ON THE VALIDATION OF VESTA 2.2.0 USING THE ARIANE - GU3 SAMPLE

The validation of the VESTA 2.2.0 Monte Carlo depletion code has been initiated using the Spent Fuel Isotopic Composition Database (SFCOMPO). The work presented in this paper is limited to one fuel sample, the GU3 PWR - UOX sample from the ARIANE program, which has a reported burn up of 52.5 MWd.kgHM - 1 . The chemical analyses of the studied fuel sample were performed by 2 independent laboratories at the end of irradiation and cooling time. US and European evaluated nuclear data libraries, namely ENDF/B - VII.1 and JEFF -3.2, but also the more recent ENDF/B - VIII.0 and JEFF - 3.3 are used for the VESTA 2.2.0 calculations. The isotopic concentration results are compared to experimental data and the C/E agreement is analyzed in the light of the previous VESTA 2.1.5 validation results obtained using ENDF/B - VII.0 and JEFF - 3.1 nuclear data libraries.


INTRODUCTION
VESTA is a Monte Carlo depletion interface code developed by IRSN. From its inception, VESTA is intended to be a "generic" interface code so that it will ultimately be capable of using any Monte Carlo code or depletion module and that can be tailored to the user's needs. VESTA 2.1.5 [1] allowed for the use of any version of MCNP(X) [2] with ORIGEN 2.2 [3] or PHOENIX [1] depletion module, developed by IRSN. A new version of the code, VESTA 2.2.0, has been recently finalized. This new version introduces several new features and improvements, like the output of the secondary particles spectra and the integration of the MORET 5 [4] Monte-Carlo code. Depletion codes such as VESTA are widely used to calculate the evolution of a material subjected to radiation (be it neutrons or another type of particle) for a wide variety of applications in the fields of nuclear safety, radiation protection and environmental health safety. Other important applications for this type of codes include fuel cycle studies, criticality safety, nuclear safeguards, waste characterization, etc. For application in these fields, experimental validation is paramount. VESTA 2.1.5 has been validated for an isotopic inventory (of 65 isotopes) from 41 samples of radiochemical measurements and 52 measurements of residual power, using the JEFF-3.1 and ENDF/B-VII.0 nuclear data libraries [5,6,7]. It contains a wide variety of reactors (PWR, BWR, VVER) and fuel (UOX, MOX) types with and without burnable absorbers. It is validated using MCNPX-2.6.0 as the transport module and PHOENIX as the depletion module. For the experimental validation of VESTA 2.2.0, new samples with radiochemical analysis are being added, namely JAERI BWR UO2-Gd2O3 fuel samples and REGAL PWR UO2-Gd2O3 fuel samples measured at low burn up. VESTA 2.2.0 experimental validation, using the JEFF-3.2 and ENDF/B-VII.1 nuclear data libraries, is ongoing. It will be based basically on several fuel samples used for VESTA 2.1.5 validation, in particular the GU3 sample of the ARIANE program [8] from the SFCOMPO database. For this paper, we will limit ourselves to this GU3 sample extracted from a PWR UO2 assembly.

The VESTA code
VESTA is a code developed by IRSN, which couples a Monte Carlo neutron transport code with a depletion module. In VESTA, both the Boltzmann and Bateman equations are coupled and form a closed and complete system of equations that describes the transport of particles as a function of time. The basic tasks of VESTA can be summarized as follows for every step in the evolution calculation:  Perform a steady state transport calculation and update the spectral and spatial averaged one group reaction rates, cross sections (for every possible nuclide and reaction in the transmutation chains), flux values -and other relevant data for use by the depletion module;  Solve the Bateman equations for every burn up zone in the problem using the data derived by the transport calculation;  Pass on the new material composition for the next time step. VESTA allows the depletion of materials for different geometrical scales, and the calculation of the material isotopic compositions during irradiation at the fuel rod level. VESTA 2.2.0 allows for the use of either MCNP(X/6) or MORET 5 transport codes, and both ORIGEN 2.2, PHOENIX, or FISPACT [9] depletion modules. MCNP6 is used by default when choosing MCNP transport module with VESTA 2.2.0 version. We will limit ourselves to the use of MCNP6 and PHOENIX for the validation case presented in this paper.

