Frequency-domain and time-domain finite element modeling of the sound attenuation of a silicone earplug under industrial impulsive noise

Authors:

Abstract
Workers in many industrial environnements are exposed to impulsive noise from operations such as hammering or pneumatic tools. Hearing protection attenuation under such exposures is commonly assessed using acoustic test fixtures (ATFs), which finite element (FE) models can complement by accessing configurations and excitations difficult to test experimentally. Although transient problems can be solved in either the time or frequency domain for linear systems, these formulations have not been directly compared for industrial impulsive noise. This paper compares them for a custom-molded silicone earplug mounted in an ATF and subjected to two representative industrial impulses. Earplug insertion-loss linearity is first verified experimentally up to 140 dB SPL under continuous and impulsive excitation, confirming that linear constitutive laws are appropriate. The mechanical properties of the earplug and earcanal skin are then identified using impedance-tube measurements. A single-branch Maxwell model describes material damping consistently in both domains, while a rational approximation represents tympanic impedance. Both FE formulations predict the ATF eardrum pressure and provide spectral and impulsive attenuation indicators. Their predictions agree within a few hundredths of a pascal for the waveforms and within 1 dB for the spectra, while reproducing measured peak level, overall level, and waveform durations. The frequency-domain approach is recommended when attenuation indicators are the main objective because it provides the full spectral transfer function at much lower computational cost, although impulse-specific indicators require transforming the excitation. The time-domain formulation is more demanding but directly resolves transient pressures throughout the model and remains preferable for nonlinear hearing protectors.

This publication has 15 references indexed in Scilit: