Coupled Melt Pool Convection and Vessel Creep Failure: The FOREVER Program

The FOREVER (Failure Of REactor VEssel Retention) program is concerned with the phenomena of melt-vessel interactions during a postulated severe accident in a light water reactor. The objectives of the FOREVER program are to obtain data and develop validated models on (i) the melt coolability process inside the vessel, in the presence of water (in particular on the efficacy of the postulated gap cooling to preclude vessel failure) and (ii) the lower head failure due to the creep process in the absence of water inside or outside of the lower head.
Integral experiments were performed in a 1/10 scaled carbon steel vessels of 0.4 m diameter, 15 mm thickness and 60 mm height. Up to 20 litres of binary oxide melt (30wt% CaO -70wt% B2O3 ) was poured into the vessel and maintained at about 1300 °C by a specially designed electrical heater operating at about 40 kW. The melt pool undergoes natural con-vection as it would in the prototypic scenario and the vessel wall temperatures vary from ~ 600 to 1000 °C azimuthally. The pressure inside the vessel was maintained at about 2.6 MPa. The main diagnostics were several types of thermocouples and linear position transducers (LPT) which measure the displacement of the vessel due to its creep deformation.
The first three experiments, described here, are focussed on vessel creep and failure.The FOREVER/CI test was performed with the German steel (15Mo3) lower head, at about 25 bars of internal pressure and at the input power level of 22 kW. The maximum essel wall temperature was about 800 °C. A sizable database was obtained for the creep deformation rates over a period of 24 hours and the maximum creep strain obtained was about 5 %. The second test FOREVER/C2 employed a French reactor vessel steel (16MND5) lower head, and a power level of 40 kW. The maximum vessel wall temperature was measured to be about 1000 °C and the maximum creep strain obtained was about 10 %. The test ended without the failure of the lower head as the heater failed due to its unconvering from the melt as the melt level receded due to vessel expansion.
The third test, the EC-FOREVER-I, incorporated higher initial melt level to avoid heater failure. Although, the heater power level was maintained same, at about 40 kW, the internal pressure was increased to 28 bars in order to obtain the lower head failure. The lower head failure was achieved, albeit, at a lower than expected creep strain of only 6%. The failure site was located just above the welding joint between the hemispherical part and the cylindrical part. The cause for this mode of failure is being investigated.
A coupled thermal-structural analysis of these FOREVER tests was performed with the ANSYS Multiphysics code. An improvised creep model was incorporated into this code which avoids the use of a single creep law for the entire lower head. Instead, a three dimensional array was developed where the creep strain is evaluated according to the actual total strain, temperature and equivalent stress for each element. The material damage is evaluated considering the creep and the prompt plastic deformations. The calculated results for the creep strain and vessel failure time are in good agreement with those from the experiments.