Gloria Patricia Carrascal Pérez

Universidad del Atlántico, Colombia

Title: Estimation of Solar Radiation in Colombia Using the Artificial Neural Network Algorithm

---

Biography

Gloria is a senior in Physics At Universidad del Atlántico, working in her research project inthe field of applied physics, she is an Engineer intern in a startup, and student at MIT IDSS'Data Science and Machine Learning: Making Data-Driven Decisions Program.

She has apassion for climate data, oceanography, quantum computing, research and technology ingeneral.Additionally, she is fluent in Spanish (native language) and English (proficiencylevel) and has developed skills in art, diving and volunteer work as ambassador in groups tohelp animals and empower women in her city and country.

Abstract

In this research, the region of Colombia with the highest average daily radiation intensity inthe studied time interval will be identified. Artificial intelligence techniques, such as Feedforward back propagation (FFBP), and Wavelet recurrent neural network (WNN) algorithmswill be explored. The modeled data provided by the CAMS open data source for the timeseries of solar radiation, air quality and surface temperature from January 1, 2020 to July 31,2020, will be used to train the algorithms and the tests data is carried out with the data fromJanuary 2, 2020 to August 1, 2020 of the time series of solar radiation. With this informationand the identification of the variables with the greatest impact on the variability of thehorizontal and global diffuse solar irradiance, the values of the following day will bepredicted. From the results, the algorithm with a low uncertainty in the prediction and underdifferent levels of pollution and temperature will be used to generate the new radiation datafor each of the regions and identify the region with the highest mean daily radiation intensityin the studied time interval.