This section defines the polynomial expansions used on each of the defined geometric composites and variables. Expansion entries specify the number of modes and the expansion type, or a full list of data of basis type, number of modes, points type and number of points. The short-hand version has the following form
1<E COMPOSITE="C[0]" NUMMODES="5" FIELDS="u" TYPE="MODIFIED" />
or, if we have more then one variable we can apply the same basis to all using
1<E COMPOSITE="C[0]" NUMMODES="5" FIELDS="u,v,p" TYPE="MODIFIED" />
The expansion basis can also be specified in detail as a combination of one-dimensional bases, and thus the user is able to, for example, increase the quadrature order. For tet elements this takes the form:
1<E COMPOSITE="C[0]" 2 BASISTYPE="Modified_A,Modified_B,Modified_C" 3 NUMMODES="3,3,3" 4 POINTSTYPE="GaussLobattoLegendre,GaussRadauMAlpha1Beta0,GaussRadauMAlpha2Beta0" 5 NUMPOINTS="4,3,3" 6 FIELDS="u" />
and for prism elements:
1<E COMPOSITE="C[1]" 2 BASISTYPE="Modified_A,Modified_A,Modified_B" 3 NUMMODES="3,3,3" 4 POINTSTYPE="GaussLobattoLegendre,GaussLobattoLegendre,GaussRadauMAlpha1Beta0" 5 NUMPOINTS="4,4,3" 6 FIELDS="u" />
The expansions can be defined with a list of <E>
elements (e.g., to represent different
polynomial orders for different variables or to address different composites). The user can
define a default expansion field by entering <E>
tags without the FIELDS
attribute. The default
expansion is used to define any variables not explicitly listed in the <E>
entries. In the
following example, the default expansion is used to define the expansions for the composites
C[0], C[1] and C[2]:
1<E COMPOSITE="C[0-2]" NUMMODES="5" TYPE="MODIFIED" /> 2<E COMPOSITE="C[3]" NUMMODES="4" TYPE="MODIFIED" FIELDS="u,v"/> 3<E COMPOSITE="C[3]" NUMMODES="3" TYPE="MODIFIED" FIELDS="p"/>
The expansions of each field should be defined only once for each composite.