Contribution à la commande prédictive des systèmes à paramètres répartis non linéaires

Abstract : This work deals with the control of nonlinear distributed parameter systems described by nonlinear partial differential equations (PDE). The aim of this study is to extend the use of the model predictive control (MPC) strategy from the classical case where systems are described by ordinary differential equations to the case where systems are described by a class of nonlinear partial differential equations. The interest in the use of MPC strategy is to be able to solve various conduct problems by explicitly handling functioning constraints. We have chosen an approach that consisted in directly considering real problems in process engineering, where many knowledge models are described by strongly nonlinear PDE: a painting drying by infrared radiation (including a real-time application), an autoclave curing of thick composites, a catalytic destruction of volatile organic compounds. The particularity of this study is to propose an internal model predictive control structure leading to the seeked process trajectory from a pre-defined trajectory or a handling constraints generated trajectory. For the optimization problem introduced by the MPC approach, we propose an off-line algorithm (trajectory generation) and an on-line algorithm.
Complete list of metadatas
Contributor : Pascal Dufour <>
Submitted on : Thursday, January 22, 2009 - 2:40:18 PM
Last modification on : Friday, May 17, 2019 - 3:38:40 PM
Long-term archiving on : Saturday, November 26, 2016 - 4:52:25 AM


  • HAL Id : tel-00337724, version 3



Pascal Dufour. Contribution à la commande prédictive des systèmes à paramètres répartis non linéaires. Automatique / Robotique. Université Claude Bernard - Lyon I, 2000. Français. ⟨tel-00337724v3⟩



Record views


Files downloads