Statistical methods for control loop performance assessment

Control loop performance assessment (CLPA) has attracted considerable interest in recent years. Several methods for CLPA are proposed by researchers in last decade. Statistical methods are one of most important methods in CLPA. This paper addresses CLPA issue through a comprehensive evaluation of statistical methods to efficiently explore their capabilities to assess for different control loop scenarios. For this purpose, an integrated computational framework will be developed, including an ASPEN environment to simulate debutanizer unit as case study. A control loop performance assessment software package is incoporporated utilizing MATLAB to mimic as an independent supervisory computational platform. To provide an interface link between these two different software environments, an OLE (Object Linking Embedding) protocol has been employed. Extensive simulation test studies have been prepared to explore the performance assessment of each individual control loop in the simulated debutanizer unit against valve stiction, random disturbance as the most probable malfunctions in the study.