The MUTBR design has been extensively simulated with MCNP. In 2020, researchers from Oak Ridge National Lab ran MCNP6 burnup analyses of a preliminary MUTBR configuration we provided. They concluded that from a neutronics perspective the configuration provided worked well enough with LEU fuel but was not notably advantageous compared to traditional reactors. The configuration provided did not have a useful fuel life with fuel that was only UNF. However, the MCNP6 burnup analysis had to assume no removal of fission products except for a small adjustment to the results to adjust for the effect of removal of the inert gas fission products xenon and krypton. The ORNL analyses also provided insights that greatly enhanced our ability to use MCNP and transition from MCNP5 to MCNP6 with burnup analysis. Based on this and our own improved simulation results we have made several improvements to the design configuration and to the fuel choices. These include the use of fuel which is a mixture of LEU and UNF as a possibility in the SMR and grid scale reactors as shown in the simulation results below.
The tables below show some results of MNCP simulations of some different MUTBR configurations and initial fuels. Neutron physics considerations limit the size range of the fuel tubes so higher power reactors have more fuel tubes. The thermal power per fuel tube is around 20 MW. The reactor configurations simulated are a 1 tube micro reactor, a 19 tube Small Modular Reactor (SMR), and a large 169 tube grid scale reactor. The initial fuel can be either very Low Enriched Uranium (LEU) or a mixture of standard (4.95% U-235) LEU and Used Nuclear Fuel (UNF). The first table is simulations run for 60 years of fuel life, in the second table the simulations are continued to a 100 year fuel life. The MUTBR reactors are controlled with control cavities which are partly full of steam with the rest heavy-water moderator. If the fuel is very reactive the control cavities are mostly steam and little liquid moderator. If the fuel is less reactive the steam bubble is smaller. The steam bubble size (shown in thousandths of the cavity size) shows the fuel reactivity. The first two columns in each table are the starting fuel reactivity and the ending fuel reactivity. Columns 3 and 4 are the starting and ending fuel fissile content, column 5 is the reactor size, and six is the initial fuel used.
s-react . e-react . s-fissile . e-fissile .... size .... fuel (60 year burn = 60 g/kg)
710 ....... 425 ....... 1.90% ..... 1.84% ........ grid ..... 100% LEU (1.90% U235), 0% UNF
549 ....... 383 ....... 2.13% ..... 1.93% ........ grid ....... 20% LEU (4.95% U235), 80% UNF
742 ....... 475 ....... 2.90% ..... 2.36% ........ SMR ... 100% LEU (2.90% U235), 0% UNF
620 ....... 431 ....... 3.01% ..... 2.39% ........ SMR ..... 45% LEU (4.95% U235), 55% UNF
796 ....... 513 ....... 8.00% ..... 5.78% ........ Micro .. 100% LEU (8.00% U235), 0% UNF
585 ......... 18 ....... 6.50% ..... 4.59% ........ Micro .. 100% LEU (6.50% U235), 0% UNF
s-react . e-react . s-fissile . e-fissile .... size .... fuel (100 year burn = 100 g/kg)
710 ....... 162 ....... 1.90% ..... 1.83% ........ grid ..... 100% LEU (1.90% U235), 0% UNF
549 ....... 144 ....... 2.13% ..... 1.91% ........ grid ....... 20% LEU (4.95% U235), 80% UNF
742 ......... 96 ....... 2.90% ..... 2.12% ........ SMR ... 100% LEU (2.90% U235), 0% UNF
620 ......... 39 ....... 3.01% ..... 2.16% ........ SMR ..... 45% LEU (4.95% U235), 50% UNF
796 ......... 59 ....... 8.00% ..... 4.69% ........ Micro .. 100% LEU (8.00% U235), 0% UNF
In the tables it can be seen that the fuel reactivity and the fissile content decrease with time in all cases. The decrease in the fuel reactivity is much greater than the decrease in fissile content because of the gradual buildup of neutron absorbing fission products with time. The decrease of both is less and the fuel initial fissile content is less in the larger reactors because fewer neutrons escape from the larger core.
