We D 02

Features of the Transient Electromagnetic Survey

Applying in Situation of Highly Conductive
Sediments in Southeast Asia
I.V. Egorov (Irkutsk Electroprospecting Company), M.V. Sharlov (Irkutsk
Electroprospecting Company), I.V. Buddo* (Irkutsk Electroprospecting
Company), Y.A. Agafonov (Irkutsk Electroprospecting Company) & A.V.
Pospeev (Irkutsk Electroprospecting Company)

SUMMARY
Electroprospecting works have been performed over one of the blocks in the Southeast Asia. Two
techniques were used: Transient Electromagnetic Soundings in Time Domain (TEM) and Induced
Polarization Electromagnetic method (EM-IP). The objective is to conduct a comprehensive study of the
sedimentary cover using TDEM. This includes mapping of zones with potential reservoirs properties from
TEM, and to qualify these areas with Induced Polarization (IP) anomalies, usually related to presence of
hydrocarbon system.
From TEM investigations was found that sediments of surveyed area are highly conductive. Total
conductivity exceeds 2000 S. Furthermore, the high noise level is common for the survey area. The most
influence on the EM data has 50 Hz noise due to its unstable parameters: amplitude and frequency.
Nevertheless EM technique has shown sensitivity to geoelectric parameters changes to the depth up to 4
km. It was possible with development of special algorithms of EM signals processing.
Overall investigations show possibility of TEM and EM-IP application even in highly conductive
environment and complicated noise situation.

6th Saint Petersburg International Conference & Exhibition – Geosciences – Investing in the Future
Saint Petersburg, Russia, 7-10 April 2014
Introduction

Electroprospecting works have been performed over one of the blocks in the Southeast Asia. Two
techniques were used: Transient Electromagnetic Soundings in Time Domain (TEM) and Induced
Polarization Electromagnetic method (EM-IP). The objective is to conduct a comprehensive study of
the sedimentary cover using TDEM. This includes mapping of zones with potential reservoirs
properties from TEM, and to qualify these areas with Induced Polarization (IP) anomalies, usually
related to presence of hydrocarbon system.

From TEM investigations was found that sediments of surveyed area are highly conductive. Total
conductivity exceeds 2000 S. Furthermore, the high noise level is common for the survey area. The
most influence on the EM data has 50 Hz noise due to its unstable parameters: amplitude and
frequency. Nevertheless EM technique has shown sensitivity to geoelectric parameters changes to the
depth up to 4 km. It was possible with development of special algorithms of EM signals processing.

Overall investigations show possibility of TEM and EM-IP application even in highly conductive
environment and complicated noise situation.

Survey area and applied EM techniques

Survey area is located on an island in Southeast Asia. The terrain is almost flat, with a large number
of rivers and mangrove swamps. Most of the densely populated area of work, covered with numerous
farms and artificial lakes, is meant for fish and shrimp. Uninhabited land area covered by jungle. The
area of work covered by a dense network of country roads on which the electromagnetic survey were
carried out.

During survey two techniques were applied: Transient Electromagnetic Soundings in Time Domain
(TEM) and Induced Polarization Electromagnetic method (EM-IP).
For TEM the ungrounded square loops with side length of 500 and 750 m were used as a source of
electromagnetic field (Tx). Receiver loops (Rx) were 18×18 m size. The profile step was 250 and 350
m, depending on the prospected area and transmitter loop size. Based on the field test result, TEM
template geometry was decided: 750×750 m – transmitter loop sizes (Tx loop), 18×18 m – receiver
loop sizes (Rx loop), spacing – 350 m. TEM sounding was performed with digital telemetric electric
prospecting system SGS-TEM “Picket – 2” (CJSC IERP, Irkutsk).

The template used for EM-IP survey on block: The length of both transmitter lines (Tx) and receiver
lines (Rx) were 600 m. The distance between transmitter and receiver was 600 m, the length of profile
line step – 600 m. For data acquisition in the field, two Rx and one Tx line were used simultaneously.
Current was injected into Tx lines in bipolar pulses with period of 18 s and current value from 25 to
50 A. Current pulses formed were meander shape, as well as in TEM.
Electromagnetic signals (EMF measured on MO and ON lines) were registered by «ERID-2» (CJSC
IERP, Irkutsk) equipment.

