Let’s look at the most popular way to use nuclear power in space. This does not directly using the heat of the reactor like in a nuclear thermal rocket.
Most people that have looked into in-space nuclear power know that using a nuclear reactor for electrical power is nothing new. The US has flown one nuclear reactor (SNAP-10a); and the USSR and Russia have flown over thirty, using two different designs. All of these systems were designed to provide electrical power for a satellite. In the case of the US it was an experimental version of the Agena spacecraft; and for the USSR and Russia, it was used for RORSATs (Radar Ocean Reconnaissance SATellites). Radar is a notorious power hog, since resolution and power supply are directly proportional to each other. When these designs were first flown, photovoltaic panels were not nearly as efficient as they are today, so they weren’t considered a practical option, or the altitude of the satellite was so low that the air resistance on the panels would cause it to deorbit too soon to be useful.
So how is nuclear electricity generated in space?
Nuclear Electricity Production On Earth
The vast majority of people are familiar with the way a nuclear power plant works on Earth: a reactor heats up water from the heat generated by fission in fuel elements (rods, bundles, or pellets, usually), which then goes through a steam generator. This steam then drives a set of steam turbines, which spin a generator and produce electricity. That water is then run through a cooling system (either cooling towers or a body of water), before being returned to the reactor core to start the process over again. Most reactors use what’s known as a two-loop system, meaning that the water that goes through the generator never enters the reactor core, instead the water from the core is run through a heat exchanger that transfers this heat to a second loop of water (the primary and secondary loops, respectively), to prevent the generator from becoming radioactive over time, while greatly simplifying the maintenance of the power plant. This is known as a two-loop pressurized water reactor (PWR), and is the most common type of nuclear reactor in the world. Another version actually produces the steam for the generators in the reactor vessel itself, and then runs this through the generator. This much simpler system is called a boiling water reactor (BWR), and has the advantage of mechanical simplicity; but adds the challenge of the turbine becoming radioactive over time, making maintenance more difficult. This design is most common in Japan. However, the systems are very similar: heat water, produce steam, run turbines, cool the water, return it to the core.
One notable exception to this general trend for deployed reactors is in the UK, where gas cooled reactors are the most common type of reactor (in this case the AGR, or Advanced Gas Cooled Reactor). We’ve been looking at gas cooled reactors (sometimes known as HTGRs, or “High Temperature Gas Cooled Reactors”) a lot in the previous few months, so this concept should be familiar to regulars of Beyond NERVA. If you aren’t familiar, check out our Nuclear Thermal Rocket page for more information. The fuel elements have a different shape, and helium is used instead of hydrogen (as we’ve seen, if you can get away with NOT using hydrogen, that’s a good thing).
However, as many readers of this blog are familiar, nuclear power is going through a renaissance worldwide right now, with many reactor concepts that were proposed, but never deployed, being heavily investigated once again. Terms like “Liquid Metal Fast Breeder Reactor,” “Aqueous Homogeneous Reactor,” the ubiquitous “Molten Salt Reactor,” and others are being discussed, each with their own set of advantages and disadvantages over the current PWRs and BWRs that have been deployed worldwide. Even the HTGR is being talked about more, despite the fact that it’s been deployed in the UK fleet for decades, for many of the same reasons these other reactor designs are being investigated: it’s possible to make each of these designs “walk-away safe”, among other safety features inherent in their design, making nuclear accidents far more unlikely.
This doesn’t mean that these reactors are maintenance-free, however. Many of the components, especially the generators, require regular, sometimes quite difficult, maintenance. Some of the more advanced designs also have unique subsystems that will require maintenance or replacement (which is why many of the reactor designers are making modular designs, where the core itself, or peripheral components, can be disconnected, loaded onto a barge or truck, and delivered to a dedicated maintenance facility, while a replacement module is installed at the power plant) on a timeline that isn’t always clear – but will be once these reactors are being deployed for power generation. This brings us to the biggest driving factor in space nuclear power supply design, and leads the conversation to what we’re all here for: in-space nuclear power.
In-Space Fission Power Systems
With the notable exception of the Hubble Space Telescope; a few military and reconnaissance satellites retrieved by the now-long-retired Space Transport System (otherwise known as the Space Shuttle); and the various space stations launched by the US, the USSR/Russia, and now the Chinese; if something breaks after it’s launched into space, then it stays broken. Satellite operators have gotten very good over the decades at working around malfunctions, but the fact remains that on-orbit repair is by far the exception, not the rule. It’s difficult, if not impossible, to send a repair crew out to make repairs; and often those repairs would be impossible, even if it were possible to get to the satellite, due to the design considerations that went into the satellite in the first place.
