Laminar and turbulent steady flow in an S-Bend

I am still having a look at STAR-CCM+’s tutorial, and while I will not post here every single tutorial I follow, I just came across a particular one which happens to be quite similar to one simulation I ran some months ago. So, let’s see if we can compare both solutions despite the differences.

While the simulation I ran in FLUENT was to study the effects of sand erosion in pipe bends, STAR-CCM+ focuses its tutorial in assessing the differences between a laminar and a turbulent flow within a pipe. Nonetheless, pressure and velocity distributions should look alike. Lets take a look.

Figure 1. S-Bend
Figure 1. S-Bend in STAR-CCM+. Laminar flow, Re = 500.

Although geometric dimensions of the pipe are different, the velocity profile —at least along the first segment of the pipe— varies parabolically across the flow. Well, sort of. I guess it is not exactly like the Poiseuille flow because here we impose an inlet flow velocity and the Poiseuille flow is driven uniquely by a pressure gradient. In such a flow, velocity across the flow would be

$$ u = \frac{1}{2\mu} \left( \frac{dp}{dx} \right) (y^2 – Dy) $$

where $\mu$ is the dynamic viscosity, $D$ is the diameter of the pipe, $x$ and $y$ are the Cartesian coordinates and $p$ stands for pressure.

Anyway, the velocity of the slurry analysed with FLUENT of is 10 m/s which makes the flow become turbulent. So, lets take a look at the velocity profile in STAR-CCM+’s turbulent case.

Figure 1. S-Bend in STAR-CCM+. Turbulent flow, Re = 50 000.
Figure 2. S-Bend in STAR-CCM+. Turbulent flow, Re = 50 000.

In the turbulent case, the recirculation bubble after the second bend is smaller and the pressure exerted on the wall is greater, more in line with was seen in FLUENT’s simulation.