Spectral principal component analysis of dynamic process data

Abstract

This article describes principal component analysis (PCA) of the power spectra of data from chemical processes. Spectral PCA can be applied to the measurements from a whole unit or plant because spectra are invariant to the phase lags caused by time delays and process dynamics. The same comment applies to PCA using autocovariance functions, which was also studied. Two case studies are presented. One was derived from simulation of a pulp process. The second was from a refinery involving 37 tags. In both cases, PCA clusters were observed which were characterised by distinct spectral features. Spectral PCA was compared with PCA using autocovariance functions. The performance was similar, and both offered an improvement over PCA using the time domain signals even when time shifting was used to align the phases. r

Process performance analysis in large-scale systems integrating different sources of information

“…Thornhill et al (2002) used the scores of spectral PCA to extract smaller clusters of measurements having the same disturbance.…”

Detecting and isolating multiple plant-wide oscillations via spectral independent component analysis

“…Spectral ICA can be thought of as an extension or improvement of spectral PCA as proposed by Thornhill et al (2002). The disadvantage of spectral PCA is that more than one peak may appear in the spectrum-like principal components (PCs) and different PCs can contain the same peaks, or peaks at the same position but of different sign.…”

“…Oscillations increase variability and can prevent a plant from operating close to optimal constraints so a key requirement of an industrial control engineer is for an automated means for (a) the detection of the presence of a plant-wide oscillation and (b) to give an indication of where to seek the root cause so that maintenance effort can be directed efficiently. Thornhill, Shah, Huang, and Vishnubhotla (2002) have demonstrated that principal component analysis of power spectra (spectral PCA) provides a means of detecting the presence of a plant-wide oscillation. However, little has been done to ଁ This paper was not presented at any IFAC meeting.…”

Detection and isolation of oscillations using the dynamic causal digraph method

“…Also, a variety of multivariate methods, such as principal component analysis [8] and nonnegative matrix factorization [9], have been applied to solve this diagnosis task. A recent trend has been to introduce process information into the diagnosis of plant-wide oscillations.…”