Google

C*******************************************************************
C**
C**    v e m g e n 2 d :
C**
C**  generation of a triangulation of 2-dimensional domain
C**  with well-known boundaries. The program can easily changed for
C**  other problems. The mesh data are read by vemu02.
C**
C**   by L. Grosz                               Karlsruhe, Oct. 1994
C**
C*******************************************************************
C**
C**                    B3
C**              (0,1)    (1,1)
C**                 *------*
C**                 |      |
C**              B4 |      | B2
C**                 |      |
C**                 *------*
C**  xi2 ^       (0,0)    (1,0)
C**      |             B1
C**      --xi1
C**
C**
C**  This FORTRAN program generates an order 2 subdivision of a
C**  two dimensional domain into triangle elements. The domain
C**  has a parametrical representation whose domain is the unit
C**  cube [0,1]^2 in the variables (xi1,xi2). There are four
C**  boundary portions numbered from 1 to 4. Boundary 1 is
C**  represented by xi2=0, boundary 2 is represented by xi1=1,
C**  boundary 3 is represented by xi2=1 and boundary 4 is
C**  represented  by xi1=0. The parametrical representation of
C**  the domain is computed from the parametrical representation
C**  F1 and F4 of boundary 1 and 4 by a linear interpolation.
C**  The code can be adapted to other domains by entering suitable
C**  statements for F1 and F4 into the node generation.
C**
C**  The boundaries are subdivided into line elements of order 2
C**  and additionally all boundary nodes are specified as nodes
C**  with Dirichlet conditions. Since the Dirichlet conditions
C**  has the higher priority, the line elements will be useless
C**  in the vecfem application. But if you want to activate
C**  special boundary elements, you can remove the
C**  corresponding boundary nodes from the generation of the
C**  Dirichlet conditions.
C**
      PROGRAM VEMEXM
C**
C**-----------------------------------------------------------------
C**
      IMPLICIT NONE
C**
C**-----------------------------------------------------------------
C**
C**    some parameters which may be chanced:
C**
C**    ELEM1 = number of elements in x1 direction
C**    ELEM2 = number of elements in x2 direction
C**    NK    = number of solution components
C**    MESH  = name of the mesh file
C**
      INTEGER        ELEM1,ELEM2,NK
      CHARACTER*80   MESH
      PARAMETER (NK=1)
C***
      INTEGER          Z1,Z2,N1,N2,S,ELMID,D,IND
      DOUBLE PRECISION X1,X2,X3,XI1,XI2,XI,F11,F12,F31,F32
      IND=1
C**
C**-----------------------------------------------------------------
C**
      PRINT*,'Enter number of elements in x1- and x2- direction:'
      READ(5,*) ELEM1,ELEM2
      PRINT*,'Name of the mesh file :'
      READ(5,'(80A)') MESH
C**
C**-----------------------------------------------------------------
C**
C**   open output file :
C**
      OPEN (99,FILE=MESH,STATUS= 'UNKNOWN',FORM='FORMATTED')
      PRINT*,'opened file : ',MESH
C**
C**-----------------------------------------------------------------
C**
C**** the generation of the geometrical nodes :
C**   ---------------------------------------
C**
C**  The grid is rectangular in the domain [0,1]^2 of the
C**  parametrical representation of the actual domain, where Ni
C**  is the number points in XIi direction (i=1,2):
C**
      N1=2*ELEM1+1
      N2=2*ELEM2+1

      WRITE(99,*) N1*N2
C**
      DO 10 Z2=1,N2
        DO 10 Z1=1,N1
          XI1=DBLE(Z1-1)/DBLE(N1-1)
          XI2=DBLE(Z2-1)/DBLE(N2-1)
C**
	  XI=XI1
C**
C**-----------------------------------------------------------------
C**
C**   here you can change the shape of the domain :
C**
C**            B3
C**    F3(0)        F3(1)    if you set :
C**        *-------*
C**        |       |           F11=XI  F12=0
C**        |       |           F31=XI  F32=1
C**        |       |
C**        *-------*        you will get the unit cube [0,1]^2.
C**    F1(0)       F1(1)
C**           B1
C**
	  F11=XI
	  F12=0
	  F31=XI
	  F32=1
C**
C**-----------------------------------------------------------------
C**
	  X1=(1.-XI2)*F11+F31*XI2
	  X2=(1.-XI2)*F12+F32*XI2
	  X3=0
          WRITE(99,*) Z1+N1*(Z2-1),X1,X2,X3
 10   CONTINUE
C**
      PRINT*,'written nodes : ',N1*N2
C**
C**-----------------------------------------------------------------
C**
C**** the generation of the elements :
C**   -------------------------------
C**
C**   The domain is covered by triangle elements of order 2
C**   and consequently the boundaries are described by line elments
C**   of order 2. The following picture illustrates the
C**   construction of the triangle elements with lower node S
C**   and their boundary elements which are only generated
C**   if they are a subset of the boundary of the domain:
C**
C**       S+2*N1   S+2*N1+1    S+2*N1+2
C**             3-----2-----1
C**           1 3-----6-----1 2
C**           | | \         | |
C**           | |   \       | |
C**      S+N1 3 6     5     4 3  S+N1+2
C**           | |       \   | |
C**           | |         \ | |
C**           2 1-----4-----2 1
C**             1-----3-----2
C**            S     S+1        S+2
C**
C**   There are two different element types:
C**
      WRITE(99,*) 2
C**
C**   These are the triangle elements:
C**
C**   ELMID is the element id number, which allows to recognize
C**   an element after the distribution to the processors.
C**
      WRITE(99,*) 2*ELEM1*ELEM2,2,3,6

