Subaerial landslide-induced waves investigated with an adaptively mesh refined multiphase granular flow model

Abstract

Impulsive waves generated by subaerial landslides pose a significant threat to coastal or enclosed basin environments. However, simulating the intricate mechanisms involved is challenging due to the multiphase nature of the process, involving air, water, and granular materials interactions. While multiphase computational fluid dynamics (CFD) models offer realistic physics for improved simulations of landslide-induced waves, their high computational cost restricts their use to small-scale laboratory cases or domains. This study presents a multi-variable adaptive mesh refinement (AMR) approach integrated into a multiphase-CFD granular flow model to simulate subaerial landslide-induced wave phenomena. The AMR integration achieves dynamically evolving mesh resolutions in key areas of interest, significantly reducing the total cell count and mitigating computational overhead. The model’s performance is assessed across three laboratory-scale scenarios varying in landslide masses, particle sizes, water depths, and domain sizes. Results demonstrate that AMR maintains static-mesh model accuracy while improving computational efficiency, particularly in high cell count scenarios. Key findings highlight the AMR-enhanced model’s ability to capture both landslide and wave dynamics, showing grid-independence behavior and substantial reduction in computational time. The study emphasizes selecting appropriate AMR parameters, such as the refinement interval, to balance model accuracy and computational efficiency. Additionally, the detailed analysis of landslide dynamics reveals critical influences on wave generation, emphasizing the role of landslide deformation and water penetration in the leading and secondary wave characteristics. Several limitations and computational issues arising from AMR implementation are identified, with recommendations for future improvements. Overall, this study provides valuable insights into the potential of AMR-enhanced multiphase-CFD models for accurately and efficiently simulating landslide-induced waves, offering significant implications for coastal engineering applications.