Multi-frequency sequential and simultaneous parametric shape optimization of reactive silencers

Abstract

Silencers are widely used to attenuate unwanted noise emitted by engines, exhaust, and ventilation systems or from intermittent gas flow in pipes. In recent years, the study of noise control using optimization techniques applied to silencers has been growing. The main objective of the present work is to parametrically optimize the shape of a compact reactive silencer with a single expansion chamber, modifying its construction dimensions. Sequential and simultaneous optimizations are applied to the shapes of three reactive filter types so that their sound transmission losses are maximized over pre-established frequency ranges. The optimized reactive filters are the expansion chamber with modified surface profile, extended ducts and the bi-partition of the chamber. The optimization process is carried out with the aid of the finite element method using a formulation for axisymmetric problems and with the models being discretized with linear quadrilateral elements. The optimizations were evaluated in 16 frequency range combinations totaling 98 possible results. The results showed that both optimization techniques are efficient in improving the efficiency of silencers above the first TL drop frequency of the base model, which is composed of a single expansion chamber. However, the sequential strategy presented optimized models with more regular final shapes. In addition, the methodologies are validated using the traditional three-dimensional Finite Element Method and with the experimental verification of two optimized models.