A rover-deployable lunar reactor and its honest mass bill

What the work claims

This is a conceptual design study, commissioned by the NASA Fission Surface Power (FSP) project to answer a deployment question: can a 40 kWe fission surface power system be landed by a crew-class cargo lander and moved, by a pre-deployed six-wheel rover chassis originally designed to carry a pressurized cabin, to a location at least one kilometer from the users it serves?¹ The authors' answer is a qualified yes. The system divides into three sled-mounted elements (reactor system, control system, cable and spool system) that the rover delivers in two trips, with a one km transmission line bridging reactor and users. The qualification, stated in the conclusions without hedging: the roughly 10 t design is far above the 6 t goal, and including the rover on the same lander does not fit, so the mobility system needs its own lander.¹

How it works

The reactor is a Los Alamos-assessed high-assay low-enriched uranium design: about 20 percent enriched uranium nitride pellets, a yttrium hydride moderator, and sodium-molybdenum heat pipes, producing about 250 kW thermal to deliver 40 kWe for ten years.¹ Eight 6.2 kWe gas-bearing Stirling convertors run as four dual-opposed pairs; there are deliberately no spares, and the failure of a convertor pair degrades output by about one quarter. The thermodynamic chain is accounted honestly: 18 percent reactor thermal loss, 2 percent loss from reactor to the Stirling hot-end interface, 26.1 percent convertor efficiency, 87 percent downstream power management and distribution efficiency, giving 18.1 percent end-to-end thermal-to-electric.¹ The Stirling hot end is capped at 700 C by superalloy material limits, and the cold end is held to 150 C rather than the 187 C that would maximize specific power, because alternator organic materials are not yet qualified for the hotter cold end.

Shielding is converted into distance. A lithium hydride and tungsten shadow shield of 1,250 kg protects the convertors at 1 m (tolerating 5 x 10^14 n/cm2 above 100 keV and 25 MRad silicon gamma), the control electronics sit 50 m away (10 m tolerance class: 5 x 10^11 n/cm2 and 25 kRad), and the crew stays one kilometer out, where exposure computes below 5 rem per year at 100 percent occupancy inside a one km wide corridor.¹ Those radiation figures are design assumptions in the study's distance-tolerance table, not measured doses from a built system.

The radiator is the quiet giant: 126,400 W of waste heat rejected through 133.4 m2 of accordion-deployed, double-sided, ISS-heritage radiator panels at the lunar pole, growing 62 percent to 216.2 m2 if the same plant moves to the equator.¹ Fully deployed, the radiator stands about 16 m tall and the whole reactor element reaches 17.6 m.

Power leaves the reactor element as 240 VAC, 50 Hz single-phase on eight redundant pairs of 10 AWG copper runs over 50 m, is rectified to 400 VDC, boosted by DC-DC converter units to plus or minus 2800 VDC (5,600 V line-to-line), travels one km down an aluminum cable, and is bucked to 120 VDC for the user: eight 5 kWe feeds.¹ The cable is where the voltage physics pays: holding line efficiency at 95 percent for 43.5 kWe, conductor-plus-insulation mass minimizes near 45 kg at the plus-or-minus 2800 V level; the same study notes a 3 km link would weigh 3,300 kg at 500 V but only 240 kg at 2,800 V.¹ That cable mass excludes micrometeoroid shielding and redundancy, an exclusion the authors state explicitly. The chain from Stirling terminals to user load runs at about 78 percent end-to-end.¹ A 4 kWh lithium-ion battery (4,000 W for one hour, 80 percent maximum depth of discharge, about 173 Wh/kg with commercial LG 18650 MJ1 cells in a 34S-13P string) provides reactor startup power, flowing backward through the auxiliary converter.¹

Steelman: the best case for this design

The strongest argument for the concept is that it spends distance instead of mass, and says so in numbers. Burying the reactor near the base would need a prepared hole and heavy equipment on the Moon; placing it one kilometer away with a cable trades construction machinery for 45 kg of optimized aluminum conductor plus power electronics the program needs anyway. Every thermal choice leans on flight heritage: the radiator and pump loop are ISS designs, the Stirling line descends from the Kilopower work, and the PMAD sizing uses the Metcalf model built from ISS power components.¹ Just as valuable is the accounting culture: masses carry AIAA-standard growth allowance plus a uniform 15 percent system margin, the convertor sparing philosophy and its one-quarter output degradation per failed pair are stated openly, and the conclusions admit the mass goal is missed rather than quietly redefining it. A study that reports its own shortfall this plainly is exactly the kind of document a flight program can build on.

Where a skeptic should push

First, maturity. This is a paper architecture: no element of this specific system exists as hardware, and the 18.1 percent end-to-end efficiency is an assembled chain of estimates (18 percent thermal loss, 2 percent interface loss, 26.1 percent convertor, 87 percent PMAD), each plausible, none measured here.¹ The radiation tolerance table is a set of allocation assumptions. Second, the mass miss is structural, not marginal: 10,046 kg of power elements after growth allowance and margin, plus a 2,080 kg mobility system, against a 6,000 kg goal the study says was significantly exceeded.¹ If fission's selling point is specific power, a 67 percent overshoot on the goal is the number to interrogate before any architecture debate. Third, the sparing philosophy: eight convertors, no spare, and a quarter of output lost per failed pair is a bet on Stirling reliability that a ten-year unmanned lunar night will test. Fourth, the cable mass excludes micrometeoroid protection and redundancy, so 45 kg is a floor, not a prediction. Fifth, the operational constraint is easy to miss: the crew must live inside a one km wide corridor one km from the reactor to keep the shadow shield mass that low; move the habitat or widen the footprint and the shielding grows.¹

What rover deployment changes for lunar power

Implication: the study quietly demonstrates that on the Moon, distance is a shielding and a safety technology with a known price. The 1,250 kg shadow shield only works because the one km cable at plus-or-minus 2800 V costs about 45 kg; that pairing, reactor mass traded against conductor mass at kilovolt voltage, is the actual architecture, and either half without the other fails. The non-obvious corollary for any remote lunar load, an ISRU plant or a polar water extractor, is that the cheapest kilowatt near the load may be a long thin high-voltage line from a distant source, not a local generator.

Threat: the 6 t goal versus roughly 10 t delivered reality is the early warning sign for the whole lunar fission case. Solar-plus-storage at the pole does not miss its mass goal by 67 percent in quite this way, and every additional ton of fission plant is a ton the cargo lander cannot spend on payload. If follow-on design cycles do not pull mass toward the goal, the architecture risks becoming the expensive fallback rather than the baseline, whatever the uninterrupted-power argument says.

Opportunity: the clearest named lever is the convertor cold end. The study states the system specific power optimum wants a 460 K cold end but settles for 420 K because alternator organics are not qualified.¹ Qualifying high-temperature organics for Stirling alternators is a narrow, unglamorous materials task, and it directly buys specific power for every future fission and radioisotope system, at kilowatt and megawatt scale alike.

The bottom line

As architecture, this is credible and unusually honest: the deployment concept closes, the efficiency and mass chains are transparent, and the authors publish their own shortfall. Nothing here is flight hardware; the correct maturity label is conceptual point design with heritage-derived subsystems (high confidence in the architecture logic, moderate confidence in the absolute masses until components are built). What would confirm it: a convertor pair life test under lunar-relevant duty and a radiation-tolerance validation of the electronics at the 50 m class. What would break it: a Phase 2 design that cannot approach the 6 t goal, or a Stirling reliability record that makes the no-spares philosophy untenable.