Nuclear Magnetic Resonance Logging While Drilling

Document Sample
Nuclear Magnetic Resonance Logging While Drilling Powered By Docstoc
					                      Nuclear Magnetic Resonance Logging
                      While Drilling

                      Innovative drilling and measurements technologies now provide increasingly
                      comprehensive borehole and formation-evaluation data in real time. Recent
                      developments in nuclear magnetic resonance logging while drilling are helping
                      operators make more informed drilling and completions decisions, reduce risk and
                      nonproductive time and optimize wellbore placement and productivity.

R. John Alvarado      Nuclear magnetic resonance (NMR) logging while                    In this article, we review basic NMR con-
Houston, Texas, USA   drilling (LWD) represents a significant advance-                cepts, introduce developments in NMR logging
                      ment in geosteering and formation-evaluation                   while drilling and discuss how operators are
Anders Damgaard       technology, bringing the benefits of wireline NMR               using this technology for wellbore placement and
Pia Hansen            to real-time drilling operations. Critical petro-              formation evaluation in real time.
Madeleine Raven
                      physical parameters, such as permeability and
Maersk Oil
                      producibility estimates, can now be obtained                   Development of Wireline NMR
Doha, Qatar
                      while drilling, providing information that helps               In the decade that NMR logs have been available,
Ralf Heidler          petrophysicists, geologists and drillers achieve               they have undergone continual improvement.2
Robert Hoshun         optimal wellbore placement within a reservoir.                 The CMR Combinable Magnetic Resonance tool
James Kovats              Real-time while-drilling measurements are                  family, beginning with the introduction of the
Chris Morriss         especially important in high-cost and time-                    CMR-A service in 1995, provided measurements
Sugar Land, Texas     sensitive drilling environments. With rig costs                of effective porosity, bound-fluid volume (BFV),
                      running as high as USD 175,000 per day,                        permeability and T2 distributions, a concept
Dave Rose             errors in well placement, formation evaluation                 described later in this article. The CMR-200
Doha, Qatar           or well-completion design can result in                        Combinable Magnetic Resonance tool introduced
                      significant additional well costs or the drilling               advances in electronics that provide an increased
Wayne Wendt
                      of expensive sidetracks.1                                      signal-to-noise ratio (S/N) while shorter echo
Houston, Texas
                      For help in preparation of this article, thanks to Emma Jane      and Wright B: “Avoiding Drilling Problems,” Oilfield
                      Bloor, Jan Morley, Marwan Moufarrej and Charles                   Review 13, no. 2 (Summer 2001): 32–51.
                      Woodburn, Sugar Land, Texas, USA; Kevin Goy, Doha,             2. Kenyon B, Kleinberg R, Straley C, Gubelin G and
                      Qatar; Mohamed Hashem, Shell, New Orleans, Louisiana,             Morriss C: “Nuclear Magnetic Resonance Imaging—
                      USA; Martin Poitzsch, Clamart, France; Joe Senecal,               Technology for the 21st Century,” Oilfield Review 7, no. 3
                      Maersk Oil, Doha, Qatar; and Brett Wendt, ConocoPhillips,         (Autumn 1995): 19–33.
                      Houston, Texas.
                                                                                        Allen D, Crary S, Freedman B, Andreani M, Klopf W,
                      CMR (Combinable Magnetic Resonance), CMR-200,                     Badry R, Flaum C, Kenyon B, Kleinberg R, Gossenberg P,
                      CMR-Plus, IDEAL (Integrated Drilling Evaluation and               Horkowitz J, Logan D, Singer J and White J: “How to
                      Logging), MDT (Modular Formation Dynamics Tester),                Use Borehole Nuclear Magnetic Resonance,” Oilfield
                      PowerDrive, PowerPulse, proVISION and VISION are                  Review 9, no. 2 (Summer 1997): 34–57.
                      marks of Schlumberger.
                                                                                        Allen D, Flaum C, Ramakrishnan TS, Bedford J, Castelijns K,
                      1. Aldred W, Plumb D, Bradford I, Cook J, Gholkar V,              Fairhurst D, Flaum C, Gubelin G, Heaton N, Minh CC,
                         Cousins L, Minton R, Fuller J, Goraya S and Tucker D:          Norville MA, Seim MR and Pritchard T: “Trends in NMR
                         “Managing Drilling Risk,” Oilfield Review 11, no. 2             Logging,” Oilfield Review 12, no. 3 (Autumn 2000): 2–19.
                         (Summer 1999): 2–19.
                                                                                        For more on the history and development of NMR logging:
                         Bargach S, Falconer I, Maeso C, Rasmus J, Bornemann T,         Dunn KJ, Bergman DJ and LaTorraca GA: Nuclear
                         Plumb R, Codazzi D, Hodenfield K, Ford G, Hartner J,            Magnetic Resonance—Petrophysical and Logging
                         Grether B and Rohler H: “Real-Time LWD—Logging for             Applications, Seismic Exploration No. 32. Amsterdam,
                         Drilling,” Oilfield Review 12, no. 3 (Autumn 2000): 58–78.      The Netherlands: Pergamon Press (2002): 3–10.
                         Bratton T, Edwards S, Fuller J, Murphy L, Goraya S,         3. Allen et al (2000), reference 2.
                         Harrold T, Holt J, Lechner J, Nicolson H, Standifird W

40                                                                                                                             Oilfield Review
spacing, on the order of 200 µs, improved petro-                                                    Dance of the Protons
physical measurement quality, including total                                                       NMR logging measures the magnetic moment of
porosity. Further improvements led to the CMR-                                                      hydrogen nuclei (protons) in water and hydro-
Plus logging tool with high-speed capability to                                                     carbons. Protons have an electrical charge and
acquire data at logging rates up to 2400 ft/hr                                                      their spin creates a weak magnetic moment.
[730 m/hr] for full porosity logging and 3600 ft/hr                                                 NMR logging tools use large permanent magnets
[1100 m/hr] for bound-fluid logging, rates three to                                                  to create a strong, static, magnetic-polarizing
five times faster than the CMR-200 tool.3                                                            field inside the formation. The longitudinal-
   To date, more than 7000 CMR logging jobs           questions concerning the presence, type and   relaxation time, T1, describes how quickly the
have been performed. For many applications,           producibility of reservoir fluids. For many   nuclei align, or polarize, in the static magnetic
NMR measurements are superior to other logging        operators, NMR logging has become a routine   field. Full polarization of the protons in pore
techniques and can provide critical answers to        service in typical logging programs.          fluids takes up to several seconds and can be

