Tooth surface micromorphology greatly affects the vibration characteristics of gear systems. By strategically modifying the tooth micromorphology, it is possible to optimize vibration characteristics and enhance overall transmission efficiency. In this context, a new method dual-driven by dynamic model and machine learning is proposed for the analysis and optimization of vibration characteristics. Specifically, the nonlinear dynamic model of gear systems characterized by 3D anisotropic fractal tooth surfaces is constructed and the vibration characteristics yielded by the micromorphology parameters are analyzed. Two vibration response prediction models are then developed by training an eXtreme Gradient Boosting (XGBoost) with the dynamic responses. Based on the prediction models, an optimization problem for fractal tooth surface parameters is formulated by minimizing dynamic transmission error amplitude and maintaining periodic system motion, and the particle swarm optimizer is adopted to realize the optimization for vibration characteristics of gear systems. Under the representation of 3D anisotropic fractal methodology, the tooth surface can more comprehensively depict its microscopic characteristics. The XGBoost-based vibration response prediction models perform excellently in processing small samples, which ensures the precision, efficiency, and reliability of the method, making it well-suited for guiding the research on noise and vibration reduction of gear systems.
LiGWangZHZhuWD. Prediction of surface wear of involute gears based on a modified fractal method. J Tribol2019; 141: 015201. DOI: 10.1115/1.4041587
2.
YuGMaoHJiangL, et al. Fractal contact mechanics model for the rough surface of a beveloid gear with elliptical asperities. Appl Sci2022; 12: 4071.
3.
QiaoWZhangJ. Fractal modeling of gear tooth surface profile and research on meshing stiffness of planetary gear system with high-power density. In: 2021 12th International conference on mechanical and aerospace engineering (ICMAE), 2021, pp.251–255. New York: IEEE.
4.
CaiWZhangYLiL, et al. Cutting mechanics and efficiency of forward and reverse multidirectional turning. Int J Mech Sci2023; 242: 108031.
5.
XuHKahramanAAndersonNE, et al. Prediction of mechanical efficiency of parallel-axis gear pairs. J Mech Des2007; 129: 58–68.
6.
ChenQWangYZhouJ, et al. Research on characterization of anisotropic and isotropic processing surfaces by characteristic roughness. J Mater Process Technol2020; 275: 116277.
7.
BerryMVLewisZV. On the Weierstrass-Mandelbrot fractal function. Math Phys Sci1980; 370: 459–484. https://about.jstor.org/terms (1980).
YuXSunYLiH, et al. An improved meshing stiffness calculation algorithm for gear pair involving fractal contact stiffness based on dynamic contact force. Eur J Mech A/Solids2022; 94: 104595.
10.
ZhaoZHanHWangP, et al. An improved model for meshing characteristics analysis of spur gears considering fractal surface contact and friction. Mech Mach Theory2021; 158: 104219.
11.
JiangSZhengYZhuH. A contact stiffness model of machined plane joint based on fractal theory. J Tribol2010; 132(1): 7.
12.
FengZXCaiYJLiZ. Experimental research on the relation between processing parameters and fractal dimensions of the machined surface. Adv Mater Res2012; 602-604: 2011–2016.
13.
ChenQWangYTianW, et al. An improved nonlinear dynamic model of gear pair with tooth surface microscopic features. Nonlinear Dyn2019; 96: 1615–1634.
14.
YuXSunYLiH, et al. Nonlinear characteristics of gear pair considering fractal surface dynamic contact as internal excitation. Int J Non Linear Mech2022; 143: 104027.
15.
YiYHuangKXiongY, et al. Nonlinear dynamic modelling and analysis for a spur gear system with time-varying pressure angle and gear backlash. Mech Syst Signal Process2019; 132: 18–34.
16.
LiuDGuDLiuZ. Coupled vibration modeling and dynamic characteristics of noncircular face gear drive system with time-varying instantaneous center excitation. Proc IMechE, Part C: J Mechanical Engineering Science2019; 233: 4947–4959.
17.
FatourehchiEMohammadpourMKingPD, et al. Effect of mesh phasing on the transmission efficiency and dynamic performance of wheel hub planetary gear sets. Proc IMechE, Part C: J Mechanical Engineering Science2018; 232: 3469–3481.
18.
LiuGParkerRG. Nonlinear, parametrically excited dynamics of two-stage spur gear trains with mesh stiffness fluctuation. Proc IMechE, Part C: J Mechanical Engineering Science2012; 226: 1939–1957.
19.
ChenQMaYHuangS, et al. Research on gears’ dynamic performance influenced by gear backlash based on fractal theory. Appl Surf Sci2014; 313: 325–332.
20.
ChenQZhouJKhushnoodA, et al. Modelling and nonlinear dynamic behavior of a geared rotor-bearing system using tooth surface microscopic features based on fractal theory. AIP Adv2019; 9: 015201. DOI: 10.1063/1.5055907
21.
