Digital twin as a tool for evaluating and optimizing flow behavior in encapsulating proccesses

The capillary underfill process is a widely studied encapsulation method, whereas numerical simulation has been mainly used to analyze flow dynamics and distribution of underfill fluids in bump arrays with uniform gap geometry. In contrast., the present work focuses on packages in the context of power electronics with larger and variable gap heights along with complex, irregular gap geometries. Comsol and Star-CCM+ solvers are used for the case of capillary two-phase flow to evaluate the performance of interface-capturing methods such as Level-Set (LS) and Volume of Fluid (VOF). Parameter variations of the High-Resolution Interface Capturing (HRIC) scheme were further analyzed to obtain a robust and computationally efficient VOF model for accurate prediction of filling time and air entrapment. The effects of various conditions on underfill flow, including material properties and geometric features, are investigated. A novel approach for experimental test modules is presented to verify the numerical model, allowing for flow observation while still reflecting realistic wetting properties. The normalized difference of filled gap volume averaged over time serves as a benchmark for comparative considerations. The results are revealing good agreement between simulated and experimental flow behavior in terms of flow propagation and flow velocity (1.48 - 3.04 %). Reasonable agreement (2.55 - 3.34 %) is also achieved by transferring the simulation model to geometries with estimated gap heights and using time scaling. The techniques presented thus provide guidance for CFD model optimization and flow assessment in terms of experimental simulation validation.