IDENTIFICATION AND CONTROL DESIGN ISSUES IN NON-LINEAR DIFFUSION-CONVECTION REACTION PROCESSES A.A. Alonso 1 , J.R. Banga 1 , C.G. Moles 1 ,E. Balsa-Canto 2 1 Process Engineering Group, IIM-CSIC, Eduardo Cabello 6, 36208 Vigo, Spain e-mail:antonio@iim.csic.es 2 CIMNE (International Center for Numerical Methods in Engineering), Gran Capita, s/n, C1-Campus Nord UPC, 08034, Barcelona, Spain e-mail:ebalsa@cimne.upc.es Keywords: Distributed Process Systems, Con- vex Extensions, Sensor/Actuator Placement, In- put Constraints Abstract In this contribution, we will discuss some recent results in state reconstruction and robust con- trol of diffusion-convection-reaction systems. The approach settles its roots on the dissipative na- ture of systems of conservation laws with con- vex extensions and passivity, as it is understood in systems theory. Dissipation ensures the exis- tence of a low dimensional subspace that captures most of the relevant dynamic features of the dis- tributed process system. This representation will be used to design robust and asymptotically con- vergent state observers. Passivity conditions, on the other hand, will be derived by making use of convex extensions to construct general classes of storage functions. Thus, this formalism will constitute the basis to set up robust non-linear control design guidelines. The proposed framework will also be employed to address control implementation issues. In partic- ular, the selection of appropriate sensor/actuator placements and stability preservation under input constraints. 1 Introduction The control of distributed process systems has re- ceived considerable attention by the control com- munity over the last years from both the theoret- ical and application points of view [5]. These sys- tems play a central role in chemical and material processing industries as many of its operations in- volve convection diffusion and reaction phenom- ena. Interesting examples include, to name a few, catalytic reactors, chemical vapor deposition units, crystallization or thermal processing. A widely accepted approach to control dis- tributed process systems relies on a state-space- like representation of the original infinite dimen- sional system through projection of the partial differential equations on appropriate basis func- tion sets. Finite differences, finite elements and spectral decomposition schemes are the most common examples. This structure is then em- ployed to design the complete control scheme - see for instance [6] and [3]. A complementary ap- proach, recently proposed by [2] to develop pas- sive stabilizing controls for distributed process settles its roots on the second law of thermody- namics and passivity, as it is understood in sys- tems theory [11]. The second law in the exergy form gives convexity which in turns provides a general answer to the question of finding Lya- punov function candidates to assess system’s evo-