Optimization of Rectangular Multi-Chamber Plenums Equipped with Multiple Extended Tubes Using the BEM, Neural Networks, and the Genetic Algorithm

Abstract

AbstractThe focal point of this paper is to uncover, by analyzing the higher order wave effect, an improved mechanism for space-constrained rectangular plenums using a simplified objective function in conjunction with a genetic (GA). Three kinds of rectangular mufflers hybridized with extended tubes will be assessed: Plenum A: A two-chamber plenum equipped with an extended tube; plenum B: A three-chamber plenum with two extended tubes; and plenum C: A two-chamber plenum equipped with three extended tubes. In order to shorten the numerical assessment, a simplified objective function (OBJ) is established using a boundary element model (BEM) in conjunction with the neural network model (NNM). To expediently approach an optimal plenum, the best OBJ will be numerically searched using a genetic algorithm (GA). However, before the GA operation is performed, the accuracy of the BEM is verified using analytical data. And, because the simplified objective function (OBJ) is seen to be in agreement with the BEM, the numerical cases of sound elimination relative to the various parameter sets and pure tones (300, 750, and 1300Hz) can be carried out.Results reveal that the maximum value of the sound transmission loss (STL) can be accurately obtained at the desired frequencies. Additionally, the acoustical performance of the lower frequencies will be improved if the number of chambers and rectangular tubes are increased. However, the acoustical performance of the higher frequencies will decrease when the number of chambers and rectangular tubes are decreased. Consequently, the algorithms proposed in this study will efficiently develop optimal rectangular plenums with multiple rectangular extended tubes.