For the sake of verification of the own methods (I and II), proposed in the first part of this work [1], the pertinent algorithms have been worked out and on their basis - the computer programs. These programs have been employed for the calculation of mass transport in the ternary gas mixture: n-propane, n-undecane, hydrogen and the ternary liquid mixture: acetone, benzene and carbon tetrachloride. The obtained results have been compared with those based on the methods presented in [8, 9] and [10]. For the gas phase the algorithm of method I simplifies, since the matrix [GAMMA]n-1 is equal to the identity matrix [I]n-1 and the mass transfer coefficients k(ij) are constant along the mass transport path. On the other hand, the algorithm of method II becomes identical with that of KRISHNA and STANDART [8]. The computations have been performed for the set of boundary conditions listed in Tab. 2. As it follows from Tab. 4, the mean error of the own method proved to be in all cases smaller than for the method [10]. It came also out that for appropriately selected "film partition" coefficient L the error values of the own method are close to 0. The analysis of the results for the liquid phase, obtained - similarly as in the case of gas phase - at the iteration parameter h equal to 0.5 and 1.0 leads to the conclusion (Tab. 7) that methods I and II yield in this case almost identical values of molar fluxes but significantly different from those obtained by the methods [9] and [10], the latter results being almost identical, too. The computations, both for the gas phase and for the liquid one, have shown the convergence of the proposed algorithms. This convergence has been compared with that of the hitherto known methods (Tabs. 3 and 6). The results presented in Tab. 7 do not give a possibility of the estimation of the accuracy of suggested methods, since the method of KRISHNA [9] is for this case an approximate method. Such an estimation becomes possible only when based on the experimental data of the liquid-phase mass transport or in a more complex system, in which the liquid-phase resistance appears as a component. In this connection a comparison has been made of the results of calculations of mass transport for a real process of condensation of a binary vapour in the presence of an inert gas with the experimental data of MODINE [15]. In this system (acetone-benzene-hydrogen or helium) significant cross-effects appear and the liquid-phase resistance cannot be neglected. The exemplary results of mass fluxes computations are presented in Figs. 1-4. The analysis of computational results (Tab. 8) for the experimental data Of MODINE [15] entitles to state that the own methods (I and 11) enable one to determine the values of mass fluxes with a significantly better accuracy than the methods [9] and [10]. It is true for these cases for which a considerable variability of the elements of matrices [PHI]n-1 and [PSI]n-1 along the process path occurs. This variability is caused not only by the concentration difference but also by the variability of binary mass transfer coefficients in a multicomponent system, the values of these coefficients depending both on concentration and temperature. The proposed own methods are easy for programming, since they consist mainly in multiplication, addition, subtraction and inversion of real matrices. They do not require the determination of eigenvalues of matrices.