Summer 2003                                                                                                                                       41
done while the logging tool is moving, but the        providing precise high-resolution NMR measure-        signal-processing algorithm is implemented
nuclei must remain exposed to the magnetic field       ments under the harsh conditions typically            downhole to perform the critical T2 inversion
for the duration of the measurement. The rela-        encountered while drilling. Similar to the CMR        process. As a result of this inversion, important
tionship between T1 and increasing pore size is       tool, the proVISION LWD tool delivers measure-        petrophysical measurements can be derived
direct, yet inverse, to formation fluid viscosity.     ments that include mineralogy-independent             in real time, namely: lithology-independent
    A series of timed radio-frequency (rf) pulses     porosity, bound-fluid volume (BFV), free-fluid          porosity, T2 spectral distributions, bound- and
from the logging-tool antenna can be used to          volume (FFV), permeability, hydrocarbon detec-        free-fluid volumes, permeability and information
manipulate proton alignment. The aligned protons      tion and T2 distributions.                            about fluid saturations and characteristics.
are tilted into a plane perpendicular to the static       Flexible design allows engineers at the well-     However, because of telemetry bandwidth limita-
magnetic field. These tilted protons precess          site to modify the measurement sequence and           tions, real-time data transmission is limited to
around the direction of the strong induced            operational characteristics of the tool for one of    magnetic resonance-derived porosities, BFV,
magnetic field. The precessing protons create          three drilling modes: rotating, sliding or station-   FFV, motion-dependent quality control parame-
oscillating magnetic fields, which generate a          ary. The tool can be programmed manually or           ters and T2LM, or logarithmic mean of the T2 dis-
weak but measurable radio signal. However,            set to switch automatically based on drilling         tribution. These are used in conjunction with the
since this signal decays rapidly, it has to be        conditions (below). Engineers can program the         standard formation evaluation and survey mea-
regenerated by repeatedly applying a sequence         tool to measure T1, T2, or both simultaneously.       surements to optimize wellbore placement
of radio-frequency pulses. The precessing protons     Although both measurements can generate NMR           within the reservoir.
in turn generate a series of radio-signal pulses or   formation-evaluation data, the proVISION sys-             Transmission of T2LM, BFV or FFV and poros-
peaks known as spin echoes. The rate at which         tem relies primarily on T2 measurements, which        ity allows calculation of permeability using the
the proton precession decays, or loses its align-     produce higher statistical repeatability and ver-     Schlumberger-Doll Research (SDR) or Timur-
ment, is called the transverse-relaxation time, T2.   tical resolution.                                     Coates equations.6 Although T2 distributions
    T1 and T2 processes are affected predomi-             Both T1 and T2 measurements sample an             themselves can be provided in real time, teleme-
nantly by interaction between pore-fluid              exponential time evolution process. T1 measure-       try bandwidth limitations require prioritization
molecules, or bulk-relaxation characteristics,        ments sample an exponential buildup and T2            of data; less critical information is stored in
and from pore-fluid interactions with the grain        measurements, an exponential decay. The T1            memory for later processing.
surfaces of the rock matrix, also known as            measurement consists of a few samples on this             Data are transmitted to surface in real time
surface-relaxation characteristics. In addition, in   buildup, each of which requires an additional         by the PowerPulse MWD telemetry system. As
the presence of a significant magnetic-field          wait time depending on the point measured. The        with other VISION Formation Evaluation and
gradient within the resonant zone, there is relax-    T2 measurement, on the other hand, captures the       Imaging While Drilling LWD tools, maximum
ation by molecular diffusion that influences only      complete decay within a single Carr-Purcell-          environmental conditions for the proVISION tool
T2 processes.4                                        Meiboom-Gill (CPMG) measurement after only            are 300°F [150°C], 20,000 psi [138 MPa], and
                                                      one wait time, resulting in a greater number of       dogleg severity of 8°/100 ft [8°/30 m] while rotat-
NMR While Drilling                                    echoes per measurement. Thus the T2 measure-          ing and 16°/100 ft [16°/30 m] while sliding.
Following the widespread acceptance of wireline       ment can be taken more quickly leading to either          The proVISION opposing-dipole magnet
NMR, development and field-testing of LWD NMR          a higher sample rate or to more averaging and,        design produces a symmetric magnetic field. The
tools began in the late 1990s.5 Research and          therefore, enhanced data quality.                     vertically oriented tubular samarium-cobalt
development efforts and lessons learned from              For LWD NMR measurements to be available          permanent magnets are stable within the operat-
wireline-conveyed NMR logging ultimately led to       in real time, they must be transmitted to the         ing temperature range of the tool. A predictable
the introduction of the proVISION real-time           surface by mud-pulse telemetry. From the raw          and repeatable NMR measurement is produced
reservoir steering service in 2001, capable of        measurements performed by the tool, an optimal        (next page, top).
                                                                                                                The interaction of the rf field and static mag-
                                                                                                            netic field produces a resonant region, or shell,
                                                                                                            with a diameter of 14 in. [36 cm] and height of
                Anticipated           Wait time,         Repetitions          Number of                     6 in. [15 cm] (next page, bottom). Magnetic-field
                productivity            sec                                    echoes
                                                                                                            strength within the shell is approximately 60
                 Oil                     6.00                 2                  500                        gauss, with a field gradient of about 3 gauss per
                                         0.60                 2                  300                        centimeter. The width of the measurement shell
                                         0.04                40                   20                        allows formation measurement in slightly
                 Oil and gas            13.00                 2                  500                        enlarged or deviated wellbores and when the tool
                                         0.60                 2                  300                        is eccentered. The formation depth of investiga-
                                         0.04                40                   20                        tion (DOI) varies with borehole diameter. For
                                                                                                            example, in an 81⁄2-in. diameter borehole, the DOI
            > The proVISION tool pulse-sequence parameters. The tool’s programmability
                                                                                                            is 23⁄4 in. [7 cm]. At a drilling rate of 50 ft/hr
            is demonstrated in this triple-wait-time acquisition sequence that was used
            to evaluate oil-productive (upper set) and oil- and gas-productive (lower set)                  [15 m/hr], vertical resolution is 3 to 4 ft [0.9 to
            intervals in a deepwater Gulf of Mexico well, USA.                                              1.2 m] after data stacking.

