Coordinating multiple model predictive controllers for multi-reservoir management

Modeling and coordinated control of large interconnected multi-purpose reservoirs spread across a country is usually challenging. These systems are generally characterized by coupled water quantity and quality variables and multiple control objectives. Efficient control of these systems can be achieved through appropriate and accurate modeling of reservoir dynamics in tandem with an accurate weather forecasting system. In these systems, the process variables are influenced by various physical and chemical phenomena occurring at different time scales. Also the water quality and quantity control objectives may be in conflict posing a difficult control problem. For such large-scale systems, Model Predictive Control (MPC) is an attractive control strategy and can be implemented in centralized or decentralized configurations. However, it has been shown that to achieve a flexible and reliable control structure with optimum overall system operations, individual decentralized controllers have to be coordinated and driven towards the performance of a centralized controller. In this work, two coordination strategies that have been reported in the literature viz. cooperation based coordination and price driven coordination are evaluated for controlling a network of reservoirs. These algorithms are evaluated on the basis of their robustness, stability and performance in comparison to that of a centralized MPC implementation. Ability to deal with a variety of model uncertainties and the coordination between the controllers within and across a hierarchy are important aspects that have been evaluated.