| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 5.15 MB | Adobe PDF |
Authors
Abstract(s)
No âmbito do projeto ICARUS pretende-se desenvolver um veículo autónomo de salvamento marítimo. O objetivo é que este constitua o primeiro meio de socorro a náufragos. Enquadrado neste projeto, proponho-me na modelação dos hélices que equiparão esta plataforma. Para o efeito, irá ser utilizado uma abordagem por “Surrogate Modeling” com recurso a Redes Neuronais Artificiais. Este método irá consistir na análise dos dados de diversas provas realizadas com diferentes hélices a várias condições. Após o tratamento dos mesmos, chegar-se-á, idealmente, a um hélice que se encontrará perfeitamente
adaptado à plataforma e à sua especificidade a nível de missão.
In accordance with ICARUS project, the intention is to develop an autonomous vehicle suitable for maritime Search and Rescue. The purpose of this unmanned vehicle, is to constitute the first mean of assistance to people who are in distress or imminent danger. Framed to this project, it is proposed the modeling of the propellers that will equip this vehicle. It will be investigated an innovative way of dealing with this problem, “Surrogate modeling” using Artificial Neural Networks. This method consists on the analysis of data from a series of tests performed with different propellers under a variety of conditions. After processing this data, it will be possible to get a propeller perfectly adapted for the platform and its specificity in terms of the nature of the missions it is supposed to execute.
In accordance with ICARUS project, the intention is to develop an autonomous vehicle suitable for maritime Search and Rescue. The purpose of this unmanned vehicle, is to constitute the first mean of assistance to people who are in distress or imminent danger. Framed to this project, it is proposed the modeling of the propellers that will equip this vehicle. It will be investigated an innovative way of dealing with this problem, “Surrogate modeling” using Artificial Neural Networks. This method consists on the analysis of data from a series of tests performed with different propellers under a variety of conditions. After processing this data, it will be possible to get a propeller perfectly adapted for the platform and its specificity in terms of the nature of the missions it is supposed to execute.
Description
Keywords
Redes Neuronais; Hélice; Modelação; Otimização; Surrogate Modeling Neural Networks; Propeller; Modelling; Optimization; Surrogate Modeling