This thesis addresses the problem of optimizing the electrical configuration of large-scale photovoltaic systems through a multi-objective approach. The primary goal is to identify design solutions that achieve a balance between maximizing energy production and minimizing Capital Expenditures (CAPEX), a critical consideration in the design of utility-scale photovoltaic plants. An optimization framework was developed utilizing the Non-Dominated Sorting Genetic Algorithm II, integrated through a fully automated workflow in which MATLAB manages the optimization, PVsyst performs the energy simulations, and AutoHotkey enables their integration. The model was used to evaluate a comprehensive set of design variables, including module type, layout configuration, number of strings and inverters, as well as geometrical parameters such as tilt and azimuth. Both fixed-tilt and single-axis tracking systems were tested, considering the use of monofacial and bifacial PV modules to allow for a broad exploration of alternative configurations. Two technological scenarios were defined and analyzed, representing different levels of component performance and market prices, in order to assess the impact of technological evolution on system optimization. The methodology was applied to a real photovoltaic project currently under construction in the Campania region, with a nominal power of 26.596 MWp. This thesis was developed in collaboration with the Engineering Division of Edison Spa, which provided technical support and access to reference data for the case study. Simulation results demonstrated that the algorithm was able to identify configurations that outperformed the original system design in terms of both energy yield and cost-effectiveness. The proposed model's validity was confirmed through comparative analysis and the use of the hypervolume indicator, underscoring its reliability and practical applicability in real-world photovoltaic plant design.
La presente tesi affronta il problema dell'ottimizzazione della configurazione elettrica di impianti fotovoltaici di grande scala attraverso un approccio multiobiettivo. L'obiettivo principale è individuare soluzioni progettuali capaci di bilanciare la massimizzazione della produzione energetica con la minimizzazione del costo d'investimento (CAPEX), un aspetto cruciale nella progettazione di impianti fotovoltaici a scala industriale. È stato sviluppato un framework di ottimizzazione basato sul Non-Dominated Sorting Genetic Algorithm II, integrato tramite un flusso di lavoro completamente automatizzato in cui MATLAB gestisce l'ottimizzazione, PVsyst esegue le simulazioni energetiche e AutoHotkey ne consente l'integrazione. Il modello è stato impiegato per valutare un insieme esaustivo di variabili progettuali, tra cui tipologia di modulo, configurazione del layout, numero di stringhe e inverter, nonché parametri geometrici come tilt e azimuth. Sono stati analizzati sia sistemi a inclinazione fissa che a inseguimento monoassiale, considerando l'impiego di moduli fotovoltaici monofacciali e bifacciali per consentire un'ampia esplorazione di configurazioni alternative. Sono stati definiti e analizzati due scenari tecnologici, rappresentativi di diversi livelli prestazionali dei componenti e prezzi di mercato, al fine di valutare l'impatto dell'evoluzione tecnologica sull'ottimizzazione del sistema. La metodologia è stata applicata a un progetto fotovoltaico reale attualmente in fase di costruzione nella regione Campania, con una potenza nominale di 26.596 MWp. La tesi è stata sviluppata in collaborazione con la Divisione Ingegneria di Edison Spa, che ha fornito supporto tecnico e accesso ai dati di riferimento per il caso studio. I risultati delle simulazioni hanno dimostrato che l'algoritmo è stato in grado di individuare configurazioni capaci di superare il progetto originale in termini sia di resa energetica che di convenienza economica. La validità del modello proposto è stata confermata tramite analisi comparative e l'utilizzo dell'indicatore di ipervolume, evidenziandone l'affidabilità e l'applicabilità pratica nella progettazione di impianti fotovoltaici reali.
Multi-objective design optimization of a grid-connected photovoltaic power plant using NSGA-II algorithm
Raeli, Angelo
2024/2025
Abstract
This thesis addresses the problem of optimizing the electrical configuration of large-scale photovoltaic systems through a multi-objective approach. The primary goal is to identify design solutions that achieve a balance between maximizing energy production and minimizing Capital Expenditures (CAPEX), a critical consideration in the design of utility-scale photovoltaic plants. An optimization framework was developed utilizing the Non-Dominated Sorting Genetic Algorithm II, integrated through a fully automated workflow in which MATLAB manages the optimization, PVsyst performs the energy simulations, and AutoHotkey enables their integration. The model was used to evaluate a comprehensive set of design variables, including module type, layout configuration, number of strings and inverters, as well as geometrical parameters such as tilt and azimuth. Both fixed-tilt and single-axis tracking systems were tested, considering the use of monofacial and bifacial PV modules to allow for a broad exploration of alternative configurations. Two technological scenarios were defined and analyzed, representing different levels of component performance and market prices, in order to assess the impact of technological evolution on system optimization. The methodology was applied to a real photovoltaic project currently under construction in the Campania region, with a nominal power of 26.596 MWp. This thesis was developed in collaboration with the Engineering Division of Edison Spa, which provided technical support and access to reference data for the case study. Simulation results demonstrated that the algorithm was able to identify configurations that outperformed the original system design in terms of both energy yield and cost-effectiveness. The proposed model's validity was confirmed through comparative analysis and the use of the hypervolume indicator, underscoring its reliability and practical applicability in real-world photovoltaic plant design.| File | Dimensione | Formato | |
|---|---|---|---|
|
2025_07_Raeli_Tesi_01.pdf
accessibile in internet per tutti
Descrizione: Testo tesi
Dimensione
11.59 MB
Formato
Adobe PDF
|
11.59 MB | Adobe PDF | Visualizza/Apri |
|
2025_07_Raeli_Executive Summary_02.pdf
accessibile in internet per tutti
Descrizione: Testo executive summary
Dimensione
1.74 MB
Formato
Adobe PDF
|
1.74 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/10589/239612