In recent years, the energy transition has accelerated significantly. However, the widespread deployment of non-dispatchable renewable energy sources, such as solar and wind, poses several challenges to the stability of the electrical grid: the intermittency of generation can indeed reduce the quality of the supplied energy and generate profound discrepancies between supply and demand. To ensure a high penetration of renewable sources within current energy mixes, it is therefore essential to implement storage technologies capable of accumulating production surpluses and returning them during periods of need. In parallel, the thermal sector is also undergoing a decarbonization process, driven primarily by the electrification of heating and cooling loads. In this scenario, Carnot batteries, and in particular Pumped Thermal Energy Storage (PTES) systems, represent a highly promising technological solution. Unlike pumped-storage hydroelectricity and compressed air energy storage (CAES) systems, they do not have geographical constraints and, compared to electrochemical batteries, they do not depend on critical materials. Therefore, they offer flexible and widespread installation, utilize low-cost materials, and enable large-scale energy storage for varying durations. Although several studies on the sizing and operational management of these storage systems exist in the literature, a significant gap remains regarding the global optimization of the plant. To fill this gap, the present work proposes an innovative bi-level optimization approach. The upper level is represented by a Particle Swarm Optimization (PSO) heuristic algorithm, tasked with identifying the optimal sizing of the components. The lower level, on the other hand, is formulated using a Mixed Integer Linear Programming (MILP) algorithm, aimed at determining the optimal hourly management of energy flows over the horizon of an entire year, with the objective of minimizing operational costs. Applied to a case study, the developed model confirmed the effectiveness of the methodology. The results demonstrate that the proposed storage system allows fora a drastic reduction in the users' dependence on external sources (specifically, the national electrical grid and natural gas supply), thereby guaranteeing substantial economic savings, quantified between 3.58 and 12.99 million euros depending on the analyzed scenario. Through the study of different scenarios, it was demonstrated that the profitability of the system is strongly linked to the simultaneity between generation and load, identifying the tertiary sector as the most profitable context. Furthermore, the sensitivity analysis conducted on different energy markets highlighted how the algorithm is capable of implementing highly complex system management strategies. Finally, the study certified the importance of sector coupling. The system's ability to recover waste heat derived from electrical conversion processes guarantees a considerable coverage of the thermal demand, maximizing the energy independence and economic savings of the entire system.
Negli ultimi anni la transizione energetica ha sperimentato una forte accelerazione. Tuttavia, la massiccia diffusione di fonti rinnovabili non programmabili, come solare ed eolico, pone severe sfide alla stabilità della rete elettrica: l'aleatorietà della generazione può infatti ridurre la qualità dell'energia fornita e generare profonde discrepanze tra domanda e offerta. Per garantire un’elevata penetrazione delle fonti rinnovabili all’interno degli attuali mix energetici, è dunque indispensabile implementare tecnologie di stoccaggio capaci di accumulare i surplus produttivi e restituirli nei periodi di necessità. Parallelamente, anche il settore termico sta affrontando un processo di decarbonizzazione, guidato soprattutto dall’elettrificazione dei carichi per il riscaldamento e raffrescamento. In questo scenario, le batterie di Carnot, e in particolare i sistemi PTES, rappresentano una soluzione tecnologica altamente promettente. A differenza degli impianti di pompaggio idroelettrico e dei sistemi di accumulo basati sull'aria compressa, essi non presentano vincoli geografici e, rispetto alle batterie elettrochimiche, non dipendono da materiali critici. Offrono pertanto un’installazione flessibile e diffusa, sfruttano materiali a basso costo e consentono l’accumulo su larga scala per tempi più o meno lunghi. Sebbene in letteratura siano presenti diversi studi sul dimensionamento e sulla gestione di questi sistemi di accumulo, permane una carenza significativa per quanto riguarda l’ottimizzazione globale dell’impianto. Per colmare tale mancanza, il presente lavoro propone un innovativo approccio di ottimizzazione a doppio livello (bi-level). Il livello superiore è rappresentato da un algoritmo euristico di Particle Swarm Optimization (PSO), incaricato di individuare le taglie ottimali