质量相关的带钢热连轧过程监控

Quality-Related Monitoring and Control in Hot Strip Mill Process

带钢热连轧过程控制是钢铁制造过程极其复杂的过程,近年来随着市场对带钢产品质量要求的日益提高,提高热连轧带钢质量具有广泛的经济和社会效益。为了确保热连轧过程安全运行,同时提高产品质量,有必要对热连轧过程的异常状况或故障进行检测、诊断和消除。以多元统计过程监控技术(MSPM)为理论指导,以主元分析(PCA)和偏最小二乘方法(PLS)为依托,研究和分析了PCA和PLS以及二者与核函数结合构成的核主元分析方法(KP-CA)和核偏最小二乘方法(KPLS)在热连轧机质量相关的故障分析与检测,通过现场数据及实验验证,在厚度质量相关的故障检测与诊断中取得较好的效果。

The problem of detecting,diagnoseing and eliminating the abnormal condition or fault in hot rolling process is dealt with to ensure the safe operation and improve the product quality.The multivariate statistical process monitoring technology（MSPM）,the principal component analysis（PCA）,and the partial least squares（PLS） are introduced and their combination with kernel function,kernel principal component analysis（KPCA） and the kernel partial least squares（KPLS）,are studied in details and applied to the quality-related fault analysis and detection of rolling process.The experiments show the good results of the fault detection and diagnosis in the quality related to the thickness.