Flooding of a large, passive, pressure-tube light water reactor

Abstract

A reactor concept has been developed which can survive loss of coolant accidents without scram and without replenishing primary coolant inventory, while maintaining safe temperature limits on the fuel and pressure tubes. The proposed concept is a pressure tube type reactor of similar design to CANDU reactors, but differing in three key aspects. First, a solid SiC-coated graphite fuel matrix is used in place of fuel pin bundles to enable the dissipation of decay heat from the fuel in the absence of primary coolant. Second, the heavy water coolant in the pressure tubes is replaced by light water, which also serves as the moderator. Finally, the calandria tank, surrounded by a graphite reflector, contains a low pressure gas instead of heavy water moderator, and this normally-voided calandria is connected to a light water heat sink. The cover gas displaces the light water from the calandria during normal operation, while during loss of coolant or loss of heat sink accidents it allows passive calandria flooding. Calandria flooding also provides redundant and diverse reactor shutdown. This paper describes the thermal hydraulic characteristics of the passively initiated, gravity driven calandria flooding process. Flooding the calandria space with light water is a unique and very important feature of the proposed pressure-tube light water reactor (PTLWR) concept. The flooding of the top row of fuel channels must be accomplished fast enough so that in the total loss of coolant, none of the critical components of the fuel channel, i.e. the pressure tube, the calandria tube, the matrix and the fuel, exceed their design limits. The flooding process has been modeled and shown to be rapid enough to maintain all components within their design limits. Two other considerations are important. The thermal shock experienced by the calandria and pressure tubes has been evaluated and shown to be within acceptable bounds, while simultaneously cooling the graphite fuel matrix. This has been demonstrated for the bounding scenarios of sustained film boiling and rapid quenching to yield nucleate boiling on the calandria tube outer surface. Finally, although complete flooding renders the reactor deeply subcritical, various steam/water densities can be hypothesized to be present during the flooding process which could cause reactivity to increase from the initially voided calandria case. One such hypothesis which leads to the maximum possible density of the steam/water mixture in the still unflooded calandria space is entrainment from the free surface. It is shown that the steam/water mixture density yielding the maximum reactivity peak cannot be achieved by entrainment because it exceeds thermohydraulically attainable densities of steam/water by an order of magnitude.