My first
experience with a horizontal well was in a gas storage field in 1996 where I
was a mud logger. This was a well in the Oriskany Sandstone near Charleston,
West Virginia. The targeted part of the porous sandstone was clean, well-rounded,
well-sorted, and fairly unconsolidated. We got to take samples from a shale
shaker, which was rare in those days in Appalachia. Drilling was very slow when
we got horizontal, so we took samples every two feet. We were especially making
note of a glauconitic zone of characteristic green color. We correlated this
zone to a soft spot of high ROP just above the main porosity zone. We were
helping to steer the well by noting the recurring presence of this zone and the
clean sand below it. However, after a few days of this very slow drilling we
saw no glauconite and no clean sand but a more consolidated finer grained sand.
We were out of the porosity zone. Apparently, the main geologist who had come
from Texas had miscalculated the dip. Because of this the project was scrapped
since at the time short-radius horizontal drilling was less flexible in
changing inclinations. The goal of the project was to increase deliverability
from the storage field in times of high winter demand. They sought
deliverability from these horizontal wells of up to 50MMFC per day. Horizontal
wells can especially increase deliverability late in the storage draw season
when volumes and pressure in the storage reservoir are lower. This well did not
succeed due to inadequate steering but others in the field did. Other
advantages of horizontal wells in gas storage fields include the ability to
develop less favorable parts of the reservoir, reduce the level of base gas
required to pressure the working gas, allow fewer surface sites, and allow less
pipe and surface equipment.
In the late
1980’s the Cretaceous Austin Chalk was a big play in Texas and Louisiana.
Vertical drilling made occasional good wells but there were also many dry
holes. Lateral drilling through the reservoir was quite successful in some
fields but less so in others. It was thought that lateral wells could better
connect with the natural fracture networks which had a strong vertical
component. Eventually these wells came to be often steered with gamma ray in
the new technique of logging-while-drilling (LWD), or measurement-while
drilling (MWD). Incidentally, the source of Austin Chalk oil is now thought to
be the prolific Eagle Ford Shale below it, although it was once thought to be
self-sourced. Some of the horizontal gas wells in the deeper Austin Chalk trend
in Texas had phenomenal production, some exceeding 50MMCF per day in line.
During this time period the first company devoted to geosteering as a
specialty, Horizontal Solutions International (HSI), came about in the
mid-1990’s.
Horizontal
drilling of coal bed methane wells became common in the late 1990’s and early
2000’s. CBM wells in the Appalachian Basin were drilled on air, mainly in order
to de-gas the coals before mining the same seams. Drilling was very fast as
coals are quite soft relative to the rock around them. This contrast is an
advantage to geosteering as the bit tends to stay in the soft rock and deflect
off of the hard rock. Even so, properly orienting the motor was still important
to staying in the coal. In coal, in the absence of MWD gamma one can steer on
the basis of ROP. Samples can also help and with coal. Even looking at the
color of what is coming out of the flow line can verify if the bit is in the
coal or not. I once made an ROP log of a well drilling vertically through a
coal seam by watching the drill string come down past the equally spaced horizontal
metal on the drilling mast and counting how long it took. I was also running a
gas detector, so I wanted to correlate gas to depth since there was no
geolograph on that rig to correlate ROP with depth. When we ran e-logs on the
well the match was very good to my makeshift ROP log. Later, we drilled some
horizontal wells in these coal seams in Southern West Virginia. I was able to
work for a short time on-site at one of these wells in 2001 learning a bit of
geosteering by using Excel worksheets of survey data and calculating dips. The
main issue when the bit exited the coal bed was simply to determine whether it
was above or below it. That is often a main issue in geosteering today as well.
If there was a contrast in rock above and below that was certainly preferred to
a transition into similar rock character in terms of ROP, gamma, and samples.
Beginning in late
2005 the company I was working for decided to try a Devonian Lower Huron Shale
well in Southern Ohio drilled with stiff foam, which proved to be a mistake.
