Browse by author
Lookup NU author(s): Shallon Stubbs, Dr Jie ZhangORCiD
Full text for this publication is not currently held within this repository. Alternative links are provided below where available.
There have been much reported successes with the use of state space models for process identification, control and monitoring of dynamic processes with several different approaches to deriving the state variables and a few variants of the state space model representation having been documented over the years. Typically the form of state-space model adopted is one that requires the estimation of five matrices to fully parameterize the model. This paper proposes a simplification of the state-space model for the specific purpose of process monitoring. The simplification of the representation achieved via a modified definition of the past vector of inputs and output, facilitates a simpler and more efficiently estimation of a reduced set of state space matrices. The proposed approach in conjunction with a devised filtering method gives improved fault detection performance over the established state space monitoring methods and other multivariate statistical methods.
Author(s): Stubbs S; Zhang J; Morris J
Publication type: Conference Proceedings (inc. Abstract)
Publication status: Published
Conference Name: 19th European Symposium on Computer Aided Process Engineering
Year of Conference: 2009
Pages: 339-344
Publisher: Elsevier BV
Library holdings: Search Newcastle University Library for this item
ISBN: 9780444534330
