BR9910426B1 - process for optimizing gravity gradiometer measurements. - Google Patents

Description

"PROCESS TO OPTIMIZE GRAVITY GRADIOMETER MEASURES"

Foundation Technique

The present invention relates to a process for optimizing gravity gradiometer measurements in the context of secondary oil recovery and more particularly, a process for optimizing gravity gradiometer measurements for secondary oil recovery. where the subsurface boundary or interface between the oil to be recovered and the reservoir drive fluid is detected and controlled to optimize recovery and even more particularly with a process for optimizing the measurements of the gravity gradiometer in which anomalies within the gravitational field caused by density changes and contrasts as a result of the movement through the time of the subsurface boundary between the oil to be recovered and the expulsion or countervailing fluid is monitored by a gravity gradiometer.

Natural gas oil and hydrocarbon reservoirs are formed as a consequence of the transformation of organic matter into various types of hydrocarbon materials, including coal, tar, oils and natural gas. Oil and gas reservoirs are believed to form as lighter hydrocarbon molecules filter toward the earth's surface until they are trapped in a relatively permeable layer beneath a relatively impermeable layer that " cover "the permeable layer. Lighter hydrocarbon molecules continue to accumulate, often accompanied by water molecules, within relatively large subsurface reservoirs. Since reservoirs exist at various depths within the earth, they are often under substantial geostatic pressure.

Hydrocarbon resources have been extracted from surface and subsurface deposits by mining solid resources (coal and tar) and by pumping or otherwise removing natural gas and liquid oil from subsurface deposits. naturally occurring.

In the last century, natural gas and oil have been extracted by drilling a borehole in subsurface reservoirs. In general, most reservoirs were naturally pressurized by the presence of free natural gas that accumulated above the liquid oil layer and often by water that accumulated below the liquid oil layer. Since naturally occurring crude oil has a lower density than water (ie in the range of 0.7 for "light" crude to 0.9 for "heavy" crude), Crude oil accumulates above the water permeated layer and below the gas permeated layer. Therefore, a probe bore terminating within the oil permeated layer would produce oil that obtains its expulsion energy from an overlying gas permeable layer and / or an underlying water permeable layer. In general, "primary" crude oil recovery occurs during this time when the natural pressurization of a reservoir causes crude oil to be driven up through the wellbore. At some point in the operating life of the reservoir, naturally occurring pressurization is effectively exhausted. Several different methods, commonly known as secondary recovery methods, have been developed to extract crude oil after natural pressurization is depleted. In general, secondary recovery involves repressurizing the reservoir with a fluid (i.e. a liquid or gas) to decrease oil viscosity and / or direct the remaining crude oil in the oil-permeable layer to the surface through one or more wells. Drive fluid is introduced into the reservoir by injection wells that pump pressurized drive fluid into the reservoir to displace and thereby direct the oil towards and into production wells.

Several schemes have been developed for injection well placement. For example, a line of injection wells may be placed adjacent to or near a known reservoir boundary to direct crude oil toward and into production wells. As the boundary between the pressurization fluid moves past the production wells, these production wells may be capped or, if desired, converted to injection wells. In another arrangement, the injection wells are interspersed between the production wells to direct the oil in the oil permeable layer away from the direct injection point and immediately adjacent to the production wells.

Various fluids, including water at various temperatures, steam, carbon dioxide and nitrogen, have been used to repressurize the reservoir and displace the desired crude oil from its rock or sand matrix to the wells. production. In the flooding technique, room temperature water is injected into a reservoir to direct the oil towards and into production wells. Injected water accumulates under the crude oil and as a consequence, the lighter density crude oil floats upwards and into the production well mine bore. In such cases where the oil-permeable layer is relatively thin from a geological perspective, and is also confined between two relatively less permeable layers (ie a waterproof reservoir roof and a more permeable reservoir subsoil), the water is injected at relatively high pressure and volume to effect a "shore steering" whereby crude oil is pushed toward the oil production wells. Injected water is sometimes heated to help lower the viscosity of the oil and thereby help to displace the crude oil from the pores of the permeable sand or rock. The flooding technique is also suitable for directing the natural gas captured within the relatively low permeability rock pores to a production well. In the steam flooding technique, steam is used to displace and direct oil from the sand or oil-containing rock towards and into production wells. Steam, which may initially be overheated, is injected into the oil-permeable layer to cause reservoir repressurization. As the steam moves away from its initial injection point, its temperature drops and the quality of the steam decreases with steam eventually condensing into a layer of hot water.

Additionally, some lighter hydrocarbon may be distilled out of the crude oil as it supports displacement at the interface between steam / hot water and crude oil. Steam injection may be continuous or on an intermittent start-and-stop basis.

In addition to the use of water and steam to repressurize the reservoir and expel crude oil into production wells, carbon dioxide and nitrogen have also been used for the same purpose.

A problem associated with water, steam, or gas expulsion techniques is the identification of the limit or interface between the expulsion fluid and crude oil. Optimally, the boundary between the expulsion fluid and the crude oil to be displaced moves in a manner that can be predicted through the reservoir from the injection points to the production wells to maximize the production of the oil. crude oil. The geology of a reservoir is often complex and inhomogeneous and often contains regions or zones of relatively higher permeability sand or rock; These higher permeability zones can function as low impedance paths to. the pressurized expulsion fluid. Pressurized expulsion fluid sometimes forms low impedance channels, known as "theft" zones, through which pressurized fluid "passes" into a production well, thereby significantly decreasing recovery efficiency.

The ability to identify the position and frequently undefined interface or boundary between the crude oil to be displaced and the pressurized expulsion fluid, to track the velocity and morphology of this boundary, and to control it would substantially enhance secondary recovery. of the oil.

Several techniques have been developed to gain an understanding of the geology configuration of the subsurface of an oil-containing reservoir. The dominant technique involves seismic echo in which a pressure wave is directed downward to the subsurface layers. The initial energy of the interrogation wave is typically created by the detonation of explosives or specialized earth impact machines. The interrogation wave radiates from its source point with its transmission speed affected by the elastic modulus and the density of the material through which it passes. As with all wave energies, the interrogation wave is subject to the reflection, refraction, scattering, and muffling effects caused by the material through which it passes and from which it is reflected. The reflected wave energy is detected by geophones spaced from the seismic source point and subjected to processing to produce a reservoir model. This technique is highly developed and well suited for detecting subsurface structures that may be favorable for oil or gas accumulation.

Other techniques for investigating subsurface geology include the use of gravimeters to detect minor changes in the magnitude of the gravity vector for the purpose of detecting subsurface structures that may be favorable for oil or gas accumulation. .

