MXPA99009936A - Apparatus and method for calculating a spin-es relaxation time distribution - Google Patents

Abstract

The present invention relates to a method for calculating a spin-spin relaxation time distribution. The spin echo amplitudes are obtained by integrating the equipment of the receiver's voltages over a time advantage. A linear operator is used to map the relaxation time distribution for spin echoes, produce a simple value decomposition (DVS) of the linear operator, determine the vectors of the DVS, and compress the spin echo data using the vectors. To eliminate a telemetry bottleneck, the T2 spectrum is calculated in the perforation and transmitted to the surface

Description

APPARATUS AND METHOD FOR CALCOIAR ONA DISTRIBUTION OF RELAXATION TIME ESPÍN-ESPÍN Cross-references This is a continuation in part of the U.S. Pat. Serial No. 09 / 033,965 (File of Lawyer No. 24,878), filed on March 3, 1998.

Background of the Invention The present invention relates in general to an apparatus and method for measuring nuclear magnetic resonance properties of a land formation traversed by a hole drilling and more particularly to an apparatus and method for computing a spin-spin relaxation time distribution. It is well known that atomic particles are a ground formation that have a nuclear magnetic angular momentum different from zero, for example protons, have a dry linear tendency in an aesthetic magnetic field imposed on the formation. Such a magnetic field could be generated naturally, as is the case for the magnetic field of the earth, BE. An RF pulse that applies a second transverse magnetic field to BE creates a magnetization component in the transverse plane (perpendicular to BE) that precesses around the vector BE with a characteristic resonance known as the Larmor frequency, ΔL, which depends on the intensity of the static magnetic field and the gyromagnetic relation of the particle. Hydrogen nuclei (protons) that precess around a magnetic field BE of 0.5 gauss, for example, have a characteristic frequency of about 2kHz. If a population of hydrogen nuclei were caused to precess in phase, the combined magnetic fields of the protons can generate a detectable oscillating voltage in a receiver coil, conditions which are known to those skilled in the art as a decomposition of free induction in a spin echo. The hydrogen nuclei of water, and the hydrocarbons that occur in pores in rocks, produce nuclear magnetic resonance (NMR) signals different from the signals that arise from other solids. U.S. Pat. Nos. 4,717,878 issued to Taicher et al. And 5,055,787 issued to Kleinberg and others, describe NMR tools that use permanent magnets to polarize hydrogen nuclei and generate a static Bo magnetic field, and RF antennas to excite and detect nuclear magnetic resonance at In order to determine the porosity, the ratio of free fluids and the permeability of a formation. The atomic nuclei are aligned with the applied field, B0, with a constant of time i. After a period of polarization, the angle between the nuclear magnetization and the applied field can be changed by applying an RF field, Bi, perpendicular to the static field B0, at the Larmor frequency FL =? B0 / 2p, where d is the ratio gyromagnetic proton and B0 designates the intensity of the static magnetic field. After the conclusion of the RF pulse, the protons precess in the plane perpendicular to B0. An RF pulse refocus sequence generates a sequence of spin echoes that produces a detectable NMR signal on the antenna. 'The U.S. patent No. 5,280,243 issued to Melvin Miller describes a nuclear magnetic resonance tool for the evaluation of the formation while it is being drilled. The tool includes a probe section consisting of a permanent magnet arranged in an annular groove extending longitudinally out of the piercing collar and an antenna disposed in a non-conductive magnetic sleeve outside the piercing collar. The gradient of the magnitude of the static magnetic field is in the radial direction. The antenna produces an RF magnetic field substantially perpendicular to both, the longitudinal axis of the tool and the direction of the magnetic field. With the apparatus? 243 the magnet must be long in its axial dimension as compared to its diameter so that the magnetic fields approximate their intended 2-D dipolar behavior. U.S. Pat. No. 5,757,186 issued to v Taicher et al., Describes a tool for measurement while drilling which includes a detection apparatus for making nuclear magnetic resonance measurements of the ground formation. The NMR detection device is mounted on. an annular groove formed in the outer surface of the collar for drilling. In one implementation -JO physical, a closure to the magnetic flux is inserted into the slit. A magnet is placed on the outer radial surface of the closure to the flow. The magnet is constructed of a series of radial segments that magnetize radially outward from the longitudinal axis of the tool. The closure to • Flow is required to provide adequate directional orientation of the magnetic field. The tools revealed in patents * 243 and? 186 suffer from common problems: Both tools require the use of a non-conductive magnet and place the magnet 0 outside the drill collar. For tool 243 the outer surface of the drill collar must contain a split area to accommodate the non-conductive magnet. For the tool? 186, the outer surface of the drill collar must contain a split area to accommodate flow closure, non-conductive magnet and antenna. Because the existence of the drill collar is a function of its spokes, reducing the outer diameter to accommodate only the magnet or closing the flow results in an unacceptably weak section of the drill collar, which could be bent or broken during the drilling operation. U.S. Pat. No. 5,557,201 issued to Kleinberg et al., Describes a pulsed nuclear magnetism tool for evaluation of the formation while drilling. The tool includes a drill bit, a drill string and a nuclear magnetic resonance device by impulse housed inside a collar for drilling made of a non-magnetic alloy. The tool includes a channel within the drill spring and the impulse NMR device through which the drilling mud is pumped into the hole bore. The pulsed NMR device comprises two tubular magnets which are mounted as poles facing each other, surrounding the channel and an antenna coil mounted on an outer surface of the piercing spring between the magnets. This tool is designed to resonate the nuclei in a measurement region known to those skilled in the art as the saddle point. U.S. Pat. No. 5,705,927 issued to Sezginer et al., Describes a nuclear impulse magnetism tool for evaluation of the formation while it is being drilled. The tool includes magnets in the form of wedges, located either inside or outside the tool, which suppress the magnetic resonance signal of the fluids in the hole bore, raising the magnitude of the static magnetic field in the hole bore so that the Larmor frequency in said hole is above the frequency of the oscillating field produced by an RF antenna located in a split area of the tool. Magnets in the form of wedges also produce gradients in the static magnetic field in the research region.

