US20020175682A1

US20020175682A1 - Rapid nmr multi-frequency t1 and t2 acquisition for earth formations evaluation with mwd or wireline tools

Info

Publication number: US20020175682A1
Application number: US09/863,568
Authority: US
Inventors: Songhua Chen; Dan Georgi; Thomas Kruspe
Original assignee: Individual
Current assignee: Baker Hughes Holdings LLC
Priority date: 2001-05-23
Filing date: 2001-05-23
Publication date: 2002-11-28
Also published as: EP1397707A4; US6600316B2; US20030001569A1; WO2002095450A1; CA2448056A1; NO20035212D0; EP1397707A1; CA2448056C

Abstract

Nuclear magnetic resonance measurement of a medium are made using a tool with a static magnetic field having a gradient. The nuclear spins of the medium are magnetically polarized using the static magnetic field. Data are acquired from at least three sensitive volumes of the medium using an interleaved sequence of pulses at three different non-overlapping frequencies. At each frequency, a number of sub-sequences of a saturation pulse, a recovery pulse and at least one refocusing pulse are used. The saturation times between the saturation pulse and the recovery pulse for the sub-sequences are selected from a distribution of values between a minimum time to a maximum time to enable determination of T₁and T₂spectra over a wide range of values. Repetitions of sub-sequences having the same saturation at one or more frequencies may be carried out with phase alternation of the recovery pulse phases.

Description

CROSS-REFERENCES TO RELATED APPLICATIONS

This application is related to U.S. patent application Ser. No. 09/514,724.[0001]

FIELD OF THE INVENTION

The invention is related to the field of nuclear magnetic resonance (“NMR”) sensing methods and measuring techniques. More specifically, the invention is related to making NMR measurements during well logging or during Measurement-While-Drilling (“MWD”) within earth formations surrounding a wellbore. The invention also relates to methods for using NMR measurements to determine petrophysical properties of reservoir rocks and properties of fluids in the earth formations surrounding the wellbore.

BACKGROUND OF THE INVENTION

The description of the background of this invention, and the description of the invention itself are approached in the context of well logging because well logging is a well known application of NMR measurement techniques. It is to be explicitly understood that the invention is not limited to the field of well logging.

An apparatus described in U.S. Pat. No. 4,710,713 issued to Taicher et al is typical of NMR instruments used to measure certain petrophysical properties of earth formations from within a wellbore drilled through the earth formations. NMR well logging instruments such as the one disclosed by Taicher et al typically include a magnet for polarizing nuclei in the earth formations surrounding the wellbore along a static magnetic field, and at least one antenna for transmitting radio frequency (“RF”) energy pulses into the formations. The RF pulses reorient the spin axes of certain nuclei in the earth formations in a predetermined direction. As the spin axes precess and reorient themselves into alignment with the static magnetic field, they emit RF energy that can be detected by the antenna. The magnitude of the RF energy emitted by the precessing nuclei, and the rate at which the magnitude changes, are related to certain petrophysical properties of interest in the earth formations.

There are several principal operating parameters in NMR well logging which should be optimized for efficient operation of an NMR well logging instrument. These parameters include the logging speed (speed of motion of the instrument along the wellbore), the average and the peak power supplied to the instrument and transmitted as RF pulses, and the signal-to-noise ratio (“SNR”). Other parameters of interest include the vertical resolution of the instrument and the radial depth of investigation of the measurements made by the instrument within the formations surrounding the wellbore. The last two of these parameters are primarily determined by the antenna and magnet configurations of the NMR logging instrument. Improvements to these two parameters are the subject of numerous patents and other publications. Providing more flexibility in the instrument's peak power requirements, and limitations on the logging speed necessitated by the physics of NMR measurement have been more difficult to overcome.

Proton NMR relaxation time varies for different fluid types in earth formations. In addition, pore-size distributions dictate relaxation rate of wetting-fluid protons, due to the fast surface relaxation and the diffusional mixing of protons near the matrix-fluid interface with those in the middle of pores. Thus, in general, a distribution of NMR relaxation times is often observed for protons of fluids in earth formation. A large number of data points acquired in the same dynamic range is crucial to improve the accuracy and resolution of the relaxation time distribution, particularly because MWD and wireline data are known to be contaminated with high levels of random noise. There is a difference between the longitudinal relaxation time T ₁distribution and the apparent transverse relaxation time T₂distribution of fluids filling the pore spaces of the porous medium. The difference is due primarily to the effect of diffusion in the presence of magnetic field gradients. For example, light hydrocarbons and natural gas may have T₁relaxation times of the order of several seconds, while the apparent T₂relaxation times may be only about {fraction (1/100)} that amount because of diffusion when measurements are made in strong gradient magnetic fields. These field gradients can arise from the nonuniformly applied static magnetic field or from the earth formations themselves. The latter gradients are caused by differences in magnetic susceptibility between the solid portion of the earth formation (referred to as the rock “matrix”) and the fluids filling the pore spaces: both the magnitude and direction of these gradients are difficult to predict).

In wireline NMR logging, the types of NMR measurements affects the logging speed. NMR measurements often require the nuclei of the material be polarized by the static magnetic field for more than three times the longest T ₁relaxation time of any individual component within the material. This requires very slow logging speeds and, in many circumstances, are unacceptable.

Typical NMR well logging measurements use pulsed NMR techniques in which RF energy is transmitted to the measurement sensitive volume in the form of a series of pulses. The most commonly used pulse sequence for logging application is the Carr-Purcell-Meiboom-Gill (“CPMG”) pulse sequence. For well logging applications known in the art, the CPMG pulse sequences are about 0.01 to 1 seconds in total duration, depending on the number of individual pulses and the time span (“TE”) between the individual RF pulses. Each series of CPMG pulses can be referred to as a “measurement set”.