This is just as true for an in-space nuclear reactor as any other system, and malfunctions have occurred on several missions involving fission power systems in space. Possibly the best known on-orbit failure is that of SNAP-10a, the only American nuclear reactor to ever fly in space. Shortly after achieving orbit, and after initial check-out of the spacecraft, the reactor was activated, and operated normally… for 14 minutes. Then, an electrical bus on the spacecraft (not part of the nuclear reactor) failed, and the entire mission was lost, including telemetry on reactor behavior. It still follows its original polar orbit, and will continue to do so for thousands of years, or until someone retrieves it at some future date (a challenging prospect for any satellite, and this was a larger-than-average one). Other failures have occurred as well, but we’ll look at these as we look at the individual systems that have been used in space so far, as well as proposed and tested systems.
There are two main concerns for space systems engineers: first, the satellite has to survive launch and deployment on orbit; and, second, everything on the spacecraft has to be as reliable as possible. The first consideration isn’t something that we are going to be examining in depth on this blog (at least not the near future). The second part, on the other hand, is something that affects most design decisions that are made in fission power supply design for space missions of all types. Maintenance is not an option in space.
Historical Fission Power Plant Programs
There have been many programs over the years looking at astronuclear power plants, from both the US and the USSR/Russia, as well as a few lower-key, less prolific programs in other countries.
The USSR has the most extensive program in terms of launches, with four major power plant designs: Romashka, BES-5, TAL-5, and ENISY. Romashka was an early prototype design that never flew. BES-5, better known as Buk, was the most successful astronuclear design ever, having flown 31 times powering the US-A RORSAT satellites. TEU-5, also known as TOPOL or TOPAZ-I in the West, flew twice on the PLASMA-A testbed satellites. Finally, ENISY was ready to fly at the collapse of the Soviet Union, and in an ironic bit of nuclear diplomacy, several reactors were purchased by the US, potentially to power American lunar bases or an experimental nuclear electric spacecraft. While it never flew, it is the only astronuclear fission reactor to be certified as flight-ready by both the US and Russia, an impressive feat.
In the US, there were quite a few programs looking at astronuclear power plants. The biggest of these programs was the Systems for Nuclear Auxiliary Power, or SNAP, program, which ran from 1958-1978. There were four major reactor programs within the SNAP umbrella: SNAP-2, SNAP-8, SNAP-10, and SNAP-50. These ended up blending, evolving, and in some cases replacing each other: SNAP-10 became SNAP-10A after the reactor core from the SNAP-2 was used instead of the original core, the SNAP-8 became the Advanced ZrH Reactor, and SNAP-50 became SPUR.
Many other designs were proposed over the years, with most Department of Energy National Laboratories all having their own programs. Each one was developed to a greater or lesser extent, depending on funding and institutional support. However, the majority of these programs did not enjoy strong support, and as such weren’t very well developed. Perhaps the best-developed design in the late 1970s to early 1990s was the SP-100 reactor.
Since the mid-1990s, high-powered fission power plants for nuclear electric spacecraft have come back into vogue, as it were. In the US, a good deal of work was put into Project Prometheus, which was a power plant designed to power the Jupiter Icy Moons Observer, or JIMO, mission. In Russia, the TEM power plant is continuing to be developed to power an astronuclear electric propulsion tug. Both of these designs are very powerful, with many advanced features, and we’ll look at them in depth soon.
Perhaps the biggest news in fission power plants, though, is Kilopower. This compact, simple reactor became the first new fission power plant of any sort to be tested in the US since the 1970s in January 2018. Since then, the Kilopower team has continued to refine the fine parts of the design, prepare the studies to prove flight-readiness, and work on the first mission that the power plant will be used on – be it a lunar surface mission or an orbiter mission.
Check out our reporting on Kilopower here! An update on the Kilopower program should be coming late 2018/early 2019!
Components in a Fission Power Plant
Every nuclear power system dedicated to electricity production has a number of parts in common: the reactor core and its associated control systems and primary coolant, the power conversion system (which often has its own working fluid or coolant), a power conditioning unit that ensures the correct wattage, voltage, and amperage are delivered to the different components of the spacecraft, and a heat rejection system. If the power plant is going to be used for nuclear electric propulsion, the electricity from the power conversion system is then sent to an electric propulsion system.
