You are on page 1of 5

We-04-06

Land Controlled-source Electromagnetic


Surveying for Viscous Oil Characterization in
Kuwait
R. Kumar (Kuwait Oil Company), M.A. Al-Saeed (Kuwait Oil Company), A.
Khalid (Kuwait Oil Company), A. Lovatini* (WesternGeco), M. Pezzoli
(WesternGeco), A. Battaglini (WesternGeco), F. Ceci (WesternGeco) & J.
Roth (WesternGeco)

SUMMARY
The Controlled Source Electromagnetic (CSEM) method offers an effective imaging tool for targets
characterized by a distinctive resistivity signature.
We present how onshore acquisition, processing and modeling of CSEM data, including 3D inversion, can
provide effective complementary
information to seismic data in the exploration,
characterization and potentially production of viscous oil from the shallow reservoir of a Kuwait field.

75th EAGE Conference & Exhibition incorporating SPE EUROPEC 2013


London, UK, 10-13 June 2013

Introduction
The controlled-source electromagnetic (CSEM) method offers an effective imaging tool for targets
characterized by a distinctive resistivity signature, such as lithological changes or different fluid pore
content. We present how acquisition, processing, and modeling of CSEM data can provide effective
complementary information to other geophysical methods to be used in the exploration,
characterization, and potentially, production of viscous oil from a shallow reservoir (Figure 1) in a
Kuwait field.
CSEM feasibility study for target reservoir
The presence of viscous-oil-bearing layers in the shallow reservoirs of Kuwait has a distinct signature
on resistivity logs of oil-saturated channels with respect to water-saturated channels (Figure 2).
However, it is difficult to predict their presence before drilling based only on seismic surface
measurements utilized for structural imaging and characterization of the deeper reservoirs. The CSEM
method, using an electric dipole source, is very sensitive to thin resistive layers and, therefore, more
effective for hydrocarbon-saturated layer characterization. For this reason, it was decided to utilize
this technique to provide complementary information on the saturation of middle Miocene units that
could be used to optimize future drilling, thereby reducing the exploration uncertainty.

Figure 1 Lithostratigraphic columns for the Figure 2 On the left, resistivity log examples from
Kuwait near-surface (after Al Sulaimi and Al water versus hydrocarbon-saturated channels. On the
right, three different resistivity distribution scenarios
Ruwaih, 2004).
simulated through CSEM forward modeling.

Figure 3 CSEM soundings simulated at three


different frequencies for the three resistivity
scenarios. The upper panel shows the electric
field amplitude responses while the lower
panel shows the phases.

Figure 4 Normalized CSEM soundings with respect


to the scenario with no resistive target (C). Note how
the normalized amplitude response increases with
respect to a non-resistive target case, and for
increasing target thickness.

75th EAGE Conference & Exhibition incorporating SPE EUROPEC 2013


London, UK, 10-13 June 2013

CSEM sensitivity was assessed through a synthetic simulation study that used several well logs
acquired in the shallow reservoirs of Kuwait. In this work, we present three example scenarios (Figure
2), upscaled from well logs, that were simulated through 1D CSEM forward modelling to show how
different layer characteristics (e.g., thickness or resistivity) from the area could be reflected in the
CSEM data curves (Figure 3). Scenarios A and B have a resistive target layer of 15 m and 7 m of
thickness, respectively, while scenario C has no resistor at the target depth. Multiple frequencies and
multiple source-receiver offsets were investigated through normalized amplitude and phase difference
of the predicted electric fields from scenarios A and B with respect to those from scenario C (Figure
4). The feasibility study highlighted the good sensitivity of CSEM to resistive targets at the reservoir
depth (Figure 4). It also allowed identifying the optimal frequency and source-receiver offset ranges
to be used in the acquisition.
CSEM acquisition and data processing
In the acquisition layout, the source position started offset from the receiver spread, progressing
through the spread and out the opposite side by the same distance, to cover the desired offset range.
The electric field data were recorded on 100-m dipoles, at a continuous 100-m interval, with 100 m
between receiver lines. Transmitter dipoles were also 100 m, but spaced at 300-m intervals along the
line, run between the receiver lines, with 400 m between successive source lines. The total
transmission time on each Tx dipole was determined by the desired frequency spectrum and by the
environmental noise characteristics to meet a target signal-to-noise ratio for all harmonics.
Data processing was carried out in the frequency domain, similar to the marine CSEM method.
Source current monitor and receiver electric field time series were synchronically segmented and then
transformed to the frequency domain. Source and sensor calibration functions were included in the
process, as well as all geometry data. A complex (amplitude and phase) earth impulse response
(Figure 5) was obtained for each source and receiver combination at multiple frequencies, the base
transmitted frequencies, and all their harmonics. Measured amplitude and phase values varied
depending on the source-receiver separation, on the frequency, and more importantly, on the
subsurface resistivity distribution between source and receiver.

