Analysis of a Finite Element Method: PDE/PROTRANSpringer Science & Business Media, 6 dic 2012 - 154 pagine This text can be used for two quite different purposes. It can be used as a reference book for the PDElPROTRAN user· who wishes to know more about the methods employed by PDE/PROTRAN Edition 1 (or its predecessor, TWODEPEP) in solving two-dimensional partial differential equations. However, because PDE/PROTRAN solves such a wide class of problems, an outline of the algorithms contained in PDElPROTRAN is also quite suitable as a text for an introductory graduate level finite element course. Algorithms which solve elliptic, parabolic, hyperbolic, and eigenvalue partial differential equation problems are pre sented, as are techniques appropriate for treatment of singularities, curved boundaries, nonsymmetric and nonlinear problems, and systems of PDEs. Direct and iterative linear equation solvers are studied. Although the text emphasizes those algorithms which are actually implemented in PDEI PROTRAN, and does not discuss in detail one- and three-dimensional problems, or collocation and least squares finite element methods, for example, many of the most commonly used techniques are studied in detail. Algorithms applicable to general problems are naturally emphasized, and not special purpose algorithms which may be more efficient for specialized problems, such as Laplace's equation. It can be argued, however, that the student will better understand the finite element method after seeing the details of one successful implementation than after seeing a broad overview of the many types of elements, linear equation solvers, and other options in existence. |
Dall'interno del libro
Risultati 1-5 di 38
Pagina 3
... zero ( " plane stress " , usually assumed to hold in thin plates ) or that all out - of - plane strains are zero ( " plane strain " , usually assumed in thick plates ) . In the plane stress case , these " constitutive " relations are ...
... zero ( " plane stress " , usually assumed to hold in thin plates ) or that all out - of - plane strains are zero ( " plane strain " , usually assumed in thick plates ) . In the plane stress case , these " constitutive " relations are ...
Pagina 4
... zero ( simply supported boundary ) then is zero on the boundary . In either case , the boundary integrals disappear , while in R the arbitrari- ness of forces : v2 ( DV2u ) = f in R It should be pointed out that the introduction of the ...
... zero ( simply supported boundary ) then is zero on the boundary . In either case , the boundary integrals disappear , while in R the arbitrari- ness of forces : v2 ( DV2u ) = f in R It should be pointed out that the introduction of the ...
Pagina 8
... zero source term , which can be rewritten in the form : Pt + + upx = Vpy -p ( Ux + vy ) The left hand side ( see above ) is the time derivative of the density of a moving volume of fluid , and is therefore zero for an incompressible ...
... zero source term , which can be rewritten in the form : Pt + + upx = Vpy -p ( Ux + vy ) The left hand side ( see above ) is the time derivative of the density of a moving volume of fluid , and is therefore zero for an incompressible ...
Pagina 9
... zero , but proportional to p . That is , pa ( Ux + Vy ) , where a is very large . This eliminates p and the divergence equation at the same time and it could even be argued that the resulting equations are closer to reality than the ...
... zero , but proportional to p . That is , pa ( Ux + Vy ) , where a is very large . This eliminates p and the divergence equation at the same time and it could even be argued that the resulting equations are closer to reality than the ...
Pagina 10
... zero in reality ) , the effects of the initial conditions will die out slowly , but once they do disappear the solution will be periodic with the same frequency and can be written u = A ( x , y ) sin ( wt ) ( Problem 1.5 ) . The ...
... zero in reality ) , the effects of the initial conditions will die out slowly , but once they do disappear the solution will be periodic with the same frequency and can be written u = A ( x , y ) sin ( wt ) ( Problem 1.5 ) . The ...
Sommario
1 | |
10 | |
Elliptic ProblemsForming the Algebraic Equations | 22 |
Elliptic ProblemsSolving the Algebraic Equations | 50 |
Parabolic Problems | 77 |
Altre edizioni - Visualizza tutto
Parole e frasi comuni
A.UX algorithm approximating assumed backward difference method band solver basis functions boundary conditions BOUNDARY FORCE calculated components conjugate gradient method convergence Crank-Nicolson Crank-Nicolson method cubic curved D3EST Default defined in GLOBAL derivatives diagonal dimensional discretization discretization error DISPLACEMENTS DOUBLE PRECISION Expression dxdy eigenfunction eigenvalue problem elastic error evaluated Expression involving constants ƏR₁ Figure filename finite difference method finite element formula FORTRAN frontal method frontal solver Galerkin given grid hyperbolic problems iarcI initial triangulation integration inverse power method Jacobian matrix keyword Lanczos Iteration Lanczos method linear system Newton iteration Newton's method nonlinear nonzero normalized triangle NOUPDATE output PDE/PROTRAN PDE2D piecewise polynomial plot Pn+1 positive definite PRECISION Expression involving PRECISION MATRIX quadratic quartic elements REAL or DOUBLE satisfies Section shown solution solve specified step Structure symmetric updated UPRINT values variables defined vector velocity VERTICES VSOL XGRID xk+1 zero