The various devices and techniques used to analyze subsurface layers have led to significant advances in the ability to create a three-dimensional reservoir model or simulation. However, existing perception technologies are not able to detect the location and morphology of the boundary or interface between the pressurized expulsion fluid and the oil or natural gas in these reservoirs supporting secondary recovery. Information on the position, morphology and speed of the limit would be of substantial value when optimizing the recovery of hydrocarbons being recovered, especially in the efficient use of expulsion fluids.

Disclosure of the Invention In view of the above, it is an object of the present invention, among others, to provide a process for improving gravity gradiometer measurements in the context of the recovery of fluid hydrocarbons such as oil and natural gas from a oil and / or gas reservoir in which the reservoir is supporting repressurization.

It is another object of the present invention to provide a process for optimizing gravity gradiometer measurements in the context of secondary hydrocarbon recovery in which a pressurized fluid is used to direct reservoir oil and / or natural gas to a well. of production.

It is still another object of the present invention to provide a process for secondary oil recovery in which the boundary or interface between the oil to be recovered and a pressurized fluid driving the oil to be recovered can be identified.

It is a further object of the present invention to provide a process for improving the accuracy of gravity gradiometer measurements in the context of secondary hydrocarbon recovery in which the boundary or interface between the hydrocarbon to be recovered and a pressurized fluid directing The hydrocarbon can be identified and subsequently controlled to maximize recovery.

In view of these and other objects, the present invention provides a process for improving or optimizing the measurement of gravity gradients in the context of secondary oil or gas recovery in which a reservoir is pressurized with liquid or exhaust gas. and the boundary or interface between the expulsion fluid and the hydrocarbon material to be displaced is monitored over time to perceive changes in density across the boundary with a gravity gradiometer. Perceived changes in position, extent, velocity, and limit morphology, including the formation of incipient theft zones, allow injection fluid control to optimize recovery efficiency.

Gravity gradient measurements taken by an accelerometer-type gradiometer are optimized by tilting the instrument's measurement plane by a selected angle above and below the horizontal to obtain data that can be differentiated or otherwise processed to remove the bias from the instrument and by collecting data in a first and then a second orthogonal azimuth direction to obtain unrelated absolute gradient measurements.

A hydrocarbon reservoir undergoing secondary recovery is subjected to an initial gravity gradient survey in which a gravity gradiometer measures the gradient on the surface above the reservoir to define an initial data set. At some point in the future, a second gravity gradient survey is done to provide a second data set. Differences between the first and second data sets produce information about changes in sub-surface density associated with gas or oil displacement and their replacement by the expulsion fluid. Subsequent gravity gradient lifts similarly shifted in time during expulsion fluid injection produce additional information regarding interface position, morphology, and speed allowing an oil field manager to control the number. injection sites and the temperature, pressure and volume of the injected fluid to thereby optimize recovery efficiency. The manager can also determine the desirability of drilling new wells, their locations and their segmentation, and the desirable orientations of each segment.

In the preferred embodiment of the invention, a plurality of gravity gradient measurement stations are established on the surface above an oil or gas reservoir undergoing secondary recovery. A rotary accelerometer-type gravity gradient measuring instrument is positioned at each station in series and gravity gradient indicative data at each station is taken in accordance with the present invention to provide a first data set. This first data set provides the data constituting baseline information about the reservoir gravity gradient affected by surface and subsurface density variations, including density contrast affecting gravity at the interface between the expulsion fluid and the oil or gas being displaced during the time measurements are being made. At some time subsequent to the measurements that produced the first data set, that is, a time period measured in months or years, the measurements are repeated to produce a second data set. The common data between the first and second data sets are indicative of substantially non-variable fixed data, representative of the effect on the gravity gradient of surface and subsurface geology while non-common data between the first and second data sets. These data are indicative of a time-dependent change in the gravity gradient as a result of time-driven movement of the interface between the expulsion fluid and displaced oil or gas and the possible effects of geological noise.

After mitigating the effects of geological noise, information about interface movement or boundary is used by an oil field manager to control the number of injection points including volume, pressures, and temperatures to control and improve hydrocarbon recovery. .

Advantageously, the present invention provides a process for the secondary recovery of oil which permits observation via measurement of gravity gradients associated with the limit between the expulsion fluid and the hydrocarbon material to be recovered from oil. such that recovery efficiency can be optimized.

Other objects and the further scope of the applicability of the present invention will become apparent from the following detailed description, taken in conjunction with the accompanying drawings, in which, parts are designated by like reference characters. .

Brief Description of the Drawings

The present invention is described below, by way of example, with reference to the accompanying drawings, where:

FIG. 1 is a representative cross-sectional view of an oil production retainer or reservoir undergoing secondary recovery showing the interface between the expulsion fluid and the oil to be displaced in an ideal manner;

FIG. IA is an idealized graphic representation of the interface between the expulsion fluid and the oil to be displaced;

FIG. IB is an idealized representation of the density contrast at the interface of FIG. IA;

FIG. 2 is a top view of the reservoir of FIG. 1 showing an oil limit as a dotted band in the context of a direct line drive configuration;

FIG. 3 is the view of FIG. 2 showing the displaced oil limit of the position shown in FIG. 2;

FIG. 4 is an example of a five point recovery configuration;

FIG. 5 illustrates the manner in which a uniform gravity field is disturbed by a mass;

FIG. 5A is a top view of the field of FIG. 5; FIG. 6 is an isometric view of a preferred gravity gradiometer with selected portions thereof broken for clarity;

FIG. 7 is a functional block diagram showing the manner in which the accelerometer output of the gravity gradiometer of FIG. 6 is processed;

FIG. 8 shows the mathematical derivation for a device-specific protocol for canceling instrument trends and aut gradients;

FIG. 9 presents the mathematical derivation for a device-specific protocol for eliminating both the horizontal gradients of the gradient curvature estimates and the horizontal gradients, both combined with the cancellation instrument trend and the self-gradients;

FIGS. IOA and IOB are flowcharts of a test protocol for obtaining multiple data sets; and

FIG. 11 is a graphical representation of gradient values in an interface.

Best Mode for Carrying Out the Invention

An idealized and exemplary geological formation having oil-containing layers is shown in FIG. 1. As shown, an oil permeable layer 10 is bordered at the top by a relatively impermeable roof layer 12 (known as a "seal") and at the bottom by a relatively permeable layer 14. The layer permeated with Oil 10 is typically a fine grain or coarse grain sand that is impregnated with crude oil that typically accumulates between the particles. In a typical formation, the layers may form a low or anticline dome under which oil accumulates; The oil is often accompanied by natural gas or water.

