During a severe accident, the lower head of a reactor pressure vessel (RPV) can be subjected to significant thermal and pressure loads. The lower head has the potential to fail, releasing large amounts of molten corium into containment. The Three Mile Island accident involved the melting of about 20 tonnes of corium, which collapsed into the lower head of the RPV. Despite the presence of water, the lower head reached temperatures of ~1300 K for 30 minutes in an area with an equivalent diameter of 1 metre. During this period, the reactor cooling system (RCS) pressure was at 10 megapascal (MPa).
Although the Three Mile Island vessel did not fail, code analyses conducted in the course of the Three Mile Island Reactor Pressure Vessel Investigation Project (TMI-VIP) predicted creep rupture in prevailing conditions. This implies that the then state-of-the art modelling of lower head failure was not mature because it did not take full account of the effect of thermal loading. These methodologies have been further developed since the TMI-VIP project to analyse existing and next generation reactors from the perspective of accident assessment, management and mitigation. In order to improve and validate structural analysis codes, there was a need for experimental data on lower head deformation and failure phenomena.
The objective of the project was to investigate the timing and size of lower head failure under conditions of low reactor coolant system pressure and large differential temperatures across the lower head wall. This objective was achieved through a series of experiments at the Sandia National Laboratories in Albuquerque, New Mexico, United States, completed in June 2002.
Sandia National Laboratories completed eight United States Nuclear Regulatory Commission (US NRC)-sponsored tests on lower head failure (LHF). These tests were specifically designed to address lower head faillure issues with prototypic material and geometry. The Sandia Lower Head Failure Project extended the US NRC/Sandia National Laboratories/LHF programme to address issues such as lower RCS pressure (representative of depressurised or partially depressurised conditions) and pressure transients. These tests also represented an improvement over previous tests by simulating a large temperature gradient across the lower head wall of the RPV. The temperature gradients addressed in the tests were representative of conditions without ex-vessel cooling.
A final report was issued 2002, available below.
Belgium, Czech Republic, Finland, France, Germany, Spain, Sweden and United States.
September 1998-June 2002
USD 1.9 million