Compendium of the Invention The aforementioned disadvantages of the above method are solved by means of the subject invention for an apparatus and method for calculating a spin-spin relaxation time distribution. A static axisymmetric magnetic field is applied in a formation penetrated by a well borehole. An oscillating magnetic field is also applied to the formation. Nuclear magnetic resonance signals are detected from the formation and transmitted to a signal processor located in the well borehole. The signal processor computes a spin-spin relaxation time distribution of the detected signals. Spin-spin relaxation times can be transmitted to a surface of the well borehole (hole above). The plurality of signals are detected having a signal amplitude plus noise, Aj, characterized by the following relation: where? j is the noise in the Aj measurement, ai is the amplitude of the T2 distribution taken in T2; 1 represents the elements of the matrix X, where tw is time and c is a constant (the relation T1 / T2),? t is the echo spacing, and j = l, 2, ... N, where N is the number of echoes collected in a single sequence of impulses. In matrix notation, the equation becomes the noise,?, Is the unknown, therefore it approaches finding a minimum of the function can add a regularization term to the function and the function is minimized using a suitable iterative minimization algorithm, Brief Description of the Drawings The advantages of the present invention will become apparent from the following description of the appended drawings. It should be understood that the drawings should be used only for illustrative purposes and not as a definition of the invention. In the drawings: Fig. 1 illustrates an apparatus that records while it is being drilled; Fig. 2 represents a low gradient probes; Figs. 2a-2d illustrate contour lines I Bol corresponding to four configurations of low gradient magnets; Figs. 3a-3d represent the gradient contour lines I VBol corresponding to four configurations of low gradient magnets; Fig. 4 represents the high gradient probe; Fig. 4a represents the contour lines I B0I corresponding to the configuration of high gradient magnets; Fig. 4b represents the contour lines of the gradient I VBJ corresponding to the configuration of high gradient magnets; Fig. 5 represents the simple data acquisition mode; Fig. 6 represents the interleaved data acquisition mode; Fig. 7 represents the mode of data acquisition by bursts; and Fig. 8 represents a block diagram of the impulse programmer.

Detailed Description of the Preferred Fisisa Implementation Referring to Fig. 1, a nuclear magnetic resonance (NMR) tool for recording while drilling is illustrated. The tool 10 comprises a drill bit 12, a drill string 14, a series of RF antennas 36, 38, and at least one gradient coil 56. The tool 10 further comprises electronic circuits 20 housed within the drill collar 22. The electronic circuits 20 comprise RF resonance circuits for the antennas 36, 38, a microprocessor, a digital signal processor and a low voltage bus. The tool 10 further comprises a series of tubular magnets 30, 32 and 34 which are biased in a direction parallel to the longitudinal axis of the tool 10 but opposite each other, e.g., with the magnetic poles equal to one another. The magnets 30, 32 and 34 comprise either conductive or non-conductive material. The configuration of the magnets 30, 32 and 34 and the antennas 36 and 38 provide at least two research NMR regions 60, 62, with a substantially axisymmetric and RF static static field. A means for drilling a hole for a hole 24 in the formation comprises a drill bit 12 and a collar for drilling 22. The drill collar 22 may include stabilization means (not shown) to stabilize the radial movement of the drill. 10 drilling tool for the hole during drilling, however, the stabilization means are not required and therefore the tool 10 can operate unstable or stabilized. The sleeve 28 for the mud flow defines a channel 90 for carrying the drilling fluid through the drill string 1. A drive mechanism 26 rotates the drill bit 12 and the drill string 14. This drive mechanism is suitably described in US Pat. No. 4,949,045 granted to Clark and others. However, a mud motor can be placed in the drill string as a drive mechanism 26.

It is contemplated in the subject invention to combine N + l magnets to have at least N research regions in the array. The combinations contemplated by this invention include, without being limited thereto, regions of low gradient gradient, high gradient-high gradient, high gradient-low gradient, low-gradient high gradient, or a combination of high gradient, low gradient and Saddle point. The combinations of high and low gradient static field regions in the training offer several advantages. For example, the high gradient region may have a higher signal-to-noise ratio, but may experience signal loss when the tool 10 undergoes lateral movement in the hole bore. On the other hand, the low gradient region has a lower susceptibility to signal loss problems when the tool 10 is in motion. Also, with a moderate movement of the tool, trains of longer echoes can be acquired in the low gradient region than in the high gradient region, thus providing information about permeability, retained and free fluid, and types. of hydrocarbons. Moreover, the combination of data acquired from both regions of gradients can provide quantitative information about the amount of lateral movement experienced by the tool 10 and can be used to correct by movement the NMR data or at least control the data in quality . Measurements of devices such as strain gauges, accelerometers or magnetometers, or any combination of these devices could be integrated with NMR information to control the quality of the data or make corrections to the spin-echo train. With the combination of static magnetic fields of high and low gradient, the high gradient region exhibits a greater diffusion effect and therefore is of greater interest for hydrocarbon typing techniques than the low gradient region. Finally, the low gradient region has a static magnetic field that has a low amplitude and therefore this region with its lower frequency of Larmor is less affected by the conductivity of the fluids in the formation Y in the perforation for the hole.

