Utilising HPC resources for computational fluid dynamics (CFD) and computational solid mechanics (CSM) modelling of wave-structure interactions
Hu Zhengyu and Prof Pearl Li Yuzhu, Civil and Environmental Engineering, NUS
Sea-level rise, extreme marine events, together with increasing flood risk due to climate change put coastal communities at a growing danger. Seawalls, which serve as a major coastal defense approach can effectively contribute to dissipating local hydrodynamic energy and protecting the shoreline from erosion. The existing seawalls are often designed to be hard structures. However, traditional seawalls have to suffer costly maintenance and repair, especially after extreme storms. In many instances, seawalls are old and poorly maintained, which increases the coastal vulnerability. Moreover, as hard structures can fully reflect waves, scours at the toe of the seawall can usually be severe and inevitable. Therefore, investigating and optimising the characteristics of seawalls to improve their resistance against wave action is important.
This work investigates the interaction between breaking waves and a flexible vertical seawall of novel concept on basis of an open-source CFD software, OpenFOAM. The present model consists of a two-phase CFD solver for the wave motion, a CSM solver for the flexible seawall response, and a two-way coupling fluid-structural interaction scheme. This fully coupled tool is able to dynamically simulate the stresses and strains of a deformable structure under the impact of free-surface wave motions. The present study verifies the numerical model for simulating the responses of a flexible vertical wall in waves, and conduct novel simulations on breaking wave interaction with a flexible seawall on a sloped bed. As the breaking waves have strong turbulent features, a stabilised two-equation turbulence model, i.e., the stabilised 𝑘-𝜔 model is utilised as turbulence closure to the Reynolds-averaged Navier–Stokes (RANS) equations.
As opposed to our workstation (which takes about a week), NUS HPC clusters only take about 2 days to complete the simulation case for the passage of 30 waves when using 24 processes. As such, this is a significant way in which we can utilise HPC resources for computational fluid dynamics (CFD) and computational solid mechanics (CSM) modelling of wave-structure interactions!