Diffraction imaging for basement fault-fracture prediction: Application to an oil field in Cuu Long basin

PETROLEUM EXPLORATION & PRODUCTION

PETROVIETNAM JOURNAL

Volume 10/2020, p. 4 - 11

ISSN 2615-9902

DIFFRACTION IMAGING FOR BASEMENT FAULT-FRACTURE

PREDICTION: APPLICATION TO AN OIL FIELD

IN CUU LONG BASIN

Ta Quang Minh, Nguyen Danh Lam, Duong Hung Cuong, Pham Van Tuyen, Mai Thi Lua, Pham The Hoang Ha

Vietnam Petroleum Institute

Email: minhtq@vpi.pvn.vn

https://doi.org/10.47800/PVJ.2020.10-01

Summary

Improvement to the image of fractured granite basements is among the most sought-after goals for processing seismic data in

Cuu Long basin, the most proliferous petroleum basin. Unlike a clear layering structure of the sediment, fuzzy images of the granite

basement are often the source of confusion for interpreters to identify which structures are presented inside it. In such a low signal-to-

noise ratio(SNR) environment, extracting geological information such as fault systems and fracture becomes challenging. In this study,

diﬀraction imaging is employed in an eﬀort to identify and enhance the fault system inside the basement. The comparison of the study

result with various standard post-stack attribute approaches shows the eﬀectiveness of the diﬀraction imaging method.

Key words: Seismic processing, seismic imaging, seismic diﬀraction, faults inside basement.

1. Introduction

In order to determine and identify fault systems from

The immediate challenge is the identiﬁcation of diﬀraction

signals (normally much weaker than reﬂection signals) in

the usually low signal-to-noise ratio (SNR) environment

of the basement, which directly impacts the possibility to

perform diﬀraction imaging and identiﬁes fault structure

from the results.

seismic data, traditional approaches usually use post-

stack processing to generate geometrical attributes that

correlate to the characteristics of faults/fractures [1]. Most

of the approaches include coherence-based methods [2],

structure analysis [3] and curvature-based approaches

[4]. The major assumption of these approaches is that

the data contains images of “reﬂectors”, and each method

identiﬁes either region of lacking reﬂections or the

curvature/geometry of the reﬂections. These approaches

work very well for sedimentary zones due to the layering

nature of the sequence stratigraphy. However, when

using the same approaches for the basement, the

assumption is either weak or no longer valid. A better

approach to study fault systems in the basement is to

directly identify pre-migrated seismic events associated

with the discontinuities in the medium, which, under

seismic excitation, should generate diﬀraction waves.

Thus, diﬀraction imaging techniques emerge as a possible

direct approach to study the fault system in the basement.

In this study, the 3D seismic acquisition dataset for

a small ﬁeld in East Cuu Long basin (Figure 1), covering

oil producing basement, was employed for the imaging/

special seismic attribute study. The seismic stack data

quality (PSDM and CBM) was too fuzzy inside the

basement to identify any fault system. Seismic data was

reprocessed using the diﬀraction imaging workﬂow and

the diﬀraction signatures were compared to the well data

(including FMI). The very encouraging results from the

study showed that diﬀraction imaging provided a much

improved image of fault systems near and inside the

basement, especially in comparison with those coming

from various traditional post-stack attribute approaches.

2. Theory and methodology

The mechanism of post-stack seismic attributes for

fault prediction involves either detecting the discontinuity

of the reﬂection features (based on the geometrical

analysis of continuous structural events) or identifying

Date of receipt: 19/8/2020. Date of review and editing: 19/8 - 18/9/2020.

Date of approval: 18/9/2020.

PETROVIETNAM - JOURNAL VOL 10/2020

4

PETROVIETNAM

anticline features (based on the curvature of the events). These detections

are to determine the fault/fracture signature indirectly (by the absence

of continuity of reﬂection planes or the calculation of their geometrical

curving properties) and the approach probably works well for sediment

layers. On the contrary, the basement might not be the best place for post-

stack attribute fault detection because the continuity of reﬂections is not

apparent.

Stemming from the works of

diﬀraction wave characterisation [6],

a new form of seismic image called

diﬀraction imaging [7], tries to separate

diﬀraction wave from the reﬂection

wave before the ﬁnal seismic migration.

