STATISTICAL MONITORING OF MULTIVARIABLE DYNAMIC PROCESSES WITH STATE-SPACE MODELS

Industrial continuous processes may have a large number of process var iables and are usually operated for extended periods at fixed operatin g points under closed-loop control, yielding process measurements that are autocorrelated, cross-correlated and collinear. A statistical pro cess monitoring (SPM) method based on multivariate statistics and syst em theory is introduced to monitor the variability of such processes. The statistical model that describes the in-control variability is bas ed on a canonical-variate (CV) stare-space model that is an equivalent representation of a vector autoregressive moving-average rime-series model. The CV state variables obtained from the state-space model are linear combinations of the past process measurements that explain the variability of the future measurements the most. Because of this disti nctive feature, the CV state variables are regarded as the principal d ynamic directions A T-2 statistic based on the CV state variables is u sed for developing an SPM procedure. Simple examples based on simulate d data and an experimental application based on a high-temperature sho rt-time milk pasteurization process illustrate advantages of the propo sed SPM method.