Vibrations affected the tools which could not get good reliable survey and
gamma readings. We sent the data to a geosteering specialist via email who then
sent back an interpretation. This continued as we drilled horizontal wells in
the Marcellus Shale beginning in 2006. I would correlate and steer with paper
logs. After a certain amount of time, I was able to rather routinely contradict
some of the interpretations of our specialist. I then devised a way to present
the data graphically. It was fairly accurate but tedious and slow. Soon
thereafter the company acquired licenses for Stoner Engineering geosteering
software (SES). I was now able to utilize it for faster correlations and
perhaps more importantly in the structurally complex Marcellus, to test several
different correlation scenarios quickly and determine the correct one. We no
longer needed our specialist.
By the late
2000’s geosteering had become well established in the shale plays and the
structural, depositional, and petrophysical variations of each play were
becoming apparent. With better well control the regional and local dips were
getting a little more predictable but in some plays like the Niobrara,
Woodford, Marcellus, and others, small faults undetectable by seismic resolution
had to be interpreted by the geosteerer, typically after a significant amount
of section had been drilled past the fault. In addition, since the survey and
gamma were typically close to 50 ft. behind the bit this means that by the time
the post-fault correlation could be determined a significant amount of hole was
necessarily drilled out of zone before adjustments could be made. In some cases,
it was worse in the Marcellus as high dips, sometimes very high dips up to 45
degrees, could be encountered near reverse faults, or high-angle thrusts. These
zones were typically avoided but sometimes they were not seen on old 2D seismic
and sometimes the curve was drilled in the fault zone to try and maximize
lateral well length between the large fault zones. This made geosteering quite
challenging at times as unexpected dip changes and folds were not uncommon.
Some plays and areas of plays are relatively calm geologically. In those areas
it is of course easier to maximize in-target footage even in small target
windows.
By 2007 around
400 rigs were drilling horizontally in multiple shale plays: mainly the
Barnett, the Bakken, the Bossier and Haynesville, the Fayetteville, the
Woodford, and the Marcellus. The Niobrara, Eagle Ford, Mississippi Lime, SCOOP,
STACK, Tuscaloosa Marine Shale, Utica-Point Pleasant, Burket, and multiple
Permian plays would come later. Dr. Mike Stoner, the creator of SES software
defined some of the terms and parameters of “technical geosteering.” One is the
idea of Relative Stratigraphic Depth (RSD) which he defined as “stratigraphic
distance relative to an “arbitrary” reference point [or marker bed]” This
software and others can stretch or squeeze a gamma or other MWD log section to
match a control well TVD log and/or the adjusted TVD log just landed in the
well being drilled. No software can be perfect as it difficult to account for
changes in stratigraphic thickness of different zones over the length of a
lateral, for deflection off hard zones which can temporarily distort the
inclination and make it appear that the dip has changed, and differences in
gamma character due to inadequate sampling rates as a result of very fast
drilling. A good geosteerer can see through these issues and adjust
accordingly, thus optimizing the effectiveness of the software.
By around 2010
geosteering had taken four main forms. 1) Dedicated geosteering centers were
established where geologists typically work 12-hour shifts, sometimes on
multiple wells. The expense of the software made this economic as software
license keys could be optimized. 2) There was on-site geosteering, sometimes
combined with mudlogging. As access to software increased and geosteering was
in demand more geologists were learning geosteering, some while learning
mudlogging. This was probably not wise at the time as that is lot to manage for
a young geologist, although there are likely some outfits that with experience
do fairly well at it now. Some plays may be better geosteered on-site. One is
perhaps the Bakken as I have heard but I have no experience with it. Plays
where gamma is less distinct and where ROP and samples are more definitive, are
probably good candidates for on-site geosteering. 3) Remote, or off-site
geosteering has also been a successful format. The advantages to remote
geosteering include avoiding travel expenses and travel times, avoiding another
body on location, and avoiding the distractions of the rig environment. With
WITSML well information systems such as Pason, all of the important parameters
of the well can be monitored in real-time from anywhere. Phone apps make these
portals quite portable as well. 4) In-House geosteering has some advantages if
the geologist is involved in updating structure maps, isopach maps, in
well-planning, or needs a second opinion from another geologist who may be present.