Challenges and solution ways

Sedimentary cover of the study area is composed of young Paleogene-Neogene sediments. According
to the log studies rocks resistivity varies from 0.5 – 1 to 10 – 15 Ohm·m. According to the projected
depth studies - 4 km, the total longitudinal conductivity of sediments exceeds 2000 S. Typically, TEM
technique is applied when the total longitudinal conductivity of the sedimentary cover is up to 500 S.
Thus, TEM investigations in such circumstances fraught with difficulties.

As the results of test works, the use of Tx with a side length of 500 m and a current of 150 Amp
allows to study the sedimentary cover to a depth up to 3 km. Beyond which the SNR starts to get low
and longer pulse/pause period would not be very reliable. The 750 m Tx Loop, acquired with 24 s
pulse/pause period, shows much cleaner data with higher SNR up to 4 km. The

6th Saint Petersburg International Conference & Exhibition – Geosciences – Investing in the Future
Saint Petersburg, Russia, 7-10 April 2014
resistivity/conductivity trend at the depth of interest can be clearly inferred. Since the main objectives
were highlighted at the depth interval of 1.7 and 2.5 – 3.8 km the size of Tx 750 m was chosen.

A B

Figure 1 TEM curves comparison for different Tx Loops: A – apparent resistivity depends on time of
sounding; B – apparent conductivity depends on apparent depth (right).Color legend: Red – 750 m Tx
size, Black – the 500 m.
The required depth of penetration occurs in all the curves and is estimated as approx 4 km.
Selection of pulse width determines the length of the off-time (recording time) and the time of current
injecting into Tx. The longer is the pulse the greater is the penetration depth. Same time the longer is
the stacking the less is the influence of random noise on the data (spherics, environmental, system)
hence better signal to noise ratio. A combination of pulse width and stacking duration will control the
time for each reading/station (integration time). Optimal length of the transmitted pulse depends on
the depth of investigation, the conductivity of the sediments and the SNR.
When using pulse/pause period 8 s it is possible to penetrate to the depth up to 2 km (4000 ms), 16 s –
up to 3000 m (8000 ms). For the current survey 24 s guarantees reaching 4000 m (12000 ms).

Attenuation of noise in TEM data from the survey territory

The TEM responses collected in the survey territory are distorted mainly by atmospheric noise and
industrial noise with the main frequency 50 Hz. Atmospheric noise is easy to eliminate during
processing with special robust filters, but the industrial noise deteriorates strongly the data quality
because it changes abruptly its frequency and amplitude [1].

For 50 Hz noise attenuation the special combined processing algorithm was used. The early-time
transient responses are processed using direct noise subtraction. The procedure consists of three steps:
estimating the noise parameters to a pre-specified accuracy, generating respective synthetic noise, and
subtracting the synthetic noise from the distorted signal. If its parameters are estimated to a high
accuracy, the noise can be almost fully removed from the resulting signal. However, highly accurate
noise estimation consumes too much of computing resources to be applied in the field. The direct
noise subtraction algorithm is markedly advantageous over other filter types (rejection, low-frequency
cutoff, or lowpass filtering) as it implies subtraction of noise from a complicated response rather than
mere cutting off the 50 Hz signal. As a result, the signal retains the useful information acquired at the
50 Hz frequency.

The record at a high current density is times longer than that at low current, and the noise frequency
may, in turn, vary strongly with time. Therefore, it is impossible to estimate the noise parameters to a
satisfactory accuracy over the entire record length. The late-time transient responses can be processed
with a lowpass filter. This is a low-frequency (LF) filter which cuts off all frequencies above its
lowpass. Transient responses being exponential and covering the full range of frequencies, the

6th Saint Petersburg International Conference & Exhibition – Geosciences – Investing in the Future
Saint Petersburg, Russia, 7-10 April 2014
lowpass filtering distorts irreversibly the early-time portion of data where the frequencies are higher
than the lowpass. However, it works perfectly for late times and can improve considerably the
signal/noise ratio.

At the final step, the early- and late-time records (before and after 200 ms, respectively) processed
separately are spliced into a single final curve. The final curves thus have high signal/noise ratios over
the entire range of delay times.

Verification of TEM/EM-IP data processing

To prove the effectiveness of TEM/EM-IP data processing verification was conducted using synthetic
data, in the case of distortion caused by 50 Hz interference.

