转炉炼钢中的炉口火焰线性回归分析

【作者】 张晓光；

【导师】 赵琦；

【作者基本信息】 南京理工大学 ， 光学工程， 2008， 硕士

【摘要】 提出了将转炉炉口火焰光强及图像信息相结合,用于转炉炼钢的终点判断的新方法。构建了相应的实验研究系统。该系统主要包括炉口火焰光强信息采集、炉口火焰图像信息采集、炉口火焰光强及图像数据分析、处理软件等几部分。利用望远系统以及数字摄像系统构建了炉口火焰图像采集系统,利用光纤谱分复用和DirectShow技术设计了炉口辐射多频道信息获取系统,选择了特定的火焰图像数据处理方法。对所涉及的硬件(炉口火焰图像采集、光学系统设计、数据提取),软件(炉口火焰图像特征值分析)系统进行调试后用于现场试验。通过对实验数据分析并发现,光强与图像特征信息在炼钢吹炼过程中存在着初期逐步变大、中期剧烈振荡,临近终点时两者出现了相反的变化趋势。在此基础上,运用多元回归方法建立起不同炉次终点时刻的数学模型。模型的显著性检验结果为:样本决定系数为0.98,F检验值为416.19远大于临界值2.36,表明模型显著且拟合程度好。模型的实验预测精度为91.7%,适合转炉在线终点控制和快速出钢模型的要求。更多还原

【Abstract】 To judge steelmaking end-point of the converter online, a new method based on the unifying of the flame radiation intensity and the image information from converter mouth is presented in this dissertation. And the corresponding experimental system has been constructed. This system is mainly consist of a multi-frequency information acquisition system,which was designed based on multi-single spectrum by use of optical fiber and DirectShow. The experiment results show that the light intensity and the image characteristic information have the remarkable varying tendency in the steelmaking blowing process. Based on this phenomenon, mathematical model of the end-point time had been established by use of the regression analysis method. The significance test results show that the sample determination coefficient is 0.98 and the F-test value is 416.19 bigger than the critical value 2.36. Therefore the model has good fitting degree and the regression equation is significant. The prediction precision of the model is 91.7%, and it suits the converter end-point control online and the fast tapping model request.更多还原