To examine the wide range of the control method, simulations were run with fuel extremes or control cavity steam bubble size extremes. With the SMR using LEU (4.09% U-235) as its starting fuel, keff = 1.00002 +- .00055 when the steam bubble in the control cavity is 1% of the cavity. Using LEU (1.64% U-235) as its starting fuel, keff = 0.99984 +- .00047 when the steam bubble in the control cavity is 99% of the cavity, showing a wide rang in the possible initial fuel enrichment. With the SMR using LEU (2.90% U-235) as its starting fuel, when the steam bubble in the control cavity is 1% of the cavity, keff = 1.13119 +- .00062 and when the control cavity is 99% steam, then keff is .92030 +- .00057 showing that the control method has a control range of around 21,089 +- 119 pcm.
MCNP is very good at determining the fission multiplication factor (keff) and the fuel evolution of solid fuel reactors. Analysis of the fuel evolution in the MUTBR is complicated by several factors. The fuel is molten and continuously circulating so it is always well mixed. Some of the fuel is outside of the reactor core in the heat exchanger or the pipes which transport the fuel between the core and the heat exchanger. Some delayed neutrons are emitted far from the place where the fission occurred. The control method changes the level of liquid moderator in the control cavities to keep keff = 1.00000 by negative feedback. The fuel circulation system is built to remove some fission products from the circulating fuel.
MCNP does not adjust its calculations for these factors so for the MUTBR simulations MCNP is run in two modes. The first is a static kcode run to determine keff for an estimate of the level of liquid moderator in the control cavities. If keff is above or below 1.00000 by more than a small margin, a new estimate of the liquid moderator level is made based on how far from 1.00000 keff was and the run is repeated until a value that is close enough is found. Then MCNP burnup is run to determine the fuel composition after two years of full power operation. Some of the fission products are thrown away from the simulated fuel composition at the end of the run to mimic the physical removal of some fission products in the physical reactor. Then both steps are repeated for another two years until the fuel reactivity goes out of the range the control method can handle (steam bubble size below 10 or above 990 parts per thousand). This approach does not deal with the delayed neutron problem but it is expected to be much less significant than the fission product and control method issues.
Almost nothing is known about the physical removal of fission products from circulating molten uranium at 1400 degrees C. It is expected that the principle separation method will be evaporation from the liquid surface of the circulating fuel. Only 12 elements with atomic numbers in the range 30 to 70 (the fission product range) have melting points above 1400 C. so evaporation is expected to be reasonably efficient. For the simulation results shown above, at the end of each two year simulation 50% of the fission products were discarded from the simulated fuel. In addition, for the simulated cases with UNF in the initial fuel, 25% of the fission products in the initial fuel were discarded before starting the simulations. For each two years this allows the fission products produced to cause unimpeded neutron loss unlike the physical reactor which has continuous fission product removal. The simulated approach is also highly biased against removal of the strong absorbers because most of them will have already absorbed a neutron in the MCNP simulation and be transmuted to some other isotope before the end of the two year simulation. For these reasons the approach to simulating fission product removal seems relatively conservative rather than overly optimistic. In any case the simulation results highlight the need for physical experimentation on fission product removal.
The simulations of the reactor are important because they provide a way to evaluate proposed design, engineering, and fuel decisions. If the simulations are not accurate they do not lead to optimum design choices. While the simulation results shown above do not prove that any of the reactor configurations simulated is viable for a 100 year fuel life, they do suggest that burning mostly used fuel from existing light water reactors may be viable for a very long time. The widespread use of MUTBR technology would greatly reduce the need to mine and process uranium for civilian nuclear power and completely change the discussion about used nuclear fuel. By using UNF as fuel, the MUTBR puts existing and future UNF into reactors where it is safe and is generating power. Very little new UNF would be created and the problem of disposal of UNF is largely deferred until nuclear power is no longer used to generate electricity.
The MUTBR is now a proposed reactor concept. The most urgent need is for a trusted entity such as a national lab (e.g. ORNL) to run advanced simulations of the MUTBR concept. These simulations need to include various plausible rates of fission product removal to determine the sensitivity of the design to the fission product removal rate. These results could then justify the expensive experiments necessary to determine actual fission product removal rates.