Abstract
In the conceptual design study of the Japan Sodium-cooled Fast Reactor (JSFR), sodium fire accident would be eliminated by adopting design measures against sodium leakage. Nevertheless, the thermodynamic consequences of sodium fire accident might still need to be evaluated in order to confirm the safety margin of the plant. The authors are therefore initiating systematic verification and validation (V&V) activity for sodium fire evaluation tools: i.e. the computational codes for sodium fire. The V&V activity is in progress with the main focuses on already developed sodium fire analysis codes SPHINCS and AQUA-SF. The events to be evaluated are hypothetical sodium spray, pool, or combined fire accidents followed by thermodynamic behaviors postulated at an air atmosphere in a containment building of the JSFR. The present paper describes that a preliminary 'Phenomena Identification and Ranking Table (PIRT)' is developed at first for clarifying the important validation points in the sodium fire analysis codes, and that an 'assessment matrix' is proposed which summarizes both separate effect tests (SET) and integral effect tests (IET) for validating the computational models or whole code for important phenomena. Furthermore, the paper shows an individual validation with SET in which the spray droplet combustion model of SPHINCS and AQUA-SF predicts the burned amount of a falling sodium droplet with the error mostly less than 30%.
