| Title: |
Reconstruction of few-group homogenized cross section by kernel method and active learning |
| Authors: |
Szames, Ea.; Tomatis, D.; Ammar, K.; Martinez, Jm. |
| Contributors: |
CEA-Direction des Energies (ex-Direction de l'Energie Nucléaire) (CEA-DES (ex-DEN)); Commissariat à l'énergie atomique et aux énergies alternatives (CEA) |
| Source: |
M and C 2019 ; https://hal.science/hal-02411076 ; M and C 2019, Aug 2019, Portland, United States |
| Publisher Information: |
CCSD |
| Publication Year: |
2019 |
| Collection: |
HAL-CEA (Commissariat à l'énergie atomique et aux énergies alternatives) |
| Subject Terms: |
Few-Group Cross Sections; Kernel Methods; Active Learning; Spline Interpolation; [PHYS.NEXP]Physics [physics]/Nuclear Experiment [nucl-ex]; [PHYS.NUCL]Physics [physics]/Nuclear Theory [nucl-th] |
| Subject Geographic: |
Portland; United States |
| Description: |
International audience ; This work deals with the approximation of homogenized few-groups cross sections by kernel methods. Different kernels types are used in conjunction with pool active learning to optimize the cross section's support. They are compared to multi-variate splines and multi-linear interpolation, similar to industry. A standard PWR fuel assembly is provided by the OECD-NEA Burn-up Credit Criticality Benchmark, Phase-IID, to evaluate their performances . |
| Document Type: |
conference object |
| Language: |
English |
| Availability: |
https://hal.science/hal-02411076; https://hal.science/hal-02411076v1/document; https://hal.science/hal-02411076v1/file/201800004324.pdf |
| Rights: |
info:eu-repo/semantics/OpenAccess |
| Accession Number: |
edsbas.DDB2BFAC |
| Database: |
BASE |