Optimization of multiple-curve-tube mufflers using neural networks, the boundary element method and GA method. Issue 5 (3rd September 2018)

• Record Type: Journal Article
• Authors: Chiu, Min-Chie; Chang, Ying-Chun; Cheng, Ho-Chih; Lin, Ting-Yang
• Is Part Of: Journal of statistics & management systems. Volume 21:Issue 5(2018)
• Page Start: 787
• Page End: 806
• Publication Date: 2018-09-03
• Subjects: Boundary element method -- Herschel-Quincke tube -- Polynomial neural network model -- Multi-connected tube -- Optimization -- Genetic algorithm; Statistics -- Periodicals; Mathematical models -- Periodicals; Mathematical models; Statistics; Periodicals; 519.5
• DOI: 10.1080/09720510.2018.1475127
• Languages: English
• ISSNs: 0972-0510

Abstract

Abstract: Based on the phase-cancellation technique, multi-connected-tube mufflers have been applied in the noise elimination for industrial venting noise. On the basis of the transfer matrix method, most researchers have explored noise reduction effects. Yet, the maximum noise reduction of multi-connected-tube mufflers within a constrained space has been ignored. Therefore, the optimum design of mufflers becomes essential. In previous study, a muffler equipped with three internally connected curved tubes at the same side has been investigated. In order to explore an advanced muffler with new acoustical effect, a muffler equipped with three connected curved tubes at different sides with rectangular sections within a fixed length has been proposed and assessed in this paper. To facilitate the assessment of optimal mufflers, a simplified objective function is established by linking the boundary element model with a polynomial neural network fitted with a series of real data input design data (curved tubes' dimensions) and output data (approximated by the boundary element model) in advance. A genetic algorithm ( GA ) is adopted as an optimizer during the optimization process. Before the GA operation can be carried out, the accuracy of the mathematical models has been checked using the experimental data. Optimal results reveal that the maximum value of the sound transmission loss ( STL ) can be improved at the targeted frequencies.