42                                                                                                                                             Oilfield Review
    For geosteering purposes, field engineers can
place the tool directly behind the downhole                                                                       Magnetic field
motor or PowerDrive rotary steerable system or                                                           (14-in. diameter x 6-in. height)
directly above the bit sub. To further enhance
geosteering capabilities, the proVISION antenna                                                                    Mud flow
section, which contains the permanent magnets,
is located at the bottom of the tool, placing the
measurement point as close to the bit as possible.
    The existence of powerful magnets within the
bottomhole assembly (BHA) has the potential to
adversely affect azimuthal magnetic-survey
instruments used for determining spatial coordi-                            Tubular
                                                                         samarium-cobalt                      Optional stabilizer
nates of the borehole. However, Schlumberger                                magnets
engineers have demonstrated through modeling
and experimentation that the axially symmetric                 > The proVISION tool design. Housed within a 37 ft [11.3 m] long, 6 3⁄4-in.
magnetic field of the proVISION tool has little                 [17.1-cm] diameter drill collar, the tool’s outside diameter is 7 3⁄4 in. [19.7 cm].
influence on azimuthal magnetic measurement.                    When configured with no external upsets and with wearbands in place, the
Since the magnitude of magnetic-field interfer-                 tool can be run in boreholes ranging from 8 3⁄8 in. up to 10 5⁄8 in. diameter. On-
ence is small and directly proportional to the                 site field engineers may attach a screw-on stabilizer to reduce lateral motion
                                                               and centralize the tool in a borehole. Telemetry connections on both ends of
intensity of the magnetic field produced by the                 the tool assembly allow configuration to any section of a bottomhole assem-
proVISION tool, errors are significant only when                bly (BHA). The tool is turbine-powered, rather than battery-powered, and can
the proVISION tool is placed directly above the                accommodate flow rates ranging from 300 to 800 gal/min [1136 to 3028 L/min].
survey instrument. Based on numerical models
and physical measurements, Schlumberger
engineers have developed survey correction
algorithms for NMR magnetic interference.
These algorithms are included in the IDEAL
Integrated Drilling Evaluation and Logging well-                                     14 in.
site software.
                                                                                                   2 3⁄4 in. DOI
4. For more on T2 relaxation mechanisms: Kenyon et al and                                                                     Diameter of investigation
   Allen et al (2000), reference 2.                                                                                                    14 in.
5. Prammer MG, Drack E, Goodman G, Masak P, Menger S,
   Morys M, Zannoni S, Suddarth B and Dudley J: “The
   Magnetic Resonance While-Drilling Tool: Theory and
   Operation,” paper SPE 62981, presented at the SPE
   Annual Technical Conference and Exhibition, Dallas,
                                                                                       Mud flow

   Texas, USA, October 1–4, 2000
                                                                                                       Resonant zone
                                                                 6 in.
   Drack ED, Prammer MG, Zannoni SA, Goodman GD,
   Masak PC, Menger SK and Morys M: “Advances in LWD
   Nuclear Magnetic Resonance,” paper SPE 71730, pre-                                                      Magnetic field
   sented at the SPE Annual Technical Conference and
   Exhibition, New Orleans, Louisiana, USA, September 30–
   October 3, 2001.
   Horkowitz J, Crary S, Ganesan K, Heidler R, Luong B,                                                Annular magnet                  8 1⁄2-in.
   Morley J, Petricola M, Prusiecki C, Speier P, Poitzsch M,                                                                          borehole
   Scheibal JR and Hashem M: “Applications of a New
   Magnetic Resonance Logging-While-Drilling Tool in a
   Gulf of Mexico Deepwater Development Project,”                                    8 1⁄2-in.
   Transactions of the SPWLA 43rd Annual Logging                                    borehole
   Symposium, Oiso, Japan, June 2–5, 2002, paper EEE.
   Morley J, Heidler R, Horkowitz J, Luong B, Woodburn C,      > Cross sections of the proVISION tool. The axial section through the
   Poitzsch M, Borbas T and Wendt B: “Field Testing of a       antenna (left) illustrates the symmetric tool design. The dark blue bars are
   New Magnetic Resonance Logging While Drilling Tool,”
   paper SPE 77477, presented at the SPE Annual Technical
                                                               hollow cylindrical magnets. Lines of constant field strength (blue) indicate a
   Conference and Exhibition, San Antonio, Texas, USA,         gradient magnetic field that decays away from the tool. The section through
   September 29–October 2, 2002.                               the coaxial wound antenna coil is shown in black. The interaction of the
6. Akbar M, Vissapragada B, Alghamdi AH, Allen D,              antenna and the magnets produces a cylindrical resonant shell (red stripes)
   Herron M, Carnegie A, Dutta D, Olesen J-R,                  that is 6 in. [15 cm] long, 0.4 in. [10 mm] thick, with a 14-in. [36-cm] diameter of
   Chourasiya RD, Logan D, Stief D, Netherwood R,              investigation. The transverse section through the coaxial wound antenna coil
   Russell SD and Saxena K: "A Snapshot of Carbonate           (right) illustrates the axisymmetric resonant shell (red). The resonant shell is
   Reservoir Evaluation," Oilfield Review 12, no. 4
   (Winter 2000/2001): 20–41.                                  the only place the measurement is made—no measurement is made
                                                               between the tool and the resonant shell or from the resonant shell farther
                                                               into the formation. The formation depth of investigation (DOI) in an 8 1⁄2-in.
                                                               [21.5-cm ] diameter borehole is 2 3⁄4 in. [7 cm].

