Research Information System
Thesis Type: Postgraduate
Institution Of The Thesis: Gazi Üniversitesi, Fen Bilimleri Enstitüsü, Turkey
Approval Date: 2019
Student: MUSTAFA ALTINSOY
Supervisor: GÜNGÖR BAL
Open Archive Collection: AVESIS Open Access Collection
Abstract:Supplying energy from renewable sources as long as possible is one of the most important requirements for sustainable development. Wind energy has an important role in Turkey in terms of the existing potential. A significant part of the consumed energy in our country is supplied by fossil fuels and imported sources. This situation adversely affects our country in terms of strategical and economical manner. As in all over the world, renewable energy investments are also increasing in our country. The correct determination of renewable energy potentials will prevent the investment to fall into an idle position. In this study, the meteorological dataset containing past 30 years data of wind speed, humidity, pressure, temperature and precipitation obtained from the Turkish State Meteorological Service are used. Wind speed estimation is studied for sample districts located in the Ankara city using this data set. In Matlab, different artificial neural network learning algorithms have been used to construct the ANN models. In case of looking at the results of the instantaneous wind speed prediction; the mean absolute percentage errors (MAPE) for testing data were found as %8,52 (for Çubuk), %7,71 (for Keçiören), % 8,89 (for Gölbaşı), %7,27 (for Kızılcahamam), %6,09 (for Elmadağ), %6,92 (for Polatlı), %8,18 (for Bala), %7,50 (for Şereflikoçhisar) and %7,85 ( for Haymana). In case of looking at the results of the future-term wind speed prediction; the mean absolute percentage errors (MAPE) for testing data were found as %9,48 (for Çubuk), %7,77 (for Keçiören), %7,88 (for Polatlı), %6,83 (for Bala), %8,02 (for Şereflikoçhisar) and %5,41 ( for Haymana). In case of comparing the results with the similar studies, it is shown that wind speed can be predicted using the data set containing month, temperature, pressure, relative humidity and precipitation data.