Low Gradient Probe Referring to Fig. 2, in a section of the tool, hereinafter referred to as the low gradient probe, a central magnet 30 is axially spaced from a lower magnet 32. These magnets 30, 32 generate a static axisymmetric magnetic field. which is radial in its polarization and, along a reasonably long cylindrical shell, the static magnetic field has a reasonably constant magnitude. It is contemplated within the subject invention to excite a series of spin cylindrical shells in the formation, where each shell is resonant at a different NMR frequency and sequentially interrogate each shell with RF pulse sequences. The area between the magnets 30, 32 is suitable for housing elements such as electronic components, an RF antenna and other similar items. For example, a series of electronic pockets 70 can form an integral part of the sludge sleeve 28. These pockets 70 can accommodate the RF circuits (e.g., Q-switch, duplexer and preamplifier), preferably in close proximity to the RF antenna. In a preferred implementation of the invention, the pockets 70 form an integral part of the magnetically permeable member 16. In that case, to maintain the axial symmetry of the magnetic field, a magnetically permeable layer is placed over each pocket 70. The magnetically permeable member 16 it is placed inside the piercing collar 22 between the magnets 30, 32. The member 16 may consist of a single piece or of a series of combined sections between the magnets 30, 32. The member 16 is constructed of a magnetically suitable permeable material., such as ferrite, permeable steel or other alloy of iron and nickel, corrosion-resistant permeable steel, or permeable steel that has a structural function in the design of the member such as stainless steel Ph 15-5. The magnetically permeable member 16 focuses the magnetic field and may also carry either drilling fluid through the drill string or provide structural support to the drill collar. Moreover, the member 16 improves the formation of the static magnetic field generated by the magnets 30, 32 and minimizes the variations of the static magnetic field due to the vertical and lateral movement of the tool during the period of acquisition of the NMR signal. The sleeve segment 28 between the magnets 30, 32 may comprise a magnetically permeable member 16. In that case the segments of the lower sleeve 28 and the magnets 30, 32 will consist of a non-magnetic member. Alternatively, a magnetically permeable frame surrounding the segment of the sleeve 28 between the magnets 30, 32 defines a member 16.

In this case the segment may consist of a magnetic or non-magnetic material. It is contemplated in this invention to integrate the frame and the segment to form the member 16. The magnets 30, 32 are biased in a direction parallel to the longitudinal axis to the tool 10 with the magnetic poles equal facing each other. For each magnet 30, 32 the induction magnetic lines move outwardly from the ends of the magnet 30, 32 into the formation along the axis of the tool 10 and move inward to the other end of the magnet 30, 32 In the region between the central magnet 30 and the lower magnet 32, the induction magnetic lines move from the center outward entering the formation creating a static field in a direction substantially perpendicular to the tool axis 10. The magnetic lines of induction ae then move inwardly symmetrically above the central magnet 30 and below the lower magnet 32 and converge in the longitudinal direction within the sleeve 28. Due to the separation the magnitude of the static magnetic field in the central region between the central magnet 30 and the interior 32 is spatially homogeneous in comparison to a saddle point field. The amount of the separation between the magnets 30, 32 is determined based on several factors: (1) selecting the required characteristics of intensity and homogeneity of the magnetic field; (2) generating a field having small radial variations in the region of interest so that the echoes received during a pulse sequence (e.g.

