A feasibility study was performed on a natural circulation cooled small nuclear reactor. The study included neutronics calculations to investigate the minimum dimensions required to obtain a critical system during the complete fuel cycle length, a determination of coolant temperature and core temperature reactivity coefficients and an investigation of the thermal hydraulics to asses the possibilities for natural circulation cooling.

The Nuclear Battery is a very small inherently safe, self regulating nuclear reactor (20 MWth) for propulsion, electricity generation or process heat applications. It can be operated for fuel cycle periods of 10 years or longer without refuelling, minimal operation cost, and is proliferation resistant. Because the reactor should be removable after shutdown, the primary circuit is incorporated into a transportable sealed container. The allowable height width and length of the primary system should fit within 3.5 x 3.5 x 20 m.

The Nuclear Battery is graphite moderated and uses TRISO coated UO2 particle fuel with enrichment of 20%. TRISO particles retain the fission products up to a fuel temperature of 1600 oC for limited periods of time [1]. A burnup of 10% FIMA should be reached. To reduce neutron leakage, the outer dimensions of the reactor core (without reflector) where chosen to minimise the geometrical buckling. A prismatic core geometry was selected because of the freedom in volume fractions of fuel, moderator and coolant. As a primary coolant liquid salt is used to allow operation at ambient pressure. The primary coolant candidates for this design are the liquid fluoride salts 7Li-Na-Zr, 7Li-Na-K, Na-Be, Na-Zr and 7Li-Be. Also liquid tin is investigated as a primary coolant based on the proposition in [2].

The full article will present the detailed results of a parameter study that was performed to assess the feasibility of the Nuclear Battery, within the boundary conditions of 1) a fuel cycle of 10 years, 2) the use of 20% enriched fuel, 3) a burnup of 10% FIMA and 4) a core diameter smaller than 3.5 m. Besides these conditions other important parameters are the core volume, reflector thickness, and the coolant. The effect of these parameters on the feasibility have been assessed by burnup calculations during a desired fuel cycle length and a keff calculation at the end of the fuel cycle (EOC). When the keff at EOC is smaller than one, the design is not feasible.

In FIG. 1. the reflector thickness and core volume combinations that give keff = 1 at EOC is shown for all salts and tin. The coolant volume fraction is 10% for the liquid salt and 3.5% for liquid tin. Also shown in the figure is the total core diameter (i.e. core plus outer reflector diameter) in a contour plot. On the lower left side of the curves shown the keff at the end of the 10 year fuel cycle is less then one (not feasible) while on the upper right side of the curves the keff is larger than one (feasible). It can be seen that for each case a wide range of core volumes and reflector thicknesses can be found within the 3.5 m diameter constraint. The 7Li-Be fluoride salt provides largest area of feasible combinations (between the 3.5 diameter constraint and the keff = 1 curve). Although design freedom is less for Na-Zr fluoride and tin, both can be promising coolant candidates due to the absence of the toxic Be and expensive isotopic separation of 7Li.

FIG. 1. The reflector thickness and core volume combinations that give keff = 1 at EOC for the salt coolants. In all cases the cycle length was 10 y with a nominal power of 20 MWth. The coolant volume fraction (cvf) was 10% for the liquid salts and 3.5% for liquid tin.

For safe operation of the reactor it is necessary that the use of the coolant does not lead to positive voiding or positive temperature reactivity effects. Unfortunately the void coefficients are positive for Na-Zr fluoride and tin. For Na-Zr the Doppler temperature effect of the fuel compensates the reactivity increase due to complete voiding by a relatively small core temperature increase. For tin the temperature increase to compensate complete voiding reactivity is too large for safe operation, therefore, in that case, measures must be taken to prevent complete voiding at all times.

The possibilities for natural convection were investigated by performing calculations in a 1D natural circulation model and heat transfer calculations in the core. It was shown that tin is well suited for natural circulation cooling. For the liquid salt coolants, it is difficult to obtain a turbulent flow in a natural circulation driven cooling system. As a consequence the heat transfer coefficient along the walls of the coolant channels will decrease leading to higher fuel temperatures. A "longer" core may have benefits for the natural circulation possibilities at the expense of increased neutron leakage.

There is great potential for a liquid cooled Nuclear Battery. Feasible neutronic core designs can be made. Thermal hydraulics calculations show that cooling by natural circulation is possible. Future work will focus on the thermal hydraulics, burnup and shielding calculations, and further analyses of the reactor physics including passive reactivity control and several accident scenarios such as loss of coolant in combination with a passive decay heat removal assessment.

References

For more information, please contact j.l.kloosterman@tudelft.nl.