Estimation of the level and temperature liquid of a two-tank coupled system subject to disturbances

Abstract

This Thesis examines the application of estimation techniques to determine non-measurable state variables in a two-tank coupled system, a widely used dynamic model in industrial processes. It emphasizes the Luenberger observer algorithm as an effective estimation method. Given the inherent limitations in direct measurement, state estimation is vital for ensuring accurate process control and fault detection. The study encompasses control system theory, the modeling and linearization of the two-tank system, and the use of the Luenberger observer to estimate both measurable and non-measurable variables. Results indicate that while full-state observation delivers precise estimates, partial-state observation provides reliable approximations, highlighting the value of estimation in maintaining robust control under varying conditions and disturbances.