DIAGRAFIA FOR NUCLEAR MAGNETIC RESONANCE OF NATURAL GAS IN DEPOSITSBACKGROUND OF THE INVENTIONThe present invention relates to a method for determining a characteristic of a porous earth formation containing a hydrocarbon gas, the formation of earth is penetrated by a sounding.
BACKGROUND OF THE INVENTIONIn the exploitation of hydrocarbon deposits, well logging is used to determine the amount of recoverable hydrocarbons. The tools of diagrafia such as the registers or profiles of the density, of the neutrons, and of the resistivity, have been developed for the measurement of properties of the deposit such as the porosity and the saturation of water and hydrocarbons of the space of the pores. These tools are not widely used in the petroleum industry. However, to determine the exact porosity using these tools, you must know the lithology of the rock. Many other properties of rock and fluids, such as salinity, the cement factor, the saturation exponent, and the shaleRef. 24472, must be known to determine the saturation of hydrocarbons using a profile or record of the resistivity. In addition, there are no known methods for estimating pore size or permeability in a continuous record or profile, ie without taking fluid samples. Nuclear Magnetic Resonance ("NMR") well logging tools capable of determining the liquid content of the pore volume within a deposit, and a method for using these tools are described in, for example, U.S. Pat. Nos. 5,309,098; 5,291,137; 5,280,243; 5,212,447; 4,717,878; 4,717,877; 4,717,876; and 4,710,713. In particular, U.S. Pat. No. 5,291,137 discloses a pulse sequence Carr-Purcell-Meiboom-Gill ("CPMG") and the echo response and a method to obtain the porosity of the free fluid, the porosity by total NMR, the porosity with respect to the bound fluid, the relaxation time (ratio of viscosity to the elastic modulus) spin-spin (which is related to the pore size distribution in the sandstone), and continuous profiles or permeability records. The recovery times between the CPMG pulse trains are typically between 0.5 and 1.5 seconds. Because the relaxation time (ratio of viscosity to elastic modulus) of methane under typical deposit conditions is greater than three according to two, the total NMR porosity measured in this method may not include the volume occupied by the hydrocarbon gas. Recently, a new diagnostic tool, the MRIL (TM NUMAR Corp. of Marvern, Penn) has been introduced to determine the filled porosity of the liquid in an independent way with respect to the lithology, ie the response of the tool does not require a correction of lithology to determine porosity. MRIL uses pulsed nuclear magnetic resonance of mobile protons in the pore space. However, according to the manufacturer of the tool, it is not capable of measuring the content of the hydrocarbon gas in the pore space. For example, Chandler et al. of NUMAR in SPE 28635"Improved Log Quality with a Dual-Frequency Pulsed NMR Tool" states that "Gas causes an incorrigible reduction of MRIL porosities". This "gas effect" is a feature of all NMR logging tools. "This" gas effect "is a major disadvantage because the main part or only the hydrocarbons in many tanks is natural gas. The pore in these deposits consists only of brine and natural gas.The inability to measure the gas means that the NMR logging tool will not measure gas-filled porosity and should be dependent on other logging tools, such as profiles. or records of density and neutrons, to measure porosity. It is therefore an object of the present invention to provide a method for determining at least one of the pore size of the formation of the earth and the pore volume occupied by the hydrocarbon gas in the formation, using a profile or record of nuclear magnetic resonance where the properties of the formation, different from the pressure and temperature of the pore, do not have to be known. In another aspect of the present invention, it is an object to determine the content of the hydrocarbon gas of a formation wherein the content of the gas can be determined without knowledge of the rock properties of the formation. It is a further object of the present invention to provide a method for determining the pore size of the formation that is not affected by the clay content of the formation. These and other objects are effected by a method to determine a characteristic of a porous earth formation containing hydrocarbon gas, the formation of the earth is penetrated by a sounding, the method comprises the steps of: obtaining a profile by NMR using a recovery time that is equal to or greater than the longitudinal relaxation time of the hydrocarbon gas; determine, from the NMR profile or graph, a distribution of the transversal relaxation times including the transverse relaxation times, attributable to the hydrocarbon gas; and estimate at least one of the pore size of the formation of the earth and the volume of the pore that is occupied by the hydrocarbon gas. In a preferred embodiment, the volume that is occupied by the hydrocarbon gas is estimated by additionally obtaining a second NMR profile or graph using a recovery time that is significantly shorter than the longitudinal relaxation time of the hydrocarbon gas. The amplitudes of the total signal from the first and second graphs or profiles by NMR are subtracted and the rest divided by the hydrogen index of the gas under the conditions of the deposit to determine the