Maestro En Sistemas Computacionales

Adaptive Modeling of the Near Shore Wave Behavior of the Baja California Sur Region
Carmona T. Javier A.1, Saúl Martínez D. 1, Marco A. Castro L. 1
1

División de Estudios de Posgrado e Investigación of Instituto Tecnológico de La Paz, La Paz., México Phone (612) 12-107-05 Fax (612) 12-112-95 E-mail: jcarmona@prodigy.net.mx

Abstract.-Wave behavior modeling is a highly important matterconcerning the planning, design and construction of near-shore commercial, protection or living-space edifications. A good prediction of future wave conditions is essential to avoid human and material losses. Currently there has been an increasing interest on the application of cooperative use of intelligent strategies such as Neural Networks, Fuzzy Systems, Genetic Algorithms and Rough Sets, commonlyreferred as soft-computing to solve complex systems modeling and control. This paper proposes the use of an adaptive Neural Fuzzy Inference System on the modeling of shore wave behavior on the south region of Baja California Sur, México. Current status of the field, the modeling strategy along with present results and its analysis are presented

investigations were directed hydrodynamicproblems [1].

by

the

need

of

solving

The first models treated water movement on a big scale basis; such as tidal movements on near-shore zones and currents on the superficial layer of the ocean. With the early success of this line of work it began the search for models that could account for more complex scenarios such as sedimentation and pollution shifts on rivers and on specificpoints of the shore lines. One of the processes that has gathered interest of the experts on the field is ocean surge from its origin on deep water and its propagation on to the shore line where it releases its energy affecting the coast and near shore edifications. Wave behavior models can be classified on three categories: discrete spectral models, hybrid parametric models and discrete coupledmodels. Discrete models began its use around half of the last century on Europe those first models had three key terms on their mathematical structure, an advection term, a wave generator term, and an energy dissipation term. The models were adjusted empirically to reproduce wave conditions that depended on the duration of the winds and the amount of fetch. The inputs for those models where the windintensity and direction, to obtain punctual results that were used to build wave propagation diagrams that can be used to calculate spectral energy exchange rates. The growth of the spectrum was simultaneous and it was assumed that all the wave components grew independently to their equilibrium value given by the Pierson-Moskowitz spectrum. Several modifications to the above mentioned model whereintroduced focusing on the energy dissipation and transfer aspects, that went through hybrid parametric models and discrete coupled models, in which the swell and the local winds are taken into account. One of the pioneering models of this precision and detail level when it comes to the flux structure is MIKE 21 NSW developed in Europe. The model is based on the Holthuijsen equations [3]. Johnsonused MIKE 21 NSW to study the growth of the wind waves on shallow waters with limited fetch; the simulation can account for the wave behavior taking into account the wind influence, the bottom friction, the spectral form and the break.

Key words—Fuzzy System, Neural Network, Prediction, Wave behavior.

I. Introduction
Wave behavior modeling on the near shore zones is a very important matterdue to the current growing human tendency of setting down near the coast lines. Currently, complex non-lineal and dynamic systems identification remains as a highly interesting area for the scientific community, the use of various techniques coming from the artificial intelligence field, such as, neural networks, fuzzy logic, genetic algorithms and rough sets is gathering a considerable amount of...