In such reservoirs that include natural gas, oil and water, natural gas tends to form a layer or region above the oil while water tends to form a layer or region below the oil. Depending on the geostatic pressure in the oil-permeable layer, some of the gas may come into solution with the oil. In general, interfacial boundaries are typically indistinct, although in some cases the boundaries may be geologically distinct.

As shown on the left in FIG. 1, the layers are shifted in a vertical direction along a fault line 16 so that the displaced rock layer 14 creates a plug 18 that defines a lateral boundary of the oil permeated layer 10. In a manner Similarly, the backing of a salt dome 20, often found in association with the oil layers, defines another side boundary of the oil-permeable layer 10. In general, the laterally limited oil-permeable layer 10 confined to one another. Ceiling rock below is defined as a reservoir and can occur at depths of several tens to several thousand feet below the earth's surface. The representation of FIG. 1 illustrates a reservoir at a depth of several hundred feet and is simply an example of the wide variety of geological configurations in which oil reservoirs can be found. In FIG. 1, four cranes are presented, each with a mine bore that penetrates through several layers within the oil permeated layer 10. When a naturally pressurized oil reservoir is initially penetrated by a mine bore, the oil is directed through the mine hole to the surface. However, in time, the pressure in the reservoir decreases to the point where mass transport to the surface ceases or falls to an unacceptable flow rate. At this point, oil flow can be induced by using surface pumps to extract oil or by repressurizing the reservoir by injecting water, steam or a gas (ie carbon dioxide or nitrogen) into the reservoir. through injection wells. In the example of FIG. 1, the two left wells are injection wells injecting a repressurization fluid into the oil containing layer 10 while the two right wells are production wells through which crude oil is removed. The all-black part of the oil-permeable layer on the right represents the available crude oil while the dotted part of the layer on the left represents this part of the layer 10 in which the oil has been displaced and replaced by the injected extrusion fluid.

In FIG. 1, the transition from the dotted area to the black area represents the interfacial boundary or "boundary" between the pressurized expulsion fluid advancing from the left and the crude oil being shifted thereby to the right to the production wells. FIG. IA is an idealized representation of the transition between the expulsion fluid and the oil to be displaced and illustrated a physical phenomenon that is not fully understood and may include variables or aspects not shown. Assuming that the oil permeated layer is fully saturated with oil, and in the case where the expulsion fluid is steam, the vapor, which may initially be overheated, over a reduction in temperature as it moves away from the bridge. injection and produces its latent heat. At some point, the quality of the steam decreases (that is, the water content increases) as the heat of vaporization is transferred to the surrounding oil. At this point, steam and / or heated condensed water may be forcibly mixed with the oil and fractional vapor-induced distillation may occur, during which some of the light hydrocarbons in the oil to be displaced are vaporized. to mix with the steam. At some point in the process, the steam quality drops to zero or close to zero (ie steam condenses in hot water). In FIG. 1A, the gradually denser dotting from left to right represents the displacement of oil from its pores. In general, about 90% of the oil is displaced from any arbitrarily defined volume unit with 10% of the oil remaining as waste oil; The remaining volume is typically replaced by 30% steam and 60% water. In FIG. IA, the vapor is presented as over cutting the oil as the vapor tends to dissipate within the oil permeated layer. The boundary is often indistinct and its morphology may change as a result of mixed water in the lower parts of the oil-permeable layer and natural gas in the upper part of the oil-permeable layer.

While the example of FIG. 1 shows the secondary recovery of crude oil, the same configuration exists for the secondary recovery of natural gas trapped within the pores of the gas permeated layer. In the case where the reservoir of FIG. 1 is a gas holding device, the injected fluid (typically water) would displace the captured natural gas towards and into the production wells.

FIG. 2 is a plan view of the field of FIG. 1 showing a rectangular set of sixteen wells located in the center of the defined blocks identified with row and column numbers. The marking 22 at the bottom of FIG. 2 represents the location of the fault line of sub-surface 16 of FIG. The curvilinear marking 24 at the top of FIG. 2 represents the periphery contour of the salt dome 20 of FIG. 1. In FIG. 2, the wells in blocks 11, 12, 13 and 14 are injection wells injecting an expulsion fluid into the oil permeable layer 10 while the remaining wells are conventional oil outlet wells. The dotted zone passing portionwise through blocks 21 and 11 and entirely through blocks 12, 13 and 14 from left to right in FIG. 2 represents the position of the sub-surface edge between the oil and the pressurized expulsion fluid. As can be appreciated, the boundary or "boundary" is indistinct and not uniform, reflecting variations in the permeability of the oil permeable layer 10. In general, the boundary will move over time across the field away from the wells. injection into production wells.

The configuration of FIG. 2 is known as a direct line drive configuration provided that a line from the injection wells in blocks 11, 12, 13 and 14 pumps the expulsion fluid into the oil permeated layer. Alternative configurations include the placement of injection wells between production wells as shown, for example, by the 5-point configuration of FIG. 4. In the 5-point configuration, an injection well is placed in the center of a series of four production wells. The centrally located injection well injects the expulsion fluid to create a subsurface boundary that extends around the injection well to dredge, displace or direct the oil towards and into production wells. Variations in the 5-point configuration include 7-point and 9-point configurations (not shown) in which the centrally located injection well is respectively surrounded by seven or nine production wells.

FIG. 3 is a representation of the field of FIG. 2 time shifted to show the migration of the boundary between the expulsion fluid and the oil to be displaced over a period of time. In general, the time difference between FIG. 2 and FIG. 3 can be measured in weeks, months or years, depending on the size of the reservoir involved and the injection and production rates. As shown, the boundary moved further away from the injection wells with different parts of the boundary moving at different speeds to substantially alter the boundary morphology. More specifically, the left boundary moved past the wells in blocks 21 and 31 while the right boundary moved only past the wells in block 24. Additionally, a portion of the boundary between the wells in blocks 22 and 23 moved in and toward the well in block 33. The particular morphology shown between the wells in blocks 22 and 23 within block 33 is highly suggestive of a theft zone, that is, a volume of material of Relatively higher permeability that allows the pressurized expulsion fluid to form a channel therethrough that can drill into a production well and substantially reduce recovery efficiency.

Information about the position, extent, morphology and speed of the boundary over time would be valuable when managing injection wells to optimize the dredged area by the moving boundary and thus optimize recovery. Information relating to changes in oil layer time is useful in predicting downward trends, identifying the location and extent of remaining resources, and providing information on the location of new injection and production wells. optimally increase production.

