Estimación del potencial energético fotovoltaico y eólico mediante agrupamiento no supervisado por k-medias: caso de estudio Universidad de La Salle (Colombia)

Abstract

This research study proposes a methodology based on data science to estimate the technical feasibility of implementing hybrid solar-wind power systems by calculating the plant factor, also known as capacity factor (CF). Data is gathered from the NOVALYNX 110-WS-18 weather station, between the years 2018 and 2020, to determine the energetic potential available for the Candelaria campus at the La Salle University (Colombia). A k-means data clustering algorithm, developed in Matlab, is applied over the following descriptors: time, solar radiation, wind speed, and temperature. The results show that the photovoltaic and wind power systems have a CF of 36.69% and 2.15%, respectively. It is concluded that it is technically feasible to implement a photovoltaic system at the La Salle University, but not a wind power system as it does not meet the minimum CF required.