Research On Continuous Energy Monte Carlo Homogenization And Group Constant Generation Based On RMC

With the use of continuous energy point cross-sections and its powerful geometric processing capabilities,Monte Carlo codes can generate more accurate few-group constants for reactor simulations with respect to deterministic codes.The Monte Carlo homogenization is the development trend of reactor physics,and there are still many problems that need further study.In this paper,the group constants are calculated under the two-stage framework.First,the fine group constants are calculated by the track length method and the collision probability method.Then the few group constants are obtained by using the fine group flux spectra and the fine group constants.It can make the direct method and the two-step method completely equivalent to calculate the scattering matrix and the high-order scattering matrix by adjusting the transition probability when being collapsed and avoid the two-stage method to introduce errors.Removing assemblies from the reactor core and replacing the true boundary conditions with reflective boundary conditions will inevitably introduce errors,and the actual calculating method of the group constants cannot guarantee the three principles of conservation.Therefore,the generated group cons tants need to undergo equivalent and correction processing.Firstly,this paper studies the BN correction theory and combines the B1 multigroup equation and the buckling iteration method to obtain the flux weight spectrum of an arbitrary multiplication coefficient system.The core of the theory is to adjust the leakage rate of the calculation object so that the real flux spectrum of the assembly can be obtained with the use of the core multiplication coefficient.What's more important is that the theory is also insensitive to the shape of the assembly.Then,the paper studies generalized equivalence theory and super-homogenization method which are very mature to the deterministic method and applies them to the Monte Carlo homogenization method.Finally,in order to completely solve the error introduced by the approximate processing of the boundary conditions,the theory of full core homogenization has been proposed and preliminary research of the theory has been implemented into the code.In order to use the high-precision continuous energy Monte Carlo homogenization group constants,the DONJON finite-element core program has been successfully coupled with RMC using the Contour algorithm,and the Newton iteration method was proposed to construct the two-step program system of the RMC homogenization calculation-MCNP multigroup Monte Carlo core calculation,and the coupling analysis system of the RMC homogenization calculation-RMC multigroup Monte Carlo function was designed.In this paper,on the basis of continuous energy Monte Carlo random geometry explicit modelling transport and Monte Carlo homogenization group constant generation methods,the random geometry Monte Carlo homogenization and group constant generation theory are studied.Through the above research,the following conclusions are drawn: For the infinite medium weight spectrum,the one-step method and the two-step method which are used to calculate the group constants are completely equivalent,which avoids the introduction of errors in the collapsi ng process when using the two-step method to calculate the few-group constants.In this paper,by studying the B1 correction theory,the equivalent homogenization theory and the whole reactor homogenization theory,the approximate processing of the boundary conditions of the homogenized assembly is completely solved,and the accuracy of the core analysis is greatly optimized.This paper has successfully coupled the RMC homogenization calculation-DONJON FEM core calculation code system,constructed the RMC homogenization calculation-multigroup Monte Carlo core code MCNP and the multigroup function of RMC.Finally,the theory of homogenization and group constant generation of random geometry is studied,which greatly improves the analysis accuracy of stochastic geometry systems.