Coming down to Earth3 October 2002
The 200kWe uranium nitride fuelled lithium cooled fast reactor "RAPID-L", originally conceived of as a power source for Lunar colonies, can be operated unmanned without refuelling for up to ten years. By Mitsuru Kambe, Hirokazu Tsunoda, Kaichiro Mishima and Takamichi Iwamura
All fast reactor operations including startup, shutdown and power control are achieved using B4C control rods. Such control systems involve a number of issues: reliability of actuation and drive mechanisms; scram capability under seismic conditions; and unprotected transient overpower (UTOP) mitigation potential. Despite several attempts to enhance inherent safety by self-actuated shutdown systems, only inherent secondary shutdown systems (ISSS) and gas expansion modules (GEM) do not require absorber rods.
The ISSS concept consists of small tantalum alloy balls that are hydraulically raised by the reactor coolant and drop into the core zone during an unprotected loss of flow (ULOF) incident. The GEMs are empty assemblies, sealed at their top end and connected to the inlet coolant plenum at the bottom end. In case of ULOF, the neutron scattering sodium is removed and a neutron leakage path out of the core is provided. Both the ISSS and GEMs are, by their own nature, effective for ULOF but not for transients like UTOP. In addition, long-life reliability is as yet unconfirmed.
Another issue to be considered in future fast reactors is the simplicity of operation. So far, skilful operators are essential for reactor startup in conventional plants. Fully automated reactor startup has never been attempted except for spacecraft nuclear reactors.
A significant advantage of the RAPID-A (Refueling by All Pins Integrated Design) 60MWe fast reactor design - which does not require the use of control rods - is the introduction of the innovative reactivity control systems: lithium expansion module (LEM), lithium injection module (LIM) and lithium release module (LRM). LEM is the most promising candidate for improving inherent reactivity feedback. LEMs could realise burnup compensation and partial load operation. LIMs assure sufficient negative reactivity feedback in unprotected transients. LRMs enable an automated reactor startup by detecting the hot standby temperature of the primary coolant. All these systems use 6Li as liquid poison and are actuated by highly reliable physical properties (volume expansion of 6Li for LEM, and freeze seal melting for LIM and LRM). However the operator-free reactor concept has not been completely confirmed yet in this reactor design.
The RAPID-L operator-free fast reactor, designed for Lunar-based activities, enables quick and simplified refuelling two months after reactor shutdown. Emphasis has been placed on demonstrating inherent safety features by using LEM, LIM and LRM innovative reactivity control systems. Potential uses for RAPID-L are in power plants in urban areas of industrialised nations to relieve the peak load, and for developing countries where remote regions cannot be conveniently connected to the main grid and where it is economical to provide local generation capacity. In addition RAPID-L can be used for seawater desalination.
overall plant design
The conceptual design of RAPID-L consists of a 5000kWt-200kWe uranium nitride fuelled, lithium cooled, fast spectrum reactor with lithium inlet and outlet temperature of 1030ºC and 1100ºC, respectively. The reactor structure and the fuel assembly are shown in Figure 1. The reactor is essentially a loop configuration with a reactor vessel of 2m diameter and 6.5m depth. The difference from a conventional pool type reactor is the integrated fuel assembly. The core consists of approximately 2700 fuel elements, which, with a core support grid and several spacer grids, are combined into a fuel assembly. In this particular design, the reactor has neither a diagrid nor a core support structure because these are integrated in the fuel assembly.
The integrity of the core support structure is of utmost importance. Thus the most important in-service inspection (ISI) requirement is validation of the core support structure, as this is entirely immersed in the coolant and therefore presents a difficult, and currently unresolved, problem. The RAPID concept makes it easier by adopting a combination of replaceable diagrid/core support structure and a permanent simple reactor vessel. This system makes ISI easier than it would be with the more complex diagrid and core support structure. ISI of the reactor vessel can be conducted in each refuelling. Unlike conventional reactors, which require a design lifetime of around 40 years, the design lifetime of the replaceable core support grid as well as the fuel assembly can be reduced to 10 years, for those located in the relatively high flux regions (1022n/cm2 in 10 years).
On the other hand, the reactor vessel can be designed as a permanent structure exposed to decreased neutron flux. To keep the final fluence of the reactor vessel within the design limit (1021n/cm2 in 20 years), neutron shielding materials must be effectively provided in the lower section of the fuel assembly.
