Chaotic behavior in a system simulating the pressure balanced injection system. Analysis of passive safety reactor behavior. JAERI's nuclear research promotion program, H12-012 (Contract research)

The pressure Balanced Injection System (PBIS) was proposed in a passive safety reactor. Pressurizing Line (PL) connects the Reactor Vessel (RV) and the gas area in the Contain Vessel (CV), and Injected Line (IL) connects two vessels at relatively lower position. In an accident, the two lines are passively opened. The vapor generated by the residual heat pressed downward the water level in the RV. When the level is lower than the inlet of the PL, vapor is ejected into the CV through the PL attaining the pressure balance between the vessels. Then boron water in the CV is injected into the RV through the IL by the static head. This process is repeated by the succeeding vapor generation. In an experiment, the oscillating system was replaced by water column in a U-shaped duct. The vapor generation was simulated by cover gas supply to one end of the duct, while the other end was open to the atmosphere. When the water level reached a certain level, electromagnetic valves opened and the cover gas was ejected. The gas pressure decreased rapidly, resulting in a surface rise. When the water level reached another level, the valves closed. The cover gas pressure increased again, thus, gas ejection occurred intermittently. The interval of the gas ejection was not constant but fluctuated widely. Mere stochastic noise could hardly explain the large amplitude. Then was expressed the system using a set of linear equations. Various types of piecewise linear model were developed to examine the cause of the fluctuation. There appeared tangential bifurcation, period-doubling bifurcation, period-adding bifurcation and so on. The calculated interval exhibited chaotic features. Thus the cause of the fluctuation can be attributed to chaotic features of the system having switching. Since the piecewise linear model was highly simplified the behavior, a quantitative comparison between the calculation and the experiment was difficult. Therefore, numerical simulation code considering nonlinear effect was developed to describe the system behavior in the case there was no stochastic noise. The comparison between the simulation and the experiment revealed that there are some bifurcation-like phenomena being not real bifurcations. Real bifurcations are discovered in the simulation as well as in the experiment. Confirmation of chaotic behavior of the system in the experiment is a subject for a future study. (author)