volume of pore occupied by the hydrocarbon gas. carbides. Alternatively, the volume that is occupied by the hydrocarbon gas can be estimated from the first and second NMR photographs by subtracting the distributions of the transverse relaxation times of the first or NMR profile from that of the second by NMR and dividing the rest between an effective hydrogen index of the gas at the deposit conditions to determine the volume occupied by the hydrocarbon gas. In the practice of this embodiment of the present invention, if the oil is also present, the difference between the two NMR profiles or profiles can be integrated over the T2 ranges within which the gas responses are expected to differentiate between gas and oil. Using magnetic field gradients either constant or pulsed, the diffusion coefficient of the gas can be measured and the amount of its restriction from the volumetric diffusion through the confining pores can be used to estimate the pore size and permeability. The saturation of the hydrocarbon gas, or the hydrocarbon gas content of the pore volume, can be accurately measured using pulsed NMR scanning tools, using an NMR pulse sequence that includes a recovery time that exceeds the time of longitudinal relaxation of the gas. Depending on the content of the other fluid in the formation, a second NMR scan using a pulse sequence that includes a recovery time that is equal to or less than the longitudinal relaxation time of the gas, a density chart and / or a chart - 1 - gamma rays may also be required to determine the content of the formation fluid. A significant aspect of the present invention is that the gas within a formation is always non-humectant. Therefore, the longitudinal relaxation time, TI, of the gas, will always be that of the volumetric gas, not shortened or reduced by surface relaxation, as are the humectant liquids such as brine. The TI of the volumetric hydrocarbon gas depends only on its temperature and on the pressure, which are known exactly for most deposits. In addition, the TI of the hydrocarbon gas is generally longer than the TI of the other reservoir fluids, that is, the brine and the crude oil. The NMR signal from the hydrocarbon gas can be measured even in hermetic, very schistose formations, for which the water signal relaxes too quickly to be measured. The great diffusivity of natural gas leads to the fact that the diffusion coefficient measured is restricted in the pores of most rocks for short or short values of the inter-echo time. In shale rocks where the aqueous phase has a brief TI and a short transverse relaxation time, T2, diffusion measurements for the aqueous phase are not possible because the inter-echo time may have to be too long compared to IT and T2. Nevertheless, if natural gas is used as the diffusional tracer, the restricted diffusion coefficient can be measured even in very schistose or slate rocks. The restricted diffusion coefficient is an indication of the pore size. For the MRIL version C diagnostic tool (field gradient of 17 gauss / cm and an inter-echo time of 1.2 ms), in rocks that have a porosity of approximately 30 PU and greater than the pore diameters of 1000 micras, there will be no restriction on the diffusion of methane. In the pore diameters of 20 microns a substantial restriction will be observed, and at a pore size of 0.1 microns or less, the diffusivity will be totally restricted to (D / tortuosity), where Do is the volumetric diffusion coefficient. The invention will be described hereinafter in greater detail and by way of example with reference to the appended drawings, in which Figure 1 is a graph of TI of methane as a function of pressure for different temperatures. Figure 2 shows a typical CPMG echo sequence (with the alternation of the 90 ° pulse phase) used for the measurement of the transverse relaxation time, T2. Figure 3 is a graph of the hydrogen index for natural gas as a function of pressure for different temperatures. Figure 4 is a graph of T2 for natural gas for a T2 for unrestricted diffusion, as measured by a MRIL C logging tool, as a function of pressure for different temperatures. Figure 5 is a graph of the T2 log or profile for natural gas for a methane T2 for relaxation restricted in diffusion for different porosities as a function of the inverse of the pore diameter as measured by a tool of MRIL C log. Figure 6A shows a T2 histogram for a sandstone containing brine and natural gas, where the pores are 1000 microns in diameter and the methane diffusion coefficient is not restricted from volumetric diffusion. Figure 6B shows a histogram of T2 for a sandstone containing brine and natural gas, where the pores are 100 microns in diameter and the diffusion coefficient of methane is slightly restricted from volumetric diffusion. Figure 6C shows a T2 histogram for a sandstone containing brine and natural gas, where the pores are 20 microns in diameter and the methane diffusion coefficient is significantly restricted from volumetric diffusion. Figure 6D shows a T2 histogram for a sandstone containing brine and natural gas, where the pores are less than 0.1 microns in diameter so that the methane diffusion coefficient is restricted with respect to l / (tortuosity). Figure 7 is a well log obtained by the method of the present invention. Figures 7A to 7C, 8A to 8C, and 9A to 9C show well logs that include NMR scans according to the method of the present invention.