Summer 2003                                                                                                                                               43
                                                                                                                                                                                                   Making Measurements
                                                                                   13 1⁄2 in.                                                         13 1⁄2 in.                                   The proVISION tool operates in a cyclic mode
                                                                                                                                                                                                   rather than a continuous mode. The operating
                                                                                                                                                                                                   cycle consists of an initial polarization wait time
                                                                                                                                                                                                   followed by the transmission of the high-
                                                                                                                                                                                                   frequency rf pulse and then the reception of the
                                                                                                                                                                                                   coherent echo signal, or echo train. The cycle of
                                                                                                                                                                                                   pulsing and echo reception is repeated in
                                                                                                                                                                                                   succession until the programmed number of
                                                                                                                                                                                                   echoes has been collected. Typically, the acquisi-
                                                                                                           Borehole wall
                                                         Resonant region

                                                                                                                                                                                                   tion is defined by the Carr-Purcell-Meiboom-Gill
                                                                                                                                                                                                   (CPMG) sequence. An initial 90° pulse followed
                                                                                                                                                                                                   by a long series of timed 180° pulses character-
                                                                                                                                                                                                   izes the CPMG sequence. The time interval
                                                                                                               Polarized                                                                           between the successive 180° pulses is the echo
                                                                                                                region                                                                             spacing and is generally on the order of hundreds
                                                                                                          Resonant region                                                                          of microseconds.
                                                                                                                                                                                                       To cancel the intrinsic noise in a CPMG
                                                                                                                                                                                                   sequence, the CPMGs are collected in pairs. The
                                                                                                                                                                                                   first of the pair is a signal with positive phase.
                                                > Effect of lateral motion on the proVISION NMR measurement. The tool is
                                                                                                                                                                                                   The second of the pair is collected with an 180°
                                                centered in the borehole at the beginning of the measurement cycle (left).                                                                         phase shift, also known as the negative phase.
                                                Subsequent to the initial polarization, drillstring motion causes the tool to rest                                                                 The two CPMG sequences are then combined to
                                                against the borehole wall, partially outside the polarized region (right). Ideally,                                                                give a phase-alternated pair. Compared with the
                                                the tool would not move during the course of a CPMG pulse-echo sequence.                                                                           individual CPMG sequence, the combined or
                                                However, lateral motion of the tool during rotation causes the measurement
                                                shell, or resonant region, to move out of the polarized region of investigation.                                                                   stacked CPMG sequence has an improved S/N.
                                                This can result in T2 amplitude and distribution errors.                                                                                               Measurements of T1 and T2 and their distri-
                                                                                                                                                                                                   butions are key elements of NMR logging. The
                                                                                                                                                                                                   primary T1 quantity measured is signal amplitude
Revolutions per min. (RPM)

                             600                                                                                                                                                                   as a function of polarization recovery time. The
                                                                                                                                                                                                   primary T2 quantities measured are echo-signal
                                                                                                                                                                                                   amplitudes and their decay. Pulse parameters
                                                                                                                                                                                                   such as echo spacing, wait times and the NMR
                                                                                                                                                                                                   measurement cycle define all aspects of the NMR
                             -200                                                                                                                                                                  measurement and are completely programmable
                                       0             2                        4           6           8           10                        12         14          16            18           20
                                                                                                                Time, s                                                                            in the proVISION tool.

                                                         Lateral velocity 15 mm/s                                                                     Lateral velocity 33 mm/s                     Drillstring Dynamics and NMR Measurements
                              5                                                                                                   5                                                                NMR measurements are not instantaneous. Tool
                              4                                                                                                   4                                                                movement may cause the resonant or excited
                              3                                                                                                   3                                                                region to move during data acquisition (above
                                                                                                                                                                                                   left). The proVISION tool is equipped with sen-
                              2                                                                                                   2
                                                                                                                                                                                                   sors that measure the amplitude and velocity of
Position, mm

                                                                                                                   Position, mm

                              1                                                                                                   1                                                                lateral motion, and instantaneous revolutions
                              0                                                                                                   0                                                                per minute (rpm).
                             -1                                                                                                   -1                                                                   Tool movement can affect both T1 and T2 mea-
                                                                                                                                                                                                   surements. Motion-induced decay primarily
                             -2                                                                                                   -2
                                                                                                                                                                                                   affects long T2 values, resulting in faster echo
                             -3                                                                                                   -3                                                               decays that may reduce the accuracy of NMR
                             -4                                                                                                   -4                                                               measurement, particularly in light hydrocarbon
                             -5                                                                                                   -5                                                               and carbonate formations. These motion effects
                                  -5       -4   -3       -2                -1 0       1       2   3   4    5                           -5   -4   -3   -2    -1 0       1    2         3   4    5
                                                                                                                                                                                                   are most severe when the measurement shell is
                                                                           Position, mm                                                                     Position, mm
                                                                                                                                                                                                   thin in relation to the tool displacement, often
> Lateral drillstring motion plots. During the 20-sec time interval, the lower left and right panels show
                                                                                                                                                                                                   resulting in movement of the resonant shell out
examples of benign and severe motion recorded by the proVISION tool while rotary drilling. Intervals
of motion amplitude less than 1 mm (bottom left) correspond to the low-rpm intervals shown (top) and                                                                                               of the region of investigation, even for small tool
represent a nearly stationary condition. Violent motion occurs during the remaining time intervals,                                                                                                movements. A high-gradient static magnetic field
when the tool is spinning freely and has lateral motion amplitudes up to 5 mm.

44                                                                                                                                                                                                                                    Oilfield Review
results in a thin measurement shell, which
rapidly decays with distance away from the tool.                              400
In contrast, the proVISION tool has a low gradi-
ent design that results in a thick measurement                                200

shell and insensitivity to tool motion.
    Since lateral motion can potentially shorten                                0
T2 decay rates, understanding this motion is
critical for developing data quality-control                                  -200
                                                                                     0   5                    10                        15                         20
techniques. To assess motion-induced effects,
                                                                                                            Time, s
engineers must know the frequency, amplitude,
trajectory and timing of the motion.7 Rapid-                                                                                                   6 wraps ahead

