Flutter Speed Prediction

Flutter Speed Prediction (aerodynamics/aeroelasticity/flutter-speed-prediction)

Use when the task is the classical flutter condition of a two-degree-of-freedom wing section: the flutter speed where the bending-torsion modes lose stability, found with the V-g method, the frequency coalescence check, and the flutter margin against the design dive speed for flutter clearance in the FAR 25.629 context. This leaf is the dynamic counterpart of aerodynamics/aeroelasticity/divergence-speed (static divergence); it is also distinct from structures/fem/modal-analysis (mode extraction) and aerodynamics/high-speed/swept-wing-aerodynamics (sweep effects), which feed and shape multi-mode flutter work.

Domain quick reference

• Binary typical section: plunge h (positive down) and pitch theta (positive nose-up) about an elastic axis at a*b from the mid-chord (positive aft; a = -0.2 is the elastic axis at 40 percent chord), with semi-chord b. The structural parameters are the mass ratio mu = m / (pi * rho * b^2), the static unbalance x_theta (the distance from the elastic axis to the center of gravity divided by b, the inertia coupling), the radius of gyration squared r_theta^2 (inertia about the elastic axis I_theta = m * r_theta^2 * b^2), and the uncoupled bending and torsion frequencies omega_h and omega_theta (stiffnesses K_h = m * omega_h^2 and K_theta = I_theta * omega_theta^2).
• Unsteady aerodynamics: the Theodorsen loads with the complex lift-deficiency function C(k) = H1^(2)(k) / (H1^(2)(k) + i H0^(2)(k)) built from Bessel J and Y series (Abramowitz and Stegun 9.1.10 to 9.1.11), evaluated at harmonic motion with the reduced frequency k = omega * b / V. Limits: C = 1 in steady flow (k = 0) and C = 1/2 at high reduced frequency. Quasi-steady aerodynamics (C = 1) is known to give erroneous pitch damping and is not used here.
• V-g method: for each reduced frequency k the flutter determinant is solved with an artificial structural damping g added to the stiffness, det[K(1 + i g) - lambda E(k)] = 0, for real lambda = omega^2 and real g. g < 0 means the aerodynamic damping stabilizes the mode (negative artificial damping would be needed to sustain harmonic motion), g = 0 is the neutral condition, and g > 0 means the mode is unstable. The flutter speed is the lowest airspeed where a mode's g rises through zero as V increases. This is a simple p-k style eigenvalue approach on the 2 x 2 aeroelastic system, a first-estimate typical section model, not a production flutter clearance tool.
• Flutter boundary: the flutter speed grows with the structural stiffness (a stiff section flutters at much higher speed, and the rigid limit is enormous), depends on the inertia coupling x_theta (no static unbalance leaves no classical flutter mechanism), on the frequency ratio omega_h / omega_theta, on the elastic axis location a, and on the aerodynamic stiffness (the circulatory V^2 terms that also drive static divergence).
• Frequency coalescence: in the classical mechanism the two modal frequencies converge strongly as the airspeed approaches the flutter boundary; the coalescence check reports the minimum frequency gap and the speed where it occurs, and the exact branch merge sits at or just above the flutter speed.
• Divergence versus flutter: flutter is the dynamic bending-torsion oscillation that loses stability at the damping crossing; divergence is the static torsional instability where the torsion stiffness vanishes at a higher dynamic pressure. For the typical section the torsion frequency collapses to zero at the divergence speed; static divergence analysis belongs to aerodynamics/aeroelasticity/divergence-speed.
• Flutter margin and clearance: margin = V_F / V_D against the design dive speed; clearance practice keeps the margin at or above 1.15 x V_D. FAR 25.629 (and the European counterpart CS 25.629) is the airworthiness context for transport category flutter clearance; the standards are referenced by name only, never reproduced (standards-map.yaml far-25 and cs-25, both reference-only).

Workflow

1. Gather the section inputs: mass ratio mu, static unbalance x_theta, radius of gyration squared r_theta^2, uncoupled frequencies omega_h and omega_theta, elastic axis location a, semi-chord b, and the flight density rho (the ISA sea level default 1.225 kg/m^3).
2. Confirm the inertia coupling: x_theta must be non-negative (center of gravity aft of the elastic axis); without it there is no classical flutter mechanism.
3. Sweep the reduced frequency with vg_damping_crossing over a range such as k in [0.05, 4] and inspect the per-mode rows (airspeed, frequency, g): both modes must be damped (g < 0) at the low-speed end, and the flutter branch must rise through g = 0 as the airspeed grows.
4. Locate the flutter speed with flutter_speed_binary, which bisects on the g = 0 crossing and returns the flutter speed, the flutter frequency, the reduced frequency, and the other mode's frequency at the flutter point. A None result means no crossing inside the scanned reduced frequency range.
5. Check the mechanism with frequency_coalescence_check: the modal frequency gap should shrink well below its low-speed value near the flutter boundary (coalescing True), confirming the classical frequency-coalescence flutter mechanism.
6. Assess the clearance with flutter_margin(v_f, v_design): a margin at or above the 1.15 clearance practice (FAR 25.629 context) is acceptable; below it the section needs stiffness or damping changes, and the analysis is re-run from step 3.
7. Confirm the deterministic behavior with the contract test scripts/test_flutter_speed_prediction.py.

