A generalized digital computer approach to analyze the loss-of-coolant accident in pressurized water reactors requires a systematic specification of the plant geometric, physical, and topological characteristics and initial conditions. The solution of the problem is hampered by numerical stability and convergence problems which can be remedied by first classifying the problem variables into three categories: 1) numerically-integrated; 2) analytically-integrated; and 3) auxiliary algebraic variables. Second, in view of the occurrence of the acoustic wave phenomenon, the maximum allowable integration time step should be kept much smaller than the subharmonics present in the solution. Another distinctive feature of this study is the treatment of stratified elements, such as the pressurizer or the steam generator secondary. Allowance for mass exchange between the top and bottom control volumes in these elements is made by the introduction of bubble rise and condensate drop velocity concepts. Furthermore, to eliminate unrealistic pressure fluctuations in the ruptured elements at the onset of two-phase blowdown, critical flow models including inertia effects are introduced. To verify the sensitivity of the solution to various two-phase frictional loss correlations, five well-known correlations are reviewed. A comparison of the analytical results against LOFT experimental data demonstrates a good agreement and shows that a more accurate prediction would require the introduction of metastability analysis.