UC Berkeley

The Resource-renewable Boiling Water Reactors (RBWRs) are a set of light water reactors (LWRs) proposed by Hitachi which use a triangular lattice and high void fraction to incinerate fuel with an epithermal spectrum, which is highly atypical of LWRs. The RBWRs operate on a closed fuel cycle, which is impossible with a typical thermal spectrum reactor, in order to accomplish missions normally reserved for sodium fast reactors (SFRs) – either fuel self-sufficiency or waste incineration. The RBWRs also axially segregate the fuel into alternating fissile “seed” regions and fertile “blanket” regions in order to enhance breeding and leakage probability upon coolant voiding.

This dissertation focuses on thorium design variants of the RBWR: the self-sufficient RBWR-SS and the RBWR-TR, which consumes reprocessed transuranic (TRU) waste from PWR used nuclear fuel. These designs were based off of the Hitachi-designed RBWR-AC and the RBWR-TB2, respectively, which use depleted uranium (DU) as the primary fertile fuel. The DU-fueled RBWRs use a pair of axially segregated seed sections in order to achieve a negative void coefficient; however, several concerns were raised with this multi-seed approach, including difficulty with controlling the reactor and unacceptably high axial power peaking. Since thorium-uranium fuel tends to have much more negative void feedback than uranium-plutonium fuels, the thorium RBWRs were designed to use a single elongated seed to avoid these issues.

A series of parametric studies were performed in order to find the design space for the thorium RBWRs, and optimize the designs while meeting the required safety constraints. The RBWR-SS was optimized to maximize the discharge burnup, while the RBWR-TR was optimized to maximize the TRU transmutation rate. These parametric studies were performed on an assembly level model using the MocDown simulator, which calculates an equilibrium fuel composition with a specified reprocessing scheme. A full core model was then created for each design, using the Serpent/PARCS 3-D core simulator, and the full core performance was assessed.

The RBWR-SS benefited from a harder spectrum than the RBWR-TR; a hard spectrum promotes breeding and increases the discharge burnup, but reduces the TRU transmutation rate. This led the RBWR-SS to have a very tight lattice, which has a lot of experimental uncertainty in the thermal hydraulic correlations. Two different RBWR-SS designs were created assuming different thermal hydraulic assumptions: the RBWR-SSH used the same assumptions as Hitachi used for the RBWR-AC, while the RBWR-SSM used more conservative correlations recommended by collaborators at MIT. However, the void feedback of the pure Th-fed system was too strongly negative, even with a single elongated seed. Therefore, instead of using just thorium, the self-sustaining designs were fed with a mix of between 30% and 50% DU and the rest thorium in order to keep the void feedback as close to zero as possible. This was not necessary for the RBWR-TR, as the external TRU feed fulfilled a similar role.

Unfortunately, it was found that the RBWR-SSM could not sustain a critical cycle without either significantly downgrading the power or supplying an external feed of fissile material. While the RBWR-SSH and the RBWR-TR could reach similar burnups and transmutation rates to their DU-fueled counterparts as designed by Hitachi, the thorium designs were unable to simultaneously have negative void feedback and sufficient shutdown margin to shut down the core. The multi-seed approach of the Hitachi designs allowed their reactors to have much lower magnitudes of Doppler feedback than the single-seed designs, which helps them to have sufficient shutdown margin. It is expected that thorium-fueled RBWRs designed to have multiple seeds would permit adequate shutdown margin, although care would need to be taken in order to avoid running into the same issues as the DU fueled RBWRs. Alternatively, it may be possible to increase the amount of boron in the control blades by changing the assembly and core design.

Nonetheless, the uncertainties in the multiplication factor due to nuclear data and void fraction uncertainty were assessed for the RBWR-SSH and the RBWR-TR, as well as for the RBWR-TB2. In addition, the uncertainty associated with the change in reactor states (such as the reactivity insertion in flooding the core) due to nuclear data uncertainties was quantified. The thorium RBWRs have much larger uncertainty of their DU-fueled counterparts as designed by Hitachi, as the fission cross section of 233U has very large uncertainty in the epithermal energy range. The uncertainty in the multiplication factor at reference conditions was about 1350 pcm for the RBWR-SSH, while it was about 900 pcm for the RBWR-TR. The uncertainty in the void coefficient of reactivity for both reactors is between 8 and 10 pcm/% void, which is on the same order of magnitude as the full core value.

Finally, since sharp linear heat rate spikes were observed in the RBWR-TB2 simulation, the RBWR-TB2 unit cell was simulated using a much finer mesh than is possible using deterministic codes. It was found that the thermal neutrons reflecting back from the reflectors and the blankets were causing extreme spikes in the power density near the axial boundaries of the seeds, which were artificially smoothed out when using coarser meshes. It is anticipated that these spikes will cause melting in both seeds in the RBWR-TB2, unless design changes – such as reducing the enrichment level near the axial boundaries of the seeds – are made.