Desktop version

Home arrow Engineering

  • Increase font
  • Decrease font

<<   CONTENTS   >>

Future Prospects

Numerical modelling of nearshore (coastal) wave processes with the Navier-Stokes (NS) equations is a scientific area that has been studied for several decades. The research work in this field, reviewed in this book, has resulted in state-of-the-art numerical methods and techniques to solve diverse and complex processes such as wave transformation, nonlinear wave-wave and wave-structure interactions or sediment transport processes. There have been a large number of recent developments, most of which have been made possible by easy access to powerful and cheap computational CPU resources, and GPUs, which possess thousands of cores (unlike CPUs, which can only have few tens of them), and enable users to have access to an affordable and very compact computational cluster. Computational fluid dynamics (CFD), however, still presents challenges that need to be overcome in order to progress yet further. This chapter aims to present, and summarize, a broad series of topics that will likely draw significant interest in the future.

Undoubtedly, despite the apparent demise of Moore’s law, the future will bring wider access to more powerful computational resources, which in turn will open new simulation possibilities. First, this fact will enable the simulation of larger or higher resolution cases, and generalize the use of more detailed turbulence modelling/simulation approaches. Although the Reynolds-Averaged Navier-Stokes (RANS) approach is currently dominant, other methods presently used but still not fully developed, such as Large Eddy Simulation (LES) or Direct Numerical Simulation (DNS), will become more commonly applied. Perhaps DNS, which is computationally far too demanding to apply in practical simulations presently, may help provide further insight into turbulence in the future.

Another approach that will benefit from the increase in computational power is multi- physics modelling, e.g., Latham et al. (2013); Ulrich et al. (2013). This type of modelling usually requires a lot of computational resource, because such models are able to simulate wave and structure interactions, including the solid mechanics within the structures. Currently, however, these models are still in an early stage of development.

Finally, in the longer term, quantum computing is knocking at the door; this radically new computational framework will need to be explored in depth in the context of CFD model development, which will doubtlessly result in new computational methods and simulation approaches.

In a shorter time frame, there are some existing approaches that, although not extensively used at this time, possess significant advantages and encouraging results, and therefore posses the potential for benefitting the field of CFD for coastal engineering significantly.

The lattice Boltzmann method

The lattice Boltzmann method (LBM) (Succi, 2001; Frandsen, 2008; Janssen and Krafczyk, 2011) is a promising approach which is derived from the Lattice Gas Automata (LGA) method, a cellular automaton molecular dynamics model developed in the late 1980s. Despite being a discrete method, the traditional macroscopic NS equations can be derived with LGA, in which the particles in LBM move on a fixed lattice, and at every time step they undergo two processes: propagation and collision.

LBM offers significant advantages with respect to other CFD approaches, such as Eule- rian and Lagrangian (Smooth Particle Hydrodynamics, SPH) methods. For example, since the particle flow is accounted for, it can deal with complex boundaries (Yin and Zhang, 2012) and boundary conditions (Chang et ah, 2009) in a more consistent way than SPH. Also, this approach is suitable for performing massive simulations very efficiently, as it can be parallelized in GPUs (Janssen and Krafczyk, 2011). Finally, LBM has been successfully coupled with the Discrete Element Method (DEM) (Liu and Wu, 2019), to solve flow within porous media generated by solid mechanical interactions, which is an example of a multiphvsics model.

LBM presents some disadvantages too. Perhaps the best-known is the limitation to model high-speed (i.e., high Mach number) flows, although this is not likely to present an issue in coastal engineering applications, except perhaps with the exception of modelling impulsive wave breaking on structures. Another disadvantage, which is shared with other simulation approaches, is the diffusivity of the interface in multiphase cases, which impacts the numerical stability of such models, and the difficulty of dealing with the large density gradient (e.g., air-water, 1:1000) in that region (Yang and Boelc, 2013).

<<   CONTENTS   >>

Related topics