The efficiency of NMR logging is affected by the following three aspects. Firstly, the wait time between the two acquisitions cycle is dictated by the formation rock and fluid properties. Thus, maximizing the number of data and experiments to be acquired within one measurement cycle is desired. Secondly, power transmission duty cycle, defined as the ratio of the RF transmitting time vs. total time, is limited by the instrumentation design and the efficiency of heat dissipation. In some existing NMR logging tool, the duty cycle is as low as 3-4%. When the measurements are limited by duty cycle, the tool can not repeat the experiment as fast as the formation wait time allows. Thus, use of pulses or pulse sequences that minimize the RF power consumption yet provides the same information is desired. Thirdly, the number of repeated measurements depends on the required signal-to-noise ratio. Formation properties, such as porosity, affect the signal strength. The conductivity of formation and/or borehole affects the RF energy transmission efficiency and, consequently, the strength of noise. Different porosity distributions may also require different SNR in order to achieve a desired accuracy of porosity estimates. Specifically, faster relaxing components require higher SNR data compared to slower relaxing components. Thus, the number of experiment repeats is desired to be higher for the portion of the signal that represents fast relaxing protons most.

The CPMG sequence is commonly used for well logging applications because it acquires a series of NMR signal amplitudes of a vital decay range, time-spaced equally, within a single polarization cycle. Although TE is desired to be as short as the instrumentation permits, the short TE is beneficial primarily for resolving fast relaxing components. For slowly relaxing components, the choice of TE must be balanced with power requirements to avoid limiting the number of echoes acquired. It is desirable to choose the time series in accordance with the relaxation distribution scale at which one wants to resolve the spectrum, rather than taking the data equally-spaced in time. Although CPMG is efficiency in terms of a large number of echoes that can be acquired within a single polarization cycle, it is not an efficiency way to use available RF energy because the data are acquired linearly equally time-spaced while the relaxation components are logarithmically time-spaced.

For MWD, where high frequency vibrations limit experiment time, the saturation-recovery sequence for T ₁measurement (Fukushima, and Roeder, p. 169, Experimental Pulse NMR, Addison-Wesley, 1981; Taicher and Reiderman, '205; Prammer, et al., SPWLA paper #EEE, Dallas, Tex., Jun. 7, 2000) is preferred to the CPMG sequence as it can be designed to be less sensitive to vibrations by using broadband saturation pulses. Despite being a relatively fast acquisition sequence compared to the inversion-recovery sequence, the saturation-recovery MWD acquisition still takes long compared to a CPMG T₂acquisition and, thus, limits the number of measurements to less than ten in almost all cases. The limited sampling of the T₁recovery limits ones ability to decompose the recovery times into a relaxation spectrum or, even worse, differentiate between slow and fast relaxing components.

Besides being insensitive to vibration, T ₁measurements have several other merits for formation evaluation. Firstly, unlike T₂, T₁measurements are insensitive to proton self-diffusion in the presence of a magnetic field gradient, which appears as an additional decay mechanism in T₂measurements. Therefore, interpretation of formation and fluid properties from T₁is simpler. On the other hand, comparison of T₁and T₂information from the same formation system makes it possible to distinguish a fluid component that is diffusion dominant, such as hydrocarbon gas.

U.S. Pat. No. 6,049,205 to Taicher et al. (“Taicher '205”) teaches a method for determining the nuclear magnetic resonance longitudinal relaxation time T ₁of a medium. The method is a time-efficient version of the inversion-recovery sequence for T₁measurements when using multiple frequencies to excite protons in a mutually non-overlapping sensitive volumes within one polarization cycle. In the method, a number, N, of 180° “inversion” pulses, one pulse at each of the excitation frequencies, f₁through f_N, are transmitted first to invert proton spins in each of the corresponding sensitive volumes. There need be negligibly small wait time between these inversion pulses because the corresponding sensitive volumes are non-overlapping. The 180° “inversion” pulses are then followed by a first (shortest) recovery time R₁, after which a first read-out CPMG pulse sequence with a duration T_tr. A second CPMG follows, transmitted with a second frequency. The procedure is then extended to all N frequencies.

Taicher '205 also teaches the determination of the transverse relaxation time distribution of the medium with an improved signal-to-noise ratio. The medium is polarized along a static magnetic field. A first CPMG echo train is acquired from within a first sensitive volume. The first CPMG train has an inter-echo spacing and a duration long enough to determine the presence of slowly relaxing components in the medium. Then a plurality of additional CPMG echo trains is acquired. Each of the additional echo trains corresponds to a different sensitive volume, and each of the additional CPMG echo trains has an inter-echo spacing and a duration less than the duration and echo spacing of the first CPMG echo train. Different sensitive volumes are measured by transmitting each additional CPMG sequence at a different radio frequency. In the preferred embodiment, the additional echo trains have a duration and inter-echo spacing adapted to determine the presence of components in the formation having a transverse relaxation time less than about 33 milliseconds. The total duration of all the additional echo trains is about equal to the duration of the first echo train. In the preferred embodiment, the total radio frequency power transmitted in the all the additional echo trains is approximately equal to the radio frequency power transmitted in the first echo train.

SUMMARY OF THE INVENTION

The present invention includes a pulse sequence scheme to obtain T ₁and T₂data that each emphasizes a specific portion of the relaxation time distribution and porosity distribution, and combining T₁and T₂data to obtain the porosity distribution formation rock and fluid properties. The pulse sequence is a rapid method of making nuclear magnetic resonance measurement of a medium, using an NMR tool that has a well-defined static magnetic field gradient, so that the magnetic filed strength varies spatially. The nuclear spins of the medium are magnetically polarized under the static magnetic field. Data are acquired at at least two different radio frequencies from at least two different sensitive volumes of the medium. In a preferred embodiment of the invention, more than two frequencies are used to speed the data acquisition and improve the data quality. For each of the at least two sensitive volumes, an RF pulse sequence is defined, each pulse sequence comprising at least one sub-sequence having a saturation pulse, a recovery pulse that follows the saturation pulse after a defined saturation time, and at least one refocusing pulse. In a preferred embodiment of the invention, a series of refocusing pulses is used so that a series of echoes are acquired from which a T₂distribution, corresponding to either fully or partially polarized signals, can be derived. The wait times for the sub-sequences of the various sensitive volumes are selected from a distribution of values between a minimum time to a maximum time. In a preferred embodiment of the invention, the data acquisitions for T₁and T₂decay data are interleaved among multiple frequencies. Specifically, during the wait time between the saturation pulse and the recovery pulse of the first frequency activation, data acquisition of one or more additional frequencies are started and data acquisition of these frequencies may be completed before or after the data acquisition of the first frequency. The interleaving pattern can be either regular or irregular. The interleaving may be nested. Typically, the minimum and maximum wait times are set at 0.1 ms. and 10 seconds. The distribution of wait times may follow a power law relationship approximately to be in accordance with the desired resolution scale of the T₁and T₂spectrum. The wait times for two sub-sequences at a particular frequency may be the same: in such a case, the corresponding refocusing pulses are phase alternated, making it possible to reduce the effects of ringing.

