The lights stay on.
The heat gets out.

The first power trade uses 20 kW average electrical demand and a provisional 30% allocation margin, giving 26 kW for planning. Actual loads must be built from equipment duty cycles, startup surges, EVA operations, science and contingency modes. Separate essential loads from equipment that can be shed. A single average cannot establish bus, inverter or cable sizing.

At 20 kW for 72 hours, delivered energy is 1,440 kWh. At 90% discharge-path efficiency and 80% usable depth of discharge, nominal battery capacity would be 2,000 kWh before aging, temperature effects and reserve. At an illustrative pack-level 150 Wh/kg, that is about 13.3 tonnes of batteries alone. These assumptions reveal why site choice, shorter crew visits, lower loads, fuel cells and eventual fission deserve a formal trade.

Tracking arrays need an illumination and shading model, dust allowances, thermal derating, deployment reliability and recharge time. To recover energy after a blackout, generation must serve the live load and charge storage. Fission is a credible development direction, but an ARLO reactor is neither procured nor qualified. No reactor output is counted as available station power.

Most electricity used inside ultimately becomes heat. In vacuum, external heat rejection relies primarily on radiation. A screening model uses Q = emissivity × Stefan–Boltzmann constant × area × (radiator temperature⁴ − effective sink temperature⁴). The calculator assumes a cold idealized sink and emissivity 0.9. Real design must include terrain infrared, sunlight, view factors, dust, fluid loops and failure isolation; effective radiating area is not necessarily panel footprint.