Моделювання еволюції нейтронних спектрів за допомогою Монте-Карло коду GEANT4

Abstract

The knowledge of neutron spectrum is essential in many di erent physical problems. First of all, in nuclear reactor physics it is required for the improvement and development of the new generation reactors, as well as for the study of various emergency states. It is especially interesting for the so-called wave nuclear reactors (in which the nuclear ssion reactions take place in a narrow layer which slowly and gradually propagates along the reactor core), which is a very promising technology being actively developed around the world. In order to study the operation modes, and even the very possibility of their existence, it is necessary to have a detailed knowledge of the form of the neutron spectrum in them. In the tasks involving a substantial change in the nuclear fuel parameters with time (e.g. temperature, chemical and isotopic composition etc.), the neutron spectrum can also change drastically, which may lead to some exotic phenomena such as the so-called blow-up modes, when the parameters may grow to in nity in a nite time. In ssile materials such phenomena are known to depend on the nuclear reaction cross-sections together with the neutron spectrum. In addition, the form of neutron spectrum becomes more important in medical applications, especially considering the rising interest to such methods as boron neutron capture therapy (BNCT). Here, using the GEANT4 Monte Carlo code, we develop software able to simulate the process of neutron moderation in various media and track the evolution of the neutron spectrum with time. As an example, we show the calculation results for four di erent media hydrogen, carbon, aluminum and iron, and two di erent types of neutron source a monoenergetic source and the one having a ssion spectrum. We analyze the behavior of a number of statistical parameters of the neutron spectrum, particularly, the entropy. This lets us single out several stages of the neutron spectrum formation, which manifest themselves in all considered cases. We estimate the characteristic time of each of these stages onset in di erent media.