DETAILED DESCRIPTION OF THE INVENTIONNuclear magnetic resonance imaging tools can measure four properties: Mo, TI, T2, and D, where Mo is the nuclear magnetization of equilibrium, TI is the longitudinal relaxation time, T2 is the transverse relaxation time, and D It is the diffusion coefficient. The first three properties do not require a magnetic field gradient, while the measurement of D requires a magnetic field gradient.
The gradient can be a gradient either permanent or pulsed. NMR imaging is normally restricted to the measurement of hydrogen (H) because of its strong signal and large gyromagnetic ratio. The NMR data of the present invention is altered from that typically used in the petroleum industry by the extension of the recovery time between the pulse sequences up to or beyond the longitudinal relaxation time of the hydrocarbon gas within the formation which is graphed. A CPGM echo sequence such as that described in U.S. Pat. No. 5,291,137 can be used, in the company of an NMR-based diagnostic tool such as MRIL C available from NUMAR Corp. of Malvern, Penn. The MRIL tool is capable of detecting properties of a portion of a formation that is as much as 10.16 cm or 12.7 cm (four or five inches) from the borehole wall. This is preferred because rocks within 5.08 cm or 7.62 cm (two or three inches) from the wall of the borehole can be cleaned with a discharge of drilling fluids and may not be representative of the formation in general. The longitudinal relaxation times, TI, of gases such as methane, are only a function of temperature and pressure, and not of other properties of the formation. TI for methane is described in, for example, C. J. Gerritsma, et al., "Proton Spin Latice Relaxation and Self Diffusion in Methanes-Paper 2", Physica, v. 5, 392 (1971). IT is considered to be proportional to density and varies with absolute temperature according to:ln (Tl) - A - B [__1_ 1 (1) Twhere: A and B are constant and T is the absolute temperature. Natural gas is composed predominantly of methane and light alkanes. Typically, up to 75% by volume of natural gas is methane. The properties of the hydrocarbon gas within a formation can therefore be estimated with sufficient accuracy for the practice of the present invention assuming a hypothetical hydrocarbon composition, such as a hydrocarbon composition C 1, .1 -, H 4y 2. . Referring now to Figure 1, an IT graph for a stream of natural gas having a composition of C1, ... 1.H4, .2"is shown as a function of the pressure for different temperatures. Equation 1 can be used for extrapolation at other temperatures. Lines a to f represent TI, in seconds, for temperatures from 37.77 ° C (100 ° F) to 176.66 ° C (350 ° F) in increments of fifty degrees, respectively. As an example of typical tank conditions, methane can have a density of approximately 0.2 g / cc and at a temperature of approximately 93.33 ° C (200 ° F), leading to an IT of approximately four seconds. A recovery time of six seconds will generally exceed TI, and leads to an NMR imagery that is useful in the practice of the present invention. Accordingly, in order not to completely saturate the methane signal, the recovery time (TD) in the CPMG sequence must be greater than four seconds, and preferably between approximately six and twelve seconds, which is two to three times the CT of the gas. The TI of the natural gas is between approximately 3 and 6 seconds for the typical conditions of the deposit. Referring now to Figure 2, an exemplary CPMG pulse sequence with alternation of the 90 ° pulse phase is shown. This impulse sequence is used to measure the transverse relaxation time, T2. The sequence consists of an RF pulse of 90 degrees (at the Larmor frequency), followed in time by tcp p * or a train of 180-degree pulses equally spaced. The pulse spacing of 180 degrees is 2tcp. For example, in MRIL C this time can be as short as 1.2 ms. A spin echo, e, was obtained between each of the 180 degree pulses. The sequence is repeated at time TR later with an RF pulse of 90 ° from the opposite phase (relative to the preceding 90 ° RF pulse). The subsequent train of negative echoes, e, is subtracted from the previous train, thus accumulating coherent signals and canceling the instrument's artifacts. When there are multiple fluids in the pore space, and a range of pore sizes, the NMR signal, A (t), represents a sum of exponential decays:n t A (t -S aie T2. (2) i-1where a1. they are constant and T2i. they are constant constants of relaxation times, and n is an integer where n T2x.s are selected at equal logarithmic intervals. Typically, thirty-five to fifty intervals, n, lead to an adjustment for the echo data. The time domain data can be inverted using a multi-exponential inversion program to give a histogram, or a graph of a., As a function of T2. This invention is described in, for example, U.S. Pat. No. 5,291,137. The relaxation times of the components that are bound to the solids are in general significantly shorter than those of the components that are not bound to the solids. The surfaces in the formations are usually either water or oil but they are not wet gas. Therefore, the effects of surface relaxation are negligible for gases. A multi-exponential inversion of an echo train from an NMR response such as that shown in Figure 2 can therefore be expressed as a relaxation time distribution T2. The ordinate could be the amplitude of the signal associated with each time constant of T2i. adjusting ai. for the given T2i.s. In Figure 2, the sequence is repeated after a recovery time, TR. If TR is greater than three times TI, then almost complete relaxation will occur. If TR is not significantly greater than TI, a correction factor, OC, is applied to account for partial saturation. This correction factor is given by:1 - . 