Application of the COBRAG subchannel code to fuel design

Global Nuclear Fuel (GNF) has developed a subchannel analysis code, COBRAG, to predict critical power, bundle pressure drop and void/flow distributions in Boiling Water Reactor (BWR) fuel bundles. COBRAG solves a set of two-phase, multi-field, time dependent governing equations along with various constitutive models to accurately predict local conditions at the subchannel-level. COBRAG was applied to optimize the performance of an advanced fuel design, GNF3, which was developed to deliver more power and reduced overall fuel cycle costs relative to GNF2, while improving operational reliability and flexibility. During GNF3 development, various lattice configurations were evaluated for nuclear and thermal-hydraulic (TH) performance. Key TH parameters such as flow/void distribution and critical power were evaluated using COBRAG and a final GNF3 lattice configuration was selected to optimize the bundle’s nuclear and TH performance. The projected improvement in the TH performance was confirmed by full-scale critical power testing. COBRAG was also utilized as a higher order benchmark tool of the GE critical quality versus boiling length (GEXL) critical power (empirical) correlation that was developed for GNF3. For this purpose, COBRAG was validated against a wide range of benchmark data, including local void distributions, bundle pressure drops, and both steady-state and transient critical power data. Comparisons of the COBRAG predictions against data shows excellent agreement. COBRAG was then used as the reference solution to assess the accuracy of GEXL where data is not readily available, such as bundle configurations with untested (e.g., double-humped) axial power shapes, partially controlled bundles, or the change in critical power ratio during transients with simulated nuclear feedback. This paper presents the use of COBRAG for GNF3 fuel development and COBRAG validation against various benchmark data.