Back to search

National Technical University of Athens (NTUA)

Πιθανοτική βραχυπρόθεσμη πρόβλεψη παραγωγής από αιολικά και φωτοβολταϊκά συστήματα με τεχνικές μηχανικής μάθησης

Abstract

dc:description

This PhD thesis contributes to the penetration of wind and photovoltaic power generation in power systems, by developing original forecasting models of wind and photovoltaic generation. Initially, the categories of forecasting models that have been developed over the years for the successful and accurate prediction from wind and photovoltaic generation systems are presented, as well as the main factors that affect the accuracy of the forecasts. Particular emphasis is placed on the analysis of Artificial Neural Networks, which in the context of the evolution of the machine learning industry have played a decisive role in improving the performance of renewable energy forecasting models. In this context and focusing on understanding the uncertainty involved in an energy production forecasting process, an innovative, hybrid model of short-term wind generation forecasting based on machine learning techniques is proposed. An optimized swarm algorithm is adapted and applied to the proposed methodology to optimize the hyperparameters of the proposed architecture. The effectiveness of the model is proven by comparing the results obtained with state-of-the-art probabilistic wind power generation prediction models. Then, focusing on power prediction from photovoltaic systems, an innovative, hybrid model for short-term, probabilistic forecasting is developed, suitable for photovoltaic power forecasting. A complex, improved chicken swarm type algorithm was adapted to the architecture of the proposed model, managing to improve its performance, while avoiding increasing the computational cost of the methodology. In addition, a prey-predator mechanism is integrated into the final hybrid model to further improve the search capabilities of the proposed model for optimal predictive outcomes. The model is applied to real data of solar parks. The final results highlight both the superiority of the proposed forecasting model over state-of-the-art models to which the available data were applied, as well as its applicability to real-world photovoltaic power forecasting problems.In conclusion, the proposed models introduce an innovative framework of methodologies that contribute to the penetration of wind and photovoltaic power generation in electric power systems.

Degree

thesis:*
Grantor dc:publisher
National Technical University of Athens (NTUA)
Year dc:date
2024

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Bazionis, Ioannis
  • Μπαζιώνης, Ιωάννης

Subjects

dc:subject × 14

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/58044

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Bazionis, Ioannis; Μπαζιώνης, Ιωάννης. Πιθανοτική βραχυπρόθεσμη πρόβλεψη παραγωγής από αιολικά και φωτοβολταϊκά συστήματα με τεχνικές μηχανικής μάθησης. National Technical University of Athens (NTUA), 2024. http://hdl.handle.net/10442/hedi/58044