Worked example

Classic typical section benchmark: b = 1 m semi-chord, elastic axis at a = -0.2 (40 percent chord), center of gravity 0.2 m aft of the elastic axis (x_theta = 0.2), r_theta^2 = 0.24, mu = 20, omega_h = 30 rad/s, omega_theta = 50 rad/s, sea level air (rho = 1.225 kg/m^3).

• V-g sweep: at k = 4 (about 14 m/s) both modes are damped, g = -0.011 for the torsion branch and g = -0.017 for the bending branch; the torsion branch rises through g = 0 between k = 0.5 (g = -0.025, about 85.6 m/s) and k = 0.46 (g = +0.002, about 89.1 m/s).
• flutter_speed_binary gives the flutter point where the damping crosses zero: V_F = 88.85 m/s at reduced frequency k_F = 0.462, flutter frequency omega_F = 41.07 rad/s, with the bending branch at 30.10 rad/s. Normalized: V_F / (b * omega_theta) = 1.78 and omega_F / omega_theta = 0.82, the classic typical section result.
• frequency_coalescence_check: the modal frequency gap shrinks from 27.7 rad/s at low speed to 8.0 rad/s at the coalescence station (about 94 m/s, 30.7 rad/s), a 71 percent convergence: coalescing True, the classical mechanism.
• flutter_margin: with a design dive speed V_D = 80 m/s the margin is 88.85 / 80 = 1.111, below the 1.15 clearance practice, so the section is flagged at flutter risk; with V_D = 70 m/s the margin is 1.269 and the clearance closes.
• Divergence context: the torsion frequency collapses to zero at the static divergence speed of about 141 m/s, above the flutter speed, so flutter is the critical instability for this section.

Related leaves

• aerodynamics/aeroelasticity/divergence-speed: the static counterpart (divergence dynamic pressure, divergence speed, and divergence margin for the same typical section geometry).
• aerodynamics/high-speed/swept-wing-aerodynamics: spanwise sweep couples bending and torsion through the flow and shifts the flutter boundary of the wing.
• structures/fem/modal-analysis: structural mode shapes and frequencies that feed multi-mode flutter analyses beyond the two-DOF typical section.

Pitfalls

• Using quasi-steady aerodynamics (C = 1) for the damping search: it is documented to give erroneous pitch damping, so the flutter crossing is computed with the full Theodorsen C(k) - do not shortcut the deficiency function.
• Reading a mode as stable because g is negative at one speed: the V-g sweep must show both modes damped at the low-speed end and the torsion branch rising through g = 0 as airspeed grows; a single point is not a stability verdict.
• Treating a None flutter speed as "no flutter": flutter_speed_binary returns None when no g = 0 crossing exists inside the scanned reduced frequency range - widen the k sweep before concluding the section is flutter-free.
• Neglecting the static-unbalance requirement: with x_theta = 0 there is no inertia coupling and no classical flutter mechanism, so a zero-unbalance section must not be pushed through the binary search as if it could flutter; a negative x_theta (elastic axis aft of the center of gravity) is rejected because the mechanism assumption breaks.
• Clearing against the margin without checking the mechanism: the frequency_coalescence_check verdict (gap shrinking from 27.7 to 8.0 rad/s in the benchmark) confirms the coalescence mechanism - a clearance margin alone does not prove the flutter boundary was found.
• Forgetting that flutter and divergence are different instabilities: the torsion frequency collapsing to zero at about 141 m/s is static divergence (divergence-speed leaf), not the 88.85 m/s flutter crossing this leaf locates.

Behavior contract (gate 3)

Run the deterministic contract test (stdlib unittest, offline):

python3 scripts/test_flutter_speed_prediction.py

The test covers the Bessel reference values, the published C(k) values and limits, low-speed stability (negative g), the torsion damping crossing, the benchmark flutter speed and its normalized classic values, the no-crossing-within-range and rigid edge cases, the coalescence verdict, the flutter margin verdicts, and invalid-input edge cases.

Compliance

• The typical section flutter analysis is public-domain textbook methodology (Theodorsen 1935, NACA Report 496; Bisplinghoff, Ashley and Halfman, Aeroelasticity; Hodges and Pierce, Introduction to Structural Dynamics and Aeroelasticity); the airworthiness context is FAR 25.629 and CS 25.629, referenced by name only, summary-only per standards-map.yaml (both reference-only).
• compliance: STANDARDS-REF, gated: false.