Review of Modelling for Coastal Hydraulics and Engineering by K. W. Chau

Numerical models are routinely used today to simulate metocean (meteorological and oceanographic) and hydroenvironmental processes in coastal engineering applications. Conventional air–sea interaction, hydrodynamic, and water quality modelling techniques comprise finite difference, finite element, and finite volume methods and combinations of these hybrid methods. Numerical implementation of these techniques in practice-oriented codes involves well-established computational methods, as well as more recent novel algorithms including knowledge-based systems, genetic algorithms, artificial neural networks, and fuzzy inference systems. Predictive tools used in coastal engineering practice are highly specialized and problem-specific. Numerical models involve certain assumptions and limitations, with significant gaps in understanding and expectations between the developers and users of models. To bridge this gap, there has been an increasing push toward integration of artificial intelligence (AI) and knowledge management (KM) advancements in the soft computing technologies with the conventional hydraulic algorithmic modeling systems in an attempt to lessen the reliance of ordinary day-to-day application users on model developers and experts in selection, application, and manipulation of various advanced mathematical models. This book provides a cursory review of conventional coastal modeling and describes potential integration of various artificial intelligence technologies into coastal modeling systems. The author first tries to show that if the soft computing methods are used properly, they could help improve the skills of more reliable prediction of extremely complex, intertwined, and scientifically challenging coastal processes. Second, the author demonstrates that combining these techniques with classic engineering methods and harnessing their benefits may lead to more advanced and powerful predictive tools in coastal engineering.

This book consists of 11 chapters plus the reference and index sections. The chapters appear in this order: “Introduction” (2 pages), “Coastal Modelling” (5 pages), “Conventional Modelling Techniques for Coastal Engineering” (10 pages), “Finite Difference Methods” (35 pages), “Finite Element Methods” (38 Pages), “Soft Computing Techniques” (19 pages), “Artificial Neural Networks” (23 pages), “Fuzzy Inference Systems” (17 pages), “Evolutionary Algorithms” (28 pages), “Knowledge-Based Systems” (27 pages), “Conclusions” (3 pages), References (18 pages), and Index (5 pages).

Each chapter has its own introduction and conclusions. Mathematical, theoretical, and numerical treatment is very elementary, presented as a brief review in descriptive style. Only in a few chapters, the basic concepts and governing equations are included without an in-depth coverage of principles involved. The parts of this book that are unique are Chapters 6 through 10, which provide a concise review of the state of the art and suggestions for integration of different artificial intelligence technologies into coastal modeling. The algorithms and methods discussed include a broad range of topics such as soft computing, data-driven machine learning algorithms, knowledge-based expert systems, genetic algorithms, artificial neural networks (ANNs), and fuzzy inference systems. Chapter 7 on ANNs includes two example applications, Chapter 9 on evolutionary algorithms has three example applications, and Chapter 10 on knowledge-based systems has two case studies.

It is also apparent from the treatment presented that the writer has placed more emphasis on knowledge-based systems because these can assist in a more defensible intelligent manipulation of numerical modeling calibration parameters. ANNs have been increasingly used with metocean models to fill in gaps in data and for forecasting estimates. Unfortunately, ANNs ignore physical processes, decisions are derived from trends in limited data, and their use in engineering design should be carefully guarded. Nonetheless the integrated model framework considered in this book is attractive, and the combination of different techniques presented may lead to more versatile practice-oriented numerical modeling technology in coastal engineering. For these reasons, this book should serve as a first-level introductory reference to students, researchers, and engineers interested in integration of traditional modeling techniques with various soft computing methods, which today are equally popular.