Navegando por Autor "BERNHARD, Gustavo Galdino Rodrigues"

Agora exibindo 1 - 2 de 2

Modelo de Previsão de Chuva Usando Redes Neurais Artificiais
(Instituto Nacional da Propriedade Industrial (INPI), 2022-09-20) BERNHARD, Gustavo Galdino Rodrigues; LIMA, Alexandra Amaro de; NASCIMENTO, Manoel Henrique Reis; ALENCAR, David Barbosa de; http://lattes.cnpq.br/4890967546423188; http://lattes.cnpq.br/6915958689972413
This document describes a computer program registered under number BR512023002419-8, whose objective is to predict the occurrence of rain using artificial neural networks. The model developed applies artificial intelligence techniques to analyze meteorological data, allowing for more accurate and efficient forecasts. The tool is useful for meteorologists, researchers and institutions that need reliable climate forecasts for planning and decision making. The program was created to improve the accuracy of rain forecasts, contributing to the mitigation of risks associated with extreme weather events. Program registration is valid for 50 years from January 1, 2023.
Modelo de previsão de chuva usando redes neurais artificiais
(Instituto de Tecnologia e Educação Galileo da Amazônia, 2022-09-20) BERNHARD, Gustavo Galdino Rodrigues; LIMA, Alexandra Amaro de; http://lattes.cnpq.br/6915958689972413
Precipitation is important in maintaining the environment and life of living beings. Through their studies and increasingly accurate forecasts, we can reduce the impacts related to floods, environmental disasters, and losses in the agricultural and tourism sectors. However, climate change has made the analysis of this variable difficult. In this article, we will present an hourly rain forecast model using Artificial Neural Networks, using the information on instantaneous, maximum and minimum temperature, relative humidity, wind, and precipitation through the automatic weather stations of the Instituto Nacional de Meteorologia (INMET), As a methodology, we will present it in three steps, called assembly, because it defines the type of network, architecture (layers and hidden commands), activation functions, type of control, learning algorithm and other parameters. Training, where the ANN captures all the relevant characteristics of the selected data set being divided, so that 70% of the information made available for the model to carry out the network learning, and 15% are saved to carry out the validation and between 15% not network test and a third stage takes place the RNA tests, where it is to perform the precipitation forecast and corrections of the values found by the network. The forecasts made with the model had strong results, showing that the model was able to reproduce the same behavior as the observed observation for the predicted day, presenting practically the same total millimeters (mm), mainly on the rainiest days. On the other hand, in cases where the observed showed a characteristic of convective precipitation, the model failed to capture the intensity, which shows that this should be tested with other astrological variables.