NASA breaks a composite truss-braced wing on purpose: failure at about 127% of limit load

NASA has loaded a composite truss-braced wing until it broke, and it failed at roughly 127% of its design limit load. NASA Armstrong Flight Research Center reported the result in an article published on 17 July 2026. The 4.6 m (15 ft) test article is SWEET-15 (Structural Wing Experiment Evaluating Truss-bracing), tested in the Flight Loads Laboratory at Edwards, California.

The wing is a long, slender span supported by an aerodynamic strut, with a secondary jury strut tying into the main structure. It is based on NASA’s earlier Transonic Truss-Braced Wing concept. NASA Langley built it with the ISAAC robot, combining five different advanced composite manufacturing and assembly technologies. NASA says this is the first time a representative composite truss-braced wing configuration has undergone this type of structural evaluation.

What the test did

Phase	What NASA reports
Loading	Over several months engineers intentionally bent the wing, applying forces progressively to represent flight conditions
Measurement	Numerous strain and load sensors, including fibre-optic strain sensors, placed throughout the structure
Correlation	Sensor data confirmed the predictions made by NASA’s computer models
Test to failure	Loads increased beyond the design limits; failure at roughly 127% of design limit load
Failure location	Visible damage near the back edge of the wing and in the upper wing cover

The work sits within NASA’s Subsonic Flight Demonstrator project.

Why it matters for design engineers

Limit versus ultimate. For certified transport aircraft, 14 CFR 25.303 applies a factor of safety of 1.5 to limit load, and 25.305 requires the structure to carry ultimate load without failure for at least 3 seconds. A research article failing at about 1.27 times limit is therefore not a certification-grade result, and NASA does not present it as one. Its value is the correlation: a full failure point, with dense strain data, to check a model against. If you run a static test, plan the test-to-failure step with the same care. That one data point tells you where your analysis is conservative and where it is not.

The load path moves. A cantilever wing takes bending to the root. A braced wing splits it. The strut reacts a large axial force and the inboard wing sees a much lower root moment. The wing box, though, now carries axial compression between strut attachment and root. That brings column and panel stability into the sizing, and the strut itself sees compression in negative-g or gust-down cases. For a quick check on a slender member, start with the classic Euler load in our Euler buckling calculator. Then knock it down for imperfections and end-fixity uncertainty.

Fittings and joints get busier. Every strut adds two primary attachment points, plus the jury strut. On composite structure those fittings are usually bolted metallic lugs into laminate, where preload scatter and bearing stress govern. Stress raisers at fastener rows and cut-outs are where tests like this usually find the first damage.

The upper cover. Failure in the upper cover is consistent with a compression-critical skin in upward bending. For stiffened composite panels, local buckling, skin-stringer separation and cut-out edges are the usual suspects.

If you	Check
Design a slender strut or brace	Euler load with realistic end fixity, Euler buckling calculator
Bolt a metallic fitting into a composite or thin skin	Preload and scatter, bolt torque and preload, and the nut factor guide
Have holes, joggles or cut-outs in a compression cover	Local peak stress, stress concentration factor $K_t$
Plan a static test	Instrument for correlation, not just pass or fail, and decide beforehand whether to take it to failure

The 127% figure matters less than the trade behind it. Bracing a wing turns a bending problem into a joint and stability problem, and NASA now has test data to calibrate models for that trade.