In order to improve passive safety of Sodium-cooled Fast Reactors (SFR) in case of unprotected transients such as Unprotected Loss of Flow (ULOF) or Unprotected Transient Overpower (UTOP), The French Alternative Energies and Atomic Energy Commission proposed a CADOR concept - a new design of SFR core with enhanced Doppler effect. One of the most important design features is the addition of moderating materials inside fuel assemblies to decrease the average neutron energy by around 40%. The solution leads to roughly three times higher magnitude of Doppler effect due to the increase of resonance neutron population. On the other hand, the softened neutron spectrum changes other core properties. It increases the importance of low-energy neutron scattering and absorption. Moreover, the heterogeneous moderator placement in the assembly may cause an uneven reaction rate distribution and a risk of power peaks not observed in standard SFRs. To demonstrate the safety of CADOR design, it is essential to first evaluate the performance of calculation tools following a Verification, Validation and Uncertainty Quantification (VVUQ) - a process that must be applied to calculation codes and methods to show their reliability. The aforementioned changes in the neutron balance put into question the applicability of standard fast reactor neutronic calculation schemes to the case of CADOR. The purpose of this thesis, therefore, is to establish an accurate neutron transport calculation scheme, in line with VVUQ principles, that takes into account all relevant physical phenomena related to atypical properties of the CADOR core.A two-step calculation scheme of deterministic neutron transport code APOLLO3 was defined as a basis for the analysis. The CADOR cores with two different moderator types, Be and ZrH2, were used. The elements of the scheme and their possible improvements were studied through direct comparison with the reference Monte Carlo code TRIPOLI-4. The systematic biases of numerical models, such as: different spatial homogenization approaches or resonance upscattering treatment, different energy and spatial mesh definitions, were studied with respect to accuracy of multiplication factor, Doppler effect and reaction rates. The most important sources of uncertainties were identified and quantified. Finally, as a first estimation of the sensitivity of the multiphysics calculation scheme, the impact of the uncertainties on simulations UTOP and ULOF transients was evaluated via coupling with MACARENa, a calculation code for transient analysis in SFRs.The results indicate that the accuracy of calculation scheme can be improved by applying exact scattering treatment, notably in case of core with ZrH2 moderator where utilization of simplified scattering kernel leads to underestimation of Doppler effect of up to 5.2 %. With exact scattering treatment the global bias of the calculation scheme of APOLLO3 was estimated at approximately 500 pcm for core with Be moderator and 460 pcm for core with ZrH2 moderator. The biases in case of CADOR are of the same order of magnitude as for conventional SFR designs. By preserving more heterogeneous description of the fissile zone during homogenization process the global bias can be further reduced by 110-280 pcm depending on the studied level of heterogeneity; however this approach has a drawback of significantly higher computational complexity. The sensitivity analysis performed in MACARENa suggests that the uncertainties of neutronic calculations have minor impact on the progression of simulated transients. This work shows that the methods available in APOLLO3 provide a good accuracy of calculation of SFRs, even in case of less conventional designs. The low uncertainties of the calculation scheme indicate robustness of the numerical models used; the calculation scheme provides sufficient accuracy to be applied in fast reactor design and safety studies.