In a preferred embodiment of the invention, the bandwidth of the saturation pulse or pulses for a sensitive volume is greater than or equal to the bandwidth of the corresponding recovery and refocusing pulses. The frequency separation of the RF pulses is chosen to be greater than the bandwidth of the RF excitation pulses in the sequence.

In another embodiment of the invention, echo trains acquired with different wait times can be stacked to improve the signal to noise ratio. The resultant echo train is inverted to obtain a T ₂spectrum. Preferably, only the fully polarized T₂components (fast relaxing components) may be analyzed and partially polarized components are discarded. The number of echoes (NE) following individual recovery pulse can be the same or different. Even if the individual NE parameters are different, they can still be averaged. The T₂data are used to interpret earth formation and reservoir fluid properties that have characteristically fast relaxation times while the T₁spectrum is used to interpret properties that have characteristically longer relaxation times (e.g., greater than 3 ms). The total porosity is obtained from the echo train acquired after the longest wait time after the saturation pulse. This wait time is sufficiently long to ensure the achievement of full polarization of all relaxation-time components. The total porosity is used as a constraint in the process to combine T₁and T₂spectra.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows a graph of amplitude of the static magnetic field of the magnet in a gradient NMR well logging apparatus used with the invention. [0018]
FIG. 2 shows an example of an acquisition scheme according to the present invention. [0019]
FIGS. 3A and 3B show examples of regular and irregular interleaving of frequencies. [0020]
FIG. 4 is a flow chart illustrating the use of the interleaved pulse sequences for obtaining T[0021] ₁and T₂spectra.
FIG. 5 shows the result of using the sum of echos, SE, as an approximation of free induction decay, M[0022] ₀, in T₁analysis.

