Quality classification of ultra-narrow gap welding based on GAF-ResNet

doi:10.1109/IAI59504.2023.10327577

	Quality classification of ultra-narrow gap welding based on GAF-ResNet
	Ma, Peijie; Zhang, Aihua; He, Weilong; Wang, Ping; Ma, Jing
	2023
会议名称	5th International Conference on Industrial Artificial Intelligence, IAI 2023
会议录名称	2023 5th International Conference on Industrial Artificial Intelligence, IAI 2023
会议日期	August 21, 2023 - August 24, 2023
会议地点	Shenyang, China
出版者	Institute of Electrical and Electronics Engineers Inc.
摘要	Ultra-narrow gap welding with flux band-constrained arc is a welding method with high efficiency and low heat input, but its welding process is complex, and it is difficult to realize online prediction of welding quality. The traditional signal feature extraction method can not make full use of time series information. In addition, the use of convolutional neural network training will have the problem of gradient disappearance as the number of network layers increases. Because of the above problems, this paper proposes a welding quality classification prediction method based on GAF-ResNet, carries out ultra-narrow gap welding experiments, and performs model accuracy verification and performance analysis. The one-dimensional time series is encoded into a two-dimensional image by GAF image, which retains the time dependence of the time series. The ResNet network is used to deeply mine the time series information in the image array to ensure the depth of feature extraction. The analysis results show that the experimental results of GASF are better than those of GADF. The accuracy and F1 value of the GAF-ResNet model reached 88.163 % and 88.109 %, respectively. Compared with other modes, the overall performance of the model is better than that of the control group. © 2023 IEEE.
关键词	Computerized tomography - Convolutional neural networks - Extraction - Feature extraction - Quality control - Time series GAF - High-low - Higher efficiency - Narrow gap welding - Quality classification - Resnet - Time series informations - Ultra-narrow gap welding - Welding method - Welding quality
DOI	10.1109/IAI59504.2023.10327577
收录类别	EI
语种	英语
EI入藏号	20235115240121
EI主题词	Network layers
EI分类号	723 Computer Software, Data Handling and Applications - 723.5 Computer Applications - 802.3 Chemical Operations - 913.3 Quality Assurance and Control - 922.2 Mathematical Statistics
原始文献类型	Conference article (CA)
引用统计	无
文献类型	会议论文
条目标识符	https://ir.lut.edu.cn/handle/2XXMBERH/169326
专题	电气工程与信息工程学院
通讯作者	Ma, Peijie
作者单位	Lanzhou University of Technology, College of Electrical and Information Engineering, Lanzhou, 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
推荐引用方式 GB/T 7714	Ma, Peijie,Zhang, Aihua,He, Weilong,et al. Quality classification of ultra-narrow gap welding based on GAF-ResNet[C]:Institute of Electrical and Electronics Engineers Inc.,2023.