Challenges in Oil and Gas Discovery

ruralrompSoftware and s/w Development

Dec 2, 2013 (3 years and 9 months ago)

111 views

Challenges in Oil and Gas Discovery

Summary

Headwave’s
Visualization

technology

has

a

big impact on the challenges in Oil and Gas
I
ndustr
y. It improves t
he quality and speed of business decisions. It

brings the
seismic
prestack domain to the
i
nterpreters
a
nd allows the processors full visualization of pre
-

and poststack.

The Headwave tools work both on a regional scale as well as on reservoir
studies.

They

provide

the security
to know

that the data has been thoroughly vetted and
that decisions made on this
data
are

well founded.
It drives increased ROI

and ROE.

Introduction

Challenges in Oil and Gas discovery are not new. The pressures and temperatures
experienced by drillers and producers have caused many headaches. With the increasing
difficulty
to
find ne
w
reserves
, the
Oil & Gas
Industry is facing the equivalent pressures

in finding seismic tools which work at depth
. Now
Seismic I
maging is struggling with
locating and viewing potential resource plays at seven to ten mile depths in the earth,
while generat
ing those images from seismic sources which are up to two miles from the
sea floor. All of this combines to make the signals sought fainter than ever before. This,
in turn, requires massive increases in processing power. Where will it be acquired?
What
kin
d of a budget is that going to take?

This
is

all addressed

by Headwave in

an

innovative
fashion
.

Deep Views with

Prestack Software

Headwave has developed the first
highly
accelerated
,

prestack visualization
and
computation
software package. While prestack
has been long used for
AvO
and RMO
analysis, the

interpretation

process
has
been used has been arduous. With multi
-
fold
gathers, interpreters have struggled to get a clear overview of what was going on. Thirty
fold gathers were barely manageable on a 24” d
isplay, but 50
-
200 fold gathers were
really out of the question. Headwave has changed that with the ability to volumetrically
view the prestack data. No matter how large the prestack dataset


we’re talking > 1
terabyte datasets


the user can access it fr
om the desktop.
Headwave’s user
-
selectable
compression delivers multi
-
terabyte datasets to the desktop and makes using the data
feasible.

Interactively work with a terabyte dataset on a standard PC workstation. See
vast swaths of data in a manner never bef
ore presented.


Figure 1. Two views of the same inline and associated prestack data. On the left a gather
is displayed and a prestack probe is also shown for the inline data. On the right, the probe
has had the shader modified to make all but the trough
peaks transparent (blue = trough)
in the probe. The nature of the far offset booming is clear, and from the gather data, a
revisiting of the muting of the prestack data might be in order.

Volumetric Visualization

Volumetric visualization means the data i
s seen as a volume in space.
Moreover, the data
is seen in context. Each gather is associated with the poststack trace

generated from the
gather. This makes it easy to relate individual prestack traces to the poststack,
additionally gather flatness (NMO qu
ality), muting

quality
, and
RMO quality
are easily
seen. It’s possible to scroll though ten MCS blocks in a single afternoon on a
n

inline

by
-
inline basis viewing each gather.
Users no longer need to avoid looking at a 1000 sq Km
region in prestack. The abi
lity to load such a dataset on a single workstation and do QC is
unprecedented.
This means that full processing QC is possible in a single day.
To do so,
the poststack data is loaded into the workspace
and linked to
the associated prestack data.
Now the pr
estack data may be displayed

in many ways:

one
gather at a time,

offset slices

or as a prestack
probe
.
Even a HyperCursor can be used to see the prestack info at a
special location. In this way,

the viewer can walk th
r
ough the prestack volume
and
inspect
a
ll traces at a given offset, or each gather in a given volume.

Having assured oneself and the group that the data processing results are as expected,
now the power of volumetric visualization can be fully utilized.
The user can easily
adjust the transpare
ncy of the prestack data volume. In doing so,
AvO
indications can be
seen. By inserting a probe into the volume, the specific regions of interest can be viewed
in more detail.
Closer inspection can lead to selections of gathers for custom stacking.
Near, m
iddle, and far stacking can be done on the fly.
Stacking can also be done on a
trace range basis.

Prestack Horizon Picking

Picking a horizon in poststack data is commonplace, it’s not so common to pick a horizon
in prestack data, and even less common to us
e a poststack horizon to seed the prestack
pick.
T
he picked prestack horizon can be used to map various attributes. Attributes such
as horizon curvature,
time delta (the difference from the time of the event at
offset
0 to
the time of the event at
offset
n
), trace count (how many traces of the gather contri
buted
to the poststack trace), G
au
s
sian curvature
,

or

others
can be derived or loaded and
mapped onto the prestack horizon.
T
he user can step th
r
ough the entire poststack volume
viewing the prestack
horiz
ons associated with that inline.

