The results of Phase-1 neutronics benchmark analysis of CEFR start-up tests was presented in the 2nd Research Coordination Meeting (RCM-2), held at Beijing, China during 28th October to 1st November 2019. Participation of this benchmark exercise has provided us a wide international forum to inter-compare the method, computer codes and cross section data employed at IGCAR for the physics design calculations of sodium cooled fast reactors by keeping the experimental data as the reference. The neutronics code systems used at IGCAR are FARCOB (deterministic) and OpenMC (Monte Carlo). In general, our simulation results are agreeing well for most of the work packages. However, few discrepancies observed for some simulations which are being resolved by detailed analysis. In this paper, the methods of neutronics simulation for 4 work packages (WP-1, 2, 3 and 4) and inter-comparison of simulated results (Phase-2) with the experimental value are briefly discussed.
The FARCOB code solves finite difference neutron diffusion theory in HEX-3D geometry using 26-group self-shielded cross sections from ABBN-93 cross section library. The neutron transport equation is solved in 2-D cylindrical model by first flight collision probability method by using COHINT code for homogenizing absorber rod cross sections. Monte Carlo calculations using OpenMC code uses realistic whole core pin-wise model of all sub-assemblies. Multiple nuclear data libraries, viz. ENDF/B VIII.0, JEFF 3.3, JENDL-4, ROSFOND-2010, CENDL-3 and TENDL-2017, have been processed using NJOY-21 and used for this analysis.
For the critical core, FARCOB under predicts the core reactivity by ~375 pcm. The criticality is predicted within 100 pcm with JEFF-3.3 by OpenMC. The spread in criticality prediction by various ENDF-6 libraries is studied using OpenMC. Absorber rod worth is predicted within 1-2 % of the measured value by these codes. The maximum deviation of temperature coefficient of reactivity with measured value is 10% and 5 % by FARCOB and OpenMC codes respectively. It is to be noted that statistical error of 0.04 pcm/K is achieved in OpenMC estimation by performing simulations for 1010 particles. Sodium void measurements using a special assembly was also simulated as exactly as possible and the estimated results are found to be within 10 % of measured values. Results of optional calculations for various reactivity coefficients are also discussed.
|Speaker's email email@example.com|
|Affiliation/Organization||Indira Gandhi Centre for Atomic Research, Kalpakkam, India|