Abstract

In order to decrease the soot particles’ harmful emission, it is necessary to use a diesel particulate filter (DPF) on cars. Recent years, many countries standardize using DPF on cars. In this paper, a multi-objective design optimization study is performed by using the nondominated sorting genetic algorithm (NSGA-II) to obtain an optimum DPF geometry. In the present study, two objective functions were determined to obtain an optimum DPF. The first was to maximize sound transmission loss (TL) values to increase the acoustical properties of the DPF, while the second was to minimize the back pressure to cause the diesel engine to work efficiently and consume less energy. The optimization problem has been constructed in the form of maximizing the sum of the values of TL at all frequencies, as optimizing the problem for each frequency separately means a new DPF at each frequency, which has a different geometry. In the literature, optimization of the DPF has been carried out for a few parameters by just changing one parameter value and keeping others stable. However, in this paper, optimization of the acoustic performance of DPFs was performed for more than one parameter in the range of their specific values simultaneously.

摘要

为了减少烟尘颗粒的有害排放，有必要在汽车上使用柴油微粒过滤器（DPF）。近年来，许多国家对汽车上使用DPF进行了标准化。在本文中，通过使用非支配排序遗传算法（NSGA-II）进行多目标设计优化研究，以获得最佳DPF几何形状。在本研究中，确定了两个目标函数以获得最佳 DPF。第一个是最大化声音传输损失 (TL) 值以增加 DPF 的声学特性，第二个是最小化背压以使柴油发动机高效工作并减少能耗。优化问题以最大化所有频率下 TL 值之和的形式构建，因为分别针对每个频率优化问题意味着在每个频率上都有一个新的 DPF，它具有不同的几何形状。在文献中，仅通过改变一个参数值并保持其他参数值稳定，对少数参数进行了 DPF 优化。然而，在本文中，DPF 的声学性能优化是同时针对其特定值范围内的多个参数进行的。