System transient analysis code development for low pressure and low power

Abstract

A real time reactor system analysis code, ARTIST, based on drift flux model has been developed to investigate the transient system behavior under low pressure, low flow and low power conditions with noncondensible gas present in the system. The governing equations of the ARTIST code consist of three mass continuity equations (steam, liquid and noncondensibles), two energy equations (gas and mixture) and one momentum equation (mixture) constituted with the drift flux model. The capability of ARTIST in predicting two-phase flow void distribution in the system has been validated against experimental data. The results of the ARTIST axial void distribution at low pressure and low flow, are far better than the results of both the homogeneous model of TASS code and the two-fluid model of RELAP5/MOD3 code. Also, RELAP5/MOD3 calculation shows the large amplitude of void fraction oscillations at low pressure. These results imply that interfacial momentum transfer terms in the two-fluid model formulation should be carefully constituted, especially for the low pressure condition due to the big density differences between steam and water. Thermal-hydraulic state solution scheme is developed when noncondensible gas exists. Numerical consistency and convergence of obtaining equilibrium state is tested with the ideal problems for various situations including very low partial pressure conditions. Calculated thermal-hydraulic state for each test shows consistent and expected behaviour. A new multi-layer back propagation network algorithm for calculating the departure from nucleate boiling ratio (DNBR) is developed and adopted in ARTIST code in order to have real-time DNBR evaluation by eliminating the tandem procedure of the transient DNBR calculation. The algorithm trained by different patterns generated by latin hypercube sampling method on the performance space is tested for the randomly sampled untrained data and the transient DNBR data. The uncertainty of the algorithm i...