Computer modeling of neutron-physical and thermohydraulic processes has been used extensively in the analysis of the Chernobyl accident. In most works the first phase of the accident (up to the moment of destruction of the fuel) is studied. These studies have revealed serious deficiencies of RBMK reactors that resulted in a nonroutine state with catastrophic consequences during operation: a large positive steam reactivity coefficient and a defect in the construction of the control rods, specifically, the possibility of a reactivity increase as a result of displacement of water to the core bottom when the safety and control rods are inserted in the process of stopping the reactor. The important role of the last factor has been noted in many investigations [1][2][3]. In some works [4-6] a large power burst was obtained during modeling, neglecting the effect of the rods, as a result of other external actions, for example, cavitation. Calculations performed using three-dimensional neutron-thermohydraulic programs with maximum use of accessible initial information have shown that both factors are important [1, 7]. If one of them is artificially excluded, it is impossible to obtain the power burst which explains the explosion.The hypothesis of a second neutron power burst has not been adequately proved. Starting with the first report [8], it has been shown in some works that a second, more powerful burst occurred -1 sec after the first burst. We note that the existence of two bursts is characteristic, as a rule, for calculations which employ the simplest (point, one-dimensional) models [4, 6, 9, 10].The second phase of the accident, associated with the destruction of the fuel elements, has been studied much less. Analysis of the destructive forces on the basis of experiments on impulsive irradiation of fuel elements at NSRR is presented in [l 1]. A scenario of the accident from the first power burst up to burning of the graphite is described in [12]. The key point is the dispersing of the fuel, i.e., rapid destruction of the fuel into small particles. The effect of destructive processes on the reactivity and the neutron power has not been considered.The effect of a change in the fuel geometry on the reactivity was studied in [6, 13, i4]. It was shown that under certain conditions destruction of the fuel can result in an increase of reactivity. One variant is rapid cooling of the dispersed fuel in water. As a result of the Doppler effect, the reactivity associated with boiling of the coolant is added to the positive reactivity. Moreover, for a dried channel with dispersed fuel, a decrease in the amount of fuel to 60-65 % of the initial amount results in a positive reactivity exceeding 1/3 [6], where/3 is the effective fraction of delayed neutrons.Computational analysis of the accident at the fuel destruction stage requires a complicated model which includes a three-dimensional description of the neutron-physical, thermohydraulic, thermomechanical, and chemical processes in the core taking into account their couplings. A...