Angle of Attack

Angle of attack is the angle between a body or lifting-surface reference line and the incoming flow direction. For aircraft it is usually written as \alpha and interpreted relative to the local relative wind, not relative to the horizon.

The reference line must be stated. It may be a wing chord line, fuselage reference line, body axis, airfoil zero-lift reference or a calibrated sensor axis. Changing the reference shifts the numerical angle even when the physical flow is unchanged.

For small longitudinal angles, a useful screening relation is:

where \theta is pitch attitude and \gamma is flight-path angle. This relation is only a simplified geometric approximation. Real aircraft require reference-line definitions, air-data calibration, local flow corrections, sensor position effects and configuration control.

Engineering role

Angle of attack matters because it drives lift coefficient, drag, pitching moment, stall margin, buffet onset, control effectiveness, short-period response, envelope protection and flight-test interpretation. A safe airspeed alone does not prove stall margin if load factor, icing, maneuvering or configuration change the required lift coefficient.

Angle of attack is also a measurement problem. Vanes, differential-pressure probes, inertial estimates and sensor-fusion systems can disagree during sideslip, high pitch rate, icing, turbulence, flow separation, ground effect or damaged configurations. A protection law that uses angle of attack must state how the value is measured, filtered, validated and handled when sensors disagree.

Lift curve and stall margin

An aircraft is in a climb with:

Parameter	Value
Pitch attitude, \theta	9.0^\circ
Flight-path angle, \gamma	3.0^\circ
Lift-curve slope, C_{L_\alpha}	5.5\ \text{rad}^{-1}
Zero-lift angle, \alpha_{L=0}	-2.0^\circ
Critical angle of attack, \alpha_{crit}	15.0^\circ
Protection trigger angle	12.0^\circ

Estimate angle of attack with the small-angle relation:

Convert the aerodynamic angle above zero lift to radians:

Estimate lift coefficient in the linear range:

The margin to the protection trigger is:

The margin to the critical angle is:

Engineering comment: this is a useful screening calculation, but not a release decision. The linear lift-curve model can fail near stall, with ice contamination, in compressible flow, with high-lift devices, during rapid pitch motion or when local sensor flow differs from wing flow. A real review must also check load factor, Mach number, Reynolds number, configuration, sensor calibration, filtering, uncertainty and validated stall-warning or protection thresholds.

Load factor and alpha demand

For a given weight W, dynamic pressure q, reference area S and load factor n, the required lift coefficient is:

In the linear range:

so the estimated required angle of attack is:

This relation explains why stall margin is not controlled by speed alone. Increasing load factor, reducing dynamic pressure, adding ice, changing configuration or degrading lift-curve slope can all move the aircraft closer to critical angle of attack.

Distinction from related angles

Angle of attack is not pitch attitude. Pitch attitude is the aircraft body’s orientation relative to a reference frame such as local level. An aircraft can have a high pitch attitude and low angle of attack in a steep climb, or a modest pitch attitude and high angle of attack during a slow approach.

Angle of attack is not flight-path angle. Flight-path angle describes the direction of the aircraft velocity vector relative to the horizon or inertial frame. Angle of attack is measured between the reference line and the incoming flow.

Angle of attack is also different from sideslip angle. Sideslip is the lateral angle between the aircraft body or velocity vector and the relative wind. Both angles matter for force and moment prediction, but they affect different axes and control problems.

Sensors, bias and filtering

An angle-of-attack signal used by a display, warning system or flight-control law is not the same thing as the aerodynamic definition. A simplified measurement model is:

where b_\alpha is bias and \epsilon_\alpha is noise, turbulence, quantization, flow distortion or processing error. Filtering can reduce noise but adds lag:

where \mathcal{F} represents the filtering or estimator logic. A few degrees of bias can matter if the protection threshold, stall warning or flight-test limit is close. Sensor disagreement logic should define voting, monitor thresholds, persistence time, degraded mode and pilot or control-law response.

What changes angle-of-attack interpretation

Angle-of-attack interpretation depends on:

• reference line, aircraft datum, wing incidence and local probe alignment;
• airspeed, Mach number, Reynolds number and dynamic pressure;
• flap, slat, gear, spoiler, store and ice configuration;
• pitch rate, gusts, turbulence, sideslip and unsteady flow;
• boundary-layer state, separation, buffet and stall progression;
• sensor location, calibration, vane friction, probe heating and icing;
• filtering, latency, estimator logic and sensor-fusion validity;
• control-law mode, protection thresholds and degraded operation.

Because angle of attack is both an aerodynamic state and a measured signal, engineering documents should separate the aerodynamic definition from the sensor or estimator implementation.

Validation evidence

Angle of attack can be measured with calibrated vanes or probes, inferred from inertial and air-data states, estimated in simulation, swept in wind-tunnel tests, or reconstructed from flight-test maneuvers. A defensible value states reference line, sign convention, units, sensor source, calibration, filtering, flight condition, configuration and uncertainty.

Useful validation evidence includes vane calibration, rig alignment, wind-tunnel alpha sweep, flight-test maneuver reconstruction, air-data consistency, sideslip range, sensor heat and icing protection, filtering delay, sample rate, control-law mode, stall-warning threshold evidence, protection trigger evidence, fault-injection tests and degraded-sensor behavior. For certification or release, nominal behavior is not enough; the evidence must include credible disagreement, bias, icing, high-sideslip, turbulence and maintenance-error cases.

Common mistakes

Common mistakes include:

• using pitch attitude as if it were angle of attack;
• quoting angle of attack without a reference line or sign convention;
• applying a linear lift-curve slope too close to stall;
• ignoring load factor when interpreting stall margin;
• trusting one angle-of-attack sensor without plausibility checks;
• comparing wind-tunnel, CFD and flight values with different reference definitions;
• validating nominal protection thresholds while leaving icing, damaged, high-sideslip or degraded-sensor cases untested.