The research theme of this thesis work is the techno-economic optimization of integrated utility systems, Rankine cycles and Heat Exchanger Networks (HEN) for given process heat sources/sinks, aimed at devising solutions with improved efficiency, flexibility and economic performance. A general and systematic synthesis methodology has been developed to optimize simultaneously the utility systems, Rankine cycles and HENs at the nominal condition (“single-period”) or considering different off-design operating conditions (“multi-period”). The great advantage of the simultaneous optimization model is that it allows to generate several options for Rankine cycle designs (i.e., selection of number of pressure levels, components and mass flow rates, with fixed pressure and temperatures) and for heat inte-gration with the process heat sources/sinks, while considering the techno-economic con-straints (and the off-design efficiency, in the multi-period optimization) at the same time. From a modelling and mathematical programming point of view, the major contribution of this work are the general superstructure for complex utility systems and Rankine cycles, and the ad-hoc bilevel decomposition method, specifically developed to address this class of problems. The method combines three state-of-the-art relaxation techniques (outer-approximation linearization, McCormick relaxation, and adaptive piecewise linearization of the cost functions) with two original additions: the inclusion of heat cascade constraints to tighten the problem relaxations and the use of nested integer cuts to speed up the con-vergence. For the considered class of problems, the ad-hoc methodology performs way better than general-purpose solvers (such as BARON) and a proposed heuristic algorithm. Indeed, the results over a set of test cases show that the computational time of the bilevel decomposition method is up to 2 orders of magnitude shorter than the heuristic approach. The method is effective in solving extremely challenging problems, as it has been success-fully applied to problems with up to 35 streams, 11,700 variables (2,840 binaries) and 17,000 equations. The proposed methodology has been applied to relevant case studies, related to the design of efficient power plants and industrial processes, such as heat recovery ORCs, NGCCs, IGCCs, ISCCs and poly-generation plants.
Il tema di ricerca della tesi riguarda l’ottimizzazione tecno-economica di utilities, cicli Rankine e reti di scambiatori di calore integrati per un dato processo, con l’obiettivo di trovare delle soluzioni efficienti, flessibili ed economiche. Una metodologia generale e sistematica è stata sviluppata per l’ottimizzazione simultanea di utilities, cicli Rankine e reti di scambiatori di calore per la condizione nominale di pro-getto (“single-period”) e per diverse condizioni operative previste (“multi-period”). Il vantaggio del metodo di ottimizzazione simultaneo è che permette di valutare molteplici opzioni per il design dei cicli Rankine (scelta del numero di livelli di pressione, dei componenti e delle portate del ciclo, considerando pressioni e temperatue fissate) e per l’integrazione termica con le sorgenti/i pozzi di calore di un dato processo, considerando allo stesso tempo i vincoli tecno-economici del problema (e l’efficienza in diverse condizioni operative, nel caso multi-period). Dal punto di vista matematico e modellistico, i maggiori contributi della tesi sono la super-struttura generale per la modellizzazione di utilities complesse e cicli Rankine, e l’algoritmo di decomposizione bilivello, sviluppato specificamente per risolvere la classe di problemi trattata. L’algoritmo combina tre tecniche di rilassamento di vincoli nonlineari (outer-approximation linearization, McCormick relaxation, e adaptive piecewise lineariza-tion) con due aggiunte innovative: l’inclusione di vincoli di “heat cascade” per rinforzare il rilassamento e l’uso di integer cut annidati per migliorare la velocità di convergenza. Per la classe di problemi trattata, la metodologia proposta ha mostrato performance superiori rispetto a risolutori globali (come BARON) e ad un metodo euristico proposto. I risultati indicano che il tempo computazionale del metodo di decomposizione bilivello è fino a 2 ordini di grandezza inferiore rispetto a quello richiesto dal metodo euristico. Il metodo è stato applicato efficacemente per la risoluzione di problemi di grande scala fino a 35 flus-si, 11,700 variabili (2,840 binarie) e 17,000 equazioni. La metodologia proposta è stata applicata a casi di studio rilevanti per il design di impianti di potenza efficienti e processi industriali avanzati, come cicli ORC a recupero, cicli com-binati, IGCC, cicli combinati integrati con fonte solare e impianti di poli-generazione.
Systematic optimization of heat exchanger networks with integrated thermodynamic cycles
ELSIDO, CRISTINA
Abstract
The research theme of this thesis work is the techno-economic optimization of integrated utility systems, Rankine cycles and Heat Exchanger Networks (HEN) for given process heat sources/sinks, aimed at devising solutions with improved efficiency, flexibility and economic performance. A general and systematic synthesis methodology has been developed to optimize simultaneously the utility systems, Rankine cycles and HENs at the nominal condition (“single-period”) or considering different off-design operating conditions (“multi-period”). The great advantage of the simultaneous optimization model is that it allows to generate several options for Rankine cycle designs (i.e., selection of number of pressure levels, components and mass flow rates, with fixed pressure and temperatures) and for heat inte-gration with the process heat sources/sinks, while considering the techno-economic con-straints (and the off-design efficiency, in the multi-period optimization) at the same time. From a modelling and mathematical programming point of view, the major contribution of this work are the general superstructure for complex utility systems and Rankine cycles, and the ad-hoc bilevel decomposition method, specifically developed to address this class of problems. The method combines three state-of-the-art relaxation techniques (outer-approximation linearization, McCormick relaxation, and adaptive piecewise linearization of the cost functions) with two original additions: the inclusion of heat cascade constraints to tighten the problem relaxations and the use of nested integer cuts to speed up the con-vergence. For the considered class of problems, the ad-hoc methodology performs way better than general-purpose solvers (such as BARON) and a proposed heuristic algorithm. Indeed, the results over a set of test cases show that the computational time of the bilevel decomposition method is up to 2 orders of magnitude shorter than the heuristic approach. The method is effective in solving extremely challenging problems, as it has been success-fully applied to problems with up to 35 streams, 11,700 variables (2,840 binaries) and 17,000 equations. The proposed methodology has been applied to relevant case studies, related to the design of efficient power plants and industrial processes, such as heat recovery ORCs, NGCCs, IGCCs, ISCCs and poly-generation plants.File | Dimensione | Formato | |
---|---|---|---|
2019_09_PhD_Elsido.pdf
non accessibile
Descrizione: Testo della tesi
Dimensione
3.18 MB
Formato
Adobe PDF
|
3.18 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/10589/148438