Guest post by Russian commenter Syndroma who was trained in IT and now works in his family business. The BN-800 was commissioned this week.
The Russian fast neutron reactor program was started in the early days of the nuclear age. Scientists back then realised that thermal neutron reactors are no more than a stopgap solution. Yes, they’re simple, but ugly from the scientific point of view. They’re inefficient, burning no more than 1% of natural uranium, and they leave a lot of highly radioactive long-term waste, which can’t be dealt with meaningfully without fast reactors. Scientists, being a bit idealistic, couldn’t believe someone would want to deal with thermal reactors in the long term.
The first attempt at fast neutron reactor with non-zero power was called BR-2 (‘B’ for fast, ‘R’ for reactor). Its power was just 100 kW, it was fueled by metallic plutonium and cooled by mercury. Just a few months after the start of operations in 1955, mercury revealed itself as a terrible coolant for a reactor. It was highly corrosive, damaging fuel cladding and pipes. Mercury leaks were a common occurence, dealt with casually. After a year of operations the experiment was declared a failure. The rector was dismantled, the mercury was decontaminated and sold for civilian use. The same building and infrastructure were used for BR-5 reactor, which was commisioned in 1959. It was a different project with power output of 5 MW, later upgraded to 10 MW, fueled by plutonium oxide and cooled by sodium. 
Figure 1: The evolution of sodium-cooled fast neutron reactors in Russia. 
BR-5/10 was operational until 2002 and proved to be a great success, providing the first operational experience and a lot of scientific data on behaviour of fuels and materials in fast neutron field. Still, it was limited in its geometry and power, so a bigger sibling BOR-60 (‘O’ for experimental) was commissioned in 1969. BOR-60 is still working to this day, being used for materials research by Russian and international teams. It is expected to be replaced around 2020 by MBIR (multipurpose fast research reactor), which is under construction right now.
The success of research reactors paved the way for the first commercial fast neutron project – BN-350 (‘B’ for fast, ‘N’ for sodium). It was built in Kazakhstan near a uranium mine, its power output was 350 MW shared between electricity generation, district heating and desalination. The first years of BN-350 operations were extremely troublesome.  Nobody could predict the issues caused by scaling up. The core of BN-350 contained 300 times more sodium than BR-5. Sodium was leaking from unexpected places. Sodium was freezing in unexpected places. But the worst problems were related to steam generators. Sodium is lighter than water, so it doesn’t leak into water, water leaks into sodium and reacts violently. The steam generators were prone to small leaks which were hard to detect promptly, contaminating entire loop with products of the reaction. The problem was so severe that the reactor was shut down and all heat exchanging pipes in the steam generators completely replaced.
After so much trouble with BN-350, BN-600 project had to prove that large-scale sodium reactors could be operated safely and reliably. It was accepted that leaks are inevitable, the focus was placed on detection and mitigation. Steam generators were partitioned into small sections which could be cut off and repaired. Early leak detection systems and fire suppression methods were implemented. There were at least 27 leaks with up to a tonne of sodium spilling into the air, none of them causing any significant downtime.  But there was a price to pay – BN-600 required a lot more piping than any other commercial reactor, severely damaging its economics. Still, 36 years after commissioning it remains the most successful fast neutron reactor in the world, and its operational lifetime is likely to be extended.
BN-800 reactors were supposed to be the first serially built fast neutron reactors. Their construction started in late 80s, 125 km away from the place where the modern BN-800 was built. But first Chernobyl, then the collapse of the USSR interfered with the plans. Construction was halted, leaving giant holes in the ground. By the 2000s, when the new Russia was ready to return to construction, BN-800 design was considered outdated. There were new, exciting possibilities. A significant number of scientists thought that sodium is a waste of time and the future belongs to lead. They proposed a revolutionary, lead-cooled fast neutron reactor BREST-300. They’re still committed to building it, sometime next decade. Other scientists were reluctant to build just a bigger version of BN-600, they wanted to build a revolutionary BN-1200 with completely different layout. But the cooler heads noted, that Russia has not built a fast neutron reactor in a generation, and given past experiences anything revolutionary tends to fail spectacularly. They proposed to build an updated conservative design just to prove that Russia still has the capability and talent to do so.
Figure 2: BN-800 scheme
Nobody saw it coming, but the US played a major positive role in the fate of BN-800. Russia and the US reached an agreement to destroy excess amounts of plutonium to make the world a safer place. The US has choosen to burn it in conventional reactors, Russia has choosen to use it as a fuel for fast reactors. BN-600 was not suitable for the task because by 2018, when the destruction of plutonium had to be started, it’d be nearing its end of design lifetime. And the only other project which had a chance to be implemented before 2018 was BN-800. The downside of the agreement for Russia was restriction to use breeding blanket around the core and restriction on reprocessing of the spent fuel. BN-800 was forbidden from operating as a Pu breeder, only as a Pu burner. The restrictions were irritating, and that was one of the reasons why Russia recently withdrew from the agreement.
The BN-800 project was dusted off, updated to modern standards and materials, its nominal electric power output was increased to 880 MW. Its construction started in 2006.
Figure 3: BN-800 reactor under construction
A 25 years long break took a heavy toll on the nuclear industry. Some things had to be re-learned. Some things had gone for good. For example, the producer of pure sodium for BN-600 has long disassembled the equipment due to lack of customers. Sodium is bought once for the entire lifetime of a reactor. The producer agreed to restore the production line only if guaranteed orders, the company wasn’t interested in a one-time job. Luckily, the French, who had a respectable sodium reactor program of themselves, kept their production. French sodium was transported 4 thousand km, reheated and pumped into the BN-800.
