波形识别技术在漏钢预报中的应用  

Application of waveform recognition technology in breakout prediction

作　　者：武冬蕾[1] 赵辉[1] 

机构地区：[1]天津理工大学自动化与能源工程学院,天津300191

出　　处：《天津理工大学学报》2005年第5期25-28,共4页Journal of Tianjin University of Technology

基　　金：天津市自然科学基金重点资助项目(033803311);天津市教委科技发展基金资助项目(20041705)

摘　　要：介绍板坯连铸结晶器拉坯阻力在线监测的功率法原理,在对漏钢事故前拉坯阻力时域异常特征分析总结的基础上,利用线性最小二乘法进行数字曲线的线性逼近,通过分段识别的方法进行波形识别,达到提取拉坯阻力时域异常特征的目的.最后介绍了波形识别算法及其在连铸漏钢预报上的应用,现场试验数据证明了算法的可靠性.This Paper introduces the theory of measuring mould friction on-line with power of mould oscillation motor. After analyze the abnormal time-region characteristics in the friction curve before breakout, the linear least square method is used to get the linear approxmation for digital curves. The method identification and segmentation of digital curves is used to recognize primitive combinations represcning the breakout. At last, a the realization of waveform recognition algorithm is introduced, which is verified by tests.

关 键 词：漏钢预报 波形识别 最小二乘法 

分 类 号：TP806.3[自动化与计算机技术—检测技术与自动化装置] TF341.6[自动化与计算机技术—控制科学与工程]