CPMG, CPl, or other sequences) is less sensitive to lateral movement of the tool; (3) depth of the investigation; and (4) minimize the interference between the resonance circuits and the low voltage telemetry bar in order to improve the isolation of the receiving antenna which detects the NMR signals of the array. As the separation between the magnets 30, 32 decreases, the magnetic field becomes stronger and less homogeneous. Conversely, as the separation between magnets 30, 32 increases, the magnetic field becomes weaker and more homogeneous. Figs. 2a-2b illustrate the contour lines of I Bol corresponding to four configurations modeled in the laboratory of central magnets 30 and lower 32. These modeled results were computed using a tool having a preselected diameter (a constant diameter was used to model all the configurations). The configuration corresponding to Fig. 2a comprises a magnetically non-permeable member separating a central magnet 30 and a lower magnet 32 in 25 inches. The configuration corresponding to Fig. 2b comprises a non-magnetically permeable member that separates a central magnet 30 and a lower magnet 32 by 18 inches. The low gradient probe corresponding to Fig. 2d comprises a magnetically permeable member 16 separating a central magnet 30 and a lower magnet 32 by 25 inches. the dimensions mentioned above were modified simply to illustrate the effect of distance and / or a member that is not magnetically permeable or magnetically permeable in I B0I. Figs. 3a-3d represent the contour lines of the gradient I VBol corresponding respectively to the configurations illustrated in Figs. 2a-2d. In the low gradient probe, the magnetically permanent member is in parallel with a significant portion of the magnetic flux entering the center of the tool 10. To illustrate the magnitude of the B0 field shown in Fig. 2d at a distance of about seven inches radially of the longitudinal axis of the tool 10 is twice as large as the field B0 shown in Fig. 2a which was generated by the same configuration of magnets separated by a non-magnetically permeable member. Moreover, low gradient probes produce a greater and more uniform extension of the static magnetic field in the axial direction. The NMR signal measured in this section of the tool is substantially less sensitive to the vertical movement of the tool. Referring to Fig. 3d, with the low gradient probes, a relatively small gradient, approximately 3 Gauss / cm., is measured at a distance of approximately seven inches radially from the longitudinal axis of the tool. This low gradient results in a measured NMR signal that is substantially less sensitive to lateral movement of the tool 10. When the movement is moderate, longer echo trains can be acquired in this region, thus providing more information about the permeability, fluid retained and free and type of hydrocarbons. In the case of low gradient probes, as is the case with other gradient designs, the regions of the proton-rich hole bore that surround tool 10 will resonate only at frequencies greater than those that are being applied to the volume of the investigation , eg, there is no proton signal in the hole drilling. Other NMR-sensitive nuclei found in mud for drilling, such as sodium-23, resonate at static magnetic field intensities significantly higher than that of hydrogen when excited at the same NMR frequency. For low gradient probes, these higher field strengths do not occur in the region of the hole bore surrounding the tool, nor near the antenna where such unwanted signals could be detected.

High Gradient Probe Referring to FIG. 4, in another section of the tool, hereinafter referred to as the high gradient probe, a central magnet 30 is axially spaced from an upper magnet 34. The magnets 30, 34 are polarized in a direction parallel to the longitudinal axis of the tool 10 with the poles equal one opposite the other. These magnets 30, 34 generate a substantially axisymmetric static magnetic field that is radial in its polarization and, through a reasonably long cylindrical shell, the static magnetic field has a reasonably constant magnitude. It is contemplated within the invention to excite a series of cylindrical spin shells in the formation where each shell is resonant at a different RF frequency. As illustrated in Fig. 2c, if the separation of the magnets between 30 and 34 is about eight inches, the contour lines of the static magnetic field strength are substantially straight and the intensity of I Bol is greater than the intensity of the static magnetic field of the low gradient region. However, the gradient I VBol becomes larger as illustrated in Fig. 3c, at a distance of approximately seven inches radially from the longitudinal axis of the tool. The contour lines I VBol are curves which indicates gradient variation in the axial direction. The high gradient probe is improved by inserting a magnetically permeable member between the magnets 30, 34. The Fig. 4a represents the contour lines of I B0I corresponding to a configuration where the magnetically permeable member 16 separates the upper and central magnets 30 by eight inches. The contour lines of Fig. 4a show a slightly more intense field indicating a better signal-to-noise ratio and less curvature in the axial direction than the contour lines of Fig. 2c. Also, as illustrated in FIG. 4b, the magnetically permeable member 16 produces a more constant gradient I VBol in the axial direction that could simplify the interpretation of the NMR measurements influenced by diffusion. In the case of the high gradient probes, as well as with other gradient designs, the region of the proton-rich hole perforation surrounding the tool 10 will only resonate at higher frequencies than those that are being applied to the volume of the investigation, eg, there is no proton signal in the hole drilling. The high gradient probe is sensitive to a small part of the sodium from the hole drilling fluid. For a concentration of 30% NaCl in the hole-hole fluid, possibly the worst case, the estimated porosity error due to the sodium signal is approximately 0.08 inch. In the low gradient probe the sodium signal is substantially lower than that in the high gradient probe. Therefore, the sodium signal is negligible for both NMR probes.

