3D actual microstructure-based modeling of non-isothermal infiltration behavior and void formation in liquid composite molding

2021 
Abstract In the liquid composite molding (LCM) non-isothermal infiltration process, the effect of the complex interaction between viscous pressure and capillary pressure on the fluid invasion morphology and void formation mechanism is not well understood. Based on high-resolution, three-dimensional (3D) computed microtomography (µCT) images of porous silicon carbide particulate-reinforced composite (SiCp) preforms, a pore-scale, non-isothermal two-phase flow numerical model was implemented to describe the flow displacement pattern that evolves with the decrease in displacement velocity during the infiltration process. The simulation approach was performed by coupling Cahn–Hilliard phase field and heat equations using a robust finite element solver. The numerical results indicate that (i) the displacement pattern experiences a transition from stable displacement to capillary fingering, which mainly depends on the transition of complex yet intriguing pore-scale events responsible for local meniscus dynamics, from viscous self-correcting smoothing transition to the noncooperative Haines jump; (ii) the void trapping stability is higher in the capillary fingering pattern, and it is difficult for the macrovoids generated by the bypass trapping mechanism to migrate; (iii) there is a critical capillary number (Ca) value, and as long as the Ca of the displacement front during infiltration process is not lower than this value, the local macroscopic void concentration can be avoided; (iv) the non-isothermal effect in the LCM simulation cannot be ignored, where the change in two-phase physical properties caused by the evolution of the temperature field affects the local pore-scale flow behavior, i.e., the disappearance of the partial interface pinning phenomenon.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    36
    References
    1
    Citations
    NaN
    KQI
    []