During the blowdown phase of a large-break loss-of-coolant accident (LOCA), high-powered regions of a nuclear reactor core exceed the critical heat flux, and peak-cladding temperatures (PCTs) in the range from 1033 to 1255°C (1400 to 1800°F) can occur before high flow rates terminate the temperature rise. Following the blowdown PCT, the core is cooled by a dispersed droplet flow at moderately high pressures. Heat transfer during this blowdown cooling period removes a significant portion of stored energy from the core and can lead to quench. Energy removal during this period is an important feature in the AP600 advanced passive plant design because the design includes good communication between the core and water in the reactor vessel upper head.

For realistic LOCA calculations performed according to the revised 10 CFR 50.46 regulation using the code scaling, applicability, and uncertainty methodology, it becomes necessary to quantify those physical processes that have a dominant effect on the transient. Blowdown cooling heat transfer determines the initial condition of the core at the start of reflood and thus is a major contributor to the propagation of uncertainty in later periods of the transient.

The approved Westinghouse best-estimate LOCA methodology uses the WCOBRA/TRAC-MOD7A code to predict the large-break LOCA transient. Quantification of blowdown cooling heat transfer coefficients (HTCs) was performed by development of a driver-plotter routine based on the WCOBRA/TRAC heat transfer package. Known fluid conditions corresponding to experimental data were then used to generate predictions of blowdown cooling HTCs. A distribution, developed from comparisons to several experimental tests, of predicted versus measured HTCs shows the average bias and uncertainty in the heat transfer package.

Based on driver-plotter routine calculations of blowdown HTCs measured in the Oak Ridge National Laboratory blowdown tests and Idaho National Engineering Laboratory high-pressure single-tube tests, a new correlation for direct-contact heat transfer is proposed. Use of the new model was found to provide significantly improved agreement between predicted and measured HTCs.