Due to the necessity of 24-hour coverage
in-house steering may also require a remote geosteering component for
off-hours. Sometimes in-house geologists will follow along geosteering along
with field or remote geosteerers for duplicate coverage and cross-checking,
which is not a bad idea, especially in structurally complex plays.
Other potential
innovations involve utilizing resistivity logs where gamma ray is indistinct or
has very little variable character. This can be important in some play, mainly
carbonates. It has been noted that in the Utica-Point Pleasant shale play in
Ohio that imprecise steering may lead to less optimization of preferred target
section as well as missing the structure along the well path which may affect
well production. Another technique of chemostratigraphy – correlating
geochemically by powdering samples to determine elemental constituents by
XRD/XRF (X-Ray diffraction and X-ray fluorescence). While this is potentially
useful in some circumstances it is also slow and not very amenable to being
useful at current ROPs. The technique itself is quite useful and probably has
better application to compare wells after they are drilled. Another method is
detailed Mass Spectrometry to tag specific mud gas signatures to specific
zones. Yet another technique that may have niche applications in complex
trapped conventional reservoirs, offshore reservoirs, and faulted
unconventional reservoirs is seismic-while-drilling (SWD). The main goal of SWD
is to image what is ahead of the bit. Schlumberger has used the technique quite
a bit but not much is known about its success rate. I had heard years ago of
the idea of using resisitivity to see ahead of the bit when attached to a BHA
close to the bit. If I recall correctly it was because it penetrates deeper
than other logs and can be pointed ahead of the bit. Another new technique I
just read about is Baker's VisiTrak which utilizes "extra deep reading
azimuthal resistivity..," in combination with omnidirectional sensors
(360deg) that can image bed boundaries up to 100ft ahead of the bit. The
sensors are very close to the bit attached to the BHA with a rotary steerable
motor.
All geosteerers
make correlation mistakes. The likelihood of making a mistake can be influenced
by several factors: gamma quality, look-alike zones, unexpected large dip
changes, a different expectation of geology, and most commonly – similar gamma
character above and below a known well-correlated zone. More well control leads
to refined structure maps and thus better predictability of regional dip. Pad
wells drilled parallel are typically easiest to steer but can have some
variability. Some of that variability has to do with whether the wells are
drilled perpendicular, oblique, or parallel with regional dip, as well as
changes in structure. Distance from the control well is important as thickness
changes relative to that well can be quite significant. This is more common in
the curve which traverses far more stratigraphic distance than a target window.
It can also be a factor in the target interval but there is usually not more
than a foot or two variation. Even so, this can be a factor when drilling a small
target window in areas of depositional thinning on marine shelves. Sometimes
deflections off of hard zones or depositional variations can look like small
faults or sometimes they may actually be small faults. Thus, I think it is
difficult to determine for sure if offsets less than about 3-4 feet or so are
actual faults or false faults. If a similar offset is encountered near the same
VS in one or more parallel well(s) then it is more likely a fault. In such
cases the actual orientation (azimuth) of the fault can be determined. If a
large gas show is encountered without changing zones and without changes in mud
properties (weight, viscosity, water loss, cuttings size and shape) then it is
possible that a significant gas-filled fracture was encountered.
