Redução de ordem de um modelo de um veículo de carga ferroviário

• Ano de defesa: 2018
• Autor(a) principal: Amorim, André Guilherme Alves
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Espírito Santo; BR; Mestrado em Engenharia Elétrica; Centro Tecnológico; UFES; Programa de Pós-Graduação em Engenharia Elétrica
• Palavras-chave em Português: Freight cars; Railroad; Vehicle Simulation; Model Order Reduction; Genetic Algorithms; Vagões; Simulação Veicular; Redução de Ordem; Ferrovias - Simulação por computador; Algorítmos genéticos; Transporte ferroviário de carga - Vagões; Modelos matemáticos; Medidas Elétricas, Magnéticas e Eletrônicas; Instrumentação; 621.3
• Link de acesso: http://repositorio.ufes.br/handle/10/10965

In railway industries there are many applications for systems modelling and dynamic simulations of models. Especially, in railway transportation department it is important to simulate computationally the dynamics of freight trains models in order to ensure design quality and safe operation conditions for wagons. The models that reproduce the behavior of the real system involve a great number of degrees of freedom, thus, more computationally complex and more difficult to implement active control suspension systems. In this context, the present work develops a 30 state linear model of the rail-vehicle interaction to evaluate the vertical and lateral dynamic response of the vehicle. It is proposed, initially, the simplification of this full model by analytical model order reduction techniques present in the literature. These techniques use similarity transformations to obtain realizations of the system in the balanced form and modal form in order to prioritize the most influent states in the final response of the model, eliminating the least contributory states through truncation or singular perturbation elimination (SPE). Then, it is applied a genetic algorithm to the result of the model order techniques, aiming the optimization of the mathematically reduced models. The comparison between both models, full and reduced, is based on the AAR (American Association of Railroads) manual of standards and recommended practices section C – part II: Design, Fabrication, and Construction of Freight Cars, where are indicated tests to evaluate security and the respective requirements for the approval of new models of freight cars. The results obtained by simulation shows that it is possible to acquire reduced models that are good approximations of the full model, specially using the singular perturbation based techniques. It is also proven the potential of application of genetic algorithms to optimize reduced models. Therefore, the results obtained in this work, can make easier the implementation of active control suspension systems and simplify the security operational analysis of new suspension projects.