Abstract
Quadruple Tank Process (QTP) is a multivariable system consisting of four inter connected tanks and two control valves. The quadruple tank system has Minimum Phase (MP) or Non-Minimum Phase (NMP) configurations depending on the two control valve openings. In the QTP system, sensors measure the lower tank levels, while the upper tank levels are not directly measured. State estimation algorithms are employed to monitor the liquid level in four tanks of the QTP. As QTP exhibits multivariable interaction and non-linear dynamics, state estimation is crucial for control. This paper investigates separately the nonlinear state estimation methods for a quadruple tank system under ideal and process uncertainty conditions using Extended Luenberger Observer (ELO) and Extended Kalman Filter (EKF) algorithms. In this paper, the state estimation for a quadruple tank system is carried out for Minimum and Non-Minimum Phase configurations. The ELO and EKF are employed under ideal circumstances and 10%, 20%, and 30% uncertainty are added to the parameters, as changes in the area of drain in tank 1 (a1) and tank 3 (a3). Mean Square Error (MSE), Integral Square Error (ISE) and Integral Absolute Error (IAE) are used to evaluate the state estimator's performance in terms of its accuracy in state tracking and its ability to process uncertainty. Simulation results show that under ideal circumstances, EKF and ELO offer good state estimates; however, EKF performs better when there are uncertainties in the process.
DOI: 10.61416/ceai.v27i4.9412