The separation may come in the form

of detecting continuous reﬂection

event and remove it from the seismic

using structure-oriented ﬁltering such

as plane-wave-destruction (PWD) [8],

or multi-focusing [9]. Another way

to accomplish the goal is to embed

the reﬂection-diﬀraction separation

as a part of the migration itself (such

as side view seismic location - SVSL

[10], velocity continuation method

Tertiary basement depth structure

Quarternary sediments

Neogene-Quarternary basalt

Late Jurassic-Cretaceous

volcanics

Late Jurassic-Cretaceous

plutonic rocks

Early-Middle Jurassic sediments

Triassic sediments

[11] or local angle domain [12]).

A

seismic cube of diﬀraction index is

the result of such separation process.

This cube of diﬀraction index can be

further enhanced with post-stack fault

prediction routines such as ant-tracking.

Devonian-Carboniferous

Cambrian-Ordovician

Inferred major

Inferred faults

Figure 1. Tertiary basement map of Cuu Long basin and the study [5].

Forthisdiﬀractionimagingstudy,we

employed the plane-wave-destruction

ﬁltering approach of Fomel [8] for

waveﬁeld separation in combination

with the velocity continuation imaging

[11]. Summary of a diﬀraction imaging

workﬂow (highly simpliﬁed) employed

at VPI is shown in Figure 2. Here, the

diﬀraction wave generated by a half

plane model is properly separated and

imaged just as described by Landa [7].

The improvement to the diﬀraction

imaging of the actual ﬁeld data is

possible by pre-stack processing and

other advanced migration techniques

(proprietary right of VPI). However, even

with the crude approach of velocity

continuation imaging, we can observe

the diﬀraction signature of fault systems

with very high conﬁdence.

Wavepropagation

Geological model

Premigration

processing

Common oꢀset sections

Diꢀrac

t

ion/Reꢁection

separation*

Reꢁection Signal

DiꢀractionSignal

Image Construction

(Migration)

Diꢀraction-Oriented

Image Construction*

Post-Processing*

Again, we note that: 1. The lack

of continuous reﬂection events in the

basement makes the study of diﬀraction

highly important. 2. Unlike post-stack

attributes, diﬀraction imaging attribute

Reꢁection Image

DiꢀractionImage

Figure 2. A simpliﬁed diﬀraction imaging workﬂow (* are VPI proprietary processing routines).

PETROVIETNAM - JOURNAL VOL 10/2020

5

PETROLEUM EXPLORATION & PRODUCTION

cubes are the direct indicator of discontinuous events

(faults, fractures). 3. The amplitude of diﬀraction signal

is several orders of magnitude less than the reﬂection,

making the method applicability questionable for the

basement, hence the real data of an oil ﬁeld needs to be

checked for eﬀectiveness.

The seismic data employed by the study include post-

stack data (PSDM) to generate the traditional attributes, as

well as the pre-migrated seismic (slightly preconditioned)

to be input for the diﬀraction image processing. Results

of the processing and comparison between the two

approaches are presented next.

3. Application to a basement ﬁeld data in Cuu Long ba-

sin, oﬀshore Vietnam

3.2. Initial diﬀraction imaging results in the sediment

near top basement

3.1. Studied area and the database

As a test case, a sedimentary section is examined

(Figure 3), where large faults can readily be seen and

interpreted from the seismic sections. Diﬀraction imaging

indicators show a good match between those of high

amplitudes and the seismic-stack-based interpreted

faults. A time-slice through a region with faults also

emphasises the fact that the diﬀraction imaging indicator

can highlight individual faults whereas those from the

stack are not so clear.

The target of the diﬀraction study is the granite

fractured reservoir of an oil ﬁeld situated to the east of Cuu

Long basin (Figure 1). Map of the top Tertiary basement

in this area indicates some horsts divided by E-W faults.

