Técnicas de identificação de forças no domínio do tempo

• Ano de defesa: 2000
• Autor(a) principal: Silva, Leandro Afonso da
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Uberlândia; Brasil; Programa de Pós-graduação em Engenharia Mecânica
• Palavras-chave em Português: Identificação de forças; Deconvolução; Problemas Inverso; Identificação; Force identification; Deconvolution; Inverse problems; Identification; CNPQ::ENGENHARIAS
• Link de acesso: https://repositorio.ufu.br/handle/123456789/29077; http://doi.org/10.14393/ufu.te.2001.1

The research work is reported herein is devoted to the study of techniques intended for identification of externation forces from the system´s vibratory responses. Such a study is motivated by the the fact that, in many cicumstances, the excitation forces cannot be easily measured using load transducers. In the context of linear elastodynamics, three methods are based on the exploration of temporal responses, each of them using a particular type of model of the mechanical system. They are: the Modal Coordinates method, which uses a set of self-solutions (natural frequencies, modal damping factors, generalized masses and components of auto vectors), the method based on Deconvolution in the time domain, in which the model is constituted by a set of impulse response functions, and the method based on the Inverse Structural Model, which uses a state space model, formulated from the direct model interchanging the input and output vectors. The study comprises the development of the formulation, in different variants, of each of the studied methods, in addition to the evaluation of its operational characteristics and performance, through the performance of several applications, both to numerically simulated structures and to real mechanical systems tested in the laboratory. . In these applications, we seek to represent the difficulties typically present in practical situations, such as the presence of experimental noise, the use of incomplete data and the occurrence of poor numerical conditioning of the estimation equations. In the study of each method, proposals for procedures are made aiming at the best adaptation of the identification methods to the practical problems of Engineering, through the treatment of some of its inconveniences and difficulties.