The geometry of the reactor core is often quite different from that seen in terrestrial reactor designs. The biggest difference is that it will be far smaller; often only a couple meters long and wide, or smaller; as opposed to the dozens or hundreds of meters that large terrestrial fission plants have. This is, of course, to save mass of the spacecraft, and to get more power out of less mass and volume. We’ll examine the different options for core geometry more as we go through this blog series, but the shape ranges from plugs of uranium oxide or carbide arranged in a row, surrounded by a power conversion system that we’ll look at later in this post; to cylindrical, square, or hexagonal fuel elements with coolant channels running through them (as we’ve seen with the Kilopower reactor, although that’s unique in that there’s just one fuel element, not several); to more exotic options, which all depend on a number of factors.
The biggest challenge in designing a reactor core is balancing the power output from the reactor – in the form of a given amount of heat, measured in thermal kilowatts, or kWt – with the temperature of the coolant as it’s leaving the reactor core to go to the power conversion system. The fuel elements can only get so hot before they break down, and can only conduct so much heat to the coolant at a given speed, so these two considerations drive the size of the reactor core more than most.
A nuclear reactor doesn’t produce electricity directly. In most cases, what it does produce is heat (caused by the collision of the fission fragments and neutrons as they slow down inside the fuel element, or by hitting reactor components), which then has to be converted into electricity. There are a number of ways to do this, many more than are used on Earth in nuclear power plants. For the terrestrial reactors, it’s often a matter of using off-the-shelf technology for other power plant types – a nuclear reactor builder will likely buy their steam turbines from a company that makes steam turbines for coal-fired power plants, for instance (although for a single-loop system they have to be manufactured with much higher tolerances, because working on them will be much harder; so they have to break less, and require less maintenance).
In space, the field is much more wide open, partially because there’s very little off-the-shelf going on in space. This means that a reactor designer has a lot more options when it comes to their power conversion system. However, these designs tend to fall into three broad categories: heat engines, thermal-electric materials conversion, and charged particle conversion.
A heat engine is a mechanical device that converts thermal energy, like heat, into mechanical energy. This usually involves turning heat into either a piston or a rotating shaft. After that, this mechanical energy is used to run an electric generator.
Thermal-electric materials conversion has been the most popular form of power conversion, since it doesn’t use any moving parts. the downside is that the power conversion efficiency (how much thermal energy is converted into electrical energy) tends to be very low; for many missions, though, this is fine. The heat is used to keep different thermally-sensitive components at the proper temperature, and having no moving parts means that the power conversion system – as a general rule – is more reliable.
Finally, there are systems that use moving charged particles (or particles that act as tiny magnets even if they don’t have an overall charge) to generate electricity. These particles move through magnetic fields and generate electricity. There are a number of different ways to do this, from magnetohydrodynamic power conversion systems to Alkali Metal Thermal Electric Convertors, each with their own advantages and disadvantages.
For more information, check out our Power Conversion Systems page!
As we all know, fission generates a lot of heat very efficiently, that’s the whole reason that it’s so attractive. Unfortunately, that heat all has to go somewhere, or two things will happen: the power conversion system stops working because the hot and cold ends approach the same temperature; and, your spacecraft gets baked until it’s destroyed. Heat always transfers from a higher temperature to a lower one, and how much heat is transferred depends on the temperature difference, the surface area of the thermal difference, and the thermal conduction of the materials involved. There are three mechanisms for heat to transfer: conduction (through physical contact with another material), convection (through gas heating up and then rising away from the heat source), and radiation (the infrared photons that actually make up the heat leave the substance in straight lines). Sadly, in the vacuum of space only radiation works, unless you’re throwing large amounts of gas overboard (as in an NTR, which operates on convection and radiation), which weighs a fair bit and takes up a lot of volume – and the container weighs a fair bit, too. This means that if we want to generate electricity, we have to get rid of heat through radiators, which heat a working fluid or other substance, and then use convection or conduction to an area with a large surface area to maximize the amount of surface area available for radiation. Radiators emit heat as a function of the temperature of the liquid (or gas) inside them, the surface area of the radiating surface, and several other factors, known as the Stefan-Boltzmann law (more on that in the heat rejection blog post coming up in a few posts). On the practical side of a radiator, if the substance used as a working fluid freezes at ambient temperatures, the designer needs to make sure it’s not going to freeze anywhere in the radiator and gum up the works.
