论文部分内容阅读
针对高炉料面煤气流分布难于检测的问题,提出了一种基于多源信息的高炉料面煤气流分布识别方法。在对高炉机理分析的基础上,确定了高炉中几个重要的过程参量。同时采用区域划分的方法对十字测温数据进行处理,并提取3个主要位置的特征值。根据布料模型建立O/C计算模型,并提取O/C特征指数。最后综合多种检测信息,运用自组织神经网络对高炉料面典型煤气流分布形态进行识别。实验结果表明:本文提出的基于多源信息的识别方法,识别率比传统方法提高了5%～10%,也避免了误识别的情况。
Aiming at the problem of gas flow distribution in blast furnace surface being difficult to detect, a gas flow distribution identification method based on multi-source information is proposed. Based on the analysis of the blast furnace mechanism, several important process parameters in the blast furnace were determined. At the same time, the method of zonation is used to deal with the cross temperature data and extract the eigenvalues of the three main locations. The O / C calculation model is established according to the cloth model and the O / C characteristic index is extracted. Finally, based on a variety of testing information, a self-organizing neural network is used to identify the typical gas flow patterns in blast furnace surface. The experimental results show that the recognition method based on multi-source information proposed in this paper can improve the recognition rate by 5% -10% compared with the traditional method and avoid the misidentification.