After definition of the global sections the system needs more detailed information about the mesh-structure and how to initialize the quantities that are to be calculated. Therefore a grid section has to be defined supporting the following syntax
Grid = <gridname> # Read a grid named <gridname> from # the defined source file { <source definition> # overwrite global source definition file <model definition> # initialize the analytical model <output definition> # overwrite global output definition file }
where Grid is a keyword and <gridname> must be a valid name of a mesh within the chosen source.
As can be seen a local source and output definition is possible but not obligatory. If they are defined, they overwrite the previously defined global sources. The <model definition section> is a means to map the analytical model defined in the AMI onto the selected grid. To have access to the analytic variables of the model a function-like syntax is used:
Model = <ModelName>(<Q1>,<Q2>,<Q3>,<Q4>,<...>) { <initialize quantities> # initializing quantities <scale quantities> # scaling of quantities <limit quantities> # set limitations <epsilon criterion> # Newton iteration limit <quantity update-mode> # selective update modes <auxiliary definition> # auxiliary definition <limits for grid-adaptation> # limits for refinement # and coarsement }
Again Model is a keyword. <ModelName> is the name of the discretized analytical model defined in the Analytical Model Interface (described in detail in section 4.3). <Q1>,<Q2> to <Qn> are the user defined quantity names that have to be unambiguous for later boundary definitions. Various additional features are definable but only the initialization of quantities is obligatory.