These horsts show the separation with adjacent grabens

by NE-SW faults, many of them are thrust faults which

are considered as the results of inversion activity in

Oligocene. The conventional seismic allows to conﬁdently

deﬁne only some fault systems with large displacement

observed at the top of the basement. There are several

wells penetrating inside the basement conﬁrming the

very good oil ﬂow in this reservoir. However, dry wells in

the same anticline structure imply the oil and gas might

be accumulated in diﬀerent special spaces that are mainly

controlled by the faults/fractures inside the Tertiary

basement since the granular porosity is considered

negligible.

3.2.1. Intra-basement analysis at a well location

FMI data from three wells indicated fault systems with

very high dip angle (56 - 88^o). Most of the FMI indicates the

occurrence of complex conjugate faults. Among them, a

vertical well (Well #1) is employed to illustrate the study.

Figure 4 indicates seismic proﬁle (PSDM) going

through exploration Well #1 superimposed with the faults

from FMI. By FMI and well information (from the FMI and

b

c

a

PSDM

PSDM + diﬀraction

Diﬀraction

d

f

e

Diﬀraction -

PSDM - Timeslice

2550ms

Timeslice 2550ms

Figure 3. Diﬀraction imaging in the sediment section. (a) PSDM stack (b) Diﬀraction index (c) Diﬀraction index coincides with the identiﬁed faults on the overlaid PSDM section. Time slice

of (d) PSDM cube over a region with faults (e, f) Diﬀraction image with fault interpretation.

PETROVIETNAM - JOURNAL VOL 10/2020

6

PETROVIETNAM

well reports), at least three zones have been identiﬁed related to faults

and fractures. There are three fault zones interpreted from diﬀerent

depth levels including Zone #1 (~3,725 - 3,860 mTVDSS), Zone #2

(~4,320 mTVDSS) and Zone #3 (~4,480 mTVDSS).The ﬁrst zone contains

two main fault systems, ENE-WSW striking sets conjugate with striking

N-S sets. The second zone is characterised

by a NW-SE strike with the dip of 73^o. Two

conjugate faults are imaged by FMI data with

strike of NE-SW and NW-SE in the Zone #3,

the dips of these faults range from 31^oto 39^o.

In Figure 4, the faults identiﬁed by FMI in each

zone are displayed by diﬀerent colours.

Top of basement

34.69

Examining the conventional PSDM

section, hints of the faults inside the

basement can be barely spotted (Figure 4),

however, the interpretation could be quite

ambiguous. In other words, most of the fault

zones are not quite observable from the

standard (reﬂection) seismic section.

Dip of fault

from FMI

76.6

Zone#1

73.18

Zone#2

3.2.2. Traditional approaches - standard post-

stack attributes to identify faults

34.69

31.35

39.7

Fault zones from FMI

Previous post-stack attribute study has

been carried out to identify the fault system

(Figure 5) where geometrical extraction

properties (such as chaos and variance) are

used before using the “ant-tracking” routine

Zone#3

Well#1

Figure 4. FMI data indicates the characters of fault system inside the basement.

a

b

c

Well#1

Ant tracking (variance)

PSDM

Variance

e

d

f

Well#1

Chaos

Ant tracking (Diﬀraction)

Diﬀraction

Figure 5. (a) Seismic proﬁle (same with Figure 4) and several popular seismic attributes including (b) variance (c) ant-tracking (d) chaos and (e) diﬀraction

and (f) ant-tracking of diﬀration data.

PETROVIETNAM - JOURNAL VOL 10/2020

7

PETROLEUM EXPLORATION & PRODUCTION

Well#1

Zone 1

Zone 2

Zone 3

Dip-azimuth calculated from

diﬀraction data

Diﬀraction

Ant-tracking diﬀraction

Original seismic

Figure 6. Seismic proﬁle going through vertical Well#1. The faults based on FMI in three zones are demonstrated by colour sticks. Rose diagrams indicate the strike and dip faults based on

the diﬀraction data.

to extract the fault system. Most of the attributes show the

evidence of uppermost faults (yellow and purple lines),

however these attributes indicate neither clearly (chaos,

variance) nor fully this fault (ant-tracking). A mismatch

between the dip, azimuth of the other estimated faults

and the actual values from the FMI logs indicates an

unsuccessful attempt. Also, many vertical events appear

with the dip of nearly 90^owhich is not suitable with the

observation from the FMIs.

by the large reﬂection seismic phase, hence the faults

become invisible. On the contrary, the strong reﬂector

related to the top of the basement is removed in

diﬀraction data and the conjugate alignments are shown.

