论文部分内容阅读
针对基于动态主元分析的故障检测方法存在的主元个数较多以及计算效率低等问题,本文提出基于混合动态主元分析(Hybrid Dynamic Principal Component Analysis,HDP-CA)的复杂过程故障检测方法。该方法采用分步策略消除数据之间的自相关和互相关性,提高了故障检测的精度和效率。对TE过程典型故障和热连轧过程中断带故障检测结果表明:HDPCA方法提取的主元个数少于DPCA方法提取的主元个数。并且,基于HDPCA的T2和SPE统计量的检测性能和检测精度都由于基于DPCA的统计量。因此,本文提出的方法可以准确有效地检测出故障。
In order to solve the problems of fault detection based on dynamic principal component analysis (PCA), such as a large number of principal components and low computational efficiency, this paper proposes a method of complex process fault detection based on Hybrid Dynamic Principal Component Analysis (HDP-CA) . The method uses a step-by-step strategy to eliminate the autocorrelation and cross-correlation between the data and improve the accuracy and efficiency of fault detection. The results of the fault detection of the typical TE process and the interruption process of the hot rolling process show that the number of principal components extracted by the HDPCA method is less than the number of principal components extracted by the DPCA method. Moreover, both the detection performance and the detection accuracy of HDPCA-based T2 and SPE statistics are due to the DPCA-based statistics. Therefore, the proposed method can detect faults accurately and effectively.