This thesis involves application of a novel, online, model-based optimization algorithm to both industrial processes and home appliances, with the aim of insuring optimal operating conditions as well as minimizing environmental impact. Modeling dynamic systems through transport equations is fundamental to describe processes and represent them mathematically. Assumptions and constitutive laws must be considered. This approach leads to systems composed by ordinary differential equations (ODEs) and differential algebraic equations (DAEs). This operation allows simulating the dynamic response of the process. Dynamic optimization algorithms have been improved considerably in the last years. The target is to explore the ability of systems to track dynamic set points and to react to interferences and disturbances. In this thesis we have analyzed, applied, and tested an optimization algorithm, named CSMBO&C (Continuous Simultaneous Model-Based Optimization & Control), which provides both NMPC (Non-linear Model Predictive Control) and DRTO (Dynamic Real-Time Optimization) functionalities. These two different approaches have been demonstrated on as many application scenarios. The models for the cases studied, analyzed in this thesis, have been implemented in C++. The numerical engine, used in all of these tests, is the BzzMath library© (Copyright 1991-2009 by Guido Buzzi-Ferraris - All rights reserved). All the simulations, which have been carried out, suggest that CSMBO&C can efficiently and promptly reject various types of disturbances and perturbations, can be easily applied to different types of processes and can solve several categories of online optimization and control problems.
Questa tesi prevede l'applicazione di un nuovo algoritmo di ottimizzazione online basato su modelli matematici sia per i processi industriali che domestici, con l'obiettivo di assicurare condizioni operative ottimali e ridurre al minimo l'impatto ambientale. La modellizzazione dei sistemi dinamici tramite equazioni di trasporto è fondamentale per descrivere i processi e rappresentarli matematicamente. Assunzioni e leggi costitutive devono essere introdotte. Questo approccio porta a sistemi composti da equazioni differenziali ordinarie (ODE) ed equazioni algebriche differenziali (DAE). Questa operazione consente di simulare la risposta dinamica del processo. Gli algoritmi di ottimizzazione dinamica sono stati notevolmente migliorati negli ultimi anni. L'obiettivo è esplorare la capacità dei sistemi di tracciare i set point dinamici e reagire alle interferenze ed ai disturbi. In questa tesi abbiamo analizzato, applicato e testato un algoritmo di ottimizzazione, denominato CSMBO&C (Continuous Simultaneous Model-Based Optimization & Control), che fornisce funzioni NMPC (Non-linear Model Predictive Control) e DRTO (Dynamic Real-Time Optimization). Questi due diversi approcci sono stati dimostrati in altrettanti scenari applicativi. I modelli per i casi studiati, analizzati in questa tesi, sono stati implementati in C ++. Il motore numerico, utilizzato in tutti questi test, è la libreria BzzMath© (Copyright 1991-2009 di Guido Buzzi-Ferraris - Tutti i diritti riservati). Tutte le simulazioni, che sono state eseguite, suggeriscono che CSMBO&C può respingere in modo efficiente e tempestivo vari tipi di disturbi e perturbazioni, può essere facilmente applicato a diversi tipi di processi e può risolvere diverse categorie di problemi di ottimizzazione e controllo online.
Biphasic continuous stirred tank reactor and domestic water heater optimal control
CARUSO, ROBERTO;VALSECCHI, STEFANO
2017/2018
Abstract
This thesis involves application of a novel, online, model-based optimization algorithm to both industrial processes and home appliances, with the aim of insuring optimal operating conditions as well as minimizing environmental impact. Modeling dynamic systems through transport equations is fundamental to describe processes and represent them mathematically. Assumptions and constitutive laws must be considered. This approach leads to systems composed by ordinary differential equations (ODEs) and differential algebraic equations (DAEs). This operation allows simulating the dynamic response of the process. Dynamic optimization algorithms have been improved considerably in the last years. The target is to explore the ability of systems to track dynamic set points and to react to interferences and disturbances. In this thesis we have analyzed, applied, and tested an optimization algorithm, named CSMBO&C (Continuous Simultaneous Model-Based Optimization & Control), which provides both NMPC (Non-linear Model Predictive Control) and DRTO (Dynamic Real-Time Optimization) functionalities. These two different approaches have been demonstrated on as many application scenarios. The models for the cases studied, analyzed in this thesis, have been implemented in C++. The numerical engine, used in all of these tests, is the BzzMath library© (Copyright 1991-2009 by Guido Buzzi-Ferraris - All rights reserved). All the simulations, which have been carried out, suggest that CSMBO&C can efficiently and promptly reject various types of disturbances and perturbations, can be easily applied to different types of processes and can solve several categories of online optimization and control problems.| File | Dimensione | Formato | |
|---|---|---|---|
|
2018_04_Caruso_Valsecchi.pdf
solo utenti autorizzati dal 05/04/2019
Descrizione: Testo della tesi
Dimensione
5.93 MB
Formato
Adobe PDF
|
5.93 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/139725