In the ﬁrst zone, the diﬀraction and ant-tracking from this

data showed quite well the two alignment systems with

E-W orientation, the NW-SE system was quite faint. In the

second zone, the diﬀraction mainly depicts faults of E-W

direction and a small part of N-S direction, the instruction

at FMI was mainly for NW-SE direction. Third zone’s

diﬀraction index clearly shows two NW-SE & conjugate

systems which is similar to FMI.

The reﬂector associated with the top basement

disappeares in the diﬀraction section, this dataset

describes several oriented events. There are two portions

around well’s location, the ﬁrst part (yellow rectangular)

witnesses the dominance of anomalies which shares the

same trend with fault Zones #1 and #2. The conjugates

of fault Zone #3 are mapped in the second part (white

rectangular).

3.3. Comparisons of post-stack attributes and the

diﬀraction index on regions beyond wells

Figure 7 shows sections with two proven faults

(yellow dash lines) conﬁrmed by the movement of seismic

phase observed (on the seismic section) at the top of the

basement and nearby wells. Standard attributes (variance,

chaos, and positive curvature) are computed but they all

show an ambiguous image of faults inside the basement,

implying poor SNR in conventional seismic images.

3.2.3. Fault identiﬁcation using the diﬀraction index

Figure 6 compares the section of an exploration well

versus PSDM seismic, diﬀraction data and the ant-tracking

based on the diﬀraction data.

Events on the variance and chaos attribute section

look like some localised spots without any systematic

In the PSDM section, the top of the basement is tied

to a strong reﬂector, all FMI-based faults are overpowered

PETROVIETNAM - JOURNAL VOL 10/2020

8

PETROVIETNAM

Chaos

Positive curvature

PSDM

Ant-tracking variance

Variance

Diﬀraction

Figure 7. Comparison among standard seismic attributes, two dash lines indicate some proven faults.

a

Diﬀraction

PSDM

c

b

Diﬀraction

Ant-tracking diﬀraction

Figure 8. (a) Diﬀraction section (b) with interpreted fault and (c) ant-tracking of diﬀraction index.

coherence to indicate fault features.The positive curvature

section provides better lineament highlighted events

related to the fault on the left but not the one on the right.

Vertical events with the dip of nearly 90^oare still present

and diﬃcult to explain from the geological perspective.

Figure 8 shows the diﬀraction section, where the

presence of at least two conjugated fault systems are

observed. The indices have good coherency that can link

together to form a fault. The fault on the right is clearly

shown on the diﬀraction section. More illustration of the

PETROVIETNAM - JOURNAL VOL 10/2020

9

PETROLEUM EXPLORATION & PRODUCTION

PSDM

Time-slice 2900ms

Diﬀraction

Figure 9. A time slice through the basement segment to compare the PSDM and diﬀraction image.

eﬀectiveness of the diﬀraction approach vs common

migrated section (lack of fault signature) is demonstrated

in Figure 9, where a time-slice extracted from diﬀraction

shows clear evidence of coherence faults, which is

invisible on PSDM slice.

4. Conclusions

Diﬀraction imaging is a new approach of extracting

wave-related information from seismic data (excluding

reﬂectionsignals),whichiscrucialinimagingdiscontinuous

geological events including faults and fractures. Given the

low SNR of the seismic signal inside the basement, a crude

calculation diﬀraction imaging eﬀort was carried out. The

results highlight the fact that it is possible to improve

the image of the fault system inside the basement using

diﬀraction imaging, which is conﬁrmed by the well data

and seismic phase displacement. Further post-stack

enhancement to the diﬀraction imaging technique will

certainly improve images for reliable prediction of the

fault/fracture system within the granite basement.

Figure 10 shows the sections from some more

advanced post-stack seismic attributes including similarity,

instantaneous dip, Kingdom’s similarity variance and

dip variance, which appear to show better fault features.

However, some of the faults that can be readily seen/

identiﬁed from the seismic stacks and the diﬀraction section

are still missing from the attributes. Moreover, since these

advanced post-stack seismic attributes are still indirect

indicators of faults, their amplitude do not carry the same

interpretable property as the amplitude of the diﬀraction

index, which may connect to the petroleum potentials.