Figure 5 Example of observed land CSEM


sounding. The upper panel shows electric field
amplitude response at multiple receiver locations
for a single source position; the lower panel
shows the corresponding phase values. Three
frequencies are shown (1, 2, and 4 Hz).

Figure 6 CSEM source-receiver survey layout


(blue dots are receiver locations, gray symbols
are source locations), and electric field amplitude
samples at 1 Hz at multiple offsets are displayed
with pseudodepth common midpoint convention
(amplitude values decrease from red to green).

CSEM data imaging


Initially, a 3D earth resistivity model mesh was built importing seismic depth horizons and upscaling
available resistivity logs to build a reference model with the general resistivity trend, but no target in
it. This model was simulated by means of anisotropic CSEM 3D forward modelling. The observed
electric field amplitude and phase data, organized in a pseudodepth cube (Figure 6) deviate from the
synthetic data proportionally as much as the true subsurface resistivity distribution deviates from the
75th EAGE Conference & Exhibition incorporating SPE EUROPEC 2013
London, UK, 10-13 June 2013

reference model. The pseudosections in Figure 7 show for two frequencies, the measured electric field
normalized amplitude (on the left) and phase difference (on the right) lateral variations along one
profile within the survey area. Resistivity variations along each section extracted from the 3D data
pseudodepth cube can be inferred by the respective variations visible in the normalized amplitude or
phase difference data. To be noted is how the resistivity increases towards the western end for the
section displayed below.
W-001

W-002

Figure 8 North-south tie line section across two


Figure 7 Observed data amplitude and phase wells extracted from the 3D inversion resistivity
pseudosections at two different frequencies model.
along an east-west line within the acquired 3D
spread.
CSEM data inversion
A second essential step in CSEM data analysis consists of inverting the data to infer a resistivity
model that could fit the observed measured data. We followed an incrementally more complex
workflow from 1D CSEM laterally constrained anisotropic inversions run on a PC workstation to 3D
anisotropic CSEM inversion using a fully parallelized code on Linux clusters.
3D inversion of collocated magnetotelluric (MT) data and 1D CSEM inversion helped to build a
reliable reference a priori model. 2.5D anisotropic CSEM inversions (Abubakar et al. 2008) were
carried out on adjacent lines and subsequently upscaled to optimize the computation efficiency and
start the 3D inversion (Mackie et al. 2007) from a model closer to the final solution. For the final 3D
inversion, a one-million-cell model was used, inverting simultaneously more than two hundred
receivers at multiple frequencies. The output anisotropic resistivity earth model showed good
consistency with the previous step qualitative imaging results and with the available resistivity logs
within the survey area (Figures 8 and 10).
Conclusions
The inversion output resistivity cube was analysed to assess its consistency with respect to the
existing geological information. Results from the modelling of collocated MT and gravity soundings
showed a high level of correlation. The CSEM resistivity cube showed a very good match between the
resistor present in the model and the depth of the expected middle Miocene units that could be
saturated.
These considerations led to the interpretation of the cube through techniques such as geobody
extraction to evaluate the presence of saturated channels and their geometry that could be used
subsequently to plan drilling both for exploration and production purposes targeting the zone with
expected higher saturation. In Figure 9, an example geobody is shown, extracted selecting an isoresistivity surface and displaying it in a 3D visualizer. This approach enabled us to identify lateral
variations within the investigated area that could be due to different saturation percentages that could
not be inferred with acoustic methods only, in particular, at the depth of these Kuwait shallow
reservoirs.
The availability in the same area of electromagnetic as well as acoustic properties also paved the way
to a more reliable quantitative estimation of petrophysical properties (saturation and porosity) to be
used in the future for the development plan and monitoring of the reservoirs.
75th EAGE Conference & Exhibition incorporating SPE EUROPEC 2013
London, UK, 10-13 June 2013

W-001

Figure 9 Resistivity amplitude geobody extracted at an arbitrary iso-resistivity value from the 3D
inversion output corresponding to the shallow reservoir depth.

Figure 10 On the left, resistivity profile extracted from the 3D inversion output model at the well
location of the resistivity log depicted on the right.
Acknowledgements
The authors thank Kuwait Oil Company and WesternGeco for permission to publish this work, and
several contributors that participated in the project execution from planning, acquisition to the
interpretation phase.
References
Abubakar, A., Habashy, T.M., Druskin, V.L., Knizherman, L., and Alumbaugh, D. [2008] 2.5D
forward and inverse modeling for interpreting low-frequency electromagnetic measurements.
Geophysics, 73, F165-F177.
Al Sulaimi, J.S., and Al Ruwaih, F.M. [2004] Geological, structural and geochemical aspects of the
main aquifer systems in Kuwait: Kuwait Journal of Science and Engineering, 31, 149-174.
Mackie, R., Watts, M.D., and Rodi, W. [2007] Joint 3D inversion of marine CSEM and MT data. 77th
Annual International Meeting, SEG, Expanded Abstracts, 26, 574-578.
75th EAGE Conference & Exhibition incorporating SPE EUROPEC 2013
London, UK, 10-13 June 2013

You might also like