The increasing use of renewable energy sources necessitates accurate forecasting models for generation scheduling. Amongst the renewable sources, solar and wind have gained acceptance and are being increasingly used in distributed generation. The main problem with these sources is the dependence of their power output on natural environmental parameters which are difficult to predict. Along with the discussion, focusing on the solar power, this thesis addresses the problem of estimation of the Solar Radiation (SR). Different setups of the time series model, and with their combinations with the weather forecast services using ensemble methods have been evaluated in medium-term prediction of the day-ahead regional SR. Moreover, several considerations including Support Vector Machine methods are also adopted at this stage, to cluster data. At the end, the validation of the approach is performed by using a SR data from a meteorological station and its nearby meteorological service which both situated in Northern Italy.
Lo sviluppo delle energie da fonti rinnovabili richiede accurate predizioni di radiazioni solari per mezzo di modelli predittivi. Tra le fonti rinnovabili che hanno avuto più successo ci sono: energia solare ed eolica. Il problema che più interessa la produzione di energia rinnovabile è la sua dipendenza dai fattori climatici che di per se sono molto variabili e difficili da prevedere. Questa tesi propone dei modelli che, sulla base di dati storici, predice la radiazione solare con un orizzonte di un giorno in avanti. Le predizioni che ogni modello presentato restituisce vengono confrontate con le predizioni di una servizio di predizione privato e con i reali valori misurati da una stazione meteo nella zona di Vipiteno. Varie tipologie di modelli predittivi sono stati proposti, quali: modelli auto regressivi (ARIMA), ibridi, combinazioni di vari modelli detti Ensemble, e l'introduzione di sistemi di Machine Learning.
Ensemble methods on improving quality of solar radiation prediction and their comparison with a commercial forecasting service
La CARRUBBA, DARIO
2018/2019
Abstract
The increasing use of renewable energy sources necessitates accurate forecasting models for generation scheduling. Amongst the renewable sources, solar and wind have gained acceptance and are being increasingly used in distributed generation. The main problem with these sources is the dependence of their power output on natural environmental parameters which are difficult to predict. Along with the discussion, focusing on the solar power, this thesis addresses the problem of estimation of the Solar Radiation (SR). Different setups of the time series model, and with their combinations with the weather forecast services using ensemble methods have been evaluated in medium-term prediction of the day-ahead regional SR. Moreover, several considerations including Support Vector Machine methods are also adopted at this stage, to cluster data. At the end, the validation of the approach is performed by using a SR data from a meteorological station and its nearby meteorological service which both situated in Northern Italy.File | Dimensione | Formato | |
---|---|---|---|
ThesisLaCarrubbaDario.pdf
accessibile in internet per tutti
Descrizione: Thesis
Dimensione
3.09 MB
Formato
Adobe PDF
|
3.09 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/10589/147974