Problem Class

The settings defined in this section are the most important to achieve the desired model conceptualization through the FEFLOW model. The section consist in different sub-sections described below.

Title

A model name used for identification purposes only. The title can be set optionally, but has no meaning for the model run.

Projection

FEFLOW supports 2D and 3D models. Finite elements of a lower dimension (1D in 2D models, 1D/2D in 3D models), so-called Discrete Features Elements can be added, representing for example fractures or boreholes.

A newly generated finite-element mesh always represents a 2D model. Two-dimensional models can be of horizontal, vertical, or axisymmetric projection.

 

Symbol	Projection	Description
	Horizontal	2D horizontal model. Implies use of the standard groundwater-flow equation. The third dimension is eliminated by averaging vertically over the aquifer. By default, confined conditions are assumed. To model unconfined or partially confined aquifers, use the setting Unconfined conditions in the Simulate flow via standard (saturated) groundwater-flow equation section.
	Vertical, planar	2D vertical cross-sectional model. Confined if flow is simulated via the standard (saturated) groundwater-flow equation.
	Vertical, axisymmetric	Characterizes a formulation in the vertical cross-section for problems with a radial symmetry. The axis of rotation corresponds to the y-axis (x-coordinate is zero). Confined if flow is simulated via the standard (saturated) groundwater-flow equation.

 

A typical application for horizontal 2D models is the setup of regional water-management models without significant vertical flow components. Vertical models are used, for example, for the simulation of unsaturated flow and saltwater intrusion. Axisymmetric models have a radial symmetry such as the cone of a pumping or injection well. Essential for the suitability of an axisymmetric model are material properties and outer boundary conditions that are homogeneous along the circumference of the well cone or mound.

 

As soon as a finite-element mesh has been generated, FEFLOW considers a 2D horizontal problem under confined conditions by default. Since the Data panel depends automatically on the settings defined in the Problem Class, it is strongly recommended to initiate the model parametrization only if the Problem Settings have been defined.

 

If the finite-element mesh includes already a 3D extension, the options of projections are not anymore available in the Problem Settings dialog. Instead the option of Gravity Settings appears in case the major coordinate directions have to be modified.

Simulate Flow

FEFLOW provides several combinations for handling the flow solution under different model projections and model extensions (2D and 3D). The entire overview of possibilities is described in detail in section Flow Problem Class.

Standard (saturated) groundwater-flow equation

The Darcy equation is applied to simulate flow. For 2D horizontal models, the simulation of unconfined conditions can be activated in this section. For 3D models, unconfined/confined aquifers are controlled via the Free Surface settings.

 

Using the Standard (saturated) groundwater-flow equation does not necessarily mean that the entire model domain is simulated as fully saturated / confined. Additional techniques are available in Free Surface to simulate aquifers with phreatic surfaces.

 

Models with a partially or fully unstructured mesh cannot use any of the options for unconfined aquifer (s). They need to be run with the unsaturated/variably saturated flow option using Richards' equation instead.

Richards' equation (unsaturated or variably saturated media):

For unsaturated/variably saturated flow, FEFLOW solves Richards’ equation based on a single-phase approach (i.e. dominant phase is the fluid) and assumes a stagnant air phase that is at atmospheric pressure everywhere. This problem class is the preferred one in case of complex combinations of confined/unconfined, e.g. perched aquifers. In general content information (e.g. fluid volume stored in elements) is better approximated with the Richards' equation than the Free Surface settings. The reason is the large degree of flexibility provided by the Richards' approach, where in theory each individual finite-element can have its own parametric relationship (i.e. saturation vs. pressure and conductivity vs. pressure) and its own saturation limits (e.g. residual water content).

As the FEFLOW implementation of Richards’ equation also includes the proper terms for saturated flow, it is generally applicable to variably saturated conditions.

 

Due to its nonlinearity, this approach typically requires significantly higher resolution and computational effort. However in terms of physical representation is preferred compared to the Phreatic option in FEFLOW.

Include Transport

FEFLOW supports the numerical calculation of both flow and transport problems. A transport simulation is always performed in conjunction with a flow simulation. FEFLOW provides capabilities for single-species and multispecies solute-transport simulation, groundwater-age simulation with the possible species age, lifetime expectancy and exit probability as well as heat-transport simulation. A combined mass-and-heat (so-called thermohaline) transport problem can be simulated, also together with groundwater age. The following options can be activated in the Problem Settings dialog to include transport:

• Mass
• Age
• Heat

In flow and transport mode, FEFLOW provides the means to do flow only or any combinations of flow and transport simulations.

 

For 3D unconfined problems, it is not possible to include transport of BOTH dissolved constituents/age and heat at the same time.

 

If the mass/age or heat transport options are not available in your copy of FEFLOW, check the license you are using. Flow, mass, and heat transport functionality can be licensed separately. Age transport requires a mass transport license.

State (Steady State or Transient)

Transient simulations proceed from an initial condition and cover a specified time period. In contrast, a steady-state solution can be obtained directly and represents the state of a system having been subject to fixed boundary conditions and material properties for an infinitely long time. It is possible to combine a steady-state flow with a transient transport simulation. In such a case, the flow system is solved once at the beginning with all storage terms set to zero to obtain a steady-state solution as the basis for the transient transport calculation.

Pure flow simulations can be done in steady-state, calculating an equilibrium status, or for a specified time period (transient). In the latter case, all temporal settings are defined in Simulation-Time Control.

The flow and transport simulation can either be done in steady-state for flow and transport, as a transient simulation for both flow and transport, or as a steady-state flow simulation and transient transport calculation. If Richards' equation is used to simulate flow (unsaturated or variably saturated media), a steady-state simulation or a transient simulation for both flow and transport have to be applied.

 

The combination of steady flow and transient transport options assumes that the initial conditions for the flow simulation reflect steady-state conditions. During the simulation, only one iteration for the flow model will be calculated, and the resulting flow field is then used as the basis for the transient transport simulation.

 

In many cases it is useful to set up and run a model as a flow model first. The model can then later on be extended by transport components without losing the flow properties. The same procedure can be used for starting with a steady-state model that is later converted to do a transient simulation.

 

Switching from transient to steady-state requires to reduce time-dependent boundary conditions, constraint conditions and material properties to time-constant values. A dialog opens up asking for the time stage of the simulation time of the transient conditions which should be used time-constant value for the boundary and constraint conditions. A separate dialog appears for the transient material properties. All time-dependent data is lost by this procedure and time-constant boundary conditions, constraints and material properties are set based on the values of the time-dependent functions at the selected time.