A turbulence-based prediction technique for flow-generated noise produced by in-duct elements in a ventilation system

Abstract

A number of investigators have tried to develop a generalized prediction method for flow-generated noise produced by in-duct elements in a ventilation system. Most of these prediction methods relied on limited data obtained from conventional measurement techniques that require the use of an expensive and specially combined acoustic and aerodynamic experimental facility. An alternative to using a specialised and aerodynamic facility that is currently gaining favour in building engineering is using computational fluid dynamics (CFD) software packages. CFD is a powerful design tool that is able to predict the behaviour of fluid flow regimes. With the aid of CFD, Mak and Oldham have developed a predictive technique that is based on the relationship between the acoustic power radiated, due to the interaction of air flow and a spoiler, and the turbulent kinetic energy generated in the region of the spoiler. Based on the results of CFD simulation of relevant configurations, the technique has been adopted to normalize the published experimental data of Nelson and Morfey, who produced a normalized spectrum for predicting the sound power level of flow-noise produced by the strip spoilers in a rectangular air duct. In this paper, the theoretical basis of this technique was reviewed and revised. A collapse of data from the simulation models were obtained against the experimental data of Oldham and Ukpoho. The data collapse for a damper were generally excellent at most Strouhal numbers. The data collapse for an orifice plate were generally excellent at lower Strouhal numbers but was less efficient at higher Strouhal numbers where considerable scatter was observed.