                                                        Number of rotations
sampling accelerometer and magnetometer
systems measure real-time drillstring motion
(previous page, bottom). Motion data are pro-
cessed in 20-sec snapshots. Raw snapshot data
                                                                                                                          7 wraps behind
are compressed and can be stored in memory,
while the processed results are recorded contin-                                     0   5                    10                        15                         20
uously to provide an uninterrupted log of lateral                                                           Time, s
motion. The theoretical maximum T2 value              > An example of extreme stick-slip. The upper graph shows instantaneous rotation (rpm). At about
resolvable during motion is calculated and a flag      8 sec into the time interval, the BHA becomes stuck for about 7 sec until the continued buildup in torque
indicating NMR data quality is transmitted with       releases the BHA and the stored energy accelerates the drillpipe to over 300 rpm after which the BHA
the real-time data set.                               becomes stuck again. The lower graph shows the number of cumulative rotations. The number of rota-
    Motion data obtained with the proVISION           tions increases until the BHA becomes stuck, at which point the topdrive continues turning and builds
                                                      seven wraps in the drillstring before the BHA breaks free. The BHA releases the built-up energy, and
tool have broad independent utility. These data       inertia causes it to overrotate and advance six wraps ahead of the topdrive, potentially unscrewing
can alert the driller to excessive lateral motion,    sections of the BHA.
an unfavorable resonant mode or excessive
shocks allowing corrective action to be taken to
reduce potential BHA or drill-bit damage and to
optimize drilling rates, improving drilling effi-
ciency. Timely response to excessive drillstring      at their Wytch Farm project located in the south           At Wytch Farm, BP’s method was relatively
motion can also minimize borehole enlargement         of England. Geological studies of the Sherwood         simple to implement. The Sherwood sandstone is
(above right).                                        sandstone oil reservoir established that reservoir     not highly cemented and grain size, porosity and
                                                      productivity is a function of permeability, and        permeability have a clearly defined relationship.
Optimizing Well Productivity                          that permeability is controlled by grain size          Also, well cores were available for model calibra-
Proper well placement and completion design           and porosity. Core data were used to create            tion. In many other reservoirs, the petrophysical
are key to optimizing productivity. To accomplish     permeability bulk-density transforms for each          characteristics are less straightforward. While
this, drillers must place wellbores in the most       grain-size class and these, in turn, were used to      similar processes might provide comparable
productive part of a target reservoir, and engi-      estimate PI. As drilling progressed, a permeabil-      results while drilling in more complex reservoirs,
neers must design completions to maximize oil         ity log was generated in real time using grain size    the petrophysical community wanted a more
production and recovery while simultaneously          obtained from sieve analysis of drill cuttings and     accurate and complete formation-evaluation
limiting water production. Real-time LWD NMR          combining porosity measurements from a litho-          solution. NMR in real time can provide this
logging provides the data necessary for informed      density-neutron logging tool. Petrophysicists          information and help in optimizing wellbore
decision-making.                                      then calibrated the model against offset wells.        placement and completion design.
    Determining which intervals of a reservoir             Engineering and petrophysical teams used
                                                                                                             7. Speier P, Crary S, Kleinberg RL and Flaum C: “Reducing
should be completed requires an estimate of a         these early real-time permeability-productivity           Motion Effects on Magnetic Resonance Bound Fluid
well’s productivity index (PI). Traditionally, this   estimates to model and optimize a well’s eco-             Estimates,” Transactions of the SPWLA 40th Annual
                                                                                                                Logging Symposium, Oslo, Norway, May 30–June 3, 1999,
question has been addressed after completion of       nomic potential in several ways. Decisions to             paper II.
drilling, wireline logging and production testing.    adjust well trajectory were based on real-time         8. Blosser WR, Davies JE, Newberry PS and Hardman KA:
The PI is based on a permeability profile, which       productivity predictions. By optimizing perfora-          “Unique ESP Completion and Perforation Methodology
                                                                                                                Maximises Production in World Record Step-Out Well,”
is the product of reservoir permeability and          tion intervals, the team maximized production             paper SPE 50586, presented at the SPE European
vertical thickness. These measurements are            and minimized the potential for water break-              Petroleum Conference, The Hague, The Netherlands,
                                                                                                                October 20–22, 1998.
obtained from well logs, formation tests, or both.    through. These data were used to estimate                 Harrison PF and Mitchell AW: “Continuous Improvement
    For more than a decade, operators have            reserves remaining in wells where intervals had           in Well Design Optimises Development,” paper SPE
                                                                                                                30536, presented at the SPE Annual Technical
sought real-time estimates of permeability and        been plugged back for water shutoff.8                     Conference and Exhibition, Dallas, Texas, USA,
PI. In 1994, BP engineers successfully experi-                                                                  October 22–25, 1995.
mented with real-time PI determination methods                                                                  Hogg AJC, Mitchell AW and Young S: “Predicting Well
                                                                                                                Productivity from Grain Size Analysis and Logging While
                                                                                                                Drilling,” Petroleum Geoscience 2, no. 1 (1996): 1–15.

Summer 2003                                                                                                                                                         45
NMR in Real Time
Modern NMR logs measure mineralogy-                                                                                                  proVISION Porosity
                                                            Washout                                                                                                     Total Porosity
independent porosity and provide an estimate of                                                                                0.6         ft3/ft3         0
                                                                                                                                                                0.6         ft3/ft3       0
permeability and bound-fluid volumes. They can               Caliper                                                                   Hydrocarbon Flag
                                                                                                    Phase Resistivity                                                 Bulk Volume Water
also detect the presence of hydrocarbons. When        -2      in.       4                                                               Bulk Density            0.6         ft3/ft3       0
                                                           Gamma Ray                          0.2        ohm-m          2000
combined with other LWD measurements,                                                                                          1.65        g/cm3         2.65
                                                      0       API     150       Rotation       Attenuation Resistivity                                                  Bound Water               T2 Distribution
NMR data can be used to generate potential                                                    0.2        ohm-m          2000
                                                                                                                               Thermal Neutron Porosity
                                                      Rate of Penetration         RPS                                                                                      Water
production estimates in real time.                                                                                             0.6         ft3/ft3         0                                  0                     40
                                                                            0            10     Real-time proVISION
    In 2002, BP engineers applied the proVISION       Rate of Penetration                           Permeability                      Hydrocarbon Flag                 proVISION BFV              proVISION T2LM
                                                      0.25 ft/sec       0       Sample        0.2         mD            2000 1             ft3/ft3        -10 0.6           ft3/ft3       0 1         msec     10,000
system on a deepwater project in the Gulf of
Mexico, USA (right). During drilling with oil-base                               XX650
mud, real-time NMR logs were obtained in three
separate 81⁄2-in. diameter wells. The proVISION
pulse sequence consisted of a single wait time
and burst sequence. A relatively long wait time of
12 sec was used to ensure adequate polarization
of the light hydrocarbons that were expected
in this reservoir. Six hundred echoes were col-
lected after the long wait time. The burst
sequence consisted of 20 echoes following a                                      XX700
0.08-sec wait time. Echoes were collected with
spacing of 0.8 and 1.2 msec. The overall NMR
cycle time was about 30 sec at a drilling rate of
approximately 70 ft [21 m] per hour. This combi-
nation of cycle time and rate of penetration
(ROP) gave a depth sample rate of about 0.75 ft
[0.23 m] per phase-alternated pair.
    To determine BFV, a T2 cutoff of 90 msec was
chosen. This T2 cutoff value was based on experi-                                XX750
ence with wireline NMR measurements in this
field. Evaluation by the petrophysical team indi-
cated that neutron, density and NMR porosity
were in agreement through the sandstone, which
has a porosity of about 28 p.u. In addition to NMR
data, the proVISION data set provided the opera-      > Formation analysis in deepwater Gulf of Mexico, USA. The proVISION resistivity-independent oil-
                                                      indicator information, bound-fluid volume data and permeability data are integrated with wireline
tor with drilling performance, lateral motion and
                                                      log-derived water-saturation information to deliver key producibility estimates while the well is being
downhole RPM logs to detect erratic drilling con-     drilled. Tracks 1 through 4 are available as real-time data channels. Changes in the signature of the
ditions, such as stick-slip motion, and allowed       recorded mode T2 distribution (Track 5) confirm the oil/water contact. The hash marks in the depth
the driller to take corrective actions, potentially   track are NMR raw-data sample points.
extending the life of the bottomhole assembly
and optimizing ROP.