In the context of FIG. IA, the density of an arbitrary volume unit is a function of liquid and gas fluid density, liquid and gas fluid saturations, rock matrix density, and rock porosity. The change in density Δρ for a reservoir volume unit Δν may be represented as:

<formula> formula see original document page 21 </formula>

where pf is the liquid density, Af is the change in liquid saturation, pg is the gas density and Δ9 is the change in gas saturation.

In general, the rock has a typical density of about 1.9 to 3.0 g / cm3 and the oil has a density of between 0.7 g / cm3 for light oil and 0.9 g / cm3. cm3 for heavier oil; consequently, the oil permeable material can be viewed as having a compound density. As the expulsion fluid shifts and replaces the captured oil, the compound density will change as a consequence of the oil displacement and pore occupancy by the waste oil and the expulsion material. In the case of an arbitrary volume unit at point "A" in FIG. IA, density is a function of the rock matrix, residual oil (about 10%) and the most recent steam introduced (about 30%) and water (about 60%). Therefore, the density difference between the volume units at point "A" and point "B" in FIG. IA represents a density contrast that will affect the local gravity gradient so that changes in this gradient can be verified on the surface. In general, the density contrast will be less than a few tenths of a gram / cm3; It is believed that the lateral dimension of the transition zone through which the adjustment of the. Density occurs is in the range of hundreds and thousands of centimeters (between tens and hundreds of feet).

The earth's gravity field ranges from a low of about 97.8 m / s2 to 9.83 m / s2 (978 gals in the equator to about 983 gals) at poles with gradient gradients in Eotvos units, where an Eotvos unit equals 10-9sec-2. For an idealized homogeneous sphere, an equipotential surface outside the sphere is also spherical. However, density homogeneities in the sphere give rise to a non-spherical equipotential surface; For such a surface, the curvature of any point is different in different directions. The two directions along which the curvature is maximum and minimum are termed as main directions; The difference in curvature along these two directions is called differential curvature, as explained more fully below. In the earth context, local variations in gravity are caused by deviations on the earth's surface from a geometric sphere, surface geology, water tides, atmospheric trends, and the change in the relative position of the earth, from the moon of of the sun. For any relatively small unit of volume in free space, an idealized field of gravity can be viewed as a set of unidirectional field lines aligned along the local vertical and having zero magnitude in x, y directions. Should a mass be placed within this volume unit, the gravity field will be disturbed. For example and as shown in FIG. 5, a cylindrically shaped dense mass M having hemispheroidal ends and a finite length along a major axis located in the center of an arbitrary volume unit will disturb the local gravity field within this volume unit causing these field lines closest to mass M bend toward mass M and the near field lines removed from the nearest field lines bend toward mass M somewhat less.

For any observation point within an arbitrary volume unit, the field of gravity at this observation point can be resolved into x, y, componentes components of which the vector ζ will have the greatest magnitude and the vectors x, y will have the respective magnitudes that are a function of the location of the observation point in relation to the mass that is disturbing M. For the observation points in a plane above or below the mass M in FIG. 5, the vector information at this observation point will provide the declination or slope information as the field disturbance. In the context of the example of FIG. 5A, a sufficient number of observations in a plane above the mass M (or a plane below the mass M) will create a dataset from which one or more isopotential surfaces can be obtained, which graphically defines the perturbation in the field. of gravity caused by the introduction of the M-mass. In general, the gravity field along the z-axis can be measured by single-axis gravimeters of which a common type uses lasers and a high-precision clock to time a mass falling between two points vertically spaced in an evacuated space. Gradiometers, unlike gravimeters, measure curvature gradients (or differential curvature or ellipity of equipotential gravity surfaces), horizontal gradients (or the rate of change of gravity increase in the horizontal direction), or vertical gradients (or the rate of gravity increase in the vertical direction). Various processes and instruments have been developed to measure the gravity gradient. These processes and instruments include lateral bending of a suspended mass from a rope (Bouguer's method) and torsional flexion exerted on a horizontally suspended beam carrying unequal masses at each end (Cavendish and Eotvos beams). Modern gravity gradiometers use force restoration accelerometers to measure the x, y side components of the gravity gradient. In general and in its simplest form, an accelerometer uses a suspended mass at the end of a beam that can sag. Any deviation in the position of the mass from a null position caused by an acceleration experienced by the mass is detected and the mass restored to its null position by a magnetic field applied by a restoration coil. The current flow in the restoration coil is proportional to the acceleration experienced by the mass. Some modern gravity gradiometers use several pairs of accelerometers that move at a constant velocity along an orbital path around an axis. The information from each accelerometer at any angular position in the orbit provides the information about lateral acceleration experienced by acceleration. In the context of FIG. 5A, an accelerometer moving at a constant angular velocity around an orbital path in the plane of observation above mass M will experience positive and negative accelerations in the x directions, and emit a sinusoidal waveform that is modulated with the severity anomaly information in this observation plan. Where the observation plane is normal to the local vertical, the accelerometer output does not contain a representative component of the z axis. Conversely and as explained below in the context of the preferred test protocol, if the accelerometer is in an observation plane that is inclined to the field lines, the accelerometer output will also be modulated with the z-axis information. .

A gravity gradiometer suitable for the present invention includes a gravity gradient instrument (GGI) sold by Lockheed Martin Corporation (Buffalo, NY, USA); Lockheed Martin's GGI, whose basic structure is shown in FIG. 6, is preferred in the context of the present invention. The structure and basic operation of the Lockheed Martin GGI is described in United States Patent No. 5,357,802, issued October 25, 1994, to Hofmeyer and Affleck entitled "Rotating Accelerometers Gradiometer", the disclosure of which is incorporated. hereby by reference.

As shown in FIGS. 6 and 7, the GGI includes eight common radius accelerometers 100 mounted around the periphery of a rotor assembly 102 which is rotated at a constant angular velocity and controlled around an axis of rotation Ax. The rotor assembly 102 includes the rotor 104 carried on a support shaft 106 for rotation therewith. Rotor assembly 102 is rotatably mounted on spherical bearings 108 and in turn carried in a flexible vibration isolation mounting assembly 110. Electronic processing components 112 are mounted on rotor 104 adjacent each accelerometer 100 to process their output signal as explained below in the context of FIG. 7. An internal housing 114 contains rotor assembly 102 and is designed to rotate with rotor assembly 102. An external housing 116 contains interior components and includes one or more heaters 118 designed to operate the instrument at any controlled temperature above ambient. (i.e. about 46,10 ° C (115 ° F)) and also includes a magnetic field shield 120. A slip ring assembly 122 at the upper end of the mounting shaft 106 provides the electrical / signal interface with the rotor assembly 102 and the active devices therein. A shaft encoder 124 at the lower end of the mounting shaft. 106 cooperates with a pick off of encoder 126 to provide rotary position information. The pick off output of encoder 126 is provided for a computer and a speed controller, which in turn controls a drive motor 128 at the upper end of the unit to provide a controlled rotational speed.

