Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.


Model Predictive Control for cascaded river power plants


Felix Althaus

Master Thesis, FS 08

In this project we consider the cascade of power plants situated in the course of the river Main. For environmental and navigational reasons the authorities impose restrictions on the operations of the power plants such that the water level at the specified points of the river must be kept within tight bounds. Currently implemented controllers work only locally, thus they do not take the interactions between the power plants into account. As a result, small fluctuations of the water level at the beginning of the cascade are amplified to undesired oscillations at the end of it. Previous master theses projects on the application of Model Predictive Control (MPC) to the cascades of power plants showed that MPC surpasses the currently implemented conventional PI controllers due to its enhanced abilities, e.g. explicit constraint handling, taking the interactions between the power plants into account and so on. The cascades investigated in these works consisted of 4 and 5 power plants. The purpose of this work is to develop a Model Predictive Control scheme for the complete cascade of 36 power plants in the course of the river Main. The main challenge in this application is the computational complexity of the controller that stems from the large scale of the cascade. The complexity of the controller is further increased due to the logic statements included in the control model, that are required to achieve another control objective; minimizing the number of control moves, i.e. the number of applied turbine discharge changes, in order to decrease the turbine wear-outs. In a recent work the computational feasibility of MPC for the complete cascade was investigated. It was shown that a centralized MPC is computationally infeasible and hence a decentralized MPC needs to be developed. The project will be accomplished in collaboration with SCIETEC, a company that develops advanced river simulation software, called Floris, and E.ON, the operator of the power plant cascade. This project is mainly composed of two parts. In the first part, conceptual problems of designing a decentralized MPC scheme will be studied based on a simplified Floris model of the river Main. These problems include designing computationally feasible local MPC schemes, development of a criterion to decompose the complete cascade into smaller parts and development of a coordination algorithm between the local MPC schemes to improve the performance. In the second part of the project, a decentralized MPC scheme will be designed and tuned based on a detailed Floris model of the river Main to be provided by SCIETEC.

Supervisors: Prof. M. Morari, Dr. Akin Sahin


Type of Publication:

(12)Diploma/Master Thesis

M. Morari

No Files for download available.
% Autogenerated BibTeX entry
@PhdThesis { Xxx:2008:IFA_3197
Permanent link