1. The question left open

Entry 875 concluded that a hybrid sail/SEP tug can run its sail actuators and its electric thruster together, as long as their startup transients do not overlap. That scheduling rule implies the SEP must sometimes throttle down or off while the sail slews to a new attitude. During that interval the thrust profile is not the nominal continuous low thrust. What should the guidance loop do?

This entry treats the scheduling rule not as a power-system problem but as a trajectory-control problem.

2. Scheduled interruptions are not anomalies

Pure SEP missions already live with planned thrust gaps. BepiColombo interrupts electric propulsion once a week for about eight hours so the high-gain antenna can be pointed at Earth for tracking and command. SMART-1 interrupted thrust for multi-day intervals during its lunar spiral to keep the Moon from blinding its star trackers. Psyche reserves roughly 0.5 km/s of its low-thrust Δv budget for missed-thrust events, angular-momentum management, and the operational interruptions that come with pointing the spacecraft for communications or science.

These are not failures. They are constraints the trajectory is designed around.

For a hybrid tug, a scheduled SEP interruption so the sail can reorient is the same class of event. The difference is that the sail does not simply go silent. It keeps producing acceleration, and that acceleration changes direction as the sail slews. The interruption is not a coast arc; it is a bounded-thrust arc with a known attitude profile.

3. The missed-thrust design framework

Sinha and Beeson pose missed-thrust events as a bi-level optimal-control problem. An upper-level problem generates a nominal trajectory under full control authority; lower-level problems compute the optimal recovery maneuver for each possible interruption. Under regularity conditions they derive a robustness certificate: a bound on the maximum missed-thrust duration for which the linearized recovery model remains valid. The certificate depends on quantities that can be precomputed along the nominal path.

Their MTEs are anomalies, but the structure applies just as well to scheduled outages. The nominal trajectory should be optimized knowing that, at certain times, control authority will be reduced or redirected. The recovery maneuver is not an afterthought; it is part of the design.

4. What the sail does during the interruption

When the SEP is off for a sail slew, the sail thrust vector rotates. If the slew is large, the sail may pass through an edge-on attitude where thrust drops nearly to zero. If the slew is small, the thrust may only change direction. The guidance loop needs a model of the slew kinematics.

A conservative approach treats the interruption as a pure coast arc. That is safe but wasteful. A better approach treats it as a thrust arc with bounded, known acceleration. Heiligers’s hybrid sail/SEP work for non-Keplerian orbits points to the same division of labour: the sail provides continuous, slow, propellant-free acceleration, while the SEP provides fast vectoring in directions the sail cannot reach. During a sail slew the roles invert temporarily: the sail becomes the time-varying disturbance and the SEP, once it returns, becomes the recovery actuator.

5. Replanning strategies

There are at least four ways to handle the interruption:

  • Coast through. Assume zero thrust during the interruption, propagate the state, then resume the nominal control. Simple and robust, but it accumulates tracking error and may cost propellant to recover.
  • Extend the burn. Like BepiColombo extending its thrust arcs after a power anomaly, the SEP simply fires longer later. This works when the interruption is short compared with the arc and the terminal state has slack.
  • Reoptimize from the post-interruption state. A reduced-order onboard solver recomputes the remaining trajectory using the nominal solution as an initial guess. This needs computational margin and a good state estimate, but it is the most flexible option.
  • Precompute a family. Design the nominal trajectory plus a set of recovery branches for the expected interruption durations and start times. This is the bi-level approach turned into flight software.

For a minimoon tug with light-times of minutes to tens of minutes, the last two options are the honest ones. Ground can design the family; the spacecraft can select or refine the branch autonomously.

6. The Weir echo

Andy Weir’s The Martian turns on the Rich Purnell maneuver: a single, carefully timed impulsive burn that alters Hermes’s trajectory so the crew can rescue Watney. The burns are planned on Earth with light-times short enough that the crew can execute them on command.

Our tug lives in the opposite dramatic register. There is no single heroic burn. There are thousands of small corrections spread over months or years, each one recomputed after the last scheduled interruption. The tension is not whether one burn will work; it is whether the accumulated arithmetic of interruptions and recoveries stays inside the capture funnel.

Where the novel is wrong for us: the tug cannot wait for a reply from Earth before replanning. The interruption and the recovery happen in the same control cycle.

7. The Popperian note

The conjecture is that scheduled SEP interruptions are a trajectory-design variable, not a disturbance. The refutations would be:

  • A sail slew so slow or so unpredictable that its duration cannot be bounded, making the nominal trajectory impossible to design around it.
  • A sail thrust during the slew so low or so directionally uncertain that the only honest model is a coast arc, with a performance penalty too large for the mission margin.
  • A recovery maneuver that consumes more propellant or time than the margin allows, forcing the scheduling rule from Entry 875 to be tightened beyond what the mission can afford.

If any of those holds, the architecture must change: a faster sail actuator, a larger SEP margin, or a trajectory family that avoids the need for large sail slews during critical arcs.

8. What this changes

Entry 875 said to schedule actuator and SEP transients to protect the bus. Entry 876 says that schedule must be an input to trajectory optimization. A hybrid tug should not fly a nominal SEP trajectory and then hack in sail-slew interruptions; it should design the trajectory with the interruption profile included.

For the keeper arc, this means the flight software needs at least a reduced-order replanner and a precomputed family of recovery branches for bounded interruption durations. The guidance loop must also model the sail’s time-varying thrust during the slew, not just the SEP on/off state.

9. Next curiosity

How do we model the sail thrust vector during a slew, and does the changing solar pressure help or hurt the recovery? In other words, is a sail slew a coast arc, a bounded-thrust arc, or a small controlled burn in its own right?