Each accelerometer 100 provides a sinusoidally variant analog output that is a function of the acceleration experienced by each accelerometer as the accelerometer orbits the rotary axis. For a gradiometer having its spin axis aligned along the field lines in an ideally uniform and undisturbed field of gravity, each accelerometer experiences the same acceleration forces as it continues along its orbital path. However, where the local gravity field is disturbed by the presence of one or more masses and / or the rotary axis is inclined relative to the local vertical field lines, each accelerometer will experience different accelerations throughout its orbit. The quantitative output of each accelerometer coupled with its rotary position provides information related to local gravity gradients.

At any point of view, the gravity gradient is a second-order derivative of the gravity potential scale adjuster Γ and is represented by a symmetrical nine-component second-order tensor Γi as follows:

<formula> formula see original document page 27 </formula> <formula> formula see original document page 28 </formula>

The components ΓΧ; Ζ and Γγ, ζ are approximately equal to the variation of the force of gravity along the χ and y directions, respectively, and are known as the horizontal gradient and Γ2 components (Ζ is known as the gradient gravity curvature Differential curvature is related to ΓΧ; Χ, ry, y and ΓΧ (Υ, as follows:

<formula> formula see original document page 28 </formula>

where F is the force of gravity.

In addition to the differential curvature of Eq. 2, a curvature vector whose magnitude is equal to the differential curvature is also defined by the value λ as follows:

<formula> formula see original document page 28 </formula>

where λ is the angle of the differential curvature vector relative to the x axis.

As is known, the diagonal elements are scalar in- variant and conform to the Laplacian relationship:

<formula> formula see original document page 28 </formula>

from which it follows: <formula> formula see original document page 29 </formula>

Additionally, three pairs of nine elements are symmetrically equal, that is, ΓΧ / z = Γz, x, Ty, z = rz, y and finally ΓΧ (γ = ry, x so that the tensor is characterized by five components independent.

The gradients Ty, y - Γx, x and 2Γx, y in Eq. 1 are two curvature gradient components, while Γχ, z and Ty, z are the two horizontal gradient components; ,Z, z is the vertical gradient component.

The output of the accelerometers is processed according to the block diagram of FIG. 7; processing can be performed by separate solid state functional devices, software, or firmware-controlled microprocessors or computers, an application-specific integrated circuit (ASIC), or any combination thereof.

As shown, the preprocessed outputs of the eight accelerometers 100 of the GGI gravity gradient instrument of FIG. 6 are divided into two groups Ά 'and λΒ' of four with each group subdivided into two subgroups, namely (A1, A2), (A3, A4), (B1, B2) and (B3, B4).

Accelerometer outputs A1, A2 are shown for summation device inputs SUM (A1 + A2) and outputs A3, A4 are also shown for summation device SUM (A3 + A4). The summed outputs of the SUM (A1 + A2) and SUM (A3 + A4) devices are displayed for a SUB-A subtractor. Similarly, the accelerometer outputs Bl, B2 are shown for summation device inputs SUM (BI + B2) and outputs B3, B4 are also shown for summation device SUM (B3 + B4). The summed outputs of the SUM (BI + B2) and SUM (B3 + B4) devices are displayed for a SUB-B subtractor. The sum of the diametrically opposed accelerometer signals 100 effectively cancels the acceleration component due to any displacement of the rotor assembly in the XY plane. The difference operation in the SUB-A and SUB-B subtraction circuits removes the effect of any angular acceleration of the rotor assembly that may occur in response to angular velocity correction signals sent to the motor by its motor. speed controller.

A set of four demodulators demodulate the outputs of the SUB-A and SUB-B subtraction devices in response to phase and quadrature reference signals at twice the rotor mounting rotational frequency and by a reference signal source. . The reference signal source may include a phase locked repeat oscillator locked with the rotor assembly revolutions 102 in response to the pickoff output of encoder 126. The sine of the phase reference signal 2ΩΤ is applied to DEMOD-SA and DEMOD-SB demodulators connected respectively to the output terminals of the SUB-A and SUB-B subtraction circuits. Similarly, the quadrature reference signal cosine 2ΩΤ connected to the DEMOD-CA and DEMOD-CB demodulators, also connected respectively to the output terminals of the subtraction circuits SUB-A and SUB-B. The output signals of the four DEMOD-SA, DEMOD-SB, DEMOD-CA, and DEMOD-CB demodulators take the form of square values of the sine and cosine signals. More specifically, DEMOD-SA outputs a value 2r (Γχχ-Γγγ) seno22QT and the associated 4ΩΤ terms, the DEMOD-CA demodulator outputs a 4Rrxyco-seno22Qt value and the terms 4Qt, the DEMOD-SB demodulator outputs a value minus 4Rrxyseno22Qt and the terms 4Qt and finally, the DEMOD-CB demodulator outputs a 2R (Γχχ-Γγγ) value cosine22Qt and the terms 4Qt. The term 2R is the distance between sets of opposite accelerometers, that is, the distance between the pair of accelerometers Al and A2.

The sum circuit SUM-A accepts the output signal from the DEMOD-SA and DEMOD-CB demodulators and provides the summed output through an LP-A low band filter. Similarly, the SUM-B sum circuit accepts the output signal from the DEMOD-SB and DEMOD-CA demodulators and provides the summed output through an LP-B lowband filter.

The output signals from the DEMOD-SA, DEMOD-SB, DEMOD-CA, and DEMOD-CB demodulators include the high frequency components in terms of frequencies equal to four times the rotational frequency of the rotor assembly. Gradient data, independent of a scale factor, is provided by both square sinusoid and square co-sinusoid components of demodulated signals. These components sum together in the sum circuits SUM-A and SUM-B to provide a DC (direct current) value of the gradient data, along with the high frequency components. The low-band filters LP-A and LP-B function, respectively, to filter the output signals of the sum circuits SUM-A and SUM-B and attenuate the high frequency components to provide the desired DC components. of the signals representing the gradient data. Gradient data is output from the LP-A lowband filter as an expression that includes the term 2R (Txx-Tyy) and from the LP-B lowband filter as an expression of 4R (Txy) .