The Quest for Carbonate Evaluation
Hydrocarbons in the Al Shaheen field, offshore
Qatar, are currently produced from three              only 3000 ft [914 m] in true vertical depth.9 In                                           NMR techniques can help determine reser-
Cretaceous formations, the Kharaib, Shuaiba and       such wells, drillpipe cannot be rotated in the                                         voir fluid flow and permeability characteristics.
Nahr Umr. The Kharaib and Shuaiba reservoirs          hole with logging cable attached. Frictional                                           These characteristics may vary significantly with
are carbonate, while the Nahr Umr comprises           effects eventually prohibit sliding beyond about                                       changes in geologic facies. Detection of facies
thin sandstones (next page, top).                     13,000 ft [3962 m]. Thus, wireline-conveyed log-                                       variation is critical to reservoir understanding
    Maersk Oil operating the Al Shaheen field in       ging tools are typically unable to reach the far-                                      and optimal wellbore placement. Often, particu-
cooperation with Qatar Petroleum is developing        thest part of a horizontal section. LWD tools are                                      larly in carbonate reservoirs, the lack of consis-
these complex formations with extended-reach          conveyed over the entire length of the borehole                                        tent relationships between porosity and
horizontal wells that occasionally exceed             while providing data for geosteering and primary                                       permeability on a reservoir scale limits LWD
30,000 ft [9144 m] measured depth (MD) while          formation evaluation.

46                                                                                                                                                                                                Oilfield Review
petrophysical characterization using porosity
logs. Conventional wireline-conveyed NMR logging                                                                                                                                           IRAN
has improved the characterization of geologic
facies and other petrophysical carbonate proper-
ties such as permeability (bottom).
     Drilling extended-reach wells in the Al
Shaheen field is challenging. Rotary steerable
BHAs are typically used for directional control
in the drilling of the long horizontal sections. The
petrophysical team was concerned about
diminished LWD NMR data quality due to                                                                                                                                          Al Shaheen
                                                                                                                             SAUDI ARABIA                                           field
motion-dependent T2 decay resulting from
the typically high levels of BHA shock, stick-slip                                                                                                                     QATAR
and lateral tool motion during drillstring rota-
tion. With ROPs occasionally in excess of
                                                                                  IRAQ                       IRAN                AFGHANISTAN
500 ft/hr [152 m/hr], further data-quality loss                                                                         0            50             100         150 miles
was expected.
                                                                                                                        0             100
                                                                                                                                                          200         300 km      UNITED ARAB EMIRATES
     Carbonate rocks typically have lower surface-
relaxation times, which leads to extended T2
                                                                           SAUDI ARABIA
times. Since much of the important petrophysi-
cal information is contained in the later echoes,                                                                   OMAN                                   > Location of the Al Shaheen field
acquisition sequences in carbonates typically                                                                                                              operated by Maersk Oil Qatar AS in
                                                                                                                                                           cooperation with Qatar Petroleum.
require a longer wait time and a greater number
of echoes than in clastic formations. It was                                             YEMEN
unknown whether the late T2 components typi-
cally seen in the Al Shaheen carbonate rocks
would be detected under the expected difficult
drilling conditions.
                                                                                                                                     Bound Fluid
     Engineers attempted to alleviate as many
variables as possible during prejob planning. To                                                                                      Free Fluid
                                                                Bins 1-2
improve the S/N, raw echo stacking was also                                                                                      Total CMR Porosity
planned. Since facies changes typically occur                    Bin 3                                                       0.6          m3/m3            0

over tens or hundreds of feet in extended-reach                                                                                    CMR Free Fluid
                                                                 Bin 4
                                                                                                                             0.6          ft3/ft3          0
wells, and the detection of small-scale variations
                                                                 Bin 5                                                               Bulk Density
was not the main objective, a loss of resolution in
                                                                 Bin 6                                                       1.7          g/cm3           2.7     NMR T2 Distribution          Deep Image
exchange for improved S/N was acceptable.                                                           SDR Permeability               Neutron Porosity

     The world’s first proVISION deployment in a                 Bins 7-8                                                     0.6          m3/m3            0 0                      29            ohm-m
carbonate reservoir was in an extended-reach,                  Gamma Ray             Depth,      Timur-Coates Permeability       Photoelectric Effect                   T2LM                Image Orientation
81⁄2-in. diameter horizontal well, drilled to more         0      API       100        ft                                    2                            12 0.3         msec     6000 U      R     B     L     U
than 24,000 ft [7315 m] MD with water-base mud.
A rotary steerable assembly controlled trajectory
while LWD NMR data were obtained in real time
along the entire borehole length.
     Limited amounts of core material were avail-                                                                                                                                            Sliding–no image
able from this particular section of the Shuaiba
reservoir. Historically, carbonate facies identifi-
cation and interpretation were based on a com-
bination of drill cuttings, thin sections and log

9. Damgaard A, Hansen P, Raven M and Rose D: “A Novel
                                                                                                                                                                                             Sliding–no image
   Approach to Real Time Detection of Facies Changes in
   Horizontal Carbonate Wells Using LWD NMR,”
   Transactions of the SPWLA 44th Annual Symposium,
   Galveston, Texas, USA, June 22–25, 2003, paper CCC.
                                                          > Identifying changes in the Shuaiba limestone reservoir with wireline NMR data. The NMR data show
                                                          a large decrease in free fluid, an increase in bound fluid (Track 3, shown shaded yellow) and a decrease
                                                          in NMR permeability (Track 2) from a depth of XN010 to XN070. It would be difficult, if not impossible,
                                                          to identify these changes with standard porosity (Track 3, neutron porosity in blue and bulk density in
                                                          red) and gamma ray logs (Track 1, solid green curve).

