Finite Element Analysis. Barna Szabó

Чтение книги онлайн.

Читать онлайн книгу Finite Element Analysis - Barna Szabó страница 43

Автор:
Жанр:
Серия:
Издательство:
Finite Element Analysis - Barna Szabó

Скачать книгу

implementations of finite element analysis software have limitations on how high the polynomial degree is allowed to be and therefore it may not be possible to increase the polynomial degree sufficiently to achieve the desired accuracy. In such cases the mesh has to be refined. Uniform refinement may not be optimal in all cases, however. Consider, for example, the following problem:

      (1.118)minus epsilon squared u Superscript double-prime Baseline plus c u equals f left-parenthesis x right-parenthesis comma u left-parenthesis 0 right-parenthesis equals u prime left-parenthesis script l right-parenthesis equals 0

      where epsilon less-than less-than c, and f is a smooth function. Intuitively, when epsilon squared is small then the solution will be close to u equals f slash c however, because of the boundary condition u left-parenthesis 0 right-parenthesis equals 0, has to be satisfied, the function u left-parenthesis x right-parenthesis will change sharply over some interval 0 less-than x less-than d left-parenthesis epsilon right-parenthesis less-than less-than script l.

      Letting c equals 1 and f left-parenthesis x right-parenthesis equals 1 the exact solution of this problem is

Graph depicts the solution uEX(x), given by eq. (1.119), in the neighborhood of x=0 for various values of ε.
, given by eq. (1.119), in the neighborhood of
for various values of epsilon.

      The optimal discretization scheme for problems with boundary layers is discussed in the context of the h p‐version in [85]. The results of analysis indicate that the size of the element at the boundary is proportional to the product of the polynomial degree p and the parameter epsilon. Specifically, for the problem discussed here, the optimal mesh consists of two elements with the node points located at x 1 equals 0, x 2 equals d, x 3 equals script l, where d equals upper C p epsilon with 0 less-than upper C less-than 4 slash e.

      A practical approach to problems like this is to create an element at the boundary (in higher dimensions a layer of elements) the size of which is controlled by a parameter. The optimal value of that parameter is then selected adaptively.

      1.6.2 The exact solution lies in bold upper H Superscript bold k Baseline left-parenthesis bold upper I right-parenthesis, bold k minus bold 1 bold less-than-or-equal-to bold p

      In this section we consider a special case of the problem stated in eq. (1.103):

      (1.120)integral Subscript 0 Superscript script l Baseline u prime v Superscript prime Baseline d x equals upper F left-parenthesis v right-parenthesis comma for all v element-of upper E Superscript 0 Baseline left-parenthesis upper I right-parenthesis

      with the data u left-parenthesis 0 right-parenthesis equals u left-parenthesis script l right-parenthesis equals 0, script l equals 1 and upper F left-parenthesis v right-parenthesis defined such that the exact solution is

      (1.121)u Subscript upper E upper X Baseline equals x Superscript alpha Baseline left-parenthesis 1 minus x right-parenthesis comma alpha greater-than 1 slash 2 comma 0 less-than x less-than 1

      that is,

      (1.122)upper F left-parenthesis v right-parenthesis equals integral Subscript 0 Superscript script l Baseline left-parenthesis alpha x Superscript alpha minus 1 Baseline minus left-parenthesis alpha plus 1 right-parenthesis x Superscript alpha Baseline right-parenthesis v Superscript prime Baseline d x period

      On integrating by parts, we get the following expression which is better suited for numerical evaluation:

      (1.123)

Скачать книгу