1 - e IT a = (3) Mowhere M is the nuclear magnetization of equilibrium or as measured by the NMR image. However, if TR is significantly smaller than TI, then the magnetization will be completely saturated and a signal will not be detected from the gas. Because TI is significantly shorter for pore water and most crude oils, hydrocarbon gas can be differentiated from water and crude oils by making two NMR diag- noses, one with a shorter recovery time. that the gas relaxation time, and one with a recovery time equal to or greater than the gas relaxation time. The gas can be identified as the difference between the two graphs. This difference between the two graphs representing the gas contribution to the NMR responses, can then be used to determine the pore size within the formation. This is possible because the mechanism of relaxation is predominant for hydrogen in a gas phase as opposed to a liquid phase. The NMR reports respond to the hydrogen content. In the interpretation of NMR radiographs, a hydrogen index ("HI") is used to convert the NMR results to a base in volume. HI is the density of hydrogen atoms relative to liquid water at standard conditions.
At the tank conditions, HI is approximately one for water and hydrocarbon liquids. HI for hydrocarbon gases are already known and are available from, for example, Schlumberger Log Interpretation Principles / Applications, available from Schlumberger Educational Services, Houston, Texas, (1987), in particular, Figures 5-17, p. 45. Where the density and HI of natural gas slightly heavier than methane pressure will be between approximately 140.7 and 703.7 kg / cm * (2000 and 10,000 psi) and the temperature will be between 37.77 ° C (100 ° F) ) to 176.66 ° C (350 ° F), leading to gas densities of between about 0.1 and about 0.3 g / cc and a HI of between about 0.2 and about 0.6. Although this HI is less than one, it is still large enough and makes the hydrocarbon gas measurable with the NMR pulse sequence of the present invention. Referring now to Figure 3, a graph of HI for a natural gas having a composition of C. H, ~ as a function of the pressure for different temperatures. Lines g to l represent HI at temperatures from 37.77 ° C (100 ° F) to 176.66 ° C (350 ° F) in increments of fifty degrees respectively. The "effective HI" is referred to here as the OC and HI product.
The NMR therefore measures the density of the hydrogen atoms in the fluid phase in the portion of the formation that is examined. The density of hydrogen atoms can be converted to a fraction by volume, 0 NMR, by dividing it by HI. The HI of both the water and the hydrocarbon liquids is almost unity, and thus the HI is only applied to the gas phase in the following descriptions .. If the free liquid water and the free liquid hydrocarbons are not present in the formation, A single NMR scan can be obtained using a recovery time that is equal to or longer than the longitudinal relaxation time of the gas. The signals corresponding to transver-sial relaxation times greater than a gas-water cut of, for example, 32 ms, are attributed to the gas while the signals corresponding to lesser relaxation times than this water-gas cut are attributed to water. When the free liquid water is present, but not the free liquid hydrocarbons, the volume of the present gas can still be found with an NMR image but a density graph and a gamma ray image are required (for the determination of the clay content). Of the information). For this method, we start with an equation for the total density, or volumetric density, (R, of the formation.) If the formation contains clay, water, and gas but not liquid hydrocarbons, it can be expressed as the sum of the density of the times of the components in the fraction in volume of the components, as follows:PB * (l-Vc? -fw-fg) Pma + vclPcl + FwPw + FgPg < 4 >where Vci- is the volume fraction of the clay,^ w is the volume fraction of water, ^ is the volume fraction of the gas, ma is the density of the matrix of the rock,, is the density of the clay, »w is the density of the water not bound with the clay , and K. is the density of the gas. Because the NMR graph responds to fluids, the volume fraction measured by the total NMR is:Frmn = w + < * lFgHIg > < 5 >Combining equations 4 and 5 to eliminate> w and solving for leads to the following equation:The densities of the individual components can be estimated with relatively good accuracy. The volumetric density can be determined from a graph, such as a gamma-gamma ray image, and the volume fraction of the clay can be determined from a gamma-ray image. Therefore, equation 6 can be used to estimate the volume fraction of the gas with an NMR graph, the graph is run or it is carried out with a recovery time greater than the longitudinal relaxation time of the gas present, a graph of density such as a gamma-gamma log, and a gamma-ray log. When two NMRs are run or carried out to determine the amount of gas in the formation, the volume of gas is determined directly from the two graphs without any other information being needed. For each chart, an iMrn, and an OC are obtained; ? 