A multiscale thermo-fluid computational model for a two-phase cooling system

Abstract

In this paper, we describe a mathematical model and a numerical simulation method for the condenser component of a novel two-phase thermosiphon cooling system for power electronics applications. The condenser consists of a set of roll-bonded vertically mounted fins among which air flows by either natural or forced convection. In order to deepen the understanding of the mechanisms that determine the performance of the condenser and to facilitate the further optimization of its industrial design, a multiscale approach is developed to reduce as much as possible the complexity of the simulation code while maintaining reasonable predictive accuracy. To this end, heat diffusion in the fins and its convective transport in air are modeled as 2D processes while the flow of the two-phase coolant within the fins is modeled as a 1D network of pipes. For the numerical solution of the resulting equations, a Dual Mixed-Finite Volume scheme with Exponential Fitting stabilization is used for 2D heat diffusion and convection while a Primal Mixed Finite Element discretization method with upwind stabilization is used for the 1D coolant flow. The mathematical model and the numerical method are validated through extensive simulations of realistic device structures which prove to be in excellent agreement with available experimental data

Blood flow mechanics and oxygen transport and delivery in the retinal microcirculation: multiscale mathematical modeling and numerical simulation

“…Numerical techniques are used to obtain approximate solutions of the system. Lowest order finite elements in primal-mixed form on the network curvilinear axial coordinate are used to discretize the PDE model for VTN-blood and O 2 (Sacco et al 2014). The PDE system for O 2 , in the physiological range of parameters, is strongly dominated by convection, and this leads to a large local Péclet number if the spatial discretization parameter is not extremely small.…”

A Poroelastic Mixture Model of Mechanobiological Processes in Tissue Engineering. Part II: Numerical Simulations

“…9), because in the present model the variable p is not a Lagrange multiplier (as in the Stokes system), rather, it is the solution of the elliptic Darcy problem (1b)-(1c). In the case of the ADR equation we employ for the approximation of the concentration and of the cellular volume fractions the primal-mixed finite element discretization scheme with exponential fitting stabilization proposed and investigated in [26]. This choice is taken because it ensures that the computed numerical solutions satisfy a strict positivity property even in the case of a strongly advective regime.…”

Development of computational models for microtesla-level magnetic brain scanning: a novel avenue for device development

“…As such, computational models in neuroscience have gained a great deal of traction because they allow rapid prototyping and testing without the need for live human subjects, thereby reducing the risk of patient harm while advancing the device’s iterative designing process [ 1 , 2 ]. In addition, computational models allow for theoretical testing of completely novel devices, highlighting yet unexplored avenues for device production [ 3 , 4 ].…”