CASMO5

As a result of steady advances of computer power, continuous-energy Monte Carlo depletion analysis is attracting considerable attention for reactor burnup calculations. The typical Monte Carlo analysis is set up as a combination of a Monte Carlo neutron transport solver and a fuel burnup solver. Note that the burnup solver is a deterministic module. The statistical errors in Monte Carlo solutions are introduced into nuclide number densities and propagated along fuel burnup. This paper is towards the understanding of the statistical implications in Monte Carlo depletions, including both statistical bias and statistical variations in depleted fuel number densities. The deterministic Studsvik lattice physics code, CASMO5, is modified to model the Monte Carlo depletion. The statistical bias in depleted number densities is found to be negligible compared to its statistical variations, which, in turn, demonstrates the correctness of the Monte Carlo depletion method. Meanwhile, the statistical variation in number densities generally increases with burnup. Several possible ways of reducing the statistical errors are discussed: 1) to increase the number of individual Monte Carlo histories; 2) to increase the number of time steps; 3) to run additional independent Monte Carlo depletion cases. Finally, a new Monte Carlo depletion methodology, called the batch depletion method, is proposed, which consists of performing a set of independent Monte Carlo depletions and is thus capable of estimating the overall statistical errors including both the local statistical error and the propagated statistical error.

During LWR Reactivity Initiated Accidents (RIA), the accurate evaluation of the Doppler reactivity feedback depends on the Doppler coefficient computed by the lattice physics code (e.g. CASMO5), and on the effective Doppler temperature computed by the transient code (e.g. SIMULATE3-K) using the non-uniform intra-pellet temperature profile. CASMO5 has many new features compared with its predecessor. Among them, the replacement of the L-library (based primarily on ENDF/B IV data) by the latest available nuclear data (ENDF/B VII.0), and the Monte Carlo based resonance elastic scattering model to overcome deficiencies in NJOY modeling have a significant impact on the fuel temperature coefficient, and hence on LWR RIA. The Doppler temperature effect in thermal reactors is driven by the 238U absorption. The different effective Doppler temperature definitions, available in the literature, try to capture the considerable self- shielding of the 238U absorption that occurs in the pellet surface by defining an appropriate fuel temperature to compute cross-sections. In this work, we investigate the effect of the nuclear data generated by CASMO5 on RIA, as well as the impact of different effective Doppler temperature definitions, including one proposed by the authors. It is concluded: 1) LWR RIA evaluated using CASMO5 cross section data will be milder because the energy released is ~10% smaller; 2) the prompt enthalpy rise is barely affected by the choice of the Doppler temperature definition; and 3) the peak fuel enthalpy is affected by the choice of the Doppler temperature definition, the under- prediction of the Doppler reactivity by the ‘NEA’ Doppler temperature results in a conservative estimate of the peak fuel enthalpy.

The huge absorption cross sections of Gd-155 and Gd-157 cause strong spatial shielding effects in Gd-bearing pins. A quadratic depletion model has been implemented in CASMO5 in order to address the issue of small depletion steps typically required for Gd-bearing fuel assemblies. In this model, the microscopic absorption reaction rates of gadolinium isotopes are assumed to be quadratic functions of the number density of Gd-155. This quadratic function assumption models the variations of the spatial shielding effects over the depletion step and therefore improves the overall accuracy of depletion calculations. With this new model, a depletion step size four times larger than the step size for a conventional predictor-corrector method can be used for Gd-bearing assemblies without compromising accuracy.

A new ENDF/B-VII gamma library for 214 nuclides/materials in 18 gamma/70 neutron energy groups has been built for inclusion with the CASMO5 lattice physics code. The methodology for developing this library uses NJOY-99.259 and other newly developed processing codes in contrast to the nuclear data and the methods of 1980’s employed for the current CASMO4 gamma library. Validation and verification of the new library is mainly done through CASMO5 and MCNP-5 comparisons. Gamma energy depositions are also analyzed for various test problems and gamma TIP full-core SIMULATE3 predictions are compared for the new and the (old) production CASMO4 gamma libraries.

Reactor shutdown refueling is a crucial task for BWR operations. In this paper, based on a realistic BWR core refueling sequence, the SIMULATE3 and SIMULATE5 shutdown margin (SDM) predictions are benchmarked against CASMO5-M results. The 2D core in this exercise is established from the middle plane of a typical 748-bundle BWR core. There are thousands of fuel move steps from the end of previous cycle to the beginning of next cycle, most of which are so-called partially loaded core configurations where empty fuel locations are filled with water. During this process, eigenvalue predictions from CASMO5-M and SIMULATE3/5 are compared for the all-rod-in (ARI) case and the ARI except the maximum-worth rod (ARI-M) case. Based on more advanced neutronic models, SIMULATE5 shows sizable improvements in agreeing with the CASMO5-M reference solution over the SIMULATE3 results for partially loaded cores.

