Please use this identifier to cite or link to this item: https://scholarbank.nus.edu.sg/handle/10635/227009
DC FieldValue
dc.titleVERY SHORT-TERM GLOBAL HORIZONTAL IRRADIANCE (GHI) FORECASTING USING MACHINE LEARNING
dc.contributor.authorKILAPARTHI AISHWARYA
dc.date.accessioned2022-06-13T07:12:18Z
dc.date.available2022-06-13T07:12:18Z
dc.date.issued2022
dc.identifier.citationKILAPARTHI AISHWARYA (2022). VERY SHORT-TERM GLOBAL HORIZONTAL IRRADIANCE (GHI) FORECASTING USING MACHINE LEARNING. ScholarBank@NUS Repository.
dc.identifier.urihttps://scholarbank.nus.edu.sg/handle/10635/227009
dc.description.abstractAs climate change is growing into a more serious issue, countries are increasingly shifting towards more renewable sources such as solar, wind and nuclear power from traditional power sources. Photovoltaic (PV) power generation is generally affected by various factors such as cloud cover and high temperatures that are relative to a country. Hence, the incorporation of solar energy into a country’s energy mix can affect the safety and stability of power grid generation due to sudden voltage fluctuations. With climate change and the transition to renewable energy sources, accurate solar forecasts are required to ensure that power grids are safe. Global Horizontal Irradiance (GHI) forecasting is conducted using single regression and ensemble models: Linear Regression (LR), Decision Trees (DT), Support Vector Regression (SVR), K Nearest Neighbours (KNN), Random Forest (RF), Extreme Gradient Boosting (XG), Bagging (BAG) and AdaBoost (ADA). The meteorological data spans over 2 years at every minute from 8am to 6pm and is taken for Folsom, California from the National Solar Radiation Database (NSRDB). The prediction is done on a Very Short-Term (VST) horizon timeframe ranging from 1 minute to 30 minutes. Model performance is analysed based on four evaluation metrics: Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Coefficient of Determination (??2) and Mean Absolute Percentage Error (MAPE). Based on the results, ensemble models outperform single regression models where ADA and RF were able to achieve a MAPE of 5.26% and 5.33% respectively.
dc.subjectSolar Energy
dc.subjectMachine Learning
dc.subjectEnergy Prediction
dc.subjectForecasting
dc.subjectclimate change
dc.typeDissertation
dc.contributor.departmentDEPT OF THE BUILT ENVIRONMENT
dc.contributor.supervisorYAN KE
dc.description.degreeBachelor's
dc.description.degreeconferredBACHELOR OF SCIENCE (PROJECT AND FACILITIES MANAGEMENT)
Appears in Collections:Bachelor's Theses

Show simple item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
Kilaparthi Aishwarya DBE_Aishwarya Kilaparthi 1.pdf2.32 MBAdobe PDF

RESTRICTED

NoneLog In

Page view(s)

28
checked on Nov 17, 2022

Download(s)

2
checked on Nov 17, 2022

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.