Estudio de los Problemas Numéricos en un Modelo de Evaporación en Tubos Verticales

Abstract

This paper addresses the numerical problems appearing during the simulation of the evaporation of a fluid inside vertical tubes. The significative differences existing between magnitudes of vapor and liquid phisical properties seem to be the principal reason for the ill-condition, which frequently arises when modeling detailed two-phase flow systems. Round-off and truncation errors collaborate to increase the mismatch between the desired response and the result. Two completly different numerical strategies are suggested here to solve the fluid evaporation model, 1) the explicit derivates, obtained after a main matrix inversion are integrated using the semi-implicit Runge-Kutta method with variable integration interval and, 2) a set of nonlinear equations, obtained by a two point weighted discretization, is solved through an implicit
optimization.