A New Modified k-ɛ Turbulence Model for Predicting Compressible Flow in Non- Symmetrical Planar-Curvature Converging-Diverging Supersonic Nozzle
Keywords:
Dissipation rate, k-ε RNG, k-ε standard, Turbulent diffusion flame; Lagrangian PDF; mixing model; modified EMST model; RSM model, turbulence kinetic energyAbstract
Convergent-Divergent (CD) nozzle of a compressible fluid is a common device to
accelerate fluid flow to a higher supersonic speed and to direct or modify the fluid flow.
CD nozzle has been applied in wide range of fluid equipment such as turbine power,
chemical mixing equipment, turbo-jet engine, and rocket. The performance of CD
nozzle is strongly affected by its geometry at certain pressure ratio and the flow
characteristic. In the case of complicated flow phenomena within a supersonic flow,
especially in the turbulence flow, many studies use the computational simulation to
obtain the detailed behavior and properties of flow. However, the k-epsilon turbulence
model has limitations in predicting the effect of dissipation due to the viscous friction.
This study aims to propose the new modified k-ε turbulence model in planar- curvature
Convergent-Divergent (CD) nozzle of a compressible fluid. Two equations model of
modified Standard k-ε for predicting the compressible flow within planar- curvature CD
nozzle was discussed. The simulation model was run in 2D and steady, while fluid was
assumed as an ideal gas with domain size was 0.65 m length, 0.071 m width. In
addition, it has been discretized in 3510 structured independent grid cells. The results
depicted that in the divergent section of the nozzle (supersonic region), the fluid
expansion caused the change in fluid parameters such as time-average of pressure,
temperature, density, and velocity. This study found that the expanded cross-sectional
area with non-symmetrical planar curvature affected the turbulence behavior and
properties. Furthermore, the new modified constants of c2 in dissipation equation and
cμ of eddy viscosity model could give a better prediction than the original constant of
the k-ε turbulence model.