Artificial intelligence based state observer in polymerization process

Abstract

Observers or state estimators are devices used to estimate immeasurable key parameters that are due to noise, disturbances and mismatch. It is important to identify those variables prior to construct a control system and avoid fault or process disruption. In certain chemical processes, such observer usage produced unsatisfactory results therefore hybrid approached is the appropriate solution. Hybrid observers are combination of two or more conventional observers mainly to enhance the estimator's performance and overcoming their limitations. In advanced cases, Artificial Intelligence algorithm is applied. This paper develops two hybrid observers namely sliding mode and extended Luenberger observers with fuzzy logic for approximating the monomer concentration in a polymerization reactor. It was found that the sliding mode observer- fuzzy combination is better based on noise handling with less oscillation.