Abstract
We present the results of the simulation of two-phase CO2 flows at low-Mach number, obtained through a pressure-based Baer-Nunziato type model. The underlying full non-equilibrium model enables the description of each phase with its own thermodynamic model, so it circumvents the requirement of the definition of the speed of sound of the vapor-liquid mixture. The primitive formulation, combined with a special pressure scaling to correctly capture the behavior in the zero-Mach limit, is well-suited to model weakly compressible flows, and makes easier the use of arbitrary thermodynamic models. At the interfaces, the phasic velocity and pressure are driven toward the equilibrium by means of relaxation processes, whose velocities are controlled by user-defined parameters. The set of seven partial differential equations describing the flow evolution is discretized through a finite-volume scheme in space and an hybrid implicit-explicit time discretization, to avoid the stringent time step limitation imposed by the acoustics. We compare the results of a shock-tube problem, initially containing saturated CO2, obtained according to the stiffened gas model and to the Peng-Robinson equation of state.
1 INTRODUCTION Among the technologies able to contrast the global warning, carbon capture and storage (CCS) is regarded as a crucial and effective approach. Consequently, the numerical investigation of carbon dioxide (CO2) flows under the different conditions we can encounter within the CCS framework is becoming more and more important. In this work, we focus in particular in unsteady weakly compressible twophase flows. Such kind of flows may occur in the transport pipelines, because of fluctuating in the CO2 supply, impurities, or during transient events, such as start-up, shut-down or depressurization [1]. From a numerical point of view, these flows present different challenging aspects. First of all, the weak
compressibility—that is the condition where the flow velocity is considerably smaller than the speed of sound but compressibility effects cannot be neglected—makes inefficient and inaccurate the standard compressible solvers. Second, the multitude of spatial scales and the presence of dynamic interfaces that separate the different phases call for an effective modeling that avoids the full resolution of the flow field but takes into consideration the relevant flow features. Third, a flexible implementation of the thermodynamic modeling for the CO2 is recommended to be able to customize it according to the different applications.
Re, Barbara