Real-time fault tolerant control of a Reverse Osmosis desalination plant based on a hybrid system approach

Many applications of Reverse Osmosis desalination plants (RO plants) require a fault tolerant system, in particular when human life depends on the availability of the plant for producing fresh water. However, RO plants are little studied from the control engineering point of view: modeling, design of control algorithms and real-time experiments are scarcely reported in the literature. The present work is a study on a real RO plant in order to discover possible faults, to analyze potential methods for Fault-Tolerant Control (FTC) and the real-time experimentation. In order to implement model based control, the plant is identified in several operating points. Model Predictive Control (MPC) is used as control law and a hybrid supervisor is proposed to combine different methods, which perform better for different kind of faults. Satisfactory results are obtained for the real-time operation.