American Nuclear Society
Home

Home / Publications / Journals / Nuclear Science and Engineering / Volume 140 / Number 2 / Pages 152-164

Three-Dimensional Modeling of a Steam-Line Break in a Boiling Water Reactor

Hernan Tinoco

Nuclear Science and Engineering / Volume 140 / Number 2 / February 2002 / Pages 152-164

Technical Paper

Because of weld problems, the core grids of Units 1 and 2 at the Forsmark nuclear power plant have been replaced by grids of a new design, consisting of a single machined piece without welds. The qualifying structural analysis has been carried out considering dynamic loads, which implies that even loss-of-coolant accidents have to be included. Therefore, a detailed time description of the loads acting on the different internal parts of the reactor is needed. To achieve sufficient space and time resolution, a computational fluid dynamics (CFD) analysis was considered to be a viable alternative.

A CFD analysis of a steam-line break in the boiling water reactor of Unit 2 is the subject of this work. The study is based on the assumption that the timescale of the transient analysis is smaller than the relaxation time of the water-steam system. Therefore, a simulation of only the upper, steam part of the reactor with no two-phase effects (flashing) is feasible.

The results obtained display a rather complex behavior of the decompression process, forcing the analysis of the pressure field to be accomplished through animation. In contrast, the computed instantaneous forces over different internal parts oscillate regularly and are approximately twice the forces estimated in the past by simpler methods, with frequencies of 30 to 40 Hz; top amplitudes of ~1.64 MN; and relatively low damping, ~25% after 0.5 s.

According to the present results, this type of modeling is physically meaningful for simulation timescales smaller than the water-steam relaxation time, i.e., ~0.5 s at reactor conditions. At larger times, a two-phase model is necessary to describe the decompression process since two-phase effects are dominant. The results have not yet been validated with experiments, but validation computations will be run in the future for comparison with results of the Marviken tests.

 
Questions or comments about the site? Contact the ANS Webmaster.
advertisement