|
ON-LINE ESTIMATION OF UNMEASURED INPUTS FOR ANAEROBIC PROCESS DESCRIBED BY INTERPOLATED L.T.I. MODELS
D. Theilliol(*), C. Aubrun(*), J-C Ponsart(*) and J. Harmand(¤)
(*) Centre de Recherche en Automatique de Nancy - CNRS UMR 7039
BP 239- 54506 Vandoeuvre Cedex-France.
(¤) Laboratoire de Biotechnologie de l’Environnment - INRA
Avenue des Etangs - 11100 Narbonne - France.
email: didier.theilliol@cran.uhp-nancy.fr
Phone: + 33 383 912 701 - Fax: + 33 383 912 030
A method for unknown input estimation in nonlinear stochastic system is presented. A key problem in bioprocess systems is the absence, in some cases, of reliable on line measurements for real time monitoring applications. In this paper, a software sensor for an anaerobic digester is presented. Unmeasured components of the influent are estimated from available on line measurements. Based on a multiple model scheme, a bank of unknown input Kalman filters are discussed to estimate a probabilistic weighting state and unknown input of the process. The performances of the method are tested in simulation using a validated model of an anaerobic fixed bed pilot plant.
Keywords: Interpolated L.T.I. models, unknown inputs, Kalman filter, multiple model algorithm, nonlinear systems, biological process.
Session slot T-Mo-A09: Observer Design for Bioprocesses/Area code 7d : Control of Biotechnological Processes
 | 