DESCRIPTION OF THE PREFERRED EMBODIMENT

An NMR well logging apparatus which is suitable for use with this invention is described, for example, in U.S. Pat. No. 5,712,566 to Taicher et al (“Taicher '566”). The apparatus described in Taicher '566 includes a magnet for inducing a static magnetic field in the earth formations. The static magnetic field includes an amplitude gradient directed radially inwardly towards the longitudinal axis of the instrument. The apparatus disclosed in Taicher '566 includes an antenna through which pulses of RF power are conducted to excite nuclei of the earth formations surrounding the instrument. The antenna includes a wire coil wound around a high magnetic permeability ferrite. The ferrite includes a frequency control coil wound thereon. By passing a selectively controllable DC voltage through the frequency control coil, the tuning frequency of the antenna can be selectively controlled, making transmission and reception of RF energy at the selected frequency. The apparatus disclosed in Taicher '566 can make NMR measurements at a plurality of different frequencies. Since the static magnetic field imparted by the magnet disclosed in Taicher '566 includes an amplitude gradient, conducting NMR measurements at different frequencies will result in these different frequency NMR measurements taking place in different sensitive (excitation) volumes. [0023]
It is to be clearly understood that the apparatus disclosed Taicher '566 is not the only apparatus that can be used for this invention. For purposes of this invention it is only necessary that the NMR apparatus be able to selectively excite different sensitive volumes to nuclear magnetic resonance, and selectively receive NMR signals from each of the selectively excited sensitive volumes. Using multiple frequencies for individual NMR measurement sequences in a gradient static magnetic field is a particularly convenient means by which to carry out the method of this invention, and so the apparatus disclosed in Taicher '566 is a particularly convenient instrument, but not the exclusive instrument by which to carry out the method of this invention. [0024]
FIG. 1 shows a graph of the amplitude of the static magnetic field, with respect to distance from the magnet, for the well logging apparatus described in Taicher '566. The amplitude of the static magnetic field generally decreases with respect to the lateral distance from the magnet. As is well known in the art, nuclear magnetic resonance conditions occur when a radio frequency magnetic field is applied to materials polarized along a static magnetic field where the frequency of the RF magnetic field matches the product of the static magnetic field strength and the gyromagnetic ratio of the nuclei being polarized by the static magnetic field, this product being referred to as the Larmor frequency. As can be inferred from the graph in FIG. 1, by adjusting the frequency of the RF magnetic field, the distance from the magnet at which nuclear magnetic resonance conditions occur can be changed corresponding to the static magnetic field amplitude at that particular distance from the magnet. For example, if frequency f[0025] ₁is the highest frequency, resonance will occur at the smallest distance to the magnet, and so on through lower frequencies f₂through f_N. Because nuclear magnetic resonance only occurs where the static magnetic field strength matches the RF magnetic field frequency, nuclear magnetic resonance measurements can be conducted within a number of different non-overlapping sensitive volumes by inducing nuclear magnetic resonance at different frequencies. A particular set of non-overlapping sensitive volumes which would result when using the apparatus described in Taicher '566, for example, would comprise thin annular cylinders each having an average radius corresponding to the particular static magnetic field amplitude in which nuclear magnetic resonance would occur at a particular RF magnetic field frequency. The thickness of each annular cylinder would be related to the bandwidth of the transmitted RF pulses, the bandwidth of a receiver circuit in the NMR instrument and the rate at which the static magnetic field changes in amplitude.
This feature of the static magnetic field, and the selectable frequency capability for the RF magnetic field in the apparatus described in Taicher '566 makes it possible to conduct time-overlapping NMR measurements within different sensitive volumes. By time-overlapping NMR experiments in different sensitive volumes, it is possible to more efficiently use the RF transmitting components in the apparatus. The manner in which the RF transmitting components are used more efficiently will now be explained. [0026]
By way of example, the interleaving of frequencies in the present invention is illustrated using a three-frequency acquisition sequence. The pulse sequences at the three frequencies are interleaved. The method of the present invention requires a minimum of two frequencies in order to provide the benefit of a reduced total acquisition time compared to prior art methods. Referring now to FIG. 2, three defined [0027] pulse sequences 103, 105 and 107 are illustrated. These correspond to different RF frequencies f₁, f₂and f₃and different regions of investigation as illustrated in FIG. 1. For clarifying the invention, the three different defined pulse sequences are shown separately, though in actual practice they would be interleaved. The abscissa 101 is time, and, to simplify the illustration, the scale is logarithmic.
At each of the three frequencies, the defined pulses are depicted according to a code: this too is for the purpose of simplifying the illustration. The three different pulse types, saturation, recovery, refocusing pulses are represented by different line types. The long, thin lines such as [0028] 121 a, 121 b, 131 c represents saturation RF pulses; the short, thin lines such as 131 a, 131 b, 131 c represent the recovery RF pulses; and the short, thick lines such as 141 a, 141 b, 141 c represents one or more refocusing pulses. For simplifying the illustration, the individual refocusing pulses within each group such as 141 a, 141 b, 141 c are not shown. The saturation pulse comprises one or more 90° pulses while the recovery pulse is a 90° pulse.
Those versed in the art would recognize that each of the refocusing pulses with a group such as [0029] 141 a, 141 b, 141 c is followed by an echo (also not shown). These echos are received by a suitable antenna on the NMR instrument. The pulses of same types are identical in pulse shapes and the flip angles for all frequencies. In a preferred embodiment of the invention, the time τ at each frequency between a recovery (R) pulse and the first refocusing pulses (F) pulse is fixed though the time τ may be different for the different frequencies. The time T_wbetween a saturation (S) pulse and the following recovery pulse is variable. Each frequency is individually calibrated: as would be known to those versed in the art, data acquired from the same sample, e.g. in a water tank, may have different quality factor (Q), different bandwidth, etc. This can result in signal strength differences when making measurements on the same sample. The calibration procedure compensates for these differences. This is a routine procedure for multi-frequency tools.
The defined pulse sequence corresponding to a single frequency such as f[0030] ₁may then be viewed as comprising a plurality of sub-sequences such as 133 a, 135 a. Each of the sub-sequences comprises a saturation pulse followed by a recovery pulse delayed by a time T_wrelative to the saturation pulse. The recovery pulses is, in turn, followed by at least one refocusing pulse after a time τ. The time delay T_wmay be referred to as a saturation time.
The RF pulses are frequency-selective pulses. The S pulses preferably are broader band, broader than the R and F pulses. However, in order to avoid interference with pulses having an adjacent frequency, the bandwidth of the S pulses must be less than the frequency separations between two adjacent frequencies. The F pulses could be between 90° and 180°. The use of such refocusing pulses of less than 180° for reducing power consumption has been taught in U.S. Pat. No. 6,163,153 to Reiderman at al. In a preferred embodiment of the invention, the phase of the R pulses at a particular frequency is alternated between + and −90°. In a preferred embodiment of the invention, the time T is between 0.2-0.5 ms but other values may be used when appropriate. [0031]