      DO 20 Z2=1,ELEM2
        DO 20 Z1=1,ELEM1
          S=2*(Z1-1)+2*(Z2-1)*N1+1

	  ELMID=Z1+ELEM1*(Z2-1)
	  WRITE(99,*) ELMID,IND,S,S+2,S+2*N1,S+1,S+N1+1,S+N1

	  ELMID=Z1+ELEM1*(Z2-1)+ELEM1*ELEM2
	  WRITE(99,*) ELMID,IND,
     &                      S+2*N1+2,S+2,S+2*N1,S+N1+2,S+N1+1,S+2*N1+1

 20   CONTINUE
C**
      PRINT*,'written triangle elements : ',2*ELEM1*ELEM2
C**
C**-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
C**
C**   These are the line elements:
C**
C**   The orientation of the line elements is counterclockwise.
C**
      WRITE(99,*) 2*(ELEM1+ELEM2),1,2,3
C**
      DO 31 Z1=1,ELEM1
C**
C****  elements on boundary 1:
C**
        S=2*(Z1-1)+1
	ELMID=Z1+2*ELEM1*ELEM2
	WRITE(99,*) ELMID,IND,S,S+2,S+1
C**
C****  elements on boundary 3:
C**
        S=2*(Z1-1)+2*(ELEM2-1)*N1+2*N1+1
	ELMID=Z1+ELEM1+2*ELEM1*ELEM2
	WRITE(99,*) ELMID,IND,S+2,S,S+1
C**
 31   CONTINUE
      DO 32 Z2=1,ELEM2
C**
C****  elements on boundary 2:
C**
        S=2*(ELEM1-1)+2*(Z2-1)*N1+3
	ELMID=Z2+ELEM2+2*ELEM1+2*ELEM1*ELEM2
	WRITE(99,*) ELMID,IND,S,S+2*N1,S+N1
C**
C****  elements on boundary 4:
C**
        S=2*(Z2-1)*N1+1
	ELMID=Z2+2*ELEM1+2*ELEM1*ELEM2
	WRITE(99,*) ELMID,IND,S+2*N1,S,S+N1

 32   CONTINUE
C**
      PRINT*,'written line elements : ',2*(ELEM1+ELEM2)
C**
C**-----------------------------------------------------------------
C**
C**   generation of the nodes with Dirichlet conditions :
C**   -------------------------------------------------
C**
C**   The real parameter is used to notify the boundary id number
C**   where the node with Dirichlet condition belongs to. Here
C**   all nodes on the boundary gets a Dirichlet condition for
C**   all components, so that the boundary elements are useless.
C**   It is easy to change the program, so that special boundary
C**   portions get no Dirichlet conditions for a component. Then
C**   boundary condition of the Neuman type can be considered.
C**
      WRITE(99,*) NK
C**
      DO 40 D=1,NK
C**
        WRITE(99,*) 2*(N1+N2-2)
C**
        DO 41 Z1=1,N1-1
          WRITE(99,*) Z1,DBLE(1)
          WRITE(99,*) N1*N2+1-Z1,DBLE(3)
   41   CONTINUE
        DO 42 Z2=1,N2-1
          WRITE(99,*) Z2*N1,DBLE(2)
          WRITE(99,*) (N2-Z2)*N1+1,DBLE(4)
   42   CONTINUE
C**
        PRINT*,2*(N1+N2-2),' Dirichlet conditions for component ',d
40    CONTINUE
C**
C**-----------------------------------------------------------------
C**
C**** end of calculation
C**   ------------------
C**