The thermal-hydraulic design of the accelerator production of tritium (APT) tungsten neutron source is presented. A carefully engineered thermal-hydraulic design is required to remove the deposited power effectively during normal operations and remove the decay power during plant shutdown and postulated accidents. For steady-state operations and operational and anticipated transients, the design criterion is to maintain single-phase flow conditions with a margin to onset of nucleate boiling. The margin is determined based on phenomenological and geometric uncertainties associated with the design. A large margin to thermal excursion limits, such as critical heat flux and onset of flow instability, also is maintained during normal operations. In general, a very robust thermal-hydraulic design can be accomplished using the traditional models and correlations available in the engineering literature. However, two issues require further attention: maintaining adequate flows in a parallel network of flow channels and minimizing the volume fraction of heavy water to maximize tritium production.

The design uses ladderlike structures that contain clad tungsten cylinders in the rungs that have coolant supplied and removed by the vertical ladder rails. Because the power density drops in the beam direction, the thickness of the tungsten cylinders is increased with increasing beam penetration length. The cooling requirement is determined using a conservative criterion where the minimum wall subcooling inside the rungs is at least 40°C and the minimum Reynolds number is 6000. Initial flow distribution tests were conducted with a full-scale model of an APT ladder assembly based on a preliminary design. Flow distributions can be made more even by using a larger riser than downcomer and also by increasing the flow resistance across each rung. The calculations discussed assume nominal dimensions, even though the power deposition and removal use a conservative approach. The effect of manufacturing tolerances will be investigated in future research. Also, the applicability of the critical heat flux and onset of flow instability models to small coolant channels is being verified experimentally. Further design optimization will be possible when these studies are completed.