For a medium that can be represented by a discrete T[0032] ₁spectrum with components T_1k, the total polarization following a time T_wis given by $\begin{matrix} M (T_{w}) = \sum_{k} [M_{0 k} (1 - e^{- T_{w} / T_{1 k}})] & (1) \end{matrix}$
where M[0033] _0krepresents the density of the T₁spectrum.
As an example, to cover a range of relaxation times from 1 ms. to 10 seconds, a set of 22 T[0034] _wvalues can be chosen as 3^kms, where k=[0, 0.4, 0.8, . . . 8, 8.4] by the following values:
[1, 1.55, 2.41, 3.74, 5.80, 9.0, 14.0, 21.7, 33.6, 52.2, 81.0, 126.0, 195.0, 303.0, 470.0, 729.0, 1131.0, 1756.0, 2724.0, 4228.0,6571.0, 10181.0][0035]
These values then can be assigned among all frequencies. The way to assign these T[0036] _wvalues is arbitrary, and it is not necessary that the long and short T_wvalues are approximately equally divided among these frequencies. The longest T_wdata acquisition spans across several short T_wintervals. The actual T_wvalues only need to be approximately those shown above; exact values are not required. The flexibility is important to avoid overlapping of the rf pulses from different frequencies to occur at an exact time instance.
The power law dependency for the times T[0037] _wresults in a substantially constant resolvability of the T₁spectrum. A conceptually similar yet different realization method has been discussed with respect to resolving the T₂spectrum in co-pending U.S. patent application Ser. No. 09/778,554 of Edwards, having the same assignee as the present application and follows from a discussion in an article entitled “Quantitative Interpretation of NMR Relaxation Data” in the Journal of Magnetic Resonance, 1989, pp 134-152, Whitall and Mackay (1989). In that approach, a variable number of repeated experiments are used in averaging CPMG echo data, with higher multiples used for early echoes than the later ones. Such an approach is applicable for measurements where the time interval between data points being a constant (i.e., linear scale) but the solution of the relaxation spectrum to be resolved is not in linear scale (i.e., logarithmic scale). In the present art, the T_wvalues are not linearly time-spaced but chosen to be commensurate to the scale in which the underlying T₁spectrum is resolved.
If more frequencies are used, the T[0038] _wvalue can be chosen more closely spaced, e.g., using 3^kms, where k [0, 0.2, 0.4, 0.6 . . . 8.0, 8.2, 8.4] for 43 values covering the same range (1 ms to 10 s). The actual T_wvalues should be chosen in accordance with the formation characteristics, the in-situ fluid types, and the time available to complete the experiment.
In a preferred embodiment of the invention, the frequency separation between adjacent RF frequencies is chosen such that negligible overlap between the adjacent frequency bands occurs. If more than two frequencies are used, the individual separations between the adjacent frequencies can be the same or different. [0039]
In an alternate embodiment of the invention, the pulse acquisition method described above is generalized to include repeats of a few T[0040] _wvalues and the repeats can be acquired with the same or different frequencies. The repetition makes it possible to improve the SNR corresponding to short T_wvalues for the weakest signals. This is useful in analyzing earth formations in which the short relaxation time components (e.g., capillary bound fluid) are important.
The acquisition pulse sequence used in the invention is then obtained by interleaving the defined pulse sequences for the individual frequencies f[0041] ₁, f₂and f₃. As noted above, the pulse sequences for the individual frequencies f₁, f₂and f₃in turn, are made up of sub-sequences comprising a saturation pulse, a recovery pulse and one or more refocusing pulses. As shown in FIG. 2, the interleaving is done in such a manner that the recovery pulse, the at least one refocusing pulse, and the subsequent saturation pulse at a particular frequency fall within the duration of a saturation time T_wof another frequency.
As noted above, the interleaving of frequencies may be regular or irregular. FIG. 3[0042] a shows an example of regular interleaving of frequencies f₁. . . f₇. FIG. 3b shows an example of irregular interleaving of frequencies. The same coding as in FIG. 2 is used, i.e., a single tall pulse is a saturation pulse, a single short pulse is a recovery pulse and a block represents a series of refocusing pulses and corresponding echos.
In an alternate embodiment of the invention, the methods described above (single values of T[0043] _wor repetitions of T_w) can be further generalized to include T₂measurements of components in the very short T₂range. This is achieved by specifically incorporating several refocusing pulses with an equal spacing of TE=2τ and analyzing the corresponding echoes. The number of these refocusing pulses is generally in the neighborhood of ten. The purpose of these short echo trains are for the estimation of fast decay components, i.e., to estimate the T₂spectra from the minimally recoverable T₂component to a pre-defined T_2cIn the present embodiment, T_2ccan be chosen as 3 ms, the commonly used cutoff for clay bound water porosities. However, the value of T_2cis not limited as such. The echo signal Echo₁may be denoted by $\begin{matrix} {Echo}_{l} = \sum_{j} M_{0 j} e^{- l \cdot TE / T_{2 j}} + \sum_{k} M_{0 k} e^{- l \cdot TE / T_{2 k}} (1 - e^{- T_{w} / T_{1 k}}) & (2) \end{matrix}$
The first term on the right hand side of the eq. (2) represents the contributions from the fast relaxing components that are fully polarized following a wait time of T[0044] _wwhile the second term on the right hand side of eq. (2) represents the contributions from the slower relaxing components that are only partially polarized following a wait time of T_w. The number of relaxation components that are fully polarized is of course dependent on T_w, however, since the purpose of using the T₂spectra is to estimate those components with T₂≦T₂, the polarization status for the slower relaxation components with T₂≧T_2c, is not important. Consequently, the echo train acquired after the individual T_wmvalues can be averaged to improve signal-to-noise ratio as long as the averaging process includes only those echo trains acquired with T_wm≧T_wcwhere T_wcis sufficiently long to allow porosity components corresponding to T₂≦T2c being substantially fully polarized. Summing eq. (2) over the echo trains corresponding to the plurality of wait times gives the result $\begin{matrix} \sum_{m} {Echo}_{l} (T_{wm}) = \sum_{m} [\sum_{j} M_{0 j} e^{- l \cdot TE / T_{2 j}} + \sum_{k} M_{0 k} e^{- l \cdot TE / T_{2 k}} (1 - e^{- T_{wm} / T_{1 k}})] . & (3) \end{matrix}$
Turning now to FIG. 4, a flow chart of various steps that may be taken in the analysis of single or multi-frequency data using the interleaved sequences is shown. Denoting by E[0045] _i,j,kthe j-th echo for the i-th frequency with the k-th wait time, the following steps may be carried out for the calibrated data 200 for the individual frequencies.
The branch corresponding to [0046] 300 relates to the first echo E_i,j,kfor the one or more frequencies (index i) and the one or more wait times (index k). The pulse echo data are averaged vertically 301 if necessary to improve SNR and then the data from the different frequencies and different wait times are combined to form time domain polarization buildup data for T₁decay analysis 311 as shown in the box 302. These are indicative of saturation recovery. The abscissa is the wait time T_wand the ordinate is the amplitude of the corresponding first pulse echo. Inversion of this curve using prior art methods gives an estimate of the T₁spectrum of the formation.
In another embodiment, all echoes, E[0047] _i,j,kacquired in each frequency (index i) are used to estimate $\begin{matrix} E_{i, 0, k} = \sum_{u} M_{0 u} + \sum_{v} M_{0 v} (1 - e^{- T_{wk} / T_{1 v}}) & (4) \end{matrix}$
by extrapolation of echo to j=0. After this, E[0048] _i,0,kis used to replace E_i,l,kin box 300. The use of a summation of echos has been described in Taicher '205 in a method of analyzing T₁from SR/CPMG and IR/CPMG. In that method, only the sum of the echoes in each CPMG sequence need be measured in order to increase the overall signal-to-noise ratio. The substitution of the sum of echos, SE, for free induction decay, M₀, in T₁estimation is applicable for systems that contain a single component T₁and T₂or can be approximated with a single component. However, the accuracy may not be as good when the relaxation time distribution is broad, as illustrated in FIG. 5. In this figure, Plots (a) through (c) represent single exponential decay and (d) through (f) a broad, bimodal distribution. The leftmost plots show the distribution, the middle ones the sum of echoes (SE) divided by the free induction decay (M₀), and the rightmost the SE and M(0) as functions of wait time TW. The ratio of SE/M₀is constant for the single exponential case but varies for multi components. For the latter, one can not use the summation of echoes to substitute the FID signals.
