France’s nuclear research agency, Commissariat à l’énergie atomique (CEA), in September confirmed it abandoned plans to build a prototype Advanced Sodium Technological Reactor for Industrial Demonstration (ASTRID), a sodium-cooled fast reactor (SFR), in the “short or medium term.” CEA cited the “current energy market situation,” noting that currently, “industrial development of fourth-generation reactors is not planned before the second half of this century.”
Before the French government confirmed it would no longer fund the project, ASTRID was one of a handful of much-watched fast-reactor projects under development globally, because it promised to validate breakthroughs in SFR technology. As industry experts have noted, interest in fast neutron reactor systems has ramped up recently because they have the potential to extract 60 times more energy from uranium compared to existing thermal reactors, and they could contribute to a significant reduction in the burden of radioactive waste. And while the variety of fast-reactor research underway globally is vast and diverse, experts generally agree that SFRs with a closed fuel cycle are the most mature of all fast-reactor concepts.
To date, only 20 SFR prototypes or demonstrations have been built throughout the world, but these have imparted more than 400 reactor-years of operation in experience and technical knowledge. Examples include BN-800 at Beloyarsk 4 in Russia, China’s CEFR, and the world’s first power-generating nuclear reactors, EBR I and II in the U.S., which came online in 1951. A number of private companies, such as GE Hitachi (for its PRISM concept) and TerraPower, have also shown interest in SFR technology.
2. Construction of the Superphénix, a sodium-cooled fast-reactor prototype on the Rhône river at Creys-Malville in France, close to the border with Switzerland, began in 1974, but owing to a series of cost overruns and public protests—including a rocket-propelled grenade attack in 1982—it wasn’t grid connected until 1986. Operations were plagued with technical faults, including corrosion and leaks of the liquid sodium cooling system, and structural damage to the turbine hall. It only produced a total of 7,494 GWh before it was decommissioned in early 1997—20 years ahead of schedule—owing to political pushback. The French Accounting Office in 1996 estimated the reactor cost the nation $10 billion. Courtesy: Yann Forget/Wikimedia Commons/CC-BY-SA
In this context, France’s SFR experiences have been especially important. The government research agency has operated three SFRs: the 40-MWth Rapsodie, which ran between 1967 and 1983; the 250-MWe Phénix, which operated between 1972 and 2009; and the 1,200-MWe Superphénix—the largest SFR built to date—which operated between 1985 and 1997, when it was abruptly ended, owing partly to politics (Figure 2). According to the CEA, these projects demonstrated the “excellent use of the uranium resource as well as the capability of these reactors to recycle the plutonium without any limitation in the number of recycling operations.” Phénix and Superphénix also showed how several material selections could be unsuitable, and shed light on SFR safety functions, fuel handling, sodium leaks, operations, maintenance, and dismantling.
CEA had envisioned ASTRID as “completely different from the fast reactors of previous generations.” Armed with a $723 million budget, work on the 600-MWe ASTRID concept began in 2010, and it moved out of the conceptual design phase in late 2015 and into the basic design phase. As a June 2016 document shows, CEA, which was the owner of the technology demonstrator project, contemplated initially using austenitic steel sub-assemblies, and later, steel strengthened by oxide dispersion—and it had already installed a “complete facility” to manufacture these materials. It would have also integrated technological breakthroughs “so its level of safety exceeds that of all the other fast reactors that would have [been] built to date, including Phénix and Superphénix,” CEA said.
One solution was to avoid a “sodium void,” which resulted in higher core reactivity in the Superphénix, by developing an enhanced safety core for ASTRID. To limit accidental contact between the liquid metal and water, CEA also planned to replace the steam generator with a heat exchanger and turbine operating with pressurized nitrogen. Finally, CEA sought to limit fire hazards arising from sodium leakage with leak detection, bunkering, and “inerting” the most-exposed zones.
These solutions were still under a major research and development effort, however. In November 2018, CEA announced it was considering slashing ASTRID’s capacity to between 100 MW and 200 MW. News that CEA was abandoning the project was first reported by French newspaper Le Monde on Aug. 30. That report, confirmed by CEA, said the project had already cost $814 million at the end of 2017—which means it was already over budget.
The project also lacked key political support amid France’s vacillation about the future of its nuclear fleet. French President Emmanuel Macron in November 2018 announced plans to reduce nuclear power’s share to 50% from the current 70%. However, as some experts noted, uncertainty had clouded ASTRID’s future as far back as November 2016, when Japan pulled the plug on its own $8.5-billion Monju prototype fast reactor due to heavy costs and announced it would instead remain committed to France’s fast-reactor efforts. In December 2018, the French government informed Tokyo that it planned to freeze ASTRID, citing an apparent uranium glut, as Japanese media reported. News reports noted Japan has so far spent $177 million on ASTRID, and the project’s cancellation effectively stalls Japan’s own nuclear fuel-cycle policy.
ASTRID’s demise was also lamented by engineers and technicians who worked on the sodium-cooled fast reactor and its predecessors. As Jean-Marie Berniolles, an engineer who teaches at the CNRS-Aix Marseille joint research center in France, wrote in a September opinion piece for the European Scientist, with ASTRID now abandoned, France’s industry will lose essential knowledge it needs to develop the mechanical and chemical properties needed for fast-reactor cladding, vessels, and pipes, which rely on much higher standards. However, the project, he wrote, may have been doomed from the beginning because its initial design was “too referential” to the Superphénix—which suffered its own “brutal ending,” owing to political sabotage. Now, even ASTRID’s “ghost” has fallen victim to France’s political environmental movement, he wrote. “So, even the ghost is gone.”
—Sonal Patel is a POWER senior associate editor.