Description of the simulation
The Actinide Research In A Nuclear Element (ARIANE) program [8] examined irradiated MOX (Mixed OXide) and LEU (Low Enriched Uranium) fuel samples in both commercial PWR (Pressurized Water Reactor) and BWR (Boiling Water Reactors) power reactors. The ARIANE.GU3 sample is a PWR UO2 sample with an estimated burn up of 52.5 MWd.kgHM -1 . It has been irradiated in the Gösgen PWR in Switzerland between 1994 and 1997 during three cycles. An overview of the sample characteristics is given in Table I. The fuel rod containing the GU3 sample was extracted from the original assembly after two cycles and inserted into another assembly for the third and final cycle. After ~2 years of cooling time, the 10 cm sample was cut from the middle of the rod, between 122.42 and 132.42 cm from the fuel bottom. It was analyzed by two of the laboratories involved in the ARIANE program, being ITU (Institute for Transuranium Elements, Germany) and SCK-CEN (Belgian Nuclear Research Center, Belgium). One eighth of the 15x15 fuel assembly containing the fuel sample has been modeled in 2D using VESTA 2.2.0. Every pin is considered separately (without any radial zones) for a total of 32 zones as illustrated in Figure 1. The surrounding assemblies were not included in the model (but their effect is simulated by applying optical reflections around the 3 surfaces). The rod position change (from F7 to G5) between the second and third irradiation cycle has been modeled explicitly. All the other fuel materials were replaced by new compositions. The water moderator in this model contains an average boron concentration of 650 ppm. The moderator and fuel pin cladding temperatures are set to 600 K and the fuel temperature at 900 K. No temperature evolution as a function of time has been modeled. The power history of the assembly is deduced from the power seen by the sample specified in the experimental report. This irradiation history was subdivided into 54 steps with a maximum burn up of 1 MWd.kgHM -1 , in accordance with the procedure laid out in [5].

Validation overview
The general calculation flow when performing an experimental validation calculation using radiochemical assay data can be split up into two steps. The first step is an iterative calibration of the irradiation history using burn up tracers to accurately assess the burn up of the sample and to provide a set of reference calculated to experiment (or C/E) values for the measured nuclides. After determining the reference C/E values, additional calculations have to be made to define an uncertainty range for every individual nuclide. This includes the measurement uncertainty but also the uncertainty introduced due to the irradiation history calibration. The combined value of 145 Nd, 146 Nd, and 148 Nd burn up tracers (for each laboratory) has been applied for the normalization of the irradiation history. According to the VESTA validation procedure [5], the calibration consists of determining a global renormalization factor which is applied to the irradiation history as a whole. This renormalization constant is determined through iteration until the relative difference between the combined C/E value for the burn up tracers to the target C/E value (being 1.000 in this case) is less than 0.1 %. The power history used for the calculation of this sample is based on the sample power from which VESTA derives the assembly power distribution. It should also be expressed that 238 U composition results are not significant because 238 U is the reference nuclide for the experimental validation of VESTA, by which all the composition results are normalized to get rid of unit issues, as explained in [5].

Validation results using ENDF/B-VII.1 and JEFF-3.2 libraries
The result of the calibration and the estimation of the burn up range for both analyses is given in Table II for both ENDF/B-VII.1 and JEFF-3.2 libraries. The burn up estimates of the sample based on the combined value of Nd isotopes are respectively of the order of 6-7 % and 0.5-1 % below the reported burn up value of the sample (which is 52.5 MWd.kgHM -1 ) for the ITU and SCK-CEN analysis for both libraries. The final error on the burn up value is about 1.5 % at 3σ for the SCK analysis, while 5.5 %  [10]. Their associated uncertainty due to burn up is not displayed on the table, for visibility reasons, but the whole uncertainty range is displayed on Figure 2 for 238 Pu.
The results are provided only for the SCK-CEN analysis, as the ITU results are suspected to exhibit some measurement issues as explained before. It should be noted here that the impact of using a new VESTA version, as well as a new MCNP version (MCNP6 instead of MCNPX-2.6.0) has been tested and found to be not significant, both on the flux, the k∞ and the isotopic concentrations, which allows to derive conclusions on impacts between libraries.       The only exceptions are: 235 U, which is decreased by 3 % for both cases but is still in a good agreement with the experiment; 238 Pu, which is increased by 8 % in JEFF-3.3 and also stays in a good agreement; 244 Pu, which is increased by 5 % in ENDF/B-VIII.0; 243 Am, which is decreased by 5 % in JEFF-3.3 and becomes in a good agreement, and finally 244 Cm, 245 Cm and 246 Cm which are increased by almost 10 % in JEFF-3.3, which improve the results. The impact on the C/E results obtained for fission products of the new ENDF/B-VIII.0 instead of ENDF/B-VII.1 is not significant. The exception is 90 Sr, which is strongly increased. For JEFF-3.3, on the other hand, a significant amount of fission products are impacted, among which europium isotopes, 155 Gd, 109 Ag, 125 Sb, 129 I and 125 Cs. However, due to their large measurement uncertainty, it remains difficult to conclude.

CONCLUSIONS
The validation of the VESTA 2.