Accuracy evaluation of pin exposure calculations in current LWR core design codes

Abstract

The deregulated utility environment and better utilization of fuel assemblies in nuclear power plants has allowed designers to burn fuel assemblies to maximum allowable exposures. Any uncertainties, associated with the technical approach and numerical methods used to perform pin exposure calculations may cause either peak power exposure to exceed the Nuclear Regulatory Commission (NRC) exposure limit or lead to excessive conservatism and thus inefficient fuel utilization. In this work, a Monte Carlo based coupled depletion code (MCNP5/ORIGEN-S) is utilized to provide reference solutions in order to assess the accuracy of pin power and pin exposure reconstruction methods in the current commercial and licensed three-dimensional (3D) nodal Light Water Reactor (LWR) core design codes. The developed at the Pennsylvania State University (PSU) MCNP5/ORIGEN-S coupled depletion code system was validated using measured data from the PSU TRIGA research reactor critical experiments. A number of test cases (modeling benchmarks) representative of LWRs were developed starting from the least complex model towards more complicated and more realistic models. The accuracy evaluation of the pin reconstruction methods was performed by using the CASMO-4 and SIMULATE-3 codes as the representative of current commercial LWR core design systems. Two-dimensional (2D) transport calculations with the TRITON module from the SCALE5 package were employed to produce the spectrum averaged cross-section libraries as a function of burnup for ORIGEN-S calculations. The burnup dependent cross-section libraries are specifically generated for each lattice configuration type. For the MCNP5 calculations continuous cross-section libraries for different isotopes at hot operating temperatures are generated and subsequently utilized. Realistic lattice configurations of the GE13 BWR fuel assemblies (unrodded and rodded) depleted under operating conditions were studied in this research because of their heterogeneous nature. The 2D model test cases are constructed prior to 3D model test cases to investigate in a consistent manner the approximations involved in pin reconstruction methods of the current commercial LWR core design codes. Discrepancies more than 11% were observed between the pin-wise power and exposure data calculated by the two different methods. The statistical uncertainties in the Monte Carlo calculations were analyzed and addressed. The statistical uncertainties in the MCNP5 results remain essentially less than 2% throughout the depletion history and there was no noticeable propagation with burnup.