Studying the Use of ANN to Estimate State-Space Variables for MIMO Systems in a NMPC Strategy

The integration of new technologies in advanced control systems continues to be a current challenge today, as control devices gain in performance and new alternatives to existing techniques are offered. This is based on the fact that the techniques from Intelligent Computing allow the development of efficient and robust control systems, without the need to fully understand the dynamics of the plant. On the other hand, the use of predictive control is a reality within the world of the process industry. It is widely used in the industrial environment due to its simplicity and robustness. The objective of this work is to study and assess the possibility of using neural models to reproduce the dynamics of complex multivariable systems, together with their applicability in control strategies that use optimization processes to generate the appropriate control actions and considering operational constraints, as is the case of the nonlinear MPC. In this work, the results obtained along the this study will be presented, where it is analyzed how to include said neuronal model in a previously established non-linear MPC strategy.