MCODE, Version 2.2: An MCNP-ORIGEN DEpletion Program

Abstract

MCODE Version 2.2 is a linkage program, which combines the continuous-energy Monte Carlo code, MCNP-4C, and the one-group depletion code, ORIGEN2, to perform burnup calculations for nuclear fission reactor systems. MCNP is used as the advanced physics modeling tool providing the neutron flux solution and detailed reaction rates in the pre-defined spatial burnup zones. ORIGEN, in turn, carries out multi-nuclide depletion calculations in each region and updates the corresponding material composition in the MCNP model. The MCNP/ORIGEN coupling follows the predictor-corrector approach. During a burnup timestep, end-of-timestep material compositions are first predicted based on the flux solution at the beginning-of-timestep. Using the predicted end-of-timestep material compositions, an MCNP run is performed to compute the neutron flux and detailed reaction rates, which are then used in a corrector burnup step. The final end-oftimestep material compositions are obtained as the average value of the results from the predictor and corrector steps. As a stand-alone code written in ANSI C, MCODE-2.2 is portable between Windows personal computers (PC’s) and UNIX/Linux machines. There are three utility programs in MCODE-2.2: (1) preproc to pre-process MCNP/ORIGEN libraries; (2) mcode as the console to run steady-state burnup/decay calculations; and (3) mcodeout to collect results from scattered data files under temporary directory and produce a detailed output. Further, there is an auxiliary program called mcnpxs, which is for the purpose of preparing a nuclide summary table of continuous energy MCNP cross section libraries. The routine usage of MCODE-2.2 only requires a tandem running of the three utility codes. The auxiliary code, mcnpxs, is intended to help users during the code installation/setup. Compared to other similar linkage codes, MCODE-2.2 emphasizes functionality, versatility and usability. Several features of the code follow: (1) The execution of MCNP and ORIGEN is in an automatic fashion. (2) All standard nuclear reaction types in ORIGEN2 are considered: capture, fission, (n,2n), (n,3n), (n,p), and (n,α). Therefore, both the nuclear fuel depletion and material irradiation/activation (e.g., boron-10 irradiation) can be handled. (3) A power history can be specified, i.e., power level at each timestep. The default depletion option is constant power depletion. Meanwhile, an iterative robust flux depletion scheme is available. In addition, decay calculations are also possible. (4) With appropriate ORIGEN one-group cross section libraries, users can rely on MCODE- 2.2 to automatically select important nuclides based on absorption ranking from ORIGEN isotope reservoir for MCNP calculations. (5) The enhanced predictor-corrector approach (consistent with CASMO-4) increases the accuracy with negligible computational cost increase. From the user’s point of view, MCODE-2.2 is an extension of normal MCNP criticality (kcode) calculations. The MCNP input inherits the MCODE-2.2 input in the form of a fourth paragraph (added at the end of the MCNP input deck) containing the burnup-related data and MCNP/ORIGEN calculation controls. A user-supplied equilibrium MCNP source file can also be provided, which might save CPU time by reducing the number of initial MCNP inactive cycles. iii The Monte Carlo burnup code has some unique characteristics, one of which is that all results are in nature stochastic. The statistical uncertainty passing through burnup calculations is one concern, which is believed by some people as the weakness or even indication of the Monte Carlo limitations to perform burnup calculations. Using a multiregion Gd-poisoned BWR 8×8 assembly depletion problem, it is shown that the random statistical uncertainties are benign and cancel each other with the burnup. In addition, a single PWR unit cell benchmark problem is documented. Comparison of results against CASMO-4 yields satisfactory agreement.