'rmnl. and O 1, for the logging using a recovery time greater than the longitudinal relaxation time of the gas, and? ^ ^ 2 For - ^ to the graph using a recovery time less than the longitudinal relaxation time of the gas. The two equations for the fluid volume measured by each NMR are as follows:(7) F rmn2"F +? T2 (FgH: [g) (8)The volume ratio of the gas can then be found from the resolution of equations 7 and 8 to obtain:* rmnl "F rmn2 Fg - (9) (a1-a2) HIgBoth the total amplitudes and the distributions of T2 resulting from the two CPMG sequences can be subtracted from each other, giving the amplitude of the signal only of the gas and the distribution of T2 respectively. When the two CPMG sequences are subtracted, a region of interest can be either double-graded, or different pulse sequences can be used in the two overrides tested by a tool such as the MRIL C diagnostic tool. MRIL C allows for separate CPMG sequences in two annulations spaced far 0.228 cm (0.09 inches) apart. One of the CPMG sequences can use aT long compared to the methIT, the other one T r r that is shorter than the methIT. For example, a T can be from six to twelve seconds while the other is 1.5 seconds, when the TI of the gas in the formation in the conditions of the formation is 4 seconds. It is preferred to use the method of the two cancellations in place of the area's chart twice because the depth shifts introduced have to re-map the same area. When the formation contains water, light oil, and gas, the volume of the pore occupied by the gas, and the light oil, can be determined, but the NMR tool that uses a pulsed gradient must be used and two NMR dia- grams. must be provided, A NMR image is provided with a relaxation time equal to or greater than the longitudinal relaxation time of the gas and an NMR image is provided with a relaxation time less than the longitudinal relaxation time of the gas. The light, gas-like oils can have a relatively long TI, and therefore do not have T2 distributions that cancel when the T2 distributions are subtracted. The NMR response attributable to the gas and light oil are identified on a graph of the difference between the two graphs by the location on a graph of the responses against T2 of the two responses. To know which response is due to gas and light oil, the relaxation mechanisms for the present components must be understood.
The relaxation mechanisms that affect Tl and T2 in rocks are (1) molecular motion in fluids, (2) surface relaxation capacity in the pore wall, and (3) molecular diffusion in field gradients. magnetic. The first mechanism, due to local movements such as molecular drumming, is called volumetric relaxation. The transverse relaxation times are equal to the longitudinal relaxation times when the predominant relaxation mechanism is volumetric relaxation. Volumetric relaxation is the mecca. predominant relaxation mechanism for liquid water and hydrocarbon liquids. TI, which is what is expected from a hydrocarbon liquid, TI, in more may be expected from a hydrocarbon liquid, TI, in more than can be estimated as a function of the viscosity,,, in cp, of the liquid of hydrocarbons according to:1200 TI, (10), 0.9The second mechanism of relaxation is surface relaxation in the pore wall, or the ratio of H nuclei when they closely approximate paramagnetic ions such as iron and mange which are located on the surfaces of the grains. This is the dominant mechanism for fluid molecules such as water that moistens rock surfaces. However, because the gas is always non-wetting and never comes close to the surface of the rock, this mechanism is negligible for the gas. This relaxation is usually very fast. The third mechanism of relaxation is the diffusion of the molecules in the gradients of the magnetic field. This relaxation mechanism affects only T2 and not TI. Therefore, when diffusion is a prominent relaxation mechanism, T2 will differ from IT. Diffusion is a predominant relaxation mechanism only for gas. Unrestricted diffusion will be the lower limit of the T2 interval within which the NMR responses of the gas are to be expected. The upper limit is fixed or placed by restricted diffusion. Using a CPMG sequence, the relaxation time due to diffusion (T2D), when the diffusion is not restricted by the pore size, is:T2D (1 1)? 