Workshop InCO4In

El prop贸sito del workshop es dar a conocer los avances y resultados m谩s relevantes聽 obtenidos en los campos del modelado, control y optimizaci贸n en el proyecto “Control y Optimizaci贸n de Planta Completa Integrado para Industria 4.0” (InCO4In).

Se desarrollar谩 online los d铆as 20 y 21 de junio de 2022 en el horario de 12:00 a 14:30 y de 11:00 a 13:30, respectivamente. Tambi茅n participar谩n dos investigadores externos que han colaborado en el proyecto e impartir谩n dos conferencias. Para inscribirse, use el siguiente formulario antes del viernes 17 de junio a las 12h.

Agenda

20 de junio: MODELADO Y SIMULACI脫N

12:00 a 12:10: Presentaci贸n y apertura del Workshop.

12:10 a 12:30: Construcci贸n sistem谩tica de modelos grises usando regresi贸n con restricciones: J.L. Pitarch.

12:30 a 12:50: Optimizaci贸n de la esterilizaci贸n t茅rmica de alimentos considerando variabilidad entre productos: C. Vilas.

12:50 a 13:10: Co-simulaci贸n de una planta de esterilizaci贸n usando FMI: S. Gal谩n.

13:10 a 13:20: Pausa

13:20 a 13:40: Simulaci贸n predictiva de la red de H2 de una refiner铆a de petr贸leo: G. Guti茅rrez.

13:40 a 14:00: 聽Desarrollo y operaci贸n de un gemelo digital en una planta de MDF: C. de Prada.

14:00 a 14:30: Advances in Algorithms for Large-Scale Stochastic Optimization: V铆ctor Zavala, College of Engineering, University of Wisconsin-Madison, zavalatejeda@wisc.edu

Abstract: In this talk, we review recent advances in the modeling and solution of large-scale stochastic optimization problems. Our work is motivated by applications in the real-time scheduling of energy systems, which requires handling of mixed-integer variables and uncertainties that span multiple timescales. We show how to manage these complexities using structure-exploiting formulations and cutting-plane algorithms and demonstrate that significant improvements in performance are possible over deterministic formulations.

21 de junio: CONTROL PREDICTIVO, SCHEDULING Y OPTIMIZACI脫N

11:00 a 11:30: Robust Dynamic Real-time Optimization for Chemical and Energy Processes: Lorenz T. Biegler, Carnegie Mellon University, Pittsburgh, PA, biegler@cmu.edu

Model Predictive Control (MPC) is widely accepted in the process industries and energy systems as a generic multivariable controller with constraint handling. More recently, MPC has been extended to Nonlinear Model Predictive Control (NMPC) in order to realize high-performance control of highly non- linear processes. In particular, NMPC allows incorporation of detailed process models (validated by off-line analysis) and also integrates with on-line optimization strategies consistent with higher-level tasks, such as scheduling and planning. NMPC for tracking and so-called 鈥渆conomic鈥 stage costs have been developed, 聽and fundamental stability and robustness properties of NMPC have been analyzed. This talk provides an overview of NMPC and eNMPC concepts and approaches, as well as the underlying optimization strategies that support the solution strategies. In addition, several process case studies are presented to demonstrate the effectiveness of these approaches.

11:30 a 11:50: Scheduling en l铆nea de la secci贸n de esterilizaci贸n en una planta conservera: C. de Prada.

11:50 a 12:10: Scheduling de crudos con incertidumbre en una refiner铆a de petr贸leo: T. Garc铆a

12:10 a 12:20: Pausa

12:20 a 12:40: M茅todos de descomposici贸n para problemas de optimizaci贸n estoc谩stica multietapa: D. Montes.

12:40 a 13:00: Operaci贸n eficiente de un secadero industrial de fibras mediante NMPC: J.L. Pitarch.

13:00 a 13:20: MPC econ贸mico usando modelos con errores estructurales, E. Oliveira.

13:20 a 13:30: Conclusiones y fin del Workshop.