Back to search

University of Patras

Πρόγνωση παραγωγής αιολικής ενέργειας: επίδραση της τοπογραφίας και των ακραίων καιρικών φαινομένων

Abstract

dc:description

During the recent years the composition of the European network of energy production is facing rapid changes, since upgraded and advanced renewable energy technologies of consist the major part of the newly added production units. Among others, wind energy is, by far, the most developed and promising, followed by the solar energy. This trend is due to the current high level of expertise and technology in the wind energy field, the bigger expansion of its resources, the decrease in the energy production costs and also because of their contribution towards the reduction of greenhouse gas emissions. Wind energy is discontinuous since its primary source, the wind, is intermittent, and generally does not match any energy consumption profile. On the contrary, its profile follows a stochastic distribution that depends on the causes creating the wind as well as the type and make of wind turbines. Wind turbines are usually installed in locations with high wind speed potential favouring power production, usually in remote areas, rather than where the demand for electric power is required. It is obvious that those characteristics of wind energy impedes its integration in the already existing electric power distribution grid, especially due to issues related to energy production balance, voltage stability and other limitations related to the network design. Therefore wind energy is a challenge not only for the complex European energy distribution network but also for a smaller regional network, that of Greece, where wind power plants increase constantly. Research on wind energy focuses, among others, to the forecast of wind energy production, either indirectly by forecasting the wind and its components or by directly forecasting the wind power production, aiming at improving forecasts performance and eventually increasing their credibility. Furthermore, in many cases it is the knowledge of the probability of appearance of extreme wind speed values that matters more. Accurate forecasts of such events, which can be combined with cost effect analyses, are of great value. The purpose of the present dissertation is to contribute to the improvement of estimation of the parameters, mainly affecting the wind energy production, through the development of methods using real data coming from operating wind energy conversion systems (WECS). The basic idea is the management and processing of the available data in such a way so that the information included in those data would be used in the most appropriate and interpretable way. Our primary focus is on the wind speed forecast in a specific location, through the understanding of its behaviour, using the available wind data either from the same site (target) or from neighbouring sites (reference), as well as the filling of missing values or even gaps of the examined time series. The wind speed forecasting at the wind turbine nacelle height and the corresponding wind energy production consist the next subject this study deals with. Finally, the relationship between severe weather phenomena, such as thunderstorms accompanied by lightning activity, and the abrupt changes-gusts of wind speed at the nacelle height, is considered, while the capability of forecasting of their occurrence, for a short temporal horizon, is also addressed. It was found that forecasts using Artificial Neural Network (ANN) models present the best overall performance either for the wind speed simulations and estimations at a pre-selected target-point, by using reference wind speed time series from neighboring sites, or during the wind speed forecasting. The key point for the development of such models is the selection of the appropriate input vector. It was found that the application of the K-fold cross-validation method combined with the Auto-Correlation Function (ACF) and Partial Auto-Correlation Function (PACF) of the examined time series, decrease substantially the required input data and, consequently, the corresponding processing time. The wind speed forecasts are further improved when, in addition to the 10-min wind speed values, wind direction are also used as inputs. The performance of the proposed method has been applied and tested in three different geographical locations of Greece. The use of the power curve of the wind turbines was found to be an essential input parameter of the models developed to forecast the produced electric energy, which combined with forecasted wind speed data leads to energy production estimations of high accuracy. The analysis of the distribution of wind gust events and the concurrent lightning discharges observed in the study area during the same period, revealed that both present a similar seasonal and daily variation; 65% of recorded gusts were found to be directly or indirectly related to the presence of cumulonimbus clouds accompanied by lightning activity. The forecast of such events for the next hour showed satisfactory results, providing thus the possibility of early warning of occurrence of such phenomena, which may be a vital tool especially for the wind farm operators.

Degree

thesis:*
Grantor dc:publisher
University of Patras
Year dc:date
2021

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Kolokythas, Konstantinos
  • Κολοκυθάς, Κωνσταντίνος

Subjects

dc:subject × 10

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/50911

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

Kolokythas, Konstantinos; Κολοκυθάς, Κωνσταντίνος. Πρόγνωση παραγωγής αιολικής ενέργειας: επίδραση της τοπογραφίας και των ακραίων καιρικών φαινομένων. University of Patras, 2021. http://hdl.handle.net/10442/hedi/50911