2G2D0 2 pwhere TF is the gyromagnetic relation of H (26, 741 radians / s-gauss), D is the unrestricted diffusion coefficient, G is the gradient of the field created by the NMR tool, and t p is half the spacing of the CPMG pulse. Do of supercritical methane as a function of temperature and density can be found, for example, in Gerritsma et al., Supra, and also in Dawson et al., AlchE Journal, Vol. 16, No. 5, 1970. Under typical deposit conditions, methane will have a density of approximately 0.2 g / cc, and a diffusion coefficient of approximately 50 times that of water, or approximately 109 x 10 cm2 / s compared to 2 x 10 -5 cm 2 / s for water. Therefore, using equation (11), if an unrestricted volumetric diffusion occurred, with the MRIL C logging tool (G = 17 gauss / cm, cp = 0.6 ms), T2D would be equal to 37.1 ms. Referring now to Figure 4, T2D for diffusion that is not restricted by pore diameters, is plotted as a function of pressure for temperatures from 37"77 ° C (100 ° F) to 176.66 ° C (350 ° F) in increments of fifty degrees like the sea lines respectively, using the parameters of the MRIL C tool. A ratio of the restricted diffusion coefficient, D ', on the unrestricted diffusion coefficient, D, approaches a limit of the inverse of tortuosity when fluids diffuse through many pores. Tortuosity is defined as the product of the resistivity factor of the formation, F, and the porosity of the formation, 0. The resistivity factor of the formation can be determined, for example, using an induction log, and the porosity it can be determined from a density or neutron log. The upper limit of the interval within which the NMR responses of the gas are to be expected, is therefore the relaxation time due to diffusion with restricted diffusion, T2D ', which isT2D '(12)? .2rG.2¿D '£ pReferring now to Figure 5, the T2D 'log is shown for methane as a function of the inverse of the pore diameter for the porosities of 10, 20 and 30% as the lines s, t, and u respectively. Given the T2 interval within which the maximum gas point is expected to fall, and the expected location of the hydrocarbon liquid peak based on the viscosity of the expected hydrocarbon liquids within the formation, the peaks or maximum the difference between the T2 distributions from the two NMRs can be separated into a peak or peak of the gas, and a peak or peak of the hydrocarbon liquid. The pore volume occupied by each of the gas and liquid hydrocarbon is then calculated by integrating the area under the peaks or maximum points and dividing the integrated area of the peak or peak of the gas between the HI of the gas to the formation conditions. Taking into account the correction factor, OC, from equation 3, based on the TI for the gas for the NMR log that has a TR that is longer than the longitudinal relaxation time of the gas, OC i ' t which is shorter than the longitudinal relaxation time of the gas, O ^ C TI), the volume of the pore occupied by the gas,? can be found as: where? P is the difference between the T2 distributions from the two NMRs, expressed as a function of T2. T2D 'r J? P (T2) dT2 T2D F (13) g HI [a (tL, Tlg) - a (ts, Tlg))The volume of the pore occupied by the liquid hydrocarbon,? , it can be found as the integrated area under the difference between the two NMRs on an interval around the T2 of T2 + where 5 is selected based on the width of the peak or maximum point of the hydrocarbon liquid, by:f J? P (T2) dT2 A T2 ° "? (14) (a (tL, Tl0) - a (ts, Tl0)] where 0 (? * Ti, T1o) is OC according to the equation3 for the NMR log that has a TR that is longer than the longitudinal relaxation time of the gas and the TI of the hydrocarbon liquid, and C ("b, TIo) is based on the NMR log with a TD that is plus. brief than the longitudinal relaxation time of the gas and TI of the hydrocarbon liquid. The MRIL C tool has a depth of 40.64 cm (16 inches). In a 20.32 cm (eight inch) hole, the formation can be examined up to 10.16 cm (four inches) from the borehole wall. The petroleum-based muds have a low invasion at this depth and are therefore the preferred drilling mud for the practice of this invention. With low-invasion oil-based muds, the gas saturation observed by the MRIL C logging tool will be uneven or unbalanced. In particular, the ESCAID 110 oil-based drilling mud with 80% ESCAID 110 and 20% water saturated with CaCl2 has been found to provide a very small invasion and is therefore a preferred system. D / Do can be expressed as a function of1/2 (D t), where D is the restricted diffusion coefficient, and T is the inter-echo time in a fixed gradii ite experiment or the pulse time of the intergradient in an NMR of the pulsed field gradient. For short times where only a small restriction occurs, it has been shown that:where S / V is the ratio of the surface to the volume of the pores. For extended echo times, D / D approaches 1 / tortuosity when fluids diffuse through many pores. Tortuosity is defined as the product of the resistivity factor of the formation, F, and the porosity of the formation,? . The resistivity factor of the formation can be determined, for example, by using an induction log, and the porosity can be determined from a neutron log. A Pade approximation can be used for the adjustment between short and long echo time behaviors. An acceptable Pade approach is:where is the porosity of the formation, F is the resistivity factor of the formation, and? is the adjustment parameter that can be approximated by the following