dei componenti. Il livello inferiore è invece formulato mediante un algoritmo di Mixed Integer Linear Programming (MILP), finalizzato a determinare la gestione oraria ottimale dei flussi energetici sull’orizzonte di un intero anno, con l’obiettivo di minimizzare i costi operativi. Applicato a un caso studio, il modello sviluppato ha confermato l'efficacia della metodologia. I risultati dimostrano che il sistema di accumulo proposto permette di ridurre drasticamente la dipendenza delle utenze dalle fonti esterne (nello specifico, la rete elettrica nazionale e la fornitura di gas naturale), garantendo così un risparmio economico notevole, quantificato tra i 3.58 e i 12.99 milioni di euro a seconda dello scenario analizzato. Tramite lo studio di diversi scenari, è stato dimostrato come la redditività del sistema sia fortemente legata alla contemporaneità tra generazione e carico, individuando nel settore terziario il contesto maggiormente remunerativo. Inoltre, l'analisi di sensitività condotta su differenti mercati energetici ha evidenziato come l'algoritmo sia in grado di attuare strategie di gestione del sistema molto complesse. Infine, lo studio ha certificato l'importanza dell'accoppiamento tra settori. La capacità del sistema di recuperare il calore residuo derivante dai processi di conversione elettrica garantisce una copertura del fabbisogno termico considerevole, massimizzando l'indipendenza energetica e il risparmio economico dell'intero sistema.
Ottimizzazione a doppio livello di sistemi energetici combinati con dispositivi di accumulo termico
PARELLI, ANDREA
2025/2026
Abstract
In recent years, the energy transition has accelerated significantly. However, the widespread deployment of non-dispatchable renewable energy sources, such as solar and wind, poses several challenges to the stability of the electrical grid: the intermittency of generation can indeed reduce the quality of the supplied energy and generate profound discrepancies between supply and demand. To ensure a high penetration of renewable sources within current energy mixes, it is therefore essential to implement storage technologies capable of accumulating production surpluses and returning them during periods of need. In parallel, the thermal sector is also undergoing a decarbonization process, driven primarily by the electrification of heating and cooling loads. In this scenario, Carnot batteries, and in particular Pumped Thermal Energy Storage (PTES) systems, represent a highly promising technological solution. Unlike pumped-storage hydroelectricity and compressed air energy storage (CAES) systems, they do not have geographical constraints and, compared to electrochemical batteries, they do not depend on critical materials. Therefore, they offer flexible and widespread installation, utilize low-cost materials, and enable large-scale energy storage for varying durations. Although several studies on the sizing and operational management of these storage systems exist in the literature, a significant gap remains regarding the global optimization of the plant. To fill this gap, the present work proposes an innovative bi-level optimization approach. The upper level is represented by a Particle Swarm Optimization (PSO) heuristic algorithm, tasked with identifying the optimal sizing of the components. The lower level, on the other hand, is formulated using a Mixed Integer Linear Programming (MILP) algorithm, aimed at determining the optimal hourly management of energy flows over the horizon of an entire year, with the objective of minimizing operational costs. Applied to a case study, the developed model confirmed the effectiveness of the methodology. The results demonstrate that the proposed storage system allows fora a drastic reduction in the users' dependence on external sources (specifically, the national electrical grid and natural gas supply), thereby guaranteeing substantial economic savings, quantified between 3.58 and 12.99 million euros depending on the analyzed scenario. Through the study of different scenarios, it was demonstrated that the profitability of the system is strongly linked to the simultaneity between generation and load, identifying the tertiary sector as the most profitable context. Furthermore, the sensitivity analysis conducted on different energy markets highlighted how the algorithm is capable of implementing highly complex system management strategies. Finally, the study certified the importance of sector coupling. The system's ability to recover waste heat derived from electrical conversion processes guarantees a considerable coverage of the thermal demand, maximizing the energy independence and economic savings of the entire system.| File | Dimensione | Formato | |
|---|---|---|---|
|
Andrea_Parelli.pdf
embargo fino al 17/04/2027
Dimensione
45.07 MB
Formato
Adobe PDF
|
45.07 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/107555