Control predictivo hibrido y no lineal

This project tries to develop algorithms for advanced process control, making use of previous results and with reference to real industrial processes. The target is to develop methods for hybrid pedictive control, where a mix of real, integer and logic variables are taken into account, applying it to the hierarchical control of the sugar and section of a beet sugar factory. Here there is a mix of continuous and bath units besides products reciycle being the aim to maximize the production avoiding bottlenecks. In addition, we try to develop algorithms for robust predictive control for non-linear systems with a reduced computational cost.Ademas se plantea el desarrollo de algoritmos de control predictivo robusto para sistemas no-lineales con coste computacional reducido, aimed to control multivariable continous precesses with difficult dynamics or tight specifications, that we will apply to the low range temperature control of the LHC particle accelarator at CERN, and to a chemical reactor. The project includes as well the development of a distributed simulation system usinf the HLA standard and a set of software tools: a Matlab Toolbox for controller design, a compenent library in the simulation language ECOSIMPro, a real time computer under Windows NT and several test in pilot plants.