Sensitivity of acoustic predictions on the mesh size of large-eddy simulation of supersonic jet

Abstract

Intense noise produced by the supersonic jet plume from a rocket can cause damage to its structural system and even a failure during launch. In order to predict the acoustic loading by rocket jet noise, large-eddy simulation (LES) is commonly employed for analysis of the jet flow field that is subsequently used as a source condition for predicting the acoustic field. Although desirable from the accuracy standpoint, LES of a supersonic jet is often burdened by an overwhelming computational cost in both runtime and memory. In this talk, we discuss the sensitivity of acoustic predictions on the spatial resolution of LES, and propose a guideline for reducing the mesh requirement. . Here, the flow field of a supersonic jet is calculated using a hybrid RANS/LES scheme on meshes of different sizes, and is fed into the Helmholtz-Kirchhoff integral for prediction of the acoustic farfield. Changes in overall sound pressure level (OASPL) and directivity of the Mach radiation are monitored as the LES mesh varies in size. The study suggests that (a) the accuracy of the acoustic prediction is not much affected by the use of a coarser mesh to a certain extent and (b) the resulting improvement in computation speed and economy thus outweighs the loss of details in LES, given the prediction of acoustic loading as the ultimate goal. (This work was conducted at High-Speed Vehicle Research Center of KAIST with the support of the Defense Acquisition Program Administration and the Agency for Defense Development under Contract UD170018CD.)Intense noise produced by the supersonic jet plume from a rocket can cause damage to its structural system and even a failure during launch. In order to predict the acoustic loading by rocket jet noise, large-eddy simulation (LES) is commonly employed for analysis of the jet flow field that is subsequently used as a source condition for predicting the acoustic field. Although desirable from the accuracy standpoint, LES of a supersonic jet is often burdened by an overwhelming computational cost in both runtime and memory. In this talk, we discuss the sensitivity of acoustic predictions on the spatial resolution of LES, and propose a guideline for reducing the mesh requirement. . Here, the flow field of a supersonic jet is calculated using a hybrid RANS/LES scheme on meshes of different sizes, and is fed into the Helmholtz-Kirchhoff integral for prediction of the acoustic farfield. Changes in overall sound pressure level (OASPL) and directivity of the Mach radiation are monitored as the LES mesh varies in s...