Recursive model predictive control strategy for multiple cascaded systems: Application to supervisory flood management of the Big Karun River

This paper proposes a model predictive control approach for the supervisory management of a multiple cascaded system using a new recursive strategy. Modeling in terms of linear state-space equations enables the application of a predictive control structure in the presence of physical and safety constraints, delay time, and prediction data. In this way, the local subsystems are completely supervised by a decision-making support system using real-time and past data from different branches. The implementation of a comprehensive and reliable flood management system for the urban Big Karun River and its watershed in Iran has been proposed in this paper as a practical application of such a strategy. The main strategic objective to be achieved is to restrain the flood before the main river and save quality water behind the cascade dams with the least waste of reserves. As a validation of the performance, a potential flooding rainfall scenario in the presence of prediction error has been simulated for the cascaded reservoir model of the six-dam water system of the case study.