Investigation of Inverse-Turbulent-Prandtl Number with Four RNG <i>k</i><i>-ε</i> Turbulence Models on Compressor Discharge Pipe of Bioenergy Micro Gas Turbine

Abstract

Inverse-Turbulent Prandtl number (α) is an important parameter in RNG k-ε turbulence models since it affects the ratio of molecular viscosity and turbulent viscosity. In curved pipe, this highly affects the model prediction to a large range eddy-scale flow. According to Yakhot & Orzag, the α range from 1-1.3929 has not been investigated in detail in curved pipe flow (Yakhot & Orszag, 1986) and specific Re. This paper varied inverse-turbulent Prandtl number α to 1-1.3 in RNG k-ε turbulence model on cylindrical curved pipe in order to obtain the optimum value of α to predict unfully-developed flow in the curve with curve ratio R/D of 1.607. Analysis was conducted numericaly with inlet specified Re of 40900 which was generated from the experiment at α 1, 1.1, 1.2, 1.3. Wall surface roughness is not considered in this paper. With assumption that thermal diffusivity is always dominant to turbulent viscosity, higher Inverse-turbulent Prandtl number represent domination of turbulent viscosity to molecular viscosity of the flow and predict to have more interaction between large scale eddy to small scale eddy as well. The results show the use of α = 1.3 has increased the turbulent kinetic energy by 7% and the turbulent dissipation by 5% compared to general inverse-turbulent Prandtl number of 1. The value difference shows that the use of higher α on RNG turbulence model described more interaction between eddies in secondary and swirling flow at pipe curve at Re = 40900.