The branch corresponding to [0049] 400 relates to those echo trains which are acquired with a wait time T_wk≧3 T_l,max, where T_l,maxis a maximum value of the expected longitudinal relaxation time of the formation. These correspond to full polarization of the spins in the formation prior to the application of the refocusing pulses. The data are vertically averaged if necessary for improving the SNR 401. If more than one values of k satisfies the condition T_wk≧3 T_l,max, then these echo trains may be averaged 402 over i and k (i.e., over frequencies and wait times) to give an averaged echo train 411 as shown in the box 403 where the abscissa is the echo number. The curve 411 corresponds to full polarization of the nuclear spins following the saturation pulse and extrapolation of the curve 411 back to the origin gives the total porosity φ₁.
The branch corresponding to [0050] 500 involves processing of all the echoes acquired using the interleaved pulse sequences described above. The data are vertically averaged if necessary for improving the SNR 401. Next, all the echo trains are averaged over the indices i and k (i.e., over frequency and wait times) to give an averaged summed pulse sequence echo sequence E (j τ) 511. Inversion of the curve 511 using prior art methods gives the T₂spectrum 513.
The portion of the T[0051] ₂spectrum 513 less than a predetermined threshold 515 corresponds to fully polarized nuclear spins with a short relaxation time T₂. Information from the inverted T₁from 303 may be combined with information from that portion of the T₂spectrum 513 less than the predetermined threshold 515 using the total porosity φ_tfrom 403 as a constraint to give a characterization of the fluid and porosity distribution in the formation. Aside from the enhanced decay due to diffusion in a gradient field, reservoir fluids T₁and T₂share the same relaxation mechanisms which forms the basis for combining the T₁and T₂information to form one single porosity spectrum. For example, the T₂data provides formation from the minimal T₂to T_2c, while the T₁data provides porosity distribution for T₁≧T_2c. However, because T₁/T₂is greater than unity and the value is difficult to predict, a small overlap between T₁and T₂distributions is possible. Two methods are used to reduce the effect. In the first method, the T₁buildup data and the T₂decay data are used together as inputs for inversion processing. In the second method, the collective porosity values from the T₁and T₂spectra: $\begin{matrix} φ = \sum_{u : T_{2 u} \leq T_{2 c}} M_{0 u} + \sum_{v : T_{1 v} > T_{2 c}} M_{0 v} & (5) \end{matrix}$
must be in agreement with the total porosity φ[0052] _testimation. If Δφ=φ′−φ_t>0 the excessive porosity Δφ is considered to be due to overlapping and the amount is subtracted from the shortest T₁components.
In MWD/LWD applications, tool motion may prevent acquiring valid T[0053] ₂data, because tool motion gives rise to an apparent faster decay of the echo signal and makes it difficult to measure the intrinsic T₂. Acquisition of a valid echo train with a substantially long acquisition time interval (e.g. is) requires minimal vibration for the entire T₂acquisition, something that is difficult to maintain during MWD/LWD operations. If the measurement is started when there is little motion, then at least the early echoes in the T₂acquisition will be valid and representative of intrinsic T₂processes. With the aid of auxiliary sensors (e.g. accelerometers) it is possible to determine how many of the initial echoes are valid. The echoes that are not corrupted by tool motion can then be used for analysis.
In the present invention, the minimization of the detrimental effects of tool vibrations to NMR measurements are achieved using two approaches, applied concurrently or separately. Firstly, the saturation pulses are made broader band than the recovery and refocusing pulses. Secondly, for T[0054] ₁measurements, an important step is the synchronization of sensitive volumes defined by the saturation pulses and the recovery/refocusing pulses. When the bandwidth of the saturation pulse is broader than the recovery and refocusing pulses, it simply means that the sensitive volume defined by the recovery and recovery pulses must be substantially inside the sensitive volume defined by the saturation pulse. The synchronization of sensitive volume is only stringent at the moments when these pulses are applied. Any tool vibration that occurs during the wait time period, i.e., the time period between the saturation pulse and the recovery pulse, has no effect on the measurements. Thus, the vibration of the tool and the position of the sensitive volume can be sensed and the sensor data can be used to trigger the RF pulsing events.
The vibration-sensor trigger events can be implemented in three different ways to control RF pulsing events at three critical time instances of any subsequence. The first time instance is the onset of the saturation pulse. The saturation pulse trigger defines the sensitive volume position. It is preferred that the event is triggered at the moment the sensitive volume is positioned close to its equilibrium position, regardless of whether or not the tool is vibrating. The second time instance is the onset of the recovery pulse. The recovery pulse trigger activates when tool vibration is minimal and the sensitive volume substantially matches that of the saturation pulse. The third time instance is to control the end of the valid echo acquisition. This trigger activates when the motion is substantial so that the sensitive volume of the last echo is asynchronized from that of the first echo in the echo train. [0055]
In one embodiment of the invention, motion triggered pulsing is performed using a method disclosed in U.S. patent application Ser. No. 09/778,295 of Slade et al, the contents of which are fully incorporated herein by reference. In the teachings of Slade, a predictive filter, such as a Kalman filter, is used to analyze the output of motion sensors on the MWD tool. The motion sensor can be used to determine when to start echo train acquisition and when to stop the echo train. Using the output of the predictive filter, the acquisition of the CPMG echo trains may be delayed or speeded up relative to values selected on the basis of the methodology discussed above. [0056]
In one embodiment of the invention, MWD/LWD are made using a sensor assembly on a non-rotating sleeve that may be clamped to the borehole wall. Such an assembly is disclosed in U.S. patent application Ser. No. 09/247,340 having the same assignee and an inventor in common with the present application, now U.S. Pat. No. ______, the contents of which are fully incorporated herein by reference. This minimizes the effect of tool motion effects on T[0057] ₂echo data.
As would be known to those versed in the art, in MWD applications the rate of movement of the tool along the borehole may be small. In MWD applications, the processing steps described above with reference to FIG. 4 may be used even at a single frequency. This may also be done with respect to wireline logging if the expense of additional logging time can be justified. [0058]
In one embodiment of the invention, the saturation pulses in each subsequences can be omitted because they are always preceded by the refocusing pulses of the prior subsequence. In the preferred embodiment, however, saturation pulses are always retained because they provide a more precise means to control the desired wait times. The precise timing become more important when the saturation pulses are also motion-sensor triggered, as described above. PCT publication WO01/13142 based on U.S. Provisional Patent Application No. 60/148,943 describes a pulse sequence in which the saturation pulses are absent. [0059]
In an alternate embodiment of the invention, the saturation pulses described above may be replaced by inversion pulses. However, saturation recovery pulse sequences acquire data faster than inversion recovery sequences and are hence preferred. [0060]
While the foregoing disclosure is directed to the preferred embodiments of the invention, various modifications will be apparent to those skilled in the art. It is intended that all variations within the scope and spirit of the appended claims be embraced by the foregoing disclosure. [0061]

