RETRAN-3D, a system analysis code currently employed by the nuclear industry in studies covering a wide variety of operational and accident scenarios, has not been extensively validated for application to loss-of-coolant accident (LOCA) scenarios.

The results of the in-depth analysis of two experimental loss-of-coolant transients, namely, Test No. 9 in the French OMEGA facility, and the International Standard Problem 26 (ISP-26) in the Japanese ROSA-IV Facility are discussed. The OMEGA test simulated the blowdown phase of a double-ended cold-leg break, whereas the ISP-26 test simulated a small break (5%) in a full height, volume (1/48), and power (~1/342) scaled facility representing a typical two (or four)-loop pressurized water reactor (PWR) system.

The RETRAN-3D results for the OMEGA test show good estimates of the important system parameters, with the best agreement corresponding to the use of the dynamic-slip flow model. A sensitivity analysis on the break flow showed that the Henry/Fauske-Isoenthalpic Expansion critical flow model yields the best results, which are significantly improved with a refined nodalization upstream of the break.

The ISP-26 was also simulated using the dynamic-slip flow model. The results indicate that the code is able to calculate a small-break LOCA with a model including the main PWR system components and to reproduce the principal physical processes in a reasonable manner.

In summary, this assessment shows the ability of RETRAN-3D to model LOCA scenarios in a reasonable way and also points to areas where further model improvement could result in more accurate simulations.