The division of the 2R (Txx-Tyy) and 4R (Xxy) outputs by 2R produces two results, (Γχχ-Γγγ) and 2 (Γχγ), which define the curvature vector; the magnitude of the curvature vector is known as the "differential curvature" or the "horizontal direction trend" and is the square root of the sum of (Tx x-Tyy) with (2Txy)

The direction λ of the curvature vector with respect to the X axis is represented by:

λ = -1/2 tan "1 (2Txy / (Tyy-Txx))

The above description with respect to FIG. 7 shows the function of GGI when the accelerometer sensitive axes are in the normal to vertical plane. In this orientation, the instrument is optimized to understand the x, y components of the gravity gradient. However, a series of errors may be introduced into the signal output as a consequence of the structure of the instrument itself. In the GGI design, mass is required to be distributed in a uniform and symmetrical manner around the axis of rotation Ax. Because GGI uses separate devices, some asymmetries of the mass come out around the rotary axis. In addition, the GGI is not symmetrical with respect to the mass above and below the rotor plane 104. While the mass asymmetry is small, the asymmetry is physically close to the accelerometers and is believed to have a influence as an error component. In accordance with the present invention and as explained below with respect to the test protocol of FIG. 10, the GGI is operated in at least two orthogonally spaced directions (i.e. 90 degrees). Since the instrument's specific bias and instrument-specific gradient errors are "fixed" to the instrument itself and its connected structures, the rotation of the GGI around its rotary axis in both orthogonal directions will not change instrument-specific trend values and instrument-specific gradient errors at the same time, concurrently, the fixed ground gradient signals are reversed. By calculating the mean differences between the sets of measurements in two orthogonal directions, the absolute values of the earth curvature gradients can be obtained since the instrument-specific trend and instrument-specific gradients will cancel out. . As explained in the mathematical presentation of FIG. 8, the GGI primary output can be characterized as an "inline" I / L component and the secondary output can be characterized as a "transverse" CR component. In the context of the functional block diagram of FIG. 7, the inline and transverse components may be designated as shown in the EQ. 1 and EQ. 2 in FIG. 8. The GGI χ axis can be aligned, for example, with the north and using the Euler Angle Sequence, the curvature gradients measured by the x, y plane of the instrument can be expressed as a function of the curvature gradients. land in the north (n) and east (e) context as represented by EQ. 3 and EQ. 4 of FIG. 8. GGI in-line exits in a north azimuth direction (0 degrees) is depicted in FIG. 8 by I / L (H = 0 °) and for orthogonal measurement by I / L (H-90 °). Similarly, the GGI cross-sectional outputs in a north and east direction are represented by CR (H = Og) and for orthogonal measurement by CR (H = 90 °) so that the gradient-free values of the gradient of curvature can be obtained from EQ 5 to EQ. 9 in FIG. 8

As explained below in relation to FIGS. IOA and 10B, the GGI is operated during the test protocol with its 'inclined' rotary axis at a small positive angle and a small negative angle to the horizontal plane to generate a horizontal component of gravity to excite the repetitions of feedback compensation control on the GGI and allow calibration of accelerometer scale factors. However, the inclined rotary axis couples the horizontal gravity gradients of the earth within the GGI curvature measurements. Obtaining measurements with the GGI inclined at a small angle above the horizontal and a small angle below the horizontal, in connection with measurements made in two azimuth directions that are 90 degrees apart, will produce measurements that bend the gradients of desired curvature and cancel out unwanted horizontal gradients.

The GGI includes orthogonal tilt and degree axes, of which the cylinder axis is held horizontal in the local x, y plane while the instrument is rotated around its cylinder axis to position the instrument up and down. down relative to the local horizontal plane. The term I / L indicates a line that is in the x, y plane of the GGI that can be aligned in an azimuth direction while the x, y plane of the instrument is placed (angled) at some small angle to the local x, y plane. . The secondary output of the gradiometer instrument is the term "transverse" CR as explained above in relation to EQ. 1 and 2 of FIG. 8. In the context of the mathematical presentation of FIG. 9, the I / L and CR gradients measured by the GGI in its x, y plane can be expressed as a function of the north, east geodetic, and south plane (n, e, d) reference tensor gradient components . The direction cosine (DCM) matrix that transforms a north, east, and south geodetic plane to the instrument plane x, y, z at a direction angle H and a slope angle P represented in EQ . 1 of FIG. 9. The outputs of GGI I / L and CR are related to geodetic gradients as depicted in the EQ. 2 and 3 of FIG. 9, where the terms "error (I / L)" and "error (CR)" include instrument self-bias and all instrument-specific gradient errors. As presented in EQ. 4 and EQ. 5, the sum of the up and down measurements are indicated where the capital letter aU "indicates a slope for" up "and a capital letter" D "indicates a" downward slope "with EQ. 6 and to EQ.7 showing the relation to the geodetic gradients By applying the measurements at H = 0 degrees and H = 90 degrees and calculating the sum of the result of the inline and transverse values, Eq. 8 and Eq. 9 to estimate Tee-Tnn and 2Tne curvature gradients without coupling of the horizontal gradients Tnd and Ted and the vertical gradient Tdd with these trend-free values and instrument-specific auto-gradient errors. and EQ.11 represent the difference between the I / L and "up" and "down" CR measurements while EQ. 12 and the EQ. 13 provide the estimate of the horizontal gradients of the measurements taken at 180 ° of separation.

According to the test protocol of the present invention and as shown in FIGS. IOA and 10B, several measurement stations η are established in the oil field. Each metering station may take the form of an area of unimpeded earth or more possibly an asphalt or cement platform. It is important that the location of the station be fixed for the period of the first and subsequent tests. The measurement stations may preferably take the form of a rectangular series of observation positions, a polar series of observation positions or a mixture thereof including observation stations that do not follow a pre-established pattern and may be considered. as randomly placed within the field under observation. For example and in the context of FIGS. 2 and 3, the measurement positions may correspond to the position of the row and column block numbers shown.

The GGI is preferably mounted on a small wheeled trailer that can be moved from station to station as the test continues. Depending on the distance between observation stations, the GGI can be transported, for example, by a wheeled or different vehicle. If desired, the GGI can be mounted on a wheeled vehicle and driven from station to station and measurements taken from the vehicle. However, the mobile trailer is preferred as the use of the trailer eliminates gravity anomalies introduced by the wheeled vehicle from the dataset. In each measuring station, the trailer is lifted by lifting control jacks.

