Page 79 - Process Modelling and Simulation With Finite Element Methods
P. 79

66         Process Modelling and Simulation with Finite Element Methods


          demarcation between regions that have already received the information, regions
         that will receive the information, and possibly regions that will never receive the
          information.  If the system is linear or quasi-linear (i.e. some coefficient depends
          on  the  dependent  variable  or  a  lower  order  partial  derivative  than  that  it
          multiplies), this classification  system and the intuition about how information is
          transported  serves as  a  robust  guide  to  second  order  systems.  For nonlinear
          systems, however,  nonlinearity  can destroy the information transport  structure.
         In nonlinear systems, information may be “bound”, i.e. never transferred, beyond
          given attractors, or it may be created from noise (one view) or lost (a different
          view) by forgetting initial conditions in a given window in time.


          2.1.1  Poisson’s equation: An elliptic PDE
          A modest variant on Laplace’s equation is the Poisson equation:

                                    V2u = f (x)

          We saw this equation in  1 -D form in (1.19) which described heat transfer  in a
          nonuniform medium with a distributed source.  Here, the thermal conductivity is
          uniform.  In order to give a different spin on (2.6), one should note that it is the
          equation satisfied by the streamfunction with an imposed vorticity profile:
                                   V2y/ = --u)  (x )


          There are  two  common  types  of  vortices  that  are  easy  to  characterize  - the
          Rankine vortex, where vorticity is constant within a region, and the point-source
          vortex, where  vorticity  falls off  rapidly  and  thus  is idealized  as point  vortex.
          One might be curious about the streamlines generated by these two vortex types.

          Start up FEMLAB and enter the Model Navigator:

                   Model Navigator
                          Select 2-D dimension
                          Select Classical PDEs + Poisson’s Equation
                          Element: Lagrange - quadratic
                          More>>


          This  application mode  gives us  one dependent  variable  u,  but  in  a  1-D space
          with coordinate x. Now we are in a position to set up our domain.  Pull down the
          Draw menu and select Circle/Ellipse.
   74   75   76   77   78   79   80   81   82   83   84