15th Triennial World Congress of the International Federation of Automatic Control
  Barcelona, 21–26 July 2002 
MODEL BASED PREDICTIVE CONTROL OF RIVER RESERVOIRS
Hartmut Linke* Eckhard Arnold**
* Technische Universität Ilmenau
P.O. Box 100565, D-98684 Ilmenau, Germany
e-mail: Hartmut.Linke@TU–Ilmenau.de
** Fraunhofer Appl. Center for Systems Technology
Am Ehrenberg 8, D-98693 Ilmenau, Germany
e-mail: and@ast.iitb.fhg.de

The operational water management of river reservoirs has to consider different and sometimes competitive objectives, e. g. navigation conditions, hydro-power generation, flood protection and environmental aspects. Despite of problem specific extensions to cover the non-linear process behavior the application of local standard (PI) controllers to each reservoir can fulfill only basic management requirements. MBPC algorithms have the capability to give a wide and time proper exploitation of the active storage of the whole cascade. A two-layer control structure, consisting of a predictive controller in the upper layer and local controllers at each barrage, allows to provide the management decisions under real time conditions. Simulation results for three reservoirs of the river Moselle show the aptitude of the proposed approach.
Keywords: river reservoir systems, model predictive control, distributed-parameter systems
Session slot T-Mo-A21: System Engineering, Management and Control Education/Area code 5a : Large Scale Systems