Observer and controller design for a methane bioconversion process

摘要

Methane is currently an emerging alternative feedstock for biological processes. In this paper, we study a particular methane to lactate bioconversion process based on the bacterium Methylomicrobium buryatense 5GB1, with the aim of designing suitable state estimators and controllers for the process. First, a nonlinear unsegregated, unstructured model, consisting of six dynamic mass flow balance equations and six state variables, is constructed and implemented as a dynamic simulator in MATLAB and Simulink. Simulation results match qualitatively with experimental data from literature. Second, an observability analysis is performed with the aim of finding suitable sensor configurations for state estimation. For each of the resulting observable configurations a linear Kalman filter is constructed, of which the performance is evaluated. This performance is not satisfactory and, therefore, extended and unscented Kalman filters are designed to overcome the nonlinear nature of the system. Third, a controllability analysis is executed to establish controllability with respect to the process inputs. Both the observability and controllability give an assessment of practical observability and controllability by using empirical Gramians. A linear quadratic regulator (LQR) with integral feedback is constructed to control the biomass and lactate concentrations. The results for a combined observer and controller scheme are promising and give reliable simulation results. (C) 2020 European Control Association. Published by Elsevier Ltd. All rights reserved.