A MSPC Technique for Identifying Biases in Industrial Processes

CEDEÑO MARCO; GALDEANO RUBÉN; RODRIGUEZ AGUILAR LEANDRO; ELWART JUAN; SANCHEZ MABEL

Salt Lake City

Congreso; 2010 Annual AIChE Meeting; 2010

American Institute of Chemical Engineers

The use of Multivariate Statistical Process Control (MSPC) strategies, devoted to bias identification and estimation for processes operating under steady-state conditions, was addressed by Sanchez et al. (2008). The technique makes use of the T2 statistic to detect the presence of biases. Besides it employs a new decomposition of this statistic to identify the faulty sensors (Alvarez et al, 2007). 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 non-redundant 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, called Original Space Strategy, succeeds in identifying single and multiple biases, and fulfils three paramount issues to practical implementation in commercial software: robustness, uncertainty and efficiency. Later on Alvarez et al. (2008) proposed a new strategy to estimate the influence of a given variable on the final value of the inflated statistic’s 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 Neighbor (NICN). The detection and identification capabilities of the NICN technique are evaluated and compared with those corresponding to the most commonly used systematic error detection and identification techniques for some benchmarks (Cedeño et al., 2009). As results indicated the technique succeeds in identifying single and multiple biases, it was applied for monitoring a petrochemical process. In this work, a comparison between the identification capabilities of both aforementioned strategies for its application to industrial processes, is presented. Also a discussion is provided about the problems that arise during its implementation and how they are solved.