Björn Kaiser

You may have a look in
<a href="https://www.mikepoweredbydhi.com/-/media/shared%20content/mike%20by%20dhi/flyers%20and%20pdf/product-documentation/feflow%20white%20papers%20vol,-d-,4.pdf">White Paper Vol IV</a>, Chapter 1.9.

A quick suggestion: Did you plot the water table (0 kPa-isoline) on several cross sections? If the shape of the water table does not look reasonable (e.g. zigzag shape) you may check if a finer vertical layer discretization helps.

You can delete or you can deactivate elements for the domain you are not interested in. Boundary conditions which are completely surrounded by inactive/deleted elements will be excluded/eliminated. Process variables (e.g. heads, pressures etc.) will be adopted for all nodes which are not completely surrounded by inactive/deleted elements. Material properties are assigned to the elements. Accordingly, if you deactivate/delete elements corresponding material properties will be excluded/eliminated. In contrast, all material properties will be maintained for all elements you do not exclude/eliminate.

If you intend to edit a Shapefile you may use a GIS. There are also free GIS available. If you want to assign a well manually you have to make an Edge Selection in the Selection Panel in case you want to assign a Multilayer Well. In case you want to assign a WellBC you have to select point(s) as the geometrical item. After that, double click on the Multilayer Well / WellBC in the Data Panel and enter required data via the Editor Toolbar.

Correct, the perforation lengths inevitably change with changing mesh geometry. Moreover, in Free&Moveable settings the Multilayer Well (MLW) is always within the saturated zone. If you want to avoid changes in the perforation lengths I suggest to switch either to the Phreatic approach or solving the Richards equation.

The diagonal in the matrix must not be <= 0. There are several possible reasons for negative entries along the principal diagonal of the matrix. One possible reason is related with strong contrasts in material properties of adjacent elements. Try to refine the mesh horizontally/vertically. Another possible reason may be related with topological errors of the mesh triggered by inconsistent input data (e.g. overlapping elements/element edges, wrong numeration of the nodes). Accordingly, the model cannot be considered as being "calibrated".

Great to hear that you managed to create a pow-file. If you need an example you may always create a "test" pow-file. Use the Time-Series Editor and generate an artificial time-series. After that, you can export the file. Load the file into an ASCII based editor and try to understand the file format.

Considering the number of wells you want to integrate in your model I suggest to make a [b]Parameter Link[/b]. The introductory tutorial shows the workflow how to assign Multilayer Wells (pages 36+37):

www.mikepoweredbydhi.com/-/media/shared%20content/mike%20by%20dhi/flyers%20and%20pdf/product-documentation/feflow-introductory-tutorial.pdf

This video may also help with the assignment. https://www.youtube.com/watch?v=Gsf6zmqY7XU

A format description about the pow-file format can be found in the FEFLOW Help. For simplicity I copy and paste the important part from the FEFLOW Help.

The text of the first comment line beginning with a '!' character is interpreted as the name of the time series (first character is skipped and 12 characters are significant).

One or multiple time series can be contained in a single file. The ID number for the functions is mandatory when having more than one series in a file, but can be skipped in case of only one series. The ID is defined by a leading line beginning with a '#' character. The syntax is:

[font=courier]# ID1
! Comment line for time series 1
! ...
x1 y1
x2 y2
...
xm ym
END
# ID2
! Comment line for time series 2
! ...
x1 y1
x2 y2
...
xn yn
END
# IDn
! Comment line for time series n
! ...
x1 y1
x2 y2
...
xo yo
END
END
[/font]

A double END statement is required at the end of the file to both end the definition of the last time series and the file.

[b]Time-series Properties[/b]

The *.pow file can also contain curve-type, time-mode, time unit, unit class and user unit properties in the comment lines that are considered at import. The properties are written in square brackets, and multiple options are separated by a semicolon:

[font=courier]
# 1
!
x1 y1
x2 y2
...
xn yn
END
# 2
!
! [type=Polylined;option=linear;timeunit=s;unitclass=LENGTH;userunit=m]
x1 y1
x2 y2
...
xo yo
END
# 3
!
! [type=Constant;option=cyclic;timeunit=d;unitclass=TEMPERATURE;userunit=^0C]
x1 y1
x2 y2
...
xp yp
END
END
[/font]

[b]Gaps[/b]

Gaps in time series are identified by the keyword GAP, as illustrated in the below example. The gap extends from the time in the value pair before the gap to the time in the value pair after the gap.

[font=courier]
# 2
! Example for gaps
0 0
0.5 1
1 0.7
GAP
10 15
100 10
END
END
[/font]

An automation to generate particle tracks is not possible in FEFLOW. In contrast, workflows to post-process computational findings as derived by age-based computations can be automatized. On top of that, age-based computations are more accurate in a physical sense than generic particle tracking techniques.

While standard streamlines are solely based on advection, other processes as such as diffusion and dispersion are simply neglected. Of course, Random Walk Particle Tracks (RWPT) mimic the additional effects diffusion and dispersion. However, both methods are based on a flow solution, because no advective-diffusive transport equation is taken into account. Moreover, these methods provide a visual inspection only. A quantitative evaluation can only be indirectly derived (second order).

Age-based computations provide a powerful alternative. Age computations do not only allow a visual inspection, but they also take budget quantities into account as driven by advection, dispersion and diffusion.

Let’s assume you are interested in delineating catchment zones. In this context you may work with the age-species Exit Propability (EP). You may prescribe Dirichlet BC’s for the EP (equal to 1.0) to nodes where water leaves the system (e.g. wells). A value of 1 at exit points corresponds to an exit probability of 100 %.

The EP is taken into account as a concentration, because internally FEFLOW solves a mass transport equation involving advection, dispersion and diffusion. During the computations FEFLOW reverses the flow field (vectors) in the advective terms of the transport equation. Accordingly, extraction wells become injection wells. These “injection wells” inject water with a “unit concentration” of 1 for the EP. This “concentration” is then transported within the porous medium and a plume may evolve. This plume can be used to delineate the capture zone.

The delineation of catchment zones can be automatized if you query the concentrations at the computational nodes.

Please let me know if you are further interested in automatizing workflows to post-process age-based computations. In this case I could provide required API functions.

Great, thanks Peter.