YuJWangHRenD. Chaos analysis of single-stage spur gear system considering backlash fractal. J Vib Eng Technol2023; 11: 3481–3491.
22.
YuXSunYLiuS, et al. Fractal-based dynamic response of a pair of spur gears considering microscopic surface morphology. Int J Mech Syst Dyn2021; 1: 194–206.
23.
YuXSunYWuS. Study on nonlinear dynamic response of spur gear under tooth surface microscopic topography. International Conference on Mechanical Design, Nature Singapore, Singapore, 2021.
24.
LeeS-HParkK-P. Development of a prediction model for the gear whine noise of transmission using machine learning. Int J Precis Eng Manuf2023; 24: 1793–1803.
25.
BonoriGBarbieriMPellicanoF. Optimum profile modifications of spur gears by means of genetic algorithms. J Sound Vib2008; 313: 603–616.
26.
TangZWangMChenZ, et al. Design of multi-stage gear modification for new energy vehicle based on optimized BP neural network. IEEE Access2020; 8: 199034–199050.
27.
TangZWangMHuY, et al. Optimal design of traction gear modification of high-speed EMU based on radial basis function neural network. IEEE Access2020; 8: 134619–134629.
28.
DongPZuoSDuS, et al. Optimum design of the tooth root profile for improving bending capacity. Mech Mach Theory2020; 151: 103910.
29.
QiuLHouXGaoS, et al. Vibration optimization of spur gear based on GSA-SA algorithm. PLoS One2023; 18: e0293460.
30.
GengZXiaoKLiJ, et al. Bifurcation and chaos of a spur gear transmission system with non-uniform wear. J Vib Acoust2021; 143: 031004. DOI: 10.1115/1.4048269
31.
ZhaoYMaiettaDMChangL. An asperity microcontact model incorporating the transition from elastic deformation to fully plastic flow. J Tribol2000; 122(1): 86–93.
32.
MengFXiaHZhangX, et al. A new tooth pitting modeling method based on matrix equation for evaluating time-varying mesh stiffness. Eng Fail Anal2022; 142: 106799.
33.
OdavićJMaliPTekićJ, et al. Application of largest Lyapunov exponent analysis on the studies of dynamics under external forces. Commun Nonlinear Sci Numer Simul2017; 47: 100–108.
34.
HuZTangJChenS, et al. Effect of mesh stiffness on the dynamic response of face gear transmission system. J Mech Des2013; 135: 071005. DOI: 10.1115/1.4024369
35.
KuboA. Stress condition, vibrational exciting force, and contact pattern of helical gears with manufacturing and alignment error. J Mech Des1978; 100: 77–84.
36.
ÖzgüvenHNHouserDR. Dynamic analysis of high speed gears by using loaded static transmission error. J Sound Vib1988; 125: 71–83.
37.
MalihaRDogruerCUOzguvenHN. Nonlinear dynamic modeling of gear-shaft-disk-bearing systems using finite elements and describing functions. J Mech Des2004; 126: 534–541.
38.
ChenTGuestrinCXGBoost. In: Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining, 2016, pp. 785-794. New York, NY: ACM.
39.
FriedmanJH. Greedy function approximation: a gradient boosting machine. Ann Stat2001; 29: 1189–1232. DOI: 10.1214/aos/1013203451
40.
MohrilRSSolankiBSKulkarniMS, et al. XGBoost based residual life prediction in the presence of human error in maintenance. Neural Comput Appl2023; 35: 3025–3039.
41.
HanYKimJEnkeD. A machine learning trading system for the stock market based on N-period Min-max labeling using XGBoost. Expert Syst Appl2023; 211: 118581.
42.
JiangHHeZYeG, et al. Network intrusion detection based on PSO-Xgboost model. IEEE Access2020; 8: 58392–58401.
43.
AsselmanAKhaldiMAammouS. Enhancing the prediction of student performance based on the machine learning XGBoost algorithm. Interact Learn Environ2021; 31: 3360–3379.
44.
AntonopoulosCG. Network inference combining mutual information rate and statistical tests. Commun Nonlinear Sci Numer Simul2023; 116: 106896.
45.
LiSZhangX. Research on orthopedic auxiliary classification and prediction model based on XGBoost algorithm. Neural Comput Appl2020; 32: 1971–1979.
46.
EberhartRKennedyJ. New optimizer using particle swarm theory. In: Proceedings of the international symposium on micro machine and human science, 1995, pp.39–43. New York: IEEE.
47.
LeeC-TLeeC-C. On a hybrid particle swarm optimization method and its application in mechanism design. Proc IMechE, Part C: J Mechanical Engineering Science2014; 228: 2844–2857.
48.
SongKYanFDingT, et al. A steel property optimization model based on the XGBoost algorithm and improved PSO. Comput Mater Sci2020; 174: 109472.
49.
ChenQXuFLiuP, et al. Research on fractal model of normal contact stiffness between two spheroidal joint surfaces considering friction factor. Tribol Int2016; 97: 253–264.
