烧结混合料粒度在线识别研究与应用
Research and application of intelligent water distribution technology for sintering based on particle size identification
续飞飞 1王晖 2安秀伟 3张涛 1赵宏博 1李永杰 1袁中洁1
作者信息
- 1. 北京智冶互联科技有限公司 北京 100144
- 2. 中冶集团 北京 100028
- 3. 青岛特殊钢铁有限公司 山东 青岛 266400
- 折叠
摘要
烧结过程中混合料配水的稳定控制对生产的稳定优质有至关重要的影响.本文基于烧结加水混料制粒的第一性原理,采用平均灰度处理、频率域法增强等技术进行了图像增强,利用Canny算子实现图像的边缘检测分割以及Halcon集合运算与区域变换算子进行处理,从而得到接近真实颗粒形态的轮廓特征,解决了混合料粒度识别的难题.通过试验对比发现试验物理筛分与图像识别结果两者偏差均在5%以内,基本可以表征混合料的整体特征;同时,验证了实时监测结果对加减水过程混合料粒度数据变化较为敏感,且对混合料加减水过程有很好的跟随性,为下一步基于混合料粒度识别的自动配水奠定了坚实基础.
Abstract
The stable control of mixing water distribution in sintering process has a vital influence on the stable and high quality of production.Based on the first principle of sintering and water mixing granulation,this paper adopts average gray level processing,frequency domain method enhancement and other technologies to carry out image enhancement processing.Canny operator is used to realize image edge detection and segmentation,Halcon set operation and region transformation operator for processing,so as to obtain contour features close to the real particle shape.The problem of particle size identification of mixture is solved.It was found that the difference between the physical screening and image recognition results was within±5%,which could basically represent the overall characteristics of the mixture.At the same time,it is verified that the real-time monitoring results are sensitive to the changes of the particle size data of the mixture during the water addition and subtraction process,and have a good followability for the automatic water distribution based on the identification of the particle size of the mixture in the next step.
关键词
烧结/粒度识别/混合料/高速工业相机Key words
sintering/particle size identification/supervised learning/intelligent water distribution/closed-loop control引用本文复制引用
出版年
2024
NETL
NSTL
万方数据