A series of single-rod LOCA tests (IFA-650) is being carried out in the Halden reactor in Norway, with particular focus on the behaviour of a high burnup fuel rod under LOCA conditions. The first two trial runs were carried out on unirradiated PWR rods, in May 2003 and May 2004 respectively. The third test was conducted on April 30th 2005 with a high burnup irradiated fuel rod.
An analysis of this test series has been undertaken at IRSN, with use of the ICARE/CATHARE computer code and its stand-alone module ICARE2. Within this frame, IRSN is participating to the OECD benchmark exercise that has been organized on the third test IFA-650.3. A blind calculation was performed with the data package provided by the Halden Reactor Project and some outcome of the trial runs performed on the previous tests IFA-650.1 and IFA-650.2. Revised calculations have been performed after the release of experimental results in July 2005.
The computer code ICARE/CATHARE which combines the capabilities of CATHARE2 for overall thermohydraulic analysis with the refined geometrical and physical core modeling capabilities implemented in ICARE 2 was particularly adapted for the analysis of the IFA-650 LOCA tests.
The CATHARE thermohydraulic model can handle the particular features of the IFA-650 tests with the establishment of an initial phase under natural circulation of water within the test train subchannels, followed by the blowdown and heatup phases simulating the LOCA transient.
The ICARE core module can describe easily the experimental test section containing one rod with concentric structures, with radiative heat exchanges between these rod and structures. It can also handle the cooling spray injection in IFA-650 tests as a hydraulic source term at the desired elevation in the concerned subchannel. The thermomechanical model is based on secondary creep velocity and burst stress derived from the EDGAR experiments. The cladding oxidation model uses the Cathcart-Pawel correlation with 2 sided oxidation after burst, over an user input length on inner side.
Preliminary calculations of the IFA-650.1 and IFA-650.2 tests were performed previously with the fast running stand alone module ICARE2, using fixed hydraulic boundary conditions corresponding to the post blowdown stabilized state. These calculations have demonstrated the dominant contribution of radiative transfers to thermal behavior and therefore the sensitivity of results to the emissivities of surface materials involved in these transfers.
Preliminary comparisons of the blind calculation results with experimental results released afterwards showed that the fuel rod thermal behavior is correctly predicted, whereas the heater temperature plateau is somewhat underpredicted, likely indicating an improper radiative transfer balance on the heater faces. As for the rod cladding mechanical behavior, calculated burst occurs earlier than in the experiment, but with a similar burst temperature. However the calculated rod internal pressure exhibits a much larger increase before burst than in the experiment, which was certainly due to an incorrect evaluation of the temperatures in the volume expansion during the transient. A revised calculation with an axial meshing of the large expansion volume extending from the hot top of fuel up to the cold top of test train has permitted to reproduce fairly well the internal rod pressure evolution and to obtain a very good agreement between calculated and measured clad temperatures.
With the same rod filling pressure as in test IFA-650.2 and with use of the mechanical properties of as-received Zry4, the calculated strain profile at burst in test IFA-650.3 is very similar to that calculated for the IFA-650.2 unirradiated rod, with maximum calculated strain reaching 32%. This value is consistent with the maximum strain values observed on ICL#1-4 tests performed at ANL, but much larger than the experimental value obtained on the IFA-650.3 rod ( 9%). However, the recently performed post-irradiation examinations of this rod have shown that the clad deformation and rupture were suspect, likely due to defects caused by the attachment of the lower TC on rod cladding. Consequently, it has been decided at the last EHPG Meeting that the IFA-650.3 test will be repeated in next experiment.