Summer 2003                                                                                                                                                                                                   47
                                                                                               Bound Fluid

                                                                                                Free Fluid
                                                                                               Bulk Density
                                                                                         1.7      g/cm3       2.7

                                                                                         Thermal Neutron Porosity
                           Binned NMR Porosity
                          Early           Late                                                                          NMR T2 Distribution
                                                                SDR Permeability            Total NMR Porosity

                              Gamma Ray           ROP                                                                       NMR T2LM             Image Orientation
                                                             Timur-Coates Permeability          BFV–NMR
                          0       API      100 500 ft/hr 0                                                          3          msec     6000 U     R    B      L   U






                         > A clear image of borehole trajectory. The LWD resistivity image (Track 5) shows the wellbore trajec-
                         tory encountering an overlying marl. The NMR data clearly show a bimodal T2 (Track 4) with the short
                         T2 peak, centered at 6 msec, coming from the argillaceous material above the borehole from XX329 to
                         XX429 ft, and the longer T2 peak, centered at 200 msec coming from the limestone below the borehole.
                         Lateral changes in the limestone are also indicated. Facies 3 occurs from XX460 to XX474 ft and XX488
                         to XX500 ft, characterized by the lower T2 LM value (Track 4).

analysis. The borehole was expected to penetrate         tool, including differentiating various carbonate                                 As expected, a high level of downhole shock
multiple carbonate facies with varying perme-            facies along the wellbore path and comparing                                   and stick-slip occurred. ROP was variable, some-
abilities and producibility characteristics.             LWD NMR log quality with that of selected                                      times exceeding 500 ft/hr. Because of tool motion
Maersk Oil hoped to gain significant reservoir            intervals of wireline-conveyed NMR logs.                                       and fast ROP, NMR LWD data had a moderate
information in real time from the proVISION                                                                                             degree of noise compared with a wireline-

48                                                                                                                                                                       Oilfield Review
                                                                                                            Bound Fluid

                                                                                                             Free Fluid
                                                                                                            Bulk Density
                                                                                                   1.7         g/cm3          2.7

                                                                                                     Thermal Neutron Porosity
                                Binned NMR Porosity
                             Early                Late
                                                                          SDR Permeability               Total NMR Porosity             NMR T2 Distribution

                                    Gamma Ray               ROP                                                                             NMR T2LM
                                                                       Timur-Coates Permeability             BFV–NMR
                         0             API         100   500 ft/hr 0                                                                3          msec       6000





                        > Facies 1 from LWD NMR. The LWD data shown indicate an interval of clean carbonate where the T2
                        (transverse relation time) distribution (Track 4) contains a significant percentage of late T2 values. The
                        solid blue line is an empirically determined T2 cutoff that is used to partition the T2 distribution into a
                        fast component representing bound fluids and a slow component indicating the free fluids. The red
                        trace represents the T2 LM distribution. The T2 LM is generally well above the T2 cutoff value, indicating
                        that most of the fluid in the pore space is free fluid. The total porosity computed from the NMR data,
                        shown as a dashed black line in Track 3, is in agreement with the conventional limestone matrix neu-
                        tron porosity in blue, and with the formation bulk density displayed in red. The yellow area represents
                        the bound-fluid volume, while light green indicates the portion of the total porosity that is filled with
                        free fluids, or the effective porosity. The longest T2 times indicate the largest pores, while the shortest
                        are attributed to the smallest pore sizes. Large pores appear to make up a significant portion of the
                        total porosity, with only a small percentage comprising small and very small pores.

conveyed NMR log. However, data stacking                        Analysis based on NMR permeabilities,                               (previous page). A low bound-fluid volume and a
improved the S/N. Results from multiple MDT                 porosities, T2LM, bound-fluid volumes and free-                          high ratio of free to bound fluid typify Facies 1
Modular Formation Dynamics Tester runs pro-                 fluid volumes discerned three distinct porosity                          (above). Facies 2 has moderate bound-fluid vol-
vided data to estimate fluid mobility and adjust             systems. The team used changes in T2 character                          ume and a lower bound- to free-fluid ratio. The
the constants in NMR permeability equations.                to map facies variation along the borehole                              average T2 of Facies 2 is shorter than that of