Antennas and Gradient Coils Referring to Figs. 2 and 4, an RF magnetic field is created in the investigation regions originated by the antennas 36, 38 which are provided in the creased areas 50, 52. The RF field can be produced by one or more RF antenna segments. transmitting and / or receiving from different circumferential sectors of the recording device. See U.S. Patent Applications. Nos. 08 / 880,343 and 09 / 094,201 (Attorney's Records Nos. 24,784 and 24,784-CIP) assigned to Schlumberger Technology Corporation. Preferably each antenna 36, 38 comprises a coil 18 wound circumferentially around the slit area 50, 52. The RF field created by such a coil arrangement is substantially axisymmetric. It is contemplated in the subject invention, to use the antenna 36, 38 to detect NMR signals. However, a separate receiver antenna can be used to detect the signals. A non-conductive material 54 is provided in the split area 50, 52 under the antenna 36, 38. The material 54 is preferably a ferrite to increase the efficiency of the antenna 36, 38. Alternatively, the material 54 may consist of plastic, rubber or a reinforced epoxy composite material. The antennas 36, 38 come into resonance by the RF circuits to create an RF magnetic field in the regions under investigation. The slit area 52 forms a shallow groove in the drill collar without reducing the internal diameter thereof, which is ordinarily done to increase the strength in a region of the drill collar where the outside diameter has been split to provide an antenna. The split area 50 has a greater depth than the split area 52. Due to mechanical constraints, it is only possible to have a deeply split area where the internal diameter of the drill collar is substantially reduced. It is contemplated in the subject invention that the split areas 50, 52 have substantially the same depth or that the split area 52 have a depth greater than the split area 50. The spin cylindrical shells in the research region can be axially segmented or, preferably, azimuthally using at least one directionally responsive gradient coil 56 disposed in the split area 50 and / or 52. In a preferred physical implementation of the invention, three gradient coils are circumferentially placed around the split area and separated in a 120 ° angular distance segment. Other quantities of gradient coils can be defined, either in a number greater than or less than three, and said coils can be separated into angular distances other than 120 ° and / or in unequal angular segments. Each coil 56 is constructed with wire loops, which conform to the curvature of the outer surface of the material 54. The magnetic field produced by each gradient coil 56 in a region of the formation facing the front of the coil is substantially parallel to the static magnetic field produced by the magnets. As is known to those skilled in the art in basic NMR measurement, a pulse sequence is applied to the formation under investigation. In U.S. Pat. No. 5,596,274 issued to Abdurrahman Sezginer and U.S. Pat. No. 5,023,551 issued to Kleinberg et al., An impulse sequence such as the Carr-Purcell-Meiboom-Gill sequence (CPMG), first applies an excitation pulse, a 90 ° impulse to the formation that rotates the spin in the transverse plane. After the spin rotates at 90 ° and begins to be shifted, the carrier of the refocus pulses, the pulses at 180 ° shifts in phase relative to the carrier of the impulse sequence at 90 ° in accordance with the next relationship: where the expression in brackets is repeated for 13 = 1,2, N, where N is the number of echoes collected in a single CPMG sequence and the spacing between echoes is tecbo = 2tcp = t? ao ° and + ti + t2. 90 ° + r indicates an RF pulse that causes the spin to rotate at an angle of 90 ° to the + x axis, as commonly defined in the rotating frame of magnetic resonance measurements (phase alternate). The time between the application of the impulse at 90 ° and the impulse at 180 ° t0, is less than t ?, half the spacing between echoes. The CPMG sequence allows the acquisition of a symmetric measurement (e.g., a measurement without using the gradient coils). The exact timing parameters t0r ti and t ?, depends on several factors (e.g., the shape of the applied pulses). In the subject invention, a current pulse applied to the gradient coil 56 generates an additional magnetic field, substantially parallel to the static magnetic field. The current pulse is applied between the first 90 ° pulse and the 180 ° phase inversion pulse. This additional field causes an additional phase shift for the spins. Since the 180 ° phase inversion pulse does not compensate for the additional phase shift, the spins of the additional field do not form an echo-spin. However, for spins not subject to the additional field, an echo-spin occurs at a time 2tcp, echo-spin occurring of amplitudes successively smaller than the tcp times after each phase inversion pulse. The sequence of impulses is where is the time between the 90 ° impulse and the gradient momentum of duration d, is the time between the gradient impulse and the inversion impulse of 180 °, and. Due to the successive pulses at 180 ° and the inhomogeneous fields, the x component of the NMR signal will be damped within a few echoes. Therefore, we will focus only on the component and the signal. Thus, depreciating the relaxation, the first NMR echo signal can be represented as - follows: Signal = where i is the imaginary complex unit; ? it is the gyromagnetic relation; and are respectively the components x and y of the magnetization at the r site at the time of the first echo in the absence of the gradient pulse; G (r) is the component of the gradient field parallel to B0 and in the same place; d is the duration of the gradient pulse and dc (r) indicates the differential sensitivity of the NMR probes. The gradient coils 56 have a number of advantages for obtaining azimuthal measurements. First, since the axisymmetric antenna detects the echo-springs, echo trains can be recorded while the tool rotates in the hole bore. Second, coil 56 simplifies the design of an NMR tool because coil 56 does not have the tuning requirements of an NMR antenna of 36, 38. Third, the same antenna 36, 38 can be used to make symmetric and axisymmetric measurements. Fourth, coil 56 can be used to obtain NMR measurements with excellent spatial resolution, particularly vertical resolution. The present invention contemplates different ways to obtain azimuth NMR measurements. For example, a "simple deterioration of magnetic properties" mode uses at least one coil 56 to deteriorate the magnetic properties of the spins in a selected quadrant where a quadrant is defined as a segment of angular distance around the periphery of the tool. 10, however, more coils 56 can be used to magnetically deteriorate a series of quadrants. In any of the cases, two measurements are obtained: a sequence of alternating symmetric phase pulses (PAPS) with a fixed waiting time followed by a gradient PAPS that has a variable waiting time, deteriorating the magnetic properties of the selected quadrant activating coil 56 in the dial. In a preferred physical implementation of the invention, the sequence of gradient pulses mentioned above is used. By subtracting the gradient measurement from the symmetric measurement, the azimuth measurement is created. In this mode, a symmetric measurement is obtained for every two PAPS and an azimuth scan is obtained for every eight PAPS. The measurement noise for the azimuth measurement is greater than the noise - in the symmetric or gradient measurement because the two measurements are combined. It is possible to reduce the contribution of noise by combining different measurements of deterioration of magnetic properties in a single quadrant. For example, four measurements of gradient PAPS can be obtained by deteriorating the magnetic properties of each quadrant. The measurements are combined to create a symmetric and synthetic azimuth measurement. By combining measurements made without activating the gradient coils 56 with the narrow measurements by activating one or more gradient coils 56, "images" resolved axially or azimutically from the formation can be generated. The acquired data, particularly in the form of azimuthal images of porosity and fixed fluid are very desirable to improve the petrophysical interpretation in drilling for highly deviated and horizontal wells and for decision making While drilling for drilling location for geologically based wells.

