Tsunami debris motion and loads in a scaled port setting: Comparative analysis of three state-of-the-art numerical methods against experiments
Justin Bonus,Felix Spröer,Andrew Winter,Pedro Arduino,Clemens Krautwald,Michael Motley,Nils Goseberg
DOI: https://doi.org/10.1016/j.coastaleng.2024.104672
IF: 5.427
2024-12-07
Coastal Engineering
Abstract:We present an international comparative analysis of simulated 3D tsunami debris hazards, applying three state-of-the-art numerical methods: the Material Point Method (MPM, ClaymoreUW, multi-GPU), Smoothed Particle Hydrodynamics (SPH, DualSPHysics, GPU), and Eulerian grid-based computational fluid dynamics (Simcenter STAR-CCM+, multi-CPU/GPU). Three teams, two from the United States and one from Germany, apply their unique expertise to shed light on the state of advanced tsunami debris modeling in both open source and professional software. A mutually accepted and meaningful benchmark is set as 1:40 Froude scale model experiments of shipping containers mobilized into and amidst a port setting with simplified and generic structures, closely related to the seminal Tohoku 2011 tsunami case histories which majorly affected seaports. A sophisticated wave flume at Waseda University in Tokyo, Japan, hosted the experiments as reported by Goseberget al. (2016b). Across dozens of trials, an elongated vacuum-chamber wave surges and spills over a generic harbor apron, mobilizing 3–6 hollow debris -modeling sea containers-, in 1–2 vertical layers against friction. One to two rows of 5 square obstacles are placed upstream or downstream of the debris, with widths and gaps of 0.66x and 2.2x of debris length, respectively. The work reports and compares results on the long wave generation from a vacuum-controlled tsunami wave maker, longitudinal displacement of debris forward and back, lateral spreading angle of debris, interactions of stacked debris, and impact forces measured with debris accelerometers and/or obstacle load-cells. Each team writes a foreword on their digital twin model, which are all open-sourced. Then, preliminary statistical analysis contrasts simulations originating off different numerical methods, and simulations with experiments. Afterward, team's give value propositions for their numerical tool. Finally, a transparent cross-interrogation of results highlights the strengths of each respective method.
engineering, civil, ocean