The quadruple-tank process: modeling and controller design
MetadataShow full item record
AbstractThe main focus of this thesis work is to model and control multivariable laboratory system that consists of four interconnected water tanks. The quadruple-tank process is a typical coupled control process with two inputs and two outputs. This type of dynamic flow control process can be easily found in chemical distillation systems and other cascaded electro-mechanical systems. The objective of the process is to adjust the two bottom tank water levels by controlling two pumps. Therefore the goal of the thesis is to explore various alternative control strategies to solve this process control problem and evaluate the merits and drawbacks of these control strategies to this control problem. In this thesis work, physical modeling of the multi variable process, experimental system identification using black box models for simulation and control purposes are presented. The detailed control structure design procedures of PI control, Internal Model Control (IMC), Model Predictive Control (MPC), and Linear Quadratic Regulator Optimal Control (LQROC) are developed, extended and applied to the multivariable system. The real time simulation results of all the controllers are presented in comparison of set point tracking performance and robustness to response to the disturbance of a sudden flow rate change. Experimental verification has been carried out by using PC interface through Wincon 3.2 real-time platform. Through this thesis work, it has been found that the PI controller presents the fastest response, the LQR optimal controller has the smallest steady state error and the MPC controller has both good properties of PI and LQR controller.
Bibliography: p. 123-128