Nuclear Science and Engineering / Volume 183 / Number 2 / June 2016 / Pages 229-246
Technical Paper / dx.doi.org/10.13182/NSE15-100
The continuous-energy Monte Carlo (MC) method is gaining attention not only for nuclear reactor statics but also for transient analysis, as computing power increases with the use of massive parallel computers. This paper presents a practical and accurate MC transient analysis method for heterogeneous, continuous-energy reactor transient problems, based on the predictor-corrector quasi-static (PCQS) method. The transient fixed-source problem of the PCQS method is solved by MC calculation with fission source iteration, where the partial current-based coarse-mesh finite difference (p-CMFD) method is used both to accelerate the convergence of the fission source distributions and to diagnose whether the fission source iteration diverges because of too large a macro-time-step size used for a positive reactivity insertion. To improve the convergence of the fission source iteration, exponential transformation is also applied. In addition, the variances of MC tallies can be reduced by increasing the number of active fission source iterations. For method and code verification, the PCQS method for the MC calculation with fission source iteration is compared with the implicit Euler method for a method-of-characteristics calculation on a two-dimensional TWIGL problem. For both multigroup energy and continuous-energy three-dimensional test problems, the proposed method efficiently reduces computing time with a large macro-time-step size, while the accuracy of the solutions is maintained, compared with those calculated with smaller macro-time-step sizes.