An asymptotic observer for batch processes with single biogas measurement

Abstract

In this paper we propose an observer for an example of systems for which the boundary of its domain is unobservable, and all trajectories converge to this boundary. The proposed case study, a bioreactor in batch operating conditions with one simple microbial growth reaction and gas production is standard and largely encountered in practical situations.Comparison of the proposed observer with a classical Luenberger observer has been performed: it shows that this later does not guarantee a convergence with no estimation error.

A positive observer for the chemostat model with biogas measurement

“…Note that the estimation in batch i.e. with D identically null has already been tackled in [16,26] where a new exact observer has been proposed, although the trajectories of the system converge to a subset of the state state that is not observable. Here we consider the "true" chemostat for which we shall assume persistently exciting inputs D.…”

A Robust Asymptotic Observer for Systems That Converge to Unobservable States—A Batch Reactor Case Study

“…Incidentally the design of the state estimate takes advantage of the fundamental reaction invariant property of reaction systems [10]. The conference paper [7] presents a new approach for the design of an asymptotic observer in this context of lack of observability, but that is unfortunately not robust to measurement errors. The present paper develops a robust extension of this first approach.…”