There are also psychological
and communication issues in geosteering. Since the process of geosteering
involves frequent interpretation of data in constantly changing conditions that
may require adjustments there are times when correlations are uncertain and
tentative. Communicating degree of certainty is important. Considering
reliability of regional dip is important as well. From my experience it is best
to have frequent contact with the directional drillers, MWD personnel, and
sometimes the Drilling Engineer, and occasionally the mudloggers for sample and
gas verifications, either in person, by phone, or through email. Email is
usually fine but if drilling out-of-zone seems possible then a phone call or
two can be warranted. Some directional drillers have particular styles of
drilling and don’t like to change inclination. Doglegs need to be considered.
Tolerances and protocols for changing inclination and by how much need to be
discussed. Structurally complex areas can make such required communication more
frequent than one might like but sometimes that is the nature of the beast.
Geosteering
results can be used to determine the amount of what I call primary
reservoir access (PRA), which is access to the most desirable reservoir
rock. If that rock is “hot” shale, then footage and percentage of the wellbore
in the hot shale can be calculated. In some source-rock mudstone reservoirs
there are thin beds of varying gamma and interbedded limestones even in small
target intervals. In such cases one may determine what I call – relative
optimized reservoir access (RORA)– which is basically how much of the
wellbore is in the best part of the target interval. Primary reservoir access
(PRA) can be compared to the microseismic delineation of stimulated reservoir
volume (SRV). With microseismic studies thin stratigraphic intervals defined by
geosteering correlations can be compared in terms of induced fracture
propagation, tendency for frac height growth, and frac half-length. Target
intervals can be tweaked or optimized on the basis of such studies.
Perhaps the most
desirable setup for geosteering, though exceedingly rare, would include gamma
ray and resistivity logs, both very close to the bit with a rotary steerable to
keep the well from moving up and down too much through the formation. Plays with
faults and/or high dips are most amenable to having gamma MWD and surveys close
to the bit in order to minimize out-of-target drilling. SES software has
improved over the years and now incorporates a nice cross-section feature that
accommodates mud logs, ROP logs, and other data. It can also accept the grid
data that makes up structure maps and seismic sections. Newer geosteering
programs like ROGII’s StarSteer can accommodate seismic lines, image logs, and
other nice graphical features.
Another current issue
with the ubiquity of multi-pad wells is wells with large turnouts for proper
spacing. This can create some distortions during drilling of the curve, but
they are typically quite manageable. A related issue is the need the change
azimuth to keep required spacing from an unleased property. Yet another is to
avoid so-called “no tag zones” near lease lines. This is an issue in
Pennsylvania where the base of the Marcellus target zone may be a mere two feet
from a no-tag zone (Onondaga Limestone) although the true rationale for this
rule is impractical and potentially unfairly punitive to an oil and gas
operator, especially if an unexpected fault, dip, or inclination would occur.
Horizontal and
directional wells requiring some degree of geosteering are also utilized in
deep geothermal drilling, where MWD tools and mud motors need to be able to
endure high temperatures for long periods. The deep, high temperature/high
pressure Haynesville Shale and Bossier plays and other oil and gas plays also
require HPHT tools. When I geosteered Haynesville wells we kept a temperature
log on our reports showing the formation often at around 300 deg F. Geosteering
may also be required in carbon sequestration wells where horizontal wells
increase both the CO2 injectivity and the CO2 storage capacity of deep brine
reservoirs. In newer closed-loop geothermal, two wells may be geosteered in
order to be connected. Typically, they are geosteered and when they get close,
they are steered to connection with magnetic ranging technology, which is more
precise with closer targets. Injection wells drilled for hot oil treatment in
the Canadian oil sands also utilize geosteering. Some oil and gas wells employ
multi-laterals where new laterals are sidetracked off of previous laterals usually
to drill a different zone. Horizontal drilling for near-surface pipeline and
utility road and stream crossings generally don’t require geosteering, although
soil horizons and rock bed dips and boundaries should be known and documented.