Claims

What is claimed is:

1. A method of nuclear magnetic resonance measurement of a medium, comprising:

(a) magnetically polarizing nuclei in said medium with a static magnetic field;

(b) defining at least two sensitive volumes of the medium, each said sensitive volume having an associated frequency of a radio frequency (RF) signal;

(c) defining a pulse sequence associated with each of the at least two sensitive volumes, each said defined pulse sequence comprising at least one sub-sequence having a saturation pulse, a recovery pulse, at least one refocusing pulse, and a wait time between the saturation pulse and the recovery pulse;

(d) defining a distribution of time intervals ranging from a minimum time to a maximum time;

(e) selecting said plurality of wait times as having values substantially equal to time intervals from said distribution of time intervals; and

(f) acquiring NMR pulse echo data using an acquisition pulse sequence comprising an interleaved sequence of said defined pulse sequences having said selected wait times.

2. The method of claim 1 wherein the each of the at least one recovery pulse, saturation pulse and refocusing pulse of any of said defined pulse sequences are substantially non-overlapping with the each of the at least recovery pulse, saturation pulse refocusing pulse of any of the other defined pulse sequences.

3. The method of claim 2 wherein a recovery pulse, the at least one refocusing pulse and a subsequent saturation pulse of at least one sub-sequence associated with one of the at least two sensitive volumes occurs during a wait time following a saturation pulse of a sub-sequence of another of the at least two sensitive volumes.

4. The method of claim 1 wherein said interleaved sequences are one of (i) regularly interleaved, and, (ii) irregularly interleaved.

5. The method of claim 1 wherein said distribution of time intervals further comprises intervals of duration 3^nkms. wherein n is an integer and k is a non-negative number.

6. The method of claim 5 wherein said distribution of time intervals further comprises between 22 and 43 time intervals.

7. The method of claim 1 wherein a time interval between a recovery pulse and a subsequent at least one refocusing pulse of each sub-sequence of at least one of said sensitive volumes has a defined value.

8. The method of claim 1 wherein a bandwidth of a saturation pulse is greater than or equal to a bandwidth of a recovery pulse and a refocusing pulse for each sub-sequence of each defined pulse sequence.

9. The method of claim 1 wherein, for each combination of the at least two sensitive volumes, a bandwidth of a saturation pulse for each of the defined pulse sequences is less than a separation of said associated frequencies.

10. The method of claim 9 wherein said separation is less than 40 kHz.

11. The method of claim 1 wherein a tipping angle of a refocusing pulse lies between 90° and 180°.

12. The method of claim 1 wherein said plurality of sub-sequences further comprises at least two sub-sequences at at least one frequency having the same wait time.

13. The method of claim 12 wherein the at least one refocusing pulse of a first and a second of said at least two sub-sequences have phases of +90° and −90°.

14. The method of claim 1 further comprising determining from said NMR pulse echo data a T₁spectrum of said medium.

15. The method of claim 14 wherein determining the T₁spectrum of said medium further comprises forming a time domain T₁decay data by combining recovery pulse signals for the at least two sensitive volumes and the associated sub-sequences.

16. The method of claim 15 wherein determining the T₁spectrum of said medium further comprises inverting said time domain T₁decay data.

17. The method of claim 1 wherein the defined pulse sequence associated with one of the at least two sensitive volumes comprises a plurality of sub-sequences having different wait times, each of said plurality of sub-sequences comprising a plurality of refocusing pulses, the method further comprising determining from said NMR pulse echo data a T₂spectrum of said medium.

18. The method of claim 17 wherein determining the T₂spectrum of said medium further comprises averaging said pulse echo data over the at least two sensitive volumes and the plurality of wait times and inverting said averaged pulse echo data.

19. The method of claim 15 wherein determining said T₂spectrum of said medium further comprises determining, for at least one of the sequence of refocusing pulses, a sum of values of an associated echo over a subset of the plurality of sub-sequences.

20. The method of claim 17 further comprising determining from said NMR pulse echo data a T₁spectrum of said medium.

21. The method of claim 20 wherein determining the T₁spectrum of said medium further comprises forming a time domain T₁decay data by combining recovery pulse signals for the at least two sensitive volumes and the associated sub-sequences.

22. The method of claim 21 wherein determining the T₁spectrum of said medium further comprises inverting said time domain T₁decay data.

23. The method of claim 22 further comprising combining the determined T₁spectrum of said medium and the determined T₂spectrum of said medium.

24. The method of claim 23 wherein combining the determined T₁and T₂spectra further comprises determining a total porosity of the medium and using said determined total porosity as a constrain.

