基于光谱信息与温度的转炉炼钢终点预测模型

转炉炼钢终点就是在出钢时通过控制使钢水的成分和温度达到一定要求。原子发射光谱法显示的元素特征光谱波长下的光谱信息的变化，能反应成分含量的变化趋势。文章介绍了基于温度和光谱信息的终点预测模型的系统，通过此系统获取转炉炼钢过程中的温度和Mn、Si光谱信息，利用多元回归分析的方法建立终点预测模型，并对模型进行显著性检验和精度分析。研究表明，此模型的终点命中率在80%左右，能比较准确预测转炉炼钢的终点。如果在预测模型中加入更多的元素光谱信息，将能进一步提高终点命中率。

Prediction model of converter steelmaking endpoint based on spectral information and temperature

Converter steelmaking endpoint is that the composition and temperature of the molten steel reaches a certain requirement by controlling while steel tapping. The changes of spectral information at elemental characteristic spectral wavelength displayed by atomic emission spectrometry can reflect the variation trend of ingredient content. In this article, a endpoint prediction model system based on temperature and spectral information is described. The temperature and Mn, Si spectral information in converter steelmaking process can be obtained through this system. The endpoint prediction model is established by using multivariate regression analysis, and the significance test and accuracy analysis of this model is done. Studies show that the hit rate of endpoints in this model is about 80%, and the converter steelmaking endpoints can be accurately predicted. If more elements spectral information is added into the prediction model, the hit rate of endpoint prediction will be further improved.