Idea 4 — Power network layout planner
15 November 2026. One of six ideas; read 00-overview.md
first. Source numbers refer to its section 8. Compiled with AI (Claude);
not re-checked by a person. "Judgement" marks an assessment.
This idea extends the working candidate from A3 (power siting with a stated confidence for one asset). It is the smallest step away from the work done so far.
Product in one line
For a set of loads and candidate positions at a site, the tool lays out generation masts, storage and distribution (cable routes or LoS power-beaming links) and returns the delivered-power availability for each consumer, with uncertainty, including consumers in shadow.
Decision it supports
A power developer decides how many masts, how high, where, and how to reach a consumer that sits in shadow. A single-asset illumination map cannot answer this, because shadows at neighbouring points are correlated.
Why now
- NASA awarded components for a small communications and solar tower in November 2026: energy storage (Honeybee Robotics), power generation (Redwire) and wireless power transfer (Yank Technologies). It expects a solicitation for integrated power systems in early 2027 [1].
- Astrobotic plans a 2028 demonstration in which a rover 500 m from a lander exchanges power with it (news) [29]. Honda and Astrobotic study 10 kW and 50 kW arrays at South Pole sites (C finding 2).
- Volta Space Technologies plans satellites that beam power by laser to receivers on landers and rovers, for the night and for PSRs (news) [30].
- Japan funds lightweight perovskite solar cells with lunar power generation as an application [7].
- The only end-customer lead of Luminous develops solar panels for the
Moon (
CLAUDE.mdsection 4).
What exists
- Research: microgrid sizing and siting near Shackleton by Aalborg University (C finding 2; A3 finding 8). A NASA Glenn tower study that maximises the LoS service area for relay and laser power [32]. A preprint that derives contact windows with terrain masking for orbital power beaming [31].
- In-house tools: LunaRay names "solar asset emplacement" (A3 finding 9).
- No reviewed tool returns a confidence on an illumination or siting result (A3 finding 9). A tool that treats several assets jointly: not found (one search beyond A3).
Data basis and checks
- Terrain: LOLA 5 m site models with 100 clones (B finding 1; A3 finding 1).
- Measured reference: dated NAC images and the polar mosaics binned by sub-solar longitude (B finding 6).
- Illumination and LoS are derived. Nothing in the core is inferred, except on interpolated terrain cells (B section 6).
Type A slice (TRL 3)
| CF | Function | Check |
|---|---|---|
| CF1 | Availability at one mast with a calibrated uncertainty | Does the predicted spread contain what dated NAC images show? (A3 section 4, point 3) |
| CF2 | LoS between two assets, and cable-route length and slope, with terrain uncertainty | LoS: predicted mutual visibility against shadow geometry in NAC images [TBD method]. Route: comparison of the 5 m model with 1 m terrain where it exists |
| CF3 | Joint availability of a layout of two or three assets | Lit or shadowed state at the asset points in the same dated image |
Out of the Type A slice: layout optimisation, mass and cost trades, dust on arrays, thermal coupling (idea 1), orbital beaming geometry.
Users outside ESA (candidates; none contacted)
- Europe, non-ESA: the Lithuanian lead [TBD name], Maana Electric, Aalborg University (C section 3).
- United States and Canada: Voyager Technologies (Astrobotic), Honda, Redwire, Honeybee Robotics, Yank Technologies, Volta Space Technologies. Several run their own tools (C section 3).
- Japan: developers of lunar solar cells under the SSF.
Four tests (judgement)
- Unique: weak. The single-asset method is published (A3 finding 1). The multi-asset step and the calibration test are the only new parts.
- Feasible: pass. Open data, a published method and measured images to check against.
- Needed: pass. Named programmes and companies on three continents.
- Fits Type A: pass. Three critical functions, one site, each with a measured check.
Link to ESA programmes and path beyond TRL 3
Argonaut Mission 1 payloads and their power supply (C finding 3). Next step: a second site, optimisation, and use by one power developer in its own design study.
Risks and unknowns
- A3 section 4, point 5: with the published clones, terrain uncertainty matters at 1 m height and is almost nil at 5 m. Ask the end customer for its mast height before fixing the scope.
- The largest users have in-house tools. Evidence that they would buy: not found (C finding 10).
- An evaluator can cite NASA's 2021 paper against novelty of CF1 (A3 section 4, point 2). The proposal must cite it first.