relationship:-2? DoslOO (17) VIf the gas saturation (S) is not equal to one, 1 / FQ) it is replaced by S / Ffl). From equations 6 and 7, an S / V can be found for a given set of fluid properties and a D measured from an NMR log with a known echo time. The pore diameter, d, for spherical pores, is a function of S / V as follows:= 6 ß -1 (18)This relationship is generally accepted as an approximation for the pore size within a formation, and can be used in the practice of the present invention to estimate the pore size based on an S / V determined from the equation 7. Figure 6A shows an exemplary T2 histogram for a 30 PU sandstone containing only brine and natural gas, where the pores are 1000 in diameter and the methane diffusion coefficient is not restricted from volumetric diffusion. The MRIL C parameters are assumed. The sandstone is at an irreducible water saturation, which for a sandstone means that the T2 of the irreducible water is less than 30 ms. Therefore, all water is bound by surface interactions and will not be produced. The T2D for methane leads to the peak or maximum value of methane in 37 ms. Assuming that the hydrogen index (HI) of the methane at the deposit conditions is 0.35, the area under the 37 ms peak is divided by 0.35, leading to a total porosity of 30 PU and a gas saturation of 70%. The location of the methane peak at 37 ms shows that the pores are 10001X and larger. Figure 6B shows an exemplary T2 histogram for a 30 PU sandstone containing only brine and natural gas, where the pores are 100 μL microns in diameter and the methane diffusion coefficient is slightly restricted from volumetric diffusion. The T2D for methane is 42.8 ms, showing a slight restriction of volumetric diffusion. Figure 6C shows an exemplary T2 histogram for a 30 PU sandstone containing only brine and natural gas, where the pores are 20C in diameter and the methane diffusion coefficient is significantly restricted from volumetric diffusion. The T2D for methane is 60 ms. Figure 6D shows an exemplary T2 histogram for a sandstone containing only brine and natural gas, where the pores are less than 0.1X. The T2D for methane is 122 ms. For this rock, D / Do has reached its prolonged value of (1 / tortuosity). Once the pore size is estimated from the methane T2D, the equations well known in the art can be used to estimate the permeability of the pore size. Some of these equations are described in U.S. Pat. No. 4,719,423, the description of which is incorporated herein for reference. In the most general cases, there may be three phases that coexist in the spaces of the pores of the rock, especially brine, crude oil, and gas. Another complexity could be if the brine phase is not in an irreducible saturation. If brine, crude oil and natural gas are present as two separate phases in the pore space, and some or all of them have a T2 relaxation superimposed, two logs, each with a different echo sequence, are obtained . A chart uses an echo sequence in which the recovery time is significantly shorter than the individual relaxation time, TI, of the gas and a second with a recovery time equal to or longer than the individual gas relaxation time. . The amplitudes of the two graphs are subtracted and the rest represents the distribution of T2 attributable to the gas. This approach is based on the fact that methane IT is significantly longer. that one for pickles and crude oils. Any of the total amplitudes of the two CPMG sequences or the T2 distributions obtained from the multiexponential inversions obtained from the data in both of the CPMG sequences can be subtracted from each other., with the difference, that is the total amplitude due to the gas or the distribution of T2 only of the gas, respectively. When subtracting two CMPG sequences, one area of interest may be either two or more times diagraphed, or different pulse sequences may be used in the two annulments or angles tested by a tool such as the MRIL C-imaging tool. The MRIL C logging tool allows for separate CPMG sequences at two angles spaced apart at 0.228 cm (0.09 inches). One of the CPMG sequences can use a prolonged RT compared to the methane IT, the other a TR that is shorter than the methane IT. For example, a TR can be from six to twelve seconds while the other is 1.5 seconds, when the gas TI in the formation, to the conditions of the formation, is 4 seconds. It is preferred to use the method of the two cancellations or angles instead of the radiography of the area twice because of the depth displacements introduced to have to x-ray the same area. The MRIL C tool has a research depth of 40.64 cm (16 inches). In a 20.32 cm (eight inch) diameter hole, the formation can be examined up to 10.16 cm (four inches) from the borehole wall. The petroleum-based muds have a low invasion at this depth and are therefore the preferred drilling mud for the practice of this invention. With low-invasion oil-based muds, the gas saturation observed by the MRIL C logging tool will be unbalanced or leveled. In particular, ESCAID 110 oil-based sludge with 80% ESCAID 110 and 20% water saturated with CaCl ", provides a very small invasion and is therefore a preferred system.
