The once through steam generator (OTSG) is a single pass counter flow heat exchanger in which primary pressurized water from the core is circulated. Main Feedwater is injected in an annular gap on the outer periphery of the steam generator shroud such that it aspirates steam to preheat the feedwater to saturation temperature. An important component of the OTSG and enhanced once through steam generator (EOTSG) is the auxiliary feedwater system (AFW), which is used during accident/transient scenarios to remove residual heat by injecting water through jets along the outer periphery of the heat exchanger core directly on to the tubes at the top of the OTSG. The intention is for the injected water, which is subcooled, to spread into the tube nest and wet as many tubes as possible. In this project, the main objectives were to use first principles Computational Fluid Dynamics to predict the number of wetted tubes versus flow rate in the EOTSG at the AFW injection location above the top tube support plate. To perform the fluid analysis, the losses in the bypass leakage flow and broached hole leakage flow were first quantified and then used to model a 1/8th sector of the EOTSG. Using user defined functions (UDF), the loss coefficients of the leakage flows were implemented on the 1/8th sector of the EOTSG computational model to provide boundary conditions at the bypass flow and leakage flow locations With this method, the number of tubes wetted in the sector of EOTSG for various AFW flow rates was found. Results showed that the number of wetted tubes was in very close agreement to that predicted by experimental-analytical methods by the sponsor, AREVA. With the maximum flow rate of 65 l/s a total of 318 tubes were wetted and the percentage of tubes wetted with broached holes was 8.7%.

The analysis on the bypass leakage flow showed that the loss coefficients was a function of the mass flow rate or the flow Reynolds number through the gap and it increased as the Reynolds number increased from 300 to 1600. The experimental and computational loss coefficients agree to within 15% of each other. In contrast, the constant loss coefficient of 1.3 used by AREVA was much higher than that obtained in this study, particularly in the low Reynolds number range. As the Reynolds number approached 3000, the loss coefficients from this study approached the value of 1.3. This value of the loss coefficient was implemented for the bypass flow leakage in the 1/8th sector of the EOTSG model.

The analysis on the broached hole leakage flow was performed using a single hole, five holes, and one, two, four and eight rows of broached holes in order to characterize the loss coefficients. The one hole and five hole computational models were validated with experiments. The computational models showed the presence of voids in the leakage flow through the tube support plate (TSP), which were not observed (visually) in the experiments. The characterization of the broached hole leakage in the one, two and four rows showed that the loss coefficient of the control broached hole increased as the number of rows increased. These results indicated that for the same height of water on the TSP, the resistance to leakage flow increased as the number of tubes increased. They also indicated that leakage flow through the broached holes was not solely a function of the height of water above the TSP but also the surrounding geometrical topology and the flow characteristics. However, the analysis done for eight rows showed that the loss coefficient became constant after a certain number of rows as the loss coefficient differed by only 5% from the results of the four rows. From these results it was determined that the loss coefficient asymptotes to an estimated value of 4.0 which was implemented in the broached hole leakage flow in the 1/8th sector of the EOTSG.

Computational models of the 1/8th sector of the EOTSG were implemented with the respective loss coefficients for the bypass and leakage flows. Results showed that as the AFW flow rate increased, the percentage wetted tubes increased. The data matched closely with AREVA's experimental-analytical model for flow rates of 14.5 l/s and higher. It was also deduced that complete wetting of the tubes is not possible at the maximum AFW flow rate of 65 l/s.