Desenvolvimento de sistema de matriz de sensores para avaliação de transdutores acústicos

• Ano de defesa: 2013
• Autor(a) principal: Gilberto Mendes
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Transdutor acústico; Diretividade; Matrizes de microfones; Aquisição de sinais; Processamento digital de sinais; Projetos mecânicos; Engenharia mecânica; Processamento de sinais; Transdutores
• Link de acesso: http://hdl.handle.net/1843/BUBD-9EGMKZ

This work involves the creation of a system for evaluating the directivity of sound sources using arrays of microphones. The arrangement as well as the choice of transducers are discussed and focus particularly on the use of two-dimensional arrays of electret microphones in evaluating of a static sinusoidal source with a fixed frequency. It is emphasized the importance of using low cost systems, due to the large number ofindividual sensors, as the use of processing techniques to reduce the processing volume to a minimum. To study of the configuration is mounted a set of experiments with a tweeter-type emitter and an array of electret microphones, enabling to vary the relative positions, frequencies, signal combinations, to evaluate relevant aspects of the method for characterizing the emitting source. The distribution of microphones is designed to allow the investigation of various geometries and compare their advantages in other works. The acoustic emitter and the sensors were mounted in an enclosure for acoustic isolation in order to eliminate interferences by the external environment. The design of signal conditioning systems is discussed in detail. The system operates with an array of 37 sensors whose signals are multiplexed and processed in sets of four channels after conversion to digital at a Digital Signal Processor development kit. As important factor in the efficiency of the system is presented and confirmed the hypothesis of temporalstationarity of the evaluated signals, i.e. the variations along the time interval of a test are neglected. This study can contribute significantly to identification applications and characterization of sound sources, enabling extensions in terms of hardware and processing.