Artificial Intelligence and Headland-Bay Beaches

• Type: Article in International Scientific Journal
• Date: 2009

Authors

G. Iglesias (Author) Other; G. Diz-Lois (Author) Other; F. Taveira-Pinto (Author) FEUP

Journal

Title: Coastal Engineering

Vol. 57

Pages: 176-183

ISSN: 0378-3839

Publisher: Elsevier

Indexing

ISI Web of Knowledge - 15 Citations

ISI Web of Science

Scopus - 14 Citations

Scientific classification

FOS: Engineering and technology > Other engineering and technologies

CORDIS: Technological sciences

Other information

Authenticus ID: P-003-99H

DOI: 10.1016/j.coastaleng.2009.10.004

Abstract (EN): Headland-bay beaches are a typical feature of many of the world's coastlines. Their curved planform has aroused much interest since the early days of Coastal Engineering. Modelling this characteristic planform is a task of great interest, not least in relation to projects of coastal structures whose effects on the shoreline must be studied from the planning stages. In this work, Artificial Intelligence is applied to this task—in particular, artificial neural networks (ANNs). Unlike conventional planform models, they are not based on a given mathematical expression of the shoreline curve. Instead, they learn from experience (from a number of training cases) how the planform of a headland-bay beach is shaped, with due regard to the obliquity of incident waves. Three artificial neural networks, with different input/output structures, are implemented and subsequently trained with a number of bays. Once trained, they are tested for validation on other headland-bay beaches. Finally, the most performing neural network is compared with a state-of-the-art planform model.

Language: English

Type (Professor's evaluation): Scientific