filmov
tv
2022-06 - Energy Dissipator - k-omega vs RNG k-epsilon
Показать описание
The simulation and discussion here are qualitative just to make some fun during the long Lunar New Year holiday with strict social distancing restrictions to curb spreading of Omicron in Hong Kong.
My previous CFD simulation "2022-05 - Energy Dissipator - Tumbling Flow in Box Culvert" was carried out using "k-omega" turbulence model. This animation compares the same energy dissipator simulation by using another commonly used turbulence model "RNG k-epsilon" in Flow 3D.
For RANS model, closure of the set of governing equations requires turbulence model. However, turbulence model is semi-empirical which requires insight to propose some terms in turbulence model and coefficients calibrated by experimental results. Therefore, different turbulence models have pros and cons in different areas of application.
Results obtained from the two turbulence models appear quite different at a first glimpse. k-omega model (upper one) gives wavy and spiky water surface, while RNG k-epsilon model (lower one) gives smooth surface showing clearly ups and downs for flow passing through roughness elements. I think that k-omega gives flow behaviour closer to reality as I tried to run LES on the same problem to obtain similar results qualitatively. RNG k-epsilon appears to average out wavy water surface to give average flow behaviour.
As a civil engineer, RNG k-epsilon model seems okay for drainage design by inspecting average flow depth and velocity. I guess unless instantaneous fluctuation of pressure or flow separation at roughness elements are of concern, k-omega model may be more suitable.
My previous CFD simulation "2022-05 - Energy Dissipator - Tumbling Flow in Box Culvert" was carried out using "k-omega" turbulence model. This animation compares the same energy dissipator simulation by using another commonly used turbulence model "RNG k-epsilon" in Flow 3D.
For RANS model, closure of the set of governing equations requires turbulence model. However, turbulence model is semi-empirical which requires insight to propose some terms in turbulence model and coefficients calibrated by experimental results. Therefore, different turbulence models have pros and cons in different areas of application.
Results obtained from the two turbulence models appear quite different at a first glimpse. k-omega model (upper one) gives wavy and spiky water surface, while RNG k-epsilon model (lower one) gives smooth surface showing clearly ups and downs for flow passing through roughness elements. I think that k-omega gives flow behaviour closer to reality as I tried to run LES on the same problem to obtain similar results qualitatively. RNG k-epsilon appears to average out wavy water surface to give average flow behaviour.
As a civil engineer, RNG k-epsilon model seems okay for drainage design by inspecting average flow depth and velocity. I guess unless instantaneous fluctuation of pressure or flow separation at roughness elements are of concern, k-omega model may be more suitable.
Комментарии