STATISTICAL PROCESS MONITORING AND DISTURBANCE DIAGNOSIS IN MULTIVARIABLE CONTINUOUS-PROCESSES

Detecting out-of-control status and diagnosing disturbances leading to the abnormal process operation early are crucial in minimizing produc t quality variations Multivariate statistical techniques are used to d evelop detection methodology for abnormal process behavior and diagnos is of disturbances causing poor process performance. Principal compone nts and discriminant analysis ave applied to quantitatively describe a nd interpret step, ramp and random-variation disturbances. All disturb ances require high-dimensional models for accurate description and can not be discriminated by biplots. Diagnosis of simultaneous multiple fa ults is addressed by building quantitative measures of overlap between models of single faults and their combinations. These measures are us ed to identify the existence of secondary disturbances and distinguish their components. The methodology is illustrated by monitoring the Te nnessee Eastman plant simulation benchmark problem subjected to differ ent disturbances. Most of the disturbances can be diagnosed correctly, the success rate being higher for step and vamp disturbances than ran dom-variation disturbances.