As shown in the flow chart of FIG. 10A, the system is initialized and the variables m and η set to 1; η being representative of the number of pre-established measuring stations where the maximum number of measuring stations is N (maximum) and m representing the number of data sets to be formed over time where the maximum number is M (maximum).

The GGI is positioned at the first measurement station and prior to the start of data collection, all objects having sufficient mass to affect measurements (transport vehicles, power sources, local computer control, personnel , etc.) are moved far enough away from the instrument to minimize any adverse effects on the instrument from these objects. In general, each preferably metering station is not located near man-made fixed stationary structures to minimize large signals that are unrelated to the boundary interface locations.

Thereafter, the GGI is initialized and the pivot axis is tilted to some preselected tilt angle (i.e. about 0.9 degrees in the case of the present invention) sufficient to provide an ζ axis response. on the accelerometers and the instrument aligned in a zero degree azimuth direction. As explained above in relation to FIG. 9, the "inclination" of the rotor assembly at any angle sufficient for the eight accelerometers to perceive the gravity vector allows the instrument to observe the output of each accelerometer for this gravity value, to identify the scale factor of each accelerometer. - meter, the differences in scaling factors between accelerometers and identifying an adjustment value so that all accelerometers provide identical apparent output. As shown in the flow chart of FIG. 10A, the instrument is also subsequently subjected to an inclination and an angle opposite the former so that the transverse couplings of the horizontal gradients can be eliminated. Therefore, measurements taken at the first inclination angle include the bias of the instrument and similarly, measurements taken at the opposite inclination angle also contain the bias errors of the instrument; Measurements for both inclination angles can then be weighted to cancel the trend error.

Then GGI takes the data from the observation station for a sufficient period of time to ensure that the sources of error are minimized; In the case of the preferred GGI, data collection with an interval of several minutes (ie about five minutes) in each direction is considered appropriate.

Thereafter, the instrument is tilted to its opposite angle (i.e. minus about 0.9 degrees in the case of the present invention) and the data collection step repeated at the initial azimuth. Once data collection is complete in the initial azimuth direction for both upward and downward inclined attitudes, the instrument is rotated azimuth by ninety degrees incremental and the upward and downward inclined data collection steps repeated . The upward and downward inclined data collection only needs to be conducted in two azimuth directions; however, upward and downward inclined data collection may be repeated in more azimuth directions, if desired, to increase the accuracy of data collection. Rotation of the instrument in the new direction includes the instrument itself, its cart and any associated structures, including any environmental enclosures. Rotation of structures associated with the operation of the instrument itself helps minimize sources of error. As explained above in connection with FIG. 8, the azimuth rotating functions to provide the information necessary to remove the sources of error associated with the mass dissymmetry of the instrument itself and its immediate surroundings. This sequence continues until data is collected in at least three orthogonal azimuth directions; however, in the case of the preferred mode, the data is taken at 0, 90, 180, 270 and then again at 0 degree. If additional information is desired, the data collection steps can be repeated in several directions.

As explained above in more detail, data collection in different azimuth directions and at different inclination angles is designed to minimize error sources and effectively increase instrument sensitivity.

After that, η is incremented by one and the GGI moved to the next successive measurement station and the sequence at this repeated station.

When data are taken from each of the η measurement stations (ie n = N (maximum), the first data set is complete. A time period (measured in weeks, months or years) is during which the oil field withstands continuous or non-continuous pressurization by the injected expulsion fluid to cause the migration of the interface between the expulsion fluid and the hydrocarbons to be recovered. , the test sequence of Fig. 10A is repeated to produce another dataset characterized as the second dataset.As may be appreciated, the third, fourth and subsequent tests may be undertaken after the interim periods of time. -test tests have elapsed to produce the third, fourth and subsequent test sets. In practice, two successive data sets (ie M (maximum) = 2) are sufficient to provide usable data. made.

Each test set includes information on the field's gravity gradient, including the effect of subsurface geology, consequent terrain variations, and man-made static stationary structures including cranes, pipelines, pumps, engines, etc., which typically occupy an oil field. In addition, this data set will include information on the effect on the gravity gradient of the interface between the expulsion fluid and the displacement supporting hydrocarbons. However, there is no way other than conjecture that the location of the interface can be precisely determined from a single set of data. As with the first dataset, the second and subsequent datasets also include information about the effect on the gravity gradient of geology, terrain and man-made structures, and the interface in their new location. Therefore, information about geology, terrain and structures made by man will represent relatively non-variable common data or common signal information between each of the data sets at the same time as information about the transient interface in time will not be common for the various data sets.

Sources of error that may adversely affect accuracy may include geological motion such as oil sump compaction and water table movement.

In order to process the first and second (and / or subsequent data sets) and as an initial step, a theoretical model of the gravity gradient relationship of the layers below the hydrocarbon reservoir of interest is developed. For any hydrocarbon reservoir supporting secondary recovery, the likelihood is that a body of geophysical data, including a reservoir model, will be available from previous acoustic surveys, mine bore data records, core samples, analysis of the test wells and a knowledge of the presence or absence (and changes thereof) of the expulsion fluid at the output of the production wells. If desired, known geophysical data may be combined with available gravity gradient data. More specifically, gradient data may be "better adjusted" with geodetic data to provide an early reservoir model and / or gradient data may be similarly "better" adjusted with geophysical data. using inversion techniques to obtain the best reservoir model and the model boundary between the expulsion fluid and the hydrocarbon to be recovered.

In order to obtain subsurface density information, early modeling methods (also known as indirect methods) or direct methods can be used. Early modeling methods begin with existing knowledge about the reservoir layers in which the injection is being performed and assume changes in the saturation levels of different fluids, incorporating the knowledge obtained from other geophysical measurements. (eg seismic) and other oil field observations such as pressure and temperature changes in the observation wells. From this initial model, the calculations of the time gradient can be made. For this purpose, Talwani's early modeling method, "Computation with help of a digital computer of magnetic anomalies caused by bodies of arbitrary shape" in Geophysics, Vol. 30, No. 25, p. 797 to 817 (1965) may be used. The calculated values are compared with the observed values of the time gradient and the initial model repeatedly modified so that the calculated values fit the observed values. Examination of the saturation values in the final model yields the boundary position between the expulsion fluid and the hydrocarbons supporting the displacement.

With direct methods, the existing knowledge about the reservoir is used as constraints and the inverted observed data to produce changes in the subsurface density, which as for the anticipated modeling methods are related to changes in fluid saturation. .

