In the last decades the energy world has been facing an intense primary energy source transition, from the fossil fuels to the renewable energy sources (RES). Just as the fossil fuel based energy industry relies on reserves, the renewable sector depends upon the assessment of resources. One of the main fields of renewable energy is related to solar energy, which is exploited in different ways, where photovoltaic (PV) systems, concentrating solar power systems, building energy systems, direct heating systems are some of the more assessed application. In order to evaluate and predict the performance of solar radiation based systems over medium-long time, given the unpredictability of the meteorological conditions, what is considered a typical period is used, known as Typical Meteorological Year (TMY). It is a yearly dataset that includes solar radiation values and meteorological data on hourly basis. In this thesis a novel procedure for TMY generation is presented, which modifies the classical methodology to account for the improved resolution of the dataset. Using this methodology a TMY for the location of Milano Bovisa is assembled. Before the application of the procedure, all meteorological data are processed by a quality control algorithm, in order to include in the analysis only measurements of proven reliability. An overview of rejected measurements and the issues faced in the data validation process are also reported. The TMY generated is later employed in Photovoltaic (PV) power forecast by means of machine learning technique. Also a classical version of TMY, being the web-available generated by Photovoltaic Geographical Information System (PVGIS) is applied to foresee PV power output. Forecast performances are then compared in perfect forecast condition, in order to assess the effectiveness of the TMY employment in PV power prediction. Moreover, further investigations are carried out about the TMY data structure influence over forecast.
Negli ultimi decenni il mondo energetico sta affrontando un'intensa transizione della fonte energetica primaria, spostandosi dai combustibili fossili alle fonti energetiche rinnovabili (RES). Uno dei principali settori delle energie rinnovabili è legato all'energia solare: sistemi fotovoltaici (PV), sistemi a concentrazione solare, sistemi energetici degli edifici, sistemi di riscaldamento diretto; sono solo alcune delle applicazioni più diffuse che si basano sulla radiazione solare come fonte energetica primaria. Al fine di analizzare e prevedere le prestazioni di sistemi basati sull’energia solare nel medio-lungo periodo, data l'imprevedibilità delle condizioni meteorologiche, in molte analisi viene utilizzato un periodo rappresentativo delle condizioni locali, chiamato anno meteorologico tipico (TMY). Si tratta di un insieme di dati aggregato in modo da costituire un anno intero, che include valori di radiazione solare ed altre variabili meteorologiche su base oraria. In questa tesi è presentata una nuova procedura per la generazione di TMY, che modifica la metodologia classica per tenere conto di una elevata risoluzione temporale dei dati. Utilizzando questa metodologia viene quindi assemblato un TMY per la località di Milano Bovisa. Tutti i dati utilizzati sono prima elaborati da un algoritmo di controllo qualità, al fine di includere nell'analisi solo misurazioni di comprovata affidabilità. Tale procedura di controllo è riportata in questo lavoro, cosi come una panoramica delle misurazioni respinte e dei problemi affrontati nel processo di convalida dei dati. Il TMY generato viene successivamente impiegato per previsioni di potenza di un modulo fotovoltaico (PV) mediante la tecnica dell’apprendimento automatico (Machine Learning). Inoltre, un secondo TMY disponibile online generato secondo la procedura classica da Photovoltaic Geographical Information System (PVGIS), viene utilizzato allo stesso scopo, e l’accuratezza delle rispettive previsioni confrontata. Il metodo adottato per valutare l'efficacia dell’uso di un TMY nella previsione di potenza di un modulo PV è quello di ‘previsione perfetta’. Vengono infine svolte ulteriori indagini sull'influenza di dati misurati e simulati del TMY sulle previsioni.
A novel procedure for a typical meteorological year composition with improved resolution adopted in the PV power forecast
STERLE, PAOLO
2018/2019
Abstract
In the last decades the energy world has been facing an intense primary energy source transition, from the fossil fuels to the renewable energy sources (RES). Just as the fossil fuel based energy industry relies on reserves, the renewable sector depends upon the assessment of resources. One of the main fields of renewable energy is related to solar energy, which is exploited in different ways, where photovoltaic (PV) systems, concentrating solar power systems, building energy systems, direct heating systems are some of the more assessed application. In order to evaluate and predict the performance of solar radiation based systems over medium-long time, given the unpredictability of the meteorological conditions, what is considered a typical period is used, known as Typical Meteorological Year (TMY). It is a yearly dataset that includes solar radiation values and meteorological data on hourly basis. In this thesis a novel procedure for TMY generation is presented, which modifies the classical methodology to account for the improved resolution of the dataset. Using this methodology a TMY for the location of Milano Bovisa is assembled. Before the application of the procedure, all meteorological data are processed by a quality control algorithm, in order to include in the analysis only measurements of proven reliability. An overview of rejected measurements and the issues faced in the data validation process are also reported. The TMY generated is later employed in Photovoltaic (PV) power forecast by means of machine learning technique. Also a classical version of TMY, being the web-available generated by Photovoltaic Geographical Information System (PVGIS) is applied to foresee PV power output. Forecast performances are then compared in perfect forecast condition, in order to assess the effectiveness of the TMY employment in PV power prediction. Moreover, further investigations are carried out about the TMY data structure influence over forecast.File | Dimensione | Formato | |
---|---|---|---|
TESI.pdf
accessibile in internet per tutti
Descrizione: Testo della tesi
Dimensione
3.72 MB
Formato
Adobe PDF
|
3.72 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/10589/148842