Abstract

techniques for wind speed forecasting. However, although there are multiple studies, none are set up for the Colombia Caribbean coast. This is a disadvantage because the potential of wind resources in this region is greater than the hydroelectric potential of the whole country, but all this potential has yet to be developed. In this paper, based on time series, Autoregressive Integrated Moving Average (ARIMA), and Multiple Regression with Ordinary Least Squares (OLS) in the study, two models are proposed and their performance for wind speed prediction is compared. The data were collected in the meteorological station located in the experimental farm of the Atlantic University, in Barranquilla, Colombia, and variables analyzed included wind speed, wind direction, temperature, relative humidity, solar radiation, and pressure. The results of the two approaches indicated that among all the involved models, the ARIMA model has the best predicting performance. Also, it is essential to highlight that through this work, decision-makers would explore the local wind potential, allowing for the possibility of predicting future wind speed, and thus giving them the ability to plan the production and the interaction of other sources of energy.