ETD

Archivio digitale delle tesi discusse presso l'Università di Pisa

Tesi etd-09122019-145415


Tipo di tesi
Tesi di laurea magistrale
Autore
PELAGAGGE, FEDERICO
URN
etd-09122019-145415
Titolo
Applicazione di algoritmi di controllo avanzato e ottimizzazione economica di processi chimici complessi
Dipartimento
INGEGNERIA CIVILE E INDUSTRIALE
Corso di studi
INGEGNERIA CHIMICA
Relatori
relatore Prof. Pannocchia, Gabriele
correlatore Dott. Vaccari, Marco
Parole chiave
  • RTO
  • Optimization
  • Offset-Free
  • Model Predictive control
  • EMPC
  • Economic MPC
  • Advanced control
Data inizio appello
04/10/2019
Consultabilità
Completa
Riassunto
Current advanced control and optimization architectures in process industries, based on a conventional hierarchy of economic optimization and control, can be far from being economically optimal. This thesis analyzes methods to reach and maintain optimal economic performance in industrial processes.
The used offset-free Economic Model Predictive Control (EMPC) technique modifies the economic optimization problem and this modification allows convergence to a point that satisfies the necessary conditions of optimality of the actual process, i.e., the actual optimum of the plant.
A major requirement of this algorithm is that, to estimate the modification required at each iteration, it is necessary to know the real gradients of the system at the current steady-state point, which is not a trivial task.
Therefore, in order to build an EMPC industrially applicable, the algorithms have been enhanced with plant gradient estimation techniques.
Five different modifier estimation techniques have been proposed. Four of them are based on steady-state perturbation methods, while, the last one is based on model identification.
A number of examples have been selected for the application of the developed offset-free EMPC algorithm.
In these examples, the actual process and the nominal MPC models are different and therefore would lead to different optimal equilibria.
As result, all proposed EMPC algorithms converge to the economic optimum of the process under determinate conditions.
File