Coupled RELAP5/GOTHIC model for IRIS SBLOCA analysis

Historically, accident analyses for nuclear power plants have been successfully performed using system thermal-hydraulics codes. In case of complex problems involving solution of thermal-hydraulics together with other disciplines such as: reactor physics, chemistry, aerosol dynamics, metallurgy; system codes are usually not enough or prediction capability can be improved by addition of separate computational models. The similar is true in case when interaction of different solution domains should be calculated (e.g. primary system and containment) with different physical models, or very different spatial or time discretization. An interesting example of such kind of interaction of different evaluation models is given by containment and reactor coolant system accident analysis of the IRIS. Previously, the different physical phenomena involved in the analyses and the need for different spatial and time discretization have led to the development of separate and specialized computer codes and evaluation models for the analysis of these two systems. The different mathematical models available are typically used independently based on external iterations and appropriate boundary conditions. In fact, the interaction of the reactor coolant system and containment is typically analyzed with two independent runs. First the mass and energy (MandE) released from the reactor versus time is calculated by a system code using a conservatively (low), bounding, containment pressure, and then the containment response is calculated for that MandE release versus time. This approach is usually sufficient for current LWR reactors. In new advanced passive reactor systems, interaction of the coolant system and containment is much more important since it impacts on the evolution of the transients. Therefore, a different modeling strategy is needed. The most straightforward approach for analyzing interacting systems would be to adopt a single evaluation model for the two coupled systems. However, the time that would be needed to develop new codes where all required models are properly incorporated, to build user experience, and to qualify the code would be prohibitively long and expensive. Therefore, the coupling of the thermal-hydraulic and containment codes can be an interesting approach and compromise between two modeling strategies mentioned above. Separate, existing computer codes can be coupled providing new capabilities without spending too much time in development and with possibility to use existing experience and perform code verification and validation only for the coupling portion of the new code. This coupling is usually performed as an extension of the classical calculation approach and it is localized at the physical points where communication between system and containment exists. For IRIS, it was decided to develop an explicit coupling of RELAP5/mod3.3, and one of the earlier versions of the GOTHIC code available at University of Zagreb, GOTHIC 3.4e; thus taking advantage of the rather large experience base in the use of the RELAP5 and GOTHIC codes as well as knowledge of their internal structure The primary goal was to explore applicability of coupled code to safety analyses of the new reactor systems where the primary system and containment closely interact. The chosen coupling strategy is simple and basic operation of constituent codes and corresponding input data are unaffected by the coupling process. This paper describes the coupled code as well as the development of the preliminary IRIS SBLOCA evaluation model and its use. Also, a discussion on the verification and validation of this methodology is provided.(author)