Optimization of Pulse Length and Operating Frequency 3? For a chosen operating RF frequency, there is an optimum duration for the 90 ° pulse, t50, as well as for the 180 ° pulses, t180 r which ensure a desired noise ratio. The search for the optimal length of the pulses can be carried out during the master calibration of the tool, so that all the pulse lengths are correctly initiated, or when the static magnetic field changes in an unpredictable manner such as a change due to to the accumulation of magnetic waste during the drilling process. See U.S. Patent Application. No. 09 / 031,926 (Attorney File No. 24,786) assigned to Schlumberger Technology Corporation. This technique can also be used to choose the frequency of ownership to meet other criteria such as maintaining constant research depth. The optimal length of the pulses can be determined by measuring the NMR response of a sample using at least two different pulse durations and using a predefined mode independent of the NMR properties of the array. Alternatively, the optimal length of the pulses can be determined using at least two different pulse durations and additionally using a mode calculated from the NMR properties of the array. In the first case, the stacking of the data improves the signal-to-noise ratio, however, the stacking procedure may require a long period to acquire data from the formation. Preferably, the measured data accumulates during a stationary time window when the tool 10 pauses in the drilling operation, such as during the time when a new section of drill pipe is added to the drill string. In the second case, if the distribution of T2 is known, a better acquisition mode can be constructed that would provide the highest signal-to-noise ratio for a unit of acquisition time and provides an optimal linear combination of the acquired echoes. Laboratory simulations show that the optimal timing for the best acquisition mode is achieved when the duration of the echo trains is approximately equal to T, max; the T2, dominant of the formation, and when the waiting time Tw, is approximately equal to 2.5 x T2, max (assuming a constant T? / T2 ratio of 1.5). The best acquisition mode determines the optimal length of the impulses within a few percentage points over several seconds. A similar technique can be used to optimize NMR signals with respect to frequency (e.g. saddle stitch design). The T2 distribution effectively helps the efficient tuning of the pulse lengths for the tool 10.

Data Acquisition Mode As described above, the tool 10 has a series of antennas 36,38. In a preferred physical implementation of the invention, these antennas 36,38 do not transmit or acquire data simultaneously. Preferably, after an antenna 36 acquires data, the other antenna 38 experiences a minimum wait time while the power source is recharged in order to be able to transmit the next pulse sequence. The subject invention contemplates transmitting or acquiring data simultaneously. Moreover, this invention contemplates data acquisition without any waiting time requirement. Based on these design references, a variety of data acquisition modes can be used. By way of example, three representative timings for the acquisition of NMR data are described below: a rapid timing suitable for wet sandstone zones with water, a slow timing appropriate for carbonaceous zones, and a very slow timing designed for zones containing hydrocarbons (or mud invasion based on oil). The timings are presented in Table 1.