For more information, check out our Heat Rejection page!
Why Use Fission Power Systems, When Most Missions Use Less Power than My Microwave?
So what needs a lot of power in space? Not much, really. The ISS uses about the same amount of power as 14 normal American houses, but those are notorious for power loss, and this is to support multiple humans and dozens of experiments. Curiosity, NASA’s current flagship mission, is an SUV-sized, six wheeled rover with: two chemistry labs; two sets of cameras; a drill; a robotic arm; and a laser that vaporizes pinpricks of rock from dozens of feet away, and analyzes the resultant color of the vaporized rock to determine the composition. All of this is done with only 110 watts (admittedly, it only does one of these things at a time, and would be challenged to travel faster than a small tortoise, but still). Instrumentation as a general rule doesn’t need a lot of power, and NASA and other space agencies over the years have gotten very good at using less and less power to do more and more.
There are really only four reasons for high power use: life support, in-situ resource utilization (ISRU), radar, and propulsion. Life support is a major focus of the Kilopower program, with a strong possibility that the first deployment of the system will be to the surface of the Moon, to support a crewed base (not a new idea, when we cover the BES-5, otherwise known as TOPAZ-II, we’ll look at a NASA mission concept that used a Soviet-built reactor to do the same thing). However, since while on-orbit in the inner solar system most life support systems can be supported by solar panels (unless there’s weather or an inconveniently long day-night cycle) those will likely be used.
In situ resource utilization is a very hot topic right now: most Mars sample return missions assume that either incredibly power-dense propellants will be used, or much of the propellant will be manufactured on Mars (often using some small portion of the raw materials brought from Earth, such as hydrogen feedstock, which is hard to get on Mars easily). Other uses include using either lunar or Martian regolith to construct habitats using 3D printing; extraction of valuable resources, either for their monetary or practical value; producing tools and other products for use on future crewed bases… the list is extensive. The other thing that’s extensive is the power requirements for this type of activity, not only for the process itself, but also for associated materials and tools needed. Many forms of chemical synthesis or extraction need precursor chemicals, solvents, reagents, and other substances that also have to be extracted, refined, and used. If a furnace is needed for metal refining, that furnace is likely going to have to be built in situ in order to be able to begin the actual end goal of refining metals, then the molds, forgings, or other associated metalworking tools to make a useful product need to be built as well. All of these things require large amounts of energy, and may require acres of solar panels (with intermittency issues, maintenance issues, and the cost of manufacturing THOSE in situ as well), or can be satisfied with a relatively small, power-dense fission power system.
In Earth orbit, radar can now be provided by solar panels (unlike the earlier RORSATs, when photovoltaics were far less efficient than today). Within the main belt asteroids, unless a high degree of reliability is needed for crewed surface installations (the focus for Kilopower at the moment), solar power is, again, usually sufficient. Out of Earth orbit, other systems have been devised to gather similar data for far less power, but these systems have limitations as far as data gathering is concerned. Given a powerful, sufficiently power-dense , power source, we may see on-orbit ground penetrating radar used to, say, map the rocky core and liquid layer of Enceladus, or probe the depths of the ice giants; but, until Kilopower becomes operational, there’s not a power supply that can support that, and radar tends to be bulky – and, even worse, heavy.
In the outer solar system, the situation is quite different; but, even then, power requirements tend to be low. Kilopower’s 1 kWe system is more than was available to Cassini/Galileo, New Horizons, or any other outer planet flagship mission. All of these (except Juno, which used experimental, and very expensive, solar panels, with a collective surface area of a tennis court) were powered by radioisotope thermoelectric generators (another upcoming topic, if not for a blog post, then for the website), which are fundamentally limited by the decay of the isotope used. These missions usually run on just a few hundred watts of power, if not less, and bring back incredible science returns.
The final reason is also the most popular one to think about: propulsion. Briefly, electric propulsion uses electric potential rather than chemical potential to accelerate a spacecraft. While it is often seen as a new thing, the concept has been around since the earliest days of spaceflight, although the details of how it’s used have changed greatly over time. The thrust and efficiency available in these systems is often dependent on having more electrical power, though, so fission power systems offer a uniquely attractive set of characteristics for this type of propulsion.
For more information on electric propulsion, including its history and different varieties of thrusters, check out our electric propulsion page!