(486i) Performance of OSS and NICN Strategies for Batch Process Monitoring

Statistical process monitoring involves three activities: detection of the out-of-control status, identification of the variable(s) that signal this condition, and diagnosis of the source cause for the abnormal behaviour.

Projections techniques like Principal Component Analysis and Independent Component Analysis have been widely applied to monitor industrial processes. Such techniques use at least two different statistics; one devoted to monitor the deterministic behaviour (D-statistic or Ie2-statistic) and other related to the random or unexplained variations (SPE-statistic).

A different approach (Mason et al.,1997) proposes to carry out both detection and identification tasks in the space defined by the original measured variables (i.e. without performing any variable projection), using only one statistic. The Hotelling's statistic ( -statistic) is first decomposed into the contributions of each measured variable, and these contributions are further inspected to identify the suspicious variables. In this sense, Alvarez et al. (2007) developed the Original Space Strategy (OSS), which is based on a unique decomposition of the -statistic that enhances the identification task.

Recently, Alvarez et al. (2008) presented a new strategy to estimate the influence of a given variable on the -statistic value. In this approach, the contribution of each variable is measured in terms of the distance between the current observation and its Nearest In Control Neighbour (NICN), which is evaluated solving an NLP optimization model.

In this work we deal specifically with identification related tasks and procedures for batch process monitoring. A performance comparative analysis between the NICN approach and the OSS strategy is presented. A benchmark fed-batch penicillin fermentation process (due to Birol et.al, 2002) is used as case study to evaluate the techniques' performance.

References

Alvarez, C.R., A. Brandolin and M.C. Sánchez (2007), Chemometrics and Intelligent Laboratory Systems, 88, 189-196.

Alvarez, C.R., A. Brandolin, M.C. Sánchez and L. Puigjaner (2008), 2008-AICHE Annual Meeting Proceedings.

G. Birol, C. Undey, A. Cinar (2002), Computers and Chemical Engineering, 26 (1), 1553?1565.

Mason, R.L., N.D. Tracy and J.C. Young (1997), Journal of Quality Technology, 29, 396-406.

Keywords: Original variables' space, Fault identification, Batch processes.