摘要
针对复杂生产过程存在着参数众多,而且工艺参数间、产品质量参数间各自呈现复杂的多变量非线性耦合关系,提出基于KPCA的非线性生产过程监控与诊断方法。结合数据重构和基于Cam加权距离的领域选取策略这两种方法,计算出故障指数的变化情况,具体分析引起过程异常的工艺参数。利用冷轧带钢热镀锌生产线的生产数据进行实验,实验表明:KPCA具有处理非线性的能力,且能准确地实现对故障的识别。
According to huge mass of parameters in complicated production process,which shows a complex multivariable nonlinear coupling relation between each processing parameters and product quality parameters,this paper proposed a method based on KPCA nonlinear monitoring and diagnosis.Combined with the data reconstruction and field selection,obtained the function of various process parameters for SPE statistics,specifically analysis process parameters that causes abnormal process.Using data of hot-dip galvanizing prove that this method not only retains the advantages of PCA,also has ability of dealing non-linear parameters,and even could identify fault.
出处
《汽车实用技术》
2015年第10期86-88,92,共4页
Automobile Applied Technology
