Abstract:
This thesis focuses on the modeling, identification, and control of a reverse osmosis (RO) desalination
system. In response to the growing scarcity of freshwater resources, RO technology offers a
viable and sustainable solution. The first phase involves the development of a dynamic model
based on experimental data, accurately capturing the interactions between key variables such as
feed pressure, pH, permeate flow rate, and conductivity. Two control strategies are explored :
Model Predictive Control (MPC), implemented on both decoupled and multivariable models,
and classical PID control, including an improved IMC-PID version. The results obtained highlight
the performance, robustness, and limitations of each control approach under model uncertainties