The workflow for verification of TEM/EM-IP data processing was:

1. Generating synthetic EM signals for a resistivity pattern typical of the surveyed block;
2. Adding real background and synthetic 50 Hz noise to the signals;
3. Processing the noisy EM curves with conventional and special filtering algorithms;
4. Comparing the EM curves processed using different filtering algorithms with the reference
initial curves (without noise);
5. Making conclusions.

The TEM and EM-IP responses were processed with SignalGenius software (designed by M.V.
Sharlov, CJSC IERP). SignalGenius is a software designed for generating synthetic signals of wanted
waveforms to be used for calibrating and testing the modern digital resistivity survey instruments, as
well as for verifying TEM-Processing algorithms. TEM and EM-IP responses were modeled using a
resistivity pattern typical of the survey region. The above model parameters were used to compute
theoretical TEM and IP responses, assuming the same system configuration as in the field. The time
grid of the curves corresponded to the sampling rates 250 µs and 1 ms used during the acquisition.
Thus obtained TEM and EM-IP curves have been taken for reference.

SignalGenius allows adding noise of two types to the reference model signals: Real background noise
recorded within the survey territory; Synthetic 50 Hz noise (Fig. 2 A).

A B

Figure 2 A - Resulting distorted model TEM signal; B – Percentage of EMF difference from the
model curve in the two filtered responses. Curve color codes are: arithmetic mean averaging
algorithm (black) and combined 50 Hz attenuation (red).
The obtained synthetic TEM and EM-IP responses were processed in two ways: by simple arithmetic
mean averaging algorithm and with a combined approach described above.
Arithmetic mean averaging was applied for each time delay discrete of a signal for processed series of
samples. The data filtered with the conventional arithmetic mean algorithm suffer from heavy noise
whereas the 50 Hz filter attenuates the noise quite effectively. The difference appears only at late
times where the signal is too weak. In Fig. 2 B the departure of the two filtered curves from the
reference is shown in percent. Thus, the average difference reaches 19 % in the case of simple

6th Saint Petersburg International Conference & Exhibition – Geosciences – Investing in the Future
Saint Petersburg, Russia, 7-10 April 2014
arithmetic mean averaging algorithm but is only 0.46 % with the combined 50 Hz algorithm. The
curves filtered with the conventional algorithm remain noisy to as much as 25 %, whereas the 50 Hz
filter attenuates the noise to a minor difference from the model curve (not exceeding 0.46 %) and does
not distort the initial signal.

The results of EM-IP data processing have been verified in a similar way. Filtering with conventional
algorithms, such as arithmetic mean or weighting, is obviously ineffective: thus filtered curves show
high variance over the entire time range. However, the use of the combined 50 Hz algorithm almost
fully removes the noise and yields smooth responses at both early and late times.
In Fig. 3 the departure of different filtering results from the reference is shown in percent.

Figure 3 Percentage of EMF difference from the reference in the curves filtered with two algorithms:
DU (A), D2U (B), and P1 (C).Curve color codes are: arithmetic mean averaging algorithm (black)
and combined 50 Hz attenuation (red).
Thus, the curves processed in the conventional way differ from the model curve for 10 - 15 %, while
the difference in those processed with the combined 50 Hz attenuation does not exceed 0.75 %.
Moreover, the 50 Hz filter does not distort the initial signal.

The above results allow several inferences:

1. Up-to-date flexible TEM technique allows to conduct investigations in a various geological
situations even in highly conductive environment when total conductivity of sediments
exceeds 2000 S.
2. Conventional simple arithmetic mean averaging algorithm cannot help such a heavy noise.
But suggested combined 50 Hz filter effectively attenuates noise both in TEM and EM-IP
data, and does not distort the initial signal.
3. Synthetic 50 Hz EM noise can be simulated with variable parameters: amplitude, phase,
frequency, and number of harmonics. Adding background noise results in a combined noise
similar to that of real field conditions at the site;
4. Verification of EM data processing has shown the high efficiency of noise attenuation in
TEM-Processing software.

References

[1] Agafonov, Y.A., Sharlov, M.V., Tokareva, O.V., Buddo, I.V., Egorov, I.V., Salleh, M.M.,
Chandola, S.K., Velayatham, T., Shukri, S., Ismail, I. and Fatimah, S. [2012] Key features of
TEM and EM-IP investigations in South-East. Petroleum Geology Conference & Exhibition,
Kuala Lumpur, Malaysia. 2012.

6th Saint Petersburg International Conference & Exhibition – Geosciences – Investing in the Future
Saint Petersburg, Russia, 7-10 April 2014