Estudio de Prefactibilidad de un Sistema Eólico Mediante Técnicas de Ciencia de Datos en la Ciudad de Bogotá

Abstract

Wind energy is an increasingly important source of clean and renewable energy. The article explores the feasibility of modeling the generation of wind power using data science techniques and machine learning in Bogotá, Colombia. The methodology employed includes clustering analysis, Analytic Hierarchy Process (AHP), and Weibull distribution to model wind power generation from historical meteorological wind data. The wind resource analysis involves analyzing multiple meteorological factors, such as wind speed, temperature, and direction, from a meteorological station.The study evaluates suitable wind turbine selection to design a renewable wind energy system, with a methodology that encompasses data collection and preprocessing, wind resource evaluation, and system design and optimization. Clustering analysis and Weibull distribution models are used to identify key characteristics following analysis of the historical meteorological wind data, whereas AHP helps during the selection of an appropriate wind turbine based on several criteria. The performance metrics consist of Capacity Factor (CF), Performance Yields (Yr), and Performance Ratio (PR), providing insights into the efficiency and output of the renewable wind energy system. The importance of data preprocessing for improved and accurate results is highlighted, along with a discussion of the challenges posed by the turbulent nature and suboptimal placement of small wind turbines in urban environments.