信息熵特征选择方法定位热轧带钢头部拉窄原因
Cause analysis of head width narrow of hot rolled strip based on feature selection of information entropy
-
摘要: 针对某热轧带钢生产线时常会出现头部拉窄现象,本文从实际生产数据出发,首先采用动态弯曲时间算法将长度不同的各变量转换为相同长度,然后采用信息熵特征选择算法分析各生产过程变量对产品质量的影响程度,找出对头部拉窄影响最大的前若干个过程变量,初步定位热轧带钢头部拉窄原因.结果表明:信息熵特征选择方法可以有效初选热轧带钢头部拉窄原因,为生产调整提供方法支撑,将来还可推广到其他质量异常定位,甚至其他工序或者其他应用领域.Abstract: The quality problem of head width narrow appears on one hot rolled strip line sometime. In order to using the real data,in this paper the dynamic time warping was firstly utilized to transfer the different length data to the same length,and then the feature selection method based on information entropy was applied to analyze the order of important about process parameters that effect on the quality. The first several process parameters were selected as the causes of head width narrow of hot rolled strip. The results show that the feature selection method based on information entropy can track the causes of the head width narrow of hot rolled strip effectively to adjust the process. In the future,the methods can be applied in other quality diagnosis,other processes or other application fields.