Fechar
Metadados

@PhDThesis{Sumida:2017:CaAnSi,
               author = "Sumida, Ivana Yoshie",
                title = "Calibra{\c{c}}{\~a}o e an{\'a}lise de simuladores de 
                         din{\^a}micas de helic{\'o}pteros por Bred Vector",
               school = "Instituto Nacional de Pesquisas Espaciais (INPE)",
                 year = "2017",
              address = "S{\~a}o Jos{\'e} dos Campos",
                month = "2017-08-24",
             keywords = "din{\^a}mica de voo, identifica{\c{c}}{\~a}o de 
                         par{\^a}metros, bred vector, flight dynamic, parameter 
                         identification, bred vector.",
             abstract = "O setor aeroespacial {\'e} fortemente ligado a diversas 
                         {\'a}reas da computa{\c{c}}{\~a}o. S{\~a}o exemplos bem 
                         conhecidos: aerodin{\^a}mica computacional, controle 
                         computadorizado de aeronaves, softwares do segmento de solo e 
                         simuladores de voo. H{\'a} um uso crescente de simuladores de voo 
                         com o objetivo de aumentar a seguran{\c{c}}a das 
                         opera{\c{c}}{\~o}es. Para que os simuladores tenham uma boa 
                         ader{\^e}ncia com a realidade {\'e} preciso calibr{\'a}-los, ou 
                         seja, identificar os par{\^a}metros que tornem a 
                         simula{\c{c}}{\~a}o o mais pr{\'o}ximo poss{\'{\i}}vel da 
                         din{\^a}mica real. Ap{\'o}s determinar tais par{\^a}metros, o 
                         simulador estar{\'a} pronto para ser usado na 
                         avalia{\c{c}}{\~a}o de foguetes, aeronaves e no treinamento de 
                         pilotos. A identifica{\c{c}}{\~a}o de par{\^a}metros, que 
                         caracteriza o comportamento aerodin{\^a}mico da aeronave, pode 
                         ser formulada como um problema de otimiza{\c{c}}{\~a}o. Um dos 
                         objetivos desse trabalho {\'e} realizar a calibra{\c{c}}{\~a}o 
                         de um simulador da din{\^a}mica de helic{\'o}ptero atrav{\'e}s 
                         da meta-heur{\'{\i}}stica: algoritmo de colis{\~a}o de 
                         m{\'u}ltiplas part{\'{\i}}culas (MPCA: Multiple Particle 
                         Collision Algorithm). Os resultados foram comparados com os 
                         encontrados pelo algoritmo gen{\'e}tico. Entretanto, uma 
                         an{\'a}lise da din{\^a}mica tamb{\'e}m {\'e} relevante para 
                         que o simulador possa bem representar a resposta da aeronave. O 
                         segundo objetivo deste trabalho {\'e} realizar a previsibilidade 
                         da din{\^a}mica. Nessa etapa, o problema da previsibilidade 
                         {\'e} formulado como um problema de classifica{\c{c}}{\~a}o. 
                         Uma t{\'e}cnica baseada em \${''}\$bred vectors\${''}\$ 
                         {\'e} empregada para avaliar um sistema din{\^a}mico n{\~a}o 
                         linear e classificadores neuro-difusos s{\~a}o utilizados para 
                         identifica{\c{c}}{\~a}o das classes de previsibilidade. Neste 
                         trabalho, a aplica{\c{c}}{\~a}o da metodologia de \${''}\$bred 
                         vectors\${''}\$ tem foco na din{\^a}mica do rotor de um 
                         helic{\'o}ptero. ABSTRACT: The aerospace industry is strongly 
                         connected to several computing areas, such as computer 
                         aerodynamics, computerized control of aircraft, ground segment 
                         software, and flight simulators. Flight simulators are being 
                         largely used to increase the safety of operations. In order that 
                         the simulators have a good adherence to reality, they need to be 
                         calibrated, that is, identify the parameters that make the 
                         simulation as close as possible to the real dynamic. After these 
                         parameters are identified, the simulator is ready to be used to 
                         evaluate rockets and aircraft, as well as in pilot training. 
                         Determining these parameters, which characterizes the aerodynamic 
                         behavior of the aircraft, can be formulated as an optimization 
                         problem. One of the goals of this work is to calibrate the 
                         simulator of a helicopter dynamics using a new meta-heuristic 
                         named MPCA (Multiple Particle Collision Algorithm). The results 
                         were compared to the ones found using a genetic algorithm. 
                         However, an analysis of the flight dynamics is also relevant so 
                         that the simulator can represent the response of the aircraft. The 
                         second goal of this work is to predict the flight dynamics of a 
                         simulator. At this stage, the problem of predictability was 
                         formulated as a classification problem. A bred vector based 
                         technique was employed to evaluate a non-linear dynamic system, 
                         and neuro-fuzzy classifiers were used to identify predictability 
                         classes. The application of bred vectors methodology focused on 
                         the dynamics of the rotor of a helicopter.",
            committee = "Ramos, Fernando Manuel (Presidente) and Campos Velho, Haroldo 
                         Fraga de (Orientador) and Sandri, Sandra Aparecida and Stephany, 
                         Stephan and G{\'o}es, Luiz Carlos Sandoval and Ritto, Thiago 
                         Gamboa",
         englishtitle = "Calibratoin and analysis of helicopter dynamics simulators by Bred 
                         Vector",
             language = "pt",
                pages = "108",
                  ibi = "8JMKD3MGP3W34P/3PDGJ58",
                  url = "http://urlib.net/rep/8JMKD3MGP3W34P/3PDGJ58",
           targetfile = "publicacao.pdf",
        urlaccessdate = "05 dez. 2020"
}


Fechar