Statistics Seminars: Recent advances in multivariate statistical process control with applications to process systems engineering
22 August 2012 14:00 in CM221
This presentation discusses recent advances on multivariate statistical-based process monitoring. To demonstrate the usefulness of these innovative enhancements, various applications studies to recorded data from a mechanical and data from an industrial process system in the chemical industry.
Commencing with an introduction of the principles of statistical process control, the presentation first motivates the need for a multivariate extension to remove the undesired effect of increased Type II errors. The talk then shows how to monitor a multivariate system when applying a reduced dimensional data representation.
Two recent innovations are then outlined that (i) relate to changes in the variable covariance structure, which may not be detectable using conventional multivariate statistical process control, and (ii) introduces an extended data structure to model industrial manufacturing systems.
Contact email@example.com for more information