PETROVIETNAM - JOURNAL VOL 10/2020

10

PETROVIETNAM

Similarity Variance

Instantaneous Dip

PSDM

Similarity

Dip Variance

Ant-tracking diﬀraction

Figure 10. Comparison of some advanced post-stack seismic attributes. Two yellow dash lines are some proven faults.

References

[7] E. Landa, V. Shtivelman, and B. Gelchinsky, "A

method for detection of diﬀracted waves on common-

oﬀset sections", Geophysical Prospecting, Vol. 35, No. 4, pp.

359 - 373, 1987. DOI: 10.1111/j.1365-2478.1987.tb00823.x.

[1] Satinder Chopra and Kurt J. Marfurt, Seismic

attributes for prospect identiﬁcation and reservoir

characterization. Society of Exploration Geophysicists,

2007. DOI: 10.1190/1.9781560801900.

[8] Sergey Fomel, "Applications of plane-wave

destruction ﬁlters", Geophysics, Vol. 67, No. 10, pp. 1946 -

1960, 2002. DOI: 10.1190/1.1527095.

[2] Mike S. Bahorich, John Lopez, Norman Lopez,

E. Nissen, and Alan Poole, "Stratigraphic and structural

interpretation with 3-D coherence", SEG Technical Program

Expanded Abstracts 1995. DOI: 10.1190/1.1887435.

[9] Boris Gelchinsky, Alexander Berkovich, and

Shmariahu Keydar, "Multifocusing homeomorphic

imaging. Part 1: Basic concepts and formulas", Journal of

Applied Geophysics, Vol. 42, pp. 229 - 242, 1999.

[3] Gijs C. Fehmers and Christian F.W. Hocker, "Fast

structural interpretation with structure-oriented ﬁltering",

Geophysics, Vol. 68, No. 4, 2003. DOI: 10.1190/1.1598121.

[10] O.L. Kouznetsov, I.A. Chirkin, I.S. Faizulline,

B.Y. Meltchouk, I.S. Dzhafarov, Y.A. Kuryanov, V.N. Nesterov,

S.I. Slenkin, G.V. Kashirin, Z.H. Mollaev, and A.P. Kozub,

"Improving the eﬃciency of seismic exploration by

applying 3D mapping of open fracturing by the SVSL

method", Conference Proceedings, 64^thEAGE Conference

& Exhibition, Florence, Italy, 27 - 30 May, 2002. DOI:

10.3997/2214-4609-pdb.5.G038.

[4] Richard J. Lisle, "Detection of zones of abnormal

strains in structures using Gaussian curvature analysis",

AAPG Bulletin, Vol. 78, No. 12, pp. 1811 - 1819, 1994.

[5] William Schmidt, Bui Huy Hoang, James W.

Handschy, Vu Trong Hai, Trinh Xuan Cuong, and Nguyen

Thanh Tung, "Tectonic evolution and regional setting of

Cuu Long basin, Vietnam", Tectonophysics, Vol. 757, pp. 36

- 57, 2019. DOI: 10.1016/j.tecto.2019.03.001.

[11] Luke Decker, Dmitrii Merzlikin, and Sergey

Fomel, "Diﬀraction imaging and time-migration velocity

analysis using oriented velocity continuation", Geophysics,

Vol. 82, No. 2, 2017. DOI: 10.1190/geo2016-0141.1.

[6] K.D. Klem Musatov and A.M. Aizenberg, "The

ray method and the theory of edge waves", Geophysical

Journal of the Royal Astronomical Society (Geophysical

Journal International), Vol. 79, No. 1, pp. 35 - 50, 1984. DOI:

10.1111/j.1365-246X.1984.tb02839.x.

[12] Igor Ravve, Zvi Koren, Allon Bartana, and Dan

Kosloﬀ, "Local angle domain in seismic imaging", 69^thEAGE

Conference and Exhibition incorporating SPE EUROPEC, 11 -

14 June, 2007. DOI: 10.3997/2214-4609.201401932.

PETROVIETNAM - JOURNAL VOL 10/2020

11