Coanda Effect Simulator — Attachment Angle of a Wall-Attached Jet
Visualize a Coanda jet attached to a curved wall. Vary jet thickness and surface radius to change the curvature parameter K=b/R and see when the jet separates.
Parameters
Jet speed U_0
m/s
Jet thickness b
mm
Surface radius R
mm
Arc extent θ_geo
°
Semi-empirical model assuming air with ρ = 1.2 kg/m³, ν = 1.5×10⁻⁵ m²/s and K_crit = 0.25.
Results
—
Curvature parameter K=b/R
—
Separation limit θ_s_max
—
Actual attachment θ_e=min(θ_geo, θ_s)
—
Vertical reaction F_y per unit width
Curved surface and jet path
Gray = arc surface / solid blue = attached jet / dashed blue = free jet after separation / red × = separation point / yellow arrow = initial and final velocity vectors
Theory & Key Formulas
A jet over a curved wall stays attached when the centripetal demand $\rho U^2/R$ is balanced by the pressure and friction. This tool models the curvature parameter and a linear separation-limit angle.
Curvature parameter K, with b = jet thickness, R = surface radius:
$$\theta_s = \theta_{\max}\left(1 - \frac{K}{K_{\text{crit}}}\right),\quad K < K_{\text{crit}}$$
Actual attachment angle θ_e and per-unit-width momentum reaction:
$$\theta_e = \min(\theta_{\text{geo}},\,\theta_s)$$
$$F_x = \rho U_0^2 b \sin\theta_e,\quad F_y = \rho U_0^2 b (1-\cos\theta_e)$$
This is a semi-empirical, simplified model. Real designs require CFD or wind-tunnel testing.
What is the Coanda Effect Simulator
🙋
When I bring the back of a spoon close to a faucet stream, the water bends as if it sticks to the spoon. What is that?
🎓
That is the Coanda effect. Roughly speaking, a jet tends to follow a nearby wall. It was reported in 1930 by the Romanian inventor Henri Coanda, which is why it carries his name. When the jet approaches a wall, it cannot entrain surrounding air on the wall side, so the pressure there drops and pulls the jet onto the surface. In the simulator, try reducing the "jet thickness b". You can see the jet bend cleanly along the curve.
🙋
Then what happens when I change the "curvature parameter K"?
🎓
K = b/R is the ratio of jet thickness to surface radius. You do not move it directly, but increasing b or decreasing R makes K larger and the jet can no longer keep up with the bend. Try lowering "surface radius R" to around 50 mm in the simulator. The separation limit θ_s_max drops sharply and a red × (separation point) appears earlier on the arc. Empirically, the limit is around K_crit ≈ 0.25; above that the jet separates immediately and never attaches.
🙋
What is the difference between the "arc extent θ_geo" and the "separation limit θ_s"?
🎓
θ_geo is the geometric extent of the surface itself, and θ_s is the fluid-dynamic limit of attachment. The actual attachment angle θ_e is whichever is smaller. For example, if the arc is only 90° but θ_s is 108°, the jet stays attached all the way to the end of the surface, so θ_e = 90°. Conversely, if the arc spans 180° but θ_s is only 50°, the jet detaches at 50° and the remaining 130° is a free jet in air.
🙋
The card at the bottom shows "F_y = 4.80 N/m". What force is that?
🎓
That is the reaction force from the change in jet momentum. When the jet redirects by θ_e, the corresponding momentum change is transferred to the wall as a reaction. $F_y = \rho U_0^2 b(1-\cos\theta_e)$, so at θ_e = 90°, $1-\cos 90° = 1$ and you get the maximum vertical reaction. The V-22 Osprey's tiltrotor control and F1 rear-wing aerodynamics actually exploit this momentum change. This tool is a semi-empirical simple model — a real design needs CFD — but the underlying mechanism is the same.
Frequently Asked Questions
The V-22 main wing sits directly under a strong rotor downwash, which costs hover efficiency and causes vibration. By shaping the wing and flap so the downwash follows the surface via the Coanda effect, designers smooth the flow and use it for thrust steering. Some research aircraft and concept studies go further and use Coanda blowing actively for thrust-vector control.