Since elapsed time gradient data (ie 4D) is obtained, numerous computer techniques are available to determine the change in the distribution of sub-surface density (layer morphology) across of the time period. Direct methods using gradient data to estimate density distributions fall within the linear problem category and direct methods that determine boundary perturbations for constant density bodies fall within the nonlinear problem category as described by DW Vasco in "Resolution and Variance Operators of Gravity and Gravity Gradiometry" published in Geophysics, vol. 54, no. 7 (July 1989), p. 889 to 899. Vasco's solution linearises the relationships between the prisms (mass density or density changes) and their associated gravity gradients and solves the inversion repeatedly using generalized inverses. Direct methods also include those described by S. K. Reamer and J. F. Ferguson in "Regularized two-dimensional Fourier gravity inversion method with application to the Silent Canyon caldera, Nevada" published in Geophysics, vol. 54, no. 4 (April 1989), p. 486 to 496 and the wavelet-Galerkin techniques described in United States Provisional Application 60 / 099,937, filed September 11, 1998, incorporated above. In general and for the Vasco technique, subjections such as a layer thickness greater than or equal to zero or all boundaries on the subsurface may be included; other subjections include acceptable ranges in density after M. Cuer and R. Bayer, as described in "Fortram Routines for Linear Inverse Problems", published in Geophysics, vol. 45, N-11 (November 1980), p. 1706 to 1719.

The initial model will be refined with increasing data and may only be the best known estimate at this time. The anticipated or inverse model is then analytically disturbed to determine the relationship between changes in gravity gradient and changes in layer morphology. Indirect methods for disrupting an early or early model include the "Directed Monte Carlo methods of Simulated Annealing and Genetic Algorithms" by M. Sen and PL Stoffa, Elseven. , Amsterdam, 1995).

Thereafter, the actual gravity gradient (from which a representative graphical representation for Γxz and Γxx is presented in FIG. 11) is compared with that predicted by the theoretical model to develop successive repetitions of the model that converge with successively smaller differences. in the comparisons. As additional data measures are taken, as shown in the test sequences of FIG. 10A and 10B, the model is successively refined. Prior to each refinement of the model, the measured data is processed to eliminate known and statistically estimated sources of error including the effects of geological "noise".

With sufficient data measurements, it is possible for the computer data processor to create a computer-displayed animation of time-bound interface motion and morphology to display on a computer monitor or similar display device. The processed data provided by the video produces information about the limit between the expulsion fluid and the hydrocarbon to be recovered. After that, the limit movement and movement rates can be graphically printed or plotted for use by the oil field manager, who can control the injection wells in terms of pressure and quantity to use the morphology. limit to maximize hydrocarbon recovery at the lowest possible secondary recovery costs. The illustration of the hydrocarbon boundaries and their changes can also be used by the oil field manager to identify possible theft zones as suggested in blocks 22 and 23 of FIG. 3. Providing the oilfield manager with this pictorial representation of the evolving situation allows him to apply his expertise to assess, correct, and compensate for changing conditions and to maintain or increase the production of the field.

Although a gradiometer is shown with the preferred instrument for detecting the gravity gradient, other devices capable of measuring or otherwise determining the local gravity gradient are equally suitable. Other devices include pair gravimeters that use falling masses in an evacuated space by accelerating the falling mass measured with a laser beam and high-precision clocks using superconducting sensors.

Industrial Applicability

The present invention advantageously provides a process for optimizing gravity gradiometer measurements to detect subsurface fluid limits in the context of secondary oil recovery by which the cost of recovery can be reduced and / or the amount of recovered hydrocarbon may be increased. The present invention is equally well suited for use in detecting migration of subsurface fluids including, for example, polluted and / or toxic fluid boundaries.

Cross Reference with Provisional Patent Applications

This application claims benefit of filing dates for United States Provisional Patent Applications No. 60 / 085,059 filed on May 12, 1998, No. 60 / 109,138 filed on November 18, 1998, No. 60 / 107,329 filed on , November 6, 1998, H-60 / 107,366 filed November 6, 1998 and N-60 / 099,937 filed September 11, 1998; the disclosures of which are hereby incorporated by reference.

As will be apparent to those skilled in the art, various changes and modifications may be made to the illustrated system and process for optimizing the gravity gradiometer measurements of the present invention without departing from the spirit and scope of the invention as set forth in the present invention. appended claims and their legal equivalents.

Claims (8)

1. A process for improving the measurement of a gravity gradient using an accelerometer gradiometer having at least one pair of accelerometers orbiting an axis of rotation in a common plane, accelerometers sensing changes in the gravity gradient. as they move along their respective orbits and provide a corresponding electrical output from them, CHARACTERIZED for understanding: position the gravity gradiometer at an observation point; tilt the common plane to a first angle below the local horizontal and align the axis to a first azimuth direction; collect data for a selected period of time in the first azimuth direction; tilt the common plane to the first angle above the local horizontal; collect data for a selected period of time in the first azimuth direction; realign the axis to a second azimuth direction offset 90 degrees from the first azimuth direction; collect data for a selected period of time at the first angle above the local horizontal and in the second azimuth direction; tilt the common plane to the first angle below the local horizontal; and collect data for a selected period of time at the first angle below the local horizontal in the second azimuth direction. Process according to claim 1, characterized in that the first angle is less than one degree. Process according to claim 2, characterized in that the first angle is 0.9 degrees. Method according to claim 1, characterized in that it further comprises determining the average of the differences between the measurement sets in the first and second azimuth directions to cancel the instrument-specific trends and gradients. 5. A process for perfecting the measurement of a gravity gradient using an accelerometer-type gradiometer having at least one pair of accelerometers orbiting an axis of rotation in a common plane, accelerometers sensing changes in the gravity gradient at as they move along their respective orbits and provide a corresponding electrical output from them, CHARACTERIZED for understanding: position the gravity gradiometer at an observation point; tilt the common plane to a first angle above the local horizontal and align the axis to a first azimuth direction; collect data for a selected period of time in the first azimuth direction; tilt the common plane to the first angle below the local horizontal; collect data for a selected period of time in the first azimuth direction; realign the axis to a second azimuth direction offset 90 degrees from the first azimuth direction; collect data for a selected period of time at the first angle below the local horizontal and in the second azimuth direction; tilt the common plane to the first angle above the local horizontal; and collect data for a selected period of time at the first angle above the local horizontal in the second azimuth direction. Process according to claim 5, characterized in that the first angle is less than one degree. Process according to claim 6, characterized in that the first angle is 0.9 degrees. Method according to claim 5, characterized in that it further comprises determining the mean of the differences between the measurement sets in the first and second azimuth directions to cancel the instrument-specific trends and gradients.