Prediction of Mechanical Properties of Cold Rolled Steel Using Fusion of Multi-Head Attention

IR > 计算机与通信学院

	Prediction of Mechanical Properties of Cold Rolled Steel Using Fusion of Multi-Head Attention
	Zhang, Qiwen1 ; Guo, Rongping 2
	2024
发表期刊	IAENG International Journal of Computer Science
ISSN	1819-656X
卷号	51 期号:1 页码:45-54
摘要	—Cold rolling of steel is a complex nonlinear process, and there exist intricate spatiotemporal correlations among its process parameters, chemical composition, and mechanical properties. However, traditional long short-term memory networks cannot distinguish the importance of different input features and often only consider the temporal information during feature extraction, thus neglecting the spatial associations of the feature data. To better extract features from cold-rolled steel data and enhance the model’s prediction accuracy, in this paper a fusion of improved multi-head attention modules is proposed. These modules enable the adaptive allocation of weights to different influencing factors, allowing the accurate distinction of the importance of different features. Additionally, a feed-forward neural network is introduced into the memory to further explore and extract the spatiotemporal features of the data. Experimental results on an industrial dataset demonstrate that the proposed model outperforms five other advanced models in terms of prediction performance. © (2024), (International Association of Engineers). All Rights Reserved.
关键词	Cold rolled steel Data mining Feedforward neural networks Forecasting Long short-term memory Metal cladding Steel sheet Chemical compositions Cold rolled steel Feed forward neural net works Index terms Index term—spatiotemporal correlation Multi-head attention Nonlinear process Prediction of mechanical properties Process parameters Spatiotemporal correlation
收录类别	EI
语种	英语
出版者	International Association of Engineers
EI入藏号	20240815598501
EI主题词	Cold rolling
EI分类号	535.1 Metal Rolling ; 535.1.2 Rolling Mill Practice ; 545.3 Steel ; 723.2 Data Processing and Image Processing
原始文献类型	Journal article (JA)
EISSN	1819-9224
文献类型	期刊论文
条目标识符	https://ir.lut.edu.cn/handle/2XXMBERH/169972
专题	计算机与通信学院
通讯作者	Guo, Rongping
作者单位	1.School of Computer and Communication, Lanzhou University of Technology, Lanzhou; 730050, China; 2.School of Computer and Communication, Lanzhou University of Technology, Lanzhou; 730050, China
第一作者单位	兰州理工大学
通讯作者单位	兰州理工大学
第一作者的第一单位	兰州理工大学
推荐引用方式 GB/T 7714	Zhang, Qiwen,Guo, Rongping. Prediction of Mechanical Properties of Cold Rolled Steel Using Fusion of Multi-Head Attention[J]. IAENG International Journal of Computer Science,2024,51(1):45-54.
APA	Zhang, Qiwen,&Guo, Rongping.(2024).Prediction of Mechanical Properties of Cold Rolled Steel Using Fusion of Multi-Head Attention.IAENG International Journal of Computer Science,51(1),45-54.
MLA	Zhang, Qiwen,et al."Prediction of Mechanical Properties of Cold Rolled Steel Using Fusion of Multi-Head Attention".IAENG International Journal of Computer Science 51.1(2024):45-54.