Finite Element Analysis. Barna Szabó

Чтение книги онлайн.

Читать онлайн книгу Finite Element Analysis - Barna Szabó страница 50

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

Скачать книгу

rel="nofollow" href="#fb3_img_img_aedb071a-5498-5740-b37d-de287e9695b6.png" alt="mu slash kappa equals 1"/>. Let script l equals 10. Explain what makes this choice of basis functions very special. Hint: Owing to the orthogonality of the basis functions, only hand calculations are involved.

      Up to this point we have been concerned with the finite element method based on the generalized formulation, called the principle of virtual work. There are many other finite element methods. All finite element methods share the following attributes:

      1 Formulation. A bilinear form is defined on the normed linear spaces X, Y (i.e. , ) and the functional is defined on Y. The exact solution lies in X and satisfies:(1.151) The normed linear spaces, , the linear functional F and the bilinear form B satisfy the respective properties listed in sections A.1.1 and A.1.2.

      2 Finite element spaces. The finite‐dimensional subspaces , () are defined and it is assumed that there are such that the sequence of functions ûi () converges in the space X to , that is:(1.152) The functions ûi are not the finite element solutions in general.

      3 The finite element solution. The finite element solution satisfies:(1.153)

      4 The stability criterion. The finite element method is said to be stable if(1.154) for all possible . The necessary and sufficient condition for a finite element method to be stable is that for every there is a so that(1.155) where is a constant, independent of i, or for every there is a so that this inequality holds. This inequality is known as the Babuška‐Brezzi condition, usually abbreviated to “the BB condition”. This condition was formulated by Babuška in 1971 [9] and independently by Brezzi in 1974 [29].

      If the Babuška‐Brezzi condition is not satisfied then there will be at least some u Subscript upper E upper X Baseline element-of upper X for which double-vertical-bar u Subscript upper E upper X Baseline minus u Subscript i vertical-bar upper F upper E Baseline double-vertical-bar Subscript upper X Baseline right-arrow with stroke 0 as i right-arrow infinity even though there may be u Subscript upper E upper X Baseline element-of upper X for which double-vertical-bar u Subscript upper E upper X Baseline minus u Subscript i vertical-bar upper F upper E Baseline double-vertical-bar Subscript upper X Baseline right-arrow 0 as i right-arrow infinity. Examples are presented in [6]. In general it is difficult, or may even be impossible, to separate those u Subscript upper E upper X for which the method works well from those for which it does not. The Babuška‐Brezzi condition guarantees that the condition number of the stiffness matrix will not become too large as i increases.

      Remark 1.15 Any implementation of the finite element method must be shown to satisfy the Babuška‐Brezzi condition otherwise there will be some input data for which the method will fail even though it may work well for other input data. The formulation based on the principle of virtual work satisfies the Babuška‐Brezzi condition.

      Exercise 1.23 Show that the finite element method based on the principle of virtual work satisfies the Babuška‐Brezzi condition.

      1.8.1 The mixed method

      Consider writing eq. (1.5) in the following form:

      and assume that the boundary conditions are u left-parenthesis 0 right-parenthesis equals u left-parenthesis script l right-parenthesis equals 0.

      (1.158)integral Subscript 0 Superscript script l Baseline left-parenthesis kappa StartFraction d u Over d x EndFraction upper G minus upper F upper G right-parenthesis d x plus integral Subscript 0 Superscript script l Baseline left-parenthesis upper F StartFraction d v Over d x EndFraction plus c u v right-parenthesis d x equals integral Subscript 0 Superscript script l Baseline upper T v d x period

      We define the bilinear form:

      (1.159)upper B left-parenthesis u comma upper F semicolon v comma upper G right-parenthesis equals Overscript def Endscripts integral Subscript 0 Superscript script l Baseline left-parenthesis kappa u prime upper G minus upper F upper G right-parenthesis d x plus integral Subscript 0 Superscript script l Baseline left-parenthesis upper F v prime plus c u v right-parenthesis d x

      (1.160)upper F left-parenthesis v right-parenthesis equals Overscript def Endscripts integral Subscript 0 Superscript script l Baseline f v d x period

      The problem is now stated as follows: Find u Subscript upper E upper X Baseline element-of upper H 0 Superscript 1 Baseline left-parenthesis upper I right-parenthesis, upper F Subscript upper E upper X Baseline element-of upper L squared left-parenthesis upper I right-parenthesis such that

      (1.161)upper B left-parenthesis u Subscript upper E upper X Baseline comma upper F Subscript upper E upper X Baseline semicolon v comma upper G right-parenthesis equals upper F left-parenthesis 
						<noindex><p style= Скачать книгу