Eg emplosA well in the Gulf of Mexico was recorded twice to demonstrate the preferred mode of the present invention. A MRIL C logging tool and a CPMG echo sequence with the alternation of the phase was used. Recovery times were six seconds and three seconds. The estimated relaxation time of any hydrocarbon gas that is expected to be in the formation was four seconds. It is known that the formation contains gas, but not liquid hydrocarbons. The investments of the CPMG data were made as described above to obtain distributions of the relaxation time. These distributions were integrated from a relaxation time of 32 ms to 1024 ms, which represents the hydrocarbon gas. Conventional charts were also obtained. Referring now to Figure 7, a neutron log (denoting porosity), 103, and a density log (gamma-gamma log), 102, are plotted as a function of depth. Conventionally, a "first convergence" neutron-density plot, shown as area 112, is considered to indicate the presence of a gas. The resistivity of the induction is shown as lines 115 and 116. The GR is shown as line 117. The results of the two NMRs are shown as lines 118 and 119, with the difference between them being the area 114 Another well in the Gulf of Mexico was plotted using a MRIL C logging tool and a CPMG echo sequence with the alternation of the phase to demonstrate an NMR modality of the present invention. The recovery time was six seconds, which exceeded the estimated four-second relaxation time of any hydrocarbon gas that is expected to be in the formation. It is known that the formation contains gas, but not liquid hydrocarbons. The investments of the CPMG data were made as described above to obtain distributions of the relaxation time. These distributions were integrated from a relaxation time of zero to 2 ms representing water bound to clay, between 2 ms and 32 ms, representing water not bound to clay, and greater than 32 ms, representing gas of hydrocarbons. Also conventional graphs were obtained. Referring now to Figure 7A, a neutron (indicative of porosity), 103, a sonic, 101, and a density (gamma-gamma), 102, graphs are plotted as a function of the foundry Conventionally, a neutron-density "first convergence" log is considered to indicate the presence of a gas. In this segment of the diagram, the lines touch on a point, but they do not cross each other. The sonic log is also an indicator of the gas in this formation. When the reading of the sonic record exceeds the level indicated by line 111, as shown by the shaded area 104, the presence of a gas is indicated. Referring now to Figure 7B, the NMR results are plotted as a function of depth, with depths that are aligned with Figure 7A. The bound water, as indicated by the signals representing the relaxation times less than about 32 ms, is plotted as the areas 105, which is the water attached to the clay, and 106, which is the water attached capillary. The gas is indicated by the area 107 from the signals representing the relaxation times from 32 ms to 1024 ms. The filtering of petroleum-based drilling mud is evident in the material that has relaxation times greater than 1024 ms. As can be seen by comparing Figure 7A with Figure 7B, the gas is present where conventional means, such as a first density-neutron convergence, do not indicate that a gas is present. The explanation for this is that the first convergence of density-neutrons has been suppressed by schistosity. The present invention can be better understood with reference to Figure 7C, in which the NMR responses representing the intervals of relaxation times are plotted as a function of depth. The depths again align with the depths of Figures 7A and 7B. The NMR responses representing the relaxation times less than 2 ms are plotted on the lower part, the responses representing relaxation times of between 2 and 4 ms are plotted on the line labeled 2 ms, with the intervals increasing. geometrically. The responses indicating relaxation times greater than 1024 ms indicate the invasion of oil-based drilling mud in the area that is examined by NMR. Figures 8A to 8C, and 9A to 9C, correspond to Figures 7A to 7C respectively, which show the graphs for different segments of the hole. Figures 8A to 8C show particular plots of a formation segment containing a significant amount of gas that correlates with a first neutron-density convergence as the area 112 for the first convergence of the neutrons-density and 107 as the indicated gas by the NMR. Additionally, Figure 8C shows a significant oil invasion as responses greater than 1024 ms for a portion of the log. Figures 9A to 9C show in particular a response of a formation containing some gas, 112 and 107, and a considerable amount of unbound water, 113. These examples demonstrate the utility of the present invention for determining the presence of the hydrocarbons in a formation using an NMR image.
It is noted that in relation to this date the best method known by the applicant to carry out the aforementioned invention, is that which is clear from the present description of the invention.
Having described the invention as above, property is claimed as contained in the following