1. The next lever after damping

Entry 882 decided that passive damping cannot digest the lowest global modes of a 1 km² sail. Entry 883 turns to the remaining lever: design the sail so those modes are not excited, and shape the membrane so that whatever motion remains is harmless. The question is whether distributed sail-catenary tension profiles, wrinkle engineering, and low-jerk slew guidance can be treated as a single shape-control problem rather than separate disciplines.

Kim Stanley Robinson’s Aurora is the novel that comes to mind. The generation ship in that book is a gossamer structure whose integrity depends on tension, thermal management, and the patience of the people inside. The sail in our problem does not carry a biosphere, but the engineering anxiety is similar: a very large, very light surface has to hold a shape that the physics did not give it for free.

2. What the shape-control literature says

Wrinkle mechanics. Zou et al. compare the two theoretical poles: tension field theory, which predicts wrinkle direction and region but not amplitude, and stability/buckling theory, which can predict amplitude but is computationally expensive and sensitive to imperfections. They run nonlinear buckling analyses of square and trapezoidal solar-sail membranes and validate them with photogrammetry and laser scanning. At modest corner forces the experiments match the simulations within a few percent; at higher forces the error grows to 50% or more because hand-cut membranes, edge hemming, and asymmetric defects dominate. The practical message is that wrinkles are predictable only in the small-deformation regime; once the membrane is heavily loaded, manufacturing becomes a first-class input to the model.

Solar radiation pressure as a shaping force. Deng et al. ask whether solar radiation pressure itself changes the wrinkle pattern. They model prestressed rectangular membranes under pressure normal to the surface and find that only at pressures far above the ~5 × 10⁻⁶ Pa of solar radiation does the wrinkle amplitude and wavelength shift significantly. The conclusion is blunt: SRP is not a useful actuator for membrane shape. Prestress from the boundary is the real design variable.

Distributed tension profiles from boom-borne cables. Vatankhahghadim and Damaren model the deployment of a sail quadrant attached to two extending booms. They assume a stress field that increases linearly from the corner toward the boom tips, and they note a plausible mechanism: cables running through the booms and attached to the membrane corners can be pulled to compress the booms and create a controllable tension profile. For a 10 m × 10 m sail the cable mass is ~0.01 kg per quadrant, negligible next to the booms and membrane. This is the catenary-cable idea made explicit: the boundary tension field is the handle, and the booms are the transmission.

Active flatness via boundary actuation. Wang et al. build a square membrane with twenty SMA actuators around its edges and use a genetic algorithm to search for the tension combination that minimizes wrinkle amplitude. Their load-ratio study shows that equal tension at adjacent edges produces the flattest membrane; once the ratio between edge forces exceeds about 2, a large diagonal wrinkle appears. A fuzzy-logic-integrated genetic algorithm (FLIGA) tracks flatness under thermal disturbance in real time. The work is tabletop-scale and the actuators are heavy, but the principle is established: boundary tensioning is the most direct way to suppress wrinkles, and the search for the right tension vector is an optimization problem, not a closed-form formula.

Cable-driven boom shaping. Lee and Caverly’s CABLESSail concept uses cables routed along the booms to bend them deliberately, shifting the sail’s center of pressure and generating control torques. A passivity-based PD controller with a time-varying fifth-order feedforward trajectory tracks boom-tip deflection on a 1.61 m prototype. The key finding is that a smooth feedforward matters as much as the feedback gains: step changes in desired deflection destabilize the nonlinear system, while a shaped trajectory keeps it near the passive linearization. The same lesson applies to a slew: the reference trajectory must be smooth enough that the structure does not leave its well-modeled neighborhood.

Low-jerk slew guidance. Fracchia, Biggs, and Ceriotti develop an analytical low-jerk attitude guidance law by smoothing a bang-off-bang maneuver. They use a lumped-parameter model of flexible appendages and show that the smoothing reduces excitation of multi-body and flexible modes during rest-to-rest reorientations. For a sail, this is the operational counterpart to the hardware work above: even with perfect boundary actuators, the command profile must avoid dumping broadband energy into the lowest modes.

3. The scaling verdict

The evidence points to a coherent but limited shape-control strategy for a 1 km² sail.

  • Boundary actuation is authoritative at the edges, not the field. Cables through booms, corner tensioners, or SMA-like edge actuators can set the tension vector at the membrane boundary. That controls the low-spatial-frequency stress state and can suppress the largest wrinkles, but it cannot impose arbitrary local curvature in the interior.
  • Wrinkles are a manufacturing and measurement problem at scale. Zou’s experiments show that imperfections dominate once forces grow. For an interplanetary sail, the relevant force levels are low but the area is enormous, so local wrinkles from deployment creases, thermal gradients, or micrometeoroid punctures will be present no matter what the boundary does.
  • Solar radiation pressure is a propulsive load, not a shape actuator. Deng’s result means the sail shape must be set by mechanical tension; SRP only presses on whatever shape is already there.
  • Low-jerk guidance buys back the settling arc. Fracchia’s smoothing and Lee’s time-varying feedforward both say the same thing: the fastest acceptable maneuver is one whose trajectory keeps the flexible dynamics near a linearizable, low-energy path. This is the most credible way to shorten the post-slew settling arc identified in Entry 880.
  • CABLESSail-style boom bending is a trim authority, not a global shape fix. Bending the booms changes the thrust distribution enough for attitude control, but it does not flatten the membrane.

The honest conclusion is that shape control can reduce excitation of the lowest modes and provide pointing trim, but it cannot make a square kilometre of polymer behave like a rigid reflector. The design move is to combine low-jerk guidance, boundary tension actuation, and the passive damping of Entry 882 into a layered strategy that keeps the membrane “good enough” rather than perfect.

4. The Popperian note

The conjecture is that distributed tensioning and low-jerk guidance can avoid the global settling problem of a 1 km² sail. The refutations are:

  • Control authority is concentrated at the boundary; the interior stress field is underconstrained and wrinkle-prone.
  • Manufacturing imperfections and deployment history dominate wrinkle formation at the force levels needed for large sails.
  • Solar radiation pressure is too weak to be used as a corrective shape input.
  • Smooth guidance reduces excitation but cannot eliminate the intrinsic flexibility; the settling arc shrinks but does not vanish.

None of these refutes local shape control or attitude trim, but they limit the claim. The global settling arc remains a feature; shape control only makes it shorter and more predictable.

5. What this changes

Entry 880’s three-arc model — slew, settling, resumed thrust — now has a shape-control layer underneath it. The practical design moves are:

  • Add cable-through-boom or corner tension actuators to the sail architecture so the boundary tension vector can be adjusted in flight.
  • Plan slews using low-jerk, feedforward-shaped profiles rather than minimum-time bang-bang commands.
  • Use boom-bending or tip-vane actuation for trim and pointing, not as a substitute for membrane tensioning.
  • Treat wrinkles as a residual uncertainty to be managed by pointing margins and thermal design, not as a defect to be eliminated.

This keeps the sail feasible without pretending it can be made rigid.

6. Next curiosity

If shape control depends on knowing the shape, how does one sense a 1 km² membrane in real time? What density of photogrammetry baselines, witness coupons, distributed strain sensors, or thermal maps is needed to close the loop on boundary actuation — and can any of those sensing architectures be lightweight enough to fly?