ENBIS: European Network for Business and Industrial Statistics

Statistical monitoring of control loops performance

23 September 2009, 10:25 – 10:45

Abstract

Submitted by

Marco P. Seabra dos Reis

Authors

Tiago M. Rato Marco S. Reis

Affiliation

University of Coimbra

Abstract

In this paper we address the problem of monitoring the performance of automatic control loops (APC) using data collected from the process. This topic has gaining importance in recent times, has a significant amount of the loss of quality and efficiency in process operations, can be traced back to malfunctions located at the level of controllers actions. The proposed index is a generalization of an historical data benchmark index [1] and presents less variation and fewer false alarms while maintaining its ability to detect the controller’s performance deterioration. It is also able to identify the affected cycles and load variables involved through control charts procedures. The results were compared to the ones obtained using some of the methods presented in the literature, supporting such claims.
The proposed index was also tested in a more realistic system, consisting of a reactor equipped with a heating jacket. The results show that the proposed index maintained its ability to detect changes in the controller’s performance about 100% of the times when they are due to perturbations occurred in the discharge coefficient. When the changes happened in the heat transfer coefficient, the index detected the disturbance 40% to 100% of the times depending on the disturbance magnitude.
The proposed index is easy to calculate and only requires the measurement of the controlled variables over time. It can also be considered a good alternative compared to other indexes found in the literature. The results obtained illustrate the potential of the approach and highlight its simple and effective structure, two important requirements for being implemented in real world scenarios.

1. Yu, J. and S.J. Qin, Statistical MIMO controller performance monitoring. Part II: Performance diagnosis. Journal of Process Control, 2008. 18: p. 297–319.

Return to programme