Verification of MPACT:
The Michigan Parallel Characteristics Transport Code
Benjamin Collins, Brendan Kochunas, Daniel Jabbay, Thomas Downar, William Martin
Department of Nuclear Engineering and Radiological Sciences
University of Michigan
Oak Ridge National Laboroatory
MPACT (Michigan PArallel Characteristics Transport Code) is a new reactor analysis tool being developed at the University of Michigan as an advanced pin-resolved transport capability within VERA (Virtual Environment for Reactor Analysis). VERA is the end-user reactor simulation tool being developed by the Consortium for the Advanced Simulation of Light Water Reactors (CASL). The focus area (FA) for Radiation Transport Methods (RTM) within CASL relies on MPACT to provide capabilities for 3-D method of characteristics (MOC), 2-D/1-D transport, time-dependent transport, and a traditional 2-D lattice physics capability. One of the major computational pieces in MPACT is the 2D MOC solver which is designed to provide an efficient, accurate, and robust flux solution over a broad range of reactor operating conditions. The VERA Core Physics Progression Problems were used to compare the accuracy of both the 2D transport solver and also the cross-section treatments.
The VERA Core Physics Benchmark Progression Problems were designed by Andrew Godrey and Scott Palmtag [1,2] of AMA and provide a progressive suite of problems to assess the performance of CASL codes as they became available. Of particular interest for evaluating the ability of MPACT to model the heterogeneity observed in reactor cores are several of the “colorset” cases in Problem 4 of the Benchmark suite. The cases have an array of 3x3 assemblies as shown in Figure 1. The center assembly has a control rod inserted for the first case and unrodded for the second. In both cases, the assemblies adjacent to the control rod have 20 Pyrex rods and a higher enrichment.
Figure 1 VERA Color-sets: Rodded Color-set
As shown in Table 1, MPACT shows very good agreement to the reference Monte Carlo solution for both the eigenvalue and pin powers of the unrodded and rodded cases. The results of the first several VERA Core Physics Benchmark problems were also performed and show comparable accuracy. In general the results provide confidence that the 2-D capability in MPACT has been implemented successfully, and that comparable accuracy can now be expected of the 3-D capability in MPACT. Along with 3D MOC capability, an initial implementation of the 2D-1D method has been implemented to provide a computationally faster alternative to 3D MOC. Results of the 3D Benchmarking of MPACT using the VERA Benchmark progression will be reported in a later CASL report.
1. Godfrey, A.T., "VERA Core Physics Benchmark Progression Problem Specifications", CASL-U-2012-0131-001, CASL, 12/21/2012.
2. Godfrey, A.T., et. al., “Analysis of Two-Dimensional Lattice Physics Verification Problems with MPACT", CASL-U-2012-0172-000, CASL, 12/21/2012.