Investigation of flashing-induced instabilities at the CIRCUS test facility - ATHLET calculations

A. Manera and F. Schäfer

The CIRCUS test facility has been built to study the start-up phase of a natural-circulation BWR. During the start-up so called flashing-induced instabilities can arise. These instabilities are induced by flashing, due to steam production in the long adiabatic riser section, which is placed above the core to enhance the flow rate. Flashing occurring in the riser causes an unbalance between driving force and pressure losses in the natural-circulation loop, giving rise to flow oscillations.

Within the European-Union 5th Framework Programme, a project, NACUSP, was started in December 2000, having as one of its main aims the understanding of the physics of the phenomena involved during the start-up phase of natural-circulation-cooled BWRs, providing a large experimental database and validating state-of-the-art thermalhydraulic codes in the low-pressure, low-power operational region of these reactors. One part of this project deals with the modeling of selected CIRCUS tests using the thermo-hydraulic code ATHLET.

CIRCUS is a natural-circulation water/steam loop built at the Delft University of Technology. The heated section consists of 4 parallel channels. On top of the heated channels a long adiabatic section (riser) is present. A steam dome, in which a mixture of water and steam is at saturation conditions, is used to simulate the steam dome of a reactor. To assure a constant core inlet temperature, a buffer vessel is used.


The calculations were performed with the Code ATHLET. The input dataset models all main parts of the test facility. The core section is modeled by 4 parallel channels. Each channel is simulated as a pipe with an electrical heater. For the core section, riser, downcomer, upper and lower horizontal parts the exact dimensions, heat capacities of glass and copper, heat losses and pressure drops are considered. The riser is modeled as an adiabatic section. The heat exchanger is represented by a simplified model with a special bypass. In all control volumes the 5-equation model (separate conservation equations for liquid and vapour mass and energy, mixture momentum equation) and the full-range drift-flux model of ATHLET are used.

Before each transient calculation a steady state calculation is performed. The steady state calculation starts with zero power. After a few seconds the core power is switched on. The transient calculation starts if the mass flow in the loop is stable. Due to the void production in the riser, the natural circulation mass flow is increased. If the temperature in the upper part of the riser reaches saturation conditions, the flashing induced instabilities occur. The period of the instabilities calculated by ATHLET show a good agreement with the experimental data. The calculated amplitude of the flow oscillations is higher than in the experiment. The ATHLET calculations show, that the period of the instabilities mainly depends on the correct modeling of heat structures and heat losses. Without heat losses and without consideration of heat structures the calculated period is appr. 20-30 % shorter.

The comparison between calculation and experiment shows:

  • the inset and period of the instabilities are calculated in a good agreement with the experimental data
  • in the experiments the period of the instabilities depends on the core power and core inlet temperature; the ATHLET calculations reproduce this behaviour in a qualitative and also quantitative way
  • in the ATHLET model the steam dome is connected to a time dependent volume; due to the simplified steam dome model the calculated mass flows are higher than in the experiment
  • the correct calculation of the instabilities requires the modeling of the heat structures and heat losses as exact as possible.


More about the ATHLET-calculations and the experiments:

Investigation of flashing-induced instabilities at the CIRCUS test facility using the code ATHLET (PDF 728 kb)