In the early 2010s, F1 teams famously used "Coanda exhausts" to steer exhaust gas along curved sidepods toward the rear diffuser inlet to generate downforce. Rear-wing flap shapes are optimized so the jet stays attached on the upper surface up to the verge of stall, squeezing the most lift (downforce) out of the wing. Managing the curvature parameter K is at the heart of the design.
Ceiling-mounted air-conditioner diffusers use the Coanda effect to keep the airflow attached to the ceiling and throw it further, promoting room-wide mixing. Dyson's bladeless fans send a fast, thin jet from a slit on the inside of a ring; the jet follows the curved outer surface and entrains 15 to 20 times its own volume of surrounding air. Both rely on the same principle of wall attachment plus entrainment.
This tool uses a semi-empirical, linear model that maps the curvature parameter K onto a separation limit angle θ_s. Real separation depends strongly on the Reynolds number, wall roughness, initial jet turbulence and external disturbances (acoustic, vibration), and a quantitative prediction needs CFD (RANS or LES) or wind-tunnel testing. The simulator is meant for qualitative understanding (more curvature means earlier separation; the momentum change sets the reaction force) and its numbers should not be used directly as design values.
Real-World Applications
High-lift devices on aircraft: STOL aircraft and carrier-based aircraft have long used "blowing flaps" that eject a jet from the leading edge of the flap and rely on the Coanda effect to keep the jet attached to the wing surface, delaying separation. Concepts such as NASA's QSRA upper-surface-blown jet flap aircraft develop enormous lift at low speeds for this reason.
Tiltrotors and thrust vectoring: Tiltrotor aircraft such as the V-22 Osprey are designed with wing surfaces and flap positions that use the Coanda effect to straighten the rotor downwash and steer thrust. Bell-Boeing studies report that Coanda-optimized wing shapes significantly improve handling response.
F1 and motorsport aerodynamics: The "Coanda exhausts" that dominated F1 in the early 2010s, the curved surfaces of rear wings and diffusers, and the sidepod profiles are all the result of engineering optimization that uses the Coanda effect to delay separation and steer air into intended directions. Managing the curvature parameter within the regulations is one of the few design levers left.
HVAC and industrial nozzles: Ceiling-mounted air conditioners, range hoods, industrial furnace gas burner nozzles, and paper-mill air knives all use the Coanda effect to carry a jet along a wall and far downstream. Dyson's bladeless fan is a textbook example: a thin jet exits a slit, follows a curved surface, and entrains a large volume of ambient air.
Common Misconceptions and Cautions
The most common misconception is to think that the Coanda effect can be explained by Bernoulli's principle alone. Popular explanations like "the wall-side flow speeds up, so the pressure drops" capture only half the picture. In reality, when the jet comes close to the wall, the entrainment of surrounding fluid is suppressed on the wall side, and the resulting low-pressure region pulls the jet onto the surface. The essential mechanism is momentum exchange via viscosity and turbulence, and Bernoulli's equation (inviscid, irrotational) cannot describe it precisely. The reason the simulator's behavior is set by K = b/R is that this is the geometric "severity of the bend" of that interaction.
The second most common error is to assume that "as long as K is small, the jet stays attached forever". The simple model in this tool caps θ_s_max at 180° (a half-circle), but in reality the jet's energy is consumed by wall friction and entrainment, so over a long enough surface the momentum decays and the jet separates. Reynolds-number dependence also matters: at low Re the laminar jet bends weakly, while at high Re turbulent mixing makes the jet more robust to bending. The linear-K model in this tool flattens all of that, so use it to understand the trend ("more curvature means earlier separation") rather than the absolute value.
Finally, note that this simulator computes only the wall-attachment limit angle and the momentum reaction force, not the jet's velocity profile or pressure distribution. Real separation-point prediction requires CFD with wall pressure gradient, viscous stress and turbulence models, and F1 and aircraft developers combine multi-million-cell RANS/LES simulations with wind-tunnel testing. This tool is a first step toward grasping the underlying phenomenon: the value F_y = 4.8 N/m is a per-unit-width theoretical maximum, not a design value for any real machine.