Repositorio Digital Marítimo, Fluvial y Costero

Tenaud Jo, Biby Ann. (2023). Pronóstico de visibilidad horizontal y temperaturas para el Aeropuerto de Trujillo mediante el uso de redes neuronales artificiales de retropropagación. (Tesis de Diploma). Universidad Nacional Agraria La Molina, Lima, Perú.

010_GRASPPE.pdf Disponible bajo licencia Creative Commons Attribution Non-commercial Share Alike. Download (2MB)

Abstract

It is almost impossible to talk about any aspect of aviation without reference to the meteorological environment in which an aircraft operates. Failure to forecast adverse weather events for aviation in a timely manner can lead to flight delays or cancellations, generating large economic costs for airlines and passengers. The new computational technologies that are being developed, allow an update of weather forecasting techniques, with artificial neural networks as a new paradigm of interest. In the present investigation, the forecast of the meteorological variables of horizontal visibility and maximum and minimum temperatures at the Trujillo airport was carried out using Backpropagation Neural Network (BNN), analyzing and identifying the climatological conditions and meteorological systems of influence that govern the place. Having as input variables for the BNN, daily data of the maximum and minimum temperature from the years 2008 to 2017 and hourly data of the horizontal visibility from the years 2013 to 2017 of the Trujillo airport, in addition to the explanatory variables such as global indices and other local weather variables. The BNNs used for maximum and minimum air temperature analysis consist of eight neurons in the input and hidden layer and an output layer of one neuron; and for horizontal visibility it consists of ten neurons in the input and hidden layer, and an output layer of one neuron. The results showed that, for the maximum and minimum temperature variables, the predicted and observed values are very close to each other with a similar trend; and for the horizontal visibility variable, significant differences are perceived, however, they do show similar trend patterns.

[error in script]

Item Type:	Thesis (Diploma)
Uncontrolled Keywords:	Redes neuronales artificiales. Retropropagación. Backpropagation. Pronóstico. Visibilidad. Temperatura. Trujillo. Aeronáutica
Geographical coverage :	Perú
Depositing User:	Dirección General Marítima
Official URL:	https://hdl.handle.net/20.500.12996/5602

Actions (login required)

View Item