The thesis focuses on the development and testing of a numerical model aimed at reproducing and optimizing the behaviour of the Italian Day-Ahead electricity market. The goal is to develop a unit commitment simulator suitable to conduct two main analyses: firstly, to carry out reasonable predictions regarding the number of equivalent hours of programmable generators and pumped-hydro storage units on a yearly basis, while providing a forecast of the price of electricity; secondly, to optimize the capacity planning of new generators and storage units under different RES penetration scenarios in a mid-term perspective. First, the evolution of the Italian electricity system is described, highlighting how the recent abrupt penetration increase of non-programmable renewables caused a attening of the average MGP hourly price curve and a significant reduction of the electricity produced by combined cycles and pumped-hydro storage plants. Starting from a literature review, the Unit Commitment model is then formulated as a MILP, classifying generators by technology on a zonal basis, and by accurately including technical constraints such as ramp-rates, start-up and shutdown times and reserve. The model, implemented in IBM-ILOG and solved by CPLEX, is compared with a well-known commercial simulation tool, Plexos, showing similar results. A sensitivity analysis regarding the optimal operation of Pumped Hydro Storage units shows how their number of equivalent hours can change significantly depending on their mark-up cost. Finally, the Unit Commitment model is included in a two-level capacity planning methodology: first, the technology and the capacity to be installed are selected among a given portfolio; then, a simplified UC model based on typical weeks optimizes the yearly dispatch and evaluates the electricity mix. This approach is aimed at estimating the least cost evolution of the the Italian generation eet and production mix in the next years in the hypothesis of two main scenarios: (a) Business As Usual; (b) High renewable penetration.
La tesi si concentra sullo sviluppo e il test di un modello numerico mirato a riprodurre e ottimizzare il comportamento del mercato elettrico italiano del giorno prima. Lo scopo ottenere un simulatore del processo di dispacciamento adatto a condurre le due seguenti analisi: in primo luogo, ottenere stime ragionevoli riguardanti il numero di ore equivalenti su base annuale per generatori a produzione programmabile e unità di pompaggio idroelettrico, fornendo inoltre una previsione del prezzo dell'elettricità; in secondo luogo, si vuole ottimizzare l'istallazione di nuovi generatori e unità di accumulo nel medio termine in diversi scenari di penetrazione di fonti rinnovabili. Si descrive inizialmente l'evoluzione del settore elettrico italiano, evidenziando come il recente e rapido incremento della produzione da fonti rinnovabili non programmabili abbia causato un appiattimento della curva oraria di prezzo sull'MGP e una signficativa riduzione dell'utilizzo di impianti a ciclo combinato e pompaggio idroelettrico. Partendo da una sintesi della letteratura presente, il modello di Unit Commitment viene quindi formulato come MILP, classificando i generatori su base tecnologica e zonale, e includendo accuratamente i vincoli tecnici quali tempi di rampa, di accensione e spegnimento e margini di riserva. Il modello, implementato in IBM-ILOG e risolto con CPLEX, viene quindi comparato con un noto software commerciale, Plexos, fornendo risultati simili. Un'analisi di sensitività riguardante la gestione ottimale degli impianti di pompaggio idroelettrico mostra come il loro numero di ore equivalenti cambi significativamente al variare del loro mark-up. Infine, il modello di Unit Commitment viene incluso in un metodo per la pianificazione della capacità a due livelli: nel primo, la tecnologia e la capacità da istallare vengono selezionate tra una serie di possibilità; nel secondo, un modello di UC semplificato, basato sulle settimane tipo per ogni stagione, ottimizza il dispacciamento annuale e valuta il mix di produzione. Questo approccio mira a stimare l'evoluzione di minor costo per la flotta di generazione italiana e per il mix produttivo dei prossimi anni nell'ipotesi di due scenari principali: (a) Business As Usual; (b) Alta penetrazione rinnovabile.
Development of a unit commitment model for the simulation of the day-ahead electricity market and capacity planning in Italy
GIANOTTI, GIANLUCA
2015/2016
Abstract
The thesis focuses on the development and testing of a numerical model aimed at reproducing and optimizing the behaviour of the Italian Day-Ahead electricity market. The goal is to develop a unit commitment simulator suitable to conduct two main analyses: firstly, to carry out reasonable predictions regarding the number of equivalent hours of programmable generators and pumped-hydro storage units on a yearly basis, while providing a forecast of the price of electricity; secondly, to optimize the capacity planning of new generators and storage units under different RES penetration scenarios in a mid-term perspective. First, the evolution of the Italian electricity system is described, highlighting how the recent abrupt penetration increase of non-programmable renewables caused a attening of the average MGP hourly price curve and a significant reduction of the electricity produced by combined cycles and pumped-hydro storage plants. Starting from a literature review, the Unit Commitment model is then formulated as a MILP, classifying generators by technology on a zonal basis, and by accurately including technical constraints such as ramp-rates, start-up and shutdown times and reserve. The model, implemented in IBM-ILOG and solved by CPLEX, is compared with a well-known commercial simulation tool, Plexos, showing similar results. A sensitivity analysis regarding the optimal operation of Pumped Hydro Storage units shows how their number of equivalent hours can change significantly depending on their mark-up cost. Finally, the Unit Commitment model is included in a two-level capacity planning methodology: first, the technology and the capacity to be installed are selected among a given portfolio; then, a simplified UC model based on typical weeks optimizes the yearly dispatch and evaluates the electricity mix. This approach is aimed at estimating the least cost evolution of the the Italian generation eet and production mix in the next years in the hypothesis of two main scenarios: (a) Business As Usual; (b) High renewable penetration.| File | Dimensione | Formato | |
|---|---|---|---|
|
2016_04_Gianotti.pdf
non accessibile
Descrizione: Testo della tesi
Dimensione
3.59 MB
Formato
Adobe PDF
|
3.59 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/10589/134420