Implications of different fault zone appearances on pressure signals in pumping tests for a typical setting of the Upper Jurassic deep geothermal reservoir in the South German Molasse Basin

Florian Konrad1, Kai Zosseder1
1 Lehrstuhl für Hydrogeologie, TU München

P 3.4 in Deep geothermal energy and deep groundwater

After the drilling and completion operations of a new geothermal well are
finished, up to three acidifications in conjunction with short airlift pumping
tests (a few hours) are generally applied in the Upper Jurassic deep geothermal
reservoir of the South German Molasse Basin (Malm aquifer). At a later stage
of the construction this is normally followed by a longer pumping test (up
to several days) in which the pressure evolution in the geothermal well for
stepwise increasing pumping rates is recorded. Those investigations are really
important because they reveal if the exploration strategy was successful and
if so how the reservoir is performing as well as what its hydraulic properties
are. [Schneider and Thomas, 2012]
To investigate the collected pressure data modern techniques are applied,
which originated from the oil industry. For the evaluation of pumping tests
they get summarized under the term ”pressure transient analysis” (PTA)
[Houze et al., 2011]. A core method in PTA is the derivative analysis also
called the Bourdet Derivative [Bourdet et al., 1983, Bourdet et al., 1989]. It
allows for identification and quantification of well, reservoir and boundary
models [Houze et al., 2011].
Ideally the effect of a fault on the pressure evolution can be seen in the
Bourdet Derivative by its slope in the early time region (linear flow). The
Malm aquifer shows in practice only rarely evidence for a flow regime around
a geothermal well dominated by faults [Schneider and Thomas, 2012]. This
doesn’t imply that there are no hydraulically active faults anywhere else at
exploration sites in this aquifer. On one hand technical effects (e.g. well
bore storage, effects due to the used pumping technique, skin) can cloud the
pressure signal in the early time region. But additionally a specific contrast
between a fault and the aquifer matrix is needed for the system to actually
show linear flow in a pumping test.
By performing parameter studies while observing the Bourdet Derivative
the impact of possible fault zone and reservoir settings on pumping tests with
respect to the Malm aquifer was investigated. Therefor a representative but
simplified, three-dimensional, numerical box model (one 70 degree inclined
fault surrounded by 500 meter thick aquifer matrix, one vertical well, 3000m
depth of the reservoir top, 1000m fault length) was built usingMeshIT [Blöcher
and Cacace, 2013] and Paraview [Ahrens et al., 2005]. By means of this
box model multiple different hydraulic simulations of idealized pumping tests
were executed. The simulation code MOOSE Framework [Gaston et al., 2009]
with the complementary application GOLEM [Cacace and Jacquey, 2017] was
used. The main parameters that were varied are the fault zone thickness,
matrix and fault zone permeability as well as storage, distance between well
and fault zone and production rates. To account for the variety of different
parameter combinations a Monte Carlo analysis was realized by application of
the RAVEN code [Alfonsi et al., 2017] on a HPC system (LRZ linux cluster).

[Ahrens et al., 2005] Ahrens, J., Geveci, B., and Law, C. (2005). ParaView: An End-User
Tool for Large Data Visualization, Visualization Handbook. Elsevier.
[Alfonsi et al., 2017] Alfonsi, A., Rabiti, C., Mandelli, D., Cogliati, J., Wang, C., Talbot,
P. W., Maljovec, D. P., and Smith, C. (2017). RAVEN Theory Manual and User Guide.
Idaho National Laboratory.
[Blöcher and Cacace, 2013] Bl¨ocher, G. and Cacace, M. (2013). MeshIt - A software for three
dimensional geometric modelling of complex fractured geological systems. Computers and
[Bourdet et al., 1989] Bourdet, D., Ayoub, J., and Pirard, Y. (1989). Use of Pressure Derivative
in Well Test Interpretation.
[Bourdet et al., 1983] Bourdet, D., Whittle, T., Douglas, A., and Pirard, Y. (1983). A new
set of type cuves simplifies well test analysis.
[Cacace and Jacquey, 2017] Cacace, M. and Jacquey, A. B. (2017). Flexible parallel implicit
modelling of coupled Thermal-Hydraulic-Mechanical processes in fractured rocks. Solid
Earth Discussions, pages 1–33.
[Gaston et al., 2009] Gaston, D., Newman, C., Hansen, G., and Lebrun-grandi´e, D. (2009).
MOOSE : A parallel computational framework for coupled systems of nonlinear equations.
In International Conference on Mathematics, Computational Methods & Reactor Physics
(M&C 2009), volume 239, pages 1768–1778.
[Houze et al., 2011] Houze, O., Viturat, D., and Fjaere, O. S. (2011). Dynamic Data Analysis.
[Schneider and Thomas, 2012] Schneider, M. and Thomas, L. (2012). Verbundvorhaben :
Wissenschaftliche und technische Grundlagen zur strukturgeologischen und hydrogeologischen
Charakterisierung tiefer geothermisch genutzter Grundwasserleiter am Beispiel des
süddeutschen Molassebeckens. Endbericht–BMU Forschungsvorhaben, page 237.