Y. Dimakos, Stochastic analysis and simulation of wind in Greece at multiple scales, MSc thesis, 214 pages, Department of Water Resources and Environmental Engineering – National Technical University of Athens, March 2014.
In this thesis it was examined the statistic - stochastic processing of wind speed, at monthly, daily and smaller time scale. Also, it was examined the spatial and temporal distribution of wind speed and it was made an effort of modelling this variable, by exporting the basic statistical characteristics, involving the entire examined area that was Greece. Furthermore, it was examined the existence of long-term persistence at the same time scale, calculation of Hurst coefficient and exportation of climacograms. Additionally, it was examined gust and wind direction, at a certain number of stations that data measurements were available, in order to depict a more complete frame of wind speed variable. Also, it was examined reanalysis data, to assess the reliability of such measurements and the importance of their use in the absence of measurements. Still, trends of wind speed were tested, for a single reporting period, in order to determine the behavior of wind for the examination period and stations with positive and negative trends were recorded, respectively. Also, it was held the estimation of the theoretical wind energy potential of data stations and resulting conclusions about which areas compose sufficient wind resources. Specific distributions were adjusted for gust and wind speed, in order to be found which one best fits to the empirical data. Additionally, synthetic time series were generated, by the above mentioned model and by specific characteristics of wind speed, which were resulted from this work. Finally, examination of wind speed revealed that wind is strongest in areas of the central Aegean and especially less potential in northern Greece. Similar was the spatial distribution of the theoretical wind energy potential of the country. Examination of climacograms was single price for the Hurst coefficient, for the entire Greece (H = 0.75), which was an important factor for a more general modelling, of the wider region of interest, by examining each individual station separately.