Modelización, monitoreo y control en procesos para producción de bioplásticos

TitleModelización, monitoreo y control en procesos para producción de bioplásticos
Publication TypeThesis
Year of Publication2016
AuthorsJamilis M
UniversityUniversidad Nacional de La Plata
Thesis TypePhD
Keywordsbioprocess control observers biotechnology
Abstract

This thesis addresses the control and estimation problem in high cell density processes. As case of study the production of bioplastics via Cupriavidus necator is studied. Novel contributions to process modelling, monitoring and on line optimization are done. With the objective of giving a more realistic and precise process description, control oriented models are developed for the volume variations in high cell density processes. Specially, for those which are fed with concentrated media and in which intracellular products are stored. The models obtained are practical in the sense that there is no dependence with difficult to obtain parameters and are more precise than other models previously reported. For process monitoring, algorithms are developed to estimate the reaction rates of the process. Observers are proposed for each of the phases of the bioplastics production process. The estimation algorithms make use of the previously developed volume models and constitute an important contribution to process monitoring in restrictive conditions, such as the ones found in high cell density processes. Along with the design of the estimators, original contributions are made to error analysis, stability proofs and conditions on the algorithm gains to achieve given decay rates in the errors. Also, experimental validation of the estimators is provided. The on line optimization of the processes is given by the use of extremum seeking controllers whose purpose is to maximize the reaction rates at, a priori, unknown operating points. A novel non linear control scheme is proposed, including a gradient estimation which allows convergence to the optimum of an objective function (the substrate to reaction rate map). For the first time, stability conditions are given for this controller, including unstructured uncertainty. Also, conditions on the design gains are obtained. The proposed control scheme improves previously reported ones in convergence speed and the smoothness of the obtained reaction rates. Moreover, it is based on the use of sensors of feasible application to high cell density processes. The control algorithm is finally applied to the bioplastics production process, obtaining satisfactory results.

URLhttp://sedici.unlp.edu.ar/handle/10915/59639
Research Line: 
Control de sistemas y procesos biológicos
Control and monitoring of bioprocesses
Control y monitoreo de bioprocesos