A new Monte Carlo method for neutron noise calculations

International audienceA new Monte Carlo algorithm which solves transport equations of the neutron noise in the frequency domain has been developed. Neutron noise equations, which are obtained by assuming small perturbations of macroscopic cross-sections around a steady state in the neutron field and then by Fourier transform in the frequency domain, are usually solved by analytical methods or by diffusion theory. It is only recently that an original stochastic method was proposed in the literature by using particles with complex-valued weights and with a weight cancellation technique, the binning procedure, in order to cancel the positive and negative values of the real and imaginary parts of weights in a large number of small regions. The new stochastic method presented here does not use any weight cancellation technique (instead we remove the implicit capture at low and high frequencies) and it is based on a real total cross-section and not on a complex total cross-section. The two Monte Carlo methods are compared with the deterministic methods (diffusion and transport) in case of a heterogenous one-dimensional system for several frequencies. We conclude that the new stochastic method is faster than the method developed in the literature, is more robust in high frequencies and is easier to implement