🌀✈️ Why is ANSYS Fluent showing POSITIVE Skin Friction where flow is clearly separated?

A NACA 0012 case with y+ < 1 + Transition SST reveals something interesting… 👀

📉 X-Wall Shear Stress becomes negative around x/c ≈ 0.645 — exactly where the separation bubble appears.

But why doesn’t Fluent’s standard Skin Friction Coefficient plot show negative values? 🤔

Is Fluent plotting |Cf|?
Or are you simply looking at the wrong component?

🔥 I break down the correct way to identify separation & reattachment using signed Cf — and why the distinction matters when comparing with literature.

#ANSYSFluent #CFD #NACA0012 #Aerodynamics #SkinFriction #FlowSeparation #TransitionSST #Airfoil #CFDAnalysis #Engineering

Yes — the X-Wall Shear Stress becoming negative is the key observation. That is the quantity you should use to identify separation and reattachment.

For this particular Fluent issue, if you can share the plot/data you mentioned, I can check the sign convention and help you make the plot directly comparable with published NACA 0012 results.

The short answer

Do not interpret Fluent's standard “Skin Friction Coefficient” plot as the streamwise signed $C_f$ unless you have verified the variable definition and wall-normal/tangent orientation.

For a 2D airfoil, separation is fundamentally identified from the wall shear stress component tangent to the surface:

$$\tau_t = \boldsymbol{\tau}_w\cdot\mathbf{t}$$

and therefore the corresponding signed skin-friction coefficient is

$$C_f=\frac{\tau_t}{\frac12\rho_\mathrm{ref}U_\mathrm{ref}^2}$$

where the sign is determined by the chosen surface tangent direction.

In your case, the fact that X-Wall Shear Stress becomes negative around $x/c\approx0.645$ is consistent with reversed near-wall flow, assuming your airfoil coordinate direction is from leading edge toward trailing edge.

Why the Fluent plot can look positive

The important distinction is between:

• wall shear-stress component in a particular direction, e.g. X-Wall Shear Stress;

• magnitude of wall shear stress;

• skin-friction coefficient, whose sign depends on how the tangential direction is defined.

A magnitude cannot become negative:

$$|\tau_w|=\sqrt{\tau_x^2+\tau_y^2}\geq0$$

So if the quantity being plotted is based on a magnitude, a separated region can still appear entirely positive. That doesn't mean separation isn't present.

However, I would be careful about saying categorically that Fluent's Wall Fluxes → Skin Friction Coefficient is always calculated as exactly $\sqrt{C_{f,x}^2+C_{f,y}^2}$. The precise Fluent field-variable definition and coordinate convention should be checked for the particular Fluent release you're using.

The safest method for your NACA 0012 case

For comparison with literature, I would create a signed streamwise $C_f$ yourself.

If the airfoil is aligned with the global X direction and you specifically want the streamwise component, use:

$$C_{f,x} = \frac{\tau_x} {\frac12\rho_\mathrm{ref}U_\mathrm{ref}^2}$$

Then plot $C_{f,x}$ versus $x/c$.

Your plot should look schematically like:

Cf
 ↑
 │       attached flow
 │      /
 │     /
0├────/───────────────
 │                  \ 
 │                   \  separation
 │                    \____
 │                         \___
 │
 └────────────────────────────→ x/c
                  ↑       ↑
              separation reattachment

The important points are:

$C_f>0$ → wall flow in the chosen positive tangential direction
$C_f=0$ → separation or reattachment point
$C_f<0$ → reversed wall flow

For a separation bubble, you should therefore see:

positive → zero → negative → zero → positive

as you move downstream.

One more important detail

For an airfoil, X-Wall Shear Stress is not always identical to the true tangential wall shear stress because the wall tangent is not necessarily parallel to global X.

For a curved NACA 0012 surface, the rigorous quantity is:

$$\tau_t=\tau_x t_x+\tau_y t_y$$

where $\mathbf t=(t_x,t_y)$ is the local unit tangent.

That becomes particularly important close to the leading edge, where the surface direction changes rapidly.

So if you want a publication-quality $C_f(x/c)$, I would use the local tangential wall shear stress, not simply X-Wall Shear Stress.

Your observation at $x/c\approx0.645$ is nevertheless very useful: if $\tau_x$ crosses zero there and the flow visualization shows a recirculation region, you have strong evidence that the separation bubble is real rather than a plotting artifact.