A permanent lunar facility requires a hybrid power architecture: fission for continuous baseload, solar arrays for low-cost daytime energy, regenerative fuel cells for eclipse and contingency storage, and radioisotope systems only for small, isolated loads. Batteries alone are not an adequate foundation for a 1,000-year settlement.
1. Design basis
A conventional lunar day-night cycle lasts approximately 29.5 Earth days: about 14.75 days of sunlight followed by 14.75 days of darkness. The lunar night can last roughly 100 hours near some polar locations and up to 367 hours near the equator, depending on terrain and site geometry.[8]
A settlement should therefore be designed around four power classes:
- Life safety: environmental control, thermal survival, communications, fault detection, and medical systems.
- Continuity loads: habitat pressure systems, computing, water processing, refrigeration, and thermal control.
- Industrial loads: excavation, oxygen extraction, metals processing, construction, and propellant production.
- Expansion loads: new habitats, manufacturing, transport infrastructure, and reserve capacity.
An initial permanent outpost should target at least 40–100 kWe continuous electrical capacity, with 10–20 kWe reserved for life safety and the remainder available for production and growth. A mature industrial settlement will require megawatts, not kilowatts.
2. Solar generation
### Output and sizing
At the Moon’s distance from the Sun, incident solar power is approximately 1,361 W/m² before conversion losses. Space-qualified photovoltaic systems commonly achieve roughly 25–35% cell efficiency, but delivered array output is lower after wiring, power electronics, dust, thermal, and pointing losses.
A practical planning value is:
- 250–350 W/m² peak electrical output
- 150–250 W/m² average usable output after system losses and operating constraints
At a favorable polar site, terrain can provide near-continuous illumination on elevated ridges, but no site should be assumed to have uninterrupted sunlight without detailed horizon mapping. Arrays should be distributed across multiple ridges or towers to reduce local shadowing and single-point failure.
For a 100 kWe daytime system, approximately 300–600 m² of active photovoltaic area is required before redundancy. With a design margin of 2–3 times for degradation, dust, shadowing, and failures, the installed area should be approximately 1,000 m².
### Degradation
Solar arrays are not a one-time installation. Their output declines through:
- Radiation damage from protons and heavy ions
- Ultraviolet exposure
- Thermal cycling
- Micrometeoroid impacts
- Electrostatic dust adhesion
- Mechanical fatigue in deployment systems
- Power-electronics and cable failures
For long-lived space photovoltaic systems, a planning assumption of approximately 0.5–1% annual power degradation is prudent until a lunar-specific environmental qualification record exists. At constant 1% annual degradation:
- After 10 years: \(0.99^{10} \approx 90\%\) of initial output
- After 25 years: approximately 78%
- After 50 years: approximately 61%
- After 100 years: approximately 37%
At 0.5% annual degradation:
- After 50 years: approximately 78%
- After 100 years: approximately 61%
These figures exclude catastrophic damage and dust-related derating. A 1,000-year plan cannot depend on the original array. Even at an unrealistically low 0.5% annual degradation, only about 0.7% of initial output would remain after 1,000 years.
### Required architecture
Solar generation should therefore be treated as a renewable industrial asset with continuous replacement, not permanent infrastructure.
Recommended provisions:
- Replaceable panel modules with standardized mechanical and electrical interfaces
- At least 3 independent array fields
- Spare modules sufficient to replace 10–20% of installed capacity
- Robotic inspection and cleaning
- Elevated or vertically oriented arrays in dusty regions
- Radiation-hardened central power electronics
- Buried or shielded cables
- Local DC microgrids to prevent one fault from disabling the settlement
- Manufacturing capability for glass, aluminum, conductors, substrates, and eventually photovoltaic cells
NASA is developing vertical solar-array concepts paired with regenerative fuel cells to deliver approximately 10 kW to permanently shadowed regions, demonstrating the relevance of distributed solar-plus-storage systems for lunar operations.[1]
3. Nuclear fission power
Fission is the essential backbone for a permanent settlement because it provides power independent of sunlight, dust, local terrain, and the 14-day night.
### Kilopower
NASA’s Kilopower program demonstrated a small fission-power concept using a uranium-235 reactor, heat pipes, and Stirling engines. The reference technology targeted approximately 1–10 kWe per unit.
Kilopower-class units are valuable for:
- Early robotic bases
- Remote mining sites
- Emergency power
- Distributed polar facilities
- Redundant life-support power
- Permanently shadowed crater operations
A settlement requiring 100 kWe should not rely on one reactor. A practical arrangement would be:
- 10 × 10 kWe units, or
- 4–6 units in the 10–25 kWe class
Distributed units improve survivability and allow maintenance without shutting down the entire base. They also permit staged construction: one reactor supports initial infrastructure; additional units follow industrial expansion.
### Fission Surface Power
NASA’s Fission Surface Power program targets approximately 40 kWe of continuous lunar electrical power per system, with a design life of roughly 10 years in the early reference architecture. Fission surface power is intended to operate through lunar day and night and is not dependent on solar illumination.[1]
A permanent facility should treat the 10-year design life as a replaceable module interval, not the life of the settlement. The reactor core, power-conversion system, radiators, control electronics, and heat pipes should be designed as separately replaceable or independently redundant subsystems.
For a 100 kWe base:
- Minimum: 3 × 40 kWe units, allowing one unit