Reduced Physics Modeling of Two-Phase Flow through High-Density Cooling Structures

Increasing power density in high performance microprocessors has driven research into high performance and energy efficient methods of cooling including both single-phase and two-phase liquid cooling. These approaches employ coldplates with complex high-density cooling structures such as fine-pitch microstructures, linked-pin-fins, metal foams, etc., to achieve high heat transfer coefficients. The development of coldplates, especially for two-phase cooling, with such complex cooling structures requires computationally efficient high fidelity thermal models to evaluate the device and system performance under different cooling configurations and operating environments. In this work, a novel reduced physics model based on porous media approximation was developed with homogeneous equilibrium model assumptions for the two-phase flow in the coolant flow domain. For an exemplar complex cooling structure, this reduced physics model produced pressure drop values close to those obtained from full-physics simulations while utilizing significantly less computational nodes and >100X less computational time.