Redes neurais artificiais aplicadas à determinação do tamanho ótimo da malha para o cálculo da intensidade útil

• Autor(a) principal: Barros, Taciano Magela de Souza Monteiro de
• Data de Publicação: 2014
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UERJ
• Texto Completo: http://www.bdtd.uerj.br/handle/1/13807
• topic: Useful acoustic intensity; Sound radiation; Artificial neural networks; Vibroacoustic; Intensidade acústica útil; Radiação sonora; Redes neurais artificiais; Vibroacústica; CNPQ::CIENCIAS EXATAS E DA TERRA::MATEMATICA::MATEMATICA APLICADA

In this paper, a study to determine the optimal size of the mesh elements, using artificial neural networks, to calculate useful intensity is presented. The main idea is training the neural networks, enabling them learning and recognizing the best size for the various superficial areas in sound sources with at geometry. The advantage of using artificial neural networks is due to the fact that they present a single size for obtaining the useful intensity, thereby significantly reducing computation time compared with the calculation time for a too fine mesh. Numerical tests with at plates - separable geometry that enables an analytical solution - are used to make comparisons. A comparative study of the computational time spent to obtain the useful intensity and the computational time spent to obtain the useful intensity using the mesh optimized via artificial neural networks is presented. A comparison of the sound power level obtained by the numerical solution in order to validate the results using neural networks is also presented.