New perturbation and sensitivity capabilities in Tripoli-4®. (November 2018)

• Record Type: Journal Article
• Authors: Terranova, Nicholas; Mancusi, Davide; Zoia, Andrea
• Is Part Of: Annals of nuclear energy. Volume 121(2018)
• Page Start: 335
• Page End: 349
• Publication Date: 2018-11
• Subjects: Perturbation theory -- Sensitivity analysis -- IFP -- Differential operator sampling -- Correlated sampling -- Monte Carlo -- Tripoli-4® -- MCNP6 -- COMAC -- SCALE; Nuclear energy -- Periodicals; Nuclear engineering -- Periodicals; 621.4805
• DOI: 10.1016/j.anucene.2018.07.031
• Languages: English
• ISSNs: 0306-4549

Abstract

Highlights: We investigate IFP-weighted perturbations and sensitivities with Tripoli -4. We compare our results with those of other codes (MCNP and KENO). We compare IFP-weighted scores with correlated sampling and Taylor expansion. Abstract: Estimating the changes in a nuclear system due to perturbations in the input nuclear data by separate Monte Carlo calculations might be extremely cumbersome for reactor applications. The Iterated Fission Probability (IFP) method has recently paved the way for the application of first-order standard perturbation theory in continuous-energy Monte Carlo codes. In this work, we detail the reactivity perturbation and k -eigenvalue sensitivity analysis capabilities of the Monte Carlo code Tripoli-4 ®. Simulation results obtained by using the newly implemented IFP algorithm of Tripoli-4 ® are compared to findings coming from other Monte Carlo methods (such as the differential operator and the correlated sampling) and codes (such as MCNP6 and KENO ). For this purpose, we select some benchmark configurations (Godiva, Stacy, Jezebel, Flattop and a fuel lattice) and we test some of the most common perturbation and sensitivity methods currently available in production codes. Their respective advantages and drawbacks are analyzed, and possible future improvements are suggested. Our main finding is that Tripoli-4 ® produces very similar results to MCNP6 when the same techniques are used. Uncertainty propagation based on the obtained sensitivity profiles and on the COMAC nuclear data covariance matrices is finally discussed.