Partial least squares (PLS) has been developed and established as one of the multivariate statistical process control (MSPC) methods. PLS is designed to identify a parametric regression matrix between the input, or predictor variables and the output, or response variables of the process. With PLS, the regression matrix is determined on the basis of a subset of the predictor variables and thus, PLS is able to reduce the number of variables to be considered. In this paper, an alternative variable reduction method is introduced, which is termed latent variable least squares or LVLS. LVLS identifies the process behavior on the basis of the score models rather than a parametric relationship between the predictor and the response variables. In similar fashion to PLS, LVLS can also be applied to monitor industrial processes. An application study, which relates to a realistic simulation of a fluid catalytic cracking unit (FCCU), is presented to demonstrate the monitoring aspect of LVLS.
|Name||Proceedings of the American Control Conference|
|Publisher||Institute of Electrical and Electronics Engineers|
|Conference||2001 American Control Conference, ACC 2001|
|Period||25/06/01 → 27/06/01|