Electromagnetic pumps and the fuel assembly are connected by the connecting tubes. The reactor subsystem is characterised by the RAPID refuelling concept to eliminate conventional fuel handling systems. A reactor block mass reduction of 60% over comparable liquid metal cooled fast reactor systems can be estimated. Its compact design makes it possible to assemble a whole structure in the workshop prior to the launch.
The overall power system is illustrated in Figure 2. The reactor is coupled to four heat exchangers placed around the reactor. Each segment has a pumped lithium heat rejection loop connected to mercury heat pipe radiators. The eight radiator panels are arranged in a vertical configuration extended radially from the power conversion segments. The reactor is located in an excavated cylindrical hole that provides shielding of gamma and neutron radiation.
Reactor core design
The reactor core is a homogeneous design with two regions. The active core region is 600mm in diameter and 600mm high with a central channel 120mm in diameter. The fuel chosen for the inner and outer core consists of 40% and 50% enriched uranium nitride, respectively. The uranium enrichments were adjusted to minimise both the burnup reactivity swing and the radial peaking factor. The core performance parameters are shown in Table 2.
RAPID refuelling concept
Refuelling is conducted every 10 years. The refuelling procedure is illustrated in Figure 3. The RAPID concept enables quick and simplified refuelling after two weeks of reactor shutdown by which time decay heat of the core is 10kW. During refuelling, a lithium filled fuel assembly is removed from the reactor and loaded into a lithium filled on-site storage cask (OSSC) waiting besides the reactor. After receiving the spent fuel, the OSSC is equipped with a heat pipe radiator for decay heat removal. It is stored in an excavated cylindrical hole to minimise the dose rate of astronauts involved. After dissipation of decay heat, the lithium will solidify in the OSSC one year after the refuelling. The spent fuel together with the OSSC could be disposed of into deep space.
Lithium expansion module
The lithium expansion module (LEM) is the most promising candidate for improving inherent reactivity feedback. The LEM (see Figure 4) is composed of an envelope of refractory metal in which liquid poison of 95% enriched 6Li is enclosed. Lithium-6 is suspended in the upper part of the envelope by surface tension exerted on the gas-liquid interface. The LEM is actuated by the volume expansion of 6Li itself. If the core exit temperature increases, the gas-liquid interface goes down and negative reactivity insertion can be achieved. The principle of LEM has been demonstrated by neutron radiography conducted at the Japan Atomic Energy Research Institute.
RAPID-L is equipped with four quick LEMs and 24 slow LEMs. A quick LEM is characterised by a quick response. It only provides a negative reactivity insertion. Three (out of four) quick LEMs ensure -50¢. Accordingly it is effective to mitigate the anticipated transient without scram. The gas-liquid interface in the nominal operation is placed at the active core top. In case the core outlet temperature decreases, the gas-liquid interface goes up and no positive reactivity insertion is expected.
Slow LEM can provide both negative and positive reactivity insertion with moderate thermal response. Reactivity varies between -2.7$ and +3.0$ for the 24 slow LEMs. Slow LEMs provide the role of automated burnup compensation. In addition, slow LEMs also realise partial load operation in accordance with the primary flow rate. The gas-liquid interface in nominal operation is placed in the active core region. In case the core outlet temperature decreases, the gas-liquid interface goes up, and positive reactivity is added, and vice-versa. To avoid quick positive reactivity addition, slow LEMs have a reservoir of double envelopes for vacuum insulation. Therefore, slow LEMs are only affected by moderate thermal transients resulting from burnup reactivity swing and primary flow rate control. The design parameters of LEMs are described in Table 3.
Transient mitigation potential
Another innovative feature of LEM is the transient mitigation potential. Because of the sensitive response of the quick LEM reservoir, sufficient negative reactivity feedback can be quickly inserted upon transients to keep the reactor running safely, even without reactor shutdown. This is preferable to avoid unnecessary reactor shutdown and to improve plant availability.
Redundancy of the LEMs
A significant characteristic of the LEM reactivity control system is its redundancy. Since RAPID-L has no control rods, UTOP transients due to fault handling of the control rods do not arise. However the boundary failures of one of the LEM envelopes should be anticipated. This failure causes positive or negative reactivity insertion, depending on the LEM situation (location of gas-liquid interface just before the failure). A maximum reactivity addition is anticipated in case of a 40% primary flowrate in the BOL core; each slow LEM is in a state of full insertion of the 6Li liquid poison. If the envelope of a slow LEM fails, liquid poison in the envelope would be quickly replaced by 7Li coolant or void. The latter yields more severe consequences: positive reactivity addition of 22.7¢ per single slow LEM. This reactivity addition is more severe than that of UTOP in conventional reactors (typically 3¢/sec), and is quickly compensated for by other quick LEMs.