25. A method of determining a parameter of interest of subsurface earth formations, comprising:

(a) conveying a logging tool into a borehole in the earth formations;

(b) using a magnet on the logging tool for magnetically polarizing nuclei in said medium with a static magnetic field;

(c) defining at least two sensitive volumes of the earth formations, each said sensitive volume having an associated frequency of a radio frequency (RF) signal;

(d) defining a pulse sequence associated with each of the at least two sensitive volumes, each said defined pulse sequence comprising at least one sub-sequence having a saturation pulse, a recovery pulse, a plurality of refocusing pulses, and a wait time between the saturation pulse and the recovery pulse;

(e) defining a distribution of time intervals ranging from a minimum time to a maximum time;

(f) selecting said plurality of wait times as having values substantially equal to time intervals from said distribution of time intervals; and

(g) using at least one antenna on the logging tool and acquiring NMR pulse echo data using an acquisition pulse sequence comprising an interleaved sequence of said defined pulse sequences having said selected wait times.

26. The method of claim 25 wherein said logging tool is conveyed into the borehole on one of (i) a wireline, (ii) a drillstring, and, (iii) coiled tubing.

27. The method of claim 26 wherein the plurality of recovery pulses, saturation pulses and the at least one refocusing pulse of any of said defined pulse sequences are substantially non-overlapping with the plurality of recovery pulses, saturation pulses and the at least one refocusing pulse of any of the other defined pulse sequences.

28. The method of claim 25 wherein a time interval between a recovery pulse and a first subsequent refocusing pulse of each sub-sequence of at least one of said sensitive volumes has a defined value.

29. The method of claim 25 wherein said logging tool is part of a bottom hole assembly used in drilling, the method further comprising:

(i) using motion sensors on the bottom hole assembly to produce signals indicative of motion of said bottom hole assembly and the logging tool, and

(ii) adjusting said selected time of at least one of: (A) the saturation pulse, (B) the recovery pulse, (C) one of the refocusing pulses, and, (D) a last one of said refocusing pulses, in response to said motion signals.

30. The method of claim 29 further comprising using a predictive filter to analyze said motion signals.

31. The method of claim 25 wherein the defined pulse sequence associated with one of the at least two sensitive volumes comprises a plurality of sub-sequences having different wait times, each of said plurality of sub-sequences comprising a plurality of refocusing pulses, the method further comprising determining from said NMR pulse echo data a T₂spectrum of said earth formations.

32. The method of claim 31 wherein determining the T₂spectrum of said earth formations further comprises averaging said pulse echo data over the at least two sensitive volumes and the plurality of wait times and inverting said averaged pulse echo data.

33. The method of claim 32 wherein determining said T₂spectrum of said earth formations further comprises determining, for at least one of the sequence of refocusing pulses, a sum of values of an associated echo over a subset of the plurality of sub-sequences.

34. The method of claim 33 further comprising determining from said NMR pulse echo data a T₁spectrum of said earth formations.

35. The method of claim 31 wherein determining the T₁spectrum of said earth formations further comprises forming a time domain T₁decay data by combining recovery pulse signals for the at least two sensitive volumes and the associated sub-sequences and inverting said time domain T₁decay data.

36. The method of claim 35 further comprising combining the determined T₁spectrum of said earth formations and the determined T₂spectrum of said earth formations.

37. A method of determining a parameter of interest of subsurface earth formations, comprising:

(a) conveying a logging tool into a borehole in the earth formations;

(c) defining a sensitive volume of the earth formations, said sensitive volume having an associated frequency of a radio frequency (RF) signal;

(e) defining a pulse sequence associated with the sensitive volume, said defined pulse sequence comprising a plurality of sub-sequences having a saturation pulse, a recovery pulse, a plurality of refocusing pulses, and a wait time between the saturation pulse and the recovery pulse; and

(f) using at least one antenna on the logging tool and acquiring NMR pulse echo data using an acquisition pulse sequence comprising an interleaved sequence of said defined pulse sequences having said selected wait times;

wherein said plurality of wait times have values substantially equal to time intervals from said distribution of time intervals.

38. The method of claim 37 further comprising determining a T₁spectrum and a T₂spectrum of the earth formations.

39. The method of claim 38 further comprising combining said T₁spectrum and said T₂spectrum of the earth formations.

40. A method of determining a parameter of interest of subsurface earth formations, comprising:

(a) conveying a logging tool into a borehole in the earth formations;

(b) using a motion sensor on the logging tool to produce signals indicative of motion thereof;

(c) using a magnet on the logging tool for magnetically polarizing nuclei in said medium with a static magnetic field;

(d) defining a sensitive volume of the earth formations, said sensitive volume having an associated frequency of a radio frequency (RF) signal;

(e) defining a distribution of time intervals ranging from a minimum time to a maximum time and selecting;

(f) defining a pulse sequence associated with the sensitive volume, said defined pulse sequence comprising a plurality of sub-sequences having a saturation pulse, a recovery pulse, a plurality of refocusing pulses, and a wait time between the saturation pulse and the recovery pulse; and

(g) using at least one antenna on the logging tool and acquiring NMR pulse echo data using an acquisition pulse sequence comprising an interleaved sequence of said defined pulse sequences having said selected wait times;

wherein said plurality of wait times have values substantially equal to time intervals from said distribution of time intervals, and wherein, for at least one sub-sequence, the plurality of refocusing pulses and their times are based in part on said motion signals.

41. The method of claim 40 further comprising using a predictive filter with said motion signals for determining times of the plurality of refocusing pulses.

42. The method of claim 40 further comprising determining from said NMR pulse echo data a T₂spectrum of said earth formations.

43. The method of claim 42 wherein determining the T₂spectrum of said earth formations further comprises averaging said pulse echo data over the plurality of wait times.

44. The method of claim 43 further comprising inverting said averaged pulse echo data.

45. The method of claim 42 further comprising determining from said NMR pulse echo data a T₁spectrum of said earth formations

46. The method of claim 45 further comprising combining the determined T₁spectrum of said earth formations and the determined T₂spectrum of said earth formations.