连铸结晶器黏结漏钢的可视化及其识别方法 |
| |
摘 要: | 基于现场实测热电偶温度数据,将结晶器铜板温度及其变化速率进行2维可视化。综合考虑黏结形成和发展过程中结晶器铜板温度随时间的变化和空间传播行为,借助计算机图像处理中的8连通区域标记和边界跟踪算法,提取了异常区域的结晶器铜板温度、位置、时间等信息,计算了结晶器铜板温度变化速率的均值、最大值、区域的面积、宽度、高度及其纵向和横向移动速率特征,对实际浇铸过程中多例漏钢样本的共性特征进行了统计和归纳。结果表明,结晶器铜板温度变化速率、几何特征与传播速率能够作为黏结漏钢在线预报的重要判据,为提高漏钢预报系统准确率提供参考,减少漏钢事故,同时为结晶器可视化、智能化监控技术开发提供方法和依据。
|
Visualization and recognition method of sticker breakout of mold during continuous casting process |
| |
Authors: | LIU Yu WANG Xu- dong YAO Man ZHANG Xiao- bing |
| |
Affiliation: | 1. School of Materials Science and Engineering, Dalian University of Technology, Dalian 116024, Liaoning, China ;2. Jiangsu Shagang Group Co. , Ltd. , Zhangjiagang 215625, Jiangsu, China |
| |
Abstract: | Based on the measured temperature of thermocouple, the two- dimensional temperature and its change velocity at copper plate of mold during continuous casting process were visualized. The propagation behaviors of the temperature at copper plate of mold in time and space were considered synthetically during the formation and development of sticker breakout. The temperature, position and time at abnormal zone were extracted by virtue of 8 connected component labeling and boundary tracking algorithm in computer image processing. The average and maximum change velocities of the temperature as well as area, width, height, characteristics of vertical and horizontal propagation velocities of the zone were also calculated. The common features of samples of sticker breakout were analyzed by statistical induction during continuous casting process. The results show that the change velocity of the temperature, geometry characteristic and propagation velocity can be used as main criterions for online prediction of sticker breakout and are helpful to enhance the accuracy of prediction system of sticker breakout which can be used to decrease sticker breakout. Meanwhile, they offer the method and basis to develop a visual and intelligent monitoring technology for mold. |
| |
Keywords: | continuous casting mold sticker breakout visualization recognition method |
本文献已被 CNKI 万方数据 等数据库收录! |
| 点击此处可从《钢铁研究学报》浏览原始摘要信息 |
|
点击此处可从《钢铁研究学报》下载全文 |
|