Fine-Tuning Solar Power Predictions: An In-Depth Analysis of Decomposition Techniques

Prabowo, Angela Oryza (0024) Fine-Tuning Solar Power Predictions: An In-Depth Analysis of Decomposition Techniques. Project Report. [s.n.], [s.l.]. (Unpublished)

Text 5025201022-Project_Report.pdf - Accepted Version Download (1MB)

Abstract

Solar power generation plays a pivotal role in Taiwan's pursuit of renewable energy, aligning with its ambitious target of generating over 27 GW by 2050 and achieving net-zero emissions by the same year. Notably, Taiwan stands as the world's second- largest producer of solar photovoltaic (PV) energy, driven by the abundance of solar radiation, particularly in southern regions where it exceeds 145 watts per square meter. Building upon previous research that attained a remarkable low Mean Absolute Error (MAE) of 0.0223 through multivariate analysis with decomposition techniques, this study aims to further refine the forecasting models by focusing on the univariate decomposition. The hypothesis posits that this approach will lead to an even lower MAE, contributing to more accurate predictions. The research methodology involves extensive data preprocessing, which includes comparing and merging external datasets with information from the Taiwan Central Weather Bureau and Open Meteo. Feature engineering techniques are employed, incorporating transformations for time, wind direction, and wind speed. Recursive feature elimination is utilized for effective feature selection, enhancing the quality of input variables. In conclusion, this project centers on the refinement of univariate decomposition techniques to optimize solar power generation forecasts in Taiwan. By improving the accuracy of the decomposition model, the anticipated outcome is a more precise overall prediction when the refined results are integrated into subsequent forecasting steps. This research contributes to the ongoing efforts to harness solar energy efficiently and aligns with Taiwan's commitment to a sustainable and renewable energy future.

Item Type:	Monograph (Project Report)
Uncontrolled Keywords:	Decomposition, Solar Power Prediction, Time Series, Univariate
Subjects:	T Technology > T Technology (General) > T174 Technological forecasting T Technology > T Technology (General) > T57.5 Data Processing
Divisions:	Faculty of Information Technology > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Angela Oryza Prabowo
Date Deposited:	24 Jan 2024 03:02
Last Modified:	24 Jan 2024 03:02
URI:	http://repository.its.ac.id/id/eprint/105591

Actions (login required)

View Item