This paper presents results from the critical experiment portion of the validation process for the Studsvik Scandpower CASMO5 lattice physics code and its ENDF/B-VII neutron data library. One major advantage of comparing to criticals is that they provide a direct method of comparison for code and library evaluation. However, one drawback of comparing to critical experiments is that they are typically only performed at room temperature (cold conditions). This paper presents not only room temperature criticals, i.e., two sets of B&W criticals: the 1810 Series which covers a wide range of burnable absorbers, and the 1484 Series (which helps verify the accuracy of radial leakage calculations), the DIMPLE criticals S06A, and S06B (which helps validate PWR reflector/baffle calculations), but also results for the Kritz-4 BWR criticals which were performed at both cold and hot conditions, with and without gadolinia, and which help to validate the CASMO5 isothermal temperature coefficient and gadolinium worth.

The radial power profile in a fuel pin is an important piece of information for transient analysis. Based on single fuel region assumption in the U-238 resonance treatment, CASMO5 relies on an empirical radial distribution function for the U-238 resonance integral. The burnup-dependent radial power profile in a typical UO2 fuel pin is examined as a numerical benchmark between CASMO5 and MCNP-5 depletions. As a result, a new set of fitting parameters for the distribution function is found for CASMO5 to better match the MCNP-5 results at high burnups. Running CASMO5 with new fitting parameters verifies the improvements in agreeing with the burnup-dependent MCNP-5 radial power results. However, except the local power profile, there are no noticeable reactivity effects during depletion. The total amount of plutonium production in the fuel is also independent of the U-238 resonance integral distribution function.

This paper details the energy release model implemented in Studsvik’s CASMO5 lattice physics code. By utilizing internally computed neutron capture rates, the capture component of the total recoverable energy is more accurately calculated, and the use of traditionally assumed constant values of energy release per capture (employed in most lattice physics codes) has been eliminated. By using problem-specific capture rates in the energy release model, the impacts of depletion, gadolinia, boron concentrations, MOX compositions, and void fraction are naturally incorporated into the energy release calculation. CASMO5 studies performed with ENDF/B-VII data show that the traditional capture energy release values for standard LWR applications are too large. Also, the implications for energy release in downstream 3-D reactor simulator codes are discussed.

The effects of accurate modeling of neutron scattering in U-238 resonances are analyzed for typical light water reactor (LWR) and next generation nuclear plant (NGNP) lattices. An exact scattering kernel is formulated and implemented in a newly developed Monte Carlo code, MCSD (Monte Carlo Slowing Down), which solves a neutron slowing down in an infinite homogeneous medium and is used to generate resonance integral data used in the CASMO5 lattice physics code. It is shown that the exact scattering kernel increases LWR Doppler coefficients by ~10% relative to the traditional assumption of asymptotic elastic downscatter for U-238 resonances. These resonance modeling improvements are shown to decrease hot full power eigenvalues by ~200 pcm for LWRs and ~450 pcm for NGNPs.

A neutron data library based upon the recently released ENDF/B-VII R0 evaluation has been generated for the LWR advanced lattice physics code CASMO5. This paper details some CASMO5 comparisons with this new data library for the B&W 1810 series of criticals.

Depletion analysis based on a Monte Carlo transport solution is attracting increased interest in recent years. A heavily gadolinium-poisoned BWR fuel assembly is used here as a challenging problem to explore the coupled MCNP-5/ORIGEN-2.2 (via utility program, MCODE-2.2) depletion analysis. A series of MCODE sensitivity runs are performed to investigate the effects of neutron histories, depletion step size, burnup corrector calculation, and a second end-of-step MCNP flux calculation. Eigenvalues and pin fission rates are compared between the converged MCODE solution and the deterministic lattice physics code, CASMO5, result.

CASMO5 is the latest code version in the Studsvik Scandpower series of lattice physics codes for general LWR analysis and has capabilities, features and numerical models not previously available in CASMO4. Among the new capabilities of CASMO5 are a “quadratic gadolinium depletion model” and the ability to perform 2D transport calculations in hundreds of energy groups.

This paper briefly describes CASMO5 capabilities and the role of CASMO5 in generating cross section and discontinuity factor data for SIMULATE5. This paper also describes the new CASMO5 data libraries (586 group neutron and 18 group gamma), and details recent work to validate CASMO5 2D transport calculations for several simple criticals.