A multivariate statistical process control procedure for BIAS identification in steady‐state processes

Abstract

in Wiley InterScience (www.interscience.wiley.com).In this article, a multivariate statistical process control (MSPC) strategy, devoted to bias identification and estimation for processes operating under steady-state conditions, is presented. The technique makes use of the D statistic to detect the presence of biases. Besides, it uses a new decomposition of this statistic to identify the faulty sensors. The strategy is based only on historical process data. Neither process modeling nor assumptions about the probability distribution of measurement errors are required. In contrast to methods based on fundamental models, both redundant and nonredundant measurements can be examined to identify the presence of biases. The performance of the proposed technique is evaluated using data-reconciliation benchmarks. Results indicate that the technique succeeds in identifying single and multiple biases and fulfills three paramount issues to practical implementation in commercial software: robustness, uncertainty, and efficiency.

A new approach to estimate variable contributions to Hotelling's statistic

“…If the monitoring statistical technique reveals the process behavior is abnormal, those variable contributions to the inflated T 2 -statistic that are greater than a given control-limit, calculated using the reference population, are identified as suspicious. This strategy has been applied for monitoring batch processes [7] and to identify faulty sensors [8].…”