At any point the attribute being
displayed may be changed and the stepping continued.
This is an example of how the
tools available in Headwave can be applied.
Contrast this to the previous methods of
viewing prestack gathe
rs on a large dataset and it’s obvious that this is a huge leap
forward for ease of use. Prior methods required either a system with memory larger than
the prestack dataset, or required a laborious loading of 2D panes of gathers, up to 10 on
the screen if
you were lucky. The major problem: the more prestack
on screen
, the less
the poststack
overview
, and forget trying to load a 100 fold gather, let alone all such
gathers related to an inline or crossline. This is not a problem
for
Headwave. Wide
Azimuth Sur
veys aren’t a problem, nor are the new circular surveys with 200 fold
coverage.


Figure 2. Here are four 36 fold gathers and a prestack probe displayed together with the
poststack inline, two time horizons are displayed as well. This display is next to
i
mpossible in a 2D display mode.


Imperfect

P
restack
D
ata

and AvO

Suppose that the prestack

data

that
is available

isn’t perfect
, h
ow can we apply AvO
?
With the computational power now available

through Headwave
,

corrections can be
applied
, interactively,
on the desktop. Maybe additional filtering would help, or maybe it
needs a minor velocity tweak

and

a test of
the
stacking
parameters

to
optimize

the
poststack trace

or improve AvO calculations
. This is all possible.
Stacking can be done
either on the full

volume, or on a selected probe. Since the probe is a subset of the entire
data volume
,

stacking can be
performed
quickly. Sometimes the prestack data won’t have
a mute applied or the applied mute
needs adjustment
.
In Headwave

th
e mute parameters
can be te
sted
.
They can be designed on
the prestack data

gathers

and the stacked result
directly inspected
.

As previously discussed, partial stacks are

also

possible, and quickly
accomplished.


After all the
interactive

QC, and with carefully applied local correcti
ons,
Headwave
offers

extraction of

the AvO parameters

on the prestack horizons

for
crucial

studies
of reservoirs and leads.

Accelerating 4D Visualization

What’s good for 3D is even better for 4D.
Full use of
prestack 4D data lets users easily
see changes,
changes in a horizon for sure, but even more, changes in volumes.
Volumetric visualization gives “visual pop” to reservoir changes
: e
valuate
s

reservoir
response with time lapse seismic. Multiple datasets can easily be merged for
comparisons, delta calculat
ions, and more.

When serious money is spent to follow changes in the reservoir, then a serious
visualization package should give full money’s worth. If all there is to
demonstrate
a
major change in the reservoir is a single thin horizon,
then
the real deta
ils
have
been lost
.

I
s there any other data
,

or

is there another

change that’s there? With Headwave you can
see, in detail
,

the full 4D results

also with all its prestack variation
. Blending of datasets
is easy, and with transparency adjustments
,

changes i
n the reservoir are
clear. The
reservoir
boundaries are clear, and more importantly,
the volume is there. If completion is
through multiple layers, then each of those layers should be clearly visible. Make sure
that the stacking that’s done optimizes the s
ignal to noise ratio, play what if, with
stacking and filtering. Drive information examination until the current state is clear and
presentable.

To that exten
t

Headwave has also developed a Prestack Plugin for Petrel
.

Saving the Planet and the Budget

Head
wave has chosen to use the power of the floating point units in graphics processor
s

to accelerate computations. Headwave is a leader in this technology.
Both NVIDIA and
AMD/ATI now offer graphics chip derived computational engines. They are fast and
it

app
ears that not only
is
computational power
useable
, but the seamless availability of
visualization with computation provide
s

substantially shortened time to discovery.

The good news is that by going to this technology, the cost, whether $ per Gigaflop, or
Watts per Gigaflop,
has gone
down, way down
,

down by a factor of 10


30 for real
world applications in Headwave.


Conclusions

Headwave has developed novel technology for the Oil & Gas
I
ndustry based on GPU
processing.
With the advent of GPU processing on
a single graphics card or even in
clusters,
its
interactive prestack visualization and computational tools are available today
for large, very large, and extremely large datasets. While
the Headwave

tools can handle
regional datasets, they are also perfect
ly suited for smaller, field
-
wide surveys. In both
cases, Headwave tools shorten the time to target.

Headwave’s tools provide the security to know that
all

data has been thoroughly
vetted and that decisions made on this data
are

well founded. It drives in
creased
ROI and ROE.