Summer 2003                                                                                                                                                                      49
                                                                                                                                                      Facies 1 and the complete data spectrum is
                                                                            Bound Fluid
                                                                                                                                                      shifted to shorter T2 values. Facies 3 is typified by
                                                                             Free Fluid
                                                                                                                                                      high bound-fluid volume and a low ratio of free to
                                                                            Bulk Density
                                                                                                                                                      bound fluid. In Facies 3, the T2 spectrum is
                                                                    1.7        g/cm3         2.7
                                                                                                                                                      shifted farther toward shorter values. Thin
                                                                          Neutron Porosity                                                            sections made from cuttings confirmed the facies
  Binned NMR Porosity
 Early           Late
                                                                                                                                                      significance of the LWD NMR T2 response.
                                           SDR Permeability           Total NMR Porosity               NMR T2 Distribution
                                                                                                                                                           LWD NMR porosity agreed with density poros-
                                                                                                                                                      ity in Facies 1 and 2 with an average 3 p.u. deficit
     Gamma Ray                          Timur-Coates Permeability            BFV–NMR                       NMR T2LM             Image Orientation
                          ft/hr                                                                                                                       in Facies 3 believed to be due to a percentage of
 0       API      100 0           500                                                              3          msec     6000 U     R    B      L   U
                                                                                                                                                      faster-decaying T2 signals. LWD NMR data indi-
                          XX400                                                                                                                       cate different T2 decay rates for each of the three
                                                                                                                                                      facies, allowing clear differentiation; this would
                                                                                                                                                      not have been possible with neutron-porosity
                                                                                                                                                      measurements alone (left).
                                                                                                                                                           To improve confidence that the LWD NMR
                                                                                                                                                      data were identifying petrophysical changes in
                                                                                                                                                      the carbonate facies, the team had to rule out the
                            Facies 2
                                                                                                                                                      possibility that the interpreted T2 response was
                                                                                                                                                      being dominated by motion-induced T2 decay.
                          XX600                                                                                                                       Measured lateral velocity data were used to con-
                                                                                                                                                      firm that the T2 data were accurate and correctly
                                                                                                                                                      indicating changes in the carbonate facies (next
                                                                                                                                                      page, top left). This particular data set shows a
                                                                                                                                                      large amount of T2 data acquired even at ele-
                                                                                                                                                      vated lateral velocities. The current proVISION
                                                                                                                                                      design does not directly allow compensation for
                                                                                                                                                      downhole tool motion in the T2 decay measure-
                                                                                                                                                      ment. However, highlighting intervals of increased
                          XX800                                                                                                                       tool motion can be used as a log-quality indicator.
                                                                                                                                                           To examine the effects of downhole tool
                            Facies 3                                                                                                                  motion on LWD NMR data, wireline CMR mea-
                                                                                                                                                      surements acquired after drilling were compared
                          XX900                                                                                                                       with real-time proVISION data. Porosity, FFV,
                                                                                                                                                      BFV, T2LM and NMR permeabilities all compare
                                                                                                                                                      favorably (next page, right). The CMR data were
                                                                                                                                                      acquired over limited intervals for comparison,
                                                                                                                                                      primarily in the proximal part of the well that
                                                                                                                                                      had been open to invasion the longest. Some
                                                                                                                                                      CMR logged intervals displayed a small decrease
                                                                                                                                                      in T2LM values consistent with the additional fil-
                                                                                                                                                      trate invasion time prior to wireline logging.
                                                                                                                                                      None of the LWD NMR intervals indicated any
> Contrasting NMR data with resistivity images. An LWD resistivity image log is shown in Track 5. The
                                                                                                                                                      identifiable motion-induced T2 decay. The favor-
image is scaled such that conductive formations are dark and more resistive formations are light with
no absolute scale. The resistivity image shows a significant change in the formation resistivity while                                                 able comparison of the late T2 components indi-
the porosity remains more or less constant, implying a possible textural change. The NMR log over the                                                 cates that downhole lateral tool motion is not a
interval identified as Facies 2 indicates some large pores. The T2 LM is above the cutoff value, but with                                              dominant T2 decay mechanism in this data set.
a broad distribution of pore sizes resulting in a significant percentage of the total porosity being occu-
                                                                                                                                                           The proVISION system was configured to
pied by bound fluid. The estimated permeability of Facies 2 is lower than that of Facies 1 (see figure,
page 49). The NMR log over the interval identified as Facies 3 indicates few, if any, large pores. The                                                 transmit porosity, T2LM and FFV in real time to
T2 LM is below the cutoff value, and most of the total porosity is occupied by bound fluid. The estimated                                              allow use of measurements for geological charac-
permeability of Facies 3 is lower than that of Facies 1 or 2.                                                                                         terization and to aid geosteering. Although further
                                                                                                                                                      evaluation will be required to completely under-
                                                                                                                                                      stand the NMR T2 response in carbonate rocks,
                                                                                                                                                      the team working in the Al Shaheen field demon-
                                                                                                                                                      strated that carefully interpreted LWD NMR data
                                                                                                                                                      can be used to help detect variation in carbonate
                                                                                                                                                      facies and their petrophysical characteristics.

50                                                                                                                                                                                       Oilfield Review
                                                                                           Bound Fluid
                          200                                                              Free Fluid


                                                                                           CMR–BVF                    NMR T2LM
                                                                                          NMR Porosity                CMR T2LM                     CMR T2LM
 Lateral velocity, mm/s

                                                                                          CMR Porosity            NMR T2 Distribution          NMR T2 Distribution
                                                                           500      0                      0                            29 0                         29


                                0   75       150     225             300
                                         T2 LM, ms

> Lack of motion-induced decay. The data acquired in this field show
no apparent reduction in T2 values associated with the lateral velocity
of the LWD NMR tool, implying that in this well, tool motion does not
affect T2 decay.


The Next Generation
The proVISION system has demonstrated its
ability to acquire real-time logs in both clastic
and carbonate reservoirs, potentially identifying
less obvious or otherwise undetected facies
changes. Even for longer T2 components in                                     XX400

carbonate formations drilled at elevated ROP,
the tool delivers sufficient data resolution for
facies determination and for permeability and                              > Agreement of wireline CMR and proVISION data. The wireline NMR poros-
bound- to free-fluid volume calculations. The                               ity is seen to follow the same trend as the LWD NMR porosity with a small
LWD proVISION tool provides essential real-                                systematic shift to lower porosity (Track 1). This difference in total porosity is
                                                                           influenced by the differing depth of investigation of the tools and the
time reservoir information and data useful for                             difference in mud-filtrate invasion related to the formation exposure time.
making geosteering decisions in complex                                    Computed bound-fluid volumes are in agreement (Track 1). The vertical, or
reservoir settings.                                                        spatial, resolution of the LWD NMR tool is reduced because of the high level
                                                                           of stacking utilized to increase the S/N. Likewise, the physics of measurement
    Severe stick-slip and BHA shock are often
                                                                           imposes a temporal, or time, resolution limit on the LWD tool relative to that
associated with drilling long horizontal sections.                         seen with the wireline sensor. The overall effect is a smoothing of the T2 dis-
Bottomhole shock, combined with high ROP, may                              tribution over time and depth. The T2 LM of the LWD NMR is shown overlaid
increase noise in the data sets. However, field                             on the CMR data (Track 2). Considering the difference in tool design, acquisi-
                                                                           tion parameters, environmental conditions, and the time lapse between
data demonstrate that the proVISION tool is                                drilling and drillpipe-conveyed wireline logging, the comparison is excellent.
sufficiently robust to handle these conditions
and provide reliable T2 data.
    Future generations of NMR tools hold great
promise. The industry can look forward to the con-
tinued evolution of LWD NMR technology, which is
expected to provide drilling engineers and petro-
physical teams with significant advancements in
real-time formation evaluation for geosteering and
productivity optimization.              —DW, SP

Summer 2003                                                                                                                                                          51