BN-800 went critical in June 2014, was connected to the grid in December 2015 and was commissioned this week.
Although the design of BN-800 was choosen to be very similiar to BN-600, it was decided early on that the fuel composition should be radically different. BN-600 runs on highly enriched uranium, BN-800 should have been run on plutonium MOX fuel. There are two kinds of plutonium: weapons-grade, which is relatively easy to handle, and reactor-grade, which is more radioactive. Russia committed to destroy excessive amounts of weapons-grade Pu, but also it has a significant stockpile of reactor-grade Pu from reprocessing of spent nuclear fuel. If the fuel assembly line was designed to handle only weapons-grade Pu, than it’d be useless for a closed nuclear fuel cycle. So, it was decided that the fuel plant should be built with all the necessary protection (and it was – inside a mountain). But then the planners stumbled upon a minor issue with a long-reaching consequence.
There were several institutions in Russia which worked on MOX fuel fabrication. And over the course of decades two competing technologies were developed. One is classical: mix plutonium and uranium powders, make pellets out of it and pack the pellets inside a tube. The other one is innovative: mixed powder is vibro-packed directly inside a tube, skipping the pellets stage. Vibro-packed MOX process claimed some important technological advantages related to utilising of defective tubes and reprocessing of spent fuel. The advantages were attractive enough for the technology to be chosen as the basic for BN-800. It turned out to be a polarizing issue, with holy wars breaking out between the opposing camps at meetings and conferences.
But reality reigns supreme. In time it had become obvious that vibro-packed MOX technology has issues at scaling up. It’d be ready a decade after the BN-800 launch. It was unacceptable. Worse, much time had been lost, and the pellet technology would not be ready either. There was no other choice than to load BN-800 mostly with the same old highly enriched uranium. ‘Mostly’ because small-scale MOX fuel lines of both kinds were capable of producing around 20% of the required fuel. And the resulting reactor core of HEU, pellet MOX and vibro-packed MOX was called “hybrid core”.
Obviously Pu and U are different elements and are expected to behave differently. Placed in the same core, they produce slightly different amounts of energy. But the coolant flows at the same speed in the reactor. To prevent thermal irregularities, small plugs were placed at the base of some assemblies to decrease the flow through them. Time was short, the modification looked simple, no proper testing was conducted. And when the fuel was loaded into the BN-800, and its giant coolant pumps started pumping liquid sodium through the core, the plugs started vibrating and unscrewing themselves. To make matters worse, when an assembly is extracted from the reactor it has to be washed out from the sodium leftovers with an interesting liquid, which makes it impossible to get it inside the reactor again. The only way to safely work with an unwashed assembly is to place it inside an inert gas atmosphere. And the closest workshop with such an atmosphere happened to be at the BN-600. Years of operational experience came to the rescue once again. 
The mishap cost the program about a year of delay. The assemblies were modified, reloaded, and the reactor went online with the hybrid core in December 2015. About the same time the pellet MOX fuel plant went online too. It is expected that BN-800 will be fully loaded with MOX fuel by 2019. Vibro-packed MOX fuel is not abandoned either, it will be used in research reactors.
Figure 4: BN-800 fuel load. 1 – low Pu content, 2 – medium Pu content, 3 – high Pu content, 4 – breeding blanket, 5 – steel shield, 6 – boron shield, 11 – spent fuel. 
Is everyone happy? Nope. The designers of BN reactors wanted neither pellet MOX nor vibro-packed MOX. The key characteristics of a fast neutron reactor directly depend on the density of the fuel. The denser it is, the more neutrons it captures, the higher reactor’s breeding ratio becomes. Research of a dense fuel is a major program of the Russian nuclear industry. So-called SNUP fuel (mixed nitride uranium-plutonium) is in transition from lab experiments to limited loads at BN-600. It’s perfomance and safety is thoroughly evaluated, and it is expected that the proposed BN-1200 will be loaded with the SNUP, eventually.
Although BN reactors have many advantages, they’re more costly than VVER (PWR) reactors. They require much more metal for construction: a bigger reactor vessel (but not pressurized) with intermediate pumps and heat exchangers, and a complicated steam generators. The main goal of BN-1200 project is to lower the cost of the unit to be competitive with conventional reactors. To achieve that goal, all systems need to be reviewed and optimized. But one system is of particular interest: the steam generators. BN-600 SGs consist of 72 sections, BN-800 SGs consist of 60 sections. BN-1200 proposes to return to integrated SGs with only 8 sections. It’s a risk, but things changed a lot since BN-350 times. Modern materials, simulation algorithms and testing capabilities could allow for much more reliable steam generators.
Figure 5: BN-800 and BN-1200 steam generators (one loop). BN-800 has 3 loops, BN-1200 has 4 loops. 
Another proposal is to change the layout of the reactor building, reducing pipeline length up to 1.8 times. If BN-800 sodium pipes look like a squid monster, BN-1200 proposes a more radially symmetrical approach.
Figure 6: BN-800 and BN-1200 sodium pipes.
The construction of BN-1200 is expected to start after 2019. The road to sustainable nuclear energy is long, but it’s definitely worth the effort. One just needs to see things in perspective.
Figure 7: BN-800
References (all in Russian)