Power systems

The power subsystem is sized by its worst case, and the worst case is never the day it launches. It is the coldest, most degraded, most eclipsed moment of the final year of the mission, and everything else follows from that.

Generation

Photovoltaic arrays supply almost all spacecraft power. Output is

with $S = 1361$ W/m² at 1 AU, $\eta$ the cell efficiency, $\theta$ the incidence angle, $K_p$ the packing factor (typically 0.85, accounting for gaps, interconnects and structure) and $K_d$ the cumulative degradation factor.

At 30% efficiency and 0.85 packing, a kilowatt of beginning-of-life power at normal incidence needs about 2.9 m² of array. That number scales as the inverse square of solar distance, so the same array at Mars, at 1.52 AU, produces 43% of what it produces at Earth, and at Jupiter about 3.7%. Beyond roughly Jupiter, photovoltaics stop being practical and radioisotope sources take over.

Degradation sets the real size

Arrays lose output over the mission, principally through displacement damage from trapped protons, as described in Space radiation. Coverglass thickness trades array mass against radiation shielding.

The array is sized for end-of-life power, which means it is oversized at beginning of life, sometimes by 25%. Managing that surplus is a real design problem, since an array producing more than the bus can absorb must be shunted or partly disconnected, and the shunted power becomes heat the thermal subsystem has to reject.

Storage

Every spacecraft in an eclipsing orbit needs batteries, and battery selection is driven by cycle life far more than by energy density.

The depth of discharge is where the orbit enters. A LEO satellite eclipses on almost every orbit, roughly 16 times a day, and accumulates around 87,000 cycles over 15 years. A GEO satellite eclipses only in two seasons around the equinoxes, roughly 90 times a year, or about 1,350 cycles over the same period.

Cycle life falls steeply with depth of discharge, so those two numbers force completely different answers:

The LEO satellite has the shorter eclipse and needs the larger battery, because it can only use a fifth of it each time. Sizing a battery on energy alone, without the cycling constraint, is a classic way to underspecify by a factor of three.

Lithium-ion replaced nickel-hydrogen on specific energy alone, and brought a cell-level failure mode with it. Cells are managed individually, with balancing, temperature control and isolation, because a single cell in thermal runaway can propagate through the pack.

Batteries also have the narrowest temperature band of anything on the vehicle, typically 0 to 30 °C, which makes them a driving constraint for thermal control.

Radioisotope power

Where sunlight is unavailable, a radioisotope thermoelectric generator converts the decay heat of plutonium-238 directly to electricity through thermocouples.

Efficiency

About 6%. An MMRTG produces roughly 110 W electrical from about 2000 W thermal, and the remaining 1890 W is waste heat that must be rejected, though some is usefully routed to keep components warm.

Decay

Plutonium-238 has an 87.7-year half-life, so the source declines about 0.8% per year. Thermocouple degradation adds to that, giving a combined fall of roughly 1.6% per year.

Longevity

The Voyager generators produced about 470 W at launch in 1977 and still supply enough to operate a reduced instrument set nearly five decades later.

Constraint

Plutonium-238 supply is the binding limit. Production restarted in the United States in 2013 at kilograms per year, and it constrains how many outer-planet missions can fly.

RTGs are used where solar is impossible rather than merely inconvenient: beyond Jupiter, on the lunar surface through a 14-day night, and in permanently shadowed terrain.

Bus architecture

Two topologies dominate, distinguished by how the array is coupled to the bus.

Direct energy transfer connects the array to the bus and shunts surplus current. It is simple and efficient at the operating point but does not extract maximum power when conditions move away from it.

Peak power tracking interposes a converter that continuously operates the array at its maximum power point. It extracts more power, particularly when the array is cold at eclipse exit or heavily degraded, at the cost of a converter in the main power path with its own losses and failure modes.

Peak power tracking wins where conditions vary widely, such as interplanetary missions with changing solar distance. Direct energy transfer wins where they do not, such as a GEO satellite operating at a nearly constant point for fifteen years.

Bus voltage rises with power level, because current and therefore harness mass scale inversely with it. Small spacecraft use 28 V; higher-power platforms use 50 or 100 V; the ISS distributes at 160 V primary. Above roughly 100 V, the plasma environment becomes a design driver, since exposed high-voltage surfaces can arc to the surrounding plasma, which connects the power subsystem directly to the charging behaviour described in The space environment.

Sizing in practice

The constraint most often missed is the charge rate one. It is not enough for the array to supply the average load; it must supply the load and recharge the battery within the sunlit portion of the orbit. In a low orbit with a 35 minute eclipse and 58 minutes of sunlight, that recharge requirement can size the array rather than the load does.

References

• Patel, M. R. Spacecraft Power Systems, CRC Press, 2004.
• Wertz, J. R., Everett, D. F. and Puschell, J. J. (eds.) Space Mission Engineering: The New SMAD, Microcosm Press, 2011, chapter 21.
• Bailey, S. and Raffaelle, R. “Space Solar Cells and Arrays”, in Handbook of Photovoltaic Science and Engineering, Wiley, 2011.
• Rayman, M. D. and Lehman, D. H. “NASA’s Radioisotope Power Systems Program”, and NASA RPS Program documentation on MMRTG performance.
• Borthomieu, Y. “Satellite Lithium-Ion Batteries”, in Lithium-Ion Batteries: Advances and Applications, Elsevier, 2014.