Suprayitno, Muhammad Yandi Pratama
Noise pollution is a serious problem as the vehicle population increases every year. The vehicle exhaust system is a major contributor to noise pollution, while mufflers play a vital role in reducing it. Among various designs, a three-pass perforated is a muffler design that is often used to reduce noise in the vehicle exhaust system. However, a good muffler design should also consider minimizing the pressure drop, where these two requirements conflict. To solve this problem, a multiobjective optimization approach was applied using a non-dominated sorting genetic algorithm (NSGA-II) to find the optimal muffler design solution. Analysis of Variance (ANOVA) was also presented in this study as a statistical tool to determine the muffler design parameters that have a significant effect. To predict the acoustic transmission loss (TL) and fluid pressure drop (PD) inside the muffler, the three-pass perforated was simulated and experimentally verified. The results present two optimal muffler designs that were selected and discussed. The best designs are TL and experiment no.26. The best design TL, produced a muffler with a noise reduction capability (TL) better than the initial design by 9 dBA or 48%, with a PD improvement of 87.3 Pa or 2%. Experimental design no. 26 offers mufflers with noise reduction capability (TL) better than the initial design by 3.2 dBA or 18%, with an improvement in PD of 19.9 Pa or 0.5%. These results offer an alternative muffler design solution that has better noise reduction capability with a small increase in PD. ANOVA results with a significance level of 0.05 show that the design hole diameter parameter has a significant influence on TL performance, as evidenced by a p-value smaller than 0.05. Meanwhile, the ANOVA results for PD performance concluded that none of the design parameters or their interactions had a significant influence on PD performance. Only the design parameter center width and its squared interaction are assumed to have a significant influence, considering the p-value is close to 0.05. © 2025 The Author(s). Published by Universiti Malaysia Pahang Al-Sultan Abdullah Publishing. This is an open access article under the CC BY-NC 4.0 International license
Department of Mechanical and Industrial Engineering, Universitas Negeri Malang, Malang, 65145, Indonesia