Table 1 Various different modes can be used with each data acquisition timing, including, but not limited to, the following: simple, interleaved, and bursts. The simplest way to acquire information, from T2 with the tool 10 is to carry out CPMG measurements with both antennas 36,38 using the same timing. Figure 5 illustrates the simple data acquisition mode that is used with the slow damping, fast damping and very slow damping timing of Table I. Each antenna 36.38 alternatively acquires a long sequence of long pulses which provides an effective measurement of the porosity of each 3638 antenna. With the interleaved mode, the high gradient antenna measures at least two cylindrical shells at two different frequencies while the low gradient antenna obtains a measurement using a single frequency. Figure 6 illustrates an interleaved measurement for fast damping samples, slow damping components and very slow damping components using the timing of Table I. The burst mode improves the signal-to-noise ratio especially for the fast damping components. -Additionally, the burst mode provides a useful measurement of the i based on the retained fluids. See WO 98/29639 assigned to Nuremar Corporation (describes a method for determining longitudinal relaxation times, Ti). See also U.S. Patent Application Ser. No.09 / 096, 320 (Attorney File No.24, 785) assigned to Schlumberger Technology Corporation (describes a method to polarize the fluid retained from a formation). Figure 7 illustrates a burst measurement for samples that dampen very quickly, components that are slowly damped and components that are slowly damped using slightly modified times compared to Table I. In addition to the simple, interleaved, and burst modes, with the subject invention it is possible to optimize the measurements for the evaluation of the formation by detecting downhole conditions that create a pause during the drilling operation, determining the drilling mode, and using the mode to control the acquisition of data. The standard rotary drilling operations contain many natural breaks in which the tool remains stationary: connection time when a new section of drill pipe is added to the drill string, circulation time when the mud is drilled and the drill pipe possibly spinning and tooling or clogging time when the drill string is stuck and has to be released before the drilling can restart. These natural pauses, which occur without interrupting normal drilling operations, or deliberately initiated pauses, are used to make NMR measurements. Drilling modes include, but are not limited to, drilling, sliding, firing, circulating, fishing tools, a short up and down movement, and drill pipe connections. The determination of the perforation mode improves the ability to obtain NMR measurements that take a long time or that benefit from a quiet environment, e.g. Ti, T2 antenna tuning and hydrocarbon typification. See US Patent Application. UU No.09 / 0312, 926 (Attorney File No.24.786), assigned to Schlumberger Technology Corporation. It is also possible to adjust the acquisition modes based on changes in the example environment (e.g. undermining by water, salinity, etc.) and / or changes in the NMR properties of the formation (e.g., long versus short Ti). The echo-spin amplitudes are obtained by integration between equipment of the receiver voltages through a time window. Tool 10 uses simple phase detection to measure the phase and quadrature components of the echo-spin-plus-noise signal amplitudes. The techniques revealed in the USA 5, 381,092 granted to Robert Freedman could be used to calculate the sum of the bottom hole windows and transmit the sum of the windows to the surface for the processing and presentation of the T2 inversion. Also, the techniques revealed in the USA. 5,363,041 granted to Abdurrahaman Sezginer can be implemented to use a linear operator to form a distribution of times of relaxation of echo-spins produce a decomposition of a singular value (SVD) of the linear operator, determine vectors of SDV, and compress the data of eco- spin using the vectors. Preferably the spectrum of T2 is computed downhole and transmitted to the surface. This offers the advantage of eliminating a telemetry bottleneck generated by transmitting the data required to calculate the T2 spectrum to the surface. A digital signal processor can be used to invert the T2 data. The amplitudes, Aj, of the echo-spin are characterized by the following relationship: where? j is the noise in the measurement Aj, a ± is the amplitude of the distribution T2 taken in T2 / i, represents the elements of the matrix X, where tw is the time and c is a constant (the relation T1 / T2),? T is the spacing between echoes, and j = l, 2, ... N, where N is the number of echoes collected in a single pulse sequence. In matrix notation, the equation becomes a "" being, the noise,?, Is the unknown, therefore it approaches a to find a minimum of the function You can add a term of regularization, function and function, is minimized using a suitable iterative minimization algorithm (e.g. Projection Method of the Conjugate Gradient) under the constraint that ai > 0 for i = l .... M. See Ron S. Dembo and Urrich Tulo itzski, On the Multiplication of Quadratic Functions Subject to Restrictions, Department of Computer Science, Yale. (September 1984) (describes the Method of Projection of the Conjugated Gradient). The time required to execute the investment of T2, using a digital signal processor is very reasonable. For example, assuming 1800 echoes and 30 samples in the Ti domain, the investment in a digital signal processor requires less than two seconds. For the basic NMR inversion with the tool 10, the electronic circuits apply a sequence of impulses to the formation under investigation. The tool 10 includes a pulse programmer 80, which adaptively selects and controls the pulse sequences applied to the array. The pulse programmer 80 establishes the pulse sequence using information found in the Measurement Control Block 82 (see FIG. 8) and the operating conditions of the tool 10. Preferably, the Measurement Control Block 82 is stored in a memory device down hole. The structure of the block 82 is set to allow the pulse programmer 80 to easily adapt and change the timing of the pulse sequences autonomously downhole. It is advantageous to divide a portion of the block 82 into a plurality of tables 84, 86 and 88. Instead of controlling all the operations of the tool that depend on the pulse sequence from the pulse programmer 80, tables 84, 86 and 88 they are used to control these operations. This allows the pulse programmer 80 to vary the pulse sequences without introducing contradictions in the tool configuration. The series of tables 84, 86 and 88 may include, but are not limited to, the following: an intermediate storage table describing the arrangement of the stacking warehouses, a description table defining the acquired signals accumulated in the warehouses intermediate, and a table of filter coefficients that prescribes the voltage filter used with a signal acquisition, a spin dynamics correction table that designates the dynamic corrections of the spins to be used for each buffer, and a table of data procedure that designates the characteristics of nuclear magnetic resonance calculated from intermediate stores acquired. The pulse programmer 80 includes a pulse sequence template 94, useful for generating pulse sequences comprising a sequence of states dependent on repetition and timing variables. These variables are calculated from sequence configuration parameters using the calculation block 92. The calculation block 92 can be implemented as an executable or interpretive structure. Based on the physical quantity that is to be measured as e.g. T2, the timing variables can be defined such as the waiting time, tw, the separation between echoes tecoS r the number of acquired echoes. The information parameters include, but are not limited to, t90, the amplitude of the pulses and the shape of the pulses. These parameters can be calculated periodically during the calibration of the tool 10 or during the operation of the tool 10 since these parameters can vary as the operating conditions of the tool 10 vary. For example, the amplitude and shape of the pulse depends on the quality factor of the antenna and therefore, on the conductivity of the formation surrounding the tool 10. Normally, after the pulse programmer 80 initiates a sequence of pulses, the sequence it runs deterministic merit until it is completed In order to implement certain azimuth measurement modes with the tool 10, the pulse programmer 80 has the ability to vary the pulse sequence during the execution of the sequence The programmer 80 can stop the execution of the sequence of impulses and entering a state of (STOP) until an external signal ends the state at time tc or until a maximum period has elapsed, tmax As previously discussed in the section of Data Acquisition Modes of these specifications since at least one of the different modes (interleaved) that can be used with the data description timing contemplate interspersed or several measurements, the programmer 80 compensates for the time that passes during the state (STOP) preferably the compensation is achieved by grouping the events (STOP) for example a group can comprise a pair of events of (STOP) where one of the events (STOP ) operates as described above and the other event of (STOP) is a normal event of duration of tma? - tc. Grouping the events allows the programmer 80 to combine sequences that have variable timing and terminology. Additionally, the sequence of states, as designated in the state sequence template 94, may comprise several alternatives by parts of the sequence. In real time, one of the alternatives (branching) is chosen depending on the external conditions of the tool (e.g., the azimuth of the tool). The foregoing description of the preferred Physical implementations and alternatives of the present invention have been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise form disclosed. Obviously, many modifications and variations will be apparent to those experts in the field. The physical implementations were chosen and described in order to better explain the principles of the invention and their practical applications, thus allowing other experts in the field to understand the invention for different physical implementations and with various modifications that are suitable for the particular use contemplated. It is intended that the scope of the invention be defined by the accompanying claims and their equivalents. "

