Multi Objective Optimization of a Concentric Tube Resonator for Maximizing Acoustic Loss at Minimum Pressure Drop

Abstract

Noise control in industrial applications is one of the highest priority considering human comfort as well as comfort of all living creatures within the reach of noise source viz. industry. Automobiles causes a large amount of noise pollution through their exhaust due to violent combustion within their engines. Practice of using acoustic filter at automobile exhaust has been adopted almost for as long as automobile existed. But use of complex geometries such as a concentric tube resonator as an acoustic filter had been avoided due to compromise with pressure drop which can cause back pressure and could heavily affect the efficiency of engine. This project suggests optimizing a concentric tube design for maximum sound transmission loss that can be achieved without producing significant amount of back pressure. Optimization of muffler has been achieved by using two objective functions for sound transmission loss and pressure drop, both containing muffler dimensions simultaneously using Genetic Algorithm. Analytical solution for calculating sound transmission loss under frequency domain analysis was developed by Sullivan where he created a discrete system of coupled equations at interface of each hole at perforated concentric tube. Analytical solution for pressure drop against mass flow rate has been developed from Flow Resistance Network theory which is developed by Elnady. Flow Resistance Network Theory helps in modeling flow of fluid equivalent electrical circuit where pressure (Stagnant) and velocity(volume) are analogous to voltage and current. Benchmarking of analytical solution has been done by comparing it with FEA solution produced by ANSYS® and experimental results available in literature for different muffler configuration. Genetic algorithm used for optimization is based on binary string based population, crossed together to form new generation at each iteration which is then selected as solution for objective function.