Modified non-Gaussian multivariate statistical process monitoring based on the Gaussian distribution transformation. (January 2020)

• Record Type: Journal Article
• Authors: Du, Wenyou; Zhang, Yingwei; Zhou, Wei
• Is Part Of: Journal of process control. Volume 85(2020)
• Page Start: 1
• Page End: 14
• Publication Date: 2020-01
• Subjects: Multivariate statistical process monitoring -- Independent component analysis -- Gaussian distribution transformation -- Electrical fused magnesia furnace; Process control -- Periodicals; Fabrication -- Contrôle -- Périodiques; Process control; Periodicals; Electronic journals; 660.281
• DOI: 10.1016/j.jprocont.2017.12.001
• Languages: English
• ISSNs: 0959-1524

Abstract

Graphical abstract: Highlights: We propose the Gaussian distribution transformation (GDT) based monitoring method. Independent components are first transformed into approximately Gaussian distribution which can be used with all kinds of non-Gaussian distributed data and achieve higher normality than box-cox transformation. The G 2 and G e 2 statistics based on Gaussian distribution transformed data are proposed which can effectively eliminate miss alarms. The ratio part of area above curve (RPAAC) is developed to evaluate the performance of fault detection performance, two types of errors (false alarm and miss alarm) are considered comprehensively in this index. We applied the proposed method for monitoring of electrical fused magnesia furnace (EFMF), the fault can be detected in time at the early stage. Abstract: Independent component analysis (ICA) has been applied for non-Gaussian multivariate statistical process monitoring (MSPM) for several years. As the independent components do not satisfy the multivariate Gaussian distribution, a missed alarm occurs when monitoring with traditional statistics. In this paper, we propose a Gaussian distribution transformation (GDT)-based monitoring method. Independent components are first transformed into approximate Gaussian distributions through the proposed nonlinear mapping. Then, we propose new statistics and their control limits to reduce missed alarms. The proposed method is particularly suitable for slight magnitude fault and early-stage fault detection. The ratio part of the area above the curve (RPAAC) is developed to evaluate the performance in fault detection. The experimental results from a synthetic example show the effectiveness of our proposed method. We also apply our method to monitor an electrical fused magnesia furnace (EFMF), and eruption and furnace wall melt faults can be detected in time.