Shape optimization of double-chamber side mufflers with extended tube by using four-pole matrix, boundary element method and genetic algorithm

Abstract

As the excellent noise reduction in dealing with most of the industrial venting noise in lower frequencies, reactive silencer has been prevailing and widely discussed, according. Most researchers explored the noise reduction effects due to flowing rate and temperature gradient based on pure plane wave theory; however, the effect of high degree wave was rarely addressed. Moreover, the maximum of silencer noise reduction under space constraint, in which frequently occurred in engineering problem, are hardly mentioned. Most of the optimal assessments were oriented with a slower optimizer, the mathematic gradient method. An issue by using a precise and quick optimizer to access the optimal noise reduction of a complicated silencer under limited space becomes essential and obligatory. In this paper, the optimal design of a double–chamber side muffler connected with internal extended tube under limited space is presented; in addition, the genetic algorithm is selected as the optimizer in searching for the best design set. In theoretical derivation, on the basis of one–dimensional plane wave theory, the system four-pole matrices under flowing effect have been established and furthermore deduced to a scalar noise reduction function. To verify the accuracy of four–pole transfer matrices, one set of silencer shape parameter has been used to construct and further acoustic test in laboratory and noise simulated in boundary element method (high degree of sound wave) have be performed individually. The results reveal that all of the four–pole matrix theory, experimental data and boundary element method are in agreement. To verify the accuracy of GA solution during optimization, an optimal design of noise reduction onto pure tone with 300 Hz is pre–run. The result reveals that the maximized STL is precisely located at 300 Hz. Consequently, a numerical case of optimal noise reduction to broadband noise source under space constraint is exemplified.