Discrete Feature Elements

Discrete feature elements (DFEs) are finite-element objects of dimension lower than that of the existing model, the latter being represented by the a standard 2D or 3D finite-element FEFLOW mesh (whose finite-elements are here referred to as "matrix elements"). They are added to an existing FEFLOW mesh in order to represent one- or two-dimensional discrete features such as tunnels, pipes, drains, faults or fractures. For instance, 1D DFEs can be added to 2D and 3D meshes along selected edges while 2D DFEs can be added to 3D meshes from element faces selections.

Until FEFLOW version 7.0, DFEs always add permeability to the model or, in other words, can contribute to the simulated processes if and only if their transmissive property is higher than the one of the matrix elements they are in connection with. This is a direct consequence of the underlaying numerical method that is based on the sharing of nodes between DFE and matrix element.

FEFLOW version 7.1 provides a higher degree of flexibility to DFEs by allowing them to feature their own nodal space and by adding a physical connection between matrix element nodes and DFE nodes. This results in an added control on the exchange fluxes between DFE and matrix element through the application of a resistance approach.

If standard DFEs are traditionally assigned along element edges (1D) or faces (2D), 1D DFEs can also be setup to connect two arbitrary nodes of the mesh (thus not necessarily matching a matrix element edge), in which case they are referred to as "arbitrary" DFEs.

DFEs can be added via the context menu in the Data panel or the Discrete Features panel. In both cases the creation considers an active edge or face selection or is triggered when an arbitrary node path has been defined.

For each DFE, geometry, flow and possibly also mass-, age- and heat-transport related properties need to be assigned. The user can choose amongst three types of flow laws:

Darcy
Hagen-Poiseuille
Manning-Strickler

For phreatic or unsaturated models using Richards’ equation, the page Other Settings in the Problem Settings dialog provides four different options for the handling of unsaturated discrete features (including Multilayer Wells). By default, the entire feature is turned off if the pressure at its nodes drops below zero. A smooth variation between this "dry" state and the fully saturated state can be used to define a relative permeability kr acting in a way similar to relative permeability in the Richards equation context. The applied smoothing function can be expressed as kr(s) = s2(1 - s), with s being the occupied fraction of positive pressure in the DFE (in a 1D DFE, s is the length of the element with positive pressure over the element length).

Alternatively, DFEs can stay fully conductive or be completely turned off under both partially and fully dry conditions. As a last option, discrete features can remain fully conductive in case of partial saturation and be turned off under dry conditions.