Abstract - Kosanovich The manufacturing cost of chemicals can be decreased by improving the operating performance of chemical reactors. One way to accomplish this is through timely control of the operating variables. For batch systems, this is important to maintaining the quality of the final product. This research is to develop new robust, feedback control formulations for electrochemical batch processes that satisfy multiple performance objectives in the face of model uncertainties and disturbances. The formulation is general and can be applied to other batch systems as well. Specifically, this project consists of three studies on new fundamental approaches to the modelling and control of batch systems: (1) development of a time-varying state space model of the non-stationery batch process for use in a model predictive control strategy; (2) system identification using multi-resolution (wavelet) analysis; and (3) formulation of a controller strategy for a multi-resolution representation of the process. The research will include and experimental demonstration of the concepts on an electrochemical batch system. Effort will also be devoted to the development of effective computational procedures. The planned career advancement activities include six months of rigorous study in the methodologies, two summer months in a cooperative industrial program, and four months devoted to educational material preparation and experimental demonstration on a laboratory scale electrochemical batch process. The methodologies that are relevant to the goals of this study are (1) multi-resolution theory, (2) model predictive and optimal control theory; and (3) theory of electrochemical systems. The experimental portimn of the activities include a cooperative program that provides hands-on experience in the operations of systems at an industrial site and collaboration with department colleagues on electrochemical systems. Other activities include active participation in workshops and seminars, both do mestic and international, assistance of a post-doctoral research fellow, and development of a graduate control course that combines theory with experiments on batch systems.
