Critical analysis of Hydraulic stimulation, geothermal reservoirs, fluid flow, electric potential and microseismicity, the Soultz-Sous-Forets Hot Dry Rock Site

1. RESERVOIR STIMULATION
CRITICAL ANALYSIS OF A RESEARCH PAPER ON
RESERVOIR STIMULATION

2. Hydraulic stimulation of geothermal reservoirs: fluid flow, electric potential and microseismicity relationships
Mathieu Darnet,∗ Guy Marquis and Pascal Sailhac
Proche Surface, IPGS, UMR 7516 CNRS-ULP, Strasbourg, France. E-mail: guy.marquis@eost.u-strasbg.fr Accepted
2006 March 30. Received 2006 March 29; in original form 2005 November 29
Problem question:
Hydraulic stimulation of geothermal reservoirs and the fluid flow, electric potential and
microseismicity caused during and after stimulation process and their relationships at the Soultz-
Sous-Forets Hot Dry Rock Site.
Objectives of study:
Objectives of the study are
 Understanding how hydraulic stimulation works in geothermal reservoirs
 Understanding the fluid flow patterns of the injected fluid
 Calculating streaming potentials caused due to electrokinetic interaction of fluid and
permeable rocks
 Calculating and finding reasons of microseismicity in reservoir due to stimulation
 Understanding the relationship between injected fluid flow, streaming potentials and
microseismicity
Hypothesis:
Reservoir stimulation is done in order to enhance the fluid transport properties of a
reservoir (i.e. permeability) to optimize its performance. This methodology has been used in
hydrocarbon reservoirs and geothermal reservoirs. A large quantity of fluid (water in case of
hydraulic stimulation) is injected in to the reservoir to achieve stimulation.
Reservoir stimulations usually generate significant microseismic activity. The simple
explanation is that the crust is in a near-critical state of stress, and so any small perturbation, such
as those induced by the injection of fluid, will provoke failure. In addition, the increase in pore
pressure brought by the injection of fluid will result in an increase in effective normal stress.

3. Stimulation and production in deep geothermal reservoirs also generate electric potential
variations. Streaming Potentials are electric potentials generated by the electrokinetic interaction
of a fluid flowing through a porous medium. Pore fluid is in equilibrium with rock matrix and this
accumulate ions at rock-fluid interface, so when a fluid flows it moves these ions. If there is no
external source this convection current is balanced by conduction current, and this conduction
current is responsible for anomalies.
∇. (σr ∇ V) = − ∇. (L ∇ P)
Left hand side of equation show divergence of conduction and left side divergence of
convection current. P shows pore fluid pressure, σr shows rock electrical conductivity and L shows
electrokinetic coupling coefficient.
The simple interpretation of these SP signals is that they originate from electrokinetic
processes as water circulates through fractures within the reservoir. Good knowledge of electrical
conductivity structure is required to map fluid circulation as ground contrasts can distort the
current lines. Streaming potentials can also be used as a time-lapse monitoring tool. In this
approach, knowledge of the ground electrical conductivity structure is not necessary because the
electrical heterogeneities, their effects on the potential are considered static.
SP is a potential-field method, each measurement integrates the effect of all electrokinetic
sources and represent overall hydrodynamics of the reservoir. So a SP survey gives information
about the fluid flow dynamics at the reservoir scale.
Methodology:
SP data reported by Marquis et al. (2002) and more recent data acquired during stimulation
experiments of the geothermal reservoir of the Soultz-sous-Forets (France) Hot Dry Rock site has
been studied. These data sets contain both high-quality SP and microseismic data for the same
stimulation. SP data along with microseismic data has been analyzed and interpreted, as both are
related to fluid flow at depth and shown that this data provide information about the fluid flow
during stimulation.
 Stimulation Experiment

4. Soultz-sous-Forets hot dry rock site is located in Rhine graben in north of Alsace. It is
studied for its geothermal potential of high temperature gradient. Main objective is to use hot
granite, 200℃ at 5km depth, as thermal exchanger to heat up water for electricity production.
Natural permeability of the reservoir is increased through hydraulic fracturing. In 2000, a
stimulation experiment was conducted in well GPK2 to develop the geothermal reservoir at
depths between 4400 and 5000 m. During operation, 23,000 tons of water were injected at flow
rates up to 50 kgs, yielding overpressures around 13 MPa. This water injection started with
the injection of 2500 tons of high-density (1.2 kgl) saturated brine in order to stimulate the
entire length of the open-hole section and was followed by the injection of fresh water. 1 week
after shut-in, a test of injectivity was performed with the injection of 4500 tons of fresh water
at flow rate up to 30 kgs. This stimulation operation induced around 30,000 microseismic
events with magnitudes up to 2.6 strongly correlated with the injection phases. 14,000 events
were localized and their envelope gives the size of the stimulated volume.
 Surface SP survey:
Surface electric potentials with Pb-PbCl2 unpolarizable electrodes at twelve sites over 1
km2 area was recorded. Each site consisted in four electrodes and their reference was chosen
at 50m from the GPK2 wellhead. SP time series (24-hr low-pass filtered to remove
electromagnetic induction) from two sites located 500m and 250m from the wellhead was used.
A long-term electric potential variation of roughly 4 mV strongly correlated with the water
injection phases was identified. The straightforward interpretation is that they are generated by
electrokinetic phenomena of the injected water circulating through the stimulated fractures,
other mechanisms (electrochemical, electro thermal) are also responsible.
 Casing Effect
Metal casing effects and channels the electric lines near the well. Metal casing is assumed
as equipotential with host medium of homogeneous electrical resistivity, although it
underestimate the electrokinetic coupling but we are interested in integrated SP response, so
the assumption is reasonable.
Electric potential distribution for a single electric current located at the shoe of a Soultz
casing is computed. The graph below shows the effect of casing on anomalies, it increases the