Claims (22)

RE IVIND I CAC I ONE S

1. A method of sounding a well, comprising the steps of: a) generating a substantially axial symmetrical static magnetic field in a formation traversed by a borehole; b) generation of an oscillating magnetic field within the formation; c) detection of nuclear magnetic resonance signals from the formation; d) provision of a signal processor in the perforation; and e) with the signal processor, the calculation of the spin-spin relaxation time distribution from the detected signals.

2. The method of Claim 1 further comprising the step of transmitting the distribution of spin-spin relaxation times to the surface of the bore.

3. The method of Claim 1 further comprising the steps of applying a sequence of magnetic field pulses of FR to the formation, and the use of the distribution of the spin-spin relaxation times to determine the optimal length of time of each impulse that is applied to the formation.

4. The method of Claim 1 further comprising the step of detecting a plurality of signals having a signal amplitude plus noise Aj, where , where is the noise in the signal, a ± is the amplitude of the spin-spin relaxation times taken in T2, i, represents elements of the matrix X, where tw is the waiting time and c is a constant, is the spacing of the echo, t j = l, 2, ... N, where N is the number of echoes collected in a sequence of pulses.

5. The method of Claim 4 wherein the Signal amplitude plus noise is and further comprises the approach step under the restriction that > 0.

The method of Claim 5 further comprising the step of determining a minimum of the functional

7. The method of Claim 5 further comprising the steps of selecting a regularization parameter, determining a minimum of the Functional

8. The method of Claim 6 further comprising the step of minimizing the functional using a conjugate gradient projection algorithm.

9. The method of Claim 7 further comprising the step of minimizing the functional using a conjugate gradient projection algorithm.

10. The method of Claim 1 further comprising the steps of applying a magnetic field pulse sequence of FR to the formation, and using the distribution of the spin-spin relaxation times to select an optimum operating frequency. .

The method of Claim 1 further comprising the steps of applying a sequence of pulses of magnetic field of FR to the formation, and the use of the distribution of the spin-spin relaxation times to maintain a substantially constant depth of research in training.

12. An apparatus for the determination of a nuclear magnetic resonance property in a region of investigation of terrestrial formations surrounding a perforation hole, comprising: a) means for the generation of a substantially axial symmetrical static magnetic field in a formation traversed by a perforation; b) means for the generation of an oscillating magnetic field in the formation; c) means for the detection of nuclear magnetic resonance signals from the formation; and d) located in the perforation, means for calculating a distribution of spin-spin relaxation times from the detected signals.

13. The apparatus of Claim 12 further comprising means for transmitting the distribution of spin-spin relaxation times to the surface of the bore.

The apparatus of Claim 12 further comprising means for applying a sequence of magnetic field pulses to the array, and means for using the spin-spin relaxation time distribution to determine an optimal time extension of every impulse that is applied to the formation.

The apparatus of Claim 12 further comprising means for applying a sequence of magnetic field pulses to the array, and means for using the spin-spin relaxation time distribution to select an optimal operating frequency.

The apparatus of Claim 15 further comprising means for applying a sequence of magnetic field pulses to the array, and means for using the spin-spin relaxation time distribution to maintain a substantially constant depth of investigation in the formation.

17. The apparatus of Claim 12 further comprising means for detecting a plurality of signals having a signal amplitude plus noise Aj, where , where is the noise in the signal, a ± is the amplitude of the spin-spin relaxation times taken in T2 / ?, represents elements of the matrix X, where tw is the waiting time and c is a constant, is the spacing of the echo, t j = l, 2, ... N, where N is the number of echoes collected in a sequence of pulses.

18. The apparatus of Claim 17 wherein the signal amplitude plus noise is and further comprises the approach step under the restriction that = 0.

The apparatus of Claim 18 further comprising a means for determining a minimum of the functional

20. The apparatus of the Claim 18 further comprising a means for selecting a regularization parameter,, and the determination of a minimum of the functional

21. The apparatus of Claim 19 further comprising means for minimizing the functional using a conjugate gradient projection algorithm.

22. The apparatus of Claim 20 further comprising means for minimizing the functional using a conjugate gradient projection algorithm.