Parallel structure of crayfish optimization with arithmetic optimization for classifying the friction behaviour of Ti-6Al-4V alloy for complex machinery applications

被引:17
作者
Chauhan, Sumika [1 ]
Vashishtha, Govind [1 ]
Gupta, Munish Kumar [2 ,3 ]
Korkmaz, Mehmet Erdi [4 ]
Demirsoz, Recep [4 ]
Noman, Khandaker [5 ]
Kolesnyk, Vitalii [6 ]
机构
[1] Wroclaw Univ Sci & Technol, Fac Geoengn Min & Geol, Grobli 15, PL-50421 Wroclaw, Poland
[2] Opole Univ Technol, Fac Mech Engn, 76 Proszkowska St, PL-45758 Opole, Poland
[3] Graph Era Deemed Be Univ, Dept Mech Engn, Dehra Dun, India
[4] Karabuk Univ, Dept Mech Engn, Karabuk, Turkiye
[5] Northwestern Polytech Univ, Sch Civil Aviat, Xian 710072, Shaanxi, Peoples R China
[6] Sumy State Univ, Dept Mfg Engn Machines & Tools, 116 Kharkivska St, UA-40007 Sumy, Ukraine
关键词
Data acquisition; Intelligent diagnosis; Machine learning; Friction forces; TITANIUM-ALLOY; WEAR; STEEL;
D O I
10.1016/j.knosys.2024.111389
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Intelligent techniques play a vital role in predicting the friction force during the wear of Ti-6Al-4V alloy under different lubricating conditions. The effective assessment of friction forces and lubricating conditions allows for the replacement of the material before catastrophic failure. However, it remains challenging to utilise friction forces under different lubrication conditions to predict the wear through intelligent techniques. In this work, an advanced technique based on artificial intelligence has been proposed to address this issue. Intially parallel structure of crayfish optimization and arithmetic optimization algorithm (PSCOAAOA) is developed to duly address the issues of slow convergence, stucking in local optima and quality of the solution. The PSCOAAOA is further implemented for finding the optimal parameters (regularization parameter and kernel function) of the Support Vector Machine (SVM). The quantitative and qualitative analysis of PSCOAAOA is carried out on CEC2014 benchmark functions to validate its efficacy and robustness. The friction force generated during wear testing under different lubricating conditions is bifurcated into training and test data. Out of which, training data trains the SVM at an optimal combination of parameters. The overall accuracy of the built SVM model is found to be 95.85% with a computation time of 26.85 s.
引用
收藏
页数:20
相关论文
共 33 条
[31]   Macroscopic and nanoscale investigation of the enhanced wear properties of medium-Mn steel processed via room-temperature quenching and partitioning [J].
Yan, Xincheng ;
Hu, Jun ;
Zhang, Xiao ;
Xu, Wei .
WEAR, 2023, 522
[32]   Research on asymmetrical edge tool wear prediction in milling TC4 titanium alloy using deep learning [J].
Yang, Yong ;
Zhao, Xuefeng ;
Zhao, Lei .
MEASUREMENT, 2022, 203
[33]   Evaluation of wear mechanism between TC4 titanium alloys and self-lubricating fabrics [J].
Zhao, Song ;
Meng, Fanyue ;
Fan, Bingli ;
Dong, Yu ;
Wang, Jibo ;
Qi, Xiaowen .
WEAR, 2023, 512-513