(270f) Robust Global Stabilization of Continuous Bioreactors

This work studies the problem of designing controllers for enlarging the stability region of continuous stirred microbial bioreactors (chemostats) by manipulating the dilution rate. A particular application of interest is the control of methanogenesis in anaerobic digestion, where the specific growth rate follows the Andrews kinetics (substrate inhibition), and the objective is to robustly stabilize the operation of the bioreactor, while maximizing methane production.

At conditions of maximal methane production, the corresponding steady state of the anaerobic digester is locally stable, but the stability region is too small to allow proper operation in the presence of disturbances. Thus, the need for control arises in the sense of enlargement of the stability region of the optimal steady state.

Earlier work by the authors (2004 DYCOPS, Boston, MA) developed a control Lyapunov function methodology for designing a globally stabilizing control law in continuous stirred microbial bioreactors. In the case of anaerobic digestion, under the assumption of continuous on-line measurement of the methane production rate, this led to a proportional output feedback controller with an appropriate choice of controller gain.

The present work will study sampling issues in the digital implementation of the globally stabilizing control law. In particular, it will be shown that:

i) global stability is preserved under sampling, even though large sampling periods result in loss of performance.

ii) the control law is robustly stable under errors in the kinetics, but sensitive to errors in the organic load of the bioreactor.

iii) incorporation of feedforward action can significantly enlarge the robustness margin.