Björn Kaiser

Documentation is provided here:
http://www.feflow.info/html/help71/feflow/mainpage.htm#t=08_ProblemSettings%2Fother_settings.html
http://www.feflow.info/html/help71/feflow/mainpage.htm#rhhlterm=discrete%20features&rhsyns=%20&t=09_Parameters%2FDFE%2Fdfe.htm

Hi Christian, I have been in contact with your colleague addressing the assignment of MLW's. The problem in your model is related with Dirichlet Boundary Conditions which are already assigned. You tried to assign MLW’s connected with nodes where Dicirchlet Boundary Conditions have been already assigned.

FEFLOW 7.0 provides the option [b]Insert slice(s) between.... [/b]. Please use this option in the 3D Layer Configurator.

Thanks GustavS.

To view the scatter plot for the heads, you need to click on [b]View[/b] and then on [b]Charts[/b] in the main menu. After that, please click on [b]Hydraulic-Head History[/b]. Right-click on the Hydraulic-Head History and choose [b]Scatter plot[/b] from the context menu. To Access the data of the scatter plot, please right-click on the Scatter plot chart and choose [b]Properties[/b] from the context menu. Now you can export the scatter points to different types of files for further processing.

There is no reason to downgrade to a previous FEFLOW version even if there is a bug. A bug can be fixed. However, was not able to reproduce your observation. I tested the Multilayer Well assignment (MLW) and it was working perfectly using all types of z-extend (from selected edges, from top to bottom elevation and from Depth to top bottom). The latter two settings require that more than the half of two connected nodes via the edge is covered by your entered values. Could you please check it?

The Unit Conversion tool of FEFLOW provides three predefined unit sets and the possibility to apply a user-defined set. In addition to the units provided by FEFLOW also user-defined units can be created. Available unit sets include, FEFLOW default, SI - international system of units and Imperial system of units. By switching to one of these unit sets, all default input units for are set according to the unit set. The default unit for each different category can be seen and changed by selecting this category from the drop-down menu. The unit changes are not only applied to the current model, but as a default setting for the FEFLOW user interface.

Yes, you are right. It's rather case specific. ~99 percent of the models I saw used the default values. Actually, for good reasons. If you pump water you trigger an advective transport component which is overprinting molecular diffusion and/or heat conduction.

The Graphical User Interface (GUI) does not provide this option. However, you can write a customized plugin (C/C++) or a Python script which is doing this job.

Here is a quick outline: In a transient simulation you have to retrieve the concentration and the time.

Use the callback [font=courier][color=blue]postTimeStep()[/color][/font]

You can retrive the concentration at current time with
[font=courier][color=blue]
double IfmGetResultsTransportMassValue
(
IfmDocument pDoc,
int nNode
);
[/color][/font]

To set the mass concentration at another node please use:
[font=courier][color=blue]
int IfmSetBcMassTypeAndValueAtCurrentTime
(
IfmDocument pDoc,
int nNode,
int bcType,
int bcUnst,
double nValue
);
[/color][/font]

Here are the correpsonding parameters you may use:
[font=courier]
IfmDocument pDoc
Handle of document data.
int nNode
Node number.
int bcType
Mass B.C. Type:
IfmBC_NONE (0) = no boundary condition
IfmBC_DIRICHLET (1) = Dirichlet (1st kind)
IfmBC_NEUMANN (2) = Neumann (2nd kind)
IfmBC_CAUCHY (3) = Cauchy (3rd kind)
IfmBC_SINGLE_WELL (4) = Single well (4th kind)
IfmBC_GRADIENT_FTYPE (5) = Gradient-type
Note: (a) if IfmBC_NONE is used an existing B.C. Is deleted,
(B) do not use a type larger than IfmBC_GRADIENT_FTYPE.
int bcUnst
1 - if unsteady, 0 - if steady-state (constant).
double nValue
Value to be set:
If bcUnst == 1 (i.e., Unsteady) the nValue represents
The power ID (an integer) of the time-dependent function,
The power ID must be in a range available in the problem,
If the power ID does not meet a valid number an error is echoed;
If bcUnst == 0 (i.e., Steady-state) the nValue is the
Constant boundary value assigned at the current time.
[/font]

I think, a DirichletBC for the mass is what you are looking for.

You may retrive the current simulation time in the callback [font=courier][color=blue]postTimeStep()[/color][/font]by using:
[font=courier][color=blue]
double IfmGetAbsoluteSimulationTime(IfmDocument pDoc);
[/color][/font]
The simulation time can be also used as a reference value to calculate the time for the time shift.
Please check also the FEFLOW Help: http://www.feflow.info/html/help71/feflow/mainpage.htm#t=13_Programming%2Fifm.htm

I assume, you made a simulation with inactive elements and used the computational findings of the process variable as an initial condition for a second run. The second run has either other inactive elements or no inactive elements at all. If you do not re-initialize the nodes, which are completely connected with inactive elements of the first run, you start your simulation with NaN (Not a Number). This is not possible in a mathematical sense. Accordingly, I suggest to re-initialize the process variables before you run your second simulation.