Zhang, Mater. Sci. Eng. A 713, 260 (2018)
[11]	Y. Qiu, J.C. Pang, C.L. Zou, M.X. Zhang, S.X. Li, Z.F. Zhang, Mater. Sci. Eng. A 724, 324 (2018)
[12]	E. Alabort, Y. Tang, D. Barba, R.C. Reed, Acta Mater. 229, 117749 (2022) DOI URL
[13]	B. Luo, H.T. Wang, H.Q. Liu, B. Li, F.Y. Peng, IEEE Trans. Ind. Electron. 66, 509 (2019) DOI URL
[14]	Z.H. Li, L. Qin, B.S. Guo, J.P. Yuan, Z.G. Zhang, W. Li, J.W. Mi, Acta Metall. Sin. -Engl. Lett. 35, 115 (2022)
[15]	N. Wagner, J.M. Rondinelli, Front. Chem. 3, 28 (2016)
[16]	C. Wen, Y. Zhang, C.X. Wang, D.Z. Xue, Y. Bai, S. Antonov, L.H. Dai, T. Lookman, Y.J. Su, Acta Mater. 170, 109 (2019) DOI URL
[17]	K. Schulz, S. Kreis, H. Trittenbach, K. Bohm, Eng. Fract. Mech. 218, 106552 (2019) DOI URL
[18]	X. Wei, S. van der Zwaag, Z. Jia, C. Wang, W. Xu, Acta Mater. 235, 118103 (2022) DOI URL
[19]	J.W. Lin, J.H. Zhang, G.C. Zhang, G.J. Ni, F.R. Bi, Chin. J. Chem. Eng. 25, 338 (2012) DOI URL
[20]	Z.X. Zhan, H. Li, Int. J. Fatigue 142, 105941 (2021)
[21]	X. Peng, S.C. Wu, W.J. Qian, J.G. Bao, Y.A. Hu, Z.X. Zhan, G.P. Guo, P.J. Withers, Int. J. Mech. Sci. 221, 107185 (2022) DOI URL
[22]	G.Q. Wang, X. Chen, Y.X. Li, Z.L. Liu, Materials 11, 1876 (2018)
[23]	S.B. Kotsiantis, Artif. Intell. Rev. 39, 261 (2013) DOI URL
[24]	T.Z. Liu, G.P. Zou, Comput. Intell. Neurosci. 2021, 2115653 (2021)
[25]	S. Calcaterra, G. Campana, L. Tomesani, J. Mater. Process. Technol. 104, 74 (2000) DOI URL
[26]	C. Fragassa, M. Babic, C.P. Bergmann, G. Minak, Metals 9, 557 (2019)
[27]	X.Y. Sun, K. Zhou, S.W. Shi, K. Song, X. Chen, Int. J. Fatigue 162, 106996 (2022)
[28]	T.Q. Chen, C. Guestrin, M. Assoc Comp, Assoc Comp, XGBoost: a Scalable tree boosting system. In: 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD). San Francisco, CA, 2016, pp. 785-794
[29]	D.H. Yu, T. Zhou, H. Zhou, P. Zhang, Y.Y. Yan, J. Mater. Eng. Perform. 31, 3962 (2022) DOI
[30]	S.N. Pan, F.Z. Zeng, N.G. Su, Z.K. Xian, J. Mater. Res. Technol. 9, 1509 (2020) DOI URL
[31]	M. Shirai, H. Yamada, Mater. Trans. 61, 176 (2020) DOI URL
[32]	Y. Murakami, T. Takagi, K. Wada, H. Matsunaga, Int. J. Int. J. Fatigue 146, 14 (2021)
[33]	V.H.C. de Albuquerque, P.C. Cortez, A.R. de Alexandria, J. Tavares, Nondestruct. Test Evaluation 23, 273 (2008)
[34]	Z.R. He, G.X. Lin, S. Ji, Mater. Sci. Eng. A 234, 161 (1997)
[35]	M.D. Cai, G.X. Sun, China Foundry 4, 103 (2007)
[36]	Y. Qiu, J.C. Pang, S.X. Li, E.N. Yang, W.Q. Fu, M.X. Liang, Z.F. Zhang, Mater. Sci. Eng. A 664, 75 (2016)
[37]	H. Mughrabi, Metall. Mater. Trans. B 40, 431 (2009)
[38]	C. Gao, M.Q. Yang, J.C. Pang, S.X. Li, M.D. Zou, X.W. Li, Z.F. Zhang, Mater. Sci. Eng. A 832, 142418 (2022)
[39]	L. Collini, G. Nicoletto, R. Konecna, Mater. Sci. Eng. A 488, 529 (2008)
[40]	H.Q. Lu, M. Liu, D.H. Yu, T. Zhou, H. Zhou, P. Zhang, H.F. Bo, W. Su, Z.H. Zhang, H. Bao, Metall. Mater. Trans. A 49, 5848 (2018)
[41]	K. Wang, F. Wang, W.C. Cui, A.L. Tian, Acta Metall. Sin. -Engl. Lett. 28, 619 (2015)