| Summary: | Impulse waves generated by subaerial landslides of a block model along a frictionless surface are investigated numerically based on a combination of immersed boundary method and lattice Boltzmann method. A wave propagation model is proposed through machine learning. The observed impulse waves are classified into Stokes-like waves, solitary-like waves, cnoidal-like waves, and bore-like waves. The influence of the slope angle and the slide front angle on the generation and propagation of impulse waves is investigated from the perspectives of the quantitative analyses on primary wave amplitude and energy transformation. The energy release, dissipation, and propagation of the impulsive waves are investigated based on energy conservation principles. The wave propagation model based on the random forest model is established upon a priori classification of wave types. By comparison with the available numerical results, the accuracy of the proposed wave propagation model in predicting free water surface elevations is demonstrated.
|
|---|