Arif Sari
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Arif Sari has a total of 5 co-authors. Here are the top co-authors by number of shared publications:
Ahmed Ah Alkurdi
Duhok Polytechnic University
1 shared publication- Biju Theruvil Sayed
Department of Computer Science, Dhofar University, Salalah, Oman
1 shared publication Karrar Hatif Mohmmed
Scientific Research Center, Al-Ayen University, Thi-Qar, Iraq
1 shared publicationShavan Askar
Erbil Polytechnic University
1 shared publicationWade Ghribi
Department of Computer Engineering, College of Computer Science, King Khalid University, Al-faraa Campus, Abha, Saudi Arabia
1 shared publication
Latest Publications
- 25 July 2024
- Published by SAGE Publications in Proceedings of the Institution of Mechanical Engineers, Part E: Journal of Process Mechanical Engineering
- Vol. 240 (2) , 1955-1969
Finite element method-enabled machine learning for analysing residual stress and plastic deformation in surface mechanical attrition-treated alloys
This paper presents a novel machine learning model designed to predict residual stress and equivalent plastic deformation in metallic alloys undergoing surface mechanical attrition treatment. The dataset used for training was generated by numerically simulating surface mechanical attrition treatment on various alloys, such as SS316L, NiTi, Ti64, Al7075, and AZ31. The regression analysis of the proposed model exhibits exceptional predictive capabilities, with high R ² values of 0.959 for residual stress and 0.911 for average equivalent plastic strain, alongside low root mean square error values of 0.035 and 0.088, respectively. Furthermore, the detailed examination of the correlation between input features and output targets revealed that the increase in values of residual stress and plastic strain in treated samples corresponded with heightened weight functions of processing parameters and material properties, respectively, within the machine learning model. A case study focusing on Al7075 was also provided, demonstrating the model's ability to adjust parameters effectively to achieve specific surface residual stress and plastic strain outcomes. Ultimately, the proposed model not only serves as a reliable predictor for the output targets but also functions as a valuable tool for characterizing the complex input–output relationships, thereby reducing the need for trial and error experimentation in real-world scenarios.
- 25 July 2024
- Published by SAGE Publications in Proceedings of the Institution of Mechanical Engineers, Part E: Journal of Process Mechanical Engineering
- Vol. 240 (2) , 1955-1969
Finite element method-enabled machine learning for analysing residual stress and plastic deformation in surface mechanical attrition-treated alloys
This paper presents a novel machine learning model designed to predict residual stress and equivalent plastic deformation in metallic alloys undergoing surface mechanical attrition treatment. The dataset used for training was generated by numerically simulating surface mechanical attrition treatment on various alloys, such as SS316L, NiTi, Ti64, Al7075, and AZ31. The regression analysis of the proposed model exhibits exceptional predictive capabilities, with high R ² values of 0.959 for residual stress and 0.911 for average equivalent plastic strain, alongside low root mean square error values of 0.035 and 0.088, respectively. Furthermore, the detailed examination of the correlation between input features and output targets revealed that the increase in values of residual stress and plastic strain in treated samples corresponded with heightened weight functions of processing parameters and material properties, respectively, within the machine learning model. A case study focusing on Al7075 was also provided, demonstrating the model's ability to adjust parameters effectively to achieve specific surface residual stress and plastic strain outcomes. Ultimately, the proposed model not only serves as a reliable predictor for the output targets but also functions as a valuable tool for characterizing the complex input–output relationships, thereby reducing the need for trial and error experimentation in real-world scenarios.