CLIP: A CUDA-Accelerated Lattice Boltzmann Framework for Interfacial Phenomena with Application to Liquid Jet Simulations (2505.12205v1)

Abstract: This work introduces CLIP, a CUDA-accelerated phase-field lattice Boltzmann framework for simulating immiscible two-phase flows with high density and viscosity ratios in both two- and three-dimensional domains. By leveraging GPU parallelism, the framework delivers substantial computational speedups, enabling large-scale simulations to be performed efficiently on standard desktop hardware without the need for high-performance computing clusters. It employs the Weighted Multi-Relaxation Time (WMRT) collision operator to enhance numerical stability and improve interface tracking under challenging multiphase conditions. The model is validated through a series of benchmark cases, including capillary wave dynamics, stationary drop tests, two-phase Poiseuille flow, shear-driven interface deformation, and Rayleigh-Taylor instability. It is further applied to simulate liquid jet breakup, capturing the transition from dripping to jetting regimes and identifying a critical Weber number of approximately 2.2. The results closely match experimental observations, offering detailed insights into breakup length, drop size distributions, and flow regime transitions. With its efficiency, accuracy, and scalability, the proposed framework serves as a powerful and accessible tool for investigating complex interfacial phenomena in multiphase flow physics.