NASA RPS Program status 2014: power as predicted

What the work claims

This is a primary program-status paper written by the NASA Headquarters and Glenn Research Center managers of the Radioisotope Power Systems (RPS) Program, which began formal implementation in December 2010. Three claims carry the paper. First, flight telemetry validates the program's power-degradation models: on five RPS missions the measured generator output matches predictions closely (Voyager 1: 265.0 W actual versus 266.4 W predicted; Voyager 2: 267.5 versus 269.2; Cassini: 647.6 W average versus 648.8; New Horizons: 213 W versus a 212 W prediction made by DOE at launch; Mars Science Laboratory: 111.8 W average versus 112).¹ Second, the Advanced Stirling Radioisotope Generator (ASRG), then the program's keystone, converts heat to electricity at on the order of 30 percent efficiency versus around 7 percent for thermoelectric designs, cutting plutonium-238 use for a given power level proportionately.¹ Third, the program's near-term pipeline is real but fragile: two additional MMRTG flight units funded, a next-generation thermoelectric RTG projected to deliver a 45 to 63 percent power increase at beginning of mission, and a plutonium-238 supply restart that in 2013 had completed only initial process demonstration steps.¹

How it works

Every radioisotope power system flown by the United States at the time of writing used the same conversion physics: the decay heat of plutonium-238 is converted directly to electricity by thermoelectric couples, a solid-state process with no moving parts and roughly 7 percent end-to-end efficiency (measurement boundary: thermal input from the heat source to electrical output at the generator, as stated in the paper).¹ The Multi-Mission RTG flying on Curiosity burns no fuel in the chemical sense; it simply harvests a decaying heat inventory. Eight General Purpose Heat Source (GPHS) modules produce a nominal 110 W-electric at beginning of mission, and output falls on a predictable curve as the Pu-238 decays (88-year half-life) and the couples degrade.¹

The ASRG replaces the thermoelectric stack with two Stirling convertors, each paired with one of two GPHS modules. A Stirling engine is a piston heat engine running a closed regenerative cycle; the convertors' alternating output feeds an ASRG Control Unit that rectifies it to the spacecraft DC bus and synchronizes the pistons to cancel vibration inputs.¹ The paper states the conversion efficiency as "on the order of 30 percent," roughly four times the thermoelectric figure, which means a Stirling generator needs about a quarter of the Pu-238 per watt. Since the heat source, not the converter, is the scarce and expensive element, efficiency translates directly into fuel inventory.¹

The prediction machinery matters as much as the hardware. Mission teams feed electrical and thermal telemetry into models that account for environmental effects, generator degradation, and natural Pu-238 decay, and use the results to project future power availability for mission planning.¹ The paper's Table 3 is the output of that loop: five missions, ten to 35-plus years of operation, and a gap between model and measurement that never exceeds about 1.7 W on a single generator.¹

The strongest case for it

The strongest reading is that Table 3 is one of the most underappreciated assets in space power engineering. Degradation models good to within half a percent, decades after launch, mean a mission planner can commit an outer-planets spacecraft to a 20-year tour and budget power the way a structural engineer budgets mass margin: against a verified curve, not a hope.¹ This is flight heritage doing what flight heritage is for. The Stirling case is equally concrete: at four times the efficiency, a Pu-238-constrained program gets roughly four times the missions from the same fuel pipeline, which is exactly the constraint the paper flags as binding, since DOE must re-establish domestic Pu-238 production and had only taken first demonstration steps by 2013.¹ The program structure, including a Multi-Mission Launch Approval Engineering element that pre-clears launch-safety documentation so PI-led missions are not gated by it, shows the office attacking schedule risk rather than chasing paper efficiency.¹

Where a skeptic should push

The single most load-bearing assumption is Stirling lifetime, and in 2014 it was exactly that: an assumption. NASA had never flown a dynamic radioisotope power system; the paper's own reliability figure for the ASRG is an estimate, and its mass and power margins are "stable" estimates of a design that had completed final design review but whose flight convertors were still being built.¹ The program's track record makes the risk vivid rather than theoretical: both Discovery 12 finalists that proposed the ASRG (the Titan Mare Explorer and the Comet Hopper) went unselected, leaving the generator with no near-term mission, its two flight units and qualification unit headed to Idaho National Laboratory for storage and long-duration testing rather than a launch pad.¹ The five-year goals table concedes that the newest round of technology investments was selected but not funded, "due to national fiscal constraints."¹ Even the good news carries a caveat: the 45 to 63 percent next-generation RTG figure is a projection for new thermocouple technology, contingent on budget appropriations and successful transfer to industry, with availability "early in the 2020s" in the paper's words.¹

Flight telemetry is an RPS design asset

The non-obvious implication is that the scarce resource in radioisotope power is not watts or even Pu-238, but verified prediction. Five concurrent missions matching model output within about a watt is what lets a program trade fuel inventory against mission duration with confidence; strip away that dataset and every outer-planets proposal carries an unpriceable power risk.¹ For anyone sizing a deep-space power budget, the lesson is to invest in telemetry-driven degradation models as infrastructure, the way this program does, rather than treating power prediction as a one-time analysis at proposal time.

The genuine threat is orphaned capability. The paper shows a generator that passed its final design review, with flight hardware in fabrication, that still needed a mission to justify existing; when the two proposing missions lost selection, the hardware path forward became storage and test.¹ Efficient hardware without a manifest decays: contractor teams disperse, qualification evidence ages, and the next mission pays rediscovery costs. The opportunity sits in the same paragraph as the risk: the next-generation thermoelectric path promises a 45 to 63 percent beginning-of-mission power increase on the same MMRTG architecture with no moving parts, and the DUFF experiment had already shown Stirling convertors coupled to a uranium-235 fission heat source through a 0.6 m stainless-steel-and-water heat pipe, with kilowatt-class fission prototype testing under consideration.¹ Efficiency without mechanism lifetime risk, and a bridge from radioisotope to fission-dynamic conversion, are both on the table this paper lays out.

The bottom line

Established by the paper, and credible because it is telemetry: radioisotope generator output on five missions tracks prediction within roughly 0.5 percent, decades into operation, and the MMRTG delivers its nominal 110 W on Mars as designed. Established as program fact: the ASRG passed final design review with flight hardware in fabrication, at a stated conversion efficiency on the order of 30 percent versus about 7 percent for thermoelectrics. Not established: Stirling lifetime in flight (no dynamic RPS had flown), the 2016 delivery date for fueled qualification testing, the projected 45 to 63 percent next-generation thermoelectric gain, or the Pu-238 supply restart, which in 2013 stood at initial process demonstrations. Confidence: high that RPS degradation modeling is an engineering strength worth copying; moderate that the ASRG would have delivered its efficiency on a mission, since ground long-duration testing was the plan precisely because flight evidence did not exist. What would have confirmed the Stirling claim: a selected mission and years of powered cruise; what broke it first, on the paper's own evidence, was not physics but selection politics and funding.