Model Predictive Control and water resource systems operation: challenges and opportunities
Jakob Luchner
DEIB PhD Student
DEIB - Seminar Room
March 28th, 2017
3.00 pm
Research Line:
Planning and management of environmental systems
DEIB PhD Student
DEIB - Seminar Room
March 28th, 2017
3.00 pm
Research Line:
Planning and management of environmental systems
Abstract
Model Predictive Control (MPC) is a well-established real-time control strategy for multivariable systems subject to constraints on decision variables and outputs. MPC has a long application history in the process industry. In recent years, MPC applications have expanded to other fields as, for example, the operation of water resources systems. Water resources systems are often characterized by multiple control structures, operation targets, and water users. MPC proved very efficient in solving such large-scale systems with numerous constraints in real-time. The challenges for MPC applications to water resources systems originate from potentially highly non-linear system dynamics, uncertainties related to numerical weather prediction, streamflow prediction and future water demand over different time scales, and multiple objectives. This presentation will provide an overview of opportunities and challenges of using MPC strategies for multipurpose water resource systems operation. Multipurpose reservoir operations often need to satisfy targets with fast dynamics (e. g. flood control, hydropower production) and slow dynamics (e. g. irrigation supply). These different operational time scales are characterized by different temporal resolutions, different degrees of uncertainty and risk aversion. A multi-scale MPC approach assimilating forecast information with different lead times and temporal resolutions could potentially provide efficient, real-time operating rules for such multipurpose water resources systems. In particular, the presentation will focus on challenges of such a multi-time layer MPC approach in the water resources context. The seminar is based on the Introductory course on MPC delivered by Prof. A. Bemporad at IMT Lucca.