Modelagem matemática e computacional da contaminação de aquíferos com uso de métodos numéricos sem malha

Abstract

In many problems of nature and a huge diversity of knowledge areas , there is a real need we model existing phenomena . Sciences like Mathematics , Physics , Chemistry, Biology , Economics and in Engineering , in general , is common among the researchers , the use of models and simulations , whi ch almost always involve fees , principles and laws , governed by Differential Equations . Problems involving fluid motion , intensity of electric current , heat propagation , population growth , among many others , are classic examples of applications of models g overned by Differential Equations , which can be differentiated as to type in Ordinary Differential Equations (ODE ) and Partial Differential Equations ( PDE). In the first , the function to be determined depends on a single variable, while in the second , the dependence of two or more independent variables occurs . Happens is that in a wide variety of problems of nature , the equations do not have well - behaved, analytic and thus solutions , it is necessary the knowledge of numerical methods such as Finite Differen ces , Finite Elements , Boundary Elements , among others, which require the discretization of the domain and therefore the creation of a mesh ( M ESH), with interactive formulas for estimating a solution and minimize the error of approximation . In this sense , t he purpose of this work is to use a very efficient and independent of mesh numerical method , called method without mesh ( MESHLESS), but specifically the method of Kansas , which makes use of Radial Basis Function ( Radial Basis Functions - RBF ) or radial sym metry , the distance between central point of the domain of the function and a generic point of the domain. The interpolating radial basis function also depends on a shape parameter " c" to be found . But the overriding question is how to determine a shape pa rameter " c" great, we can provide a consistent solution , reducing waste and therefore the existing error ? For both , modeled itself a problem of contamination of the aquifer by making use of the diffusion equation , comparing the results of its analytical so lution with the numerical solution obtained by numerical method without mesh and parameter simulated shape and optimized by SCILAB platform (version 5. 4 . 1 )