Étude des risques d'ébullition de sodium dans les réacteurs à neutrons rapides

Abstract

Click to increase image sizeClick to decrease image sizeIn a fast reactor of the Super-Phénix type the sodium temperature lies between 400 and 600°C. The saturation temperature, under the pressure of the core, being of the order of 900°C, sodium boiling is not possible under normal operating conditions. Despite the large safety margins and all the safeguard devices, hypothetical accident scenarios are taken into account in the safety studies. In these accidents the increase of the bulk temperature is large enough to reach the sodium boiling and materials melting limits. Because of positive reactivity feedbacks, the voiding rate of the core influences the energy deposition, therefore sodium boiling has been considered as a risk. Typical hypothetical situations are : - subassembly faults : inlet blockage, local blockage in the bundle, - whole core accident : pump run-down without scram... A description of sodium boiling phenomena during a slow pump rundown without scram has been chosen to illustrate the progress made in the study of two-phase systems. The particular characteristics of the thermohydraulic studies presented are that they deal with a long pump coast-down time (1/10 of the nominal flow after about 10 minutes) accompanied by a reduction of the power (about 1/3 after 10 minutes) and involve a subassembly with wire-wrap spacers. Saturations conditions, based on bulk calculations, in the single channel approximation, are reached about 10 minutes after the onset of the pump run-down during a very slow transient so that stability considerations based on the Ledinegg criterion (flow excursion) are relevant. As a quantitative description of this flow excursion was not available a special program was undertaken to understand it and to get experimental data to validate computer codes. From the series of experiments performed in sodium in the single channel approximation, covering a wide range of parameters, different heated lengths, downstream geometries, and power levels, it can he concluded that the flow excursion is a rather slow process, lasting for a period of the order of several seconds. Two distinct phases can be observed : - In the first phase, there is a progressive voiding of the channel accompanied by a quiet boiling regime (bubbly flow), - In the second phase, starting with the local dry-out of the pin there are expulsions (chugging). The main reason for the rather slow decrease of the flow during the flow excursion is that the thermal inertia of the structural material and the pins is large compared to the thermal inertia of the sodium. Although great care has been taken in the definition of the characteristics of the single channel test section a major uncertainty still remains as to how far the results obtained in such a simplified geometry are representative of a subassembly. A reply will be given by an on going program on small bundles.