Analysis and application of a novel approach based on Finsler geometry for wind speed modelling in energy systems

Today, the use of renewable energy systems are increased rapidly in parallel with the growing energy demand in order to produce clean and sustainable energy. This thesis includes novel approach models and different forecasting studies in terms of wind energy systems which is one of the most important renewable energy sources. In the first stage of the study, forecasting of wind speed is performed by time series analysis, artificial neural networks (ANN), and hybrid model which is empirical mode decomposition (EMD) neural networks. In the second stage of the study, monthly, seasonal and yearly wind speed modeling are carried out using the Weibull distribution (WD), Rayleigh distribution (RD) as well as the newly proposed inverse Weibull distribution (IWD) in the literature. In this regard, Riemann geometry based Information Geometry Method (IGM) is proposed as a novel approach for estimation of parameters. The advantages of Information Geometry Method that is a novel approach parameters estimation of the distribution function for wind speed modeling are discussed with the other methods in the literature. In the last stage of the study, two-dimensional Finsler space metric function is obtained for Weibull distribution. It is used in many applications in this area such as wind speed modeling. The metric definition for two-parameter Weibull probability density function which has shape (k) and scale (c) parameters in two-dimensional Finsler space is realized using a different approach by Finsler geometry. New probability and cumulative probability density functions based on Finsler geometry are derived for more accurate modeling. The novel approach based on Finsler geometry is analyzed comparatively with other methods, which are commonly used in the wind speed model, and evaluated in respect to energy potential analysis. The models and methods in the thesis are carried out for wind data, consisting of hourly wind speed records between 2012-2016 from Bilecik, Yalova, Gökçeada, Bozcaada, Sakarya and Bandırma stations. All the discussed tables and figures are obtained by utilizing the results comparatively in the numerical computing software tools MATLAB and Mathematica.