Recently, as the
oil & and gas industry has been contracted on and off due to low commodity
prices and increased efficiency, the need for day-rate field and operations
people has been reduced. This has created an overhang of qualified field people
that will likely stick around to some extent even as prices recover. Thus,
there is now much competition in terms of people and price. Success rates
driven by experience and effective workflows are very important in tight
margins.
Since I wrote
this short history there have been a few new developments in geosteering. Some
MWD companies provide continuous inclination data which may give suggestions
whether the bit is deflecting from a hard zone which can give a slightly
erroneous inclination. They may also provide gamma-up and gamma-down curves
which can indicate whether one is above or below a zone. I have found these to
be effective maybe 60% of the time. The problem with them is that there may be
thin beds of higher or lower gamma that only show up if the well path is very
close to dip, In any case, they are quite useful when there is uncertainty.
Many wells are now drilled with rotary steerable motors with survey and MWD data
closer to the bit so less delay means less out-of-target footage. Drilling is
consistently fast in most formations.
There are different methods of geosteering that may be more or less suitable for different plays or regions of plays. In plays or regions of plays where stratigraphy is flat or where dip is has little variability, it may be better to steer by waypoints or by calculating a spread of the stratigraphic target from a previously defined drilling target and then adjusting when the spread gets to a certain magnitude. This works where there is little change in dip. It has worked for me in the Haynesville play where a single correlation dip line may last for 500 to 1500ft. However, in plays like the Marcellus, the dip often changes in addition to changes in target interval thickness or thickness relative to type wells so that there are often several apparent dip changes within a hundred feet. In that case, we always recommend an average dip to maintain. Sometimes when drastic dip changes are expected it is wise to anticipate them by moving to one end of the target window or even out of the target window to avoid going out the other end when the dip changes in the case of the largest dip changes. Frequency of reporting depends a lot on the expected geology and may change within the well dependent on dip changes and/or faulting.
ROGII’s Star Steer has taken quite a bit of the geosteering software market share. It is a good program that is easy to use, fast to set up and adapt, and has quite a few interesting features and “bells and whistles.” Some companies now also set up geosteering reports on the cloud with automatic downloading of WITSML data straight from the rig. This saves time and gets correlations up faster. It also allows the geosteerer to focus more on correlation and less on typing, copy-pasting, and emailing. Wells today are drilled quite fast and turnaround time can be important.
Better than gamma up and gamma down in my opinion is the multitrack azimuthal gamma ray log curve which is presented as an image log to interpret whether the bit is moving up section or down section. A dynamic image gamma ray can also be generated to see if an interpretation best matches the azimuthal gamma. I have been learning the value of this log in delineating gamma lobes that look alike in the Austin Chalk of South Texas.
References:
Geosteering
Keeps Drillers on the Right Track – by Louise S. Durham, in A.A.P.G. Explorer,
Dec. 2012
Technical
Geosteering Finds the Sweet Spot – by Dr. Mike Stoner, in E & P Magazine,
November 2007, pgs. 71-77
Carbon
Dioxide Injectivity in Brine Reservoirs Using Horizontal Wells – multiple
authors, National Energy Technology Lab – U.S DOE
ENEL’s
Experience With Directional Geothermal Wells – by Bianchi, Quintavalle, and
Rossi (Italy)
Horizontal
Drilling Used in Gas Storage Programs – by Young, McDonald, and Shikari (Gas
Research Institute), in Oil & Gas Journal, April, 5, 1993.
Geosteering
as Research – by Louise S. Durham, in A.A.P.G. Explorer, Dec. 2013
Seismic
While Drilling Moves Closer to Reality – by David Brown, in A.A.P.G. Explorer,
Dec. 2013
VisiTrak
Services May Eliminate Need for Pilot Wells in Wellbore Placement - by Alex
Endress, in Drilling Contractor, drillingcontractor.org, 2016.
Guide
to Geosteering Series - Parts 1,2,3, and 4 - by Patrick Tobin, posted in
OilPro, Feb., 2016
No comments:
Post a Comment