5. effect by a ten factor so the anomalies are detected in mV and without casing SP anomalies
would have been very sensitive to measure.
Casing also controls the geometry of electric lines so SP source geometry and fluid flow geometry
are not recoverable. Another data set shows that SP values decrease with increasing distance from
well. As fluid flow is 5km deep, so we can expect that without casing effect SP anomalies would
be almost equal. Redox potential variation also contribute as charges migrate along metal casing.
Discussion:
 Origin of Surface SP during Stimulation

6. Most of the SP is generated by electrokinetic coupling process during stimulation, but
electrochemical and electrothermal processes also contribute to SP anomalies. In case of Soultz
reservoir electrokinetic coupling dominates when injection rate is greater than 30kgs but when
injection rate is below 10kgs then electrochemical and electrothermal processes dominate.
Electrokinetic coupling is warranted to cause SP anomalies during stimulation in this case.
 Post-Shut-In fluid flow
After the injection stopped it was noted that SP decay is slow as compare to pressure decay.
Electrochemical and electrothermal processes only add a small amount to SP anomaly and
most is added by electrokinetic coupling due to persisting fluid flow even after injection
ceased.
 Over-Pressure, Seismicity and Streaming Potential After Shut-in
Microseismic events were observed even after 1 month after the shut-in. This is due to
post-shut-in fluid flow. Microseismic events density start decreasing after the shut-in. As slow
SP decay indicates persisting fluid flow, it also indicates reason for microseismic events.
Post-shut-in fluid flow should also indicate slow pressure decay but that is not the case
here. As pressure measurements describe flow dynamics within hydraulically connected zones
to sensor and SP measurements integrates effect of fluid flow over whole reservoir. So after
shut-in zone connected to openhole is disconnected from rest of reservoir so only local flow
pressure around well is measured while SP measurements are of whole stimulated area.
Microseismic activity shut-in shows uncoupling between vicinity of well and rest of
reservoir. After the shut-in persistence events occur away from openhole and an aseismic zone
grows from openhole into reservoir indicating relaxation around openhole. As no microseismic
event in this area so fractures may close and uncoupling is caused.
Migration of the seismic event is most likely due to fluid diffusion. A high pressure zone
was propagating out from the well even after injection ceased. It is due to the fluid movement
away from edge of stimulated area. This fluid flow also cause slow SP decay and activates
fractures away from stimulated area.
Conclusion:

7. Electrokinetic effects related to water injected into the Soultz-Sous-Forets Hot Dry Rock
geothermal reservoir during a stimulation experiment generate surface streaming potential (SP)
anomalies of several mV. These surface SP anomalies are detectable even 5 km above the fluid
flow thanks to the electrically conducting steel casing channeling the electric current to the surface.
We have confirmed that electrokinetic phenomena dominate the observed signals. In particular,
the slow temporal SP decay observed after shut-in is related to large fluid flow persisting after the
end of the stimulation operation at the edge of the stimulated area. This flow may explain why the
microseismic activity is maintained at high level in this zone long after shut-in (until 1 month in
our case).
Furthermore, this flow is not visible on hydraulic data because it takes place in a zone
hydraulically disconnected from the open hole. Therefore, during this stimulation experiment, the
capacity of SP measurements to monitor fluid flow at the reservoir scale has revealed the existence
of a fluid flow playing a major role in the mechanical response of the reservoir to the hydraulic
stimulation. For stimulation experiments, this kind of information could prove useful for seismic
risk assessment. Furthermore, the application of such a method to natural systems in order to
observe fluid diffusion processes is useful for the understanding of the role of fluid in earthquakes
mechanics, especially during aftershocks.
Recommendation:
This stimulation method and the resulting fluid flow pattern detection should be applied in
seismic studies to determine the aftershocks. Injection rate should not be kept too high as they
cause large microseismic events which can damage the production well leading to economic loss
and on surface as well.