LEM failures in which LEMs provide positive reactivity (for example, burnup compensation in the EOL core) will cause negative reactivity insertion. This results in only a slight decrease in the reactor power.
Lithium injection module
The lithium injection module (LIM) is another innovative device installed in RAPID-L to ensure inherent safety (see Figure 5). LIM is also composed of an envelope in which 95% enriched 6Li is enclosed. In case the core outlet temperature exceeds the melting point of the freeze seal, 6Li is injected by pneumatic mechanism from upper to lower region to achieve negative reactivity insertion. In this way the reactor is automatically brought into a permanently subcritical state and temperatures are kept well below the boiling point of lithium (1330ºC). Time required for reactivity insertion of LIM is 0.24s, which is much shorter than that of free drop of the conventional scram rods (as much as 2s) on the Earth.
To provide the shutdown margin of -0.5$, the total reactivity worth of -3.7$ is required for LIMs to counterbalance the slow LEMs' feedback (+3.0$) and the power defect reactivity of the core (+0.2$). Sixteen LEMs of 20mm-diameter envelope are sufficient to achieve this. Similarly to LEMs, LIMs ensure sufficient negative reactivity feedback in unprotected transients. The role of LIM is to provide a variety and redundancy of inherent safety during unprotected transients. Either LEMs or LIMs can meet such transients independently. The difference between LEM and LIM is that the former can achieve both negative and positive reactivity feedbacks irreversibly, and the latter only negative feedback permanently.
Freeze seal design is key to ensuring the accurate injection temperature over the design lifetime. The freeze seal segment consists of CuNi alloy (trade name: L-30) to assure the injection temperature of 1240ºC.
Lithium release module
Fully automated reactor startup can be achieved by lithium release module (LRM, see Figure 6). LRM is similar to LIM; however, 6Li is reserved in the active core level prior to reactor startup. The LRM should be placed in the active core region where the local coolant void worth is positive, as is the case with LEM and LIM.
RAPID-L is equipped with an LRM bundle in which 16 LRMs are assembled. The reactivity worth of an LRM bundle is +3.7$ where 95% enriched 6Li is enclosed in each 20mm diameter envelope. An automated startup can be achieved by gradually increasing the primary coolant temperature by the primary pump circulation. The freeze seal of LRMs melt at the hot standby temperature (approximately 780ºC), and 6Li is released from lower level (active core level) to upper level to achieve positive reactivity addition. The freeze seal support material of 28:72 Ag-Cu is used to ensure the startup at 780ºC.
RAPID-L on the Lunar base can be operated for 10 years without shutdown. From the perspective of plant availability, periodic inspection of the reactor plant should be planned during a refuelling shutdown. Such an interval of periodic inspections would be reasonable in future reactors, especially in space reactors. Prior to the refuelling, the LRM bundle should be released in the lower part of the core so that liquid poison can locate in the active core region again. In this case, the LRM bundle acts as a poison rod. Once released, it is clumped at the bottom and is impossible to pull out again. This design conforms to USA space reactor safety criteria. The reactor startup is only possible by installing a new fuel assembly.
Reliability and redundancy
Innovative reactivity control systems LEM, LIM and LRM have the following advantages in terms of reliability and redundancy.
• LEM, LIM and LRM have no moving parts and depend only on highly reliable physical properties.
• LEM and LIM are functionally and mechanically diverse to each other. Either of them can ensure negative reactivity feedback on unprotected transients independently.
• LEM, LIM and LRM are effectively independent of the magnitude and direction of gravity.
• LEM, LIM and LRM have significant redundancy because there are 24 slow LEMs, 4 quick LEMs, 16 LIMs and 16 LRM in RAPID-L.
• Reactivity addition resulting from failure of LEM, LIM and LRM can be quickly compensated by serviceable LEMs. Therefore such failures would not affect plant availability.
• LEM, LIM and LRM can be replaced together with the fuel assembly during a refuelling shutdown without any effect on the plant availability.
TablesTable 1: System performance parameters Table 2: Core performance parameters Table 3: Design parameters of LEMs Table 4: RAPID-L reactivity control system