Methods for modeling the coagulation in an aerosol code developed for calculating the behavior of multicomponent aerosols are examined. A comparative analysis is made and the methods are tested on model problems for conditions characterizing the formation of aerosols of fi ssion products in the fi rst loop of VVER during an accident with depressurization of fuel elements.The primary method for performing a safety analysis of NPP and radiation contamination of the environment as a result of a beyond design basis accident is modeling, using special integral codes, the behavior of the reactor at all stages of the accident. Accurate modeling of the formation and transport of aerosols in the fi rst loop is a necessary condition for making a quantitative assessment of the radiation contamination of the rooms in the containment shell [1,2]. Aerosols are formed as a result of nucleation of vapors of fi ssion products, structural materials, and possibly uranium oxides of the fuel matrix. Particles of the corrosion and aerosols of nonvolatile fi ssion products can also be a source of heterogeneous centers. The radius of the aerosol particles in the fi rst loop during an accident ranges from 10 -9 to 10 -5 m [2-4].A code is being developed for modeling the behavior of multicomponent aerosols. A spatial cell with a uniform distribution of the gas and aerosol components and the thermodynamic parameters of the medium is studied. The approximation with composite particles whose composition is the average for all particles of a given size is being examined. In addition, the composition can be different for particles of different size. We note that the direct modeling of the dynamics of multicomponent aerosols presupposes that the dependence of the phase density of the particles on the mass of the components comprising a particle is taken into account. However, this approach is currently not implemented in the integral codes because the computational time is excessive.A universal method of modeling the dynamics of aerosol particles is the method of fractions. In this method, the particle size spectrum, for example, of a volume, is subdivided into individual fractions with fi xed or moving boundaries. As a result, the general integro-differential equation governing the dynamics of aerosols becomes a system of nonlinear fi rst-order differential equations for the concentration or mass of the particles in each fraction [2,5]. The main drawback of this method is induced numerical diffusion along the size axis as a result of the artifi cial redistribution of a defi nite fraction of the particles formed after collisions into a fraction with sizes larger than the actual size. An obvious method of minimizing the effect is to increase the number fractions considered. However, in this case, the computing time increases accordingly as the squared size of the system of equations. For this reason, it is important to use effi cient numerical algorithms to minimize the computing time while maintaining acceptable accuracy.In the present work, tw...