Jamrozik Wojciech, Górka Jacek, Kik Tomasz
Department of Fundamentals of Machinery Design, Silesian University of Technology, Konarskiego Str. 18a, 44-100 Gliwice, Poland.
Department of Welding Engineering, Silesian University of Technology, Konarskiego Str. 18a, 44-100 Gliwice, Poland.
Materials (Basel). 2021 Jan 18;14(2):442. doi: 10.3390/ma14020442.
Welding is an important process in terms of manufacturing components for various types of machines and structures. One of the vital and still unsolved issues is determining the quality and properties welded joint in an online manner. In this paper, a technique for prediction of joint hardness based on the sequence of thermogram acquired during welding process is proposed. First, the correspondence between temperature, welding linear energy and hardness was revealed and confirmed using correlation analysis. Using a linear regression model, relations between temperature and hardness were described. According to obtained results in the joint area, prediction error was as low as 1.25%, while for HAZ it exceeded 15%. Future work on optimizing model and input data for HAZ hardness prediction are planned.
就制造各种类型的机器和结构的部件而言，焊接是一个重要的过程。一个至关重要且仍未解决的问题是以在线方式确定焊接接头的质量和性能。本文提出了一种基于焊接过程中获取的热像图序列来预测接头硬度的技术。首先，通过相关性分析揭示并确认了温度、焊接线能量和硬度之间的对应关系。使用线性回归模型描述了温度与硬度之间的关系。根据接头区域获得的结果，预测误差低至1.25%，而热影响区（HAZ）的预测误差超过15%。计划未来开展优化热影响区硬度预测模型和输入数据的工作。
Zotero 插件
